3 @c Copyright 1991, 1992, 1993, 1994, 1995, 1996, 1997, 1998, 1999, 2000,
4 @c 2001, 2002, 2003 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 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 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 * TI COFF:: ld and the TI COFF
159 * Win32:: ld and WIN32 (cygwin/mingw)
162 * Xtensa:: ld and Xtensa Processors
165 @ifclear SingleFormat
168 @c Following blank line required for remaining bug in makeinfo conds/menus
170 * Reporting Bugs:: Reporting Bugs
171 * MRI:: MRI Compatible Script Files
172 * GNU Free Documentation License:: GNU Free Documentation License
180 @cindex @sc{gnu} linker
181 @cindex what is this?
184 @c man begin SYNOPSIS
185 ld [@b{options}] @var{objfile} @dots{}
189 ar(1), nm(1), objcopy(1), objdump(1), readelf(1) and
190 the Info entries for @file{binutils} and
195 @c man begin DESCRIPTION
197 @command{ld} combines a number of object and archive files, relocates
198 their data and ties up symbol references. Usually the last step in
199 compiling a program is to run @command{ld}.
201 @command{ld} accepts Linker Command Language files written in
202 a superset of AT&T's Link Editor Command Language syntax,
203 to provide explicit and total control over the linking process.
207 This man page does not describe the command language; see the
208 @command{ld} entry in @code{info}, or the manual
209 ld: the GNU linker, for full details on the command language and
210 on other aspects of the GNU linker.
213 @ifclear SingleFormat
214 This version of @command{ld} uses the general purpose BFD libraries
215 to operate on object files. This allows @command{ld} to read, combine, and
216 write object files in many different formats---for example, COFF or
217 @code{a.out}. Different formats may be linked together to produce any
218 available kind of object file. @xref{BFD}, for more information.
221 Aside from its flexibility, the @sc{gnu} linker is more helpful than other
222 linkers in providing diagnostic information. Many linkers abandon
223 execution immediately upon encountering an error; whenever possible,
224 @command{ld} continues executing, allowing you to identify other errors
225 (or, in some cases, to get an output file in spite of the error).
232 @c man begin DESCRIPTION
234 The @sc{gnu} linker @command{ld} is meant to cover a broad range of situations,
235 and to be as compatible as possible with other linkers. As a result,
236 you have many choices to control its behavior.
242 * Options:: Command Line Options
243 * Environment:: Environment Variables
247 @section Command Line Options
255 The linker supports a plethora of command-line options, but in actual
256 practice few of them are used in any particular context.
257 @cindex standard Unix system
258 For instance, a frequent use of @command{ld} is to link standard Unix
259 object files on a standard, supported Unix system. On such a system, to
260 link a file @code{hello.o}:
263 ld -o @var{output} /lib/crt0.o hello.o -lc
266 This tells @command{ld} to produce a file called @var{output} as the
267 result of linking the file @code{/lib/crt0.o} with @code{hello.o} and
268 the library @code{libc.a}, which will come from the standard search
269 directories. (See the discussion of the @samp{-l} option below.)
271 Some of the command-line options to @command{ld} may be specified at any
272 point in the command line. However, options which refer to files, such
273 as @samp{-l} or @samp{-T}, cause the file to be read at the point at
274 which the option appears in the command line, relative to the object
275 files and other file options. Repeating non-file options with a
276 different argument will either have no further effect, or override prior
277 occurrences (those further to the left on the command line) of that
278 option. Options which may be meaningfully specified more than once are
279 noted in the descriptions below.
282 Non-option arguments are object files or archives which are to be linked
283 together. They may follow, precede, or be mixed in with command-line
284 options, except that an object file argument may not be placed between
285 an option and its argument.
287 Usually the linker is invoked with at least one object file, but you can
288 specify other forms of binary input files using @samp{-l}, @samp{-R},
289 and the script command language. If @emph{no} binary input files at all
290 are specified, the linker does not produce any output, and issues the
291 message @samp{No input files}.
293 If the linker cannot recognize the format of an object file, it will
294 assume that it is a linker script. A script specified in this way
295 augments the main linker script used for the link (either the default
296 linker script or the one specified by using @samp{-T}). This feature
297 permits the linker to link against a file which appears to be an object
298 or an archive, but actually merely defines some symbol values, or uses
299 @code{INPUT} or @code{GROUP} to load other objects. Note that
300 specifying a script in this way merely augments the main linker script;
301 use the @samp{-T} option to replace the default linker script entirely.
304 For options whose names are a single letter,
305 option arguments must either follow the option letter without intervening
306 whitespace, or be given as separate arguments immediately following the
307 option that requires them.
309 For options whose names are multiple letters, either one dash or two can
310 precede the option name; for example, @samp{-trace-symbol} and
311 @samp{--trace-symbol} are equivalent. Note---there is one exception to
312 this rule. Multiple letter options that start with a lower case 'o' can
313 only be preceeded by two dashes. This is to reduce confusion with the
314 @samp{-o} option. So for example @samp{-omagic} sets the output file
315 name to @samp{magic} whereas @samp{--omagic} sets the NMAGIC flag on the
318 Arguments to multiple-letter options must either be separated from the
319 option name by an equals sign, or be given as separate arguments
320 immediately following the option that requires them. For example,
321 @samp{--trace-symbol foo} and @samp{--trace-symbol=foo} are equivalent.
322 Unique abbreviations of the names of multiple-letter options are
325 Note---if the linker is being invoked indirectly, via a compiler driver
326 (e.g. @samp{gcc}) then all the linker command line options should be
327 prefixed by @samp{-Wl,} (or whatever is appropriate for the particular
328 compiler driver) like this:
331 gcc -Wl,--startgroup foo.o bar.o -Wl,--endgroup
334 This is important, because otherwise the compiler driver program may
335 silently drop the linker options, resulting in a bad link.
337 Here is a table of the generic command line switches accepted by the GNU
341 @kindex -a@var{keyword}
342 @item -a@var{keyword}
343 This option is supported for HP/UX compatibility. The @var{keyword}
344 argument must be one of the strings @samp{archive}, @samp{shared}, or
345 @samp{default}. @samp{-aarchive} is functionally equivalent to
346 @samp{-Bstatic}, and the other two keywords are functionally equivalent
347 to @samp{-Bdynamic}. This option may be used any number of times.
350 @cindex architectures
352 @item -A@var{architecture}
353 @kindex --architecture=@var{arch}
354 @itemx --architecture=@var{architecture}
355 In the current release of @command{ld}, this option is useful only for the
356 Intel 960 family of architectures. In that @command{ld} configuration, the
357 @var{architecture} argument identifies the particular architecture in
358 the 960 family, enabling some safeguards and modifying the
359 archive-library search path. @xref{i960,,@command{ld} and the Intel 960
360 family}, for details.
362 Future releases of @command{ld} may support similar functionality for
363 other architecture families.
366 @ifclear SingleFormat
367 @cindex binary input format
368 @kindex -b @var{format}
369 @kindex --format=@var{format}
372 @item -b @var{input-format}
373 @itemx --format=@var{input-format}
374 @command{ld} may be configured to support more than one kind of object
375 file. If your @command{ld} is configured this way, you can use the
376 @samp{-b} option to specify the binary format for input object files
377 that follow this option on the command line. Even when @command{ld} is
378 configured to support alternative object formats, you don't usually need
379 to specify this, as @command{ld} should be configured to expect as a
380 default input format the most usual format on each machine.
381 @var{input-format} is a text string, the name of a particular format
382 supported by the BFD libraries. (You can list the available binary
383 formats with @samp{objdump -i}.)
386 You may want to use this option if you are linking files with an unusual
387 binary format. You can also use @samp{-b} to switch formats explicitly (when
388 linking object files of different formats), by including
389 @samp{-b @var{input-format}} before each group of object files in a
392 The default format is taken from the environment variable
397 You can also define the input format from a script, using the command
400 see @ref{Format Commands}.
404 @kindex -c @var{MRI-cmdfile}
405 @kindex --mri-script=@var{MRI-cmdfile}
406 @cindex compatibility, MRI
407 @item -c @var{MRI-commandfile}
408 @itemx --mri-script=@var{MRI-commandfile}
409 For compatibility with linkers produced by MRI, @command{ld} accepts script
410 files written in an alternate, restricted command language, described in
412 @ref{MRI,,MRI Compatible Script Files}.
415 the MRI Compatible Script Files section of GNU ld documentation.
417 Introduce MRI script files with
418 the option @samp{-c}; use the @samp{-T} option to run linker
419 scripts written in the general-purpose @command{ld} scripting language.
420 If @var{MRI-cmdfile} does not exist, @command{ld} looks for it in the directories
421 specified by any @samp{-L} options.
423 @cindex common allocation
430 These three options are equivalent; multiple forms are supported for
431 compatibility with other linkers. They assign space to common symbols
432 even if a relocatable output file is specified (with @samp{-r}). The
433 script command @code{FORCE_COMMON_ALLOCATION} has the same effect.
434 @xref{Miscellaneous Commands}.
436 @cindex entry point, from command line
437 @kindex -e @var{entry}
438 @kindex --entry=@var{entry}
440 @itemx --entry=@var{entry}
441 Use @var{entry} as the explicit symbol for beginning execution of your
442 program, rather than the default entry point. If there is no symbol
443 named @var{entry}, the linker will try to parse @var{entry} as a number,
444 and use that as the entry address (the number will be interpreted in
445 base 10; you may use a leading @samp{0x} for base 16, or a leading
446 @samp{0} for base 8). @xref{Entry Point}, for a discussion of defaults
447 and other ways of specifying the entry point.
449 @cindex dynamic symbol table
451 @kindex --export-dynamic
453 @itemx --export-dynamic
454 When creating a dynamically linked executable, add all symbols to the
455 dynamic symbol table. The dynamic symbol table is the set of symbols
456 which are visible from dynamic objects at run time.
458 If you do not use this option, the dynamic symbol table will normally
459 contain only those symbols which are referenced by some dynamic object
460 mentioned in the link.
462 If you use @code{dlopen} to load a dynamic object which needs to refer
463 back to the symbols defined by the program, rather than some other
464 dynamic object, then you will probably need to use this option when
465 linking the program itself.
467 You can also use the version script to control what symbols should
468 be added to the dynamic symbol table if the output format supports it.
469 See the description of @samp{--version-script} in @ref{VERSION}.
471 @ifclear SingleFormat
472 @cindex big-endian objects
476 Link big-endian objects. This affects the default output format.
478 @cindex little-endian objects
481 Link little-endian objects. This affects the default output format.
487 @itemx --auxiliary @var{name}
488 When creating an ELF shared object, set the internal DT_AUXILIARY field
489 to the specified name. This tells the dynamic linker that the symbol
490 table of the shared object should be used as an auxiliary filter on the
491 symbol table of the shared object @var{name}.
493 If you later link a program against this filter object, then, when you
494 run the program, the dynamic linker will see the DT_AUXILIARY field. If
495 the dynamic linker resolves any symbols from the filter object, it will
496 first check whether there is a definition in the shared object
497 @var{name}. If there is one, it will be used instead of the definition
498 in the filter object. The shared object @var{name} need not exist.
499 Thus the shared object @var{name} may be used to provide an alternative
500 implementation of certain functions, perhaps for debugging or for
501 machine specific performance.
503 This option may be specified more than once. The DT_AUXILIARY entries
504 will be created in the order in which they appear on the command line.
509 @itemx --filter @var{name}
510 When creating an ELF shared object, set the internal DT_FILTER field to
511 the specified name. This tells the dynamic linker that the symbol table
512 of the shared object which is being created should be used as a filter
513 on the symbol table of the shared object @var{name}.
515 If you later link a program against this filter object, then, when you
516 run the program, the dynamic linker will see the DT_FILTER field. The
517 dynamic linker will resolve symbols according to the symbol table of the
518 filter object as usual, but it will actually link to the definitions
519 found in the shared object @var{name}. Thus the filter object can be
520 used to select a subset of the symbols provided by the object
523 Some older linkers used the @option{-F} option throughout a compilation
524 toolchain for specifying object-file format for both input and output
526 @ifclear SingleFormat
527 The @sc{gnu} linker uses other mechanisms for this purpose: the
528 @option{-b}, @option{--format}, @option{--oformat} options, the
529 @code{TARGET} command in linker scripts, and the @code{GNUTARGET}
530 environment variable.
532 The @sc{gnu} linker will ignore the @option{-F} option when not
533 creating an ELF shared object.
535 @cindex finalization function
537 @item -fini @var{name}
538 When creating an ELF executable or shared object, call NAME when the
539 executable or shared object is unloaded, by setting DT_FINI to the
540 address of the function. By default, the linker uses @code{_fini} as
541 the function to call.
545 Ignored. Provided for compatibility with other tools.
551 @itemx --gpsize=@var{value}
552 Set the maximum size of objects to be optimized using the GP register to
553 @var{size}. This is only meaningful for object file formats such as
554 MIPS ECOFF which supports putting large and small objects into different
555 sections. This is ignored for other object file formats.
557 @cindex runtime library name
559 @kindex -soname=@var{name}
561 @itemx -soname=@var{name}
562 When creating an ELF shared object, set the internal DT_SONAME field to
563 the specified name. When an executable is linked with a shared object
564 which has a DT_SONAME field, then when the executable is run the dynamic
565 linker will attempt to load the shared object specified by the DT_SONAME
566 field rather than the using the file name given to the linker.
569 @cindex incremental link
571 Perform an incremental link (same as option @samp{-r}).
573 @cindex initialization function
575 @item -init @var{name}
576 When creating an ELF executable or shared object, call NAME when the
577 executable or shared object is loaded, by setting DT_INIT to the address
578 of the function. By default, the linker uses @code{_init} as the
581 @cindex archive files, from cmd line
582 @kindex -l@var{archive}
583 @kindex --library=@var{archive}
584 @item -l@var{archive}
585 @itemx --library=@var{archive}
586 Add archive file @var{archive} to the list of files to link. This
587 option may be used any number of times. @command{ld} will search its
588 path-list for occurrences of @code{lib@var{archive}.a} for every
589 @var{archive} specified.
591 On systems which support shared libraries, @command{ld} may also search for
592 libraries with extensions other than @code{.a}. Specifically, on ELF
593 and SunOS systems, @command{ld} will search a directory for a library with
594 an extension of @code{.so} before searching for one with an extension of
595 @code{.a}. By convention, a @code{.so} extension indicates a shared
598 The linker will search an archive only once, at the location where it is
599 specified on the command line. If the archive defines a symbol which
600 was undefined in some object which appeared before the archive on the
601 command line, the linker will include the appropriate file(s) from the
602 archive. However, an undefined symbol in an object appearing later on
603 the command line will not cause the linker to search the archive again.
605 See the @option{-(} option for a way to force the linker to search
606 archives multiple times.
608 You may list the same archive multiple times on the command line.
611 This type of archive searching is standard for Unix linkers. However,
612 if you are using @command{ld} on AIX, note that it is different from the
613 behaviour of the AIX linker.
616 @cindex search directory, from cmd line
618 @kindex --library-path=@var{dir}
619 @item -L@var{searchdir}
620 @itemx --library-path=@var{searchdir}
621 Add path @var{searchdir} to the list of paths that @command{ld} will search
622 for archive libraries and @command{ld} control scripts. You may use this
623 option any number of times. The directories are searched in the order
624 in which they are specified on the command line. Directories specified
625 on the command line are searched before the default directories. All
626 @option{-L} options apply to all @option{-l} options, regardless of the
627 order in which the options appear.
629 If @var{searchdir} begins with @code{=}, then the @code{=} will be replaced
630 by the @dfn{sysroot prefix}, a path specified when the linker is configured.
633 The default set of paths searched (without being specified with
634 @samp{-L}) depends on which emulation mode @command{ld} is using, and in
635 some cases also on how it was configured. @xref{Environment}.
638 The paths can also be specified in a link script with the
639 @code{SEARCH_DIR} command. Directories specified this way are searched
640 at the point in which the linker script appears in the command line.
643 @kindex -m @var{emulation}
644 @item -m@var{emulation}
645 Emulate the @var{emulation} linker. You can list the available
646 emulations with the @samp{--verbose} or @samp{-V} options.
648 If the @samp{-m} option is not used, the emulation is taken from the
649 @code{LDEMULATION} environment variable, if that is defined.
651 Otherwise, the default emulation depends upon how the linker was
659 Print a link map to the standard output. A link map provides
660 information about the link, including the following:
664 Where object files and symbols are mapped into memory.
666 How common symbols are allocated.
668 All archive members included in the link, with a mention of the symbol
669 which caused the archive member to be brought in.
673 @cindex read-only text
678 Turn off page alignment of sections, and mark the output as
679 @code{NMAGIC} if possible.
683 @cindex read/write from cmd line
687 Set the text and data sections to be readable and writable. Also, do
688 not page-align the data segment, and disable linking against shared
689 libraries. If the output format supports Unix style magic numbers,
690 mark the output as @code{OMAGIC}.
695 This option negates most of the effects of the @option{-N} option. It
696 sets the text section to be read-only, and forces the data segment to
697 be page-aligned. Note - this option does not enable linking against
698 shared libraries. Use @option{-Bdynamic} for this.
700 @kindex -o @var{output}
701 @kindex --output=@var{output}
702 @cindex naming the output file
703 @item -o @var{output}
704 @itemx --output=@var{output}
705 Use @var{output} as the name for the program produced by @command{ld}; if this
706 option is not specified, the name @file{a.out} is used by default. The
707 script command @code{OUTPUT} can also specify the output file name.
709 @kindex -O @var{level}
710 @cindex generating optimized output
712 If @var{level} is a numeric values greater than zero @command{ld} optimizes
713 the output. This might take significantly longer and therefore probably
714 should only be enabled for the final binary.
717 @kindex --emit-relocs
718 @cindex retain relocations in final executable
721 Leave relocation sections and contents in fully linked exececutables.
722 Post link analysis and optimization tools may need this information in
723 order to perform correct modifications of executables. This results
724 in larger executables.
726 This option is currently only supported on ELF platforms.
729 @cindex relocatable output
731 @kindex --relocatable
734 Generate relocatable output---i.e., generate an output file that can in
735 turn serve as input to @command{ld}. This is often called @dfn{partial
736 linking}. As a side effect, in environments that support standard Unix
737 magic numbers, this option also sets the output file's magic number to
739 @c ; see @option{-N}.
740 If this option is not specified, an absolute file is produced. When
741 linking C++ programs, this option @emph{will not} resolve references to
742 constructors; to do that, use @samp{-Ur}.
744 When an input file does not have the same format as the output file,
745 partial linking is only supported if that input file does not contain any
746 relocations. Different output formats can have further restrictions; for
747 example some @code{a.out}-based formats do not support partial linking
748 with input files in other formats at all.
750 This option does the same thing as @samp{-i}.
752 @kindex -R @var{file}
753 @kindex --just-symbols=@var{file}
754 @cindex symbol-only input
755 @item -R @var{filename}
756 @itemx --just-symbols=@var{filename}
757 Read symbol names and their addresses from @var{filename}, but do not
758 relocate it or include it in the output. This allows your output file
759 to refer symbolically to absolute locations of memory defined in other
760 programs. You may use this option more than once.
762 For compatibility with other ELF linkers, if the @option{-R} option is
763 followed by a directory name, rather than a file name, it is treated as
764 the @option{-rpath} option.
768 @cindex strip all symbols
771 Omit all symbol information from the output file.
774 @kindex --strip-debug
775 @cindex strip debugger symbols
778 Omit debugger symbol information (but not all symbols) from the output file.
782 @cindex input files, displaying
785 Print the names of the input files as @command{ld} processes them.
787 @kindex -T @var{script}
788 @kindex --script=@var{script}
790 @item -T @var{scriptfile}
791 @itemx --script=@var{scriptfile}
792 Use @var{scriptfile} as the linker script. This script replaces
793 @command{ld}'s default linker script (rather than adding to it), so
794 @var{commandfile} must specify everything necessary to describe the
795 output file. @xref{Scripts}. If @var{scriptfile} does not exist in
796 the current directory, @code{ld} looks for it in the directories
797 specified by any preceding @samp{-L} options. Multiple @samp{-T}
800 @kindex -u @var{symbol}
801 @kindex --undefined=@var{symbol}
802 @cindex undefined symbol
803 @item -u @var{symbol}
804 @itemx --undefined=@var{symbol}
805 Force @var{symbol} to be entered in the output file as an undefined
806 symbol. Doing this may, for example, trigger linking of additional
807 modules from standard libraries. @samp{-u} may be repeated with
808 different option arguments to enter additional undefined symbols. This
809 option is equivalent to the @code{EXTERN} linker script command.
814 For anything other than C++ programs, this option is equivalent to
815 @samp{-r}: it generates relocatable output---i.e., an output file that can in
816 turn serve as input to @command{ld}. When linking C++ programs, @samp{-Ur}
817 @emph{does} resolve references to constructors, unlike @samp{-r}.
818 It does not work to use @samp{-Ur} on files that were themselves linked
819 with @samp{-Ur}; once the constructor table has been built, it cannot
820 be added to. Use @samp{-Ur} only for the last partial link, and
821 @samp{-r} for the others.
823 @kindex --unique[=@var{SECTION}]
824 @item --unique[=@var{SECTION}]
825 Creates a separate output section for every input section matching
826 @var{SECTION}, or if the optional wildcard @var{SECTION} argument is
827 missing, for every orphan input section. An orphan section is one not
828 specifically mentioned in a linker script. You may use this option
829 multiple times on the command line; It prevents the normal merging of
830 input sections with the same name, overriding output section assignments
840 Display the version number for @command{ld}. The @option{-V} option also
841 lists the supported emulations.
844 @kindex --discard-all
845 @cindex deleting local symbols
848 Delete all local symbols.
851 @kindex --discard-locals
852 @cindex local symbols, deleting
853 @cindex L, deleting symbols beginning
855 @itemx --discard-locals
856 Delete all temporary local symbols. For most targets, this is all local
857 symbols whose names begin with @samp{L}.
859 @kindex -y @var{symbol}
860 @kindex --trace-symbol=@var{symbol}
861 @cindex symbol tracing
862 @item -y @var{symbol}
863 @itemx --trace-symbol=@var{symbol}
864 Print the name of each linked file in which @var{symbol} appears. This
865 option may be given any number of times. On many systems it is necessary
866 to prepend an underscore.
868 This option is useful when you have an undefined symbol in your link but
869 don't know where the reference is coming from.
871 @kindex -Y @var{path}
873 Add @var{path} to the default library search path. This option exists
874 for Solaris compatibility.
876 @kindex -z @var{keyword}
877 @item -z @var{keyword}
878 The recognized keywords are:
882 Combines multiple reloc sections and sorts them to make dynamic symbol
883 lookup caching possible.
886 Disallows undefined symbols in object files. Undefined symbols in
887 shared libraries are still allowed.
890 This option is only meaningful when building a shared object.
891 It marks the object so that its runtime initialization will occur
892 before the runtime initialization of any other objects brought into
893 the process at the same time. Similarly the runtime finalization of
894 the object will occur after the runtime finalization of any other
898 Marks the object that its symbol table interposes before all symbols
899 but the primary executable.
902 Marks the object that its filters be processed immediately at
906 Allows multiple definitions.
909 Disables multiple reloc sections combining.
912 Disables production of copy relocs.
915 Marks the object that the search for dependencies of this object will
916 ignore any default library search paths.
