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 @code{initfirst}, @code{interpose},
879 @code{loadfltr}, @code{nodefaultlib}, @code{nodelete}, @code{nodlopen},
880 @code{nodump}, @code{now}, @code{origin}, @code{combreloc}, @code{nocombreloc}
881 and @code{nocopyreloc}.
882 The other keywords are
883 ignored for Solaris compatibility. @code{initfirst} marks the object
884 to be initialized first at runtime before any other objects.
885 @code{interpose} marks the object that its symbol table interposes
886 before all symbols but the primary executable. @code{loadfltr} marks
887 the object that its filtees be processed immediately at runtime.
888 @code{nodefaultlib} marks the object that the search for dependencies
889 of this object will ignore any default library search paths.
890 @code{nodelete} marks the object shouldn't be unloaded at runtime.
891 @code{nodlopen} marks the object not available to @code{dlopen}.
892 @code{nodump} marks the object can not be dumped by @code{dldump}.
893 @code{now} marks the object with the non-lazy runtime binding.
894 @code{origin} marks the object may contain $ORIGIN.
895 @code{defs} disallows undefined symbols.
896 @code{muldefs} allows multiple definitions.
897 @code{combreloc} combines multiple reloc sections and sorts them
898 to make dynamic symbol lookup caching possible.
899 @code{nocombreloc} disables multiple reloc sections combining.
900 @code{nocopyreloc} disables production of copy relocs.
903 @cindex groups of archives
904 @item -( @var{archives} -)
905 @itemx --start-group @var{archives} --end-group
906 The @var{archives} should be a list of archive files. They may be
907 either explicit file names, or @samp{-l} options.
909 The specified archives are searched repeatedly until no new undefined
910 references are created. Normally, an archive is searched only once in
911 the order that it is specified on the command line. If a symbol in that
912 archive is needed to resolve an undefined symbol referred to by an
913 object in an archive that appears later on the command line, the linker
914 would not be able to resolve that reference. By grouping the archives,
915 they all be searched repeatedly until all possible references are
918 Using this option has a significant performance cost. It is best to use
919 it only when there are unavoidable circular references between two or
922 @kindex --accept-unknown-input-arch
923 @kindex --no-accept-unknown-input-arch
924 @item --accept-unknown-input-arch
925 @itemx --no-accept-unknown-input-arch
926 Tells the linker to accept input files whose architecture cannot be
927 recognised. The assumption is that the user knows what they are doing
928 and deliberately wants to link in these unknown input files. This was
929 the default behaviour of the linker, before release 2.14. The default
930 behaviour from release 2.14 onwards is to reject such input files, and
931 so the @samp{--accept-unknown-input-arch} option has been added to
932 restore the old behaviour.
934 @kindex -assert @var{keyword}
935 @item -assert @var{keyword}
936 This option is ignored for SunOS compatibility.
944 Link against dynamic libraries. This is only meaningful on platforms
945 for which shared libraries are supported. This option is normally the
946 default on such platforms. The different variants of this option are
947 for compatibility with various systems. You may use this option
948 multiple times on the command line: it affects library searching for
949 @option{-l} options which follow it.
953 Set the @code{DF_1_GROUP} flag in the @code{DT_FLAGS_1} entry in the dynamic
954 section. This causes the runtime linker to handle lookups in this
955 object and its dependencies to be performed only inside the group.
956 @option{--no-undefined} is implied. This option is only meaningful on ELF
957 platforms which support shared libraries.
967 Do not link against shared libraries. This is only meaningful on
968 platforms for which shared libraries are supported. The different
969 variants of this option are for compatibility with various systems. You
970 may use this option multiple times on the command line: it affects
971 library searching for @option{-l} options which follow it.
975 When creating a shared library, bind references to global symbols to the
976 definition within the shared library, if any. Normally, it is possible
977 for a program linked against a shared library to override the definition
978 within the shared library. This option is only meaningful on ELF
979 platforms which support shared libraries.
981 @kindex --check-sections
982 @kindex --no-check-sections
983 @item --check-sections
984 @itemx --no-check-sections
985 Asks the linker @emph{not} to check section addresses after they have
986 been assigned to see if there any overlaps. Normally the linker will
987 perform this check, and if it finds any overlaps it will produce
988 suitable error messages. The linker does know about, and does make
989 allowances for sections in overlays. The default behaviour can be
990 restored by using the command line switch @samp{--check-sections}.
992 @cindex cross reference table
995 Output a cross reference table. If a linker map file is being
996 generated, the cross reference table is printed to the map file.
997 Otherwise, it is printed on the standard output.
999 The format of the table is intentionally simple, so that it may be
1000 easily processed by a script if necessary. The symbols are printed out,
1001 sorted by name. For each symbol, a list of file names is given. If the
1002 symbol is defined, the first file listed is the location of the
1003 definition. The remaining files contain references to the symbol.
1005 @cindex common allocation
1006 @kindex --no-define-common
1007 @item --no-define-common
1008 This option inhibits the assignment of addresses to common symbols.
1009 The script command @code{INHIBIT_COMMON_ALLOCATION} has the same effect.
1010 @xref{Miscellaneous Commands}.
1012 The @samp{--no-define-common} option allows decoupling
1013 the decision to assign addresses to Common symbols from the choice
1014 of the output file type; otherwise a non-Relocatable output type
1015 forces assigning addresses to Common symbols.
1016 Using @samp{--no-define-common} allows Common symbols that are referenced
1017 from a shared library to be assigned addresses only in the main program.
1018 This eliminates the unused duplicate space in the shared library,
1019 and also prevents any possible confusion over resolving to the wrong
1020 duplicate when there are many dynamic modules with specialized search
1021 paths for runtime symbol resolution.
1023 @cindex symbols, from command line
1024 @kindex --defsym @var{symbol}=@var{exp}
1025 @item --defsym @var{symbol}=@var{expression}
1026 Create a global symbol in the output file, containing the absolute
1027 address given by @var{expression}. You may use this option as many
1028 times as necessary to define multiple symbols in the command line. A
1029 limited form of arithmetic is supported for the @var{expression} in this
1030 context: you may give a hexadecimal constant or the name of an existing
1031 symbol, or use @code{+} and @code{-} to add or subtract hexadecimal
1032 constants or symbols. If you need more elaborate expressions, consider
1033 using the linker command language from a script (@pxref{Assignments,,
1034 Assignment: Symbol Definitions}). @emph{Note:} there should be no white
1035 space between @var{symbol}, the equals sign (``@key{=}''), and
1038 @cindex demangling, from command line
1039 @kindex --demangle[=@var{style}]
1040 @kindex --no-demangle
1041 @item --demangle[=@var{style}]
1042 @itemx --no-demangle
1043 These options control whether to demangle symbol names in error messages
1044 and other output. When the linker is told to demangle, it tries to
1045 present symbol names in a readable fashion: it strips leading
1046 underscores if they are used by the object file format, and converts C++
1047 mangled symbol names into user readable names. Different compilers have
1048 different mangling styles. The optional demangling style argument can be used
1049 to choose an appropriate demangling style for your compiler. The linker will
1050 demangle by default unless the environment variable @samp{COLLECT_NO_DEMANGLE}
1051 is set. These options may be used to override the default.
1053 @cindex dynamic linker, from command line
1054 @kindex -I@var{file}
1055 @kindex --dynamic-linker @var{file}
1056 @item --dynamic-linker @var{file}
1057 Set the name of the dynamic linker. This is only meaningful when
1058 generating dynamically linked ELF executables. The default dynamic
1059 linker is normally correct; don't use this unless you know what you are
1062 @cindex MIPS embedded PIC code
1063 @kindex --embedded-relocs
1064 @item --embedded-relocs
1065 This option is only meaningful when linking MIPS embedded PIC code,
1066 generated by the -membedded-pic option to the @sc{gnu} compiler and
1067 assembler. It causes the linker to create a table which may be used at
1068 runtime to relocate any data which was statically initialized to pointer
1069 values. See the code in testsuite/ld-empic for details.
1072 @kindex --fatal-warnings
1073 @item --fatal-warnings
1074 Treat all warnings as errors.
1076 @kindex --force-exe-suffix
1077 @item --force-exe-suffix
1078 Make sure that an output file has a .exe suffix.
1080 If a successfully built fully linked output file does not have a
1081 @code{.exe} or @code{.dll} suffix, this option forces the linker to copy
1082 the output file to one of the same name with a @code{.exe} suffix. This
1083 option is useful when using unmodified Unix makefiles on a Microsoft
1084 Windows host, since some versions of Windows won't run an image unless
1085 it ends in a @code{.exe} suffix.
1087 @kindex --gc-sections
1088 @kindex --no-gc-sections
1089 @cindex garbage collection
1090 @item --no-gc-sections
1091 @itemx --gc-sections
1092 Enable garbage collection of unused input sections. It is ignored on
1093 targets that do not support this option. This option is not compatible
1094 with @samp{-r}, nor should it be used with dynamic linking. The default
1095 behaviour (of not performing this garbage collection) can be restored by
1096 specifying @samp{--no-gc-sections} on the command line.
1102 Print a summary of the command-line options on the standard output and exit.
1104 @kindex --target-help
1106 Print a summary of all target specific options on the standard output and exit.
1109 @item -Map @var{mapfile}
1110 Print a link map to the file @var{mapfile}. See the description of the
1111 @samp{-M} option, above.
1113 @cindex memory usage
1114 @kindex --no-keep-memory
1115 @item --no-keep-memory
1116 @command{ld} normally optimizes for speed over memory usage by caching the
1117 symbol tables of input files in memory. This option tells @command{ld} to
1118 instead optimize for memory usage, by rereading the symbol tables as
1119 necessary. This may be required if @command{ld} runs out of memory space
1120 while linking a large executable.
1122 @kindex --no-undefined
1124 @item --no-undefined
1126 Normally when creating a non-symbolic shared library, undefined symbols
1127 are allowed and left to be resolved by the runtime loader. This option
1128 disallows such undefined symbols if they come from regular object
1129 files. The switch @samp{--no-allow-shlib-undefined} controls the
1130 behaviour for shared objects being linked into the shared library.
1132 @kindex --allow-multiple-definition
1134 @item --allow-multiple-definition
1136 Normally when a symbol is defined multiple times, the linker will
1137 report a fatal error. These options allow multiple definitions and the
1138 first definition will be used.
1140 @kindex --allow-shlib-undefined
1141 @kindex --no-allow-shlib-undefined
1142 @item --allow-shlib-undefined
1143 @itemx --no-allow-shlib-undefined
1144 Allow (the default) or disallow undefined symbols in shared objects.
1145 The setting of this switch overrides @samp{--no-undefined} where
1146 shared objects are concerned. Thus if @samp{--no-undefined} is set
1147 but @samp{--no-allow-shlib-undefined} is not, the net result will be
1148 that undefined symbols in regular object files will trigger an error,
1149 but undefined symbols in shared objects will be ignored.
1151 The reason that @samp{--allow-shlib-undefined} is the default is that
1152 the shared object being specified at link time may not be the same one
1153 that is available at load time, so the symbols might actually be
1154 resolvable at load time. Plus there are some systems, (eg BeOS) where
1155 undefined symbols in shared libraries is normal since the kernel
1156 patches them at load time to select which function is most appropriate
1157 for the current architecture. eg. to dynamically select an appropriate
1158 memset function. Apparently it is also normal for HPPA shared
1159 libraries to have undefined symbols.
1161 @kindex --no-undefined-version
1162 @item --no-undefined-version
1163 Normally when a symbol has an undefined version, the linker will ignore
1164 it. This option disallows symbols with undefined version and a fatal error
1165 will be issued instead.
1167 @kindex --no-warn-mismatch
1168 @item --no-warn-mismatch
1169 Normally @command{ld} will give an error if you try to link together input
1170 files that are mismatched for some reason, perhaps because they have
1171 been compiled for different processors or for different endiannesses.
1172 This option tells @command{ld} that it should silently permit such possible
1173 errors. This option should only be used with care, in cases when you
1174 have taken some special action that ensures that the linker errors are
1177 @kindex --no-whole-archive
1178 @item --no-whole-archive
1179 Turn off the effect of the @option{--whole-archive} option for subsequent
1182 @cindex output file after errors
1183 @kindex --noinhibit-exec
1184 @item --noinhibit-exec
1185 Retain the executable output file whenever it is still usable.
1186 Normally, the linker will not produce an output file if it encounters
1187 errors during the link process; it exits without writing an output file
1188 when it issues any error whatsoever.
1192 Only search library directories explicitly specified on the
1193 command line. Library directories specified in linker scripts
1194 (including linker scripts specified on the command line) are ignored.
1196 @ifclear SingleFormat
1198 @item --oformat @var{output-format}
1199 @command{ld} may be configured to support more than one kind of object
1200 file. If your @command{ld} is configured this way, you can use the
1201 @samp{--oformat} option to specify the binary format for the output
1202 object file. Even when @command{ld} is configured to support alternative
1203 object formats, you don't usually need to specify this, as @command{ld}
1204 should be configured to produce as a default output format the most
1205 usual format on each machine. @var{output-format} is a text string, the
1206 name of a particular format supported by the BFD libraries. (You can
1207 list the available binary formats with @samp{objdump -i}.) The script
1208 command @code{OUTPUT_FORMAT} can also specify the output format, but
1209 this option overrides it. @xref{BFD}.
1213 @kindex --pic-executable
1215 @itemx --pic-executable
1216 @cindex position independent executables
1217 Create a position independent executable. This is currently only supported on
1218 ELF platforms. Position independent executables are similar to shared
1219 libraries in that they are relocated by the dynamic linker to the virtual
1220 address the OS chooses for them (which can vary between invocations). Like
1221 normal dynamically linked executables they can be executed and symbols
1222 defined in the executable cannot be overridden by shared libraries.
1226 This option is ignored for Linux compatibility.
1230 This option is ignored for SVR4 compatibility.
1233 @cindex synthesizing linker
1234 @cindex relaxing addressing modes
1236 An option with machine dependent effects.
1238 This option is only supported on a few targets.
1241 @xref{H8/300,,@command{ld} and the H8/300}.
1244 @xref{i960,, @command{ld} and the Intel 960 family}.
1247 @xref{Xtensa,, @command{ld} and Xtensa Processors}.
1250 On some platforms, the @samp{--relax} option performs global
1251 optimizations that become possible when the linker resolves addressing
1252 in the program, such as relaxing address modes and synthesizing new
1253 instructions in the output object file.
1255 On some platforms these link time global optimizations may make symbolic
1256 debugging of the resulting executable impossible.
1259 the case for the Matsushita MN10200 and MN10300 family of processors.
1263 On platforms where this is not supported, @samp{--relax} is accepted,
1267 @cindex retaining specified symbols
1268 @cindex stripping all but some symbols
1269 @cindex symbols, retaining selectively
1270 @item --retain-symbols-file @var{filename}
1271 Retain @emph{only} the symbols listed in the file @var{filename},
1272 discarding all others. @var{filename} is simply a flat file, with one
1273 symbol name per line. This option is especially useful in environments
1277 where a large global symbol table is accumulated gradually, to conserve
1280 @samp{--retain-symbols-file} does @emph{not} discard undefined symbols,
1281 or symbols needed for relocations.
1283 You may only specify @samp{--retain-symbols-file} once in the command
1284 line. It overrides @samp{-s} and @samp{-S}.
1287 @item -rpath @var{dir}
1288 @cindex runtime library search path
1290 Add a directory to the runtime library search path. This is used when
1291 linking an ELF executable with shared objects. All @option{-rpath}
1292 arguments are concatenated and passed to the runtime linker, which uses
1293 them to locate shared objects at runtime. The @option{-rpath} option is
1294 also used when locating shared objects which are needed by shared
1295 objects explicitly included in the link; see the description of the
1296 @option{-rpath-link} option. If @option{-rpath} is not used when linking an
1297 ELF executable, the contents of the environment variable
1298 @code{LD_RUN_PATH} will be used if it is defined.
1300 The @option{-rpath} option may also be used on SunOS. By default, on
1301 SunOS, the linker will form a runtime search patch out of all the
1302 @option{-L} options it is given. If a @option{-rpath} option is used, the
1303 runtime search path will be formed exclusively using the @option{-rpath}
1304 options, ignoring the @option{-L} options. This can be useful when using
1305 gcc, which adds many @option{-L} options which may be on NFS mounted
1308 For compatibility with other ELF linkers, if the @option{-R} option is
1309 followed by a directory name, rather than a file name, it is treated as
1310 the @option{-rpath} option.
1314 @cindex link-time runtime library search path
1316 @item -rpath-link @var{DIR}
1317 When using ELF or SunOS, one shared library may require another. This
1318 happens when an @code{ld -shared} link includes a shared library as one
1321 When the linker encounters such a dependency when doing a non-shared,
1322 non-relocatable link, it will automatically try to locate the required
1323 shared library and include it in the link, if it is not included
1324 explicitly. In such a case, the @option{-rpath-link} option
1325 specifies the first set of directories to search. The
1326 @option{-rpath-link} option may specify a sequence of directory names
1327 either by specifying a list of names separated by colons, or by
1328 appearing multiple times.
1330 This option should be used with caution as it overrides the search path
1331 that may have been hard compiled into a shared library. In such a case it
1332 is possible to use unintentionally a different search path than the
1333 runtime linker would do.
1335 The linker uses the following search paths to locate required shared
1339 Any directories specified by @option{-rpath-link} options.
1341 Any directories specified by @option{-rpath} options. The difference
1342 between @option{-rpath} and @option{-rpath-link} is that directories
1343 specified by @option{-rpath} options are included in the executable and
1344 used at runtime, whereas the @option{-rpath-link} option is only effective
1345 at link time. It is for the native linker only.
1347 On an ELF system, if the @option{-rpath} and @code{rpath-link} options
1348 were not used, search the contents of the environment variable
1349 @code{LD_RUN_PATH}. It is for the native linker only.
1351 On SunOS, if the @option{-rpath} option was not used, search any
1352 directories specified using @option{-L} options.
1354 For a native linker, the contents of the environment variable
1355 @code{LD_LIBRARY_PATH}.
1357 For a native ELF linker, the directories in @code{DT_RUNPATH} or
1358 @code{DT_RPATH} of a shared library are searched for shared
1359 libraries needed by it. The @code{DT_RPATH} entries are ignored if
1360 @code{DT_RUNPATH} entries exist.
1362 The default directories, normally @file{/lib} and @file{/usr/lib}.
1364 For a native linker on an ELF system, if the file @file{/etc/ld.so.conf}
1365 exists, the list of directories found in that file.
1368 If the required shared library is not found, the linker will issue a
1369 warning and continue with the link.
1376 @cindex shared libraries
1377 Create a shared library. This is currently only supported on ELF, XCOFF
1378 and SunOS platforms. On SunOS, the linker will automatically create a
1379 shared library if the @option{-e} option is not used and there are
1380 undefined symbols in the link.
1383 @kindex --sort-common
1384 This option tells @command{ld} to sort the common symbols by size when it
1385 places them in the appropriate output sections. First come all the one
1386 byte symbols, then all the two byte, then all the four byte, and then
1387 everything else. This is to prevent gaps between symbols due to
1388 alignment constraints.
1390 @kindex --split-by-file
1391 @item --split-by-file [@var{size}]
1392 Similar to @option{--split-by-reloc} but creates a new output section for
1393 each input file when @var{size} is reached. @var{size} defaults to a
1394 size of 1 if not given.
1396 @kindex --split-by-reloc
1397 @item --split-by-reloc [@var{count}]
1398 Tries to creates extra sections in the output file so that no single
1399 output section in the file contains more than @var{count} relocations.
1400 This is useful when generating huge relocatable files for downloading into
1401 certain real time kernels with the COFF object file format; since COFF
1402 cannot represent more than 65535 relocations in a single section. Note
1403 that this will fail to work with object file formats which do not
1404 support arbitrary sections. The linker will not split up individual
1405 input sections for redistribution, so if a single input section contains
1406 more than @var{count} relocations one output section will contain that
1407 many relocations. @var{count} defaults to a value of 32768.
1411 Compute and display statistics about the operation of the linker, such
1412 as execution time and memory usage.
1414 @kindex --traditional-format
1415 @cindex traditional format
1416 @item --traditional-format
1417 For some targets, the output of @command{ld} is different in some ways from
1418 the output of some existing linker. This switch requests @command{ld} to
1419 use the traditional format instead.
