3 @c Copyright 1991, 1992, 1993, 1994, 1995, 1996, 1997, 1998, 1999, 2000,
4 @c 2001 Free Software Foundation, Inc.
6 @include configdoc.texi
7 @c (configdoc.texi is generated by the Makefile)
14 @c Configure for the generation of man pages
47 * Ld: (ld). The GNU linker.
53 This file documents the @sc{gnu} linker LD version @value{VERSION}.
55 Copyright (C) 1991, 92, 93, 94, 95, 96, 97, 98, 99, 2000,
56 2001 Free Software Foundation, Inc.
60 Permission is granted to copy, distribute and/or modify this document
61 under the terms of the GNU Free Documentation License, Version 1.1
62 or any later version published by the Free Software Foundation;
63 with no Invariant Sections, with no Front-Cover Texts, and with no
64 Back-Cover Texts. A copy of the license is included in the
65 section entitled "GNU Free Documentation License".
67 Permission is granted to process this file through Tex and print the
68 results, provided the printed document carries copying permission
69 notice identical to this one except for the removal of this paragraph
70 (this paragraph not being relevant to the printed manual).
76 @setchapternewpage odd
77 @settitle Using LD, the GNU linker
80 @subtitle The GNU linker
82 @subtitle @code{ld} version 2
83 @subtitle Version @value{VERSION}
84 @author Steve Chamberlain
85 @author Ian Lance Taylor
90 \hfill Red Hat Inc\par
91 \hfill nickc\@credhat.com, doc\@redhat.com\par
92 \hfill {\it Using LD, the GNU linker}\par
93 \hfill Edited by Jeffrey Osier (jeffrey\@cygnus.com)\par
95 \global\parindent=0pt % Steve likes it this way.
98 @vskip 0pt plus 1filll
99 @c man begin COPYRIGHT
100 Copyright @copyright{} 1991, 92, 93, 94, 95, 96, 97, 98, 99, 2000 Free Software Foundation, Inc.
102 Permission is granted to copy, distribute and/or modify this document
103 under the terms of the GNU Free Documentation License, Version 1.1
104 or any later version published by the Free Software Foundation;
105 with no Invariant Sections, with no Front-Cover Texts, and with no
106 Back-Cover Texts. A copy of the license is included in the
107 section entitled "GNU Free Documentation License".
112 @c FIXME: Talk about importance of *order* of args, cmds to linker!
117 This file documents the @sc{gnu} linker ld version @value{VERSION}.
119 This document is distributed under the terms of the GNU Free
120 Documentation License. A copy of the license is included in the
121 section entitled "GNU Free Documentation License".
124 * Overview:: Overview
125 * Invocation:: Invocation
126 * Scripts:: Linker Scripts
128 * Machine Dependent:: Machine Dependent Features
132 * H8/300:: ld and the H8/300
135 * Hitachi:: ld and other Hitachi micros
138 * i960:: ld and the Intel 960 family
141 * TI COFF:: ld and the TI COFF
144 @ifclear SingleFormat
147 @c Following blank line required for remaining bug in makeinfo conds/menus
149 * Reporting Bugs:: Reporting Bugs
150 * MRI:: MRI Compatible Script Files
151 * GNU Free Documentation License:: GNU Free Documentation License
159 @cindex @sc{gnu} linker
160 @cindex what is this?
163 @c man begin SYNOPSIS
164 ld [ options ] objfile...
168 ar(1), nm(1), objcopy(1), objdump(1), readelf(1) and
169 the Info entries for @file{binutils} and
174 @c man begin DESCRIPTION
176 @code{ld} combines a number of object and archive files, relocates
177 their data and ties up symbol references. Usually the last step in
178 compiling a program is to run @code{ld}.
180 @code{ld} accepts Linker Command Language files written in
181 a superset of AT&T's Link Editor Command Language syntax,
182 to provide explicit and total control over the linking process.
186 This man page does not describe the command language; see the
187 @code{ld} entry in @code{info}, or the manual
188 ld: the GNU linker, for full details on the command language and
189 on other aspects of the GNU linker.
192 @ifclear SingleFormat
193 This version of @code{ld} uses the general purpose BFD libraries
194 to operate on object files. This allows @code{ld} to read, combine, and
195 write object files in many different formats---for example, COFF or
196 @code{a.out}. Different formats may be linked together to produce any
197 available kind of object file. @xref{BFD}, for more information.
200 Aside from its flexibility, the @sc{gnu} linker is more helpful than other
201 linkers in providing diagnostic information. Many linkers abandon
202 execution immediately upon encountering an error; whenever possible,
203 @code{ld} continues executing, allowing you to identify other errors
204 (or, in some cases, to get an output file in spite of the error).
211 @c man begin DESCRIPTION
213 The @sc{gnu} linker @code{ld} is meant to cover a broad range of situations,
214 and to be as compatible as possible with other linkers. As a result,
215 you have many choices to control its behavior.
221 * Options:: Command Line Options
222 * Environment:: Environment Variables
226 @section Command Line Options
234 The linker supports a plethora of command-line options, but in actual
235 practice few of them are used in any particular context.
236 @cindex standard Unix system
237 For instance, a frequent use of @code{ld} is to link standard Unix
238 object files on a standard, supported Unix system. On such a system, to
239 link a file @code{hello.o}:
242 ld -o @var{output} /lib/crt0.o hello.o -lc
245 This tells @code{ld} to produce a file called @var{output} as the
246 result of linking the file @code{/lib/crt0.o} with @code{hello.o} and
247 the library @code{libc.a}, which will come from the standard search
248 directories. (See the discussion of the @samp{-l} option below.)
250 Some of the command-line options to @code{ld} may be specified at any
251 point in the command line. However, options which refer to files, such
252 as @samp{-l} or @samp{-T}, cause the file to be read at the point at
253 which the option appears in the command line, relative to the object
254 files and other file options. Repeating non-file options with a
255 different argument will either have no further effect, or override prior
256 occurrences (those further to the left on the command line) of that
257 option. Options which may be meaningfully specified more than once are
258 noted in the descriptions below.
261 Non-option arguments are object files or archives which are to be linked
262 together. They may follow, precede, or be mixed in with command-line
263 options, except that an object file argument may not be placed between
264 an option and its argument.
266 Usually the linker is invoked with at least one object file, but you can
267 specify other forms of binary input files using @samp{-l}, @samp{-R},
268 and the script command language. If @emph{no} binary input files at all
269 are specified, the linker does not produce any output, and issues the
270 message @samp{No input files}.
272 If the linker can not recognize the format of an object file, it will
273 assume that it is a linker script. A script specified in this way
274 augments the main linker script used for the link (either the default
275 linker script or the one specified by using @samp{-T}). This feature
276 permits the linker to link against a file which appears to be an object
277 or an archive, but actually merely defines some symbol values, or uses
278 @code{INPUT} or @code{GROUP} to load other objects. Note that
279 specifying a script in this way should only be used to augment the main
280 linker script; if you want to use some command that logically can only
281 appear once, such as the @code{SECTIONS} or @code{MEMORY} command, you
282 must replace the default linker script using the @samp{-T} option.
285 For options whose names are a single letter,
286 option arguments must either follow the option letter without intervening
287 whitespace, or be given as separate arguments immediately following the
288 option that requires them.
290 For options whose names are multiple letters, either one dash or two can
291 precede the option name; for example, @samp{-trace-symbol} and
292 @samp{--trace-symbol} are equivalent. Note - there is one exception to
293 this rule. Multiple letter options that start with a lower case 'o' can
294 only be preceeded by two dashes. This is to reduce confusion with the
295 @samp{-o} option. So for example @samp{-omagic} sets the output file
296 name to @samp{magic} whereas @samp{--omagic} sets the NMAGIC flag on the
299 Arguments to multiple-letter options must either be separated from the
300 option name by an equals sign, or be given as separate arguments
301 immediately following the option that requires them. For example,
302 @samp{--trace-symbol foo} and @samp{--trace-symbol=foo} are equivalent.
303 Unique abbreviations of the names of multiple-letter options are
306 Note - if the linker is being invoked indirectly, via a compiler driver
307 (eg @samp{gcc}) then all the linker command line options should be
308 prefixed by @samp{-Wl,} (or whatever is appropriate for the particular
309 compiler driver) like this:
312 gcc -Wl,--startgroup foo.o bar.o -Wl,--endgroup
315 This is important, because otherwise the compiler driver program may
316 silently drop the linker options, resulting in a bad link.
318 Here is a table of the generic command line switches accepted by the GNU
322 @kindex -a@var{keyword}
323 @item -a@var{keyword}
324 This option is supported for HP/UX compatibility. The @var{keyword}
325 argument must be one of the strings @samp{archive}, @samp{shared}, or
326 @samp{default}. @samp{-aarchive} is functionally equivalent to
327 @samp{-Bstatic}, and the other two keywords are functionally equivalent
328 to @samp{-Bdynamic}. This option may be used any number of times.
331 @cindex architectures
333 @item -A@var{architecture}
334 @kindex --architecture=@var{arch}
335 @itemx --architecture=@var{architecture}
336 In the current release of @code{ld}, this option is useful only for the
337 Intel 960 family of architectures. In that @code{ld} configuration, the
338 @var{architecture} argument identifies the particular architecture in
339 the 960 family, enabling some safeguards and modifying the
340 archive-library search path. @xref{i960,,@code{ld} and the Intel 960
341 family}, for details.