919 Marks the object shouldn't be unloaded at runtime.
922 Marks the object not available to @code{dlopen}.
925 Marks the object can not be dumped by @code{dldump}.
928 When generating an executable or shared library, mark it to tell the
929 dynamic linker to resolve all symbols when the program is started, or
930 when the shared library is linked to using dlopen, instead of
931 deferring function call resolution to the point when the function is
935 Marks the object may contain $ORIGIN.
939 Other keywords are ignored for Solaris compatibility.
942 @cindex groups of archives
943 @item -( @var{archives} -)
944 @itemx --start-group @var{archives} --end-group
945 The @var{archives} should be a list of archive files. They may be
946 either explicit file names, or @samp{-l} options.
948 The specified archives are searched repeatedly until no new undefined
949 references are created. Normally, an archive is searched only once in
950 the order that it is specified on the command line. If a symbol in that
951 archive is needed to resolve an undefined symbol referred to by an
952 object in an archive that appears later on the command line, the linker
953 would not be able to resolve that reference. By grouping the archives,
954 they all be searched repeatedly until all possible references are
957 Using this option has a significant performance cost. It is best to use
958 it only when there are unavoidable circular references between two or
961 @kindex --accept-unknown-input-arch
962 @kindex --no-accept-unknown-input-arch
963 @item --accept-unknown-input-arch
964 @itemx --no-accept-unknown-input-arch
965 Tells the linker to accept input files whose architecture cannot be
966 recognised. The assumption is that the user knows what they are doing
967 and deliberately wants to link in these unknown input files. This was
968 the default behaviour of the linker, before release 2.14. The default
969 behaviour from release 2.14 onwards is to reject such input files, and
970 so the @samp{--accept-unknown-input-arch} option has been added to
971 restore the old behaviour.
973 @kindex -assert @var{keyword}
974 @item -assert @var{keyword}
975 This option is ignored for SunOS compatibility.
983 Link against dynamic libraries. This is only meaningful on platforms
984 for which shared libraries are supported. This option is normally the
985 default on such platforms. The different variants of this option are
986 for compatibility with various systems. You may use this option
987 multiple times on the command line: it affects library searching for
988 @option{-l} options which follow it. This
989 option also implies @option{--unresolved-symbols=ignore-all}.
993 Set the @code{DF_1_GROUP} flag in the @code{DT_FLAGS_1} entry in the dynamic
994 section. This causes the runtime linker to handle lookups in this
995 object and its dependencies to be performed only inside the group.
996 @option{--unresolved-symbols=report-all} is implied. This option is
997 only meaningful on ELF platforms which support shared libraries.
1007 Do not link against shared libraries. This is only meaningful on
1008 platforms for which shared libraries are supported. The different
1009 variants of this option are for compatibility with various systems. You
1010 may use this option multiple times on the command line: it affects
1011 library searching for @option{-l} options which follow it. This
1012 option also implies @option{--unresolved-symbols=report-all}.
1016 When creating a shared library, bind references to global symbols to the
1017 definition within the shared library, if any. Normally, it is possible
1018 for a program linked against a shared library to override the definition
1019 within the shared library. This option is only meaningful on ELF
1020 platforms which support shared libraries.
1022 @kindex --check-sections
1023 @kindex --no-check-sections
1024 @item --check-sections
1025 @itemx --no-check-sections
1026 Asks the linker @emph{not} to check section addresses after they have
1027 been assigned to see if there any overlaps. Normally the linker will
1028 perform this check, and if it finds any overlaps it will produce
1029 suitable error messages. The linker does know about, and does make
1030 allowances for sections in overlays. The default behaviour can be
1031 restored by using the command line switch @option{--check-sections}.
1033 @cindex cross reference table
1036 Output a cross reference table. If a linker map file is being
1037 generated, the cross reference table is printed to the map file.
1038 Otherwise, it is printed on the standard output.
1040 The format of the table is intentionally simple, so that it may be
1041 easily processed by a script if necessary. The symbols are printed out,
1042 sorted by name. For each symbol, a list of file names is given. If the
1043 symbol is defined, the first file listed is the location of the
1044 definition. The remaining files contain references to the symbol.
1046 @cindex common allocation
1047 @kindex --no-define-common
1048 @item --no-define-common
1049 This option inhibits the assignment of addresses to common symbols.
1050 The script command @code{INHIBIT_COMMON_ALLOCATION} has the same effect.
1051 @xref{Miscellaneous Commands}.
1053 The @samp{--no-define-common} option allows decoupling
1054 the decision to assign addresses to Common symbols from the choice
1055 of the output file type; otherwise a non-Relocatable output type
1056 forces assigning addresses to Common symbols.
1057 Using @samp{--no-define-common} allows Common symbols that are referenced
1058 from a shared library to be assigned addresses only in the main program.
1059 This eliminates the unused duplicate space in the shared library,
1060 and also prevents any possible confusion over resolving to the wrong
1061 duplicate when there are many dynamic modules with specialized search
1062 paths for runtime symbol resolution.
1064 @cindex symbols, from command line
1065 @kindex --defsym @var{symbol}=@var{exp}
1066 @item --defsym @var{symbol}=@var{expression}
1067 Create a global symbol in the output file, containing the absolute
1068 address given by @var{expression}. You may use this option as many
1069 times as necessary to define multiple symbols in the command line. A
1070 limited form of arithmetic is supported for the @var{expression} in this
1071 context: you may give a hexadecimal constant or the name of an existing
1072 symbol, or use @code{+} and @code{-} to add or subtract hexadecimal
1073 constants or symbols. If you need more elaborate expressions, consider
1074 using the linker command language from a script (@pxref{Assignments,,
1075 Assignment: Symbol Definitions}). @emph{Note:} there should be no white
1076 space between @var{symbol}, the equals sign (``@key{=}''), and
1079 @cindex demangling, from command line
1080 @kindex --demangle[=@var{style}]
1081 @kindex --no-demangle
1082 @item --demangle[=@var{style}]
1083 @itemx --no-demangle
1084 These options control whether to demangle symbol names in error messages
1085 and other output. When the linker is told to demangle, it tries to
1086 present symbol names in a readable fashion: it strips leading
1087 underscores if they are used by the object file format, and converts C++
1088 mangled symbol names into user readable names. Different compilers have
1089 different mangling styles. The optional demangling style argument can be used
1090 to choose an appropriate demangling style for your compiler. The linker will
1091 demangle by default unless the environment variable @samp{COLLECT_NO_DEMANGLE}
1092 is set. These options may be used to override the default.
1094 @cindex dynamic linker, from command line
1095 @kindex -I@var{file}
1096 @kindex --dynamic-linker @var{file}
1097 @item --dynamic-linker @var{file}
1098 Set the name of the dynamic linker. This is only meaningful when
1099 generating dynamically linked ELF executables. The default dynamic
1100 linker is normally correct; don't use this unless you know what you are
1103 @cindex MIPS embedded PIC code
1104 @kindex --embedded-relocs
1105 @item --embedded-relocs
1106 This option is only meaningful when linking MIPS embedded PIC code,
1107 generated by the -membedded-pic option to the @sc{gnu} compiler and
1108 assembler. It causes the linker to create a table which may be used at
1109 runtime to relocate any data which was statically initialized to pointer
1110 values. See the code in testsuite/ld-empic for details.
1113 @kindex --fatal-warnings
1114 @item --fatal-warnings
1115 Treat all warnings as errors.
1117 @kindex --force-exe-suffix
1118 @item --force-exe-suffix
1119 Make sure that an output file has a .exe suffix.
1121 If a successfully built fully linked output file does not have a
1122 @code{.exe} or @code{.dll} suffix, this option forces the linker to copy
1123 the output file to one of the same name with a @code{.exe} suffix. This
1124 option is useful when using unmodified Unix makefiles on a Microsoft
1125 Windows host, since some versions of Windows won't run an image unless
1126 it ends in a @code{.exe} suffix.
1128 @kindex --gc-sections
1129 @kindex --no-gc-sections
1130 @cindex garbage collection
1131 @item --no-gc-sections
1132 @itemx --gc-sections
1133 Enable garbage collection of unused input sections. It is ignored on
1134 targets that do not support this option. This option is not compatible
1135 with @samp{-r}, nor should it be used with dynamic linking. The default
1136 behaviour (of not performing this garbage collection) can be restored by
1137 specifying @samp{--no-gc-sections} on the command line.
1143 Print a summary of the command-line options on the standard output and exit.
1145 @kindex --target-help
1147 Print a summary of all target specific options on the standard output and exit.
1150 @item -Map @var{mapfile}
1151 Print a link map to the file @var{mapfile}. See the description of the
1152 @option{-M} option, above.
1154 @cindex memory usage
1155 @kindex --no-keep-memory
1156 @item --no-keep-memory
1157 @command{ld} normally optimizes for speed over memory usage by caching the
1158 symbol tables of input files in memory. This option tells @command{ld} to
1159 instead optimize for memory usage, by rereading the symbol tables as
1160 necessary. This may be required if @command{ld} runs out of memory space
1161 while linking a large executable.
1163 @kindex --no-undefined
1165 @item --no-undefined
1167 Report unresolved symbol references from regular object files. This
1168 is done even if the linker is creating a non-symbolic shared library.
1169 The switch @option{--[no-]allow-shlib-undefined} controls the
1170 behaviour for reporting unresolved references found in shared
1171 libraries being linked in.
1173 @kindex --allow-multiple-definition
1175 @item --allow-multiple-definition
1177 Normally when a symbol is defined multiple times, the linker will
1178 report a fatal error. These options allow multiple definitions and the
1179 first definition will be used.
1181 @kindex --allow-shlib-undefined
1182 @kindex --no-allow-shlib-undefined
1183 @item --allow-shlib-undefined
1184 @itemx --no-allow-shlib-undefined
1185 Allows (the default) or disallows undefined symbols in shared libraries.
1186 This switch is similar to @option{--no-undefined} except that it
1187 determines the behaviour when the undefined symbols are in a
1188 shared library rather than a regular object file. It does not affect
1189 how undefined symbols in regular object files are handled.
1191 The reason that @option{--allow-shlib-undefined} is the default is that
1192 the shared library being specified at link time may not be the same as
1193 the one that is available at load time, so the symbols might actually be
1194 resolvable at load time. Plus there are some systems, (eg BeOS) where
1195 undefined symbols in shared libraries is normal. (The kernel patches
1196 them at load time to select which function is most appropriate
1197 for the current architecture. This is used for example to dynamically
1198 select an appropriate memset function). Apparently it is also normal
1199 for HPPA shared libraries to have undefined symbols.
1201 @kindex --no-undefined-version
1202 @item --no-undefined-version
1203 Normally when a symbol has an undefined version, the linker will ignore
1204 it. This option disallows symbols with undefined version and a fatal error
1205 will be issued instead.
1207 @kindex --no-warn-mismatch
1208 @item --no-warn-mismatch
1209 Normally @command{ld} will give an error if you try to link together input
1210 files that are mismatched for some reason, perhaps because they have
1211 been compiled for different processors or for different endiannesses.
1212 This option tells @command{ld} that it should silently permit such possible
1213 errors. This option should only be used with care, in cases when you
1214 have taken some special action that ensures that the linker errors are
1217 @kindex --no-whole-archive
1218 @item --no-whole-archive
1219 Turn off the effect of the @option{--whole-archive} option for subsequent
1222 @cindex output file after errors
1223 @kindex --noinhibit-exec
1224 @item --noinhibit-exec
1225 Retain the executable output file whenever it is still usable.
1226 Normally, the linker will not produce an output file if it encounters
1227 errors during the link process; it exits without writing an output file
1228 when it issues any error whatsoever.
1232 Only search library directories explicitly specified on the
1233 command line. Library directories specified in linker scripts
1234 (including linker scripts specified on the command line) are ignored.
1236 @ifclear SingleFormat
1238 @item --oformat @var{output-format}
1239 @command{ld} may be configured to support more than one kind of object
1240 file. If your @command{ld} is configured this way, you can use the
1241 @samp{--oformat} option to specify the binary format for the output
1242 object file. Even when @command{ld} is configured to support alternative
1243 object formats, you don't usually need to specify this, as @command{ld}
1244 should be configured to produce as a default output format the most
1245 usual format on each machine. @var{output-format} is a text string, the
1246 name of a particular format supported by the BFD libraries. (You can
1247 list the available binary formats with @samp{objdump -i}.) The script
1248 command @code{OUTPUT_FORMAT} can also specify the output format, but
1249 this option overrides it. @xref{BFD}.
1253 @kindex --pic-executable
1255 @itemx --pic-executable
1256 @cindex position independent executables
1257 Create a position independent executable. This is currently only supported on
1258 ELF platforms. Position independent executables are similar to shared
1259 libraries in that they are relocated by the dynamic linker to the virtual
1260 address the OS chooses for them (which can vary between invocations). Like
1261 normal dynamically linked executables they can be executed and symbols
1262 defined in the executable cannot be overridden by shared libraries.
1266 This option is ignored for Linux compatibility.
1270 This option is ignored for SVR4 compatibility.
1273 @cindex synthesizing linker
1274 @cindex relaxing addressing modes
1276 An option with machine dependent effects.
1278 This option is only supported on a few targets.
1281 @xref{H8/300,,@command{ld} and the H8/300}.
1284 @xref{i960,, @command{ld} and the Intel 960 family}.
1287 @xref{Xtensa,, @command{ld} and Xtensa Processors}.
1290 On some platforms, the @samp{--relax} option performs global
1291 optimizations that become possible when the linker resolves addressing
1292 in the program, such as relaxing address modes and synthesizing new
1293 instructions in the output object file.
1295 On some platforms these link time global optimizations may make symbolic
1296 debugging of the resulting executable impossible.
1299 the case for the Matsushita MN10200 and MN10300 family of processors.
1303 On platforms where this is not supported, @samp{--relax} is accepted,
1307 @cindex retaining specified symbols
1308 @cindex stripping all but some symbols
1309 @cindex symbols, retaining selectively
1310 @item --retain-symbols-file @var{filename}
1311 Retain @emph{only} the symbols listed in the file @var{filename},
1312 discarding all others. @var{filename} is simply a flat file, with one
1313 symbol name per line. This option is especially useful in environments
1317 where a large global symbol table is accumulated gradually, to conserve
1320 @samp{--retain-symbols-file} does @emph{not} discard undefined symbols,
1321 or symbols needed for relocations.
1323 You may only specify @samp{--retain-symbols-file} once in the command
1324 line. It overrides @samp{-s} and @samp{-S}.
1327 @item -rpath @var{dir}
1328 @cindex runtime library search path
1330 Add a directory to the runtime library search path. This is used when
1331 linking an ELF executable with shared objects. All @option{-rpath}
1332 arguments are concatenated and passed to the runtime linker, which uses
1333 them to locate shared objects at runtime. The @option{-rpath} option is
1334 also used when locating shared objects which are needed by shared
1335 objects explicitly included in the link; see the description of the
1336 @option{-rpath-link} option. If @option{-rpath} is not used when linking an
1337 ELF executable, the contents of the environment variable
1338 @code{LD_RUN_PATH} will be used if it is defined.
1340 The @option{-rpath} option may also be used on SunOS. By default, on
1341 SunOS, the linker will form a runtime search patch out of all the
1342 @option{-L} options it is given. If a @option{-rpath} option is used, the
1343 runtime search path will be formed exclusively using the @option{-rpath}
1344 options, ignoring the @option{-L} options. This can be useful when using
1345 gcc, which adds many @option{-L} options which may be on NFS mounted
1348 For compatibility with other ELF linkers, if the @option{-R} option is
1349 followed by a directory name, rather than a file name, it is treated as
1350 the @option{-rpath} option.
1354 @cindex link-time runtime library search path
1356 @item -rpath-link @var{DIR}
1357 When using ELF or SunOS, one shared library may require another. This
1358 happens when an @code{ld -shared} link includes a shared library as one
1361 When the linker encounters such a dependency when doing a non-shared,
1362 non-relocatable link, it will automatically try to locate the required
1363 shared library and include it in the link, if it is not included
1364 explicitly. In such a case, the @option{-rpath-link} option
1365 specifies the first set of directories to search. The
1366 @option{-rpath-link} option may specify a sequence of directory names
1367 either by specifying a list of names separated by colons, or by
1368 appearing multiple times.
1370 This option should be used with caution as it overrides the search path
1371 that may have been hard compiled into a shared library. In such a case it
1372 is possible to use unintentionally a different search path than the
1373 runtime linker would do.
1375 The linker uses the following search paths to locate required shared
1379 Any directories specified by @option{-rpath-link} options.
1381 Any directories specified by @option{-rpath} options. The difference
1382 between @option{-rpath} and @option{-rpath-link} is that directories
1383 specified by @option{-rpath} options are included in the executable and
1384 used at runtime, whereas the @option{-rpath-link} option is only effective
1385 at link time. It is for the native linker only.
1387 On an ELF system, if the @option{-rpath} and @code{rpath-link} options
1388 were not used, search the contents of the environment variable
1389 @code{LD_RUN_PATH}. It is for the native linker only.
1391 On SunOS, if the @option{-rpath} option was not used, search any
1392 directories specified using @option{-L} options.
1394 For a native linker, the contents of the environment variable
1395 @code{LD_LIBRARY_PATH}.
1397 For a native ELF linker, the directories in @code{DT_RUNPATH} or
1398 @code{DT_RPATH} of a shared library are searched for shared
1399 libraries needed by it. The @code{DT_RPATH} entries are ignored if
1400 @code{DT_RUNPATH} entries exist.
1402 The default directories, normally @file{/lib} and @file{/usr/lib}.
1404 For a native linker on an ELF system, if the file @file{/etc/ld.so.conf}
1405 exists, the list of directories found in that file.
1408 If the required shared library is not found, the linker will issue a
1409 warning and continue with the link.
1416 @cindex shared libraries
1417 Create a shared library. This is currently only supported on ELF, XCOFF
1418 and SunOS platforms. On SunOS, the linker will automatically create a
1419 shared library if the @option{-e} option is not used and there are
1420 undefined symbols in the link.
1423 @kindex --sort-common
1424 This option tells @command{ld} to sort the common symbols by size when it
1425 places them in the appropriate output sections. First come all the one
1426 byte symbols, then all the two byte, then all the four byte, and then
1427 everything else. This is to prevent gaps between symbols due to
1428 alignment constraints.
1430 @kindex --split-by-file
1431 @item --split-by-file [@var{size}]
1432 Similar to @option{--split-by-reloc} but creates a new output section for
1433 each input file when @var{size} is reached. @var{size} defaults to a
1434 size of 1 if not given.
1436 @kindex --split-by-reloc
1437 @item --split-by-reloc [@var{count}]
1438 Tries to creates extra sections in the output file so that no single
1439 output section in the file contains more than @var{count} relocations.
1440 This is useful when generating huge relocatable files for downloading into
1441 certain real time kernels with the COFF object file format; since COFF
1442 cannot represent more than 65535 relocations in a single section. Note
1443 that this will fail to work with object file formats which do not
1444 support arbitrary sections. The linker will not split up individual
1445 input sections for redistribution, so if a single input section contains
1446 more than @var{count} relocations one output section will contain that
1447 many relocations. @var{count} defaults to a value of 32768.
1451 Compute and display statistics about the operation of the linker, such
1452 as execution time and memory usage.
1454 @kindex --traditional-format
1455 @cindex traditional format
1456 @item --traditional-format
1457 For some targets, the output of @command{ld} is different in some ways from
1458 the output of some existing linker. This switch requests @command{ld} to
1459 use the traditional format instead.
1462 For example, on SunOS, @command{ld} combines duplicate entries in the
1463 symbol string table. This can reduce the size of an output file with
1464 full debugging information by over 30 percent. Unfortunately, the SunOS
1465 @code{dbx} program can not read the resulting program (@code{gdb} has no
1466 trouble). The @samp{--traditional-format} switch tells @command{ld} to not
1467 combine duplicate entries.
1469 @kindex --section-start @var{sectionname}=@var{org}
1470 @item --section-start @var{sectionname}=@var{org}
1471 Locate a section in the output file at the absolute
1472 address given by @var{org}. You may use this option as many
1473 times as necessary to locate multiple sections in the command
1475 @var{org} must be a single hexadecimal integer;
1476 for compatibility with other linkers, you may omit the leading
1477 @samp{0x} usually associated with hexadecimal values. @emph{Note:} there
1478 should be no white space between @var{sectionname}, the equals
1479 sign (``@key{=}''), and @var{org}.
1481 @kindex -Tbss @var{org}
1482 @kindex -Tdata @var{org}
1483 @kindex -Ttext @var{org}
1484 @cindex segment origins, cmd line
1485 @item -Tbss @var{org}
1486 @itemx -Tdata @var{org}
1487 @itemx -Ttext @var{org}
1488 Same as --section-start, with @code{.bss}, @code{.data} or
1489 @code{.text} as the @var{sectionname}.
1491 @kindex --unresolved-symbols
1492 @item --unresolved-symbols=@var{method}
1493 Determine how to handle unresolved symbols. There are four possible
1494 values for @samp{method}:
1498 Do not report any unresolved symbols. This is the default when
1499 creating shared libraries or dynamic executables.
1502 Report all unresolved symbols. This is the default when creating
1505 @item ignore-in-object-files
1506 Report unresolved symbols that are contained in shared libraries, but
1507 ignore them if they come from regular object files.
1509 @item ignore-in-shared-libs
1510 Report unresolved symbols that come from regular object files, but
1511 ignore them if they come from shared libraries. This can be useful
1512 when creating a dynamic binary and it is known that all the shared
1513 libraries that it should be referencing are included on the linker's
1517 The behaviour for shared libraries on their own can also be controlled
1518 by the @option{--[no-]allow-shlib-undefined} option.
1520 Normally the linker will generate an error message for each reported
1521 unresolved symbol but the option @option{--warn-unresolved-symbols}
1522 can change this to a warning.
1528 Display the version number for @command{ld} and list the linker emulations
1529 supported. Display which input files can and cannot be opened. Display
1530 the linker script being used by the linker.
1532 @kindex --version-script=@var{version-scriptfile}
1533 @cindex version script, symbol versions
1534 @itemx --version-script=@var{version-scriptfile}
1535 Specify the name of a version script to the linker. This is typically
1536 used when creating shared libraries to specify additional information
1537 about the version hierarchy for the library being created. This option
1538 is only meaningful on ELF platforms which support shared libraries.
1541 @kindex --warn-common
1542 @cindex warnings, on combining symbols
1543 @cindex combining symbols, warnings on
1545 Warn when a common symbol is combined with another common symbol or with
1546 a symbol definition. Unix linkers allow this somewhat sloppy practise,
1547 but linkers on some other operating systems do not. This option allows
1548 you to find potential problems from combining global symbols.
1549 Unfortunately, some C libraries use this practise, so you may get some
1550 warnings about symbols in the libraries as well as in your programs.
1552 There are three kinds of global symbols, illustrated here by C examples:
1556 A definition, which goes in the initialized data section of the output
1560 An undefined reference, which does not allocate space.
1561 There must be either a definition or a common symbol for the
1565 A common symbol. If there are only (one or more) common symbols for a
1566 variable, it goes in the uninitialized data area of the output file.
1567 The linker merges multiple common symbols for the same variable into a
1568 single symbol. If they are of different sizes, it picks the largest
1569 size. The linker turns a common symbol into a declaration, if there is
1570 a definition of the same variable.