1422 For example, on SunOS, @command{ld} combines duplicate entries in the
1423 symbol string table. This can reduce the size of an output file with
1424 full debugging information by over 30 percent. Unfortunately, the SunOS
1425 @code{dbx} program can not read the resulting program (@code{gdb} has no
1426 trouble). The @samp{--traditional-format} switch tells @command{ld} to not
1427 combine duplicate entries.
1429 @kindex --section-start @var{sectionname}=@var{org}
1430 @item --section-start @var{sectionname}=@var{org}
1431 Locate a section in the output file at the absolute
1432 address given by @var{org}. You may use this option as many
1433 times as necessary to locate multiple sections in the command
1435 @var{org} must be a single hexadecimal integer;
1436 for compatibility with other linkers, you may omit the leading
1437 @samp{0x} usually associated with hexadecimal values. @emph{Note:} there
1438 should be no white space between @var{sectionname}, the equals
1439 sign (``@key{=}''), and @var{org}.
1441 @kindex -Tbss @var{org}
1442 @kindex -Tdata @var{org}
1443 @kindex -Ttext @var{org}
1444 @cindex segment origins, cmd line
1445 @item -Tbss @var{org}
1446 @itemx -Tdata @var{org}
1447 @itemx -Ttext @var{org}
1448 Same as --section-start, with @code{.bss}, @code{.data} or
1449 @code{.text} as the @var{sectionname}.
1455 Display the version number for @command{ld} and list the linker emulations
1456 supported. Display which input files can and cannot be opened. Display
1457 the linker script being used by the linker.
1459 @kindex --version-script=@var{version-scriptfile}
1460 @cindex version script, symbol versions
1461 @itemx --version-script=@var{version-scriptfile}
1462 Specify the name of a version script to the linker. This is typically
1463 used when creating shared libraries to specify additional information
1464 about the version hierarchy for the library being created. This option
1465 is only meaningful on ELF platforms which support shared libraries.
1468 @kindex --warn-common
1469 @cindex warnings, on combining symbols
1470 @cindex combining symbols, warnings on
1472 Warn when a common symbol is combined with another common symbol or with
1473 a symbol definition. Unix linkers allow this somewhat sloppy practice,
1474 but linkers on some other operating systems do not. This option allows
1475 you to find potential problems from combining global symbols.
1476 Unfortunately, some C libraries use this practice, so you may get some
1477 warnings about symbols in the libraries as well as in your programs.
1479 There are three kinds of global symbols, illustrated here by C examples:
1483 A definition, which goes in the initialized data section of the output
1487 An undefined reference, which does not allocate space.
1488 There must be either a definition or a common symbol for the
1492 A common symbol. If there are only (one or more) common symbols for a
1493 variable, it goes in the uninitialized data area of the output file.
1494 The linker merges multiple common symbols for the same variable into a
1495 single symbol. If they are of different sizes, it picks the largest
1496 size. The linker turns a common symbol into a declaration, if there is
1497 a definition of the same variable.
1500 The @samp{--warn-common} option can produce five kinds of warnings.
1501 Each warning consists of a pair of lines: the first describes the symbol
1502 just encountered, and the second describes the previous symbol
1503 encountered with the same name. One or both of the two symbols will be
1508 Turning a common symbol into a reference, because there is already a
1509 definition for the symbol.
1511 @var{file}(@var{section}): warning: common of `@var{symbol}'
1512 overridden by definition
1513 @var{file}(@var{section}): warning: defined here
1517 Turning a common symbol into a reference, because a later definition for
1518 the symbol is encountered. This is the same as the previous case,
1519 except that the symbols are encountered in a different order.
1521 @var{file}(@var{section}): warning: definition of `@var{symbol}'
1523 @var{file}(@var{section}): warning: common is here
1527 Merging a common symbol with a previous same-sized common symbol.
1529 @var{file}(@var{section}): warning: multiple common
1531 @var{file}(@var{section}): warning: previous common is here
1535 Merging a common symbol with a previous larger common symbol.
1537 @var{file}(@var{section}): warning: common of `@var{symbol}'
1538 overridden by larger common
1539 @var{file}(@var{section}): warning: larger common is here
1543 Merging a common symbol with a previous smaller common symbol. This is
1544 the same as the previous case, except that the symbols are
1545 encountered in a different order.
1547 @var{file}(@var{section}): warning: common of `@var{symbol}'
1548 overriding smaller common
1549 @var{file}(@var{section}): warning: smaller common is here
1553 @kindex --warn-constructors
1554 @item --warn-constructors
1555 Warn if any global constructors are used. This is only useful for a few
1556 object file formats. For formats like COFF or ELF, the linker can not
1557 detect the use of global constructors.
1559 @kindex --warn-multiple-gp
1560 @item --warn-multiple-gp
1561 Warn if multiple global pointer values are required in the output file.
1562 This is only meaningful for certain processors, such as the Alpha.
1563 Specifically, some processors put large-valued constants in a special
1564 section. A special register (the global pointer) points into the middle
1565 of this section, so that constants can be loaded efficiently via a
1566 base-register relative addressing mode. Since the offset in
1567 base-register relative mode is fixed and relatively small (e.g., 16
1568 bits), this limits the maximum size of the constant pool. Thus, in
1569 large programs, it is often necessary to use multiple global pointer
1570 values in order to be able to address all possible constants. This
1571 option causes a warning to be issued whenever this case occurs.
1574 @cindex warnings, on undefined symbols
1575 @cindex undefined symbols, warnings on
1577 Only warn once for each undefined symbol, rather than once per module
1580 @kindex --warn-section-align
1581 @cindex warnings, on section alignment
1582 @cindex section alignment, warnings on
1583 @item --warn-section-align
1584 Warn if the address of an output section is changed because of
1585 alignment. Typically, the alignment will be set by an input section.
1586 The address will only be changed if it not explicitly specified; that
1587 is, if the @code{SECTIONS} command does not specify a start address for
1588 the section (@pxref{SECTIONS}).
1590 @kindex --whole-archive
1591 @cindex including an entire archive
1592 @item --whole-archive
1593 For each archive mentioned on the command line after the
1594 @option{--whole-archive} option, include every object file in the archive
1595 in the link, rather than searching the archive for the required object
1596 files. This is normally used to turn an archive file into a shared
1597 library, forcing every object to be included in the resulting shared
1598 library. This option may be used more than once.
1600 Two notes when using this option from gcc: First, gcc doesn't know
1601 about this option, so you have to use @option{-Wl,-whole-archive}.
1602 Second, don't forget to use @option{-Wl,-no-whole-archive} after your
1603 list of archives, because gcc will add its own list of archives to
1604 your link and you may not want this flag to affect those as well.
1607 @item --wrap @var{symbol}
1608 Use a wrapper function for @var{symbol}. Any undefined reference to
1609 @var{symbol} will be resolved to @code{__wrap_@var{symbol}}. Any
1610 undefined reference to @code{__real_@var{symbol}} will be resolved to
1613 This can be used to provide a wrapper for a system function. The
1614 wrapper function should be called @code{__wrap_@var{symbol}}. If it
1615 wishes to call the system function, it should call
1616 @code{__real_@var{symbol}}.
1618 Here is a trivial example:
1622 __wrap_malloc (int c)
1624 printf ("malloc called with %ld\n", c);
1625 return __real_malloc (c);
1629 If you link other code with this file using @option{--wrap malloc}, then
1630 all calls to @code{malloc} will call the function @code{__wrap_malloc}
1631 instead. The call to @code{__real_malloc} in @code{__wrap_malloc} will
1632 call the real @code{malloc} function.
1634 You may wish to provide a @code{__real_malloc} function as well, so that
1635 links without the @option{--wrap} option will succeed. If you do this,
1636 you should not put the definition of @code{__real_malloc} in the same
1637 file as @code{__wrap_malloc}; if you do, the assembler may resolve the
1638 call before the linker has a chance to wrap it to @code{malloc}.
1640 @kindex --enable-new-dtags
1641 @kindex --disable-new-dtags
1642 @item --enable-new-dtags
1643 @itemx --disable-new-dtags
1644 This linker can create the new dynamic tags in ELF. But the older ELF
1645 systems may not understand them. If you specify
1646 @option{--enable-new-dtags}, the dynamic tags will be created as needed.
1647 If you specify @option{--disable-new-dtags}, no new dynamic tags will be
1648 created. By default, the new dynamic tags are not created. Note that
1649 those options are only available for ELF systems.
1655 @subsection Options Specific to i386 PE Targets
1657 @c man begin OPTIONS
1659 The i386 PE linker supports the @option{-shared} option, which causes
1660 the output to be a dynamically linked library (DLL) instead of a
1661 normal executable. You should name the output @code{*.dll} when you
1662 use this option. In addition, the linker fully supports the standard
1663 @code{*.def} files, which may be specified on the linker command line
1664 like an object file (in fact, it should precede archives it exports
1665 symbols from, to ensure that they get linked in, just like a normal
1668 In addition to the options common to all targets, the i386 PE linker
1669 support additional command line options that are specific to the i386
1670 PE target. Options that take values may be separated from their
1671 values by either a space or an equals sign.
1675 @kindex --add-stdcall-alias
1676 @item --add-stdcall-alias
1677 If given, symbols with a stdcall suffix (@@@var{nn}) will be exported
1678 as-is and also with the suffix stripped.
1679 [This option is specific to the i386 PE targeted port of the linker]
1682 @item --base-file @var{file}
1683 Use @var{file} as the name of a file in which to save the base
1684 addresses of all the relocations needed for generating DLLs with
1686 [This is an i386 PE specific option]
1690 Create a DLL instead of a regular executable. You may also use
1691 @option{-shared} or specify a @code{LIBRARY} in a given @code{.def}
1693 [This option is specific to the i386 PE targeted port of the linker]
1695 @kindex --enable-stdcall-fixup
1696 @kindex --disable-stdcall-fixup
1697 @item --enable-stdcall-fixup
1698 @itemx --disable-stdcall-fixup
1699 If the link finds a symbol that it cannot resolve, it will attempt to
1700 do ``fuzzy linking'' by looking for another defined symbol that differs
1701 only in the format of the symbol name (cdecl vs stdcall) and will
1702 resolve that symbol by linking to the match. For example, the
1703 undefined symbol @code{_foo} might be linked to the function
1704 @code{_foo@@12}, or the undefined symbol @code{_bar@@16} might be linked
1705 to the function @code{_bar}. When the linker does this, it prints a
1706 warning, since it normally should have failed to link, but sometimes
1707 import libraries generated from third-party dlls may need this feature
1708 to be usable. If you specify @option{--enable-stdcall-fixup}, this
1709 feature is fully enabled and warnings are not printed. If you specify
1710 @option{--disable-stdcall-fixup}, this feature is disabled and such
1711 mismatches are considered to be errors.
1712 [This option is specific to the i386 PE targeted port of the linker]
1714 @cindex DLLs, creating
1715 @kindex --export-all-symbols
1716 @item --export-all-symbols
1717 If given, all global symbols in the objects used to build a DLL will
1718 be exported by the DLL. Note that this is the default if there
1719 otherwise wouldn't be any exported symbols. When symbols are
1720 explicitly exported via DEF files or implicitly exported via function
1721 attributes, the default is to not export anything else unless this
1722 option is given. Note that the symbols @code{DllMain@@12},
1723 @code{DllEntryPoint@@0}, @code{DllMainCRTStartup@@12}, and
1724 @code{impure_ptr} will not be automatically
1725 exported. Also, symbols imported from other DLLs will not be
1726 re-exported, nor will symbols specifying the DLL's internal layout
1727 such as those beginning with @code{_head_} or ending with
1728 @code{_iname}. In addition, no symbols from @code{libgcc},
1729 @code{libstd++}, @code{libmingw32}, or @code{crtX.o} will be exported.
1730 Symbols whose names begin with @code{__rtti_} or @code{__builtin_} will
1731 not be exported, to help with C++ DLLs. Finally, there is an
1732 extensive list of cygwin-private symbols that are not exported
1733 (obviously, this applies on when building DLLs for cygwin targets).
1734 These cygwin-excludes are: @code{_cygwin_dll_entry@@12},
1735 @code{_cygwin_crt0_common@@8}, @code{_cygwin_noncygwin_dll_entry@@12},
1736 @code{_fmode}, @code{_impure_ptr}, @code{cygwin_attach_dll},
1737 @code{cygwin_premain0}, @code{cygwin_premain1}, @code{cygwin_premain2},
1738 @code{cygwin_premain3}, and @code{environ}.
1739 [This option is specific to the i386 PE targeted port of the linker]
1741 @kindex --exclude-symbols
1742 @item --exclude-symbols @var{symbol},@var{symbol},...
1743 Specifies a list of symbols which should not be automatically
1744 exported. The symbol names may be delimited by commas or colons.
1745 [This option is specific to the i386 PE targeted port of the linker]
1747 @kindex --exclude-libs
1748 @item --exclude-libs @var{lib},@var{lib},...
1749 Specifies a list of archive libraries from which symbols should not be automatically
1750 exported. The library names may be delimited by commas or colons. Specifying
1751 @code{--exclude-libs ALL} excludes symbols in all archive libraries from
1752 automatic export. Symbols explicitly listed in a .def file are still exported,
1753 regardless of this option.
1754 [This option is specific to the i386 PE targeted port of the linker]
1756 @kindex --file-alignment
1757 @item --file-alignment
1758 Specify the file alignment. Sections in the file will always begin at
1759 file offsets which are multiples of this number. This defaults to
1761 [This option is specific to the i386 PE targeted port of the linker]
1765 @item --heap @var{reserve}
1766 @itemx --heap @var{reserve},@var{commit}
1767 Specify the amount of memory to reserve (and optionally commit) to be
1768 used as heap for this program. The default is 1Mb reserved, 4K
1770 [This option is specific to the i386 PE targeted port of the linker]
1773 @kindex --image-base
1774 @item --image-base @var{value}
1775 Use @var{value} as the base address of your program or dll. This is
1776 the lowest memory location that will be used when your program or dll
1777 is loaded. To reduce the need to relocate and improve performance of
1778 your dlls, each should have a unique base address and not overlap any
1779 other dlls. The default is 0x400000 for executables, and 0x10000000
1781 [This option is specific to the i386 PE targeted port of the linker]
1785 If given, the stdcall suffixes (@@@var{nn}) will be stripped from
1786 symbols before they are exported.
1787 [This option is specific to the i386 PE targeted port of the linker]
1789 @kindex --major-image-version
1790 @item --major-image-version @var{value}
1791 Sets the major number of the ``image version''. Defaults to 1.
1792 [This option is specific to the i386 PE targeted port of the linker]
1794 @kindex --major-os-version
1795 @item --major-os-version @var{value}
1796 Sets the major number of the ``os version''. Defaults to 4.
1797 [This option is specific to the i386 PE targeted port of the linker]
1799 @kindex --major-subsystem-version
1800 @item --major-subsystem-version @var{value}
1801 Sets the major number of the ``subsystem version''. Defaults to 4.
1802 [This option is specific to the i386 PE targeted port of the linker]
1804 @kindex --minor-image-version
1805 @item --minor-image-version @var{value}
1806 Sets the minor number of the ``image version''. Defaults to 0.
1807 [This option is specific to the i386 PE targeted port of the linker]
1809 @kindex --minor-os-version
1810 @item --minor-os-version @var{value}
1811 Sets the minor number of the ``os version''. Defaults to 0.
1812 [This option is specific to the i386 PE targeted port of the linker]
1814 @kindex --minor-subsystem-version
1815 @item --minor-subsystem-version @var{value}
1816 Sets the minor number of the ``subsystem version''. Defaults to 0.
1817 [This option is specific to the i386 PE targeted port of the linker]
1819 @cindex DEF files, creating
1820 @cindex DLLs, creating
1821 @kindex --output-def
1822 @item --output-def @var{file}
1823 The linker will create the file @var{file} which will contain a DEF
1824 file corresponding to the DLL the linker is generating. This DEF file
1825 (which should be called @code{*.def}) may be used to create an import
1826 library with @code{dlltool} or may be used as a reference to
1827 automatically or implicitly exported symbols.
1828 [This option is specific to the i386 PE targeted port of the linker]
1830 @cindex DLLs, creating
1831 @kindex --out-implib
1832 @item --out-implib @var{file}
1833 The linker will create the file @var{file} which will contain an
1834 import lib corresponding to the DLL the linker is generating. This
1835 import lib (which should be called @code{*.dll.a} or @code{*.a}
1836 may be used to link clients against the generated DLL; this behavior
1837 makes it possible to skip a separate @code{dlltool} import library
1839 [This option is specific to the i386 PE targeted port of the linker]
1841 @kindex --enable-auto-image-base
1842 @item --enable-auto-image-base
1843 Automatically choose the image base for DLLs, unless one is specified
1844 using the @code{--image-base} argument. By using a hash generated
1845 from the dllname to create unique image bases for each DLL, in-memory
1846 collisions and relocations which can delay program execution are
1848 [This option is specific to the i386 PE targeted port of the linker]
1850 @kindex --disable-auto-image-base
1851 @item --disable-auto-image-base
1852 Do not automatically generate a unique image base. If there is no
1853 user-specified image base (@code{--image-base}) then use the platform
1855 [This option is specific to the i386 PE targeted port of the linker]
1857 @cindex DLLs, linking to
1858 @kindex --dll-search-prefix
1859 @item --dll-search-prefix @var{string}
1860 When linking dynamically to a dll without an import library,
1861 search for @code{<string><basename>.dll} in preference to
1862 @code{lib<basename>.dll}. This behavior allows easy distinction
1863 between DLLs built for the various "subplatforms": native, cygwin,
1864 uwin, pw, etc. For instance, cygwin DLLs typically use
1865 @code{--dll-search-prefix=cyg}.
1866 [This option is specific to the i386 PE targeted port of the linker]
1868 @kindex --enable-auto-import
1869 @item --enable-auto-import
1870 Do sophisticated linking of @code{_symbol} to @code{__imp__symbol} for
1871 DATA imports from DLLs, and create the necessary thunking symbols when
1872 building the import libraries with those DATA exports. This generally
1873 will 'just work' -- but sometimes you may see this message:
1875 "variable '<var>' can't be auto-imported. Please read the
1876 documentation for ld's @code{--enable-auto-import} for details."
1878 This message occurs when some (sub)expression accesses an address
1879 ultimately given by the sum of two constants (Win32 import tables only
1880 allow one). Instances where this may occur include accesses to member
1881 fields of struct variables imported from a DLL, as well as using a
1882 constant index into an array variable imported from a DLL. Any
1883 multiword variable (arrays, structs, long long, etc) may trigger
1884 this error condition. However, regardless of the exact data type
1885 of the offending exported variable, ld will always detect it, issue
1886 the warning, and exit.
1888 There are several ways to address this difficulty, regardless of the
1889 data type of the exported variable:
1891 One way is to use --enable-runtime-pseudo-reloc switch. This leaves the task
1892 of adjusting references in your client code for runtime environment, so
1893 this method works only when runtime environtment supports this feature.
1895 A second solution is to force one of the 'constants' to be a variable --
1896 that is, unknown and un-optimizable at compile time. For arrays,
1897 there are two possibilities: a) make the indexee (the array's address)
1898 a variable, or b) make the 'constant' index a variable. Thus:
1901 extern type extern_array[];
1903 @{ volatile type *t=extern_array; t[1] @}
1909 extern type extern_array[];
1911 @{ volatile int t=1; extern_array[t] @}
1914 For structs (and most other multiword data types) the only option
1915 is to make the struct itself (or the long long, or the ...) variable:
1918 extern struct s extern_struct;
1919 extern_struct.field -->
1920 @{ volatile struct s *t=&extern_struct; t->field @}
1926 extern long long extern_ll;
1928 @{ volatile long long * local_ll=&extern_ll; *local_ll @}
1931 A third method of dealing with this difficulty is to abandon
1932 'auto-import' for the offending symbol and mark it with
1933 @code{__declspec(dllimport)}. However, in practice that
1934 requires using compile-time #defines to indicate whether you are
1935 building a DLL, building client code that will link to the DLL, or
1936 merely building/linking to a static library. In making the choice
1937 between the various methods of resolving the 'direct address with
1938 constant offset' problem, you should consider typical real-world usage:
1946 void main(int argc, char **argv)@{
1947 printf("%d\n",arr[1]);
1957 void main(int argc, char **argv)@{
1958 /* This workaround is for win32 and cygwin; do not "optimize" */
1959 volatile int *parr = arr;
1960 printf("%d\n",parr[1]);
1967 /* Note: auto-export is assumed (no __declspec(dllexport)) */
1968 #if (defined(_WIN32) || defined(__CYGWIN__)) && \
1969 !(defined(FOO_BUILD_DLL) || defined(FOO_STATIC))
1970 #define FOO_IMPORT __declspec(dllimport)
1974 extern FOO_IMPORT int arr[];
1977 void main(int argc, char **argv)@{
1978 printf("%d\n",arr[1]);
1982 A fourth way to avoid this problem is to re-code your
1983 library to use a functional interface rather than a data interface
1984 for the offending variables (e.g. set_foo() and get_foo() accessor
1986 [This option is specific to the i386 PE targeted port of the linker]
1988 @kindex --disable-auto-import
1989 @item --disable-auto-import
1990 Do not attempt to do sophisticalted linking of @code{_symbol} to
1991 @code{__imp__symbol} for DATA imports from DLLs.