343 Future releases of @code{ld} may support similar functionality for
344 other architecture families.
347 @ifclear SingleFormat
348 @cindex binary input format
349 @kindex -b @var{format}
350 @kindex --format=@var{format}
353 @item -b @var{input-format}
354 @itemx --format=@var{input-format}
355 @code{ld} may be configured to support more than one kind of object
356 file. If your @code{ld} is configured this way, you can use the
357 @samp{-b} option to specify the binary format for input object files
358 that follow this option on the command line. Even when @code{ld} is
359 configured to support alternative object formats, you don't usually need
360 to specify this, as @code{ld} should be configured to expect as a
361 default input format the most usual format on each machine.
362 @var{input-format} is a text string, the name of a particular format
363 supported by the BFD libraries. (You can list the available binary
364 formats with @samp{objdump -i}.)
367 You may want to use this option if you are linking files with an unusual
368 binary format. You can also use @samp{-b} to switch formats explicitly (when
369 linking object files of different formats), by including
370 @samp{-b @var{input-format}} before each group of object files in a
373 The default format is taken from the environment variable
378 You can also define the input format from a script, using the command
381 see @ref{Format Commands}.
385 @kindex -c @var{MRI-cmdfile}
386 @kindex --mri-script=@var{MRI-cmdfile}
387 @cindex compatibility, MRI
388 @item -c @var{MRI-commandfile}
389 @itemx --mri-script=@var{MRI-commandfile}
390 For compatibility with linkers produced by MRI, @code{ld} accepts script
391 files written in an alternate, restricted command language, described in
393 @ref{MRI,,MRI Compatible Script Files}.
396 the MRI Compatible Script Files section of GNU ld documentation.
398 Introduce MRI script files with
399 the option @samp{-c}; use the @samp{-T} option to run linker
400 scripts written in the general-purpose @code{ld} scripting language.
401 If @var{MRI-cmdfile} does not exist, @code{ld} looks for it in the directories
402 specified by any @samp{-L} options.
404 @cindex common allocation
411 These three options are equivalent; multiple forms are supported for
412 compatibility with other linkers. They assign space to common symbols
413 even if a relocatable output file is specified (with @samp{-r}). The
414 script command @code{FORCE_COMMON_ALLOCATION} has the same effect.
415 @xref{Miscellaneous Commands}.
417 @cindex entry point, from command line
418 @kindex -e @var{entry}
419 @kindex --entry=@var{entry}
421 @itemx --entry=@var{entry}
422 Use @var{entry} as the explicit symbol for beginning execution of your
423 program, rather than the default entry point. If there is no symbol
424 named @var{entry}, the linker will try to parse @var{entry} as a number,
425 and use that as the entry address (the number will be interpreted in
426 base 10; you may use a leading @samp{0x} for base 16, or a leading
427 @samp{0} for base 8). @xref{Entry Point}, for a discussion of defaults
428 and other ways of specifying the entry point.
430 @cindex dynamic symbol table
432 @kindex --export-dynamic
434 @itemx --export-dynamic
435 When creating a dynamically linked executable, add all symbols to the
436 dynamic symbol table. The dynamic symbol table is the set of symbols
437 which are visible from dynamic objects at run time.
439 If you do not use this option, the dynamic symbol table will normally
440 contain only those symbols which are referenced by some dynamic object
441 mentioned in the link.
443 If you use @code{dlopen} to load a dynamic object which needs to refer
444 back to the symbols defined by the program, rather than some other
445 dynamic object, then you will probably need to use this option when
446 linking the program itself.
448 You can also use the version script to control what symbols should
449 be added to the dynamic symbol table if the output format supports it.
450 See the description of @samp{--version-script} in @ref{VERSION}.
452 @cindex big-endian objects
456 Link big-endian objects. This affects the default output format.
458 @cindex little-endian objects
461 Link little-endian objects. This affects the default output format.
466 @itemx --auxiliary @var{name}
467 When creating an ELF shared object, set the internal DT_AUXILIARY field
468 to the specified name. This tells the dynamic linker that the symbol
469 table of the shared object should be used as an auxiliary filter on the
470 symbol table of the shared object @var{name}.
472 If you later link a program against this filter object, then, when you
473 run the program, the dynamic linker will see the DT_AUXILIARY field. If
474 the dynamic linker resolves any symbols from the filter object, it will
475 first check whether there is a definition in the shared object
476 @var{name}. If there is one, it will be used instead of the definition
477 in the filter object. The shared object @var{name} need not exist.
478 Thus the shared object @var{name} may be used to provide an alternative
479 implementation of certain functions, perhaps for debugging or for
480 machine specific performance.
482 This option may be specified more than once. The DT_AUXILIARY entries
483 will be created in the order in which they appear on the command line.
488 @itemx --filter @var{name}
489 When creating an ELF shared object, set the internal DT_FILTER field to
490 the specified name. This tells the dynamic linker that the symbol table
491 of the shared object which is being created should be used as a filter
492 on the symbol table of the shared object @var{name}.
494 If you later link a program against this filter object, then, when you
495 run the program, the dynamic linker will see the DT_FILTER field. The
496 dynamic linker will resolve symbols according to the symbol table of the
497 filter object as usual, but it will actually link to the definitions
498 found in the shared object @var{name}. Thus the filter object can be
499 used to select a subset of the symbols provided by the object
502 Some older linkers used the @code{-F} option throughout a compilation
503 toolchain for specifying object-file format for both input and output
504 object files. The @sc{gnu} linker uses other mechanisms for this
505 purpose: the @code{-b}, @code{--format}, @code{--oformat} options, the
506 @code{TARGET} command in linker scripts, and the @code{GNUTARGET}
507 environment variable. The @sc{gnu} linker will ignore the @code{-F}
508 option when not creating an ELF shared object.
510 @cindex finalization function
512 @item -fini @var{name}
513 When creating an ELF executable or shared object, call NAME when the
514 executable or shared object is unloaded, by setting DT_FINI to the
515 address of the function. By default, the linker uses @code{_fini} as
516 the function to call.
520 Ignored. Provided for compatibility with other tools.
526 @itemx --gpsize=@var{value}
527 Set the maximum size of objects to be optimized using the GP register to
528 @var{size}. This is only meaningful for object file formats such as
529 MIPS ECOFF which supports putting large and small objects into different
530 sections. This is ignored for other object file formats.
532 @cindex runtime library name
534 @kindex -soname=@var{name}
536 @itemx -soname=@var{name}
537 When creating an ELF shared object, set the internal DT_SONAME field to
538 the specified name. When an executable is linked with a shared object
539 which has a DT_SONAME field, then when the executable is run the dynamic
540 linker will attempt to load the shared object specified by the DT_SONAME
541 field rather than the using the file name given to the linker.
544 @cindex incremental link
546 Perform an incremental link (same as option @samp{-r}).
548 @cindex initialization function
550 @item -init @var{name}
551 When creating an ELF executable or shared object, call NAME when the
552 executable or shared object is loaded, by setting DT_INIT to the address
553 of the function. By default, the linker uses @code{_init} as the
556 @cindex archive files, from cmd line
557 @kindex -l@var{archive}
558 @kindex --library=@var{archive}
559 @item -l@var{archive}
560 @itemx --library=@var{archive}
561 Add archive file @var{archive} to the list of files to link. This
562 option may be used any number of times. @code{ld} will search its
563 path-list for occurrences of @code{lib@var{archive}.a} for every
564 @var{archive} specified.
566 On systems which support shared libraries, @code{ld} may also search for
567 libraries with extensions other than @code{.a}. Specifically, on ELF
568 and SunOS systems, @code{ld} will search a directory for a library with
569 an extension of @code{.so} before searching for one with an extension of
570 @code{.a}. By convention, a @code{.so} extension indicates a shared
573 The linker will search an archive only once, at the location where it is
574 specified on the command line. If the archive defines a symbol which
575 was undefined in some object which appeared before the archive on the
576 command line, the linker will include the appropriate file(s) from the
577 archive. However, an undefined symbol in an object appearing later on
578 the command line will not cause the linker to search the archive again.
580 See the @code{-(} option for a way to force the linker to search
581 archives multiple times.
583 You may list the same archive multiple times on the command line.
586 This type of archive searching is standard for Unix linkers. However,
587 if you are using @code{ld} on AIX, note that it is different from the
588 behaviour of the AIX linker.
591 @cindex search directory, from cmd line
593 @kindex --library-path=@var{dir}
594 @item -L@var{searchdir}
595 @itemx --library-path=@var{searchdir}
596 Add path @var{searchdir} to the list of paths that @code{ld} will search
597 for archive libraries and @code{ld} control scripts. You may use this
598 option any number of times. The directories are searched in the order
599 in which they are specified on the command line. Directories specified
600 on the command line are searched before the default directories. All
601 @code{-L} options apply to all @code{-l} options, regardless of the
602 order in which the options appear.
605 The default set of paths searched (without being specified with
606 @samp{-L}) depends on which emulation mode @code{ld} is using, and in
607 some cases also on how it was configured. @xref{Environment}.