1573 The @samp{--warn-common} option can produce five kinds of warnings.
1574 Each warning consists of a pair of lines: the first describes the symbol
1575 just encountered, and the second describes the previous symbol
1576 encountered with the same name. One or both of the two symbols will be
1581 Turning a common symbol into a reference, because there is already a
1582 definition for the symbol.
1584 @var{file}(@var{section}): warning: common of `@var{symbol}'
1585 overridden by definition
1586 @var{file}(@var{section}): warning: defined here
1590 Turning a common symbol into a reference, because a later definition for
1591 the symbol is encountered. This is the same as the previous case,
1592 except that the symbols are encountered in a different order.
1594 @var{file}(@var{section}): warning: definition of `@var{symbol}'
1596 @var{file}(@var{section}): warning: common is here
1600 Merging a common symbol with a previous same-sized common symbol.
1602 @var{file}(@var{section}): warning: multiple common
1604 @var{file}(@var{section}): warning: previous common is here
1608 Merging a common symbol with a previous larger common symbol.
1610 @var{file}(@var{section}): warning: common of `@var{symbol}'
1611 overridden by larger common
1612 @var{file}(@var{section}): warning: larger common is here
1616 Merging a common symbol with a previous smaller common symbol. This is
1617 the same as the previous case, except that the symbols are
1618 encountered in a different order.
1620 @var{file}(@var{section}): warning: common of `@var{symbol}'
1621 overriding smaller common
1622 @var{file}(@var{section}): warning: smaller common is here
1626 @kindex --warn-constructors
1627 @item --warn-constructors
1628 Warn if any global constructors are used. This is only useful for a few
1629 object file formats. For formats like COFF or ELF, the linker can not
1630 detect the use of global constructors.
1632 @kindex --warn-multiple-gp
1633 @item --warn-multiple-gp
1634 Warn if multiple global pointer values are required in the output file.
1635 This is only meaningful for certain processors, such as the Alpha.
1636 Specifically, some processors put large-valued constants in a special
1637 section. A special register (the global pointer) points into the middle
1638 of this section, so that constants can be loaded efficiently via a
1639 base-register relative addressing mode. Since the offset in
1640 base-register relative mode is fixed and relatively small (e.g., 16
1641 bits), this limits the maximum size of the constant pool. Thus, in
1642 large programs, it is often necessary to use multiple global pointer
1643 values in order to be able to address all possible constants. This
1644 option causes a warning to be issued whenever this case occurs.
1647 @cindex warnings, on undefined symbols
1648 @cindex undefined symbols, warnings on
1650 Only warn once for each undefined symbol, rather than once per module
1653 @kindex --warn-section-align
1654 @cindex warnings, on section alignment
1655 @cindex section alignment, warnings on
1656 @item --warn-section-align
1657 Warn if the address of an output section is changed because of
1658 alignment. Typically, the alignment will be set by an input section.
1659 The address will only be changed if it not explicitly specified; that
1660 is, if the @code{SECTIONS} command does not specify a start address for
1661 the section (@pxref{SECTIONS}).
1663 @kindex --warn-unresolved-symbols
1664 @item --warn-unresolved-symbols
1665 If the linker is going to report an unresolved symbol (see the option
1666 @option{--unresolved-symbols}) it will normally generate an error.
1667 This option makes it generate a warning instead.
1669 @kindex --error-unresolved-symbols
1670 @item --error-unresolved-symbols
1671 This restores the linker's default behaviour of generating errors when
1672 it is reporting unresolved symbols.
1674 @kindex --whole-archive
1675 @cindex including an entire archive
1676 @item --whole-archive
1677 For each archive mentioned on the command line after the
1678 @option{--whole-archive} option, include every object file in the archive
1679 in the link, rather than searching the archive for the required object
1680 files. This is normally used to turn an archive file into a shared
1681 library, forcing every object to be included in the resulting shared
1682 library. This option may be used more than once.
1684 Two notes when using this option from gcc: First, gcc doesn't know
1685 about this option, so you have to use @option{-Wl,-whole-archive}.
1686 Second, don't forget to use @option{-Wl,-no-whole-archive} after your
1687 list of archives, because gcc will add its own list of archives to
1688 your link and you may not want this flag to affect those as well.
1691 @item --wrap @var{symbol}
1692 Use a wrapper function for @var{symbol}. Any undefined reference to
1693 @var{symbol} will be resolved to @code{__wrap_@var{symbol}}. Any
1694 undefined reference to @code{__real_@var{symbol}} will be resolved to
1697 This can be used to provide a wrapper for a system function. The
1698 wrapper function should be called @code{__wrap_@var{symbol}}. If it
1699 wishes to call the system function, it should call
1700 @code{__real_@var{symbol}}.
1702 Here is a trivial example:
1706 __wrap_malloc (int c)
1708 printf ("malloc called with %ld\n", c);
1709 return __real_malloc (c);
1713 If you link other code with this file using @option{--wrap malloc}, then
1714 all calls to @code{malloc} will call the function @code{__wrap_malloc}
1715 instead. The call to @code{__real_malloc} in @code{__wrap_malloc} will
1716 call the real @code{malloc} function.
1718 You may wish to provide a @code{__real_malloc} function as well, so that
1719 links without the @option{--wrap} option will succeed. If you do this,
1720 you should not put the definition of @code{__real_malloc} in the same
1721 file as @code{__wrap_malloc}; if you do, the assembler may resolve the
1722 call before the linker has a chance to wrap it to @code{malloc}.
1724 @kindex --enable-new-dtags
1725 @kindex --disable-new-dtags
1726 @item --enable-new-dtags
1727 @itemx --disable-new-dtags
1728 This linker can create the new dynamic tags in ELF. But the older ELF
1729 systems may not understand them. If you specify
1730 @option{--enable-new-dtags}, the dynamic tags will be created as needed.
1731 If you specify @option{--disable-new-dtags}, no new dynamic tags will be
1732 created. By default, the new dynamic tags are not created. Note that
1733 those options are only available for ELF systems.
1739 @subsection Options Specific to i386 PE Targets
1741 @c man begin OPTIONS
1743 The i386 PE linker supports the @option{-shared} option, which causes
1744 the output to be a dynamically linked library (DLL) instead of a
1745 normal executable. You should name the output @code{*.dll} when you
1746 use this option. In addition, the linker fully supports the standard
1747 @code{*.def} files, which may be specified on the linker command line
1748 like an object file (in fact, it should precede archives it exports
1749 symbols from, to ensure that they get linked in, just like a normal
1752 In addition to the options common to all targets, the i386 PE linker
1753 support additional command line options that are specific to the i386
1754 PE target. Options that take values may be separated from their
1755 values by either a space or an equals sign.
1759 @kindex --add-stdcall-alias
1760 @item --add-stdcall-alias
1761 If given, symbols with a stdcall suffix (@@@var{nn}) will be exported
1762 as-is and also with the suffix stripped.
1763 [This option is specific to the i386 PE targeted port of the linker]
1766 @item --base-file @var{file}
1767 Use @var{file} as the name of a file in which to save the base
1768 addresses of all the relocations needed for generating DLLs with
1770 [This is an i386 PE specific option]
1774 Create a DLL instead of a regular executable. You may also use
1775 @option{-shared} or specify a @code{LIBRARY} in a given @code{.def}
1777 [This option is specific to the i386 PE targeted port of the linker]
1779 @kindex --enable-stdcall-fixup
1780 @kindex --disable-stdcall-fixup
1781 @item --enable-stdcall-fixup
1782 @itemx --disable-stdcall-fixup
1783 If the link finds a symbol that it cannot resolve, it will attempt to
1784 do ``fuzzy linking'' by looking for another defined symbol that differs
1785 only in the format of the symbol name (cdecl vs stdcall) and will
1786 resolve that symbol by linking to the match. For example, the
1787 undefined symbol @code{_foo} might be linked to the function
1788 @code{_foo@@12}, or the undefined symbol @code{_bar@@16} might be linked
1789 to the function @code{_bar}. When the linker does this, it prints a
1790 warning, since it normally should have failed to link, but sometimes
1791 import libraries generated from third-party dlls may need this feature
1792 to be usable. If you specify @option{--enable-stdcall-fixup}, this
1793 feature is fully enabled and warnings are not printed. If you specify
1794 @option{--disable-stdcall-fixup}, this feature is disabled and such
1795 mismatches are considered to be errors.
1796 [This option is specific to the i386 PE targeted port of the linker]
1798 @cindex DLLs, creating
1799 @kindex --export-all-symbols
1800 @item --export-all-symbols
1801 If given, all global symbols in the objects used to build a DLL will
1802 be exported by the DLL. Note that this is the default if there
1803 otherwise wouldn't be any exported symbols. When symbols are
1804 explicitly exported via DEF files or implicitly exported via function
1805 attributes, the default is to not export anything else unless this
1806 option is given. Note that the symbols @code{DllMain@@12},
1807 @code{DllEntryPoint@@0}, @code{DllMainCRTStartup@@12}, and
1808 @code{impure_ptr} will not be automatically
1809 exported. Also, symbols imported from other DLLs will not be
1810 re-exported, nor will symbols specifying the DLL's internal layout
1811 such as those beginning with @code{_head_} or ending with
1812 @code{_iname}. In addition, no symbols from @code{libgcc},
1813 @code{libstd++}, @code{libmingw32}, or @code{crtX.o} will be exported.
1814 Symbols whose names begin with @code{__rtti_} or @code{__builtin_} will
1815 not be exported, to help with C++ DLLs. Finally, there is an
1816 extensive list of cygwin-private symbols that are not exported
1817 (obviously, this applies on when building DLLs for cygwin targets).
1818 These cygwin-excludes are: @code{_cygwin_dll_entry@@12},
1819 @code{_cygwin_crt0_common@@8}, @code{_cygwin_noncygwin_dll_entry@@12},
1820 @code{_fmode}, @code{_impure_ptr}, @code{cygwin_attach_dll},
1821 @code{cygwin_premain0}, @code{cygwin_premain1}, @code{cygwin_premain2},
1822 @code{cygwin_premain3}, and @code{environ}.
1823 [This option is specific to the i386 PE targeted port of the linker]
1825 @kindex --exclude-symbols
1826 @item --exclude-symbols @var{symbol},@var{symbol},...
1827 Specifies a list of symbols which should not be automatically
1828 exported. The symbol names may be delimited by commas or colons.
1829 [This option is specific to the i386 PE targeted port of the linker]
1831 @kindex --exclude-libs
1832 @item --exclude-libs @var{lib},@var{lib},...
1833 Specifies a list of archive libraries from which symbols should not be automatically
1834 exported. The library names may be delimited by commas or colons. Specifying
1835 @code{--exclude-libs ALL} excludes symbols in all archive libraries from
1836 automatic export. Symbols explicitly listed in a .def file are still exported,
1837 regardless of this option.
1838 [This option is specific to the i386 PE targeted port of the linker]
1840 @kindex --file-alignment
1841 @item --file-alignment
1842 Specify the file alignment. Sections in the file will always begin at
1843 file offsets which are multiples of this number. This defaults to
1845 [This option is specific to the i386 PE targeted port of the linker]
1849 @item --heap @var{reserve}
1850 @itemx --heap @var{reserve},@var{commit}
1851 Specify the amount of memory to reserve (and optionally commit) to be
1852 used as heap for this program. The default is 1Mb reserved, 4K
1854 [This option is specific to the i386 PE targeted port of the linker]
1857 @kindex --image-base
1858 @item --image-base @var{value}
1859 Use @var{value} as the base address of your program or dll. This is
1860 the lowest memory location that will be used when your program or dll
1861 is loaded. To reduce the need to relocate and improve performance of
1862 your dlls, each should have a unique base address and not overlap any
1863 other dlls. The default is 0x400000 for executables, and 0x10000000
1865 [This option is specific to the i386 PE targeted port of the linker]
1869 If given, the stdcall suffixes (@@@var{nn}) will be stripped from
1870 symbols before they are exported.
1871 [This option is specific to the i386 PE targeted port of the linker]
1873 @kindex --major-image-version
1874 @item --major-image-version @var{value}
1875 Sets the major number of the ``image version''. Defaults to 1.
1876 [This option is specific to the i386 PE targeted port of the linker]
1878 @kindex --major-os-version
1879 @item --major-os-version @var{value}
1880 Sets the major number of the ``os version''. Defaults to 4.
1881 [This option is specific to the i386 PE targeted port of the linker]
1883 @kindex --major-subsystem-version
1884 @item --major-subsystem-version @var{value}
1885 Sets the major number of the ``subsystem version''. Defaults to 4.
1886 [This option is specific to the i386 PE targeted port of the linker]
1888 @kindex --minor-image-version
1889 @item --minor-image-version @var{value}
1890 Sets the minor number of the ``image version''. Defaults to 0.
1891 [This option is specific to the i386 PE targeted port of the linker]
1893 @kindex --minor-os-version
1894 @item --minor-os-version @var{value}
1895 Sets the minor number of the ``os version''. Defaults to 0.
1896 [This option is specific to the i386 PE targeted port of the linker]
1898 @kindex --minor-subsystem-version
1899 @item --minor-subsystem-version @var{value}
1900 Sets the minor number of the ``subsystem version''. Defaults to 0.
1901 [This option is specific to the i386 PE targeted port of the linker]
1903 @cindex DEF files, creating
1904 @cindex DLLs, creating
1905 @kindex --output-def
1906 @item --output-def @var{file}
1907 The linker will create the file @var{file} which will contain a DEF
1908 file corresponding to the DLL the linker is generating. This DEF file
1909 (which should be called @code{*.def}) may be used to create an import
1910 library with @code{dlltool} or may be used as a reference to
1911 automatically or implicitly exported symbols.
1912 [This option is specific to the i386 PE targeted port of the linker]
1914 @cindex DLLs, creating
1915 @kindex --out-implib
1916 @item --out-implib @var{file}
1917 The linker will create the file @var{file} which will contain an
1918 import lib corresponding to the DLL the linker is generating. This
1919 import lib (which should be called @code{*.dll.a} or @code{*.a}
1920 may be used to link clients against the generated DLL; this behaviour
1921 makes it possible to skip a separate @code{dlltool} import library
1923 [This option is specific to the i386 PE targeted port of the linker]
1925 @kindex --enable-auto-image-base
1926 @item --enable-auto-image-base
1927 Automatically choose the image base for DLLs, unless one is specified
1928 using the @code{--image-base} argument. By using a hash generated
1929 from the dllname to create unique image bases for each DLL, in-memory
1930 collisions and relocations which can delay program execution are
1932 [This option is specific to the i386 PE targeted port of the linker]
1934 @kindex --disable-auto-image-base
1935 @item --disable-auto-image-base
1936 Do not automatically generate a unique image base. If there is no
1937 user-specified image base (@code{--image-base}) then use the platform
1939 [This option is specific to the i386 PE targeted port of the linker]
1941 @cindex DLLs, linking to
1942 @kindex --dll-search-prefix
1943 @item --dll-search-prefix @var{string}
1944 When linking dynamically to a dll without an import library,
1945 search for @code{<string><basename>.dll} in preference to
1946 @code{lib<basename>.dll}. This behaviour allows easy distinction
1947 between DLLs built for the various "subplatforms": native, cygwin,
1948 uwin, pw, etc. For instance, cygwin DLLs typically use
1949 @code{--dll-search-prefix=cyg}.
1950 [This option is specific to the i386 PE targeted port of the linker]
1952 @kindex --enable-auto-import
1953 @item --enable-auto-import
1954 Do sophisticated linking of @code{_symbol} to @code{__imp__symbol} for
1955 DATA imports from DLLs, and create the necessary thunking symbols when
1956 building the import libraries with those DATA exports. This generally
1957 will 'just work' -- but sometimes you may see this message:
1959 "variable '<var>' can't be auto-imported. Please read the
1960 documentation for ld's @code{--enable-auto-import} for details."
1962 This message occurs when some (sub)expression accesses an address
1963 ultimately given by the sum of two constants (Win32 import tables only
1964 allow one). Instances where this may occur include accesses to member
1965 fields of struct variables imported from a DLL, as well as using a
1966 constant index into an array variable imported from a DLL. Any
1967 multiword variable (arrays, structs, long long, etc) may trigger
1968 this error condition. However, regardless of the exact data type
1969 of the offending exported variable, ld will always detect it, issue
1970 the warning, and exit.
1972 There are several ways to address this difficulty, regardless of the
1973 data type of the exported variable:
1975 One way is to use --enable-runtime-pseudo-reloc switch. This leaves the task
1976 of adjusting references in your client code for runtime environment, so
1977 this method works only when runtime environment supports this feature.
1979 A second solution is to force one of the 'constants' to be a variable --
1980 that is, unknown and un-optimizable at compile time. For arrays,
1981 there are two possibilities: a) make the indexee (the array's address)
1982 a variable, or b) make the 'constant' index a variable. Thus:
1985 extern type extern_array[];
1987 @{ volatile type *t=extern_array; t[1] @}
1993 extern type extern_array[];
1995 @{ volatile int t=1; extern_array[t] @}
1998 For structs (and most other multiword data types) the only option
1999 is to make the struct itself (or the long long, or the ...) variable:
2002 extern struct s extern_struct;
2003 extern_struct.field -->
2004 @{ volatile struct s *t=&extern_struct; t->field @}
2010 extern long long extern_ll;
2012 @{ volatile long long * local_ll=&extern_ll; *local_ll @}
2015 A third method of dealing with this difficulty is to abandon
2016 'auto-import' for the offending symbol and mark it with
2017 @code{__declspec(dllimport)}. However, in practise that
2018 requires using compile-time #defines to indicate whether you are
2019 building a DLL, building client code that will link to the DLL, or
2020 merely building/linking to a static library. In making the choice
2021 between the various methods of resolving the 'direct address with
2022 constant offset' problem, you should consider typical real-world usage:
2030 void main(int argc, char **argv)@{
2031 printf("%d\n",arr[1]);
2041 void main(int argc, char **argv)@{
2042 /* This workaround is for win32 and cygwin; do not "optimize" */
2043 volatile int *parr = arr;
2044 printf("%d\n",parr[1]);
2051 /* Note: auto-export is assumed (no __declspec(dllexport)) */
2052 #if (defined(_WIN32) || defined(__CYGWIN__)) && \
2053 !(defined(FOO_BUILD_DLL) || defined(FOO_STATIC))
2054 #define FOO_IMPORT __declspec(dllimport)
2058 extern FOO_IMPORT int arr[];
2061 void main(int argc, char **argv)@{
2062 printf("%d\n",arr[1]);
2066 A fourth way to avoid this problem is to re-code your
2067 library to use a functional interface rather than a data interface
2068 for the offending variables (e.g. set_foo() and get_foo() accessor
2070 [This option is specific to the i386 PE targeted port of the linker]
2072 @kindex --disable-auto-import
2073 @item --disable-auto-import
2074 Do not attempt to do sophisticated linking of @code{_symbol} to
2075 @code{__imp__symbol} for DATA imports from DLLs.
2076 [This option is specific to the i386 PE targeted port of the linker]
2078 @kindex --enable-runtime-pseudo-reloc
2079 @item --enable-runtime-pseudo-reloc
2080 If your code contains expressions described in --enable-auto-import section,
2081 that is, DATA imports from DLL with non-zero offset, this switch will create
2082 a vector of 'runtime pseudo relocations' which can be used by runtime
2083 environment to adjust references to such data in your client code.
2084 [This option is specific to the i386 PE targeted port of the linker]
2086 @kindex --disable-runtime-pseudo-reloc
2087 @item --disable-runtime-pseudo-reloc
2088 Do not create pseudo relocations for non-zero offset DATA imports from
2089 DLLs. This is the default.
2090 [This option is specific to the i386 PE targeted port of the linker]
2092 @kindex --enable-extra-pe-debug
2093 @item --enable-extra-pe-debug
2094 Show additional debug info related to auto-import symbol thunking.
2095 [This option is specific to the i386 PE targeted port of the linker]
2097 @kindex --section-alignment
2098 @item --section-alignment
2099 Sets the section alignment. Sections in memory will always begin at
2100 addresses which are a multiple of this number. Defaults to 0x1000.
2101 [This option is specific to the i386 PE targeted port of the linker]
2105 @item --stack @var{reserve}
2106 @itemx --stack @var{reserve},@var{commit}
2107 Specify the amount of memory to reserve (and optionally commit) to be
2108 used as stack for this program. The default is 2Mb reserved, 4K
2110 [This option is specific to the i386 PE targeted port of the linker]
2113 @item --subsystem @var{which}
2114 @itemx --subsystem @var{which}:@var{major}
2115 @itemx --subsystem @var{which}:@var{major}.@var{minor}
2116 Specifies the subsystem under which your program will execute. The
2117 legal values for @var{which} are @code{native}, @code{windows},
2118 @code{console}, and @code{posix}. You may optionally set the
2119 subsystem version also.
2120 [This option is specific to the i386 PE targeted port of the linker]
2128 @section Environment Variables
2130 @c man begin ENVIRONMENT
2132 You can change the behaviour of @command{ld} with the environment variables
2133 @ifclear SingleFormat
2136 @code{LDEMULATION} and @code{COLLECT_NO_DEMANGLE}.
2138 @ifclear SingleFormat
2140 @cindex default input format
2141 @code{GNUTARGET} determines the input-file object format if you don't
2142 use @samp{-b} (or its synonym @samp{--format}). Its value should be one
2143 of the BFD names for an input format (@pxref{BFD}). If there is no
2144 @code{GNUTARGET} in the environment, @command{ld} uses the natural format
2145 of the target. If @code{GNUTARGET} is set to @code{default} then BFD
2146 attempts to discover the input format by examining binary input files;
2147 this method often succeeds, but there are potential ambiguities, since
2148 there is no method of ensuring that the magic number used to specify
2149 object-file formats is unique. However, the configuration procedure for
2150 BFD on each system places the conventional format for that system first
2151 in the search-list, so ambiguities are resolved in favor of convention.
2155 @cindex default emulation
2156 @cindex emulation, default
2157 @code{LDEMULATION} determines the default emulation if you don't use the
2158 @samp{-m} option. The emulation can affect various aspects of linker
2159 behaviour, particularly the default linker script. You can list the
2160 available emulations with the @samp{--verbose} or @samp{-V} options. If
2161 the @samp{-m} option is not used, and the @code{LDEMULATION} environment
2162 variable is not defined, the default emulation depends upon how the
2163 linker was configured.
2165 @kindex COLLECT_NO_DEMANGLE
2166 @cindex demangling, default
2167 Normally, the linker will default to demangling symbols. However, if
2168 @code{COLLECT_NO_DEMANGLE} is set in the environment, then it will
2169 default to not demangling symbols. This environment variable is used in
2170 a similar fashion by the @code{gcc} linker wrapper program. The default
2171 may be overridden by the @samp{--demangle} and @samp{--no-demangle}
2178 @chapter Linker Scripts
2181 @cindex linker scripts
2182 @cindex command files
2183 Every link is controlled by a @dfn{linker script}. This script is
2184 written in the linker command language.
2186 The main purpose of the linker script is to describe how the sections in
2187 the input files should be mapped into the output file, and to control
2188 the memory layout of the output file. Most linker scripts do nothing
2189 more than this. However, when necessary, the linker script can also
2190 direct the linker to perform many other operations, using the commands
2193 The linker always uses a linker script. If you do not supply one
2194 yourself, the linker will use a default script that is compiled into the
2195 linker executable. You can use the @samp{--verbose} command line option
2196 to display the default linker script. Certain command line options,
2197 such as @samp{-r} or @samp{-N}, will affect the default linker script.