1992 [This option is specific to the i386 PE targeted port of the linker]
1994 @kindex --enable-runtime-pseudo-reloc
1995 @item --enable-runtime-pseudo-reloc
1996 If your code contains expressions described in --enable-auto-import section,
1997 that is, DATA imports from DLL with non-zero offset, this switch will create
1998 a vector of 'runtime pseudo relocations' which can be used by runtime
1999 environment to adjust references to such data in your client code.
2000 [This option is specific to the i386 PE targeted port of the linker]
2002 @kindex --disable-runtime-pseudo-reloc
2003 @item --disable-runtime-pseudo-reloc
2004 Do not create pseudo relocations for non-zero offset DATA imports from
2005 DLLs. This is the default.
2006 [This option is specific to the i386 PE targeted port of the linker]
2008 @kindex --enable-extra-pe-debug
2009 @item --enable-extra-pe-debug
2010 Show additional debug info related to auto-import symbol thunking.
2011 [This option is specific to the i386 PE targeted port of the linker]
2013 @kindex --section-alignment
2014 @item --section-alignment
2015 Sets the section alignment. Sections in memory will always begin at
2016 addresses which are a multiple of this number. Defaults to 0x1000.
2017 [This option is specific to the i386 PE targeted port of the linker]
2021 @item --stack @var{reserve}
2022 @itemx --stack @var{reserve},@var{commit}
2023 Specify the amount of memory to reserve (and optionally commit) to be
2024 used as stack for this program. The default is 2Mb reserved, 4K
2026 [This option is specific to the i386 PE targeted port of the linker]
2029 @item --subsystem @var{which}
2030 @itemx --subsystem @var{which}:@var{major}
2031 @itemx --subsystem @var{which}:@var{major}.@var{minor}
2032 Specifies the subsystem under which your program will execute. The
2033 legal values for @var{which} are @code{native}, @code{windows},
2034 @code{console}, and @code{posix}. You may optionally set the
2035 subsystem version also.
2036 [This option is specific to the i386 PE targeted port of the linker]
2044 @section Environment Variables
2046 @c man begin ENVIRONMENT
2048 You can change the behavior of @command{ld} with the environment variables
2049 @ifclear SingleFormat
2052 @code{LDEMULATION} and @code{COLLECT_NO_DEMANGLE}.
2054 @ifclear SingleFormat
2056 @cindex default input format
2057 @code{GNUTARGET} determines the input-file object format if you don't
2058 use @samp{-b} (or its synonym @samp{--format}). Its value should be one
2059 of the BFD names for an input format (@pxref{BFD}). If there is no
2060 @code{GNUTARGET} in the environment, @command{ld} uses the natural format
2061 of the target. If @code{GNUTARGET} is set to @code{default} then BFD
2062 attempts to discover the input format by examining binary input files;
2063 this method often succeeds, but there are potential ambiguities, since
2064 there is no method of ensuring that the magic number used to specify
2065 object-file formats is unique. However, the configuration procedure for
2066 BFD on each system places the conventional format for that system first
2067 in the search-list, so ambiguities are resolved in favor of convention.
2071 @cindex default emulation
2072 @cindex emulation, default
2073 @code{LDEMULATION} determines the default emulation if you don't use the
2074 @samp{-m} option. The emulation can affect various aspects of linker
2075 behaviour, particularly the default linker script. You can list the
2076 available emulations with the @samp{--verbose} or @samp{-V} options. If
2077 the @samp{-m} option is not used, and the @code{LDEMULATION} environment
2078 variable is not defined, the default emulation depends upon how the
2079 linker was configured.
2081 @kindex COLLECT_NO_DEMANGLE
2082 @cindex demangling, default
2083 Normally, the linker will default to demangling symbols. However, if
2084 @code{COLLECT_NO_DEMANGLE} is set in the environment, then it will
2085 default to not demangling symbols. This environment variable is used in
2086 a similar fashion by the @code{gcc} linker wrapper program. The default
2087 may be overridden by the @samp{--demangle} and @samp{--no-demangle}
2094 @chapter Linker Scripts
2097 @cindex linker scripts
2098 @cindex command files
2099 Every link is controlled by a @dfn{linker script}. This script is
2100 written in the linker command language.
2102 The main purpose of the linker script is to describe how the sections in
2103 the input files should be mapped into the output file, and to control
2104 the memory layout of the output file. Most linker scripts do nothing
2105 more than this. However, when necessary, the linker script can also
2106 direct the linker to perform many other operations, using the commands
2109 The linker always uses a linker script. If you do not supply one
2110 yourself, the linker will use a default script that is compiled into the
2111 linker executable. You can use the @samp{--verbose} command line option
2112 to display the default linker script. Certain command line options,
2113 such as @samp{-r} or @samp{-N}, will affect the default linker script.
2115 You may supply your own linker script by using the @samp{-T} command
2116 line option. When you do this, your linker script will replace the
2117 default linker script.
2119 You may also use linker scripts implicitly by naming them as input files
2120 to the linker, as though they were files to be linked. @xref{Implicit
2124 * Basic Script Concepts:: Basic Linker Script Concepts
2125 * Script Format:: Linker Script Format
2126 * Simple Example:: Simple Linker Script Example
2127 * Simple Commands:: Simple Linker Script Commands
2128 * Assignments:: Assigning Values to Symbols
2129 * SECTIONS:: SECTIONS Command
2130 * MEMORY:: MEMORY Command
2131 * PHDRS:: PHDRS Command
2132 * VERSION:: VERSION Command
2133 * Expressions:: Expressions in Linker Scripts
2134 * Implicit Linker Scripts:: Implicit Linker Scripts
2137 @node Basic Script Concepts
2138 @section Basic Linker Script Concepts
2139 @cindex linker script concepts
2140 We need to define some basic concepts and vocabulary in order to
2141 describe the linker script language.
2143 The linker combines input files into a single output file. The output
2144 file and each input file are in a special data format known as an
2145 @dfn{object file format}. Each file is called an @dfn{object file}.
2146 The output file is often called an @dfn{executable}, but for our
2147 purposes we will also call it an object file. Each object file has,
2148 among other things, a list of @dfn{sections}. We sometimes refer to a
2149 section in an input file as an @dfn{input section}; similarly, a section
2150 in the output file is an @dfn{output section}.
2152 Each section in an object file has a name and a size. Most sections
2153 also have an associated block of data, known as the @dfn{section
2154 contents}. A section may be marked as @dfn{loadable}, which mean that
2155 the contents should be loaded into memory when the output file is run.
2156 A section with no contents may be @dfn{allocatable}, which means that an
2157 area in memory should be set aside, but nothing in particular should be
2158 loaded there (in some cases this memory must be zeroed out). A section
2159 which is neither loadable nor allocatable typically contains some sort
2160 of debugging information.
2162 Every loadable or allocatable output section has two addresses. The
2163 first is the @dfn{VMA}, or virtual memory address. This is the address
2164 the section will have when the output file is run. The second is the
2165 @dfn{LMA}, or load memory address. This is the address at which the
2166 section will be loaded. In most cases the two addresses will be the
2167 same. An example of when they might be different is when a data section
2168 is loaded into ROM, and then copied into RAM when the program starts up
2169 (this technique is often used to initialize global variables in a ROM
2170 based system). In this case the ROM address would be the LMA, and the
2171 RAM address would be the VMA.
2173 You can see the sections in an object file by using the @code{objdump}
2174 program with the @samp{-h} option.
2176 Every object file also has a list of @dfn{symbols}, known as the
2177 @dfn{symbol table}. A symbol may be defined or undefined. Each symbol
2178 has a name, and each defined symbol has an address, among other
2179 information. If you compile a C or C++ program into an object file, you
2180 will get a defined symbol for every defined function and global or
2181 static variable. Every undefined function or global variable which is
2182 referenced in the input file will become an undefined symbol.
2184 You can see the symbols in an object file by using the @code{nm}
2185 program, or by using the @code{objdump} program with the @samp{-t}
2189 @section Linker Script Format
2190 @cindex linker script format
2191 Linker scripts are text files.
2193 You write a linker script as a series of commands. Each command is
2194 either a keyword, possibly followed by arguments, or an assignment to a
2195 symbol. You may separate commands using semicolons. Whitespace is
2198 Strings such as file or format names can normally be entered directly.
2199 If the file name contains a character such as a comma which would
2200 otherwise serve to separate file names, you may put the file name in
2201 double quotes. There is no way to use a double quote character in a
2204 You may include comments in linker scripts just as in C, delimited by
2205 @samp{/*} and @samp{*/}. As in C, comments are syntactically equivalent
2208 @node Simple Example
2209 @section Simple Linker Script Example
2210 @cindex linker script example
2211 @cindex example of linker script
2212 Many linker scripts are fairly simple.
2214 The simplest possible linker script has just one command:
2215 @samp{SECTIONS}. You use the @samp{SECTIONS} command to describe the
2216 memory layout of the output file.
2218 The @samp{SECTIONS} command is a powerful command. Here we will
2219 describe a simple use of it. Let's assume your program consists only of
2220 code, initialized data, and uninitialized data. These will be in the
2221 @samp{.text}, @samp{.data}, and @samp{.bss} sections, respectively.
2222 Let's assume further that these are the only sections which appear in
2225 For this example, let's say that the code should be loaded at address
2226 0x10000, and that the data should start at address 0x8000000. Here is a
2227 linker script which will do that:
2232 .text : @{ *(.text) @}
2234 .data : @{ *(.data) @}
2235 .bss : @{ *(.bss) @}
2239 You write the @samp{SECTIONS} command as the keyword @samp{SECTIONS},
2240 followed by a series of symbol assignments and output section
2241 descriptions enclosed in curly braces.
2243 The first line inside the @samp{SECTIONS} command of the above example
2244 sets the value of the special symbol @samp{.}, which is the location
2245 counter. If you do not specify the address of an output section in some
2246 other way (other ways are described later), the address is set from the
2247 current value of the location counter. The location counter is then
2248 incremented by the size of the output section. At the start of the
2249 @samp{SECTIONS} command, the location counter has the value @samp{0}.
2251 The second line defines an output section, @samp{.text}. The colon is
2252 required syntax which may be ignored for now. Within the curly braces
2253 after the output section name, you list the names of the input sections
2254 which should be placed into this output section. The @samp{*} is a
2255 wildcard which matches any file name. The expression @samp{*(.text)}
2256 means all @samp{.text} input sections in all input files.
2258 Since the location counter is @samp{0x10000} when the output section
2259 @samp{.text} is defined, the linker will set the address of the
2260 @samp{.text} section in the output file to be @samp{0x10000}.
2262 The remaining lines define the @samp{.data} and @samp{.bss} sections in
2263 the output file. The linker will place the @samp{.data} output section
2264 at address @samp{0x8000000}. After the linker places the @samp{.data}
2265 output section, the value of the location counter will be
2266 @samp{0x8000000} plus the size of the @samp{.data} output section. The
2267 effect is that the linker will place the @samp{.bss} output section
2268 immediately after the @samp{.data} output section in memory
2270 The linker will ensure that each output section has the required
2271 alignment, by increasing the location counter if necessary. In this
2272 example, the specified addresses for the @samp{.text} and @samp{.data}
2273 sections will probably satisfy any alignment constraints, but the linker
2274 may have to create a small gap between the @samp{.data} and @samp{.bss}
2277 That's it! That's a simple and complete linker script.
2279 @node Simple Commands
2280 @section Simple Linker Script Commands
2281 @cindex linker script simple commands
2282 In this section we describe the simple linker script commands.
2285 * Entry Point:: Setting the entry point
2286 * File Commands:: Commands dealing with files
2287 @ifclear SingleFormat
2288 * Format Commands:: Commands dealing with object file formats
2291 * Miscellaneous Commands:: Other linker script commands
2295 @subsection Setting the Entry Point
2296 @kindex ENTRY(@var{symbol})
2297 @cindex start of execution
2298 @cindex first instruction
2300 The first instruction to execute in a program is called the @dfn{entry
2301 point}. You can use the @code{ENTRY} linker script command to set the
2302 entry point. The argument is a symbol name:
2307 There are several ways to set the entry point. The linker will set the
2308 entry point by trying each of the following methods in order, and
2309 stopping when one of them succeeds:
2312 the @samp{-e} @var{entry} command-line option;
2314 the @code{ENTRY(@var{symbol})} command in a linker script;
2316 the value of the symbol @code{start}, if defined;
2318 the address of the first byte of the @samp{.text} section, if present;
2320 The address @code{0}.
2324 @subsection Commands Dealing with Files
2325 @cindex linker script file commands
2326 Several linker script commands deal with files.
2329 @item INCLUDE @var{filename}
2330 @kindex INCLUDE @var{filename}
2331 @cindex including a linker script
2332 Include the linker script @var{filename} at this point. The file will
2333 be searched for in the current directory, and in any directory specified
2334 with the @option{-L} option. You can nest calls to @code{INCLUDE} up to
2337 @item INPUT(@var{file}, @var{file}, @dots{})
2338 @itemx INPUT(@var{file} @var{file} @dots{})
2339 @kindex INPUT(@var{files})
2340 @cindex input files in linker scripts
2341 @cindex input object files in linker scripts
2342 @cindex linker script input object files
2343 The @code{INPUT} command directs the linker to include the named files
2344 in the link, as though they were named on the command line.
2346 For example, if you always want to include @file{subr.o} any time you do
2347 a link, but you can't be bothered to put it on every link command line,
2348 then you can put @samp{INPUT (subr.o)} in your linker script.
2350 In fact, if you like, you can list all of your input files in the linker
2351 script, and then invoke the linker with nothing but a @samp{-T} option.
2353 In case a @dfn{sysroot prefix} is configured, and the filename starts
2354 with the @samp{/} character, and the script being processed was
2355 located inside the @dfn{sysroot prefix}, the filename will be looked
2356 for in the @dfn{sysroot prefix}. Otherwise, the linker will try to
2357 open the file in the current directory. If it is not found, the
2358 linker will search through the archive library search path. See the
2359 description of @samp{-L} in @ref{Options,,Command Line Options}.
2361 If you use @samp{INPUT (-l@var{file})}, @command{ld} will transform the
2362 name to @code{lib@var{file}.a}, as with the command line argument
2365 When you use the @code{INPUT} command in an implicit linker script, the
2366 files will be included in the link at the point at which the linker
2367 script file is included. This can affect archive searching.
2369 @item GROUP(@var{file}, @var{file}, @dots{})
2370 @itemx GROUP(@var{file} @var{file} @dots{})
2371 @kindex GROUP(@var{files})
2372 @cindex grouping input files
2373 The @code{GROUP} command is like @code{INPUT}, except that the named
2374 files should all be archives, and they are searched repeatedly until no
2375 new undefined references are created. See the description of @samp{-(}
2376 in @ref{Options,,Command Line Options}.
2378 @item OUTPUT(@var{filename})
2379 @kindex OUTPUT(@var{filename})
2380 @cindex output file name in linker scripot
2381 The @code{OUTPUT} command names the output file. Using
2382 @code{OUTPUT(@var{filename})} in the linker script is exactly like using
2383 @samp{-o @var{filename}} on the command line (@pxref{Options,,Command
2384 Line Options}). If both are used, the command line option takes
2387 You can use the @code{OUTPUT} command to define a default name for the
2388 output file other than the usual default of @file{a.out}.
2390 @item SEARCH_DIR(@var{path})
2391 @kindex SEARCH_DIR(@var{path})
2392 @cindex library search path in linker script
2393 @cindex archive search path in linker script
2394 @cindex search path in linker script
2395 The @code{SEARCH_DIR} command adds @var{path} to the list of paths where
2396 @command{ld} looks for archive libraries. Using
2397 @code{SEARCH_DIR(@var{path})} is exactly like using @samp{-L @var{path}}
2398 on the command line (@pxref{Options,,Command Line Options}). If both
2399 are used, then the linker will search both paths. Paths specified using
2400 the command line option are searched first.
2402 @item STARTUP(@var{filename})
2403 @kindex STARTUP(@var{filename})
2404 @cindex first input file
2405 The @code{STARTUP} command is just like the @code{INPUT} command, except
2406 that @var{filename} will become the first input file to be linked, as
2407 though it were specified first on the command line. This may be useful
2408 when using a system in which the entry point is always the start of the
2412 @ifclear SingleFormat
2413 @node Format Commands
2414 @subsection Commands Dealing with Object File Formats
2415 A couple of linker script commands deal with object file formats.
2418 @item OUTPUT_FORMAT(@var{bfdname})
2419 @itemx OUTPUT_FORMAT(@var{default}, @var{big}, @var{little})
2420 @kindex OUTPUT_FORMAT(@var{bfdname})
2421 @cindex output file format in linker script
2422 The @code{OUTPUT_FORMAT} command names the BFD format to use for the
2423 output file (@pxref{BFD}). Using @code{OUTPUT_FORMAT(@var{bfdname})} is
2424 exactly like using @samp{--oformat @var{bfdname}} on the command line
2425 (@pxref{Options,,Command Line Options}). If both are used, the command
2426 line option takes precedence.
2428 You can use @code{OUTPUT_FORMAT} with three arguments to use different
2429 formats based on the @samp{-EB} and @samp{-EL} command line options.
2430 This permits the linker script to set the output format based on the
2433 If neither @samp{-EB} nor @samp{-EL} are used, then the output format
2434 will be the first argument, @var{default}. If @samp{-EB} is used, the
2435 output format will be the second argument, @var{big}. If @samp{-EL} is
2436 used, the output format will be the third argument, @var{little}.
2438 For example, the default linker script for the MIPS ELF target uses this
2441 OUTPUT_FORMAT(elf32-bigmips, elf32-bigmips, elf32-littlemips)
2443 This says that the default format for the output file is
2444 @samp{elf32-bigmips}, but if the user uses the @samp{-EL} command line
2445 option, the output file will be created in the @samp{elf32-littlemips}
2448 @item TARGET(@var{bfdname})
2449 @kindex TARGET(@var{bfdname})
2450 @cindex input file format in linker script
2451 The @code{TARGET} command names the BFD format to use when reading input
2452 files. It affects subsequent @code{INPUT} and @code{GROUP} commands.
2453 This command is like using @samp{-b @var{bfdname}} on the command line
2454 (@pxref{Options,,Command Line Options}). If the @code{TARGET} command
2455 is used but @code{OUTPUT_FORMAT} is not, then the last @code{TARGET}
2456 command is also used to set the format for the output file. @xref{BFD}.
2460 @node Miscellaneous Commands
2461 @subsection Other Linker Script Commands
2462 There are a few other linker scripts commands.
2465 @item ASSERT(@var{exp}, @var{message})
2467 @cindex assertion in linker script
2468 Ensure that @var{exp} is non-zero. If it is zero, then exit the linker
2469 with an error code, and print @var{message}.
2471 @item EXTERN(@var{symbol} @var{symbol} @dots{})
2473 @cindex undefined symbol in linker script
2474 Force @var{symbol} to be entered in the output file as an undefined
2475 symbol. Doing this may, for example, trigger linking of additional
2476 modules from standard libraries. You may list several @var{symbol}s for
2477 each @code{EXTERN}, and you may use @code{EXTERN} multiple times. This
2478 command has the same effect as the @samp{-u} command-line option.
2480 @item FORCE_COMMON_ALLOCATION
2481 @kindex FORCE_COMMON_ALLOCATION
2482 @cindex common allocation in linker script
2483 This command has the same effect as the @samp{-d} command-line option:
2484 to make @command{ld} assign space to common symbols even if a relocatable
2485 output file is specified (@samp{-r}).