610 The paths can also be specified in a link script with the
611 @code{SEARCH_DIR} command. Directories specified this way are searched
612 at the point in which the linker script appears in the command line.
615 @kindex -m @var{emulation}
616 @item -m@var{emulation}
617 Emulate the @var{emulation} linker. You can list the available
618 emulations with the @samp{--verbose} or @samp{-V} options.
620 If the @samp{-m} option is not used, the emulation is taken from the
621 @code{LDEMULATION} environment variable, if that is defined.
623 Otherwise, the default emulation depends upon how the linker was
631 Print a link map to the standard output. A link map provides
632 information about the link, including the following:
636 Where object files and symbols are mapped into memory.
638 How common symbols are allocated.
640 All archive members included in the link, with a mention of the symbol
641 which caused the archive member to be brought in.
645 @cindex read-only text
650 Turn off page alignment of sections, and mark the output as
651 @code{NMAGIC} if possible.
655 @cindex read/write from cmd line
659 Set the text and data sections to be readable and writable. Also, do
660 not page-align the data segment. If the output format supports Unix
661 style magic numbers, mark the output as @code{OMAGIC}.
663 @kindex -o @var{output}
664 @kindex --output=@var{output}
665 @cindex naming the output file
666 @item -o @var{output}
667 @itemx --output=@var{output}
668 Use @var{output} as the name for the program produced by @code{ld}; if this
669 option is not specified, the name @file{a.out} is used by default. The
670 script command @code{OUTPUT} can also specify the output file name.
672 @kindex -O @var{level}
673 @cindex generating optimized output
675 If @var{level} is a numeric values greater than zero @code{ld} optimizes
676 the output. This might take significantly longer and therefore probably
677 should only be enabled for the final binary.
680 @kindex --emit-relocs
681 @cindex retain relocations in final executable
684 Leave relocation sections and contents in fully linked exececutables.
685 Post link analysis and optimization tools may need this information in
686 order to perform correct modifications of executables. This results
687 in larger executables.
690 @cindex relocatable output
692 @kindex --relocateable
694 @itemx --relocateable
695 Generate relocatable output---i.e., generate an output file that can in
696 turn serve as input to @code{ld}. This is often called @dfn{partial
697 linking}. As a side effect, in environments that support standard Unix
698 magic numbers, this option also sets the output file's magic number to
701 If this option is not specified, an absolute file is produced. When
702 linking C++ programs, this option @emph{will not} resolve references to
703 constructors; to do that, use @samp{-Ur}.
705 When an input file does not have the same format as the output file,
706 partial linking is only supported if that input file does not contain any
707 relocations. Different output formats can have further restrictions; for
708 example some @code{a.out}-based formats do not support partial linking
709 with input files in other formats at all.
711 This option does the same thing as @samp{-i}.
713 @kindex -R @var{file}
714 @kindex --just-symbols=@var{file}
715 @cindex symbol-only input
716 @item -R @var{filename}
717 @itemx --just-symbols=@var{filename}
718 Read symbol names and their addresses from @var{filename}, but do not
719 relocate it or include it in the output. This allows your output file
720 to refer symbolically to absolute locations of memory defined in other
721 programs. You may use this option more than once.
723 For compatibility with other ELF linkers, if the @code{-R} option is
724 followed by a directory name, rather than a file name, it is treated as
725 the @code{-rpath} option.
729 @cindex strip all symbols
732 Omit all symbol information from the output file.
735 @kindex --strip-debug
736 @cindex strip debugger symbols
739 Omit debugger symbol information (but not all symbols) from the output file.
743 @cindex input files, displaying
746 Print the names of the input files as @code{ld} processes them.
748 @kindex -T @var{script}
749 @kindex --script=@var{script}
751 @item -T @var{scriptfile}
752 @itemx --script=@var{scriptfile}
753 Use @var{scriptfile} as the linker script. This script replaces
754 @code{ld}'s default linker script (rather than adding to it), so
755 @var{commandfile} must specify everything necessary to describe the
756 output file. You must use this option if you want to use a command
757 which can only appear once in a linker script, such as the
758 @code{SECTIONS} or @code{MEMORY} command. @xref{Scripts}. If
759 @var{scriptfile} does not exist in the current directory, @code{ld}
760 looks for it in the directories specified by any preceding @samp{-L}
761 options. Multiple @samp{-T} options accumulate.
763 @kindex -u @var{symbol}
764 @kindex --undefined=@var{symbol}
765 @cindex undefined symbol
766 @item -u @var{symbol}
767 @itemx --undefined=@var{symbol}
768 Force @var{symbol} to be entered in the output file as an undefined
769 symbol. Doing this may, for example, trigger linking of additional
770 modules from standard libraries. @samp{-u} may be repeated with
771 different option arguments to enter additional undefined symbols. This
772 option is equivalent to the @code{EXTERN} linker script command.
777 For anything other than C++ programs, this option is equivalent to
778 @samp{-r}: it generates relocatable output---i.e., an output file that can in
779 turn serve as input to @code{ld}. When linking C++ programs, @samp{-Ur}
780 @emph{does} resolve references to constructors, unlike @samp{-r}.
781 It does not work to use @samp{-Ur} on files that were themselves linked
782 with @samp{-Ur}; once the constructor table has been built, it cannot
783 be added to. Use @samp{-Ur} only for the last partial link, and
784 @samp{-r} for the others.
786 @kindex --unique[=@var{SECTION}]
787 @item --unique[=@var{SECTION}]
788 Creates a separate output section for every input section matching
789 @var{SECTION}, or if the optional wildcard @var{SECTION} argument is
790 missing, for every orphan input section. An orphan section is one not
791 specifically mentioned in a linker script. You may use this option
792 multiple times on the command line; It prevents the normal merging of
793 input sections with the same name, overriding output section assignments
803 Display the version number for @code{ld}. The @code{-V} option also
804 lists the supported emulations.
807 @kindex --discard-all
808 @cindex deleting local symbols
811 Delete all local symbols.
814 @kindex --discard-locals
815 @cindex local symbols, deleting
816 @cindex L, deleting symbols beginning
818 @itemx --discard-locals
819 Delete all temporary local symbols. For most targets, this is all local
820 symbols whose names begin with @samp{L}.
822 @kindex -y @var{symbol}
823 @kindex --trace-symbol=@var{symbol}
824 @cindex symbol tracing
825 @item -y @var{symbol}
826 @itemx --trace-symbol=@var{symbol}
827 Print the name of each linked file in which @var{symbol} appears. This
828 option may be given any number of times. On many systems it is necessary
829 to prepend an underscore.
831 This option is useful when you have an undefined symbol in your link but
832 don't know where the reference is coming from.
834 @kindex -Y @var{path}
836 Add @var{path} to the default library search path. This option exists
837 for Solaris compatibility.
839 @kindex -z @var{keyword}
840 @item -z @var{keyword}
841 The recognized keywords are @code{initfirst}, @code{interpose},
842 @code{loadfltr}, @code{nodefaultlib}, @code{nodelete}, @code{nodlopen},
843 @code{nodump}, @code{now} and @code{origin}. The other keywords are
844 ignored for Solaris compatibility. @code{initfirst} marks the object
845 to be initialized first at runtime before any other objects.
846 @code{interpose} marks the object that its symbol table interposes
847 before all symbols but the primary executable. @code{loadfltr} marks
848 the object that its filtees be processed immediately at runtime.
849 @code{nodefaultlib} marks the object that the search for dependencies
850 of this object will ignore any default library search paths.
851 @code{nodelete} marks the object shouldn't be unloaded at runtime.
852 @code{nodlopen} marks the object not available to @code{dlopen}.
853 @code{nodump} marks the object can not be dumped by @code{dldump}.
854 @code{now} marks the object with the non-lazy runtime binding.
855 @code{origin} marks the object may contain $ORIGIN.
856 @code{defs} disallows undefined symbols.
859 @cindex groups of archives
860 @item -( @var{archives} -)
861 @itemx --start-group @var{archives} --end-group
862 The @var{archives} should be a list of archive files. They may be
863 either explicit file names, or @samp{-l} options.
865 The specified archives are searched repeatedly until no new undefined
866 references are created. Normally, an archive is searched only once in
867 the order that it is specified on the command line. If a symbol in that
868 archive is needed to resolve an undefined symbol referred to by an
869 object in an archive that appears later on the command line, the linker
870 would not be able to resolve that reference. By grouping the archives,
871 they all be searched repeatedly until all possible references are
874 Using this option has a significant performance cost. It is best to use
875 it only when there are unavoidable circular references between two or
878 @kindex -assert @var{keyword}
879 @item -assert @var{keyword}
880 This option is ignored for SunOS compatibility.
888 Link against dynamic libraries. This is only meaningful on platforms
889 for which shared libraries are supported. This option is normally the
890 default on such platforms. The different variants of this option are
891 for compatibility with various systems. You may use this option
892 multiple times on the command line: it affects library searching for
893 @code{-l} options which follow it.
897 Set the @code{DF_1_GROUP} flag in the @code{DT_FLAGS_1} entry in the dynamic
898 section. This causes the runtime linker to handle lookups in this
899 object and its dependencies to be performed only inside the group.