2199 You may supply your own linker script by using the @samp{-T} command
2200 line option. When you do this, your linker script will replace the
2201 default linker script.
2203 You may also use linker scripts implicitly by naming them as input files
2204 to the linker, as though they were files to be linked. @xref{Implicit
2208 * Basic Script Concepts:: Basic Linker Script Concepts
2209 * Script Format:: Linker Script Format
2210 * Simple Example:: Simple Linker Script Example
2211 * Simple Commands:: Simple Linker Script Commands
2212 * Assignments:: Assigning Values to Symbols
2213 * SECTIONS:: SECTIONS Command
2214 * MEMORY:: MEMORY Command
2215 * PHDRS:: PHDRS Command
2216 * VERSION:: VERSION Command
2217 * Expressions:: Expressions in Linker Scripts
2218 * Implicit Linker Scripts:: Implicit Linker Scripts
2221 @node Basic Script Concepts
2222 @section Basic Linker Script Concepts
2223 @cindex linker script concepts
2224 We need to define some basic concepts and vocabulary in order to
2225 describe the linker script language.
2227 The linker combines input files into a single output file. The output
2228 file and each input file are in a special data format known as an
2229 @dfn{object file format}. Each file is called an @dfn{object file}.
2230 The output file is often called an @dfn{executable}, but for our
2231 purposes we will also call it an object file. Each object file has,
2232 among other things, a list of @dfn{sections}. We sometimes refer to a
2233 section in an input file as an @dfn{input section}; similarly, a section
2234 in the output file is an @dfn{output section}.
2236 Each section in an object file has a name and a size. Most sections
2237 also have an associated block of data, known as the @dfn{section
2238 contents}. A section may be marked as @dfn{loadable}, which mean that
2239 the contents should be loaded into memory when the output file is run.
2240 A section with no contents may be @dfn{allocatable}, which means that an
2241 area in memory should be set aside, but nothing in particular should be
2242 loaded there (in some cases this memory must be zeroed out). A section
2243 which is neither loadable nor allocatable typically contains some sort
2244 of debugging information.
2246 Every loadable or allocatable output section has two addresses. The
2247 first is the @dfn{VMA}, or virtual memory address. This is the address
2248 the section will have when the output file is run. The second is the
2249 @dfn{LMA}, or load memory address. This is the address at which the
2250 section will be loaded. In most cases the two addresses will be the
2251 same. An example of when they might be different is when a data section
2252 is loaded into ROM, and then copied into RAM when the program starts up
2253 (this technique is often used to initialize global variables in a ROM
2254 based system). In this case the ROM address would be the LMA, and the
2255 RAM address would be the VMA.
2257 You can see the sections in an object file by using the @code{objdump}
2258 program with the @samp{-h} option.
2260 Every object file also has a list of @dfn{symbols}, known as the
2261 @dfn{symbol table}. A symbol may be defined or undefined. Each symbol
2262 has a name, and each defined symbol has an address, among other
2263 information. If you compile a C or C++ program into an object file, you
2264 will get a defined symbol for every defined function and global or
2265 static variable. Every undefined function or global variable which is
2266 referenced in the input file will become an undefined symbol.
2268 You can see the symbols in an object file by using the @code{nm}
2269 program, or by using the @code{objdump} program with the @samp{-t}
2273 @section Linker Script Format
2274 @cindex linker script format
2275 Linker scripts are text files.
2277 You write a linker script as a series of commands. Each command is
2278 either a keyword, possibly followed by arguments, or an assignment to a
2279 symbol. You may separate commands using semicolons. Whitespace is
2282 Strings such as file or format names can normally be entered directly.
2283 If the file name contains a character such as a comma which would
2284 otherwise serve to separate file names, you may put the file name in
2285 double quotes. There is no way to use a double quote character in a
2288 You may include comments in linker scripts just as in C, delimited by
2289 @samp{/*} and @samp{*/}. As in C, comments are syntactically equivalent
2292 @node Simple Example
2293 @section Simple Linker Script Example
2294 @cindex linker script example
2295 @cindex example of linker script
2296 Many linker scripts are fairly simple.
2298 The simplest possible linker script has just one command:
2299 @samp{SECTIONS}. You use the @samp{SECTIONS} command to describe the
2300 memory layout of the output file.
2302 The @samp{SECTIONS} command is a powerful command. Here we will
2303 describe a simple use of it. Let's assume your program consists only of
2304 code, initialized data, and uninitialized data. These will be in the
2305 @samp{.text}, @samp{.data}, and @samp{.bss} sections, respectively.
2306 Let's assume further that these are the only sections which appear in
2309 For this example, let's say that the code should be loaded at address
2310 0x10000, and that the data should start at address 0x8000000. Here is a
2311 linker script which will do that:
2316 .text : @{ *(.text) @}
2318 .data : @{ *(.data) @}
2319 .bss : @{ *(.bss) @}
2323 You write the @samp{SECTIONS} command as the keyword @samp{SECTIONS},
2324 followed by a series of symbol assignments and output section
2325 descriptions enclosed in curly braces.
2327 The first line inside the @samp{SECTIONS} command of the above example
2328 sets the value of the special symbol @samp{.}, which is the location
2329 counter. If you do not specify the address of an output section in some
2330 other way (other ways are described later), the address is set from the
2331 current value of the location counter. The location counter is then
2332 incremented by the size of the output section. At the start of the
2333 @samp{SECTIONS} command, the location counter has the value @samp{0}.
2335 The second line defines an output section, @samp{.text}. The colon is
2336 required syntax which may be ignored for now. Within the curly braces
2337 after the output section name, you list the names of the input sections
2338 which should be placed into this output section. The @samp{*} is a
2339 wildcard which matches any file name. The expression @samp{*(.text)}
2340 means all @samp{.text} input sections in all input files.
2342 Since the location counter is @samp{0x10000} when the output section
2343 @samp{.text} is defined, the linker will set the address of the
2344 @samp{.text} section in the output file to be @samp{0x10000}.
2346 The remaining lines define the @samp{.data} and @samp{.bss} sections in
2347 the output file. The linker will place the @samp{.data} output section
2348 at address @samp{0x8000000}. After the linker places the @samp{.data}
2349 output section, the value of the location counter will be
2350 @samp{0x8000000} plus the size of the @samp{.data} output section. The
2351 effect is that the linker will place the @samp{.bss} output section
2352 immediately after the @samp{.data} output section in memory
2354 The linker will ensure that each output section has the required
2355 alignment, by increasing the location counter if necessary. In this
2356 example, the specified addresses for the @samp{.text} and @samp{.data}
2357 sections will probably satisfy any alignment constraints, but the linker
2358 may have to create a small gap between the @samp{.data} and @samp{.bss}
2361 That's it! That's a simple and complete linker script.
2363 @node Simple Commands
2364 @section Simple Linker Script Commands
2365 @cindex linker script simple commands
2366 In this section we describe the simple linker script commands.
2369 * Entry Point:: Setting the entry point
2370 * File Commands:: Commands dealing with files
2371 @ifclear SingleFormat
2372 * Format Commands:: Commands dealing with object file formats
2375 * Miscellaneous Commands:: Other linker script commands
2379 @subsection Setting the Entry Point
2380 @kindex ENTRY(@var{symbol})
2381 @cindex start of execution
2382 @cindex first instruction
2384 The first instruction to execute in a program is called the @dfn{entry
2385 point}. You can use the @code{ENTRY} linker script command to set the
2386 entry point. The argument is a symbol name:
2391 There are several ways to set the entry point. The linker will set the
2392 entry point by trying each of the following methods in order, and
2393 stopping when one of them succeeds:
2396 the @samp{-e} @var{entry} command-line option;
2398 the @code{ENTRY(@var{symbol})} command in a linker script;
2400 the value of the symbol @code{start}, if defined;
2402 the address of the first byte of the @samp{.text} section, if present;
2404 The address @code{0}.
2408 @subsection Commands Dealing with Files
2409 @cindex linker script file commands
2410 Several linker script commands deal with files.
2413 @item INCLUDE @var{filename}
2414 @kindex INCLUDE @var{filename}
2415 @cindex including a linker script
2416 Include the linker script @var{filename} at this point. The file will
2417 be searched for in the current directory, and in any directory specified
2418 with the @option{-L} option. You can nest calls to @code{INCLUDE} up to
2421 @item INPUT(@var{file}, @var{file}, @dots{})
2422 @itemx INPUT(@var{file} @var{file} @dots{})
2423 @kindex INPUT(@var{files})
2424 @cindex input files in linker scripts
2425 @cindex input object files in linker scripts
2426 @cindex linker script input object files
2427 The @code{INPUT} command directs the linker to include the named files
2428 in the link, as though they were named on the command line.
2430 For example, if you always want to include @file{subr.o} any time you do
2431 a link, but you can't be bothered to put it on every link command line,
2432 then you can put @samp{INPUT (subr.o)} in your linker script.
2434 In fact, if you like, you can list all of your input files in the linker
2435 script, and then invoke the linker with nothing but a @samp{-T} option.
2437 In case a @dfn{sysroot prefix} is configured, and the filename starts
2438 with the @samp{/} character, and the script being processed was
2439 located inside the @dfn{sysroot prefix}, the filename will be looked
2440 for in the @dfn{sysroot prefix}. Otherwise, the linker will try to
2441 open the file in the current directory. If it is not found, the
2442 linker will search through the archive library search path. See the
2443 description of @samp{-L} in @ref{Options,,Command Line Options}.
2445 If you use @samp{INPUT (-l@var{file})}, @command{ld} will transform the
2446 name to @code{lib@var{file}.a}, as with the command line argument
2449 When you use the @code{INPUT} command in an implicit linker script, the
2450 files will be included in the link at the point at which the linker
2451 script file is included. This can affect archive searching.
2453 @item GROUP(@var{file}, @var{file}, @dots{})
2454 @itemx GROUP(@var{file} @var{file} @dots{})
2455 @kindex GROUP(@var{files})
2456 @cindex grouping input files
2457 The @code{GROUP} command is like @code{INPUT}, except that the named
2458 files should all be archives, and they are searched repeatedly until no
2459 new undefined references are created. See the description of @samp{-(}
2460 in @ref{Options,,Command Line Options}.
2462 @item OUTPUT(@var{filename})
2463 @kindex OUTPUT(@var{filename})
2464 @cindex output file name in linker scripot
2465 The @code{OUTPUT} command names the output file. Using
2466 @code{OUTPUT(@var{filename})} in the linker script is exactly like using
2467 @samp{-o @var{filename}} on the command line (@pxref{Options,,Command
2468 Line Options}). If both are used, the command line option takes
2471 You can use the @code{OUTPUT} command to define a default name for the
2472 output file other than the usual default of @file{a.out}.
2474 @item SEARCH_DIR(@var{path})
2475 @kindex SEARCH_DIR(@var{path})
2476 @cindex library search path in linker script
2477 @cindex archive search path in linker script
2478 @cindex search path in linker script
2479 The @code{SEARCH_DIR} command adds @var{path} to the list of paths where
2480 @command{ld} looks for archive libraries. Using
2481 @code{SEARCH_DIR(@var{path})} is exactly like using @samp{-L @var{path}}
2482 on the command line (@pxref{Options,,Command Line Options}). If both
2483 are used, then the linker will search both paths. Paths specified using
2484 the command line option are searched first.
2486 @item STARTUP(@var{filename})
2487 @kindex STARTUP(@var{filename})
2488 @cindex first input file
2489 The @code{STARTUP} command is just like the @code{INPUT} command, except
2490 that @var{filename} will become the first input file to be linked, as
2491 though it were specified first on the command line. This may be useful
2492 when using a system in which the entry point is always the start of the
2496 @ifclear SingleFormat
2497 @node Format Commands
2498 @subsection Commands Dealing with Object File Formats
2499 A couple of linker script commands deal with object file formats.
2502 @item OUTPUT_FORMAT(@var{bfdname})
2503 @itemx OUTPUT_FORMAT(@var{default}, @var{big}, @var{little})
2504 @kindex OUTPUT_FORMAT(@var{bfdname})
2505 @cindex output file format in linker script
2506 The @code{OUTPUT_FORMAT} command names the BFD format to use for the
2507 output file (@pxref{BFD}). Using @code{OUTPUT_FORMAT(@var{bfdname})} is
2508 exactly like using @samp{--oformat @var{bfdname}} on the command line
2509 (@pxref{Options,,Command Line Options}). If both are used, the command
2510 line option takes precedence.
2512 You can use @code{OUTPUT_FORMAT} with three arguments to use different
2513 formats based on the @samp{-EB} and @samp{-EL} command line options.
2514 This permits the linker script to set the output format based on the
2517 If neither @samp{-EB} nor @samp{-EL} are used, then the output format
2518 will be the first argument, @var{default}. If @samp{-EB} is used, the
2519 output format will be the second argument, @var{big}. If @samp{-EL} is
2520 used, the output format will be the third argument, @var{little}.
2522 For example, the default linker script for the MIPS ELF target uses this
2525 OUTPUT_FORMAT(elf32-bigmips, elf32-bigmips, elf32-littlemips)
2527 This says that the default format for the output file is
2528 @samp{elf32-bigmips}, but if the user uses the @samp{-EL} command line
2529 option, the output file will be created in the @samp{elf32-littlemips}
2532 @item TARGET(@var{bfdname})
2533 @kindex TARGET(@var{bfdname})
2534 @cindex input file format in linker script
2535 The @code{TARGET} command names the BFD format to use when reading input
2536 files. It affects subsequent @code{INPUT} and @code{GROUP} commands.
2537 This command is like using @samp{-b @var{bfdname}} on the command line
2538 (@pxref{Options,,Command Line Options}). If the @code{TARGET} command
2539 is used but @code{OUTPUT_FORMAT} is not, then the last @code{TARGET}
2540 command is also used to set the format for the output file. @xref{BFD}.
2544 @node Miscellaneous Commands
2545 @subsection Other Linker Script Commands
2546 There are a few other linker scripts commands.
2549 @item ASSERT(@var{exp}, @var{message})
2551 @cindex assertion in linker script
2552 Ensure that @var{exp} is non-zero. If it is zero, then exit the linker
2553 with an error code, and print @var{message}.
2555 @item EXTERN(@var{symbol} @var{symbol} @dots{})
2557 @cindex undefined symbol in linker script
2558 Force @var{symbol} to be entered in the output file as an undefined
2559 symbol. Doing this may, for example, trigger linking of additional
2560 modules from standard libraries. You may list several @var{symbol}s for
2561 each @code{EXTERN}, and you may use @code{EXTERN} multiple times. This
2562 command has the same effect as the @samp{-u} command-line option.
2564 @item FORCE_COMMON_ALLOCATION
2565 @kindex FORCE_COMMON_ALLOCATION
2566 @cindex common allocation in linker script
2567 This command has the same effect as the @samp{-d} command-line option:
2568 to make @command{ld} assign space to common symbols even if a relocatable
2569 output file is specified (@samp{-r}).
2571 @item INHIBIT_COMMON_ALLOCATION
2572 @kindex INHIBIT_COMMON_ALLOCATION
2573 @cindex common allocation in linker script
2574 This command has the same effect as the @samp{--no-define-common}
2575 command-line option: to make @code{ld} omit the assignment of addresses
2576 to common symbols even for a non-relocatable output file.
2578 @item NOCROSSREFS(@var{section} @var{section} @dots{})
2579 @kindex NOCROSSREFS(@var{sections})
2580 @cindex cross references
2581 This command may be used to tell @command{ld} to issue an error about any
2582 references among certain output sections.
2584 In certain types of programs, particularly on embedded systems when
2585 using overlays, when one section is loaded into memory, another section
2586 will not be. Any direct references between the two sections would be
2587 errors. For example, it would be an error if code in one section called
2588 a function defined in the other section.
2590 The @code{NOCROSSREFS} command takes a list of output section names. If
2591 @command{ld} detects any cross references between the sections, it reports
2592 an error and returns a non-zero exit status. Note that the
2593 @code{NOCROSSREFS} command uses output section names, not input section
2596 @ifclear SingleFormat
2597 @item OUTPUT_ARCH(@var{bfdarch})
2598 @kindex OUTPUT_ARCH(@var{bfdarch})
2599 @cindex machine architecture
2600 @cindex architecture
2601 Specify a particular output machine architecture. The argument is one
2602 of the names used by the BFD library (@pxref{BFD}). You can see the
2603 architecture of an object file by using the @code{objdump} program with
2604 the @samp{-f} option.
2609 @section Assigning Values to Symbols
2610 @cindex assignment in scripts
2611 @cindex symbol definition, scripts
2612 @cindex variables, defining
2613 You may assign a value to a symbol in a linker script. This will define
2614 the symbol as a global symbol.
2617 * Simple Assignments:: Simple Assignments
2621 @node Simple Assignments
2622 @subsection Simple Assignments
2624 You may assign to a symbol using any of the C assignment operators:
2627 @item @var{symbol} = @var{expression} ;
2628 @itemx @var{symbol} += @var{expression} ;
2629 @itemx @var{symbol} -= @var{expression} ;
2630 @itemx @var{symbol} *= @var{expression} ;
2631 @itemx @var{symbol} /= @var{expression} ;
2632 @itemx @var{symbol} <<= @var{expression} ;
2633 @itemx @var{symbol} >>= @var{expression} ;
2634 @itemx @var{symbol} &= @var{expression} ;
2635 @itemx @var{symbol} |= @var{expression} ;
2638 The first case will define @var{symbol} to the value of
2639 @var{expression}. In the other cases, @var{symbol} must already be
2640 defined, and the value will be adjusted accordingly.
2642 The special symbol name @samp{.} indicates the location counter. You
2643 may only use this within a @code{SECTIONS} command.
2645 The semicolon after @var{expression} is required.
2647 Expressions are defined below; see @ref{Expressions}.
2649 You may write symbol assignments as commands in their own right, or as
2650 statements within a @code{SECTIONS} command, or as part of an output
2651 section description in a @code{SECTIONS} command.
2653 The section of the symbol will be set from the section of the
2654 expression; for more information, see @ref{Expression Section}.
2656 Here is an example showing the three different places that symbol
2657 assignments may be used:
2668 _bdata = (. + 3) & ~ 3;
2669 .data : @{ *(.data) @}
2673 In this example, the symbol @samp{floating_point} will be defined as
2674 zero. The symbol @samp{_etext} will be defined as the address following
2675 the last @samp{.text} input section. The symbol @samp{_bdata} will be
2676 defined as the address following the @samp{.text} output section aligned
2677 upward to a 4 byte boundary.
2682 In some cases, it is desirable for a linker script to define a symbol
2683 only if it is referenced and is not defined by any object included in
2684 the link. For example, traditional linkers defined the symbol
2685 @samp{etext}. However, ANSI C requires that the user be able to use
2686 @samp{etext} as a function name without encountering an error. The
2687 @code{PROVIDE} keyword may be used to define a symbol, such as
2688 @samp{etext}, only if it is referenced but not defined. The syntax is
2689 @code{PROVIDE(@var{symbol} = @var{expression})}.
2691 Here is an example of using @code{PROVIDE} to define @samp{etext}:
2704 In this example, if the program defines @samp{_etext} (with a leading
2705 underscore), the linker will give a multiple definition error. If, on
2706 the other hand, the program defines @samp{etext} (with no leading
2707 underscore), the linker will silently use the definition in the program.
2708 If the program references @samp{etext} but does not define it, the
2709 linker will use the definition in the linker script.
2712 @section SECTIONS Command
2714 The @code{SECTIONS} command tells the linker how to map input sections
2715 into output sections, and how to place the output sections in memory.
2717 The format of the @code{SECTIONS} command is:
2721 @var{sections-command}
2722 @var{sections-command}
2727 Each @var{sections-command} may of be one of the following:
2731 an @code{ENTRY} command (@pxref{Entry Point,,Entry command})
2733 a symbol assignment (@pxref{Assignments})
2735 an output section description
2737 an overlay description
2740 The @code{ENTRY} command and symbol assignments are permitted inside the
2741 @code{SECTIONS} command for convenience in using the location counter in
2742 those commands. This can also make the linker script easier to
2743 understand because you can use those commands at meaningful points in
2744 the layout of the output file.
2746 Output section descriptions and overlay descriptions are described
2749 If you do not use a @code{SECTIONS} command in your linker script, the
2750 linker will place each input section into an identically named output
2751 section in the order that the sections are first encountered in the
2752 input files. If all input sections are present in the first file, for
2753 example, the order of sections in the output file will match the order
2754 in the first input file. The first section will be at address zero.
2757 * Output Section Description:: Output section description
2758 * Output Section Name:: Output section name
2759 * Output Section Address:: Output section address
2760 * Input Section:: Input section description
2761 * Output Section Data:: Output section data
2762 * Output Section Keywords:: Output section keywords
2763 * Output Section Discarding:: Output section discarding
2764 * Output Section Attributes:: Output section attributes
2765 * Overlay Description:: Overlay description
2768 @node Output Section Description
2769 @subsection Output Section Description
2770 The full description of an output section looks like this:
2773 @var{section} [@var{address}] [(@var{type})] :
2774 [AT(@var{lma})] [SUBALIGN(@var{subsection_align})]
2776 @var{output-section-command}
2777 @var{output-section-command}
2779 @} [>@var{region}] [AT>@var{lma_region}] [:@var{phdr} :@var{phdr} @dots{}] [=@var{fillexp}]
2783 Most output sections do not use most of the optional section attributes.
2785 The whitespace around @var{section} is required, so that the section
2786 name is unambiguous. The colon and the curly braces are also required.
2787 The line breaks and other white space are optional.
2789 Each @var{output-section-command} may be one of the following:
2793 a symbol assignment (@pxref{Assignments})
2795 an input section description (@pxref{Input Section})
2797 data values to include directly (@pxref{Output Section Data})
2799 a special output section keyword (@pxref{Output Section Keywords})
2802 @node Output Section Name
2803 @subsection Output Section Name
2804 @cindex name, section
2805 @cindex section name
2806 The name of the output section is @var{section}. @var{section} must
2807 meet the constraints of your output format. In formats which only
2808 support a limited number of sections, such as @code{a.out}, the name
2809 must be one of the names supported by the format (@code{a.out}, for
2810 example, allows only @samp{.text}, @samp{.data} or @samp{.bss}). If the
2811 output format supports any number of sections, but with numbers and not
2812 names (as is the case for Oasys), the name should be supplied as a
2813 quoted numeric string. A section name may consist of any sequence of
2814 characters, but a name which contains any unusual characters such as
2815 commas must be quoted.
2817 The output section name @samp{/DISCARD/} is special; @ref{Output Section
2820 @node Output Section Address
2821 @subsection Output Section Description
2822 @cindex address, section
2823 @cindex section address
2824 The @var{address} is an expression for the VMA (the virtual memory
2825 address) of the output section. If you do not provide @var{address},
2826 the linker will set it based on @var{region} if present, or otherwise
2827 based on the current value of the location counter.
2829 If you provide @var{address}, the address of the output section will be
2830 set to precisely that. If you provide neither @var{address} nor
2831 @var{region}, then the address of the output section will be set to the
2832 current value of the location counter aligned to the alignment
2833 requirements of the output section. The alignment requirement of the
2834 output section is the strictest alignment of any input section contained
2835 within the output section.