2487 @item INHIBIT_COMMON_ALLOCATION
2488 @kindex INHIBIT_COMMON_ALLOCATION
2489 @cindex common allocation in linker script
2490 This command has the same effect as the @samp{--no-define-common}
2491 command-line option: to make @code{ld} omit the assignment of addresses
2492 to common symbols even for a non-relocatable output file.
2494 @item NOCROSSREFS(@var{section} @var{section} @dots{})
2495 @kindex NOCROSSREFS(@var{sections})
2496 @cindex cross references
2497 This command may be used to tell @command{ld} to issue an error about any
2498 references among certain output sections.
2500 In certain types of programs, particularly on embedded systems when
2501 using overlays, when one section is loaded into memory, another section
2502 will not be. Any direct references between the two sections would be
2503 errors. For example, it would be an error if code in one section called
2504 a function defined in the other section.
2506 The @code{NOCROSSREFS} command takes a list of output section names. If
2507 @command{ld} detects any cross references between the sections, it reports
2508 an error and returns a non-zero exit status. Note that the
2509 @code{NOCROSSREFS} command uses output section names, not input section
2512 @ifclear SingleFormat
2513 @item OUTPUT_ARCH(@var{bfdarch})
2514 @kindex OUTPUT_ARCH(@var{bfdarch})
2515 @cindex machine architecture
2516 @cindex architecture
2517 Specify a particular output machine architecture. The argument is one
2518 of the names used by the BFD library (@pxref{BFD}). You can see the
2519 architecture of an object file by using the @code{objdump} program with
2520 the @samp{-f} option.
2525 @section Assigning Values to Symbols
2526 @cindex assignment in scripts
2527 @cindex symbol definition, scripts
2528 @cindex variables, defining
2529 You may assign a value to a symbol in a linker script. This will define
2530 the symbol as a global symbol.
2533 * Simple Assignments:: Simple Assignments
2537 @node Simple Assignments
2538 @subsection Simple Assignments
2540 You may assign to a symbol using any of the C assignment operators:
2543 @item @var{symbol} = @var{expression} ;
2544 @itemx @var{symbol} += @var{expression} ;
2545 @itemx @var{symbol} -= @var{expression} ;
2546 @itemx @var{symbol} *= @var{expression} ;
2547 @itemx @var{symbol} /= @var{expression} ;
2548 @itemx @var{symbol} <<= @var{expression} ;
2549 @itemx @var{symbol} >>= @var{expression} ;
2550 @itemx @var{symbol} &= @var{expression} ;
2551 @itemx @var{symbol} |= @var{expression} ;
2554 The first case will define @var{symbol} to the value of
2555 @var{expression}. In the other cases, @var{symbol} must already be
2556 defined, and the value will be adjusted accordingly.
2558 The special symbol name @samp{.} indicates the location counter. You
2559 may only use this within a @code{SECTIONS} command.
2561 The semicolon after @var{expression} is required.
2563 Expressions are defined below; see @ref{Expressions}.
2565 You may write symbol assignments as commands in their own right, or as
2566 statements within a @code{SECTIONS} command, or as part of an output
2567 section description in a @code{SECTIONS} command.
2569 The section of the symbol will be set from the section of the
2570 expression; for more information, see @ref{Expression Section}.
2572 Here is an example showing the three different places that symbol
2573 assignments may be used:
2584 _bdata = (. + 3) & ~ 3;
2585 .data : @{ *(.data) @}
2589 In this example, the symbol @samp{floating_point} will be defined as
2590 zero. The symbol @samp{_etext} will be defined as the address following
2591 the last @samp{.text} input section. The symbol @samp{_bdata} will be
2592 defined as the address following the @samp{.text} output section aligned
2593 upward to a 4 byte boundary.
2598 In some cases, it is desirable for a linker script to define a symbol
2599 only if it is referenced and is not defined by any object included in
2600 the link. For example, traditional linkers defined the symbol
2601 @samp{etext}. However, ANSI C requires that the user be able to use
2602 @samp{etext} as a function name without encountering an error. The
2603 @code{PROVIDE} keyword may be used to define a symbol, such as
2604 @samp{etext}, only if it is referenced but not defined. The syntax is
2605 @code{PROVIDE(@var{symbol} = @var{expression})}.
2607 Here is an example of using @code{PROVIDE} to define @samp{etext}:
2620 In this example, if the program defines @samp{_etext} (with a leading
2621 underscore), the linker will give a multiple definition error. If, on
2622 the other hand, the program defines @samp{etext} (with no leading
2623 underscore), the linker will silently use the definition in the program.
2624 If the program references @samp{etext} but does not define it, the
2625 linker will use the definition in the linker script.
2628 @section SECTIONS Command
2630 The @code{SECTIONS} command tells the linker how to map input sections
2631 into output sections, and how to place the output sections in memory.
2633 The format of the @code{SECTIONS} command is:
2637 @var{sections-command}
2638 @var{sections-command}
2643 Each @var{sections-command} may of be one of the following:
2647 an @code{ENTRY} command (@pxref{Entry Point,,Entry command})
2649 a symbol assignment (@pxref{Assignments})
2651 an output section description
2653 an overlay description
2656 The @code{ENTRY} command and symbol assignments are permitted inside the
2657 @code{SECTIONS} command for convenience in using the location counter in
2658 those commands. This can also make the linker script easier to
2659 understand because you can use those commands at meaningful points in
2660 the layout of the output file.
2662 Output section descriptions and overlay descriptions are described
2665 If you do not use a @code{SECTIONS} command in your linker script, the
2666 linker will place each input section into an identically named output
2667 section in the order that the sections are first encountered in the
2668 input files. If all input sections are present in the first file, for
2669 example, the order of sections in the output file will match the order
2670 in the first input file. The first section will be at address zero.
2673 * Output Section Description:: Output section description
2674 * Output Section Name:: Output section name
2675 * Output Section Address:: Output section address
2676 * Input Section:: Input section description
2677 * Output Section Data:: Output section data
2678 * Output Section Keywords:: Output section keywords
2679 * Output Section Discarding:: Output section discarding
2680 * Output Section Attributes:: Output section attributes
2681 * Overlay Description:: Overlay description
2684 @node Output Section Description
2685 @subsection Output Section Description
2686 The full description of an output section looks like this:
2689 @var{section} [@var{address}] [(@var{type})] :
2690 [AT(@var{lma})] [SUBALIGN(@var{subsection_align})]
2692 @var{output-section-command}
2693 @var{output-section-command}
2695 @} [>@var{region}] [AT>@var{lma_region}] [:@var{phdr} :@var{phdr} @dots{}] [=@var{fillexp}]
2699 Most output sections do not use most of the optional section attributes.
2701 The whitespace around @var{section} is required, so that the section
2702 name is unambiguous. The colon and the curly braces are also required.
2703 The line breaks and other white space are optional.
2705 Each @var{output-section-command} may be one of the following:
2709 a symbol assignment (@pxref{Assignments})
2711 an input section description (@pxref{Input Section})
2713 data values to include directly (@pxref{Output Section Data})
2715 a special output section keyword (@pxref{Output Section Keywords})
2718 @node Output Section Name
2719 @subsection Output Section Name
2720 @cindex name, section
2721 @cindex section name
2722 The name of the output section is @var{section}. @var{section} must
2723 meet the constraints of your output format. In formats which only
2724 support a limited number of sections, such as @code{a.out}, the name
2725 must be one of the names supported by the format (@code{a.out}, for
2726 example, allows only @samp{.text}, @samp{.data} or @samp{.bss}). If the
2727 output format supports any number of sections, but with numbers and not
2728 names (as is the case for Oasys), the name should be supplied as a
2729 quoted numeric string. A section name may consist of any sequence of
2730 characters, but a name which contains any unusual characters such as
2731 commas must be quoted.
2733 The output section name @samp{/DISCARD/} is special; @ref{Output Section
2736 @node Output Section Address
2737 @subsection Output Section Description
2738 @cindex address, section
2739 @cindex section address
2740 The @var{address} is an expression for the VMA (the virtual memory
2741 address) of the output section. If you do not provide @var{address},
2742 the linker will set it based on @var{region} if present, or otherwise
2743 based on the current value of the location counter.
2745 If you provide @var{address}, the address of the output section will be
2746 set to precisely that. If you provide neither @var{address} nor
2747 @var{region}, then the address of the output section will be set to the
2748 current value of the location counter aligned to the alignment
2749 requirements of the output section. The alignment requirement of the
2750 output section is the strictest alignment of any input section contained
2751 within the output section.
2755 .text . : @{ *(.text) @}
2760 .text : @{ *(.text) @}
2763 are subtly different. The first will set the address of the
2764 @samp{.text} output section to the current value of the location
2765 counter. The second will set it to the current value of the location
2766 counter aligned to the strictest alignment of a @samp{.text} input
2769 The @var{address} may be an arbitrary expression; @ref{Expressions}.
2770 For example, if you want to align the section on a 0x10 byte boundary,
2771 so that the lowest four bits of the section address are zero, you could
2772 do something like this:
2774 .text ALIGN(0x10) : @{ *(.text) @}
2777 This works because @code{ALIGN} returns the current location counter
2778 aligned upward to the specified value.
2780 Specifying @var{address} for a section will change the value of the
2784 @subsection Input Section Description
2785 @cindex input sections
2786 @cindex mapping input sections to output sections
2787 The most common output section command is an input section description.
2789 The input section description is the most basic linker script operation.
2790 You use output sections to tell the linker how to lay out your program
2791 in memory. You use input section descriptions to tell the linker how to
2792 map the input files into your memory layout.
2795 * Input Section Basics:: Input section basics
2796 * Input Section Wildcards:: Input section wildcard patterns
2797 * Input Section Common:: Input section for common symbols
2798 * Input Section Keep:: Input section and garbage collection
2799 * Input Section Example:: Input section example
2802 @node Input Section Basics
2803 @subsubsection Input Section Basics
2804 @cindex input section basics
2805 An input section description consists of a file name optionally followed
2806 by a list of section names in parentheses.
2808 The file name and the section name may be wildcard patterns, which we
2809 describe further below (@pxref{Input Section Wildcards}).
2811 The most common input section description is to include all input
2812 sections with a particular name in the output section. For example, to
2813 include all input @samp{.text} sections, you would write:
2818 Here the @samp{*} is a wildcard which matches any file name. To exclude a list
2819 of files from matching the file name wildcard, EXCLUDE_FILE may be used to
2820 match all files except the ones specified in the EXCLUDE_FILE list. For
2823 (*(EXCLUDE_FILE (*crtend.o *otherfile.o) .ctors))
2825 will cause all .ctors sections from all files except @file{crtend.o} and
2826 @file{otherfile.o} to be included.
2828 There are two ways to include more than one section:
2834 The difference between these is the order in which the @samp{.text} and
2835 @samp{.rdata} input sections will appear in the output section. In the
2836 first example, they will be intermingled, appearing in the same order as
2837 they are found in the linker input. In the second example, all
2838 @samp{.text} input sections will appear first, followed by all
2839 @samp{.rdata} input sections.
2841 You can specify a file name to include sections from a particular file.
2842 You would do this if one or more of your files contain special data that
2843 needs to be at a particular location in memory. For example:
2848 If you use a file name without a list of sections, then all sections in
2849 the input file will be included in the output section. This is not
2850 commonly done, but it may by useful on occasion. For example:
2855 When you use a file name which does not contain any wild card
2856 characters, the linker will first see if you also specified the file
2857 name on the linker command line or in an @code{INPUT} command. If you
2858 did not, the linker will attempt to open the file as an input file, as
2859 though it appeared on the command line. Note that this differs from an
2860 @code{INPUT} command, because the linker will not search for the file in
2861 the archive search path.
2863 @node Input Section Wildcards
2864 @subsubsection Input Section Wildcard Patterns
2865 @cindex input section wildcards
2866 @cindex wildcard file name patterns
2867 @cindex file name wildcard patterns
2868 @cindex section name wildcard patterns
2869 In an input section description, either the file name or the section
2870 name or both may be wildcard patterns.
2872 The file name of @samp{*} seen in many examples is a simple wildcard
2873 pattern for the file name.
2875 The wildcard patterns are like those used by the Unix shell.
2879 matches any number of characters
2881 matches any single character
2883 matches a single instance of any of the @var{chars}; the @samp{-}
2884 character may be used to specify a range of characters, as in
2885 @samp{[a-z]} to match any lower case letter
2887 quotes the following character
2890 When a file name is matched with a wildcard, the wildcard characters
2891 will not match a @samp{/} character (used to separate directory names on
2892 Unix). A pattern consisting of a single @samp{*} character is an
2893 exception; it will always match any file name, whether it contains a
2894 @samp{/} or not. In a section name, the wildcard characters will match
2895 a @samp{/} character.
2897 File name wildcard patterns only match files which are explicitly
2898 specified on the command line or in an @code{INPUT} command. The linker
2899 does not search directories to expand wildcards.
2901 If a file name matches more than one wildcard pattern, or if a file name
2902 appears explicitly and is also matched by a wildcard pattern, the linker
2903 will use the first match in the linker script. For example, this
2904 sequence of input section descriptions is probably in error, because the
2905 @file{data.o} rule will not be used:
2907 .data : @{ *(.data) @}
2908 .data1 : @{ data.o(.data) @}
2912 Normally, the linker will place files and sections matched by wildcards
2913 in the order in which they are seen during the link. You can change
2914 this by using the @code{SORT} keyword, which appears before a wildcard
2915 pattern in parentheses (e.g., @code{SORT(.text*)}). When the
2916 @code{SORT} keyword is used, the linker will sort the files or sections
2917 into ascending order by name before placing them in the output file.
2919 If you ever get confused about where input sections are going, use the
2920 @samp{-M} linker option to generate a map file. The map file shows
2921 precisely how input sections are mapped to output sections.
2923 This example shows how wildcard patterns might be used to partition
2924 files. This linker script directs the linker to place all @samp{.text}
2925 sections in @samp{.text} and all @samp{.bss} sections in @samp{.bss}.
2926 The linker will place the @samp{.data} section from all files beginning
2927 with an upper case character in @samp{.DATA}; for all other files, the
2928 linker will place the @samp{.data} section in @samp{.data}.
2932 .text : @{ *(.text) @}
2933 .DATA : @{ [A-Z]*(.data) @}
2934 .data : @{ *(.data) @}
2935 .bss : @{ *(.bss) @}
2940 @node Input Section Common
2941 @subsubsection Input Section for Common Symbols
2942 @cindex common symbol placement
2943 @cindex uninitialized data placement
2944 A special notation is needed for common symbols, because in many object
2945 file formats common symbols do not have a particular input section. The
2946 linker treats common symbols as though they are in an input section
2947 named @samp{COMMON}.
2949 You may use file names with the @samp{COMMON} section just as with any
2950 other input sections. You can use this to place common symbols from a
2951 particular input file in one section while common symbols from other
2952 input files are placed in another section.
2954 In most cases, common symbols in input files will be placed in the
2955 @samp{.bss} section in the output file. For example:
2957 .bss @{ *(.bss) *(COMMON) @}
2960 @cindex scommon section
2961 @cindex small common symbols
2962 Some object file formats have more than one type of common symbol. For
2963 example, the MIPS ELF object file format distinguishes standard common
2964 symbols and small common symbols. In this case, the linker will use a
2965 different special section name for other types of common symbols. In
2966 the case of MIPS ELF, the linker uses @samp{COMMON} for standard common
2967 symbols and @samp{.scommon} for small common symbols. This permits you
2968 to map the different types of common symbols into memory at different
2972 You will sometimes see @samp{[COMMON]} in old linker scripts. This
2973 notation is now considered obsolete. It is equivalent to
2976 @node Input Section Keep
2977 @subsubsection Input Section and Garbage Collection
2979 @cindex garbage collection
2980 When link-time garbage collection is in use (@samp{--gc-sections}),
2981 it is often useful to mark sections that should not be eliminated.
2982 This is accomplished by surrounding an input section's wildcard entry
2983 with @code{KEEP()}, as in @code{KEEP(*(.init))} or
2984 @code{KEEP(SORT(*)(.ctors))}.
2986 @node Input Section Example
2987 @subsubsection Input Section Example
2988 The following example is a complete linker script. It tells the linker
2989 to read all of the sections from file @file{all.o} and place them at the
2990 start of output section @samp{outputa} which starts at location
2991 @samp{0x10000}. All of section @samp{.input1} from file @file{foo.o}
2992 follows immediately, in the same output section. All of section
2993 @samp{.input2} from @file{foo.o} goes into output section
2994 @samp{outputb}, followed by section @samp{.input1} from @file{foo1.o}.
2995 All of the remaining @samp{.input1} and @samp{.input2} sections from any
2996 files are written to output section @samp{outputc}.
3024 @node Output Section Data
3025 @subsection Output Section Data
3027 @cindex section data
3028 @cindex output section data
3029 @kindex BYTE(@var{expression})
3030 @kindex SHORT(@var{expression})
3031 @kindex LONG(@var{expression})
3032 @kindex QUAD(@var{expression})
3033 @kindex SQUAD(@var{expression})
3034 You can include explicit bytes of data in an output section by using
3035 @code{BYTE}, @code{SHORT}, @code{LONG}, @code{QUAD}, or @code{SQUAD} as
3036 an output section command. Each keyword is followed by an expression in
3037 parentheses providing the value to store (@pxref{Expressions}). The
3038 value of the expression is stored at the current value of the location
3041 The @code{BYTE}, @code{SHORT}, @code{LONG}, and @code{QUAD} commands
3042 store one, two, four, and eight bytes (respectively). After storing the
3043 bytes, the location counter is incremented by the number of bytes
3046 For example, this will store the byte 1 followed by the four byte value
3047 of the symbol @samp{addr}:
3053 When using a 64 bit host or target, @code{QUAD} and @code{SQUAD} are the
3054 same; they both store an 8 byte, or 64 bit, value. When both host and
3055 target are 32 bits, an expression is computed as 32 bits. In this case
3056 @code{QUAD} stores a 32 bit value zero extended to 64 bits, and
3057 @code{SQUAD} stores a 32 bit value sign extended to 64 bits.
3059 If the object file format of the output file has an explicit endianness,
3060 which is the normal case, the value will be stored in that endianness.
3061 When the object file format does not have an explicit endianness, as is
3062 true of, for example, S-records, the value will be stored in the
3063 endianness of the first input object file.
3065 Note---these commands only work inside a section description and not
3066 between them, so the following will produce an error from the linker:
3068 SECTIONS @{@ .text : @{@ *(.text) @}@ LONG(1) .data : @{@ *(.data) @}@ @}@
3070 whereas this will work:
3072 SECTIONS @{@ .text : @{@ *(.text) ; LONG(1) @}@ .data : @{@ *(.data) @}@ @}@
3075 @kindex FILL(@var{expression})
3076 @cindex holes, filling
3077 @cindex unspecified memory
3078 You may use the @code{FILL} command to set the fill pattern for the
3079 current section. It is followed by an expression in parentheses. Any
3080 otherwise unspecified regions of memory within the section (for example,
3081 gaps left due to the required alignment of input sections) are filled
3082 with the value of the expression, repeated as
3083 necessary. A @code{FILL} statement covers memory locations after the
3084 point at which it occurs in the section definition; by including more
3085 than one @code{FILL} statement, you can have different fill patterns in
3086 different parts of an output section.
3088 This example shows how to fill unspecified regions of memory with the
3094 The @code{FILL} command is similar to the @samp{=@var{fillexp}} output
3095 section attribute, but it only affects the
3096 part of the section following the @code{FILL} command, rather than the
3097 entire section. If both are used, the @code{FILL} command takes
3098 precedence. @xref{Output Section Fill}, for details on the fill
3101 @node Output Section Keywords
3102 @subsection Output Section Keywords
3103 There are a couple of keywords which can appear as output section
3107 @kindex CREATE_OBJECT_SYMBOLS
3108 @cindex input filename symbols
3109 @cindex filename symbols
3110 @item CREATE_OBJECT_SYMBOLS
3111 The command tells the linker to create a symbol for each input file.
3112 The name of each symbol will be the name of the corresponding input
3113 file. The section of each symbol will be the output section in which
3114 the @code{CREATE_OBJECT_SYMBOLS} command appears.
3116 This is conventional for the a.out object file format. It is not
3117 normally used for any other object file format.