900 @code{--no-undefined} is implied. This option is only meaningful on ELF
901 platforms which support shared libraries.
911 Do not link against shared libraries. This is only meaningful on
912 platforms for which shared libraries are supported. The different
913 variants of this option are for compatibility with various systems. You
914 may use this option multiple times on the command line: it affects
915 library searching for @code{-l} options which follow it.
919 When creating a shared library, bind references to global symbols to the
920 definition within the shared library, if any. Normally, it is possible
921 for a program linked against a shared library to override the definition
922 within the shared library. This option is only meaningful on ELF
923 platforms which support shared libraries.
925 @kindex --check-sections
926 @kindex --no-check-sections
927 @item --check-sections
928 @itemx --no-check-sections
929 Asks the linker @emph{not} to check section addresses after they have
930 been assigned to see if there any overlaps. Normally the linker will
931 perform this check, and if it finds any overlaps it will produce
932 suitable error messages. The linker does know about, and does make
933 allowances for sections in overlays. The default behaviour can be
934 restored by using the command line switch @samp{--check-sections}.
936 @cindex cross reference table
939 Output a cross reference table. If a linker map file is being
940 generated, the cross reference table is printed to the map file.
941 Otherwise, it is printed on the standard output.
943 The format of the table is intentionally simple, so that it may be
944 easily processed by a script if necessary. The symbols are printed out,
945 sorted by name. For each symbol, a list of file names is given. If the
946 symbol is defined, the first file listed is the location of the
947 definition. The remaining files contain references to the symbol.
949 @cindex symbols, from command line
950 @kindex --defsym @var{symbol}=@var{exp}
951 @item --defsym @var{symbol}=@var{expression}
952 Create a global symbol in the output file, containing the absolute
953 address given by @var{expression}. You may use this option as many
954 times as necessary to define multiple symbols in the command line. A
955 limited form of arithmetic is supported for the @var{expression} in this
956 context: you may give a hexadecimal constant or the name of an existing
957 symbol, or use @code{+} and @code{-} to add or subtract hexadecimal
958 constants or symbols. If you need more elaborate expressions, consider
959 using the linker command language from a script (@pxref{Assignments,,
960 Assignment: Symbol Definitions}). @emph{Note:} there should be no white
961 space between @var{symbol}, the equals sign (``@key{=}''), and
964 @cindex demangling, from command line
965 @kindex --demangle[=@var{style}]
966 @kindex --no-demangle
967 @item --demangle[=@var{style}]
969 These options control whether to demangle symbol names in error messages
970 and other output. When the linker is told to demangle, it tries to
971 present symbol names in a readable fashion: it strips leading
972 underscores if they are used by the object file format, and converts C++
973 mangled symbol names into user readable names. Different compilers have
974 different mangling styles. The optional demangling style argument can be used
975 to choose an appropriate demangling style for your compiler. The linker will
976 demangle by default unless the environment variable @samp{COLLECT_NO_DEMANGLE}
977 is set. These options may be used to override the default.
979 @cindex dynamic linker, from command line
981 @kindex --dynamic-linker @var{file}
982 @item --dynamic-linker @var{file}
983 Set the name of the dynamic linker. This is only meaningful when
984 generating dynamically linked ELF executables. The default dynamic
985 linker is normally correct; don't use this unless you know what you are
988 @cindex MIPS embedded PIC code
989 @kindex --embedded-relocs
990 @item --embedded-relocs
991 This option is only meaningful when linking MIPS embedded PIC code,
992 generated by the -membedded-pic option to the @sc{gnu} compiler and
993 assembler. It causes the linker to create a table which may be used at
994 runtime to relocate any data which was statically initialized to pointer
995 values. See the code in testsuite/ld-empic for details.
998 @kindex --fatal-warnings
999 @item --fatal-warnings
1000 Treat all warnings as errors.
1002 @kindex --force-exe-suffix
1003 @item --force-exe-suffix
1004 Make sure that an output file has a .exe suffix.
1006 If a successfully built fully linked output file does not have a
1007 @code{.exe} or @code{.dll} suffix, this option forces the linker to copy
1008 the output file to one of the same name with a @code{.exe} suffix. This
1009 option is useful when using unmodified Unix makefiles on a Microsoft
1010 Windows host, since some versions of Windows won't run an image unless
1011 it ends in a @code{.exe} suffix.
1013 @kindex --gc-sections
1014 @kindex --no-gc-sections
1015 @cindex garbage collection
1016 @item --no-gc-sections
1017 @itemx --gc-sections
1018 Enable garbage collection of unused input sections. It is ignored on
1019 targets that do not support this option. This option is not compatible
1020 with @samp{-r}, nor should it be used with dynamic linking. The default
1021 behaviour (of not performing this garbage collection) can be restored by
1022 specifying @samp{--no-gc-sections} on the command line.
1028 Print a summary of the command-line options on the standard output and exit.
1030 @kindex --target-help
1032 Print a summary of all target specific options on the standard output and exit.
1035 @item -Map @var{mapfile}
1036 Print a link map to the file @var{mapfile}. See the description of the
1037 @samp{-M} option, above.
1039 @cindex memory usage
1040 @kindex --no-keep-memory
1041 @item --no-keep-memory
1042 @code{ld} normally optimizes for speed over memory usage by caching the
1043 symbol tables of input files in memory. This option tells @code{ld} to
1044 instead optimize for memory usage, by rereading the symbol tables as
1045 necessary. This may be required if @code{ld} runs out of memory space
1046 while linking a large executable.
1048 @kindex --no-undefined
1050 @item --no-undefined
1052 Normally when creating a non-symbolic shared library, undefined symbols
1053 are allowed and left to be resolved by the runtime loader. These options
1054 disallows such undefined symbols.
1056 @kindex --allow-shlib-undefined
1057 @item --allow-shlib-undefined
1058 Allow undefined symbols in shared objects even when --no-undefined is
1059 set. The net result will be that undefined symbols in regular objects
1060 will still trigger an error, but undefined symbols in shared objects
1061 will be ignored. The implementation of no_undefined makes the
1062 assumption that the runtime linker will choke on undefined symbols.
1063 However there is at least one system (BeOS) where undefined symbols in
1064 shared libraries is normal since the kernel patches them at load time to
1065 select which function is most appropriate for the current architecture.
1066 I.E. dynamically select an appropriate memset function. Apparently it
1067 is also normal for HPPA shared libraries to have undefined symbols.
1069 @kindex --no-warn-mismatch
1070 @item --no-warn-mismatch
1071 Normally @code{ld} will give an error if you try to link together input
1072 files that are mismatched for some reason, perhaps because they have
1073 been compiled for different processors or for different endiannesses.
1074 This option tells @code{ld} that it should silently permit such possible
1075 errors. This option should only be used with care, in cases when you
1076 have taken some special action that ensures that the linker errors are
1079 @kindex --no-whole-archive
1080 @item --no-whole-archive
1081 Turn off the effect of the @code{--whole-archive} option for subsequent
1084 @cindex output file after errors
1085 @kindex --noinhibit-exec
1086 @item --noinhibit-exec
1087 Retain the executable output file whenever it is still usable.
1088 Normally, the linker will not produce an output file if it encounters
1089 errors during the link process; it exits without writing an output file
1090 when it issues any error whatsoever.
1092 @ifclear SingleFormat
1094 @item --oformat @var{output-format}
1095 @code{ld} may be configured to support more than one kind of object
1096 file. If your @code{ld} is configured this way, you can use the
1097 @samp{--oformat} option to specify the binary format for the output
1098 object file. Even when @code{ld} is configured to support alternative
1099 object formats, you don't usually need to specify this, as @code{ld}
1100 should be configured to produce as a default output format the most
1101 usual format on each machine. @var{output-format} is a text string, the
1102 name of a particular format supported by the BFD libraries. (You can
1103 list the available binary formats with @samp{objdump -i}.) The script
1104 command @code{OUTPUT_FORMAT} can also specify the output format, but
1105 this option overrides it. @xref{BFD}.
1110 This option is ignored for Linux compatibility.
1114 This option is ignored for SVR4 compatibility.
1117 @cindex synthesizing linker
1118 @cindex relaxing addressing modes
1120 An option with machine dependent effects.
1122 This option is only supported on a few targets.
1125 @xref{H8/300,,@code{ld} and the H8/300}.
1128 @xref{i960,, @code{ld} and the Intel 960 family}.
1132 On some platforms, the @samp{--relax} option performs global
1133 optimizations that become possible when the linker resolves addressing
1134 in the program, such as relaxing address modes and synthesizing new
1135 instructions in the output object file.
1137 On some platforms these link time global optimizations may make symbolic
1138 debugging of the resulting executable impossible.
1141 the case for the Matsushita MN10200 and MN10300 family of processors.
1145 On platforms where this is not supported, @samp{--relax} is accepted,
1149 @cindex retaining specified symbols
1150 @cindex stripping all but some symbols
1151 @cindex symbols, retaining selectively
1152 @item --retain-symbols-file @var{filename}
1153 Retain @emph{only} the symbols listed in the file @var{filename},
1154 discarding all others. @var{filename} is simply a flat file, with one
1155 symbol name per line. This option is especially useful in environments
1159 where a large global symbol table is accumulated gradually, to conserve
1162 @samp{--retain-symbols-file} does @emph{not} discard undefined symbols,
1163 or symbols needed for relocations.