2839 .text . : @{ *(.text) @}
2844 .text : @{ *(.text) @}
2847 are subtly different. The first will set the address of the
2848 @samp{.text} output section to the current value of the location
2849 counter. The second will set it to the current value of the location
2850 counter aligned to the strictest alignment of a @samp{.text} input
2853 The @var{address} may be an arbitrary expression; @ref{Expressions}.
2854 For example, if you want to align the section on a 0x10 byte boundary,
2855 so that the lowest four bits of the section address are zero, you could
2856 do something like this:
2858 .text ALIGN(0x10) : @{ *(.text) @}
2861 This works because @code{ALIGN} returns the current location counter
2862 aligned upward to the specified value.
2864 Specifying @var{address} for a section will change the value of the
2868 @subsection Input Section Description
2869 @cindex input sections
2870 @cindex mapping input sections to output sections
2871 The most common output section command is an input section description.
2873 The input section description is the most basic linker script operation.
2874 You use output sections to tell the linker how to lay out your program
2875 in memory. You use input section descriptions to tell the linker how to
2876 map the input files into your memory layout.
2879 * Input Section Basics:: Input section basics
2880 * Input Section Wildcards:: Input section wildcard patterns
2881 * Input Section Common:: Input section for common symbols
2882 * Input Section Keep:: Input section and garbage collection
2883 * Input Section Example:: Input section example
2886 @node Input Section Basics
2887 @subsubsection Input Section Basics
2888 @cindex input section basics
2889 An input section description consists of a file name optionally followed
2890 by a list of section names in parentheses.
2892 The file name and the section name may be wildcard patterns, which we
2893 describe further below (@pxref{Input Section Wildcards}).
2895 The most common input section description is to include all input
2896 sections with a particular name in the output section. For example, to
2897 include all input @samp{.text} sections, you would write:
2902 Here the @samp{*} is a wildcard which matches any file name. To exclude a list
2903 of files from matching the file name wildcard, EXCLUDE_FILE may be used to
2904 match all files except the ones specified in the EXCLUDE_FILE list. For
2907 (*(EXCLUDE_FILE (*crtend.o *otherfile.o) .ctors))
2909 will cause all .ctors sections from all files except @file{crtend.o} and
2910 @file{otherfile.o} to be included.
2912 There are two ways to include more than one section:
2918 The difference between these is the order in which the @samp{.text} and
2919 @samp{.rdata} input sections will appear in the output section. In the
2920 first example, they will be intermingled, appearing in the same order as
2921 they are found in the linker input. In the second example, all
2922 @samp{.text} input sections will appear first, followed by all
2923 @samp{.rdata} input sections.
2925 You can specify a file name to include sections from a particular file.
2926 You would do this if one or more of your files contain special data that
2927 needs to be at a particular location in memory. For example:
2932 If you use a file name without a list of sections, then all sections in
2933 the input file will be included in the output section. This is not
2934 commonly done, but it may by useful on occasion. For example:
2939 When you use a file name which does not contain any wild card
2940 characters, the linker will first see if you also specified the file
2941 name on the linker command line or in an @code{INPUT} command. If you
2942 did not, the linker will attempt to open the file as an input file, as
2943 though it appeared on the command line. Note that this differs from an
2944 @code{INPUT} command, because the linker will not search for the file in
2945 the archive search path.
2947 @node Input Section Wildcards
2948 @subsubsection Input Section Wildcard Patterns
2949 @cindex input section wildcards
2950 @cindex wildcard file name patterns
2951 @cindex file name wildcard patterns
2952 @cindex section name wildcard patterns
2953 In an input section description, either the file name or the section
2954 name or both may be wildcard patterns.
2956 The file name of @samp{*} seen in many examples is a simple wildcard
2957 pattern for the file name.
2959 The wildcard patterns are like those used by the Unix shell.
2963 matches any number of characters
2965 matches any single character
2967 matches a single instance of any of the @var{chars}; the @samp{-}
2968 character may be used to specify a range of characters, as in
2969 @samp{[a-z]} to match any lower case letter
2971 quotes the following character
2974 When a file name is matched with a wildcard, the wildcard characters
2975 will not match a @samp{/} character (used to separate directory names on
2976 Unix). A pattern consisting of a single @samp{*} character is an
2977 exception; it will always match any file name, whether it contains a
2978 @samp{/} or not. In a section name, the wildcard characters will match
2979 a @samp{/} character.
2981 File name wildcard patterns only match files which are explicitly
2982 specified on the command line or in an @code{INPUT} command. The linker
2983 does not search directories to expand wildcards.
2985 If a file name matches more than one wildcard pattern, or if a file name
2986 appears explicitly and is also matched by a wildcard pattern, the linker
2987 will use the first match in the linker script. For example, this
2988 sequence of input section descriptions is probably in error, because the
2989 @file{data.o} rule will not be used:
2991 .data : @{ *(.data) @}
2992 .data1 : @{ data.o(.data) @}
2996 Normally, the linker will place files and sections matched by wildcards
2997 in the order in which they are seen during the link. You can change
2998 this by using the @code{SORT} keyword, which appears before a wildcard
2999 pattern in parentheses (e.g., @code{SORT(.text*)}). When the
3000 @code{SORT} keyword is used, the linker will sort the files or sections
3001 into ascending order by name before placing them in the output file.
3003 If you ever get confused about where input sections are going, use the
3004 @samp{-M} linker option to generate a map file. The map file shows
3005 precisely how input sections are mapped to output sections.
3007 This example shows how wildcard patterns might be used to partition
3008 files. This linker script directs the linker to place all @samp{.text}
3009 sections in @samp{.text} and all @samp{.bss} sections in @samp{.bss}.
3010 The linker will place the @samp{.data} section from all files beginning
3011 with an upper case character in @samp{.DATA}; for all other files, the
3012 linker will place the @samp{.data} section in @samp{.data}.
3016 .text : @{ *(.text) @}
3017 .DATA : @{ [A-Z]*(.data) @}
3018 .data : @{ *(.data) @}
3019 .bss : @{ *(.bss) @}
3024 @node Input Section Common
3025 @subsubsection Input Section for Common Symbols
3026 @cindex common symbol placement
3027 @cindex uninitialized data placement
3028 A special notation is needed for common symbols, because in many object
3029 file formats common symbols do not have a particular input section. The
3030 linker treats common symbols as though they are in an input section
3031 named @samp{COMMON}.
3033 You may use file names with the @samp{COMMON} section just as with any
3034 other input sections. You can use this to place common symbols from a
3035 particular input file in one section while common symbols from other
3036 input files are placed in another section.
3038 In most cases, common symbols in input files will be placed in the
3039 @samp{.bss} section in the output file. For example:
3041 .bss @{ *(.bss) *(COMMON) @}
3044 @cindex scommon section
3045 @cindex small common symbols
3046 Some object file formats have more than one type of common symbol. For
3047 example, the MIPS ELF object file format distinguishes standard common
3048 symbols and small common symbols. In this case, the linker will use a
3049 different special section name for other types of common symbols. In
3050 the case of MIPS ELF, the linker uses @samp{COMMON} for standard common
3051 symbols and @samp{.scommon} for small common symbols. This permits you
3052 to map the different types of common symbols into memory at different
3056 You will sometimes see @samp{[COMMON]} in old linker scripts. This
3057 notation is now considered obsolete. It is equivalent to
3060 @node Input Section Keep
3061 @subsubsection Input Section and Garbage Collection
3063 @cindex garbage collection
3064 When link-time garbage collection is in use (@samp{--gc-sections}),
3065 it is often useful to mark sections that should not be eliminated.
3066 This is accomplished by surrounding an input section's wildcard entry
3067 with @code{KEEP()}, as in @code{KEEP(*(.init))} or
3068 @code{KEEP(SORT(*)(.ctors))}.
3070 @node Input Section Example
3071 @subsubsection Input Section Example
3072 The following example is a complete linker script. It tells the linker
3073 to read all of the sections from file @file{all.o} and place them at the
3074 start of output section @samp{outputa} which starts at location
3075 @samp{0x10000}. All of section @samp{.input1} from file @file{foo.o}
3076 follows immediately, in the same output section. All of section
3077 @samp{.input2} from @file{foo.o} goes into output section
3078 @samp{outputb}, followed by section @samp{.input1} from @file{foo1.o}.
3079 All of the remaining @samp{.input1} and @samp{.input2} sections from any
3080 files are written to output section @samp{outputc}.
3108 @node Output Section Data
3109 @subsection Output Section Data
3111 @cindex section data
3112 @cindex output section data
3113 @kindex BYTE(@var{expression})
3114 @kindex SHORT(@var{expression})
3115 @kindex LONG(@var{expression})
3116 @kindex QUAD(@var{expression})
3117 @kindex SQUAD(@var{expression})
3118 You can include explicit bytes of data in an output section by using
3119 @code{BYTE}, @code{SHORT}, @code{LONG}, @code{QUAD}, or @code{SQUAD} as
3120 an output section command. Each keyword is followed by an expression in
3121 parentheses providing the value to store (@pxref{Expressions}). The
3122 value of the expression is stored at the current value of the location
3125 The @code{BYTE}, @code{SHORT}, @code{LONG}, and @code{QUAD} commands
3126 store one, two, four, and eight bytes (respectively). After storing the
3127 bytes, the location counter is incremented by the number of bytes
3130 For example, this will store the byte 1 followed by the four byte value
3131 of the symbol @samp{addr}:
3137 When using a 64 bit host or target, @code{QUAD} and @code{SQUAD} are the
3138 same; they both store an 8 byte, or 64 bit, value. When both host and
3139 target are 32 bits, an expression is computed as 32 bits. In this case
3140 @code{QUAD} stores a 32 bit value zero extended to 64 bits, and
3141 @code{SQUAD} stores a 32 bit value sign extended to 64 bits.
3143 If the object file format of the output file has an explicit endianness,
3144 which is the normal case, the value will be stored in that endianness.
3145 When the object file format does not have an explicit endianness, as is
3146 true of, for example, S-records, the value will be stored in the
3147 endianness of the first input object file.
3149 Note---these commands only work inside a section description and not
3150 between them, so the following will produce an error from the linker:
3152 SECTIONS @{@ .text : @{@ *(.text) @}@ LONG(1) .data : @{@ *(.data) @}@ @}@
3154 whereas this will work:
3156 SECTIONS @{@ .text : @{@ *(.text) ; LONG(1) @}@ .data : @{@ *(.data) @}@ @}@
3159 @kindex FILL(@var{expression})
3160 @cindex holes, filling
3161 @cindex unspecified memory
3162 You may use the @code{FILL} command to set the fill pattern for the
3163 current section. It is followed by an expression in parentheses. Any
3164 otherwise unspecified regions of memory within the section (for example,
3165 gaps left due to the required alignment of input sections) are filled
3166 with the value of the expression, repeated as
3167 necessary. A @code{FILL} statement covers memory locations after the
3168 point at which it occurs in the section definition; by including more
3169 than one @code{FILL} statement, you can have different fill patterns in
3170 different parts of an output section.
3172 This example shows how to fill unspecified regions of memory with the
3178 The @code{FILL} command is similar to the @samp{=@var{fillexp}} output
3179 section attribute, but it only affects the
3180 part of the section following the @code{FILL} command, rather than the
3181 entire section. If both are used, the @code{FILL} command takes
3182 precedence. @xref{Output Section Fill}, for details on the fill
3185 @node Output Section Keywords
3186 @subsection Output Section Keywords
3187 There are a couple of keywords which can appear as output section
3191 @kindex CREATE_OBJECT_SYMBOLS
3192 @cindex input filename symbols
3193 @cindex filename symbols
3194 @item CREATE_OBJECT_SYMBOLS
3195 The command tells the linker to create a symbol for each input file.
3196 The name of each symbol will be the name of the corresponding input
3197 file. The section of each symbol will be the output section in which
3198 the @code{CREATE_OBJECT_SYMBOLS} command appears.
3200 This is conventional for the a.out object file format. It is not
3201 normally used for any other object file format.
3203 @kindex CONSTRUCTORS
3204 @cindex C++ constructors, arranging in link
3205 @cindex constructors, arranging in link
3207 When linking using the a.out object file format, the linker uses an
3208 unusual set construct to support C++ global constructors and
3209 destructors. When linking object file formats which do not support
3210 arbitrary sections, such as ECOFF and XCOFF, the linker will
3211 automatically recognize C++ global constructors and destructors by name.
3212 For these object file formats, the @code{CONSTRUCTORS} command tells the
3213 linker to place constructor information in the output section where the
3214 @code{CONSTRUCTORS} command appears. The @code{CONSTRUCTORS} command is
3215 ignored for other object file formats.
3217 The symbol @w{@code{__CTOR_LIST__}} marks the start of the global
3218 constructors, and the symbol @w{@code{__DTOR_LIST}} marks the end. The
3219 first word in the list is the number of entries, followed by the address
3220 of each constructor or destructor, followed by a zero word. The
3221 compiler must arrange to actually run the code. For these object file
3222 formats @sc{gnu} C++ normally calls constructors from a subroutine
3223 @code{__main}; a call to @code{__main} is automatically inserted into
3224 the startup code for @code{main}. @sc{gnu} C++ normally runs
3225 destructors either by using @code{atexit}, or directly from the function
3228 For object file formats such as @code{COFF} or @code{ELF} which support
3229 arbitrary section names, @sc{gnu} C++ will normally arrange to put the
3230 addresses of global constructors and destructors into the @code{.ctors}
3231 and @code{.dtors} sections. Placing the following sequence into your
3232 linker script will build the sort of table which the @sc{gnu} C++
3233 runtime code expects to see.
3237 LONG((__CTOR_END__ - __CTOR_LIST__) / 4 - 2)
3242 LONG((__DTOR_END__ - __DTOR_LIST__) / 4 - 2)
3248 If you are using the @sc{gnu} C++ support for initialization priority,
3249 which provides some control over the order in which global constructors
3250 are run, you must sort the constructors at link time to ensure that they
3251 are executed in the correct order. When using the @code{CONSTRUCTORS}
3252 command, use @samp{SORT(CONSTRUCTORS)} instead. When using the
3253 @code{.ctors} and @code{.dtors} sections, use @samp{*(SORT(.ctors))} and
3254 @samp{*(SORT(.dtors))} instead of just @samp{*(.ctors)} and
3257 Normally the compiler and linker will handle these issues automatically,
3258 and you will not need to concern yourself with them. However, you may
3259 need to consider this if you are using C++ and writing your own linker
3264 @node Output Section Discarding
3265 @subsection Output Section Discarding
3266 @cindex discarding sections
3267 @cindex sections, discarding
3268 @cindex removing sections
3269 The linker will not create output section which do not have any
3270 contents. This is for convenience when referring to input sections that
3271 may or may not be present in any of the input files. For example:
3276 will only create a @samp{.foo} section in the output file if there is a
3277 @samp{.foo} section in at least one input file.
3279 If you use anything other than an input section description as an output
3280 section command, such as a symbol assignment, then the output section
3281 will always be created, even if there are no matching input sections.
3284 The special output section name @samp{/DISCARD/} may be used to discard
3285 input sections. Any input sections which are assigned to an output
3286 section named @samp{/DISCARD/} are not included in the output file.
3288 @node Output Section Attributes
3289 @subsection Output Section Attributes
3290 @cindex output section attributes
3291 We showed above that the full description of an output section looked
3295 @var{section} [@var{address}] [(@var{type})] :
3296 [AT(@var{lma})] [SUBALIGN(@var{subsection_align})]
3298 @var{output-section-command}
3299 @var{output-section-command}
3301 @} [>@var{region}] [AT>@var{lma_region}] [:@var{phdr} :@var{phdr} @dots{}] [=@var{fillexp}]
3304 We've already described @var{section}, @var{address}, and
3305 @var{output-section-command}. In this section we will describe the
3306 remaining section attributes.
3309 * Output Section Type:: Output section type
3310 * Output Section LMA:: Output section LMA
3311 * Forced Input Alignment:: Forced Input Alignment
3312 * Output Section Region:: Output section region
3313 * Output Section Phdr:: Output section phdr
3314 * Output Section Fill:: Output section fill
3317 @node Output Section Type
3318 @subsubsection Output Section Type
3319 Each output section may have a type. The type is a keyword in
3320 parentheses. The following types are defined:
3324 The section should be marked as not loadable, so that it will not be
3325 loaded into memory when the program is run.
3330 These type names are supported for backward compatibility, and are
3331 rarely used. They all have the same effect: the section should be
3332 marked as not allocatable, so that no memory is allocated for the
3333 section when the program is run.
3337 @cindex prevent unnecessary loading
3338 @cindex loading, preventing
3339 The linker normally sets the attributes of an output section based on
3340 the input sections which map into it. You can override this by using
3341 the section type. For example, in the script sample below, the
3342 @samp{ROM} section is addressed at memory location @samp{0} and does not
3343 need to be loaded when the program is run. The contents of the
3344 @samp{ROM} section will appear in the linker output file as usual.
3348 ROM 0 (NOLOAD) : @{ @dots{} @}
3354 @node Output Section LMA
3355 @subsubsection Output Section LMA
3356 @kindex AT>@var{lma_region}
3357 @kindex AT(@var{lma})
3358 @cindex load address
3359 @cindex section load address
3360 Every section has a virtual address (VMA) and a load address (LMA); see
3361 @ref{Basic Script Concepts}. The address expression which may appear in
3362 an output section description sets the VMA (@pxref{Output Section
3365 The linker will normally set the LMA equal to the VMA. You can change
3366 that by using the @code{AT} keyword. The expression @var{lma} that
3367 follows the @code{AT} keyword specifies the load address of the
3368 section. Alternatively, with @samp{AT>@var{lma_region}} expression,
3369 you may specify a memory region for the section's load address. @xref{MEMORY}.
3371 @cindex ROM initialized data
3372 @cindex initialized data in ROM
3373 This feature is designed to make it easy to build a ROM image. For
3374 example, the following linker script creates three output sections: one
3375 called @samp{.text}, which starts at @code{0x1000}, one called
3376 @samp{.mdata}, which is loaded at the end of the @samp{.text} section
3377 even though its VMA is @code{0x2000}, and one called @samp{.bss} to hold
3378 uninitialized data at address @code{0x3000}. The symbol @code{_data} is
3379 defined with the value @code{0x2000}, which shows that the location
3380 counter holds the VMA value, not the LMA value.
3386 .text 0x1000 : @{ *(.text) _etext = . ; @}
3388 AT ( ADDR (.text) + SIZEOF (.text) )
3389 @{ _data = . ; *(.data); _edata = . ; @}
3391 @{ _bstart = . ; *(.bss) *(COMMON) ; _bend = . ;@}
3396 The run-time initialization code for use with a program generated with
3397 this linker script would include something like the following, to copy
3398 the initialized data from the ROM image to its runtime address. Notice
3399 how this code takes advantage of the symbols defined by the linker
3404 extern char _etext, _data, _edata, _bstart, _bend;
3405 char *src = &_etext;
3408 /* ROM has data at end of text; copy it. */
3409 while (dst < &_edata) @{
3414 for (dst = &_bstart; dst< &_bend; dst++)
3419 @node Forced Input Alignment
3420 @subsubsection Forced Input Alignment
3421 @kindex SUBALIGN(@var{subsection_align})
3422 @cindex forcing input section alignment
3423 @cindex input section alignment
3424 You can force input section alignment within an output section by using
3425 SUBALIGN. The value specified overrides any alignment given by input
3426 sections, whether larger or smaller.
3428 @node Output Section Region
3429 @subsubsection Output Section Region
3430 @kindex >@var{region}
3431 @cindex section, assigning to memory region
3432 @cindex memory regions and sections
3433 You can assign a section to a previously defined region of memory by
3434 using @samp{>@var{region}}. @xref{MEMORY}.
3436 Here is a simple example:
3439 MEMORY @{ rom : ORIGIN = 0x1000, LENGTH = 0x1000 @}
3440 SECTIONS @{ ROM : @{ *(.text) @} >rom @}
3444 @node Output Section Phdr
3445 @subsubsection Output Section Phdr
3447 @cindex section, assigning to program header
3448 @cindex program headers and sections
3449 You can assign a section to a previously defined program segment by
3450 using @samp{:@var{phdr}}. @xref{PHDRS}. If a section is assigned to
3451 one or more segments, then all subsequent allocated sections will be
3452 assigned to those segments as well, unless they use an explicitly
3453 @code{:@var{phdr}} modifier. You can use @code{:NONE} to tell the
3454 linker to not put the section in any segment at all.
3456 Here is a simple example:
3459 PHDRS @{ text PT_LOAD ; @}
3460 SECTIONS @{ .text : @{ *(.text) @} :text @}
3464 @node Output Section Fill
3465 @subsubsection Output Section Fill
3466 @kindex =@var{fillexp}
3467 @cindex section fill pattern
3468 @cindex fill pattern, entire section
3469 You can set the fill pattern for an entire section by using
3470 @samp{=@var{fillexp}}. @var{fillexp} is an expression
3471 (@pxref{Expressions}). Any otherwise unspecified regions of memory
3472 within the output section (for example, gaps left due to the required
3473 alignment of input sections) will be filled with the value, repeated as
3474 necessary. If the fill expression is a simple hex number, ie. a string
3475 of hex digit starting with @samp{0x} and without a trailing @samp{k} or @samp{M}, then
3476 an arbitrarily long sequence of hex digits can be used to specify the
3477 fill pattern; Leading zeros become part of the pattern too. For all
3478 other cases, including extra parentheses or a unary @code{+}, the fill
3479 pattern is the four least significant bytes of the value of the
3480 expression. In all cases, the number is big-endian.
3482 You can also change the fill value with a @code{FILL} command in the
3483 output section commands; (@pxref{Output Section Data}).
3485 Here is a simple example:
3488 SECTIONS @{ .text : @{ *(.text) @} =0x90909090 @}
3492 @node Overlay Description
3493 @subsection Overlay Description
3496 An overlay description provides an easy way to describe sections which
3497 are to be loaded as part of a single memory image but are to be run at
3498 the same memory address. At run time, some sort of overlay manager will
3499 copy the overlaid sections in and out of the runtime memory address as
3500 required, perhaps by simply manipulating addressing bits. This approach
3501 can be useful, for example, when a certain region of memory is faster
3504 Overlays are described using the @code{OVERLAY} command. The
3505 @code{OVERLAY} command is used within a @code{SECTIONS} command, like an
3506 output section description. The full syntax of the @code{OVERLAY}
3507 command is as follows:
3510 OVERLAY [@var{start}] : [NOCROSSREFS] [AT ( @var{ldaddr} )]
3514 @var{output-section-command}
3515 @var{output-section-command}
3517 @} [:@var{phdr}@dots{}] [=@var{fill}]
3520 @var{output-section-command}
3521 @var{output-section-command}
3523 @} [:@var{phdr}@dots{}] [=@var{fill}]
3525 @} [>@var{region}] [:@var{phdr}@dots{}] [=@var{fill}]
3529 Everything is optional except @code{OVERLAY} (a keyword), and each
3530 section must have a name (@var{secname1} and @var{secname2} above). The
3531 section definitions within the @code{OVERLAY} construct are identical to
3532 those within the general @code{SECTIONS} contruct (@pxref{SECTIONS}),
3533 except that no addresses and no memory regions may be defined for
3534 sections within an @code{OVERLAY}.