3119 @kindex CONSTRUCTORS
3120 @cindex C++ constructors, arranging in link
3121 @cindex constructors, arranging in link
3123 When linking using the a.out object file format, the linker uses an
3124 unusual set construct to support C++ global constructors and
3125 destructors. When linking object file formats which do not support
3126 arbitrary sections, such as ECOFF and XCOFF, the linker will
3127 automatically recognize C++ global constructors and destructors by name.
3128 For these object file formats, the @code{CONSTRUCTORS} command tells the
3129 linker to place constructor information in the output section where the
3130 @code{CONSTRUCTORS} command appears. The @code{CONSTRUCTORS} command is
3131 ignored for other object file formats.
3133 The symbol @w{@code{__CTOR_LIST__}} marks the start of the global
3134 constructors, and the symbol @w{@code{__DTOR_LIST}} marks the end. The
3135 first word in the list is the number of entries, followed by the address
3136 of each constructor or destructor, followed by a zero word. The
3137 compiler must arrange to actually run the code. For these object file
3138 formats @sc{gnu} C++ normally calls constructors from a subroutine
3139 @code{__main}; a call to @code{__main} is automatically inserted into
3140 the startup code for @code{main}. @sc{gnu} C++ normally runs
3141 destructors either by using @code{atexit}, or directly from the function
3144 For object file formats such as @code{COFF} or @code{ELF} which support
3145 arbitrary section names, @sc{gnu} C++ will normally arrange to put the
3146 addresses of global constructors and destructors into the @code{.ctors}
3147 and @code{.dtors} sections. Placing the following sequence into your
3148 linker script will build the sort of table which the @sc{gnu} C++
3149 runtime code expects to see.
3153 LONG((__CTOR_END__ - __CTOR_LIST__) / 4 - 2)
3158 LONG((__DTOR_END__ - __DTOR_LIST__) / 4 - 2)
3164 If you are using the @sc{gnu} C++ support for initialization priority,
3165 which provides some control over the order in which global constructors
3166 are run, you must sort the constructors at link time to ensure that they
3167 are executed in the correct order. When using the @code{CONSTRUCTORS}
3168 command, use @samp{SORT(CONSTRUCTORS)} instead. When using the
3169 @code{.ctors} and @code{.dtors} sections, use @samp{*(SORT(.ctors))} and
3170 @samp{*(SORT(.dtors))} instead of just @samp{*(.ctors)} and
3173 Normally the compiler and linker will handle these issues automatically,
3174 and you will not need to concern yourself with them. However, you may
3175 need to consider this if you are using C++ and writing your own linker
3180 @node Output Section Discarding
3181 @subsection Output Section Discarding
3182 @cindex discarding sections
3183 @cindex sections, discarding
3184 @cindex removing sections
3185 The linker will not create output section which do not have any
3186 contents. This is for convenience when referring to input sections that
3187 may or may not be present in any of the input files. For example:
3192 will only create a @samp{.foo} section in the output file if there is a
3193 @samp{.foo} section in at least one input file.
3195 If you use anything other than an input section description as an output
3196 section command, such as a symbol assignment, then the output section
3197 will always be created, even if there are no matching input sections.
3200 The special output section name @samp{/DISCARD/} may be used to discard
3201 input sections. Any input sections which are assigned to an output
3202 section named @samp{/DISCARD/} are not included in the output file.
3204 @node Output Section Attributes
3205 @subsection Output Section Attributes
3206 @cindex output section attributes
3207 We showed above that the full description of an output section looked
3211 @var{section} [@var{address}] [(@var{type})] :
3212 [AT(@var{lma})] [SUBALIGN(@var{subsection_align})]
3214 @var{output-section-command}
3215 @var{output-section-command}
3217 @} [>@var{region}] [AT>@var{lma_region}] [:@var{phdr} :@var{phdr} @dots{}] [=@var{fillexp}]
3220 We've already described @var{section}, @var{address}, and
3221 @var{output-section-command}. In this section we will describe the
3222 remaining section attributes.
3225 * Output Section Type:: Output section type
3226 * Output Section LMA:: Output section LMA
3227 * Forced Input Alignment:: Forced Input Alignment
3228 * Output Section Region:: Output section region
3229 * Output Section Phdr:: Output section phdr
3230 * Output Section Fill:: Output section fill
3233 @node Output Section Type
3234 @subsubsection Output Section Type
3235 Each output section may have a type. The type is a keyword in
3236 parentheses. The following types are defined:
3240 The section should be marked as not loadable, so that it will not be
3241 loaded into memory when the program is run.
3246 These type names are supported for backward compatibility, and are
3247 rarely used. They all have the same effect: the section should be
3248 marked as not allocatable, so that no memory is allocated for the
3249 section when the program is run.
3253 @cindex prevent unnecessary loading
3254 @cindex loading, preventing
3255 The linker normally sets the attributes of an output section based on
3256 the input sections which map into it. You can override this by using
3257 the section type. For example, in the script sample below, the
3258 @samp{ROM} section is addressed at memory location @samp{0} and does not
3259 need to be loaded when the program is run. The contents of the
3260 @samp{ROM} section will appear in the linker output file as usual.
3264 ROM 0 (NOLOAD) : @{ @dots{} @}
3270 @node Output Section LMA
3271 @subsubsection Output Section LMA
3272 @kindex AT>@var{lma_region}
3273 @kindex AT(@var{lma})
3274 @cindex load address
3275 @cindex section load address
3276 Every section has a virtual address (VMA) and a load address (LMA); see
3277 @ref{Basic Script Concepts}. The address expression which may appear in
3278 an output section description sets the VMA (@pxref{Output Section
3281 The linker will normally set the LMA equal to the VMA. You can change
3282 that by using the @code{AT} keyword. The expression @var{lma} that
3283 follows the @code{AT} keyword specifies the load address of the
3284 section. Alternatively, with @samp{AT>@var{lma_region}} expression,
3285 you may specify a memory region for the section's load address. @xref{MEMORY}.
3287 @cindex ROM initialized data
3288 @cindex initialized data in ROM
3289 This feature is designed to make it easy to build a ROM image. For
3290 example, the following linker script creates three output sections: one
3291 called @samp{.text}, which starts at @code{0x1000}, one called
3292 @samp{.mdata}, which is loaded at the end of the @samp{.text} section
3293 even though its VMA is @code{0x2000}, and one called @samp{.bss} to hold
3294 uninitialized data at address @code{0x3000}. The symbol @code{_data} is
3295 defined with the value @code{0x2000}, which shows that the location
3296 counter holds the VMA value, not the LMA value.
3302 .text 0x1000 : @{ *(.text) _etext = . ; @}
3304 AT ( ADDR (.text) + SIZEOF (.text) )
3305 @{ _data = . ; *(.data); _edata = . ; @}
3307 @{ _bstart = . ; *(.bss) *(COMMON) ; _bend = . ;@}
3312 The run-time initialization code for use with a program generated with
3313 this linker script would include something like the following, to copy
3314 the initialized data from the ROM image to its runtime address. Notice
3315 how this code takes advantage of the symbols defined by the linker
3320 extern char _etext, _data, _edata, _bstart, _bend;
3321 char *src = &_etext;
3324 /* ROM has data at end of text; copy it. */
3325 while (dst < &_edata) @{
3330 for (dst = &_bstart; dst< &_bend; dst++)
3335 @node Forced Input Alignment
3336 @subsubsection Forced Input Alignment
3337 @kindex SUBALIGN(@var{subsection_align})
3338 @cindex forcing input section alignment
3339 @cindex input section alignment
3340 You can force input section alignment within an output section by using
3341 SUBALIGN. The value specified overrides any alignment given by input
3342 sections, whether larger or smaller.
3344 @node Output Section Region
3345 @subsubsection Output Section Region
3346 @kindex >@var{region}
3347 @cindex section, assigning to memory region
3348 @cindex memory regions and sections
3349 You can assign a section to a previously defined region of memory by
3350 using @samp{>@var{region}}. @xref{MEMORY}.
3352 Here is a simple example:
3355 MEMORY @{ rom : ORIGIN = 0x1000, LENGTH = 0x1000 @}
3356 SECTIONS @{ ROM : @{ *(.text) @} >rom @}
3360 @node Output Section Phdr
3361 @subsubsection Output Section Phdr
3363 @cindex section, assigning to program header
3364 @cindex program headers and sections
3365 You can assign a section to a previously defined program segment by
3366 using @samp{:@var{phdr}}. @xref{PHDRS}. If a section is assigned to
3367 one or more segments, then all subsequent allocated sections will be
3368 assigned to those segments as well, unless they use an explicitly
3369 @code{:@var{phdr}} modifier. You can use @code{:NONE} to tell the
3370 linker to not put the section in any segment at all.
3372 Here is a simple example:
3375 PHDRS @{ text PT_LOAD ; @}
3376 SECTIONS @{ .text : @{ *(.text) @} :text @}
3380 @node Output Section Fill
3381 @subsubsection Output Section Fill
3382 @kindex =@var{fillexp}
3383 @cindex section fill pattern
3384 @cindex fill pattern, entire section
3385 You can set the fill pattern for an entire section by using
3386 @samp{=@var{fillexp}}. @var{fillexp} is an expression
3387 (@pxref{Expressions}). Any otherwise unspecified regions of memory
3388 within the output section (for example, gaps left due to the required
3389 alignment of input sections) will be filled with the value, repeated as
3390 necessary. If the fill expression is a simple hex number, ie. a string
3391 of hex digit starting with @samp{0x} and without a trailing @samp{k} or @samp{M}, then
3392 an arbitrarily long sequence of hex digits can be used to specify the
3393 fill pattern; Leading zeros become part of the pattern too. For all
3394 other cases, including extra parentheses or a unary @code{+}, the fill
3395 pattern is the four least significant bytes of the value of the
3396 expression. In all cases, the number is big-endian.
3398 You can also change the fill value with a @code{FILL} command in the
3399 output section commands; (@pxref{Output Section Data}).
3401 Here is a simple example:
3404 SECTIONS @{ .text : @{ *(.text) @} =0x90909090 @}
3408 @node Overlay Description
3409 @subsection Overlay Description
3412 An overlay description provides an easy way to describe sections which
3413 are to be loaded as part of a single memory image but are to be run at
3414 the same memory address. At run time, some sort of overlay manager will
3415 copy the overlaid sections in and out of the runtime memory address as
3416 required, perhaps by simply manipulating addressing bits. This approach
3417 can be useful, for example, when a certain region of memory is faster
3420 Overlays are described using the @code{OVERLAY} command. The
3421 @code{OVERLAY} command is used within a @code{SECTIONS} command, like an
3422 output section description. The full syntax of the @code{OVERLAY}
3423 command is as follows:
3426 OVERLAY [@var{start}] : [NOCROSSREFS] [AT ( @var{ldaddr} )]
3430 @var{output-section-command}
3431 @var{output-section-command}
3433 @} [:@var{phdr}@dots{}] [=@var{fill}]
3436 @var{output-section-command}
3437 @var{output-section-command}
3439 @} [:@var{phdr}@dots{}] [=@var{fill}]
3441 @} [>@var{region}] [:@var{phdr}@dots{}] [=@var{fill}]
3445 Everything is optional except @code{OVERLAY} (a keyword), and each
3446 section must have a name (@var{secname1} and @var{secname2} above). The
3447 section definitions within the @code{OVERLAY} construct are identical to
3448 those within the general @code{SECTIONS} contruct (@pxref{SECTIONS}),
3449 except that no addresses and no memory regions may be defined for
3450 sections within an @code{OVERLAY}.
3452 The sections are all defined with the same starting address. The load
3453 addresses of the sections are arranged such that they are consecutive in
3454 memory starting at the load address used for the @code{OVERLAY} as a
3455 whole (as with normal section definitions, the load address is optional,
3456 and defaults to the start address; the start address is also optional,
3457 and defaults to the current value of the location counter).
3459 If the @code{NOCROSSREFS} keyword is used, and there any references
3460 among the sections, the linker will report an error. Since the sections
3461 all run at the same address, it normally does not make sense for one
3462 section to refer directly to another. @xref{Miscellaneous Commands,
3465 For each section within the @code{OVERLAY}, the linker automatically
3466 defines two symbols. The symbol @code{__load_start_@var{secname}} is
3467 defined as the starting load address of the section. The symbol
3468 @code{__load_stop_@var{secname}} is defined as the final load address of
3469 the section. Any characters within @var{secname} which are not legal
3470 within C identifiers are removed. C (or assembler) code may use these
3471 symbols to move the overlaid sections around as necessary.
3473 At the end of the overlay, the value of the location counter is set to
3474 the start address of the overlay plus the size of the largest section.
3476 Here is an example. Remember that this would appear inside a
3477 @code{SECTIONS} construct.
3480 OVERLAY 0x1000 : AT (0x4000)
3482 .text0 @{ o1/*.o(.text) @}
3483 .text1 @{ o2/*.o(.text) @}
3488 This will define both @samp{.text0} and @samp{.text1} to start at
3489 address 0x1000. @samp{.text0} will be loaded at address 0x4000, and
3490 @samp{.text1} will be loaded immediately after @samp{.text0}. The
3491 following symbols will be defined: @code{__load_start_text0},
3492 @code{__load_stop_text0}, @code{__load_start_text1},
3493 @code{__load_stop_text1}.
3495 C code to copy overlay @code{.text1} into the overlay area might look
3500 extern char __load_start_text1, __load_stop_text1;
3501 memcpy ((char *) 0x1000, &__load_start_text1,
3502 &__load_stop_text1 - &__load_start_text1);
3506 Note that the @code{OVERLAY} command is just syntactic sugar, since
3507 everything it does can be done using the more basic commands. The above
3508 example could have been written identically as follows.
3512 .text0 0x1000 : AT (0x4000) @{ o1/*.o(.text) @}
3513 __load_start_text0 = LOADADDR (.text0);
3514 __load_stop_text0 = LOADADDR (.text0) + SIZEOF (.text0);
3515 .text1 0x1000 : AT (0x4000 + SIZEOF (.text0)) @{ o2/*.o(.text) @}
3516 __load_start_text1 = LOADADDR (.text1);
3517 __load_stop_text1 = LOADADDR (.text1) + SIZEOF (.text1);
3518 . = 0x1000 + MAX (SIZEOF (.text0), SIZEOF (.text1));
3523 @section MEMORY Command
3525 @cindex memory regions
3526 @cindex regions of memory
3527 @cindex allocating memory
3528 @cindex discontinuous memory
3529 The linker's default configuration permits allocation of all available
3530 memory. You can override this by using the @code{MEMORY} command.
3532 The @code{MEMORY} command describes the location and size of blocks of
3533 memory in the target. You can use it to describe which memory regions
3534 may be used by the linker, and which memory regions it must avoid. You
3535 can then assign sections to particular memory regions. The linker will
3536 set section addresses based on the memory regions, and will warn about
3537 regions that become too full. The linker will not shuffle sections
3538 around to fit into the available regions.
3540 A linker script may contain at most one use of the @code{MEMORY}
3541 command. However, you can define as many blocks of memory within it as
3542 you wish. The syntax is:
3547 @var{name} [(@var{attr})] : ORIGIN = @var{origin}, LENGTH = @var{len}
3553 The @var{name} is a name used in the linker script to refer to the
3554 region. The region name has no meaning outside of the linker script.
3555 Region names are stored in a separate name space, and will not conflict
3556 with symbol names, file names, or section names. Each memory region
3557 must have a distinct name.
3559 @cindex memory region attributes
3560 The @var{attr} string is an optional list of attributes that specify
3561 whether to use a particular memory region for an input section which is
3562 not explicitly mapped in the linker script. As described in
3563 @ref{SECTIONS}, if you do not specify an output section for some input
3564 section, the linker will create an output section with the same name as
3565 the input section. If you define region attributes, the linker will use
3566 them to select the memory region for the output section that it creates.
3568 The @var{attr} string must consist only of the following characters:
3583 Invert the sense of any of the preceding attributes
3586 If a unmapped section matches any of the listed attributes other than
3587 @samp{!}, it will be placed in the memory region. The @samp{!}
3588 attribute reverses this test, so that an unmapped section will be placed
3589 in the memory region only if it does not match any of the listed
3595 The @var{origin} is an expression for the start address of the memory
3596 region. The expression must evaluate to a constant before memory
3597 allocation is performed, which means that you may not use any section
3598 relative symbols. The keyword @code{ORIGIN} may be abbreviated to
3599 @code{org} or @code{o} (but not, for example, @code{ORG}).
3604 The @var{len} is an expression for the size in bytes of the memory
3605 region. As with the @var{origin} expression, the expression must
3606 evaluate to a constant before memory allocation is performed. The
3607 keyword @code{LENGTH} may be abbreviated to @code{len} or @code{l}.
3609 In the following example, we specify that there are two memory regions
3610 available for allocation: one starting at @samp{0} for 256 kilobytes,
3611 and the other starting at @samp{0x40000000} for four megabytes. The
3612 linker will place into the @samp{rom} memory region every section which
3613 is not explicitly mapped into a memory region, and is either read-only
3614 or executable. The linker will place other sections which are not
3615 explicitly mapped into a memory region into the @samp{ram} memory
3622 rom (rx) : ORIGIN = 0, LENGTH = 256K
3623 ram (!rx) : org = 0x40000000, l = 4M
3628 Once you define a memory region, you can direct the linker to place
3629 specific output sections into that memory region by using the
3630 @samp{>@var{region}} output section attribute. For example, if you have
3631 a memory region named @samp{mem}, you would use @samp{>mem} in the
3632 output section definition. @xref{Output Section Region}. If no address
3633 was specified for the output section, the linker will set the address to
3634 the next available address within the memory region. If the combined
3635 output sections directed to a memory region are too large for the
3636 region, the linker will issue an error message.
3639 @section PHDRS Command
3641 @cindex program headers
3642 @cindex ELF program headers
3643 @cindex program segments
3644 @cindex segments, ELF
3645 The ELF object file format uses @dfn{program headers}, also knows as
3646 @dfn{segments}. The program headers describe how the program should be
3647 loaded into memory. You can print them out by using the @code{objdump}
3648 program with the @samp{-p} option.
3650 When you run an ELF program on a native ELF system, the system loader
3651 reads the program headers in order to figure out how to load the
3652 program. This will only work if the program headers are set correctly.
3653 This manual does not describe the details of how the system loader
3654 interprets program headers; for more information, see the ELF ABI.
3656 The linker will create reasonable program headers by default. However,
3657 in some cases, you may need to specify the program headers more
3658 precisely. You may use the @code{PHDRS} command for this purpose. When
3659 the linker sees the @code{PHDRS} command in the linker script, it will
3660 not create any program headers other than the ones specified.
3662 The linker only pays attention to the @code{PHDRS} command when
3663 generating an ELF output file. In other cases, the linker will simply
3664 ignore @code{PHDRS}.
3666 This is the syntax of the @code{PHDRS} command. The words @code{PHDRS},
3667 @code{FILEHDR}, @code{AT}, and @code{FLAGS} are keywords.
3673 @var{name} @var{type} [ FILEHDR ] [ PHDRS ] [ AT ( @var{address} ) ]
3674 [ FLAGS ( @var{flags} ) ] ;
3679 The @var{name} is used only for reference in the @code{SECTIONS} command
3680 of the linker script. It is not put into the output file. Program
3681 header names are stored in a separate name space, and will not conflict
3682 with symbol names, file names, or section names. Each program header
3683 must have a distinct name.
3685 Certain program header types describe segments of memory which the
3686 system loader will load from the file. In the linker script, you
3687 specify the contents of these segments by placing allocatable output
3688 sections in the segments. You use the @samp{:@var{phdr}} output section
3689 attribute to place a section in a particular segment. @xref{Output
3692 It is normal to put certain sections in more than one segment. This
3693 merely implies that one segment of memory contains another. You may
3694 repeat @samp{:@var{phdr}}, using it once for each segment which should
3695 contain the section.
3697 If you place a section in one or more segments using @samp{:@var{phdr}},
3698 then the linker will place all subsequent allocatable sections which do
3699 not specify @samp{:@var{phdr}} in the same segments. This is for
3700 convenience, since generally a whole set of contiguous sections will be
3701 placed in a single segment. You can use @code{:NONE} to override the
3702 default segment and tell the linker to not put the section in any
3707 You may use the @code{FILEHDR} and @code{PHDRS} keywords appear after
3708 the program header type to further describe the contents of the segment.
3709 The @code{FILEHDR} keyword means that the segment should include the ELF
3710 file header. The @code{PHDRS} keyword means that the segment should
3711 include the ELF program headers themselves.