1165 You may only specify @samp{--retain-symbols-file} once in the command
1166 line. It overrides @samp{-s} and @samp{-S}.
1169 @item -rpath @var{dir}
1170 @cindex runtime library search path
1172 Add a directory to the runtime library search path. This is used when
1173 linking an ELF executable with shared objects. All @code{-rpath}
1174 arguments are concatenated and passed to the runtime linker, which uses
1175 them to locate shared objects at runtime. The @code{-rpath} option is
1176 also used when locating shared objects which are needed by shared
1177 objects explicitly included in the link; see the description of the
1178 @code{-rpath-link} option. If @code{-rpath} is not used when linking an
1179 ELF executable, the contents of the environment variable
1180 @code{LD_RUN_PATH} will be used if it is defined.
1182 The @code{-rpath} option may also be used on SunOS. By default, on
1183 SunOS, the linker will form a runtime search patch out of all the
1184 @code{-L} options it is given. If a @code{-rpath} option is used, the
1185 runtime search path will be formed exclusively using the @code{-rpath}
1186 options, ignoring the @code{-L} options. This can be useful when using
1187 gcc, which adds many @code{-L} options which may be on NFS mounted
1190 For compatibility with other ELF linkers, if the @code{-R} option is
1191 followed by a directory name, rather than a file name, it is treated as
1192 the @code{-rpath} option.
1196 @cindex link-time runtime library search path
1198 @item -rpath-link @var{DIR}
1199 When using ELF or SunOS, one shared library may require another. This
1200 happens when an @code{ld -shared} link includes a shared library as one
1203 When the linker encounters such a dependency when doing a non-shared,
1204 non-relocatable link, it will automatically try to locate the required
1205 shared library and include it in the link, if it is not included
1206 explicitly. In such a case, the @code{-rpath-link} option
1207 specifies the first set of directories to search. The
1208 @code{-rpath-link} option may specify a sequence of directory names
1209 either by specifying a list of names separated by colons, or by
1210 appearing multiple times.
1212 This option should be used with caution as it overrides the search path
1213 that may have been hard compiled into a shared library. In such a case it
1214 is possible to use unintentionally a different search path than the
1215 runtime linker would do.
1217 The linker uses the following search paths to locate required shared
1221 Any directories specified by @code{-rpath-link} options.
1223 Any directories specified by @code{-rpath} options. The difference
1224 between @code{-rpath} and @code{-rpath-link} is that directories
1225 specified by @code{-rpath} options are included in the executable and
1226 used at runtime, whereas the @code{-rpath-link} option is only effective
1227 at link time. It is for the native linker only.
1229 On an ELF system, if the @code{-rpath} and @code{rpath-link} options
1230 were not used, search the contents of the environment variable
1231 @code{LD_RUN_PATH}. It is for the native linker only.
1233 On SunOS, if the @code{-rpath} option was not used, search any
1234 directories specified using @code{-L} options.
1236 For a native linker, the contents of the environment variable
1237 @code{LD_LIBRARY_PATH}.
1239 For a native ELF linker, the directories in @code{DT_RUNPATH} or
1240 @code{DT_RPATH} of a shared library are searched for shared
1241 libraries needed by it. The @code{DT_RPATH} entries are ignored if
1242 @code{DT_RUNPATH} entries exist.
1244 The default directories, normally @file{/lib} and @file{/usr/lib}.
1246 For a native linker on an ELF system, if the file @file{/etc/ld.so.conf}
1247 exists, the list of directories found in that file.
1250 If the required shared library is not found, the linker will issue a
1251 warning and continue with the link.
1258 @cindex shared libraries
1259 Create a shared library. This is currently only supported on ELF, XCOFF
1260 and SunOS platforms. On SunOS, the linker will automatically create a
1261 shared library if the @code{-e} option is not used and there are
1262 undefined symbols in the link.
1265 @kindex --sort-common
1266 This option tells @code{ld} to sort the common symbols by size when it
1267 places them in the appropriate output sections. First come all the one
1268 byte symbols, then all the two bytes, then all the four bytes, and then
1269 everything else. This is to prevent gaps between symbols due to
1270 alignment constraints.
1272 @kindex --split-by-file
1273 @item --split-by-file [@var{size}]
1274 Similar to @code{--split-by-reloc} but creates a new output section for
1275 each input file when @var{size} is reached. @var{size} defaults to a
1276 size of 1 if not given.
1278 @kindex --split-by-reloc
1279 @item --split-by-reloc [@var{count}]
1280 Tries to creates extra sections in the output file so that no single
1281 output section in the file contains more than @var{count} relocations.
1282 This is useful when generating huge relocatable files for downloading into
1283 certain real time kernels with the COFF object file format; since COFF
1284 cannot represent more than 65535 relocations in a single section. Note
1285 that this will fail to work with object file formats which do not
1286 support arbitrary sections. The linker will not split up individual
1287 input sections for redistribution, so if a single input section contains
1288 more than @var{count} relocations one output section will contain that
1289 many relocations. @var{count} defaults to a value of 32768.
1293 Compute and display statistics about the operation of the linker, such
1294 as execution time and memory usage.
1296 @kindex --traditional-format
1297 @cindex traditional format
1298 @item --traditional-format
1299 For some targets, the output of @code{ld} is different in some ways from
1300 the output of some existing linker. This switch requests @code{ld} to
1301 use the traditional format instead.
1304 For example, on SunOS, @code{ld} combines duplicate entries in the
1305 symbol string table. This can reduce the size of an output file with
1306 full debugging information by over 30 percent. Unfortunately, the SunOS
1307 @code{dbx} program can not read the resulting program (@code{gdb} has no
1308 trouble). The @samp{--traditional-format} switch tells @code{ld} to not
1309 combine duplicate entries.
1311 @kindex --section-start @var{sectionname}=@var{org}
1312 @item --section-start @var{sectionname}=@var{org}
1313 Locate a section in the output file at the absolute
1314 address given by @var{org}. You may use this option as many
1315 times as necessary to locate multiple sections in the command
1317 @var{org} must be a single hexadecimal integer;
1318 for compatibility with other linkers, you may omit the leading
1319 @samp{0x} usually associated with hexadecimal values. @emph{Note:} there
1320 should be no white space between @var{sectionname}, the equals
1321 sign (``@key{=}''), and @var{org}.
1323 @kindex -Tbss @var{org}
1324 @kindex -Tdata @var{org}
1325 @kindex -Ttext @var{org}
1326 @cindex segment origins, cmd line
1327 @item -Tbss @var{org}
1328 @itemx -Tdata @var{org}
1329 @itemx -Ttext @var{org}
1330 Use @var{org} as the starting address for---respectively---the
1331 @code{bss}, @code{data}, or the @code{text} segment of the output file.
1332 @var{org} must be a single hexadecimal integer;
1333 for compatibility with other linkers, you may omit the leading
1334 @samp{0x} usually associated with hexadecimal values.
1340 Display the version number for @code{ld} and list the linker emulations
1341 supported. Display which input files can and cannot be opened. Display
1342 the linker script if using a default builtin script.
1344 @kindex --version-script=@var{version-scriptfile}
1345 @cindex version script, symbol versions
1346 @itemx --version-script=@var{version-scriptfile}
1347 Specify the name of a version script to the linker. This is typically
1348 used when creating shared libraries to specify additional information
1349 about the version heirarchy for the library being created. This option
1350 is only meaningful on ELF platforms which support shared libraries.
1353 @kindex --warn-common
1354 @cindex warnings, on combining symbols
1355 @cindex combining symbols, warnings on
1357 Warn when a common symbol is combined with another common symbol or with
1358 a symbol definition. Unix linkers allow this somewhat sloppy practice,
1359 but linkers on some other operating systems do not. This option allows
1360 you to find potential problems from combining global symbols.
1361 Unfortunately, some C libraries use this practice, so you may get some
1362 warnings about symbols in the libraries as well as in your programs.
1364 There are three kinds of global symbols, illustrated here by C examples:
1368 A definition, which goes in the initialized data section of the output
1372 An undefined reference, which does not allocate space.
1373 There must be either a definition or a common symbol for the
1377 A common symbol. If there are only (one or more) common symbols for a
1378 variable, it goes in the uninitialized data area of the output file.
1379 The linker merges multiple common symbols for the same variable into a
1380 single symbol. If they are of different sizes, it picks the largest
1381 size. The linker turns a common symbol into a declaration, if there is
1382 a definition of the same variable.
1385 The @samp{--warn-common} option can produce five kinds of warnings.
1386 Each warning consists of a pair of lines: the first describes the symbol
1387 just encountered, and the second describes the previous symbol
1388 encountered with the same name. One or both of the two symbols will be
1393 Turning a common symbol into a reference, because there is already a
1394 definition for the symbol.
1396 @var{file}(@var{section}): warning: common of `@var{symbol}'
1397 overridden by definition
1398 @var{file}(@var{section}): warning: defined here
1402 Turning a common symbol into a reference, because a later definition for
1403 the symbol is encountered. This is the same as the previous case,
1404 except that the symbols are encountered in a different order.