3536 The sections are all defined with the same starting address. The load
3537 addresses of the sections are arranged such that they are consecutive in
3538 memory starting at the load address used for the @code{OVERLAY} as a
3539 whole (as with normal section definitions, the load address is optional,
3540 and defaults to the start address; the start address is also optional,
3541 and defaults to the current value of the location counter).
3543 If the @code{NOCROSSREFS} keyword is used, and there any references
3544 among the sections, the linker will report an error. Since the sections
3545 all run at the same address, it normally does not make sense for one
3546 section to refer directly to another. @xref{Miscellaneous Commands,
3549 For each section within the @code{OVERLAY}, the linker automatically
3550 defines two symbols. The symbol @code{__load_start_@var{secname}} is
3551 defined as the starting load address of the section. The symbol
3552 @code{__load_stop_@var{secname}} is defined as the final load address of
3553 the section. Any characters within @var{secname} which are not legal
3554 within C identifiers are removed. C (or assembler) code may use these
3555 symbols to move the overlaid sections around as necessary.
3557 At the end of the overlay, the value of the location counter is set to
3558 the start address of the overlay plus the size of the largest section.
3560 Here is an example. Remember that this would appear inside a
3561 @code{SECTIONS} construct.
3564 OVERLAY 0x1000 : AT (0x4000)
3566 .text0 @{ o1/*.o(.text) @}
3567 .text1 @{ o2/*.o(.text) @}
3572 This will define both @samp{.text0} and @samp{.text1} to start at
3573 address 0x1000. @samp{.text0} will be loaded at address 0x4000, and
3574 @samp{.text1} will be loaded immediately after @samp{.text0}. The
3575 following symbols will be defined: @code{__load_start_text0},
3576 @code{__load_stop_text0}, @code{__load_start_text1},
3577 @code{__load_stop_text1}.
3579 C code to copy overlay @code{.text1} into the overlay area might look
3584 extern char __load_start_text1, __load_stop_text1;
3585 memcpy ((char *) 0x1000, &__load_start_text1,
3586 &__load_stop_text1 - &__load_start_text1);
3590 Note that the @code{OVERLAY} command is just syntactic sugar, since
3591 everything it does can be done using the more basic commands. The above
3592 example could have been written identically as follows.
3596 .text0 0x1000 : AT (0x4000) @{ o1/*.o(.text) @}
3597 __load_start_text0 = LOADADDR (.text0);
3598 __load_stop_text0 = LOADADDR (.text0) + SIZEOF (.text0);
3599 .text1 0x1000 : AT (0x4000 + SIZEOF (.text0)) @{ o2/*.o(.text) @}
3600 __load_start_text1 = LOADADDR (.text1);
3601 __load_stop_text1 = LOADADDR (.text1) + SIZEOF (.text1);
3602 . = 0x1000 + MAX (SIZEOF (.text0), SIZEOF (.text1));
3607 @section MEMORY Command
3609 @cindex memory regions
3610 @cindex regions of memory
3611 @cindex allocating memory
3612 @cindex discontinuous memory
3613 The linker's default configuration permits allocation of all available
3614 memory. You can override this by using the @code{MEMORY} command.
3616 The @code{MEMORY} command describes the location and size of blocks of
3617 memory in the target. You can use it to describe which memory regions
3618 may be used by the linker, and which memory regions it must avoid. You
3619 can then assign sections to particular memory regions. The linker will
3620 set section addresses based on the memory regions, and will warn about
3621 regions that become too full. The linker will not shuffle sections
3622 around to fit into the available regions.
3624 A linker script may contain at most one use of the @code{MEMORY}
3625 command. However, you can define as many blocks of memory within it as
3626 you wish. The syntax is:
3631 @var{name} [(@var{attr})] : ORIGIN = @var{origin}, LENGTH = @var{len}
3637 The @var{name} is a name used in the linker script to refer to the
3638 region. The region name has no meaning outside of the linker script.
3639 Region names are stored in a separate name space, and will not conflict
3640 with symbol names, file names, or section names. Each memory region
3641 must have a distinct name.
3643 @cindex memory region attributes
3644 The @var{attr} string is an optional list of attributes that specify
3645 whether to use a particular memory region for an input section which is
3646 not explicitly mapped in the linker script. As described in
3647 @ref{SECTIONS}, if you do not specify an output section for some input
3648 section, the linker will create an output section with the same name as
3649 the input section. If you define region attributes, the linker will use
3650 them to select the memory region for the output section that it creates.
3652 The @var{attr} string must consist only of the following characters:
3667 Invert the sense of any of the preceding attributes
3670 If a unmapped section matches any of the listed attributes other than
3671 @samp{!}, it will be placed in the memory region. The @samp{!}
3672 attribute reverses this test, so that an unmapped section will be placed
3673 in the memory region only if it does not match any of the listed
3679 The @var{origin} is an expression for the start address of the memory
3680 region. The expression must evaluate to a constant before memory
3681 allocation is performed, which means that you may not use any section
3682 relative symbols. The keyword @code{ORIGIN} may be abbreviated to
3683 @code{org} or @code{o} (but not, for example, @code{ORG}).
3688 The @var{len} is an expression for the size in bytes of the memory
3689 region. As with the @var{origin} expression, the expression must
3690 evaluate to a constant before memory allocation is performed. The
3691 keyword @code{LENGTH} may be abbreviated to @code{len} or @code{l}.
3693 In the following example, we specify that there are two memory regions
3694 available for allocation: one starting at @samp{0} for 256 kilobytes,
3695 and the other starting at @samp{0x40000000} for four megabytes. The
3696 linker will place into the @samp{rom} memory region every section which
3697 is not explicitly mapped into a memory region, and is either read-only
3698 or executable. The linker will place other sections which are not
3699 explicitly mapped into a memory region into the @samp{ram} memory
3706 rom (rx) : ORIGIN = 0, LENGTH = 256K
3707 ram (!rx) : org = 0x40000000, l = 4M
3712 Once you define a memory region, you can direct the linker to place
3713 specific output sections into that memory region by using the
3714 @samp{>@var{region}} output section attribute. For example, if you have
3715 a memory region named @samp{mem}, you would use @samp{>mem} in the
3716 output section definition. @xref{Output Section Region}. If no address
3717 was specified for the output section, the linker will set the address to
3718 the next available address within the memory region. If the combined
3719 output sections directed to a memory region are too large for the
3720 region, the linker will issue an error message.
3723 @section PHDRS Command
3725 @cindex program headers
3726 @cindex ELF program headers
3727 @cindex program segments
3728 @cindex segments, ELF
3729 The ELF object file format uses @dfn{program headers}, also knows as
3730 @dfn{segments}. The program headers describe how the program should be
3731 loaded into memory. You can print them out by using the @code{objdump}
3732 program with the @samp{-p} option.
3734 When you run an ELF program on a native ELF system, the system loader
3735 reads the program headers in order to figure out how to load the
3736 program. This will only work if the program headers are set correctly.
3737 This manual does not describe the details of how the system loader
3738 interprets program headers; for more information, see the ELF ABI.
3740 The linker will create reasonable program headers by default. However,
3741 in some cases, you may need to specify the program headers more
3742 precisely. You may use the @code{PHDRS} command for this purpose. When
3743 the linker sees the @code{PHDRS} command in the linker script, it will
3744 not create any program headers other than the ones specified.
3746 The linker only pays attention to the @code{PHDRS} command when
3747 generating an ELF output file. In other cases, the linker will simply
3748 ignore @code{PHDRS}.
3750 This is the syntax of the @code{PHDRS} command. The words @code{PHDRS},
3751 @code{FILEHDR}, @code{AT}, and @code{FLAGS} are keywords.
3757 @var{name} @var{type} [ FILEHDR ] [ PHDRS ] [ AT ( @var{address} ) ]
3758 [ FLAGS ( @var{flags} ) ] ;
3763 The @var{name} is used only for reference in the @code{SECTIONS} command
3764 of the linker script. It is not put into the output file. Program
3765 header names are stored in a separate name space, and will not conflict
3766 with symbol names, file names, or section names. Each program header
3767 must have a distinct name.
3769 Certain program header types describe segments of memory which the
3770 system loader will load from the file. In the linker script, you
3771 specify the contents of these segments by placing allocatable output
3772 sections in the segments. You use the @samp{:@var{phdr}} output section
3773 attribute to place a section in a particular segment. @xref{Output
3776 It is normal to put certain sections in more than one segment. This
3777 merely implies that one segment of memory contains another. You may
3778 repeat @samp{:@var{phdr}}, using it once for each segment which should
3779 contain the section.
3781 If you place a section in one or more segments using @samp{:@var{phdr}},
3782 then the linker will place all subsequent allocatable sections which do
3783 not specify @samp{:@var{phdr}} in the same segments. This is for
3784 convenience, since generally a whole set of contiguous sections will be
3785 placed in a single segment. You can use @code{:NONE} to override the
3786 default segment and tell the linker to not put the section in any
3791 You may use the @code{FILEHDR} and @code{PHDRS} keywords appear after
3792 the program header type to further describe the contents of the segment.
3793 The @code{FILEHDR} keyword means that the segment should include the ELF
3794 file header. The @code{PHDRS} keyword means that the segment should
3795 include the ELF program headers themselves.
3797 The @var{type} may be one of the following. The numbers indicate the
3798 value of the keyword.
3801 @item @code{PT_NULL} (0)
3802 Indicates an unused program header.
3804 @item @code{PT_LOAD} (1)
3805 Indicates that this program header describes a segment to be loaded from
3808 @item @code{PT_DYNAMIC} (2)
3809 Indicates a segment where dynamic linking information can be found.
3811 @item @code{PT_INTERP} (3)
3812 Indicates a segment where the name of the program interpreter may be
3815 @item @code{PT_NOTE} (4)
3816 Indicates a segment holding note information.
3818 @item @code{PT_SHLIB} (5)
3819 A reserved program header type, defined but not specified by the ELF
3822 @item @code{PT_PHDR} (6)
3823 Indicates a segment where the program headers may be found.
3825 @item @var{expression}
3826 An expression giving the numeric type of the program header. This may
3827 be used for types not defined above.
3830 You can specify that a segment should be loaded at a particular address
3831 in memory by using an @code{AT} expression. This is identical to the
3832 @code{AT} command used as an output section attribute (@pxref{Output
3833 Section LMA}). The @code{AT} command for a program header overrides the
3834 output section attribute.
3836 The linker will normally set the segment flags based on the sections
3837 which comprise the segment. You may use the @code{FLAGS} keyword to
3838 explicitly specify the segment flags. The value of @var{flags} must be
3839 an integer. It is used to set the @code{p_flags} field of the program
3842 Here is an example of @code{PHDRS}. This shows a typical set of program
3843 headers used on a native ELF system.
3849 headers PT_PHDR PHDRS ;
3851 text PT_LOAD FILEHDR PHDRS ;
3853 dynamic PT_DYNAMIC ;
3859 .interp : @{ *(.interp) @} :text :interp
3860 .text : @{ *(.text) @} :text
3861 .rodata : @{ *(.rodata) @} /* defaults to :text */
3863 . = . + 0x1000; /* move to a new page in memory */
3864 .data : @{ *(.data) @} :data
3865 .dynamic : @{ *(.dynamic) @} :data :dynamic
3872 @section VERSION Command
3873 @kindex VERSION @{script text@}
3874 @cindex symbol versions
3875 @cindex version script
3876 @cindex versions of symbols
3877 The linker supports symbol versions when using ELF. Symbol versions are
3878 only useful when using shared libraries. The dynamic linker can use
3879 symbol versions to select a specific version of a function when it runs
3880 a program that may have been linked against an earlier version of the
3883 You can include a version script directly in the main linker script, or
3884 you can supply the version script as an implicit linker script. You can
3885 also use the @samp{--version-script} linker option.
3887 The syntax of the @code{VERSION} command is simply
3889 VERSION @{ version-script-commands @}
3892 The format of the version script commands is identical to that used by
3893 Sun's linker in Solaris 2.5. The version script defines a tree of
3894 version nodes. You specify the node names and interdependencies in the
3895 version script. You can specify which symbols are bound to which
3896 version nodes, and you can reduce a specified set of symbols to local
3897 scope so that they are not globally visible outside of the shared
3900 The easiest way to demonstrate the version script language is with a few
3922 This example version script defines three version nodes. The first
3923 version node defined is @samp{VERS_1.1}; it has no other dependencies.
3924 The script binds the symbol @samp{foo1} to @samp{VERS_1.1}. It reduces
3925 a number of symbols to local scope so that they are not visible outside
3926 of the shared library; this is done using wildcard patterns, so that any
3927 symbol whose name begins with @samp{old}, @samp{original}, or @samp{new}
3928 is matched. The wildcard patterns available are the same as those used
3929 in the shell when matching filenames (also known as ``globbing'').
3931 Next, the version script defines node @samp{VERS_1.2}. This node
3932 depends upon @samp{VERS_1.1}. The script binds the symbol @samp{foo2}
3933 to the version node @samp{VERS_1.2}.
3935 Finally, the version script defines node @samp{VERS_2.0}. This node
3936 depends upon @samp{VERS_1.2}. The scripts binds the symbols @samp{bar1}
3937 and @samp{bar2} are bound to the version node @samp{VERS_2.0}.
3939 When the linker finds a symbol defined in a library which is not
3940 specifically bound to a version node, it will effectively bind it to an
3941 unspecified base version of the library. You can bind all otherwise
3942 unspecified symbols to a given version node by using @samp{global: *;}
3943 somewhere in the version script.
3945 The names of the version nodes have no specific meaning other than what
3946 they might suggest to the person reading them. The @samp{2.0} version
3947 could just as well have appeared in between @samp{1.1} and @samp{1.2}.
3948 However, this would be a confusing way to write a version script.
3950 Node name can be omited, provided it is the only version node
3951 in the version script. Such version script doesn't assign any versions to
3952 symbols, only selects which symbols will be globally visible out and which
3956 @{ global: foo; bar; local: *; @};
3959 When you link an application against a shared library that has versioned
3960 symbols, the application itself knows which version of each symbol it
3961 requires, and it also knows which version nodes it needs from each
3962 shared library it is linked against. Thus at runtime, the dynamic
3963 loader can make a quick check to make sure that the libraries you have
3964 linked against do in fact supply all of the version nodes that the
3965 application will need to resolve all of the dynamic symbols. In this
3966 way it is possible for the dynamic linker to know with certainty that
3967 all external symbols that it needs will be resolvable without having to
3968 search for each symbol reference.
3970 The symbol versioning is in effect a much more sophisticated way of
3971 doing minor version checking that SunOS does. The fundamental problem
3972 that is being addressed here is that typically references to external
3973 functions are bound on an as-needed basis, and are not all bound when
3974 the application starts up. If a shared library is out of date, a
3975 required interface may be missing; when the application tries to use
3976 that interface, it may suddenly and unexpectedly fail. With symbol
3977 versioning, the user will get a warning when they start their program if
3978 the libraries being used with the application are too old.
3980 There are several GNU extensions to Sun's versioning approach. The
3981 first of these is the ability to bind a symbol to a version node in the
3982 source file where the symbol is defined instead of in the versioning
3983 script. This was done mainly to reduce the burden on the library
3984 maintainer. You can do this by putting something like:
3986 __asm__(".symver original_foo,foo@@VERS_1.1");
3989 in the C source file. This renames the function @samp{original_foo} to
3990 be an alias for @samp{foo} bound to the version node @samp{VERS_1.1}.
3991 The @samp{local:} directive can be used to prevent the symbol
3992 @samp{original_foo} from being exported. A @samp{.symver} directive
3993 takes precedence over a version script.
3995 The second GNU extension is to allow multiple versions of the same
3996 function to appear in a given shared library. In this way you can make
3997 an incompatible change to an interface without increasing the major
3998 version number of the shared library, while still allowing applications
3999 linked against the old interface to continue to function.
4001 To do this, you must use multiple @samp{.symver} directives in the
4002 source file. Here is an example:
4005 __asm__(".symver original_foo,foo@@");
4006 __asm__(".symver old_foo,foo@@VERS_1.1");
4007 __asm__(".symver old_foo1,foo@@VERS_1.2");
4008 __asm__(".symver new_foo,foo@@@@VERS_2.0");
4011 In this example, @samp{foo@@} represents the symbol @samp{foo} bound to the
4012 unspecified base version of the symbol. The source file that contains this
4013 example would define 4 C functions: @samp{original_foo}, @samp{old_foo},
4014 @samp{old_foo1}, and @samp{new_foo}.
4016 When you have multiple definitions of a given symbol, there needs to be
4017 some way to specify a default version to which external references to
4018 this symbol will be bound. You can do this with the
4019 @samp{foo@@@@VERS_2.0} type of @samp{.symver} directive. You can only
4020 declare one version of a symbol as the default in this manner; otherwise
4021 you would effectively have multiple definitions of the same symbol.
4023 If you wish to bind a reference to a specific version of the symbol
4024 within the shared library, you can use the aliases of convenience
4025 (i.e., @samp{old_foo}), or you can use the @samp{.symver} directive to
4026 specifically bind to an external version of the function in question.
4028 You can also specify the language in the version script:
4031 VERSION extern "lang" @{ version-script-commands @}
4034 The supported @samp{lang}s are @samp{C}, @samp{C++}, and @samp{Java}.
4035 The linker will iterate over the list of symbols at the link time and
4036 demangle them according to @samp{lang} before matching them to the
4037 patterns specified in @samp{version-script-commands}.
4040 @section Expressions in Linker Scripts
4043 The syntax for expressions in the linker script language is identical to
4044 that of C expressions. All expressions are evaluated as integers. All
4045 expressions are evaluated in the same size, which is 32 bits if both the
4046 host and target are 32 bits, and is otherwise 64 bits.
4048 You can use and set symbol values in expressions.
4050 The linker defines several special purpose builtin functions for use in
4054 * Constants:: Constants
4055 * Symbols:: Symbol Names
4056 * Location Counter:: The Location Counter
4057 * Operators:: Operators
4058 * Evaluation:: Evaluation
4059 * Expression Section:: The Section of an Expression
4060 * Builtin Functions:: Builtin Functions
4064 @subsection Constants
4065 @cindex integer notation
4066 @cindex constants in linker scripts
4067 All constants are integers.
4069 As in C, the linker considers an integer beginning with @samp{0} to be
4070 octal, and an integer beginning with @samp{0x} or @samp{0X} to be
4071 hexadecimal. The linker considers other integers to be decimal.
4073 @cindex scaled integers
4074 @cindex K and M integer suffixes
4075 @cindex M and K integer suffixes
4076 @cindex suffixes for integers
4077 @cindex integer suffixes
4078 In addition, you can use the suffixes @code{K} and @code{M} to scale a
4082 @c END TEXI2ROFF-KILL
4083 @code{1024} or @code{1024*1024}
4087 ${\rm 1024}$ or ${\rm 1024}^2$
4089 @c END TEXI2ROFF-KILL
4090 respectively. For example, the following all refer to the same quantity:
4098 @subsection Symbol Names
4099 @cindex symbol names
4101 @cindex quoted symbol names
4103 Unless quoted, symbol names start with a letter, underscore, or period
4104 and may include letters, digits, underscores, periods, and hyphens.
4105 Unquoted symbol names must not conflict with any keywords. You can
4106 specify a symbol which contains odd characters or has the same name as a
4107 keyword by surrounding the symbol name in double quotes:
4110 "with a space" = "also with a space" + 10;
4113 Since symbols can contain many non-alphabetic characters, it is safest
4114 to delimit symbols with spaces. For example, @samp{A-B} is one symbol,
4115 whereas @samp{A - B} is an expression involving subtraction.
4117 @node Location Counter
4118 @subsection The Location Counter
4121 @cindex location counter
4122 @cindex current output location
4123 The special linker variable @dfn{dot} @samp{.} always contains the
4124 current output location counter. Since the @code{.} always refers to a
4125 location in an output section, it may only appear in an expression
4126 within a @code{SECTIONS} command. The @code{.} symbol may appear
4127 anywhere that an ordinary symbol is allowed in an expression.
4130 Assigning a value to @code{.} will cause the location counter to be
4131 moved. This may be used to create holes in the output section. The
4132 location counter may never be moved backwards.
4148 In the previous example, the @samp{.text} section from @file{file1} is
4149 located at the beginning of the output section @samp{output}. It is
4150 followed by a 1000 byte gap. Then the @samp{.text} section from
4151 @file{file2} appears, also with a 1000 byte gap following before the
4152 @samp{.text} section from @file{file3}. The notation @samp{= 0x12345678}
4153 specifies what data to write in the gaps (@pxref{Output Section Fill}).
4155 @cindex dot inside sections
4156 Note: @code{.} actually refers to the byte offset from the start of the
4157 current containing object. Normally this is the @code{SECTIONS}
4158 statement, whose start address is 0, hence @code{.} can be used as an
4159 absolute address. If @code{.} is used inside a section description
4160 however, it refers to the byte offset from the start of that section,
4161 not an absolute address. Thus in a script like this:
4179 The @samp{.text} section will be assigned a starting address of 0x100
4180 and a size of exactly 0x200 bytes, even if there is not enough data in
4181 the @samp{.text} input sections to fill this area. (If there is too
4182 much data, an error will be produced because this would be an attempt to
4183 move @code{.} backwards). The @samp{.data} section will start at 0x500
4184 and it will have an extra 0x600 bytes worth of space after the end of
4185 the values from the @samp{.data} input sections and before the end of
4186 the @samp{.data} output section itself.
4190 @subsection Operators
4191 @cindex operators for arithmetic
4192 @cindex arithmetic operators
4193 @cindex precedence in expressions
4194 The linker recognizes the standard C set of arithmetic operators, with
4195 the standard bindings and precedence levels:
4198 @c END TEXI2ROFF-KILL
4200 precedence associativity Operators Notes
4206 5 left == != > < <= >=
4212 11 right &= += -= *= /= (2)
4216 (1) Prefix operators
4217 (2) @xref{Assignments}.
4221 \vskip \baselineskip
4222 %"lispnarrowing" is the extra indent used generally for smallexample
4223 \hskip\lispnarrowing\vbox{\offinterlineskip
4226 {\vrule#&\strut\hfil\ #\ \hfil&\vrule#&\strut\hfil\ #\ \hfil&\vrule#&\strut\hfil\ {\tt #}\ \hfil&\vrule#\cr
4227 height2pt&\omit&&\omit&&\omit&\cr
4228 &Precedence&& Associativity &&{\rm Operators}&\cr
4229 height2pt&\omit&&\omit&&\omit&\cr
4231 height2pt&\omit&&\omit&&\omit&\cr
4233 % '176 is tilde, '~' in tt font
4234 &1&&left&&\qquad- \char'176\ !\qquad\dag&\cr
4235 &2&&left&&* / \%&\cr
4238 &5&&left&&== != > < <= >=&\cr
4241 &8&&left&&{\&\&}&\cr
4244 &11&&right&&\qquad\&= += -= *= /=\qquad\ddag&\cr
4246 height2pt&\omit&&\omit&&\omit&\cr}
4251 @obeylines@parskip=0pt@parindent=0pt
4252 @dag@quad Prefix operators.
4253 @ddag@quad @xref{Assignments}.
4256 @c END TEXI2ROFF-KILL
4259 @subsection Evaluation
4260 @cindex lazy evaluation
4261 @cindex expression evaluation order
4262 The linker evaluates expressions lazily. It only computes the value of
4263 an expression when absolutely necessary.