3713 The @var{type} may be one of the following. The numbers indicate the
3714 value of the keyword.
3717 @item @code{PT_NULL} (0)
3718 Indicates an unused program header.
3720 @item @code{PT_LOAD} (1)
3721 Indicates that this program header describes a segment to be loaded from
3724 @item @code{PT_DYNAMIC} (2)
3725 Indicates a segment where dynamic linking information can be found.
3727 @item @code{PT_INTERP} (3)
3728 Indicates a segment where the name of the program interpreter may be
3731 @item @code{PT_NOTE} (4)
3732 Indicates a segment holding note information.
3734 @item @code{PT_SHLIB} (5)
3735 A reserved program header type, defined but not specified by the ELF
3738 @item @code{PT_PHDR} (6)
3739 Indicates a segment where the program headers may be found.
3741 @item @var{expression}
3742 An expression giving the numeric type of the program header. This may
3743 be used for types not defined above.
3746 You can specify that a segment should be loaded at a particular address
3747 in memory by using an @code{AT} expression. This is identical to the
3748 @code{AT} command used as an output section attribute (@pxref{Output
3749 Section LMA}). The @code{AT} command for a program header overrides the
3750 output section attribute.
3752 The linker will normally set the segment flags based on the sections
3753 which comprise the segment. You may use the @code{FLAGS} keyword to
3754 explicitly specify the segment flags. The value of @var{flags} must be
3755 an integer. It is used to set the @code{p_flags} field of the program
3758 Here is an example of @code{PHDRS}. This shows a typical set of program
3759 headers used on a native ELF system.
3765 headers PT_PHDR PHDRS ;
3767 text PT_LOAD FILEHDR PHDRS ;
3769 dynamic PT_DYNAMIC ;
3775 .interp : @{ *(.interp) @} :text :interp
3776 .text : @{ *(.text) @} :text
3777 .rodata : @{ *(.rodata) @} /* defaults to :text */
3779 . = . + 0x1000; /* move to a new page in memory */
3780 .data : @{ *(.data) @} :data
3781 .dynamic : @{ *(.dynamic) @} :data :dynamic
3788 @section VERSION Command
3789 @kindex VERSION @{script text@}
3790 @cindex symbol versions
3791 @cindex version script
3792 @cindex versions of symbols
3793 The linker supports symbol versions when using ELF. Symbol versions are
3794 only useful when using shared libraries. The dynamic linker can use
3795 symbol versions to select a specific version of a function when it runs
3796 a program that may have been linked against an earlier version of the
3799 You can include a version script directly in the main linker script, or
3800 you can supply the version script as an implicit linker script. You can
3801 also use the @samp{--version-script} linker option.
3803 The syntax of the @code{VERSION} command is simply
3805 VERSION @{ version-script-commands @}
3808 The format of the version script commands is identical to that used by
3809 Sun's linker in Solaris 2.5. The version script defines a tree of
3810 version nodes. You specify the node names and interdependencies in the
3811 version script. You can specify which symbols are bound to which
3812 version nodes, and you can reduce a specified set of symbols to local
3813 scope so that they are not globally visible outside of the shared
3816 The easiest way to demonstrate the version script language is with a few
3838 This example version script defines three version nodes. The first
3839 version node defined is @samp{VERS_1.1}; it has no other dependencies.
3840 The script binds the symbol @samp{foo1} to @samp{VERS_1.1}. It reduces
3841 a number of symbols to local scope so that they are not visible outside
3842 of the shared library; this is done using wildcard patterns, so that any
3843 symbol whose name begins with @samp{old}, @samp{original}, or @samp{new}
3844 is matched. The wildcard patterns available are the same as those used
3845 in the shell when matching filenames (also known as ``globbing'').
3847 Next, the version script defines node @samp{VERS_1.2}. This node
3848 depends upon @samp{VERS_1.1}. The script binds the symbol @samp{foo2}
3849 to the version node @samp{VERS_1.2}.
3851 Finally, the version script defines node @samp{VERS_2.0}. This node
3852 depends upon @samp{VERS_1.2}. The scripts binds the symbols @samp{bar1}
3853 and @samp{bar2} are bound to the version node @samp{VERS_2.0}.
3855 When the linker finds a symbol defined in a library which is not
3856 specifically bound to a version node, it will effectively bind it to an
3857 unspecified base version of the library. You can bind all otherwise
3858 unspecified symbols to a given version node by using @samp{global: *;}
3859 somewhere in the version script.
3861 The names of the version nodes have no specific meaning other than what
3862 they might suggest to the person reading them. The @samp{2.0} version
3863 could just as well have appeared in between @samp{1.1} and @samp{1.2}.
3864 However, this would be a confusing way to write a version script.
3866 Node name can be omited, provided it is the only version node
3867 in the version script. Such version script doesn't assign any versions to
3868 symbols, only selects which symbols will be globally visible out and which
3872 @{ global: foo; bar; local: *; @};
3875 When you link an application against a shared library that has versioned
3876 symbols, the application itself knows which version of each symbol it
3877 requires, and it also knows which version nodes it needs from each
3878 shared library it is linked against. Thus at runtime, the dynamic
3879 loader can make a quick check to make sure that the libraries you have
3880 linked against do in fact supply all of the version nodes that the
3881 application will need to resolve all of the dynamic symbols. In this
3882 way it is possible for the dynamic linker to know with certainty that
3883 all external symbols that it needs will be resolvable without having to
3884 search for each symbol reference.
3886 The symbol versioning is in effect a much more sophisticated way of
3887 doing minor version checking that SunOS does. The fundamental problem
3888 that is being addressed here is that typically references to external
3889 functions are bound on an as-needed basis, and are not all bound when
3890 the application starts up. If a shared library is out of date, a
3891 required interface may be missing; when the application tries to use
3892 that interface, it may suddenly and unexpectedly fail. With symbol
3893 versioning, the user will get a warning when they start their program if
3894 the libraries being used with the application are too old.
3896 There are several GNU extensions to Sun's versioning approach. The
3897 first of these is the ability to bind a symbol to a version node in the
3898 source file where the symbol is defined instead of in the versioning
3899 script. This was done mainly to reduce the burden on the library
3900 maintainer. You can do this by putting something like:
3902 __asm__(".symver original_foo,foo@@VERS_1.1");
3905 in the C source file. This renames the function @samp{original_foo} to
3906 be an alias for @samp{foo} bound to the version node @samp{VERS_1.1}.
3907 The @samp{local:} directive can be used to prevent the symbol
3908 @samp{original_foo} from being exported. A @samp{.symver} directive
3909 takes precedence over a version script.
3911 The second GNU extension is to allow multiple versions of the same
3912 function to appear in a given shared library. In this way you can make
3913 an incompatible change to an interface without increasing the major
3914 version number of the shared library, while still allowing applications
3915 linked against the old interface to continue to function.
3917 To do this, you must use multiple @samp{.symver} directives in the
3918 source file. Here is an example:
3921 __asm__(".symver original_foo,foo@@");
3922 __asm__(".symver old_foo,foo@@VERS_1.1");
3923 __asm__(".symver old_foo1,foo@@VERS_1.2");
3924 __asm__(".symver new_foo,foo@@@@VERS_2.0");
3927 In this example, @samp{foo@@} represents the symbol @samp{foo} bound to the
3928 unspecified base version of the symbol. The source file that contains this
3929 example would define 4 C functions: @samp{original_foo}, @samp{old_foo},
3930 @samp{old_foo1}, and @samp{new_foo}.
3932 When you have multiple definitions of a given symbol, there needs to be
3933 some way to specify a default version to which external references to
3934 this symbol will be bound. You can do this with the
3935 @samp{foo@@@@VERS_2.0} type of @samp{.symver} directive. You can only
3936 declare one version of a symbol as the default in this manner; otherwise
3937 you would effectively have multiple definitions of the same symbol.
3939 If you wish to bind a reference to a specific version of the symbol
3940 within the shared library, you can use the aliases of convenience
3941 (i.e., @samp{old_foo}), or you can use the @samp{.symver} directive to
3942 specifically bind to an external version of the function in question.
3944 You can also specify the language in the version script:
3947 VERSION extern "lang" @{ version-script-commands @}
3950 The supported @samp{lang}s are @samp{C}, @samp{C++}, and @samp{Java}.
3951 The linker will iterate over the list of symbols at the link time and
3952 demangle them according to @samp{lang} before matching them to the
3953 patterns specified in @samp{version-script-commands}.
3956 @section Expressions in Linker Scripts
3959 The syntax for expressions in the linker script language is identical to
3960 that of C expressions. All expressions are evaluated as integers. All
3961 expressions are evaluated in the same size, which is 32 bits if both the
3962 host and target are 32 bits, and is otherwise 64 bits.
3964 You can use and set symbol values in expressions.
3966 The linker defines several special purpose builtin functions for use in
3970 * Constants:: Constants
3971 * Symbols:: Symbol Names
3972 * Location Counter:: The Location Counter
3973 * Operators:: Operators
3974 * Evaluation:: Evaluation
3975 * Expression Section:: The Section of an Expression
3976 * Builtin Functions:: Builtin Functions
3980 @subsection Constants
3981 @cindex integer notation
3982 @cindex constants in linker scripts
3983 All constants are integers.
3985 As in C, the linker considers an integer beginning with @samp{0} to be
3986 octal, and an integer beginning with @samp{0x} or @samp{0X} to be
3987 hexadecimal. The linker considers other integers to be decimal.
3989 @cindex scaled integers
3990 @cindex K and M integer suffixes
3991 @cindex M and K integer suffixes
3992 @cindex suffixes for integers
3993 @cindex integer suffixes
3994 In addition, you can use the suffixes @code{K} and @code{M} to scale a
3998 @c END TEXI2ROFF-KILL
3999 @code{1024} or @code{1024*1024}
4003 ${\rm 1024}$ or ${\rm 1024}^2$
4005 @c END TEXI2ROFF-KILL
4006 respectively. For example, the following all refer to the same quantity:
4014 @subsection Symbol Names
4015 @cindex symbol names
4017 @cindex quoted symbol names
4019 Unless quoted, symbol names start with a letter, underscore, or period
4020 and may include letters, digits, underscores, periods, and hyphens.
4021 Unquoted symbol names must not conflict with any keywords. You can
4022 specify a symbol which contains odd characters or has the same name as a
4023 keyword by surrounding the symbol name in double quotes:
4026 "with a space" = "also with a space" + 10;
4029 Since symbols can contain many non-alphabetic characters, it is safest
4030 to delimit symbols with spaces. For example, @samp{A-B} is one symbol,
4031 whereas @samp{A - B} is an expression involving subtraction.
4033 @node Location Counter
4034 @subsection The Location Counter
4037 @cindex location counter
4038 @cindex current output location
4039 The special linker variable @dfn{dot} @samp{.} always contains the
4040 current output location counter. Since the @code{.} always refers to a
4041 location in an output section, it may only appear in an expression
4042 within a @code{SECTIONS} command. The @code{.} symbol may appear
4043 anywhere that an ordinary symbol is allowed in an expression.
4046 Assigning a value to @code{.} will cause the location counter to be
4047 moved. This may be used to create holes in the output section. The
4048 location counter may never be moved backwards.
4064 In the previous example, the @samp{.text} section from @file{file1} is
4065 located at the beginning of the output section @samp{output}. It is
4066 followed by a 1000 byte gap. Then the @samp{.text} section from
4067 @file{file2} appears, also with a 1000 byte gap following before the
4068 @samp{.text} section from @file{file3}. The notation @samp{= 0x12345678}
4069 specifies what data to write in the gaps (@pxref{Output Section Fill}).
4071 @cindex dot inside sections
4072 Note: @code{.} actually refers to the byte offset from the start of the
4073 current containing object. Normally this is the @code{SECTIONS}
4074 statement, whose start address is 0, hence @code{.} can be used as an
4075 absolute address. If @code{.} is used inside a section description
4076 however, it refers to the byte offset from the start of that section,
4077 not an absolute address. Thus in a script like this:
4095 The @samp{.text} section will be assigned a starting address of 0x100
4096 and a size of exactly 0x200 bytes, even if there is not enough data in
4097 the @samp{.text} input sections to fill this area. (If there is too
4098 much data, an error will be produced because this would be an attempt to
4099 move @code{.} backwards). The @samp{.data} section will start at 0x500
4100 and it will have an extra 0x600 bytes worth of space after the end of
4101 the values from the @samp{.data} input sections and before the end of
4102 the @samp{.data} output section itself.
4106 @subsection Operators
4107 @cindex operators for arithmetic
4108 @cindex arithmetic operators
4109 @cindex precedence in expressions
4110 The linker recognizes the standard C set of arithmetic operators, with
4111 the standard bindings and precedence levels:
4114 @c END TEXI2ROFF-KILL
4116 precedence associativity Operators Notes
4122 5 left == != > < <= >=
4128 11 right &= += -= *= /= (2)
4132 (1) Prefix operators
4133 (2) @xref{Assignments}.
4137 \vskip \baselineskip
4138 %"lispnarrowing" is the extra indent used generally for smallexample
4139 \hskip\lispnarrowing\vbox{\offinterlineskip
4142 {\vrule#&\strut\hfil\ #\ \hfil&\vrule#&\strut\hfil\ #\ \hfil&\vrule#&\strut\hfil\ {\tt #}\ \hfil&\vrule#\cr
4143 height2pt&\omit&&\omit&&\omit&\cr
4144 &Precedence&& Associativity &&{\rm Operators}&\cr
4145 height2pt&\omit&&\omit&&\omit&\cr
4147 height2pt&\omit&&\omit&&\omit&\cr
4149 % '176 is tilde, '~' in tt font
4150 &1&&left&&\qquad- \char'176\ !\qquad\dag&\cr
4151 &2&&left&&* / \%&\cr
4154 &5&&left&&== != > < <= >=&\cr
4157 &8&&left&&{\&\&}&\cr
4160 &11&&right&&\qquad\&= += -= *= /=\qquad\ddag&\cr
4162 height2pt&\omit&&\omit&&\omit&\cr}
4167 @obeylines@parskip=0pt@parindent=0pt
4168 @dag@quad Prefix operators.
4169 @ddag@quad @xref{Assignments}.
4172 @c END TEXI2ROFF-KILL
4175 @subsection Evaluation
4176 @cindex lazy evaluation
4177 @cindex expression evaluation order
4178 The linker evaluates expressions lazily. It only computes the value of
4179 an expression when absolutely necessary.
4181 The linker needs some information, such as the value of the start
4182 address of the first section, and the origins and lengths of memory
4183 regions, in order to do any linking at all. These values are computed
4184 as soon as possible when the linker reads in the linker script.
4186 However, other values (such as symbol values) are not known or needed
4187 until after storage allocation. Such values are evaluated later, when
4188 other information (such as the sizes of output sections) is available
4189 for use in the symbol assignment expression.
4191 The sizes of sections cannot be known until after allocation, so
4192 assignments dependent upon these are not performed until after
4195 Some expressions, such as those depending upon the location counter
4196 @samp{.}, must be evaluated during section allocation.
4198 If the result of an expression is required, but the value is not
4199 available, then an error results. For example, a script like the
4205 .text 9+this_isnt_constant :
4211 will cause the error message @samp{non constant expression for initial
4214 @node Expression Section
4215 @subsection The Section of an Expression
4216 @cindex expression sections
4217 @cindex absolute expressions
4218 @cindex relative expressions
4219 @cindex absolute and relocatable symbols
4220 @cindex relocatable and absolute symbols
4221 @cindex symbols, relocatable and absolute
4222 When the linker evaluates an expression, the result is either absolute
4223 or relative to some section. A relative expression is expressed as a
4224 fixed offset from the base of a section.
4226 The position of the expression within the linker script determines
4227 whether it is absolute or relative. An expression which appears within
4228 an output section definition is relative to the base of the output
4229 section. An expression which appears elsewhere will be absolute.
4231 A symbol set to a relative expression will be relocatable if you request
4232 relocatable output using the @samp{-r} option. That means that a
4233 further link operation may change the value of the symbol. The symbol's
4234 section will be the section of the relative expression.
4236 A symbol set to an absolute expression will retain the same value
4237 through any further link operation. The symbol will be absolute, and
4238 will not have any particular associated section.
4240 You can use the builtin function @code{ABSOLUTE} to force an expression
4241 to be absolute when it would otherwise be relative. For example, to
4242 create an absolute symbol set to the address of the end of the output
4243 section @samp{.data}:
4247 .data : @{ *(.data) _edata = ABSOLUTE(.); @}
4251 If @samp{ABSOLUTE} were not used, @samp{_edata} would be relative to the
4252 @samp{.data} section.
4254 @node Builtin Functions
4255 @subsection Builtin Functions
4256 @cindex functions in expressions
4257 The linker script language includes a number of builtin functions for
4258 use in linker script expressions.
4261 @item ABSOLUTE(@var{exp})
4262 @kindex ABSOLUTE(@var{exp})
4263 @cindex expression, absolute
4264 Return the absolute (non-relocatable, as opposed to non-negative) value
4265 of the expression @var{exp}. Primarily useful to assign an absolute
4266 value to a symbol within a section definition, where symbol values are
4267 normally section relative. @xref{Expression Section}.
4269 @item ADDR(@var{section})
4270 @kindex ADDR(@var{section})
4271 @cindex section address in expression
4272 Return the absolute address (the VMA) of the named @var{section}. Your
4273 script must previously have defined the location of that section. In
4274 the following example, @code{symbol_1} and @code{symbol_2} are assigned
4281 start_of_output_1 = ABSOLUTE(.);
4286 symbol_1 = ADDR(.output1);
4287 symbol_2 = start_of_output_1;
4293 @item ALIGN(@var{exp})
4294 @kindex ALIGN(@var{exp})
4295 @cindex round up location counter
4296 @cindex align location counter
4297 Return the location counter (@code{.}) aligned to the next @var{exp}
4299 @code{ALIGN} doesn't change the value of the location counter---it just
4300 does arithmetic on it. Here is an example which aligns the output
4301 @code{.data} section to the next @code{0x2000} byte boundary after the
4302 preceding section and sets a variable within the section to the next
4303 @code{0x8000} boundary after the input sections:
4307 .data ALIGN(0x2000): @{
4309 variable = ALIGN(0x8000);
4315 The first use of @code{ALIGN} in this example specifies the location of
4316 a section because it is used as the optional @var{address} attribute of
4317 a section definition (@pxref{Output Section Address}). The second use
4318 of @code{ALIGN} is used to defines the value of a symbol.
4320 The builtin function @code{NEXT} is closely related to @code{ALIGN}.
4322 @item BLOCK(@var{exp})
4323 @kindex BLOCK(@var{exp})
4324 This is a synonym for @code{ALIGN}, for compatibility with older linker
4325 scripts. It is most often seen when setting the address of an output
4328 @item DATA_SEGMENT_ALIGN(@var{maxpagesize}, @var{commonpagesize})
4329 @kindex DATA_SEGMENT_ALIGN(@var{maxpagesize}, @var{commonpagesize})
4330 This is equivalent to either
4332 (ALIGN(@var{maxpagesize}) + (. & (@var{maxpagesize} - 1)))
4336 (ALIGN(@var{maxpagesize}) + (. & (@var{maxpagesize} - @var{commonpagesize})))
4339 depending on whether the latter uses fewer @var{commonpagesize} sized pages
4340 for the data segment (area between the result of this expression and
4341 @code{DATA_SEGMENT_END}) than the former or not.
4342 If the latter form is used, it means @var{commonpagesize} bytes of runtime
4343 memory will be saved at the expense of up to @var{commonpagesize} wasted
4344 bytes in the on-disk file.
4346 This expression can only be used directly in @code{SECTIONS} commands, not in
4347 any output section descriptions and only once in the linker script.
4348 @var{commonpagesize} should be less or equal to @var{maxpagesize} and should
4349 be the system page size the object wants to be optimized for (while still
4350 working on system page sizes up to @var{maxpagesize}).
4355 . = DATA_SEGMENT_ALIGN(0x10000, 0x2000);
4358 @item DATA_SEGMENT_END(@var{exp})
4359 @kindex DATA_SEGMENT_END(@var{exp})
4360 This defines the end of data segment for @code{DATA_SEGMENT_ALIGN}
4361 evaluation purposes.