1406 @var{file}(@var{section}): warning: definition of `@var{symbol}'
1408 @var{file}(@var{section}): warning: common is here
1412 Merging a common symbol with a previous same-sized common symbol.
1414 @var{file}(@var{section}): warning: multiple common
1416 @var{file}(@var{section}): warning: previous common is here
1420 Merging a common symbol with a previous larger common symbol.
1422 @var{file}(@var{section}): warning: common of `@var{symbol}'
1423 overridden by larger common
1424 @var{file}(@var{section}): warning: larger common is here
1428 Merging a common symbol with a previous smaller common symbol. This is
1429 the same as the previous case, except that the symbols are
1430 encountered in a different order.
1432 @var{file}(@var{section}): warning: common of `@var{symbol}'
1433 overriding smaller common
1434 @var{file}(@var{section}): warning: smaller common is here
1438 @kindex --warn-constructors
1439 @item --warn-constructors
1440 Warn if any global constructors are used. This is only useful for a few
1441 object file formats. For formats like COFF or ELF, the linker can not
1442 detect the use of global constructors.
1444 @kindex --warn-multiple-gp
1445 @item --warn-multiple-gp
1446 Warn if multiple global pointer values are required in the output file.
1447 This is only meaningful for certain processors, such as the Alpha.
1448 Specifically, some processors put large-valued constants in a special
1449 section. A special register (the global pointer) points into the middle
1450 of this section, so that constants can be loaded efficiently via a
1451 base-register relative addressing mode. Since the offset in
1452 base-register relative mode is fixed and relatively small (e.g., 16
1453 bits), this limits the maximum size of the constant pool. Thus, in
1454 large programs, it is often necessary to use multiple global pointer
1455 values in order to be able to address all possible constants. This
1456 option causes a warning to be issued whenever this case occurs.
1459 @cindex warnings, on undefined symbols
1460 @cindex undefined symbols, warnings on
1462 Only warn once for each undefined symbol, rather than once per module
1465 @kindex --warn-section-align
1466 @cindex warnings, on section alignment
1467 @cindex section alignment, warnings on
1468 @item --warn-section-align
1469 Warn if the address of an output section is changed because of
1470 alignment. Typically, the alignment will be set by an input section.
1471 The address will only be changed if it not explicitly specified; that
1472 is, if the @code{SECTIONS} command does not specify a start address for
1473 the section (@pxref{SECTIONS}).
1475 @kindex --whole-archive
1476 @cindex including an entire archive
1477 @item --whole-archive
1478 For each archive mentioned on the command line after the
1479 @code{--whole-archive} option, include every object file in the archive
1480 in the link, rather than searching the archive for the required object
1481 files. This is normally used to turn an archive file into a shared
1482 library, forcing every object to be included in the resulting shared
1483 library. This option may be used more than once.
1485 Two notes when using this option from gcc: First, gcc doesn't know
1486 about this option, so you have to use @code{-Wl,-whole-archive}.
1487 Second, don't forget to use @code{-Wl,-no-whole-archive} after your
1488 list of archives, because gcc will add its own list of archives to
1489 your link and you may not want this flag to affect those as well.
1492 @item --wrap @var{symbol}
1493 Use a wrapper function for @var{symbol}. Any undefined reference to
1494 @var{symbol} will be resolved to @code{__wrap_@var{symbol}}. Any
1495 undefined reference to @code{__real_@var{symbol}} will be resolved to
1498 This can be used to provide a wrapper for a system function. The
1499 wrapper function should be called @code{__wrap_@var{symbol}}. If it
1500 wishes to call the system function, it should call
1501 @code{__real_@var{symbol}}.
1503 Here is a trivial example:
1507 __wrap_malloc (int c)
1509 printf ("malloc called with %ld\n", c);
1510 return __real_malloc (c);
1514 If you link other code with this file using @code{--wrap malloc}, then
1515 all calls to @code{malloc} will call the function @code{__wrap_malloc}
1516 instead. The call to @code{__real_malloc} in @code{__wrap_malloc} will
1517 call the real @code{malloc} function.
1519 You may wish to provide a @code{__real_malloc} function as well, so that
1520 links without the @code{--wrap} option will succeed. If you do this,
1521 you should not put the definition of @code{__real_malloc} in the same
1522 file as @code{__wrap_malloc}; if you do, the assembler may resolve the
1523 call before the linker has a chance to wrap it to @code{malloc}.
1525 @kindex --enable-new-dtags
1526 @kindex --disable-new-dtags
1527 @item --enable-new-dtags
1528 @itemx --disable-new-dtags
1529 This linker can create the new dynamic tags in ELF. But the older ELF
1530 systems may not understand them. If you specify
1531 @code{--enable-new-dtags}, the dynamic tags will be created as needed.
1532 If you specify @code{--disable-new-dtags}, no new dynamic tags will be
1533 created. By default, the new dynamic tags are not created. Note that
1534 those options are only available for ELF systems.
1540 @subsection Options specific to i386 PE targets
1542 @c man begin OPTIONS
1544 The i386 PE linker supports the @code{-shared} option, which causes
1545 the output to be a dynamically linked library (DLL) instead of a
1546 normal executable. You should name the output @code{*.dll} when you
1547 use this option. In addition, the linker fully supports the standard
1548 @code{*.def} files, which may be specified on the linker command line
1549 like an object file (in fact, it should precede archives it exports
1550 symbols from, to ensure that they get linked in, just like a normal
1553 In addition to the options common to all targets, the i386 PE linker
1554 support additional command line options that are specific to the i386
1555 PE target. Options that take values may be separated from their
1556 values by either a space or an equals sign.
1560 @kindex --add-stdcall-alias
1561 @item --add-stdcall-alias
1562 If given, symbols with a stdcall suffix (@@@var{nn}) will be exported
1563 as-is and also with the suffix stripped.
1566 @item --base-file @var{file}
1567 Use @var{file} as the name of a file in which to save the base
1568 addresses of all the relocations needed for generating DLLs with
1573 Create a DLL instead of a regular executable. You may also use
1574 @code{-shared} or specify a @code{LIBRARY} in a given @code{.def}
1577 @kindex --enable-stdcall-fixup
1578 @kindex --disable-stdcall-fixup
1579 @item --enable-stdcall-fixup
1580 @itemx --disable-stdcall-fixup
1581 If the link finds a symbol that it cannot resolve, it will attempt to
1582 do "fuzzy linking" by looking for another defined symbol that differs
1583 only in the format of the symbol name (cdecl vs stdcall) and will
1584 resolve that symbol by linking to the match. For example, the
1585 undefined symbol @code{_foo} might be linked to the function
1586 @code{_foo@@12}, or the undefined symbol @code{_bar@@16} might be linked
1587 to the function @code{_bar}. When the linker does this, it prints a
1588 warning, since it normally should have failed to link, but sometimes
1589 import libraries generated from third-party dlls may need this feature
1590 to be usable. If you specify @code{--enable-stdcall-fixup}, this
1591 feature is fully enabled and warnings are not printed. If you specify
1592 @code{--disable-stdcall-fixup}, this feature is disabled and such
1593 mismatches are considered to be errors.
1595 @cindex DLLs, creating
1596 @kindex --export-all-symbols
1597 @item --export-all-symbols
1598 If given, all global symbols in the objects used to build a DLL will
1599 be exported by the DLL. Note that this is the default if there
1600 otherwise wouldn't be any exported symbols. When symbols are
1601 explicitly exported via DEF files or implicitly exported via function
1602 attributes, the default is to not export anything else unless this
1603 option is given. Note that the symbols @code{DllMain@@12},
1604 @code{DllEntryPoint@@0}, @code{DllMainCRTStartup@@12}, and
1605 @code{impure_ptr} will not be automatically
1606 exported. Also, symbols imported from other DLLs will not be
1607 re-exported, nor will symbols specifying the DLL's internal layout
1608 such as those beginning with @code{_head_} or ending with
1609 @code{_iname}. In addition, no symbols from @code{libgcc},
1610 @code{libstd++}, @code{libmingw32}, or @code{crtX.o} will be exported.
1611 Symbols whose names begin with @code{__rtti_} or @code{__builtin_} will
1612 not be exported, to help with C++ DLLs. Finally, there is an
1613 extensive list of cygwin-private symbols that are not exported
1614 (obviously, this applies on when building DLLs for cygwin targets).
1615 These cygwin-excludes are: @code{_cygwin_dll_entry@@12},
1616 @code{_cygwin_crt0_common@@8}, @code{_cygwin_noncygwin_dll_entry@@12},
1617 @code{_fmode}, @code{_impure_ptr}, @code{cygwin_attach_dll},
1618 @code{cygwin_premain0}, @code{cygwin_premain1}, @code{cygwin_premain2},
1619 @code{cygwin_premain3}, and @code{environ}.
1621 @kindex --exclude-symbols
1622 @item --exclude-symbols @var{symbol},@var{symbol},...
1623 Specifies a list of symbols which should not be automatically
1624 exported. The symbol names may be delimited by commas or colons.