4265 The linker needs some information, such as the value of the start
4266 address of the first section, and the origins and lengths of memory
4267 regions, in order to do any linking at all. These values are computed
4268 as soon as possible when the linker reads in the linker script.
4270 However, other values (such as symbol values) are not known or needed
4271 until after storage allocation. Such values are evaluated later, when
4272 other information (such as the sizes of output sections) is available
4273 for use in the symbol assignment expression.
4275 The sizes of sections cannot be known until after allocation, so
4276 assignments dependent upon these are not performed until after
4279 Some expressions, such as those depending upon the location counter
4280 @samp{.}, must be evaluated during section allocation.
4282 If the result of an expression is required, but the value is not
4283 available, then an error results. For example, a script like the
4289 .text 9+this_isnt_constant :
4295 will cause the error message @samp{non constant expression for initial
4298 @node Expression Section
4299 @subsection The Section of an Expression
4300 @cindex expression sections
4301 @cindex absolute expressions
4302 @cindex relative expressions
4303 @cindex absolute and relocatable symbols
4304 @cindex relocatable and absolute symbols
4305 @cindex symbols, relocatable and absolute
4306 When the linker evaluates an expression, the result is either absolute
4307 or relative to some section. A relative expression is expressed as a
4308 fixed offset from the base of a section.
4310 The position of the expression within the linker script determines
4311 whether it is absolute or relative. An expression which appears within
4312 an output section definition is relative to the base of the output
4313 section. An expression which appears elsewhere will be absolute.
4315 A symbol set to a relative expression will be relocatable if you request
4316 relocatable output using the @samp{-r} option. That means that a
4317 further link operation may change the value of the symbol. The symbol's
4318 section will be the section of the relative expression.
4320 A symbol set to an absolute expression will retain the same value
4321 through any further link operation. The symbol will be absolute, and
4322 will not have any particular associated section.
4324 You can use the builtin function @code{ABSOLUTE} to force an expression
4325 to be absolute when it would otherwise be relative. For example, to
4326 create an absolute symbol set to the address of the end of the output
4327 section @samp{.data}:
4331 .data : @{ *(.data) _edata = ABSOLUTE(.); @}
4335 If @samp{ABSOLUTE} were not used, @samp{_edata} would be relative to the
4336 @samp{.data} section.
4338 @node Builtin Functions
4339 @subsection Builtin Functions
4340 @cindex functions in expressions
4341 The linker script language includes a number of builtin functions for
4342 use in linker script expressions.
4345 @item ABSOLUTE(@var{exp})
4346 @kindex ABSOLUTE(@var{exp})
4347 @cindex expression, absolute
4348 Return the absolute (non-relocatable, as opposed to non-negative) value
4349 of the expression @var{exp}. Primarily useful to assign an absolute
4350 value to a symbol within a section definition, where symbol values are
4351 normally section relative. @xref{Expression Section}.
4353 @item ADDR(@var{section})
4354 @kindex ADDR(@var{section})
4355 @cindex section address in expression
4356 Return the absolute address (the VMA) of the named @var{section}. Your
4357 script must previously have defined the location of that section. In
4358 the following example, @code{symbol_1} and @code{symbol_2} are assigned
4365 start_of_output_1 = ABSOLUTE(.);
4370 symbol_1 = ADDR(.output1);
4371 symbol_2 = start_of_output_1;
4377 @item ALIGN(@var{exp})
4378 @kindex ALIGN(@var{exp})
4379 @cindex round up location counter
4380 @cindex align location counter
4381 Return the location counter (@code{.}) aligned to the next @var{exp}
4383 @code{ALIGN} doesn't change the value of the location counter---it just
4384 does arithmetic on it. Here is an example which aligns the output
4385 @code{.data} section to the next @code{0x2000} byte boundary after the
4386 preceding section and sets a variable within the section to the next
4387 @code{0x8000} boundary after the input sections:
4391 .data ALIGN(0x2000): @{
4393 variable = ALIGN(0x8000);
4399 The first use of @code{ALIGN} in this example specifies the location of
4400 a section because it is used as the optional @var{address} attribute of
4401 a section definition (@pxref{Output Section Address}). The second use
4402 of @code{ALIGN} is used to defines the value of a symbol.
4404 The builtin function @code{NEXT} is closely related to @code{ALIGN}.
4406 @item BLOCK(@var{exp})
4407 @kindex BLOCK(@var{exp})
4408 This is a synonym for @code{ALIGN}, for compatibility with older linker
4409 scripts. It is most often seen when setting the address of an output
4412 @item DATA_SEGMENT_ALIGN(@var{maxpagesize}, @var{commonpagesize})
4413 @kindex DATA_SEGMENT_ALIGN(@var{maxpagesize}, @var{commonpagesize})
4414 This is equivalent to either
4416 (ALIGN(@var{maxpagesize}) + (. & (@var{maxpagesize} - 1)))
4420 (ALIGN(@var{maxpagesize}) + (. & (@var{maxpagesize} - @var{commonpagesize})))
4423 depending on whether the latter uses fewer @var{commonpagesize} sized pages
4424 for the data segment (area between the result of this expression and
4425 @code{DATA_SEGMENT_END}) than the former or not.
4426 If the latter form is used, it means @var{commonpagesize} bytes of runtime
4427 memory will be saved at the expense of up to @var{commonpagesize} wasted
4428 bytes in the on-disk file.
4430 This expression can only be used directly in @code{SECTIONS} commands, not in
4431 any output section descriptions and only once in the linker script.
4432 @var{commonpagesize} should be less or equal to @var{maxpagesize} and should
4433 be the system page size the object wants to be optimized for (while still
4434 working on system page sizes up to @var{maxpagesize}).
4439 . = DATA_SEGMENT_ALIGN(0x10000, 0x2000);
4442 @item DATA_SEGMENT_END(@var{exp})
4443 @kindex DATA_SEGMENT_END(@var{exp})
4444 This defines the end of data segment for @code{DATA_SEGMENT_ALIGN}
4445 evaluation purposes.
4448 . = DATA_SEGMENT_END(.);
4451 @item DEFINED(@var{symbol})
4452 @kindex DEFINED(@var{symbol})
4453 @cindex symbol defaults
4454 Return 1 if @var{symbol} is in the linker global symbol table and is
4455 defined before the statement using DEFINED in the script, otherwise
4456 return 0. You can use this function to provide
4457 default values for symbols. For example, the following script fragment
4458 shows how to set a global symbol @samp{begin} to the first location in
4459 the @samp{.text} section---but if a symbol called @samp{begin} already
4460 existed, its value is preserved:
4466 begin = DEFINED(begin) ? begin : . ;
4474 @item LOADADDR(@var{section})
4475 @kindex LOADADDR(@var{section})
4476 @cindex section load address in expression
4477 Return the absolute LMA of the named @var{section}. This is normally
4478 the same as @code{ADDR}, but it may be different if the @code{AT}
4479 attribute is used in the output section definition (@pxref{Output
4483 @item MAX(@var{exp1}, @var{exp2})
4484 Returns the maximum of @var{exp1} and @var{exp2}.
4487 @item MIN(@var{exp1}, @var{exp2})
4488 Returns the minimum of @var{exp1} and @var{exp2}.
4490 @item NEXT(@var{exp})
4491 @kindex NEXT(@var{exp})
4492 @cindex unallocated address, next
4493 Return the next unallocated address that is a multiple of @var{exp}.
4494 This function is closely related to @code{ALIGN(@var{exp})}; unless you
4495 use the @code{MEMORY} command to define discontinuous memory for the
4496 output file, the two functions are equivalent.
4498 @item SIZEOF(@var{section})
4499 @kindex SIZEOF(@var{section})
4500 @cindex section size
4501 Return the size in bytes of the named @var{section}, if that section has
4502 been allocated. If the section has not been allocated when this is
4503 evaluated, the linker will report an error. In the following example,
4504 @code{symbol_1} and @code{symbol_2} are assigned identical values:
4513 symbol_1 = .end - .start ;
4514 symbol_2 = SIZEOF(.output);
4519 @item SIZEOF_HEADERS
4520 @itemx sizeof_headers
4521 @kindex SIZEOF_HEADERS
4523 Return the size in bytes of the output file's headers. This is
4524 information which appears at the start of the output file. You can use
4525 this number when setting the start address of the first section, if you
4526 choose, to facilitate paging.
4528 @cindex not enough room for program headers
4529 @cindex program headers, not enough room
4530 When producing an ELF output file, if the linker script uses the
4531 @code{SIZEOF_HEADERS} builtin function, the linker must compute the
4532 number of program headers before it has determined all the section
4533 addresses and sizes. If the linker later discovers that it needs
4534 additional program headers, it will report an error @samp{not enough
4535 room for program headers}. To avoid this error, you must avoid using
4536 the @code{SIZEOF_HEADERS} function, or you must rework your linker
4537 script to avoid forcing the linker to use additional program headers, or
4538 you must define the program headers yourself using the @code{PHDRS}
4539 command (@pxref{PHDRS}).
4542 @node Implicit Linker Scripts
4543 @section Implicit Linker Scripts
4544 @cindex implicit linker scripts
4545 If you specify a linker input file which the linker can not recognize as
4546 an object file or an archive file, it will try to read the file as a
4547 linker script. If the file can not be parsed as a linker script, the
4548 linker will report an error.
4550 An implicit linker script will not replace the default linker script.
4552 Typically an implicit linker script would contain only symbol
4553 assignments, or the @code{INPUT}, @code{GROUP}, or @code{VERSION}
4556 Any input files read because of an implicit linker script will be read
4557 at the position in the command line where the implicit linker script was
4558 read. This can affect archive searching.
4561 @node Machine Dependent
4562 @chapter Machine Dependent Features
4564 @cindex machine dependencies
4565 @command{ld} has additional features on some platforms; the following
4566 sections describe them. Machines where @command{ld} has no additional
4567 functionality are not listed.
4571 * H8/300:: @command{ld} and the H8/300
4574 * i960:: @command{ld} and the Intel 960 family
4577 * ARM:: @command{ld} and the ARM family
4580 * HPPA ELF32:: @command{ld} and HPPA 32-bit ELF
4583 * MMIX:: @command{ld} and MMIX
4586 * MSP430:: @command{ld} and MSP430
4589 * TI COFF:: @command{ld} and TI COFF
4592 * WIN32:: @command{ld} and WIN32 (cygwin/mingw)
4595 * Xtensa:: @command{ld} and Xtensa Processors
4606 @section @command{ld} and the H8/300
4608 @cindex H8/300 support
4609 For the H8/300, @command{ld} can perform these global optimizations when
4610 you specify the @samp{--relax} command-line option.
4613 @cindex relaxing on H8/300
4614 @item relaxing address modes
4615 @command{ld} finds all @code{jsr} and @code{jmp} instructions whose
4616 targets are within eight bits, and turns them into eight-bit
4617 program-counter relative @code{bsr} and @code{bra} instructions,
4620 @cindex synthesizing on H8/300
4621 @item synthesizing instructions
4622 @c FIXME: specifically mov.b, or any mov instructions really?
4623 @command{ld} finds all @code{mov.b} instructions which use the
4624 sixteen-bit absolute address form, but refer to the top
4625 page of memory, and changes them to use the eight-bit address form.
4626 (That is: the linker turns @samp{mov.b @code{@@}@var{aa}:16} into
4627 @samp{mov.b @code{@@}@var{aa}:8} whenever the address @var{aa} is in the
4628 top page of memory).
4638 @c This stuff is pointless to say unless you're especially concerned
4639 @c with Renesas chips; don't enable it for generic case, please.
4641 @chapter @command{ld} and Other Renesas Chips
4643 @command{ld} also supports the Renesas (formerly Hitachi) H8/300H,
4644 H8/500, and SH chips. No special features, commands, or command-line
4645 options are required for these chips.
4655 @section @command{ld} and the Intel 960 Family
4657 @cindex i960 support
4659 You can use the @samp{-A@var{architecture}} command line option to
4660 specify one of the two-letter names identifying members of the 960
4661 family; the option specifies the desired output target, and warns of any
4662 incompatible instructions in the input files. It also modifies the
4663 linker's search strategy for archive libraries, to support the use of
4664 libraries specific to each particular architecture, by including in the
4665 search loop names suffixed with the string identifying the architecture.
4667 For example, if your @command{ld} command line included @w{@samp{-ACA}} as
4668 well as @w{@samp{-ltry}}, the linker would look (in its built-in search
4669 paths, and in any paths you specify with @samp{-L}) for a library with
4682 The first two possibilities would be considered in any event; the last
4683 two are due to the use of @w{@samp{-ACA}}.
4685 You can meaningfully use @samp{-A} more than once on a command line, since
4686 the 960 architecture family allows combination of target architectures; each
4687 use will add another pair of name variants to search for when @w{@samp{-l}}
4688 specifies a library.
4690 @cindex @option{--relax} on i960
4691 @cindex relaxing on i960
4692 @command{ld} supports the @samp{--relax} option for the i960 family. If
4693 you specify @samp{--relax}, @command{ld} finds all @code{balx} and
4694 @code{calx} instructions whose targets are within 24 bits, and turns
4695 them into 24-bit program-counter relative @code{bal} and @code{cal}
4696 instructions, respectively. @command{ld} also turns @code{cal}
4697 instructions into @code{bal} instructions when it determines that the
4698 target subroutine is a leaf routine (that is, the target subroutine does
4699 not itself call any subroutines).
4712 @section @command{ld}'s Support for Interworking Between ARM and Thumb Code
4714 @cindex ARM interworking support
4715 @kindex --support-old-code
4716 For the ARM, @command{ld} will generate code stubs to allow functions calls
4717 betweem ARM and Thumb code. These stubs only work with code that has
4718 been compiled and assembled with the @samp{-mthumb-interwork} command
4719 line option. If it is necessary to link with old ARM object files or
4720 libraries, which have not been compiled with the -mthumb-interwork
4721 option then the @samp{--support-old-code} command line switch should be
4722 given to the linker. This will make it generate larger stub functions
4723 which will work with non-interworking aware ARM code. Note, however,
4724 the linker does not support generating stubs for function calls to
4725 non-interworking aware Thumb code.
4727 @cindex thumb entry point
4728 @cindex entry point, thumb
4729 @kindex --thumb-entry=@var{entry}
4730 The @samp{--thumb-entry} switch is a duplicate of the generic
4731 @samp{--entry} switch, in that it sets the program's starting address.
4732 But it also sets the bottom bit of the address, so that it can be
4733 branched to using a BX instruction, and the program will start
4734 executing in Thumb mode straight away.
4747 @section @command{ld} and HPPA 32-bit ELF Support
4748 @cindex HPPA multiple sub-space stubs
4749 @kindex --multi-subspace
4750 When generating a shared library, @command{ld} will by default generate
4751 import stubs suitable for use with a single sub-space application.
4752 The @samp{--multi-subspace} switch causes @command{ld} to generate export
4753 stubs, and different (larger) import stubs suitable for use with
4754 multiple sub-spaces.
4756 @cindex HPPA stub grouping
4757 @kindex --stub-group-size=@var{N}
4758 Long branch stubs and import/export stubs are placed by @command{ld} in
4759 stub sections located between groups of input sections.
4760 @samp{--stub-group-size} specifies the maximum size of a group of input
4761 sections handled by one stub section. Since branch offsets are signed,
4762 a stub section may serve two groups of input sections, one group before
4763 the stub section, and one group after it. However, when using
4764 conditional branches that require stubs, it may be better (for branch
4765 prediction) that stub sections only serve one group of input sections.
4766 A negative value for @samp{N} chooses this scheme, ensuring that
4767 branches to stubs always use a negative offset. Two special values of
4768 @samp{N} are recognized, @samp{1} and @samp{-1}. These both instruct
4769 @command{ld} to automatically size input section groups for the branch types
4770 detected, with the same behaviour regarding stub placement as other
4771 positive or negative values of @samp{N} respectively.
4773 Note that @samp{--stub-group-size} does not split input sections. A
4774 single input section larger than the group size specified will of course
4775 create a larger group (of one section). If input sections are too
4776 large, it may not be possible for a branch to reach its stub.
4789 @section @code{ld} and MMIX
4790 For MMIX, there is a choice of generating @code{ELF} object files or
4791 @code{mmo} object files when linking. The simulator @code{mmix}
4792 understands the @code{mmo} format. The binutils @code{objcopy} utility
4793 can translate between the two formats.
4795 There is one special section, the @samp{.MMIX.reg_contents} section.
4796 Contents in this section is assumed to correspond to that of global
4797 registers, and symbols referring to it are translated to special symbols,
4798 equal to registers. In a final link, the start address of the
4799 @samp{.MMIX.reg_contents} section corresponds to the first allocated
4800 global register multiplied by 8. Register @code{$255} is not included in
4801 this section; it is always set to the program entry, which is at the
4802 symbol @code{Main} for @code{mmo} files.
4804 Symbols with the prefix @code{__.MMIX.start.}, for example
4805 @code{__.MMIX.start..text} and @code{__.MMIX.start..data} are special;
4806 there must be only one each, even if they are local. The default linker
4807 script uses these to set the default start address of a section.
4809 Initial and trailing multiples of zero-valued 32-bit words in a section,
4810 are left out from an mmo file.
4823 @section @code{ld} and MSP430
4824 For the MSP430 it is possible to select the MPU architecture. The flag @samp{-m [mpu type]}
4825 will select an appropriate linker script for selected MPU type. (To get a list of known MPUs
4826 just pass @samp{-m help} option to the linker).
4828 @cindex MSP430 extra sections
4829 The linker will recognize some extra sections which are MSP430 specific:
4832 @item @samp{.vectors}
4833 Defines a portion of ROM where interrupt vectors located.
4835 @item @samp{.bootloader}
4836 Defines the bootloader portion of the ROM (if applicable). Any code
4837 in this section will be uploaded to the MPU.
4839 @item @samp{.infomem}
4840 Defines an information memory section (if applicable). Any code in
4841 this section will be uploaded to the MPU.
4843 @item @samp{.infomemnobits}
4844 This is the same as the @samp{.infomem} section except that any code
4845 in this section will not be uploaded to the MPU.
4847 @item @samp{.noinit}
4848 Denotes a portion of RAM located above @samp{.bss} section.
4850 The last two sections are used by gcc.
4864 @section @command{ld}'s Support for Various TI COFF Versions
4865 @cindex TI COFF versions
4866 @kindex --format=@var{version}
4867 The @samp{--format} switch allows selection of one of the various
4868 TI COFF versions. The latest of this writing is 2; versions 0 and 1 are
4869 also supported. The TI COFF versions also vary in header byte-order
4870 format; @command{ld} will read any version or byte order, but the output
4871 header format depends on the default specified by the specific target.
4884 @section @command{ld} and WIN32 (cygwin/mingw)
4886 This section describes some of the win32 specific @command{ld} issues.
4887 See @ref{Options,,Command Line Options} for detailed decription of the
4888 command line options mentioned here.
4891 @cindex import libraries
4892 @item import libraries
4893 The standard Windows linker creates and uses so-called import
4894 libraries, which contains information for linking to dll's. They are
4895 regular static archives and are handled as any other static
4896 archive. The cygwin and mingw ports of @command{ld} have specific
4897 support for creating such libraries provided with the
4898 @samp{--out-implib} command line option.
4900 @item exporting DLL symbols
4901 @cindex exporting DLL symbols
4902 The cygwin/mingw @command{ld} has several ways to export symbols for dll's.
4905 @item using auto-export functionality
4906 @cindex using auto-export functionality
4907 By default @command{ld} exports symbols with the auto-export functionality,
4908 which is controlled by the following command line options:
4911 @item --export-all-symbols [This is the default]
4912 @item --exclude-symbols
4913 @item --exclude-libs
4916 If, however, @samp{--export-all-symbols} is not given explicitly on the
4917 command line, then the default auto-export behavior will be @emph{disabled}
4918 if either of the following are true:
4921 @item A DEF file is used.
4922 @item Any symbol in any object file was marked with the __declspec(dllexport) attribute.
4925 @item using a DEF file
4926 @cindex using a DEF file
4927 Another way of exporting symbols is using a DEF file. A DEF file is
4928 an ASCII file containing definitions of symbols which should be
4929 exported when a dll is created. Usually it is named @samp{<dll
4930 name>.def} and is added as any other object file to the linker's
4931 command line. The file's name must end in @samp{.def} or @samp{.DEF}.
4934 gcc -o <output> <objectfiles> <dll name>.def
4937 Using a DEF file turns off the normal auto-export behavior, unless the
4938 @samp{--export-all-symbols} option is also used.
4940 Here is an example of a DEF file for a shared library called @samp{xyz.dll}:
4943 LIBRARY "xyz.dll" BASE=0x10000000
4951 This example defines a base address and three symbols. The third
4952 symbol is an alias for the second. For the complete format
4953 specification see ld/deffilep.y in the binutils sources.
4955 @cindex creating a DEF file
4956 While linking a shared dll, @command{ld} is able to create a DEF file
4957 with the @samp{--output-def <file>} command line option.
4959 @item Using decorations
4960 @cindex Using decorations
4961 Another way of marking symbols for export is to modify the source code
4962 itself, so that when building the DLL each symbol to be exported is
4966 __declspec(dllexport) int a_variable
4967 __declspec(dllexport) void a_function(int with_args)
4970 All such symbols will be exported from the DLL. If, however,
4971 any of the object files in the DLL contain symbols decorated in
4972 this way, then the normal auto-export behavior is disabled, unless
4973 the @samp{--export-all-symbols} option is also used.
4975 Note that object files that wish to access these symbols must @emph{not}
4976 decorate them with dllexport. Instead, they should use dllimport,
4980 __declspec(dllimport) int a_variable
4981 __declspec(dllimport) void a_function(int with_args)
4984 This complicates the structure of library header files, because
4985 when included by the library itself the header must declare the
4986 variables and functions as dllexport, but when included by client
4987 code the header must declare them as dllimport. There are a number
4988 of idioms that are typically used to do this; often client code can
4989 omit the __declspec() declaration completely. See
4990 @samp{--enable-auto-import} and @samp{automatic data imports} for more
4994 @cindex automatic data imports
4995 @item automatic data imports
4996 The standard Windows dll format supports data imports from dlls only
4997 by adding special decorations (dllimport/dllexport), which let the
4998 compiler produce specific assembler instructions to deal with this
4999 issue. This increases the effort necessary to port existing Un*x
5000 code to these platforms, especially for large
5001 c++ libraries and applications. The auto-import feature, which was
5002 initially provided by Paul Sokolovsky, allows one to omit the
5003 decorations to archieve a behavior that conforms to that on POSIX/Un*x
5004 platforms. This feature is enabled with the @samp{--enable-auto-import}
5005 command-line option, although it is enabled by default on cygwin/mingw.
5006 The @samp{--enable-auto-import} option itself now serves mainly to
5007 suppress any warnings that are ordinarily emitted when linked objects
5008 trigger the feature's use.
5010 auto-import of variables does not always work flawlessly without
5011 additional assistance. Sometimes, you will see this message
5013 "variable '<var>' can't be auto-imported. Please read the
5014 documentation for ld's @code{--enable-auto-import} for details."
5016 The @samp{--enable-auto-import} documentation explains why this error
5017 occurs, and several methods that can be used to overcome this difficulty.