4364 . = DATA_SEGMENT_END(.);
4367 @item DEFINED(@var{symbol})
4368 @kindex DEFINED(@var{symbol})
4369 @cindex symbol defaults
4370 Return 1 if @var{symbol} is in the linker global symbol table and is
4371 defined, otherwise return 0. You can use this function to provide
4372 default values for symbols. For example, the following script fragment
4373 shows how to set a global symbol @samp{begin} to the first location in
4374 the @samp{.text} section---but if a symbol called @samp{begin} already
4375 existed, its value is preserved:
4381 begin = DEFINED(begin) ? begin : . ;
4389 @item LOADADDR(@var{section})
4390 @kindex LOADADDR(@var{section})
4391 @cindex section load address in expression
4392 Return the absolute LMA of the named @var{section}. This is normally
4393 the same as @code{ADDR}, but it may be different if the @code{AT}
4394 attribute is used in the output section definition (@pxref{Output
4398 @item MAX(@var{exp1}, @var{exp2})
4399 Returns the maximum of @var{exp1} and @var{exp2}.
4402 @item MIN(@var{exp1}, @var{exp2})
4403 Returns the minimum of @var{exp1} and @var{exp2}.
4405 @item NEXT(@var{exp})
4406 @kindex NEXT(@var{exp})
4407 @cindex unallocated address, next
4408 Return the next unallocated address that is a multiple of @var{exp}.
4409 This function is closely related to @code{ALIGN(@var{exp})}; unless you
4410 use the @code{MEMORY} command to define discontinuous memory for the
4411 output file, the two functions are equivalent.
4413 @item SIZEOF(@var{section})
4414 @kindex SIZEOF(@var{section})
4415 @cindex section size
4416 Return the size in bytes of the named @var{section}, if that section has
4417 been allocated. If the section has not been allocated when this is
4418 evaluated, the linker will report an error. In the following example,
4419 @code{symbol_1} and @code{symbol_2} are assigned identical values:
4428 symbol_1 = .end - .start ;
4429 symbol_2 = SIZEOF(.output);
4434 @item SIZEOF_HEADERS
4435 @itemx sizeof_headers
4436 @kindex SIZEOF_HEADERS
4438 Return the size in bytes of the output file's headers. This is
4439 information which appears at the start of the output file. You can use
4440 this number when setting the start address of the first section, if you
4441 choose, to facilitate paging.
4443 @cindex not enough room for program headers
4444 @cindex program headers, not enough room
4445 When producing an ELF output file, if the linker script uses the
4446 @code{SIZEOF_HEADERS} builtin function, the linker must compute the
4447 number of program headers before it has determined all the section
4448 addresses and sizes. If the linker later discovers that it needs
4449 additional program headers, it will report an error @samp{not enough
4450 room for program headers}. To avoid this error, you must avoid using
4451 the @code{SIZEOF_HEADERS} function, or you must rework your linker
4452 script to avoid forcing the linker to use additional program headers, or
4453 you must define the program headers yourself using the @code{PHDRS}
4454 command (@pxref{PHDRS}).
4457 @node Implicit Linker Scripts
4458 @section Implicit Linker Scripts
4459 @cindex implicit linker scripts
4460 If you specify a linker input file which the linker can not recognize as
4461 an object file or an archive file, it will try to read the file as a
4462 linker script. If the file can not be parsed as a linker script, the
4463 linker will report an error.
4465 An implicit linker script will not replace the default linker script.
4467 Typically an implicit linker script would contain only symbol
4468 assignments, or the @code{INPUT}, @code{GROUP}, or @code{VERSION}
4471 Any input files read because of an implicit linker script will be read
4472 at the position in the command line where the implicit linker script was
4473 read. This can affect archive searching.
4476 @node Machine Dependent
4477 @chapter Machine Dependent Features
4479 @cindex machine dependencies
4480 @command{ld} has additional features on some platforms; the following
4481 sections describe them. Machines where @command{ld} has no additional
4482 functionality are not listed.
4486 * H8/300:: @command{ld} and the H8/300
4489 * i960:: @command{ld} and the Intel 960 family
4492 * ARM:: @command{ld} and the ARM family
4495 * HPPA ELF32:: @command{ld} and HPPA 32-bit ELF
4498 * MMIX:: @command{ld} and MMIX
4501 * MSP430:: @command{ld} and MSP430
4504 * TI COFF:: @command{ld} and TI COFF
4507 * WIN32:: @command{ld} and WIN32 (cygwin/mingw)
4510 * Xtensa:: @command{ld} and Xtensa Processors
4521 @section @command{ld} and the H8/300
4523 @cindex H8/300 support
4524 For the H8/300, @command{ld} can perform these global optimizations when
4525 you specify the @samp{--relax} command-line option.
4528 @cindex relaxing on H8/300
4529 @item relaxing address modes
4530 @command{ld} finds all @code{jsr} and @code{jmp} instructions whose
4531 targets are within eight bits, and turns them into eight-bit
4532 program-counter relative @code{bsr} and @code{bra} instructions,
4535 @cindex synthesizing on H8/300
4536 @item synthesizing instructions
4537 @c FIXME: specifically mov.b, or any mov instructions really?
4538 @command{ld} finds all @code{mov.b} instructions which use the
4539 sixteen-bit absolute address form, but refer to the top
4540 page of memory, and changes them to use the eight-bit address form.
4541 (That is: the linker turns @samp{mov.b @code{@@}@var{aa}:16} into
4542 @samp{mov.b @code{@@}@var{aa}:8} whenever the address @var{aa} is in the
4543 top page of memory).
4553 @c This stuff is pointless to say unless you're especially concerned
4554 @c with Renesas chips; don't enable it for generic case, please.
4556 @chapter @command{ld} and Other Renesas Chips
4558 @command{ld} also supports the Renesas (formerly Hitachi) H8/300H,
4559 H8/500, and SH chips. No special features, commands, or command-line
4560 options are required for these chips.
4570 @section @command{ld} and the Intel 960 Family
4572 @cindex i960 support
4574 You can use the @samp{-A@var{architecture}} command line option to
4575 specify one of the two-letter names identifying members of the 960
4576 family; the option specifies the desired output target, and warns of any
4577 incompatible instructions in the input files. It also modifies the
4578 linker's search strategy for archive libraries, to support the use of
4579 libraries specific to each particular architecture, by including in the
4580 search loop names suffixed with the string identifying the architecture.
4582 For example, if your @command{ld} command line included @w{@samp{-ACA}} as
4583 well as @w{@samp{-ltry}}, the linker would look (in its built-in search
4584 paths, and in any paths you specify with @samp{-L}) for a library with
4597 The first two possibilities would be considered in any event; the last
4598 two are due to the use of @w{@samp{-ACA}}.
4600 You can meaningfully use @samp{-A} more than once on a command line, since
4601 the 960 architecture family allows combination of target architectures; each
4602 use will add another pair of name variants to search for when @w{@samp{-l}}
4603 specifies a library.
4605 @cindex @option{--relax} on i960
4606 @cindex relaxing on i960
4607 @command{ld} supports the @samp{--relax} option for the i960 family. If
4608 you specify @samp{--relax}, @command{ld} finds all @code{balx} and
4609 @code{calx} instructions whose targets are within 24 bits, and turns
4610 them into 24-bit program-counter relative @code{bal} and @code{cal}
4611 instructions, respectively. @command{ld} also turns @code{cal}
4612 instructions into @code{bal} instructions when it determines that the
4613 target subroutine is a leaf routine (that is, the target subroutine does
4614 not itself call any subroutines).
4627 @section @command{ld}'s Support for Interworking Between ARM and Thumb Code
4629 @cindex ARM interworking support
4630 @kindex --support-old-code
4631 For the ARM, @command{ld} will generate code stubs to allow functions calls
4632 betweem ARM and Thumb code. These stubs only work with code that has
4633 been compiled and assembled with the @samp{-mthumb-interwork} command
4634 line option. If it is necessary to link with old ARM object files or
4635 libraries, which have not been compiled with the -mthumb-interwork
4636 option then the @samp{--support-old-code} command line switch should be
4637 given to the linker. This will make it generate larger stub functions
4638 which will work with non-interworking aware ARM code. Note, however,
4639 the linker does not support generating stubs for function calls to
4640 non-interworking aware Thumb code.
4642 @cindex thumb entry point
4643 @cindex entry point, thumb
4644 @kindex --thumb-entry=@var{entry}
4645 The @samp{--thumb-entry} switch is a duplicate of the generic
4646 @samp{--entry} switch, in that it sets the program's starting address.
4647 But it also sets the bottom bit of the address, so that it can be
4648 branched to using a BX instruction, and the program will start
4649 executing in Thumb mode straight away.
4662 @section @command{ld} and HPPA 32-bit ELF Support
4663 @cindex HPPA multiple sub-space stubs
4664 @kindex --multi-subspace
4665 When generating a shared library, @command{ld} will by default generate
4666 import stubs suitable for use with a single sub-space application.
4667 The @samp{--multi-subspace} switch causes @command{ld} to generate export
4668 stubs, and different (larger) import stubs suitable for use with
4669 multiple sub-spaces.
4671 @cindex HPPA stub grouping
4672 @kindex --stub-group-size=@var{N}
4673 Long branch stubs and import/export stubs are placed by @command{ld} in
4674 stub sections located between groups of input sections.
4675 @samp{--stub-group-size} specifies the maximum size of a group of input
4676 sections handled by one stub section. Since branch offsets are signed,
4677 a stub section may serve two groups of input sections, one group before
4678 the stub section, and one group after it. However, when using
4679 conditional branches that require stubs, it may be better (for branch
4680 prediction) that stub sections only serve one group of input sections.
4681 A negative value for @samp{N} chooses this scheme, ensuring that
4682 branches to stubs always use a negative offset. Two special values of
4683 @samp{N} are recognized, @samp{1} and @samp{-1}. These both instruct
4684 @command{ld} to automatically size input section groups for the branch types
4685 detected, with the same behaviour regarding stub placement as other
4686 positive or negative values of @samp{N} respectively.
4688 Note that @samp{--stub-group-size} does not split input sections. A
4689 single input section larger than the group size specified will of course
4690 create a larger group (of one section). If input sections are too
4691 large, it may not be possible for a branch to reach its stub.
4704 @section @code{ld} and MMIX
4705 For MMIX, there is a choice of generating @code{ELF} object files or
4706 @code{mmo} object files when linking. The simulator @code{mmix}
4707 understands the @code{mmo} format. The binutils @code{objcopy} utility
4708 can translate between the two formats.
4710 There is one special section, the @samp{.MMIX.reg_contents} section.
4711 Contents in this section is assumed to correspond to that of global
4712 registers, and symbols referring to it are translated to special symbols,
4713 equal to registers. In a final link, the start address of the
4714 @samp{.MMIX.reg_contents} section corresponds to the first allocated
4715 global register multiplied by 8. Register @code{$255} is not included in
4716 this section; it is always set to the program entry, which is at the
4717 symbol @code{Main} for @code{mmo} files.
4719 Symbols with the prefix @code{__.MMIX.start.}, for example
4720 @code{__.MMIX.start..text} and @code{__.MMIX.start..data} are special;
4721 there must be only one each, even if they are local. The default linker
4722 script uses these to set the default start address of a section.
4724 Initial and trailing multiples of zero-valued 32-bit words in a section,
4725 are left out from an mmo file.
4738 @section @code{ld} and MSP430
4739 For the MSP430 it is possible to select the MPU architecture. The flag @samp{-m [mpu type]}
4740 will select an appropriate linker script for selected MPU type. (To get a list of known MPUs
4741 just pass @samp{-m help} option to the linker).
4743 @cindex MSP430 extra sections
4744 The linker will recognize some extra sections which are MSP430 specific:
4747 @item @samp{.vectors}
4748 Defines a portion of ROM where interrupt vectors located.
4750 @item @samp{.bootloader}
4751 Defines the bootloader portion of the ROM (if applicable). Any code
4752 in this section will be uploaded to the MPU.
4754 @item @samp{.infomem}
4755 Defines an information memory section (if applicable). Any code in
4756 this section will be uploaded to the MPU.
4758 @item @samp{.infomemnobits}
4759 This is the same as the @samp{.infomem} section except that any code
4760 in this section will not be uploaded to the MPU.
4762 @item @samp{.noinit}
4763 Denotes a portion of RAM located above @samp{.bss} section.
4765 The last two sections are used by gcc.
4779 @section @command{ld}'s Support for Various TI COFF Versions
4780 @cindex TI COFF versions
4781 @kindex --format=@var{version}
4782 The @samp{--format} switch allows selection of one of the various
4783 TI COFF versions. The latest of this writing is 2; versions 0 and 1 are
4784 also supported. The TI COFF versions also vary in header byte-order
4785 format; @command{ld} will read any version or byte order, but the output
4786 header format depends on the default specified by the specific target.
4799 @section @command{ld} and WIN32 (cygwin/mingw)
4801 This section describes some of the win32 specific @command{ld} issues.
4802 See @ref{Options,,Command Line Options} for detailed decription of the
4803 command line options mentioned here.
4806 @cindex import libraries
4807 @item import libraries
4808 The standard Windows linker creates and uses so-called import
4809 libraries, which contains information for linking to dll's. They are
4810 regular static archives and are handled as any other static
4811 archive. The cygwin and mingw ports of @command{ld} have specific
4812 support for creating such libraries provided with the
4813 @samp{--out-implib} command line option.
4815 @item exporting DLL symbols
4816 @cindex exporting DLL symbols
4817 The cygwin/mingw @command{ld} has several ways to export symbols for dll's.
4820 @item using auto-export functionality
4821 @cindex using auto-export functionality
4822 By default @command{ld} exports symbols with the auto-export functionality,
4823 which is controlled by the following command line options:
4826 @item --export-all-symbols [This is the default]
4827 @item --exclude-symbols
4828 @item --exclude-libs
4831 If, however, @samp{--export-all-symbols} is not given explicitly on the
4832 command line, then the default auto-export behavior will be @emph{disabled}
4833 if either of the following are true:
4836 @item A DEF file is used.
4837 @item Any symbol in any object file was marked with the __declspec(dllexport) attribute.
4840 @item using a DEF file
4841 @cindex using a DEF file
4842 Another way of exporting symbols is using a DEF file. A DEF file is
4843 an ASCII file containing definitions of symbols which should be
4844 exported when a dll is created. Usually it is named @samp{<dll
4845 name>.def} and is added as any other object file to the linker's
4846 command line. The file's name must end in @samp{.def} or @samp{.DEF}.
4849 gcc -o <output> <objectfiles> <dll name>.def
4852 Using a DEF file turns off the normal auto-export behavior, unless the
4853 @samp{--export-all-symbols} option is also used.
4855 Here is an example of a DEF file for a shared library called @samp{xyz.dll}:
4858 LIBRARY "xyz.dll" BASE=0x10000000
4866 This example defines a base address and three symbols. The third
4867 symbol is an alias for the second. For the complete format
4868 specification see ld/deffilep.y in the binutils sources.
4870 @cindex creating a DEF file
4871 While linking a shared dll, @command{ld} is able to create a DEF file
4872 with the @samp{--output-def <file>} command line option.
4874 @item Using decorations
4875 @cindex Using decorations
4876 Another way of marking symbols for export is to modify the source code
4877 itself, so that when building the DLL each symbol to be exported is
4881 __declspec(dllexport) int a_variable
4882 __declspec(dllexport) void a_function(int with_args)
4885 All such symbols will be exported from the DLL. If, however,
4886 any of the object files in the DLL contain symbols decorated in
4887 this way, then the normal auto-export behavior is disabled, unless
4888 the @samp{--export-all-symbols} option is also used.
4890 Note that object files that wish to access these symbols must @emph{not}
4891 decorate them with dllexport. Instead, they should use dllimport,
4895 __declspec(dllimport) int a_variable
4896 __declspec(dllimport) void a_function(int with_args)
4899 This complicates the structure of library header files, because
4900 when included by the library itself the header must declare the
4901 variables and functions as dllexport, but when included by client
4902 code the header must declare them as dllimport. There are a number
4903 of idioms that are typically used to do this; often client code can
4904 omit the __declspec() declaration completely. See
4905 @samp{--enable-auto-import} and @samp{automatic data imports} for more
4909 @cindex automatic data imports
4910 @item automatic data imports
4911 The standard Windows dll format supports data imports from dlls only
4912 by adding special decorations (dllimport/dllexport), which let the
4913 compiler produce specific assembler instructions to deal with this
4914 issue. This increases the effort necessary to port existing Un*x
4915 code to these platforms, especially for large
4916 c++ libraries and applications. The auto-import feature, which was
4917 initially provided by Paul Sokolovsky, allows one to omit the
4918 decorations to archieve a behavior that conforms to that on POSIX/Un*x
4919 platforms. This feature is enabled with the @samp{--enable-auto-import}
4920 command-line option, although it is enabled by default on cygwin/mingw.
4921 The @samp{--enable-auto-import} option itself now serves mainly to
4922 suppress any warnings that are ordinarily emitted when linked objects
4923 trigger the feature's use.
4925 auto-import of variables does not always work flawlessly without
4926 additional assistance. Sometimes, you will see this message
4928 "variable '<var>' can't be auto-imported. Please read the
4929 documentation for ld's @code{--enable-auto-import} for details."
4931 The @samp{--enable-auto-import} documentation explains why this error
4932 occurs, and several methods that can be used to overcome this difficulty.
4933 One of these methods is the @emph{runtime pseudo-relocs} feature, described
4936 @cindex runtime pseudo-relocation
4937 For complex variables imported from DLLs (such as structs or classes),
4938 object files typically contain a base address for the variable and an
4939 offset (@emph{addend}) within the variable--to specify a particular
4940 field or public member, for instance. Unfortunately, the runtime loader used
4941 in win32 environments is incapable of fixing these references at runtime
4942 without the additional information supplied by dllimport/dllexport decorations.
4943 The standard auto-import feature described above is unable to resolve these
4946 The @samp{--enable-runtime-pseudo-relocs} switch allows these references to
4947 be resolved without error, while leaving the task of adjusting the references
4948 themselves (with their non-zero addends) to specialized code provided by the
4949 runtime environment. Recent versions of the cygwin and mingw environments and
4950 compilers provide this runtime support; older versions do not. However, the
4951 support is only necessary on the developer's platform; the compiled result will
4952 run without error on an older system.
4954 @samp{--enable-runtime-pseudo-relocs} is not the default; it must be explicitly
4957 @cindex direct linking to a dll
4958 @item direct linking to a dll
4959 The cygwin/mingw ports of @command{ld} support the direct linking,
4960 including data symbols, to a dll without the usage of any import
4961 libraries. This is much faster and uses much less memory than does the
4962 traditional import library method, expecially when linking large
4963 libraries or applications. When @command{ld} creates an import lib, each
4964 function or variable exported from the dll is stored in its own bfd, even
4965 though a single bfd could contain many exports. The overhead involved in
4966 storing, loading, and processing so many bfd's is quite large, and explains the
4967 tremendous time, memory, and storage needed to link against particularly
4968 large or complex libraries when using import libs.
4970 Linking directly to a dll uses no extra command-line switches other than
4971 @samp{-L} and @samp{-l}, because @command{ld} already searches for a number
4972 of names to match each library. All that is needed from the developer's
4973 perspective is an understanding of this search, in order to force ld to
4974 select the dll instead of an import library.
4977 For instance, when ld is called with the argument @samp{-lxxx} it will attempt
4978 to find, in the first directory of its search path,
4989 before moving on to the next directory in the search path.
4991 (*) Actually, this is not @samp{cygxxx.dll} but in fact is @samp{<prefix>xxx.dll},
4992 where @samp{<prefix>} is set by the @command{ld} option
4993 @samp{--dll-search-prefix=<prefix>}. In the case of cygwin, the standard gcc spec
4994 file includes @samp{--dll-search-prefix=cyg}, so in effect we actually search for
4997 Other win32-based unix environments, such as mingw or pw32, may use other
4998 @samp{<prefix>}es, although at present only cygwin makes use of this feature. It
4999 was originally intended to help avoid name conflicts among dll's built for the
5000 various win32/un*x environments, so that (for example) two versions of a zlib dll
5001 could coexist on the same machine.