1626 @kindex --file-alignment
1627 @item --file-alignment
1628 Specify the file alignment. Sections in the file will always begin at
1629 file offsets which are multiples of this number. This defaults to
1634 @item --heap @var{reserve}
1635 @itemx --heap @var{reserve},@var{commit}
1636 Specify the amount of memory to reserve (and optionally commit) to be
1637 used as heap for this program. The default is 1Mb reserved, 4K
1641 @kindex --image-base
1642 @item --image-base @var{value}
1643 Use @var{value} as the base address of your program or dll. This is
1644 the lowest memory location that will be used when your program or dll
1645 is loaded. To reduce the need to relocate and improve performance of
1646 your dlls, each should have a unique base address and not overlap any
1647 other dlls. The default is 0x400000 for executables, and 0x10000000
1652 If given, the stdcall suffixes (@@@var{nn}) will be stripped from
1653 symbols before they are exported.
1655 @kindex --major-image-version
1656 @item --major-image-version @var{value}
1657 Sets the major number of the "image version". Defaults to 1.
1659 @kindex --major-os-version
1660 @item --major-os-version @var{value}
1661 Sets the major number of the "os version". Defaults to 4.
1663 @kindex --major-subsystem-version
1664 @item --major-subsystem-version @var{value}
1665 Sets the major number of the "subsystem version". Defaults to 4.
1667 @kindex --minor-image-version
1668 @item --minor-image-version @var{value}
1669 Sets the minor number of the "image version". Defaults to 0.
1671 @kindex --minor-os-version
1672 @item --minor-os-version @var{value}
1673 Sets the minor number of the "os version". Defaults to 0.
1675 @kindex --minor-subsystem-version
1676 @item --minor-subsystem-version @var{value}
1677 Sets the minor number of the "subsystem version". Defaults to 0.
1679 @cindex DEF files, creating
1680 @cindex DLLs, creating
1681 @kindex --output-def
1682 @item --output-def @var{file}
1683 The linker will create the file @var{file} which will contain a DEF
1684 file corresponding to the DLL the linker is generating. This DEF file
1685 (which should be called @code{*.def}) may be used to create an import
1686 library with @code{dlltool} or may be used as a reference to
1687 automatically or implicitly exported symbols.
1689 @cindex DLLs, creating
1690 @kindex --out-implib
1691 @item --out-implib @var{file}
1692 The linker will create the file @var{file} which will contain an
1693 import lib corresponding to the DLL the linker is generating. This
1694 import lib (which should be called @code{*.dll.a} or @code{*.a}
1695 may be used to link clients against the generated DLL; this behavior
1696 makes it possible to skip a separate @code{dlltool} import library
1699 @kindex --enable-auto-image-base
1700 @item --enable-auto-image-base
1701 Automatically choose the image base for DLLs, unless one is specified
1702 using the @code{--image-base} argument. By using a hash generated
1703 from the dllname to create unique image bases for each DLL, in-memory
1704 collisions and relocations which can delay program execution are
1707 @kindex --disable-auto-image-base
1708 @item --disable-auto-image-base
1709 Do not automatically generate a unique image base. If there is no
1710 user-specified image base (@code{--image-base}) then use the platform
1713 @cindex DLLs, linking to
1714 @kindex --dll-search-prefix
1715 @item --dll-search-prefix @var{string}
1716 When linking dynamically to a dll without an import library, i
1717 search for @code{<string><basename>.dll} in preference to
1718 @code{lib<basename>.dll}. This behavior allows easy distinction
1719 between DLLs built for the various "subplatforms": native, cygwin,
1720 uwin, pw, etc. For instance, cygwin DLLs typically use
1721 @code{--dll-search-prefix=cyg}.
1723 @kindex --enable-auto-import
1724 @item --enable-auto-import
1725 Do sophisticalted linking of @code{_symbol} to @code{__imp__symbol} for
1726 DATA imports from DLLs, and create the necessary thunking symbols when
1727 building the DLLs with those DATA exports.
1729 @kindex --disable-auto-import
1730 @item --disable-auto-import
1731 Do not attempt to do sophisticalted linking of @code{_symbol} to
1732 @code{__imp__symbol} for DATA imports from DLLs.
1734 @kindex --enable-extra-pe-debug
1735 @item --enable-extra-pe-debug
1736 Show additional debug info related to auto-import symbol thunking.
1738 @kindex --section-alignment
1739 @item --section-alignment
1740 Sets the section alignment. Sections in memory will always begin at
1741 addresses which are a multiple of this number. Defaults to 0x1000.
1745 @item --stack @var{reserve}
1746 @itemx --stack @var{reserve},@var{commit}
1747 Specify the amount of memory to reserve (and optionally commit) to be
1748 used as stack for this program. The default is 32Mb reserved, 4K
1752 @item --subsystem @var{which}
1753 @itemx --subsystem @var{which}:@var{major}
1754 @itemx --subsystem @var{which}:@var{major}.@var{minor}
1755 Specifies the subsystem under which your program will execute. The
1756 legal values for @var{which} are @code{native}, @code{windows},
1757 @code{console}, and @code{posix}. You may optionally set the
1758 subsystem version also.
1766 @section Environment Variables
1768 @c man begin ENVIRONMENT
1770 You can change the behavior of @code{ld} with the environment variables
1771 @code{GNUTARGET}, @code{LDEMULATION}, and @code{COLLECT_NO_DEMANGLE}.
1774 @cindex default input format
1775 @code{GNUTARGET} determines the input-file object format if you don't
1776 use @samp{-b} (or its synonym @samp{--format}). Its value should be one
1777 of the BFD names for an input format (@pxref{BFD}). If there is no
1778 @code{GNUTARGET} in the environment, @code{ld} uses the natural format
1779 of the target. If @code{GNUTARGET} is set to @code{default} then BFD
1780 attempts to discover the input format by examining binary input files;
1781 this method often succeeds, but there are potential ambiguities, since
1782 there is no method of ensuring that the magic number used to specify
1783 object-file formats is unique. However, the configuration procedure for
1784 BFD on each system places the conventional format for that system first
1785 in the search-list, so ambiguities are resolved in favor of convention.
1788 @cindex default emulation
1789 @cindex emulation, default
1790 @code{LDEMULATION} determines the default emulation if you don't use the
1791 @samp{-m} option. The emulation can affect various aspects of linker
1792 behaviour, particularly the default linker script. You can list the
1793 available emulations with the @samp{--verbose} or @samp{-V} options. If
1794 the @samp{-m} option is not used, and the @code{LDEMULATION} environment
1795 variable is not defined, the default emulation depends upon how the
1796 linker was configured.
1798 @kindex COLLECT_NO_DEMANGLE
1799 @cindex demangling, default
1800 Normally, the linker will default to demangling symbols. However, if
1801 @code{COLLECT_NO_DEMANGLE} is set in the environment, then it will
1802 default to not demangling symbols. This environment variable is used in
1803 a similar fashion by the @code{gcc} linker wrapper program. The default
1804 may be overridden by the @samp{--demangle} and @samp{--no-demangle}
1811 @chapter Linker Scripts
1814 @cindex linker scripts
1815 @cindex command files
1816 Every link is controlled by a @dfn{linker script}. This script is
1817 written in the linker command language.
1819 The main purpose of the linker script is to describe how the sections in
1820 the input files should be mapped into the output file, and to control
1821 the memory layout of the output file. Most linker scripts do nothing
1822 more than this. However, when necessary, the linker script can also
1823 direct the linker to perform many other operations, using the commands
1826 The linker always uses a linker script. If you do not supply one
1827 yourself, the linker will use a default script that is compiled into the
1828 linker executable. You can use the @samp{--verbose} command line option
1829 to display the default linker script. Certain command line options,
1830 such as @samp{-r} or @samp{-N}, will affect the default linker script.
1832 You may supply your own linker script by using the @samp{-T} command
1833 line option. When you do this, your linker script will replace the
1834 default linker script.
1836 You may also use linker scripts implicitly by naming them as input files
1837 to the linker, as though they were files to be linked. @xref{Implicit
1841 * Basic Script Concepts:: Basic Linker Script Concepts
1842 * Script Format:: Linker Script Format
1843 * Simple Example:: Simple Linker Script Example
1844 * Simple Commands:: Simple Linker Script Commands
1845 * Assignments:: Assigning Values to Symbols
1846 * SECTIONS:: SECTIONS Command
1847 * MEMORY:: MEMORY Command
1848 * PHDRS:: PHDRS Command
1849 * VERSION:: VERSION Command
1850 * Expressions:: Expressions in Linker Scripts
1851 * Implicit Linker Scripts:: Implicit Linker Scripts
1854 @node Basic Script Concepts
1855 @section Basic Linker Script Concepts
1856 @cindex linker script concepts
1857 We need to define some basic concepts and vocabulary in order to
1858 describe the linker script language.
1860 The linker combines input files into a single output file. The output
1861 file and each input file are in a special data format known as an
1862 @dfn{object file format}. Each file is called an @dfn{object file}.
1863 The output file is often called an @dfn{executable}, but for our
1864 purposes we will also call it an object file. Each object file has,
1865 among other things, a list of @dfn{sections}. We sometimes refer to a
1866 section in an input file as an @dfn{input section}; similarly, a section
1867 in the output file is an @dfn{output section}.