5018 One of these methods is the @emph{runtime pseudo-relocs} feature, described
5021 @cindex runtime pseudo-relocation
5022 For complex variables imported from DLLs (such as structs or classes),
5023 object files typically contain a base address for the variable and an
5024 offset (@emph{addend}) within the variable--to specify a particular
5025 field or public member, for instance. Unfortunately, the runtime loader used
5026 in win32 environments is incapable of fixing these references at runtime
5027 without the additional information supplied by dllimport/dllexport decorations.
5028 The standard auto-import feature described above is unable to resolve these
5031 The @samp{--enable-runtime-pseudo-relocs} switch allows these references to
5032 be resolved without error, while leaving the task of adjusting the references
5033 themselves (with their non-zero addends) to specialized code provided by the
5034 runtime environment. Recent versions of the cygwin and mingw environments and
5035 compilers provide this runtime support; older versions do not. However, the
5036 support is only necessary on the developer's platform; the compiled result will
5037 run without error on an older system.
5039 @samp{--enable-runtime-pseudo-relocs} is not the default; it must be explicitly
5042 @cindex direct linking to a dll
5043 @item direct linking to a dll
5044 The cygwin/mingw ports of @command{ld} support the direct linking,
5045 including data symbols, to a dll without the usage of any import
5046 libraries. This is much faster and uses much less memory than does the
5047 traditional import library method, expecially when linking large
5048 libraries or applications. When @command{ld} creates an import lib, each
5049 function or variable exported from the dll is stored in its own bfd, even
5050 though a single bfd could contain many exports. The overhead involved in
5051 storing, loading, and processing so many bfd's is quite large, and explains the
5052 tremendous time, memory, and storage needed to link against particularly
5053 large or complex libraries when using import libs.
5055 Linking directly to a dll uses no extra command-line switches other than
5056 @samp{-L} and @samp{-l}, because @command{ld} already searches for a number
5057 of names to match each library. All that is needed from the developer's
5058 perspective is an understanding of this search, in order to force ld to
5059 select the dll instead of an import library.
5062 For instance, when ld is called with the argument @samp{-lxxx} it will attempt
5063 to find, in the first directory of its search path,
5074 before moving on to the next directory in the search path.
5076 (*) Actually, this is not @samp{cygxxx.dll} but in fact is @samp{<prefix>xxx.dll},
5077 where @samp{<prefix>} is set by the @command{ld} option
5078 @samp{--dll-search-prefix=<prefix>}. In the case of cygwin, the standard gcc spec
5079 file includes @samp{--dll-search-prefix=cyg}, so in effect we actually search for
5082 Other win32-based unix environments, such as mingw or pw32, may use other
5083 @samp{<prefix>}es, although at present only cygwin makes use of this feature. It
5084 was originally intended to help avoid name conflicts among dll's built for the
5085 various win32/un*x environments, so that (for example) two versions of a zlib dll
5086 could coexist on the same machine.
5088 The generic cygwin/mingw path layout uses a @samp{bin} directory for
5089 applications and dll's and a @samp{lib} directory for the import
5090 libraries (using cygwin nomenclature):
5096 libxxx.dll.a (in case of dll's)
5097 libxxx.a (in case of static archive)
5100 Linking directly to a dll without using the import library can be
5103 1. Use the dll directly by adding the @samp{bin} path to the link line
5105 gcc -Wl,-verbose -o a.exe -L../bin/ -lxxx
5108 However, as the dll's often have version numbers appended to their names
5109 (@samp{cygncurses-5.dll}) this will often fail, unless one specifies
5110 @samp{-L../bin -lncurses-5} to include the version. Import libs are generally
5111 not versioned, and do not have this difficulty.
5113 2. Create a symbolic link from the dll to a file in the @samp{lib}
5114 directory according to the above mentioned search pattern. This
5115 should be used to avoid unwanted changes in the tools needed for
5119 ln -s bin/cygxxx.dll lib/[cyg|lib|]xxx.dll[.a]
5122 Then you can link without any make environment changes.
5125 gcc -Wl,-verbose -o a.exe -L../lib/ -lxxx
5128 This technique also avoids the version number problems, because the following is
5135 libxxx.dll.a -> ../bin/cygxxx-5.dll
5138 Linking directly to a dll without using an import lib will work
5139 even when auto-import features are exercised, and even when
5140 @samp{--enable-runtime-pseudo-relocs} is used.
5142 Given the improvements in speed and memory usage, one might justifiably
5143 wonder why import libraries are used at all. There are two reasons:
5145 1. Until recently, the link-directly-to-dll functionality did @emph{not}
5146 work with auto-imported data.
5148 2. Sometimes it is necessary to include pure static objects within the
5149 import library (which otherwise contains only bfd's for indirection
5150 symbols that point to the exports of a dll). Again, the import lib
5151 for the cygwin kernel makes use of this ability, and it is not
5152 possible to do this without an import lib.
5154 So, import libs are not going away. But the ability to replace
5155 true import libs with a simple symbolic link to (or a copy of)
5156 a dll, in most cases, is a useful addition to the suite of tools
5157 binutils makes available to the win32 developer. Given the
5158 massive improvements in memory requirements during linking, storage
5159 requirements, and linking speed, we expect that many developers
5160 will soon begin to use this feature whenever possible.
5162 @item symbol aliasing
5164 @item adding additional names
5165 Sometimes, it is useful to export symbols with additional names.
5166 A symbol @samp{foo} will be exported as @samp{foo}, but it can also be
5167 exported as @samp{_foo} by using special directives in the DEF file
5168 when creating the dll. This will affect also the optional created
5169 import library. Consider the following DEF file:
5172 LIBRARY "xyz.dll" BASE=0x61000000
5179 The line @samp{_foo = foo} maps the symbol @samp{foo} to @samp{_foo}.
5181 Another method for creating a symbol alias is to create it in the
5182 source code using the "weak" attribute:
5185 void foo () @{ /* Do something. */; @}
5186 void _foo () __attribute__ ((weak, alias ("foo")));
5189 See the gcc manual for more information about attributes and weak
5192 @item renaming symbols
5193 Sometimes it is useful to rename exports. For instance, the cygwin
5194 kernel does this regularly. A symbol @samp{_foo} can be exported as
5195 @samp{foo} but not as @samp{_foo} by using special directives in the
5196 DEF file. (This will also affect the import library, if it is
5197 created). In the following example:
5200 LIBRARY "xyz.dll" BASE=0x61000000
5206 The line @samp{_foo = foo} maps the exported symbol @samp{foo} to
5210 Note: using a DEF file disables the default auto-export behavior,
5211 unless the @samp{--export-all-symbols} command line option is used.
5212 If, however, you are trying to rename symbols, then you should list
5213 @emph{all} desired exports in the DEF file, including the symbols
5214 that are not being renamed, and do @emph{not} use the
5215 @samp{--export-all-symbols} option. If you list only the
5216 renamed symbols in the DEF file, and use @samp{--export-all-symbols}
5217 to handle the other symbols, then the both the new names @emph{and}
5218 the original names for the the renamed symbols will be exported.
5219 In effect, you'd be aliasing those symbols, not renaming them,
5220 which is probably not what you wanted.
5234 @section @code{ld} and Xtensa Processors
5236 @cindex Xtensa processors
5237 The default @command{ld} behavior for Xtensa processors is to interpret
5238 @code{SECTIONS} commands so that lists of explicitly named sections in a
5239 specification with a wildcard file will be interleaved when necessary to
5240 keep literal pools within the range of PC-relative load offsets. For
5241 example, with the command:
5253 @command{ld} may interleave some of the @code{.literal}
5254 and @code{.text} sections from different object files to ensure that the
5255 literal pools are within the range of PC-relative load offsets. A valid
5256 interleaving might place the @code{.literal} sections from an initial
5257 group of files followed by the @code{.text} sections of that group of
5258 files. Then, the @code{.literal} sections from the rest of the files
5259 and the @code{.text} sections from the rest of the files would follow.
5260 The non-interleaved order can still be specified as:
5266 *(.literal) *(.text)
5271 @cindex @code{--relax} on Xtensa
5272 @cindex relaxing on Xtensa
5274 The Xtensa version of @command{ld} enables the @option{--relax} option by
5275 default to attempt to reduce space in the output image by combining
5276 literals with identical values. It also provides the
5277 @option{--no-relax} option to disable this optimization. When enabled,
5278 the relaxation algorithm ensures that a literal will only be merged with
5279 another literal when the new merged literal location is within the
5280 offset range of all of its uses.
5282 The relaxation mechanism will also attempt to optimize
5283 assembler-generated ``longcall'' sequences of
5284 @code{L32R}/@code{CALLX@var{n}} when the target is known to fit into a
5285 @code{CALL@var{n}} instruction encoding. The current optimization
5286 converts the sequence into @code{NOP}/@code{CALL@var{n}} and removes the
5287 literal referenced by the @code{L32R} instruction.
5294 @ifclear SingleFormat
5299 @cindex object file management
5300 @cindex object formats available
5302 The linker accesses object and archive files using the BFD libraries.
5303 These libraries allow the linker to use the same routines to operate on
5304 object files whatever the object file format. A different object file
5305 format can be supported simply by creating a new BFD back end and adding
5306 it to the library. To conserve runtime memory, however, the linker and
5307 associated tools are usually configured to support only a subset of the
5308 object file formats available. You can use @code{objdump -i}
5309 (@pxref{objdump,,objdump,binutils.info,The GNU Binary Utilities}) to
5310 list all the formats available for your configuration.
5312 @cindex BFD requirements
5313 @cindex requirements for BFD
5314 As with most implementations, BFD is a compromise between
5315 several conflicting requirements. The major factor influencing
5316 BFD design was efficiency: any time used converting between
5317 formats is time which would not have been spent had BFD not
5318 been involved. This is partly offset by abstraction payback; since
5319 BFD simplifies applications and back ends, more time and care
5320 may be spent optimizing algorithms for a greater speed.
5322 One minor artifact of the BFD solution which you should bear in
5323 mind is the potential for information loss. There are two places where
5324 useful information can be lost using the BFD mechanism: during
5325 conversion and during output. @xref{BFD information loss}.
5328 * BFD outline:: How it works: an outline of BFD
5332 @section How It Works: An Outline of BFD
5333 @cindex opening object files
5334 @include bfdsumm.texi
5337 @node Reporting Bugs
5338 @chapter Reporting Bugs
5339 @cindex bugs in @command{ld}
5340 @cindex reporting bugs in @command{ld}
5342 Your bug reports play an essential role in making @command{ld} reliable.
5344 Reporting a bug may help you by bringing a solution to your problem, or
5345 it may not. But in any case the principal function of a bug report is
5346 to help the entire community by making the next version of @command{ld}
5347 work better. Bug reports are your contribution to the maintenance of
5350 In order for a bug report to serve its purpose, you must include the
5351 information that enables us to fix the bug.
5354 * Bug Criteria:: Have you found a bug?
5355 * Bug Reporting:: How to report bugs
5359 @section Have You Found a Bug?
5360 @cindex bug criteria
5362 If you are not sure whether you have found a bug, here are some guidelines:
5365 @cindex fatal signal
5366 @cindex linker crash
5367 @cindex crash of linker
5369 If the linker gets a fatal signal, for any input whatever, that is a
5370 @command{ld} bug. Reliable linkers never crash.
5372 @cindex error on valid input
5374 If @command{ld} produces an error message for valid input, that is a bug.
5376 @cindex invalid input
5378 If @command{ld} does not produce an error message for invalid input, that
5379 may be a bug. In the general case, the linker can not verify that
5380 object files are correct.
5383 If you are an experienced user of linkers, your suggestions for
5384 improvement of @command{ld} are welcome in any case.
5388 @section How to Report Bugs
5390 @cindex @command{ld} bugs, reporting
5392 A number of companies and individuals offer support for @sc{gnu}
5393 products. If you obtained @command{ld} from a support organization, we
5394 recommend you contact that organization first.
5396 You can find contact information for many support companies and
5397 individuals in the file @file{etc/SERVICE} in the @sc{gnu} Emacs
5400 Otherwise, send bug reports for @command{ld} to
5401 @samp{bug-binutils@@gnu.org}.
5403 The fundamental principle of reporting bugs usefully is this:
5404 @strong{report all the facts}. If you are not sure whether to state a
5405 fact or leave it out, state it!
5407 Often people omit facts because they think they know what causes the
5408 problem and assume that some details do not matter. Thus, you might
5409 assume that the name of a symbol you use in an example does not
5410 matter. Well, probably it does not, but one cannot be sure. Perhaps
5411 the bug is a stray memory reference which happens to fetch from the
5412 location where that name is stored in memory; perhaps, if the name
5413 were different, the contents of that location would fool the linker
5414 into doing the right thing despite the bug. Play it safe and give a
5415 specific, complete example. That is the easiest thing for you to do,
5416 and the most helpful.
5418 Keep in mind that the purpose of a bug report is to enable us to fix
5419 the bug if it is new to us. Therefore, always write your bug reports
5420 on the assumption that the bug has not been reported previously.
5422 Sometimes people give a few sketchy facts and ask, ``Does this ring a
5423 bell?'' This cannot help us fix a bug, so it is basically useless. We
5424 respond by asking for enough details to enable us to investigate.
5425 You might as well expedite matters by sending them to begin with.
5427 To enable us to fix the bug, you should include all these things:
5431 The version of @command{ld}. @command{ld} announces it if you start it with
5432 the @samp{--version} argument.
5434 Without this, we will not know whether there is any point in looking for
5435 the bug in the current version of @command{ld}.
5438 Any patches you may have applied to the @command{ld} source, including any
5439 patches made to the @code{BFD} library.
5442 The type of machine you are using, and the operating system name and
5446 What compiler (and its version) was used to compile @command{ld}---e.g.
5450 The command arguments you gave the linker to link your example and
5451 observe the bug. To guarantee you will not omit something important,
5452 list them all. A copy of the Makefile (or the output from make) is
5455 If we were to try to guess the arguments, we would probably guess wrong
5456 and then we might not encounter the bug.
5459 A complete input file, or set of input files, that will reproduce the
5460 bug. It is generally most helpful to send the actual object files
5461 provided that they are reasonably small. Say no more than 10K. For
5462 bigger files you can either make them available by FTP or HTTP or else
5463 state that you are willing to send the object file(s) to whomever
5464 requests them. (Note - your email will be going to a mailing list, so
5465 we do not want to clog it up with large attachments). But small
5466 attachments are best.
5468 If the source files were assembled using @code{gas} or compiled using
5469 @code{gcc}, then it may be OK to send the source files rather than the
5470 object files. In this case, be sure to say exactly what version of
5471 @code{gas} or @code{gcc} was used to produce the object files. Also say
5472 how @code{gas} or @code{gcc} were configured.
5475 A description of what behavior you observe that you believe is
5476 incorrect. For example, ``It gets a fatal signal.''
5478 Of course, if the bug is that @command{ld} gets a fatal signal, then we
5479 will certainly notice it. But if the bug is incorrect output, we might
5480 not notice unless it is glaringly wrong. You might as well not give us
5481 a chance to make a mistake.
5483 Even if the problem you experience is a fatal signal, you should still
5484 say so explicitly. Suppose something strange is going on, such as, your
5485 copy of @command{ld} is out of synch, or you have encountered a bug in the
5486 C library on your system. (This has happened!) Your copy might crash
5487 and ours would not. If you told us to expect a crash, then when ours
5488 fails to crash, we would know that the bug was not happening for us. If
5489 you had not told us to expect a crash, then we would not be able to draw
5490 any conclusion from our observations.
5493 If you wish to suggest changes to the @command{ld} source, send us context
5494 diffs, as generated by @code{diff} with the @samp{-u}, @samp{-c}, or
5495 @samp{-p} option. Always send diffs from the old file to the new file.
5496 If you even discuss something in the @command{ld} source, refer to it by
5497 context, not by line number.
5499 The line numbers in our development sources will not match those in your
5500 sources. Your line numbers would convey no useful information to us.
5503 Here are some things that are not necessary:
5507 A description of the envelope of the bug.
5509 Often people who encounter a bug spend a lot of time investigating
5510 which changes to the input file will make the bug go away and which
5511 changes will not affect it.
5513 This is often time consuming and not very useful, because the way we
5514 will find the bug is by running a single example under the debugger
5515 with breakpoints, not by pure deduction from a series of examples.
5516 We recommend that you save your time for something else.
5518 Of course, if you can find a simpler example to report @emph{instead}
5519 of the original one, that is a convenience for us. Errors in the
5520 output will be easier to spot, running under the debugger will take
5521 less time, and so on.
5523 However, simplification is not vital; if you do not want to do this,
5524 report the bug anyway and send us the entire test case you used.
5527 A patch for the bug.
5529 A patch for the bug does help us if it is a good one. But do not omit
5530 the necessary information, such as the test case, on the assumption that
5531 a patch is all we need. We might see problems with your patch and decide
5532 to fix the problem another way, or we might not understand it at all.
5534 Sometimes with a program as complicated as @command{ld} it is very hard to
5535 construct an example that will make the program follow a certain path
5536 through the code. If you do not send us the example, we will not be
5537 able to construct one, so we will not be able to verify that the bug is
5540 And if we cannot understand what bug you are trying to fix, or why your
5541 patch should be an improvement, we will not install it. A test case will
5542 help us to understand.
5545 A guess about what the bug is or what it depends on.
5547 Such guesses are usually wrong. Even we cannot guess right about such
5548 things without first using the debugger to find the facts.
5552 @appendix MRI Compatible Script Files
5553 @cindex MRI compatibility
5554 To aid users making the transition to @sc{gnu} @command{ld} from the MRI
5555 linker, @command{ld} can use MRI compatible linker scripts as an
5556 alternative to the more general-purpose linker scripting language
5557 described in @ref{Scripts}. MRI compatible linker scripts have a much
5558 simpler command set than the scripting language otherwise used with
5559 @command{ld}. @sc{gnu} @command{ld} supports the most commonly used MRI
5560 linker commands; these commands are described here.
5562 In general, MRI scripts aren't of much use with the @code{a.out} object
5563 file format, since it only has three sections and MRI scripts lack some
5564 features to make use of them.
5566 You can specify a file containing an MRI-compatible script using the
5567 @samp{-c} command-line option.
5569 Each command in an MRI-compatible script occupies its own line; each
5570 command line starts with the keyword that identifies the command (though
5571 blank lines are also allowed for punctuation). If a line of an
5572 MRI-compatible script begins with an unrecognized keyword, @command{ld}
5573 issues a warning message, but continues processing the script.
5575 Lines beginning with @samp{*} are comments.
5577 You can write these commands using all upper-case letters, or all
5578 lower case; for example, @samp{chip} is the same as @samp{CHIP}.
5579 The following list shows only the upper-case form of each command.
5582 @cindex @code{ABSOLUTE} (MRI)
5583 @item ABSOLUTE @var{secname}
5584 @itemx ABSOLUTE @var{secname}, @var{secname}, @dots{} @var{secname}
5585 Normally, @command{ld} includes in the output file all sections from all
5586 the input files. However, in an MRI-compatible script, you can use the
5587 @code{ABSOLUTE} command to restrict the sections that will be present in
5588 your output program. If the @code{ABSOLUTE} command is used at all in a
5589 script, then only the sections named explicitly in @code{ABSOLUTE}
5590 commands will appear in the linker output. You can still use other
5591 input sections (whatever you select on the command line, or using
5592 @code{LOAD}) to resolve addresses in the output file.
5594 @cindex @code{ALIAS} (MRI)
5595 @item ALIAS @var{out-secname}, @var{in-secname}
5596 Use this command to place the data from input section @var{in-secname}
5597 in a section called @var{out-secname} in the linker output file.
5599 @var{in-secname} may be an integer.
5601 @cindex @code{ALIGN} (MRI)
5602 @item ALIGN @var{secname} = @var{expression}
5603 Align the section called @var{secname} to @var{expression}. The
5604 @var{expression} should be a power of two.
5606 @cindex @code{BASE} (MRI)
5607 @item BASE @var{expression}
5608 Use the value of @var{expression} as the lowest address (other than
5609 absolute addresses) in the output file.
5611 @cindex @code{CHIP} (MRI)
5612 @item CHIP @var{expression}
5613 @itemx CHIP @var{expression}, @var{expression}
5614 This command does nothing; it is accepted only for compatibility.
5616 @cindex @code{END} (MRI)
5618 This command does nothing whatever; it's only accepted for compatibility.
5620 @cindex @code{FORMAT} (MRI)
5621 @item FORMAT @var{output-format}
5622 Similar to the @code{OUTPUT_FORMAT} command in the more general linker
5623 language, but restricted to one of these output formats:
5627 S-records, if @var{output-format} is @samp{S}
5630 IEEE, if @var{output-format} is @samp{IEEE}
5633 COFF (the @samp{coff-m68k} variant in BFD), if @var{output-format} is
5637 @cindex @code{LIST} (MRI)
5638 @item LIST @var{anything}@dots{}
5639 Print (to the standard output file) a link map, as produced by the
5640 @command{ld} command-line option @samp{-M}.
5642 The keyword @code{LIST} may be followed by anything on the
5643 same line, with no change in its effect.
5645 @cindex @code{LOAD} (MRI)
5646 @item LOAD @var{filename}
5647 @itemx LOAD @var{filename}, @var{filename}, @dots{} @var{filename}
5648 Include one or more object file @var{filename} in the link; this has the
5649 same effect as specifying @var{filename} directly on the @command{ld}
5652 @cindex @code{NAME} (MRI)
5653 @item NAME @var{output-name}
5654 @var{output-name} is the name for the program produced by @command{ld}; the
5655 MRI-compatible command @code{NAME} is equivalent to the command-line
5656 option @samp{-o} or the general script language command @code{OUTPUT}.
5658 @cindex @code{ORDER} (MRI)
5659 @item ORDER @var{secname}, @var{secname}, @dots{} @var{secname}
5660 @itemx ORDER @var{secname} @var{secname} @var{secname}
5661 Normally, @command{ld} orders the sections in its output file in the
5662 order in which they first appear in the input files. In an MRI-compatible
5663 script, you can override this ordering with the @code{ORDER} command. The
5664 sections you list with @code{ORDER} will appear first in your output
5665 file, in the order specified.
5667 @cindex @code{PUBLIC} (MRI)
5668 @item PUBLIC @var{name}=@var{expression}
5669 @itemx PUBLIC @var{name},@var{expression}
5670 @itemx PUBLIC @var{name} @var{expression}
5671 Supply a value (@var{expression}) for external symbol
5672 @var{name} used in the linker input files.
5674 @cindex @code{SECT} (MRI)
5675 @item SECT @var{secname}, @var{expression}
5676 @itemx SECT @var{secname}=@var{expression}
5677 @itemx SECT @var{secname} @var{expression}
5678 You can use any of these three forms of the @code{SECT} command to
5679 specify the start address (@var{expression}) for section @var{secname}.
5680 If you have more than one @code{SECT} statement for the same
5681 @var{secname}, only the @emph{first} sets the start address.
5692 % I think something like @colophon should be in texinfo. In the
5694 \long\def\colophon{\hbox to0pt{}\vfill
5695 \centerline{The body of this manual is set in}
5696 \centerline{\fontname\tenrm,}
5697 \centerline{with headings in {\bf\fontname\tenbf}}
5698 \centerline{and examples in {\tt\fontname\tentt}.}
5699 \centerline{{\it\fontname\tenit\/} and}
5700 \centerline{{\sl\fontname\tensl\/}}
5701 \centerline{are used for emphasis.}\vfill}
5703 % Blame: doc@cygnus.com, 28mar91.