5003 The generic cygwin/mingw path layout uses a @samp{bin} directory for
5004 applications and dll's and a @samp{lib} directory for the import
5005 libraries (using cygwin nomenclature):
5011 libxxx.dll.a (in case of dll's)
5012 libxxx.a (in case of static archive)
5015 Linking directly to a dll without using the import library can be
5018 1. Use the dll directly by adding the @samp{bin} path to the link line
5020 gcc -Wl,-verbose -o a.exe -L../bin/ -lxxx
5023 However, as the dll's often have version numbers appended to their names
5024 (@samp{cygncurses-5.dll}) this will often fail, unless one specifies
5025 @samp{-L../bin -lncurses-5} to include the version. Import libs are generally
5026 not versioned, and do not have this difficulty.
5028 2. Create a symbolic link from the dll to a file in the @samp{lib}
5029 directory according to the above mentioned search pattern. This
5030 should be used to avoid unwanted changes in the tools needed for
5034 ln -s bin/cygxxx.dll lib/[cyg|lib|]xxx.dll[.a]
5037 Then you can link without any make environment changes.
5040 gcc -Wl,-verbose -o a.exe -L../lib/ -lxxx
5043 This technique also avoids the version number problems, because the following is
5050 libxxx.dll.a -> ../bin/cygxxx-5.dll
5053 Linking directly to a dll without using an import lib will work
5054 even when auto-import features are exercised, and even when
5055 @samp{--enable-runtime-pseudo-relocs} is used.
5057 Given the improvements in speed and memory usage, one might justifiably
5058 wonder why import libraries are used at all. There are two reasons:
5060 1. Until recently, the link-directly-to-dll functionality did @emph{not}
5061 work with auto-imported data.
5063 2. Sometimes it is necessary to include pure static objects within the
5064 import library (which otherwise contains only bfd's for indirection
5065 symbols that point to the exports of a dll). Again, the import lib
5066 for the cygwin kernel makes use of this ability, and it is not
5067 possible to do this without an import lib.
5069 So, import libs are not going away. But the ability to replace
5070 true import libs with a simple symbolic link to (or a copy of)
5071 a dll, in most cases, is a useful addition to the suite of tools
5072 binutils makes available to the win32 developer. Given the
5073 massive improvements in memory requirements during linking, storage
5074 requirements, and linking speed, we expect that many developers
5075 will soon begin to use this feature whenever possible.
5077 @item symbol aliasing
5079 @item adding additional names
5080 Sometimes, it is useful to export symbols with additional names.
5081 A symbol @samp{foo} will be exported as @samp{foo}, but it can also be
5082 exported as @samp{_foo} by using special directives in the DEF file
5083 when creating the dll. This will affect also the optional created
5084 import library. Consider the following DEF file:
5087 LIBRARY "xyz.dll" BASE=0x61000000
5094 The line @samp{_foo = foo} maps the symbol @samp{foo} to @samp{_foo}.
5096 Another method for creating a symbol alias is to create it in the
5097 source code using the "weak" attribute:
5100 void foo () @{ /* Do something. */; @}
5101 void _foo () __attribute__ ((weak, alias ("foo")));
5104 See the gcc manual for more information about attributes and weak
5107 @item renaming symbols
5108 Sometimes it is useful to rename exports. For instance, the cygwin
5109 kernel does this regularly. A symbol @samp{_foo} can be exported as
5110 @samp{foo} but not as @samp{_foo} by using special directives in the
5111 DEF file. (This will also affect the import library, if it is
5112 created). In the following example:
5115 LIBRARY "xyz.dll" BASE=0x61000000
5121 The line @samp{_foo = foo} maps the exported symbol @samp{foo} to
5125 Note: using a DEF file disables the default auto-export behavior,
5126 unless the @samp{--export-all-symbols} command line option is used.
5127 If, however, you are trying to rename symbols, then you should list
5128 @emph{all} desired exports in the DEF file, including the symbols
5129 that are not being renamed, and do @emph{not} use the
5130 @samp{--export-all-symbols} option. If you list only the
5131 renamed symbols in the DEF file, and use @samp{--export-all-symbols}
5132 to handle the other symbols, then the both the new names @emph{and}
5133 the original names for the the renamed symbols will be exported.
5134 In effect, you'd be aliasing those symbols, not renaming them,
5135 which is probably not what you wanted.
5149 @section @code{ld} and Xtensa Processors
5151 @cindex Xtensa processors
5152 The default @command{ld} behavior for Xtensa processors is to interpret
5153 @code{SECTIONS} commands so that lists of explicitly named sections in a
5154 specification with a wildcard file will be interleaved when necessary to
5155 keep literal pools within the range of PC-relative load offsets. For
5156 example, with the command:
5168 @command{ld} may interleave some of the @code{.literal}
5169 and @code{.text} sections from different object files to ensure that the
5170 literal pools are within the range of PC-relative load offsets. A valid
5171 interleaving might place the @code{.literal} sections from an initial
5172 group of files followed by the @code{.text} sections of that group of
5173 files. Then, the @code{.literal} sections from the rest of the files
5174 and the @code{.text} sections from the rest of the files would follow.
5175 The non-interleaved order can still be specified as:
5181 *(.literal) *(.text)
5186 @cindex @code{--relax} on Xtensa
5187 @cindex relaxing on Xtensa
5189 The Xtensa version of @command{ld} enables the @option{--relax} option by
5190 default to attempt to reduce space in the output image by combining
5191 literals with identical values. It also provides the
5192 @option{--no-relax} option to disable this optimization. When enabled,
5193 the relaxation algorithm ensures that a literal will only be merged with
5194 another literal when the new merged literal location is within the
5195 offset range of all of its uses.
5197 The relaxation mechanism will also attempt to optimize
5198 assembler-generated ``longcall'' sequences of
5199 @code{L32R}/@code{CALLX@var{n}} when the target is known to fit into a
5200 @code{CALL@var{n}} instruction encoding. The current optimization
5201 converts the sequence into @code{NOP}/@code{CALL@var{n}} and removes the
5202 literal referenced by the @code{L32R} instruction.
5209 @ifclear SingleFormat
5214 @cindex object file management
5215 @cindex object formats available
5217 The linker accesses object and archive files using the BFD libraries.
5218 These libraries allow the linker to use the same routines to operate on
5219 object files whatever the object file format. A different object file
5220 format can be supported simply by creating a new BFD back end and adding
5221 it to the library. To conserve runtime memory, however, the linker and
5222 associated tools are usually configured to support only a subset of the
5223 object file formats available. You can use @code{objdump -i}
5224 (@pxref{objdump,,objdump,binutils.info,The GNU Binary Utilities}) to
5225 list all the formats available for your configuration.
5227 @cindex BFD requirements
5228 @cindex requirements for BFD
5229 As with most implementations, BFD is a compromise between
5230 several conflicting requirements. The major factor influencing
5231 BFD design was efficiency: any time used converting between
5232 formats is time which would not have been spent had BFD not
5233 been involved. This is partly offset by abstraction payback; since
5234 BFD simplifies applications and back ends, more time and care
5235 may be spent optimizing algorithms for a greater speed.
5237 One minor artifact of the BFD solution which you should bear in
5238 mind is the potential for information loss. There are two places where
5239 useful information can be lost using the BFD mechanism: during
5240 conversion and during output. @xref{BFD information loss}.
5243 * BFD outline:: How it works: an outline of BFD
5247 @section How It Works: An Outline of BFD
5248 @cindex opening object files
5249 @include bfdsumm.texi
5252 @node Reporting Bugs
5253 @chapter Reporting Bugs
5254 @cindex bugs in @command{ld}
5255 @cindex reporting bugs in @command{ld}
5257 Your bug reports play an essential role in making @command{ld} reliable.
5259 Reporting a bug may help you by bringing a solution to your problem, or
5260 it may not. But in any case the principal function of a bug report is
5261 to help the entire community by making the next version of @command{ld}
5262 work better. Bug reports are your contribution to the maintenance of
5265 In order for a bug report to serve its purpose, you must include the
5266 information that enables us to fix the bug.
5269 * Bug Criteria:: Have you found a bug?
5270 * Bug Reporting:: How to report bugs
5274 @section Have You Found a Bug?
5275 @cindex bug criteria
5277 If you are not sure whether you have found a bug, here are some guidelines:
5280 @cindex fatal signal
5281 @cindex linker crash
5282 @cindex crash of linker
5284 If the linker gets a fatal signal, for any input whatever, that is a
5285 @command{ld} bug. Reliable linkers never crash.
5287 @cindex error on valid input
5289 If @command{ld} produces an error message for valid input, that is a bug.
5291 @cindex invalid input
5293 If @command{ld} does not produce an error message for invalid input, that
5294 may be a bug. In the general case, the linker can not verify that
5295 object files are correct.
5298 If you are an experienced user of linkers, your suggestions for
5299 improvement of @command{ld} are welcome in any case.
5303 @section How to Report Bugs
5305 @cindex @command{ld} bugs, reporting
5307 A number of companies and individuals offer support for @sc{gnu}
5308 products. If you obtained @command{ld} from a support organization, we
5309 recommend you contact that organization first.
5311 You can find contact information for many support companies and
5312 individuals in the file @file{etc/SERVICE} in the @sc{gnu} Emacs
5315 Otherwise, send bug reports for @command{ld} to
5316 @samp{bug-binutils@@gnu.org}.
5318 The fundamental principle of reporting bugs usefully is this:
5319 @strong{report all the facts}. If you are not sure whether to state a
5320 fact or leave it out, state it!
5322 Often people omit facts because they think they know what causes the
5323 problem and assume that some details do not matter. Thus, you might
5324 assume that the name of a symbol you use in an example does not
5325 matter. Well, probably it does not, but one cannot be sure. Perhaps
5326 the bug is a stray memory reference which happens to fetch from the
5327 location where that name is stored in memory; perhaps, if the name
5328 were different, the contents of that location would fool the linker
5329 into doing the right thing despite the bug. Play it safe and give a
5330 specific, complete example. That is the easiest thing for you to do,
5331 and the most helpful.
5333 Keep in mind that the purpose of a bug report is to enable us to fix
5334 the bug if it is new to us. Therefore, always write your bug reports
5335 on the assumption that the bug has not been reported previously.
5337 Sometimes people give a few sketchy facts and ask, ``Does this ring a
5338 bell?'' This cannot help us fix a bug, so it is basically useless. We
5339 respond by asking for enough details to enable us to investigate.
5340 You might as well expedite matters by sending them to begin with.
5342 To enable us to fix the bug, you should include all these things:
5346 The version of @command{ld}. @command{ld} announces it if you start it with
5347 the @samp{--version} argument.
5349 Without this, we will not know whether there is any point in looking for
5350 the bug in the current version of @command{ld}.
5353 Any patches you may have applied to the @command{ld} source, including any
5354 patches made to the @code{BFD} library.
5357 The type of machine you are using, and the operating system name and
5361 What compiler (and its version) was used to compile @command{ld}---e.g.
5365 The command arguments you gave the linker to link your example and
5366 observe the bug. To guarantee you will not omit something important,
5367 list them all. A copy of the Makefile (or the output from make) is
5370 If we were to try to guess the arguments, we would probably guess wrong
5371 and then we might not encounter the bug.
5374 A complete input file, or set of input files, that will reproduce the
5375 bug. It is generally most helpful to send the actual object files
5376 provided that they are reasonably small. Say no more than 10K. For
5377 bigger files you can either make them available by FTP or HTTP or else
5378 state that you are willing to send the object file(s) to whomever
5379 requests them. (Note - your email will be going to a mailing list, so
5380 we do not want to clog it up with large attachments). But small
5381 attachments are best.
5383 If the source files were assembled using @code{gas} or compiled using
5384 @code{gcc}, then it may be OK to send the source files rather than the
5385 object files. In this case, be sure to say exactly what version of
5386 @code{gas} or @code{gcc} was used to produce the object files. Also say
5387 how @code{gas} or @code{gcc} were configured.
5390 A description of what behavior you observe that you believe is
5391 incorrect. For example, ``It gets a fatal signal.''
5393 Of course, if the bug is that @command{ld} gets a fatal signal, then we
5394 will certainly notice it. But if the bug is incorrect output, we might
5395 not notice unless it is glaringly wrong. You might as well not give us
5396 a chance to make a mistake.
5398 Even if the problem you experience is a fatal signal, you should still
5399 say so explicitly. Suppose something strange is going on, such as, your
5400 copy of @command{ld} is out of synch, or you have encountered a bug in the
5401 C library on your system. (This has happened!) Your copy might crash
5402 and ours would not. If you told us to expect a crash, then when ours
5403 fails to crash, we would know that the bug was not happening for us. If
5404 you had not told us to expect a crash, then we would not be able to draw
5405 any conclusion from our observations.
5408 If you wish to suggest changes to the @command{ld} source, send us context
5409 diffs, as generated by @code{diff} with the @samp{-u}, @samp{-c}, or
5410 @samp{-p} option. Always send diffs from the old file to the new file.
5411 If you even discuss something in the @command{ld} source, refer to it by
5412 context, not by line number.
5414 The line numbers in our development sources will not match those in your
5415 sources. Your line numbers would convey no useful information to us.
5418 Here are some things that are not necessary:
5422 A description of the envelope of the bug.
5424 Often people who encounter a bug spend a lot of time investigating
5425 which changes to the input file will make the bug go away and which
5426 changes will not affect it.
5428 This is often time consuming and not very useful, because the way we
5429 will find the bug is by running a single example under the debugger
5430 with breakpoints, not by pure deduction from a series of examples.
5431 We recommend that you save your time for something else.
5433 Of course, if you can find a simpler example to report @emph{instead}
5434 of the original one, that is a convenience for us. Errors in the
5435 output will be easier to spot, running under the debugger will take
5436 less time, and so on.
5438 However, simplification is not vital; if you do not want to do this,
5439 report the bug anyway and send us the entire test case you used.
5442 A patch for the bug.
5444 A patch for the bug does help us if it is a good one. But do not omit
5445 the necessary information, such as the test case, on the assumption that
5446 a patch is all we need. We might see problems with your patch and decide
5447 to fix the problem another way, or we might not understand it at all.
5449 Sometimes with a program as complicated as @command{ld} it is very hard to
5450 construct an example that will make the program follow a certain path
5451 through the code. If you do not send us the example, we will not be
5452 able to construct one, so we will not be able to verify that the bug is
5455 And if we cannot understand what bug you are trying to fix, or why your
5456 patch should be an improvement, we will not install it. A test case will
5457 help us to understand.
5460 A guess about what the bug is or what it depends on.
5462 Such guesses are usually wrong. Even we cannot guess right about such
5463 things without first using the debugger to find the facts.
5467 @appendix MRI Compatible Script Files
5468 @cindex MRI compatibility
5469 To aid users making the transition to @sc{gnu} @command{ld} from the MRI
5470 linker, @command{ld} can use MRI compatible linker scripts as an
5471 alternative to the more general-purpose linker scripting language
5472 described in @ref{Scripts}. MRI compatible linker scripts have a much
5473 simpler command set than the scripting language otherwise used with
5474 @command{ld}. @sc{gnu} @command{ld} supports the most commonly used MRI
5475 linker commands; these commands are described here.
5477 In general, MRI scripts aren't of much use with the @code{a.out} object
5478 file format, since it only has three sections and MRI scripts lack some
5479 features to make use of them.
5481 You can specify a file containing an MRI-compatible script using the
5482 @samp{-c} command-line option.
5484 Each command in an MRI-compatible script occupies its own line; each
5485 command line starts with the keyword that identifies the command (though
5486 blank lines are also allowed for punctuation). If a line of an
5487 MRI-compatible script begins with an unrecognized keyword, @command{ld}
5488 issues a warning message, but continues processing the script.
5490 Lines beginning with @samp{*} are comments.
5492 You can write these commands using all upper-case letters, or all
5493 lower case; for example, @samp{chip} is the same as @samp{CHIP}.
5494 The following list shows only the upper-case form of each command.
5497 @cindex @code{ABSOLUTE} (MRI)
5498 @item ABSOLUTE @var{secname}
5499 @itemx ABSOLUTE @var{secname}, @var{secname}, @dots{} @var{secname}
5500 Normally, @command{ld} includes in the output file all sections from all
5501 the input files. However, in an MRI-compatible script, you can use the
5502 @code{ABSOLUTE} command to restrict the sections that will be present in
5503 your output program. If the @code{ABSOLUTE} command is used at all in a
5504 script, then only the sections named explicitly in @code{ABSOLUTE}
5505 commands will appear in the linker output. You can still use other
5506 input sections (whatever you select on the command line, or using
5507 @code{LOAD}) to resolve addresses in the output file.
5509 @cindex @code{ALIAS} (MRI)
5510 @item ALIAS @var{out-secname}, @var{in-secname}
5511 Use this command to place the data from input section @var{in-secname}
5512 in a section called @var{out-secname} in the linker output file.
5514 @var{in-secname} may be an integer.
5516 @cindex @code{ALIGN} (MRI)
5517 @item ALIGN @var{secname} = @var{expression}
5518 Align the section called @var{secname} to @var{expression}. The
5519 @var{expression} should be a power of two.
5521 @cindex @code{BASE} (MRI)
5522 @item BASE @var{expression}
5523 Use the value of @var{expression} as the lowest address (other than
5524 absolute addresses) in the output file.
5526 @cindex @code{CHIP} (MRI)
5527 @item CHIP @var{expression}
5528 @itemx CHIP @var{expression}, @var{expression}
5529 This command does nothing; it is accepted only for compatibility.
5531 @cindex @code{END} (MRI)
5533 This command does nothing whatever; it's only accepted for compatibility.
5535 @cindex @code{FORMAT} (MRI)
5536 @item FORMAT @var{output-format}
5537 Similar to the @code{OUTPUT_FORMAT} command in the more general linker
5538 language, but restricted to one of these output formats:
5542 S-records, if @var{output-format} is @samp{S}
5545 IEEE, if @var{output-format} is @samp{IEEE}
5548 COFF (the @samp{coff-m68k} variant in BFD), if @var{output-format} is
5552 @cindex @code{LIST} (MRI)
5553 @item LIST @var{anything}@dots{}
5554 Print (to the standard output file) a link map, as produced by the
5555 @command{ld} command-line option @samp{-M}.
5557 The keyword @code{LIST} may be followed by anything on the
5558 same line, with no change in its effect.
5560 @cindex @code{LOAD} (MRI)
5561 @item LOAD @var{filename}
5562 @itemx LOAD @var{filename}, @var{filename}, @dots{} @var{filename}
5563 Include one or more object file @var{filename} in the link; this has the
5564 same effect as specifying @var{filename} directly on the @command{ld}
5567 @cindex @code{NAME} (MRI)
5568 @item NAME @var{output-name}
5569 @var{output-name} is the name for the program produced by @command{ld}; the
5570 MRI-compatible command @code{NAME} is equivalent to the command-line
5571 option @samp{-o} or the general script language command @code{OUTPUT}.
5573 @cindex @code{ORDER} (MRI)
5574 @item ORDER @var{secname}, @var{secname}, @dots{} @var{secname}
5575 @itemx ORDER @var{secname} @var{secname} @var{secname}
5576 Normally, @command{ld} orders the sections in its output file in the
5577 order in which they first appear in the input files. In an MRI-compatible
5578 script, you can override this ordering with the @code{ORDER} command. The
5579 sections you list with @code{ORDER} will appear first in your output
5580 file, in the order specified.
5582 @cindex @code{PUBLIC} (MRI)
5583 @item PUBLIC @var{name}=@var{expression}
5584 @itemx PUBLIC @var{name},@var{expression}
5585 @itemx PUBLIC @var{name} @var{expression}
5586 Supply a value (@var{expression}) for external symbol
5587 @var{name} used in the linker input files.
5589 @cindex @code{SECT} (MRI)
5590 @item SECT @var{secname}, @var{expression}
5591 @itemx SECT @var{secname}=@var{expression}
5592 @itemx SECT @var{secname} @var{expression}
5593 You can use any of these three forms of the @code{SECT} command to
5594 specify the start address (@var{expression}) for section @var{secname}.
5595 If you have more than one @code{SECT} statement for the same
5596 @var{secname}, only the @emph{first} sets the start address.
5607 % I think something like @colophon should be in texinfo. In the
5609 \long\def\colophon{\hbox to0pt{}\vfill
5610 \centerline{The body of this manual is set in}
5611 \centerline{\fontname\tenrm,}
5612 \centerline{with headings in {\bf\fontname\tenbf}}
5613 \centerline{and examples in {\tt\fontname\tentt}.}
5614 \centerline{{\it\fontname\tenit\/} and}
5615 \centerline{{\sl\fontname\tensl\/}}
5616 \centerline{are used for emphasis.}\vfill}
5618 % Blame: doc@cygnus.com, 28mar91.