1869 Each section in an object file has a name and a size. Most sections
1870 also have an associated block of data, known as the @dfn{section
1871 contents}. A section may be marked as @dfn{loadable}, which mean that
1872 the contents should be loaded into memory when the output file is run.
1873 A section with no contents may be @dfn{allocatable}, which means that an
1874 area in memory should be set aside, but nothing in particular should be
1875 loaded there (in some cases this memory must be zeroed out). A section
1876 which is neither loadable nor allocatable typically contains some sort
1877 of debugging information.
1879 Every loadable or allocatable output section has two addresses. The
1880 first is the @dfn{VMA}, or virtual memory address. This is the address
1881 the section will have when the output file is run. The second is the
1882 @dfn{LMA}, or load memory address. This is the address at which the
1883 section will be loaded. In most cases the two addresses will be the
1884 same. An example of when they might be different is when a data section
1885 is loaded into ROM, and then copied into RAM when the program starts up
1886 (this technique is often used to initialize global variables in a ROM
1887 based system). In this case the ROM address would be the LMA, and the
1888 RAM address would be the VMA.
1890 You can see the sections in an object file by using the @code{objdump}
1891 program with the @samp{-h} option.
1893 Every object file also has a list of @dfn{symbols}, known as the
1894 @dfn{symbol table}. A symbol may be defined or undefined. Each symbol
1895 has a name, and each defined symbol has an address, among other
1896 information. If you compile a C or C++ program into an object file, you
1897 will get a defined symbol for every defined function and global or
1898 static variable. Every undefined function or global variable which is
1899 referenced in the input file will become an undefined symbol.
1901 You can see the symbols in an object file by using the @code{nm}
1902 program, or by using the @code{objdump} program with the @samp{-t}
1906 @section Linker Script Format
1907 @cindex linker script format
1908 Linker scripts are text files.
1910 You write a linker script as a series of commands. Each command is
1911 either a keyword, possibly followed by arguments, or an assignment to a
1912 symbol. You may separate commands using semicolons. Whitespace is
1915 Strings such as file or format names can normally be entered directly.
1916 If the file name contains a character such as a comma which would
1917 otherwise serve to separate file names, you may put the file name in
1918 double quotes. There is no way to use a double quote character in a
1921 You may include comments in linker scripts just as in C, delimited by
1922 @samp{/*} and @samp{*/}. As in C, comments are syntactically equivalent
1925 @node Simple Example
1926 @section Simple Linker Script Example
1927 @cindex linker script example
1928 @cindex example of linker script
1929 Many linker scripts are fairly simple.
1931 The simplest possible linker script has just one command:
1932 @samp{SECTIONS}. You use the @samp{SECTIONS} command to describe the
1933 memory layout of the output file.
1935 The @samp{SECTIONS} command is a powerful command. Here we will
1936 describe a simple use of it. Let's assume your program consists only of
1937 code, initialized data, and uninitialized data. These will be in the
1938 @samp{.text}, @samp{.data}, and @samp{.bss} sections, respectively.
1939 Let's assume further that these are the only sections which appear in
1942 For this example, let's say that the code should be loaded at address
1943 0x10000, and that the data should start at address 0x8000000. Here is a
1944 linker script which will do that:
1949 .text : @{ *(.text) @}
1951 .data : @{ *(.data) @}
1952 .bss : @{ *(.bss) @}
1956 You write the @samp{SECTIONS} command as the keyword @samp{SECTIONS},
1957 followed by a series of symbol assignments and output section
1958 descriptions enclosed in curly braces.
1960 The first line inside the @samp{SECTIONS} command of the above example
1961 sets the value of the special symbol @samp{.}, which is the location
1962 counter. If you do not specify the address of an output section in some
1963 other way (other ways are described later), the address is set from the
1964 current value of the location counter. The location counter is then
1965 incremented by the size of the output section. At the start of the
1966 @samp{SECTIONS} command, the location counter has the value @samp{0}.
1968 The second line defines an output section, @samp{.text}. The colon is
1969 required syntax which may be ignored for now. Within the curly braces
1970 after the output section name, you list the names of the input sections
1971 which should be placed into this output section. The @samp{*} is a
1972 wildcard which matches any file name. The expression @samp{*(.text)}
1973 means all @samp{.text} input sections in all input files.
1975 Since the location counter is @samp{0x10000} when the output section
1976 @samp{.text} is defined, the linker will set the address of the
1977 @samp{.text} section in the output file to be @samp{0x10000}.
1979 The remaining lines define the @samp{.data} and @samp{.bss} sections in
1980 the output file. The linker will place the @samp{.data} output section
1981 at address @samp{0x8000000}. After the linker places the @samp{.data}
1982 output section, the value of the location counter will be
1983 @samp{0x8000000} plus the size of the @samp{.data} output section. The
1984 effect is that the linker will place the @samp{.bss} output section
1985 immediately after the @samp{.data} output section in memory
1987 The linker will ensure that each output section has the required
1988 alignment, by increasing the location counter if necessary. In this
1989 example, the specified addresses for the @samp{.text} and @samp{.data}
1990 sections will probably satisfy any alignment constraints, but the linker
1991 may have to create a small gap between the @samp{.data} and @samp{.bss}
1994 That's it! That's a simple and complete linker script.
1996 @node Simple Commands
1997 @section Simple Linker Script Commands
1998 @cindex linker script simple commands
1999 In this section we describe the simple linker script commands.
2002 * Entry Point:: Setting the entry point
2003 * File Commands:: Commands dealing with files
2004 @ifclear SingleFormat
2005 * Format Commands:: Commands dealing with object file formats
2008 * Miscellaneous Commands:: Other linker script commands
2012 @subsection Setting the entry point
2013 @kindex ENTRY(@var{symbol})
2014 @cindex start of execution
2015 @cindex first instruction
2017 The first instruction to execute in a program is called the @dfn{entry
2018 point}. You can use the @code{ENTRY} linker script command to set the
2019 entry point. The argument is a symbol name:
2024 There are several ways to set the entry point. The linker will set the
2025 entry point by trying each of the following methods in order, and
2026 stopping when one of them succeeds:
2029 the @samp{-e} @var{entry} command-line option;
2031 the @code{ENTRY(@var{symbol})} command in a linker script;
2033 the value of the symbol @code{start}, if defined;
2035 the address of the first byte of the @samp{.text} section, if present;
2037 The address @code{0}.
2041 @subsection Commands dealing with files
2042 @cindex linker script file commands
2043 Several linker script commands deal with files.
2046 @item INCLUDE @var{filename}
2047 @kindex INCLUDE @var{filename}
2048 @cindex including a linker script
2049 Include the linker script @var{filename} at this point. The file will
2050 be searched for in the current directory, and in any directory specified
2051 with the @code{-L} option. You can nest calls to @code{INCLUDE} up to
2054 @item INPUT(@var{file}, @var{file}, @dots{})
2055 @itemx INPUT(@var{file} @var{file} @dots{})
2056 @kindex INPUT(@var{files})
2057 @cindex input files in linker scripts
2058 @cindex input object files in linker scripts
2059 @cindex linker script input object files
2060 The @code{INPUT} command directs the linker to include the named files
2061 in the link, as though they were named on the command line.
2063 For example, if you always want to include @file{subr.o} any time you do
2064 a link, but you can't be bothered to put it on every link command line,
2065 then you can put @samp{INPUT (subr.o)} in your linker script.
2067 In fact, if you like, you can list all of your input files in the linker
2068 script, and then invoke the linker with nothing but a @samp{-T} option.
2070 The linker will first try to open the file in the current directory. If
2071 it is not found, the linker will search through the archive library
2072 search path. See the description of @samp{-L} in @ref{Options,,Command
2075 If you use @samp{INPUT (-l@var{file})}, @code{ld} will transform the
2076 name to @code{lib@var{file}.a}, as with the command line argument
2079 When you use the @code{INPUT} command in an implicit linker script, the
2080 files will be included in the link at the point at which the linker
2081 script file is included. This can affect archive searching.
2083 @item GROUP(@var{file}, @var{file}, @dots{})
2084 @itemx GROUP(@var{file} @var{file} @dots{})
2085 @kindex GROUP(@var{files})
2086 @cindex grouping input files
2087 The @code{GROUP} command is like @code{INPUT}, except that the named
2088 files should all be archives, and they are searched repeatedly until no
2089 new undefined references are created. See the description of @samp{-(}
2090 in @ref{Options,,Command Line Options}.
2092 @item OUTPUT(@var{filename})
2093 @kindex OUTPUT(@var{filename})
2094 @cindex output file name in linker scripot
2095 The @code{OUTPUT} command names the output file. Using
2096 @code{OUTPUT(@var{filename})} in the linker script is exactly like using
2097 @samp{-o @var{filename}} on the command line (@pxref{Options,,Command
2098 Line Options}). If both are used, the command line option takes
2101 You can use the @code{OUTPUT} command to define a default name for the
2102 output file other than the usual default of @file{a.out}.
2104 @item SEARCH_DIR(@var{path})
2105 @kindex SEARCH_DIR(@var{path})
2106 @cindex library search path in linker script
2107 @cindex archive search path in linker script
2108 @cindex search path in linker script
2109 The @code{SEARCH_DIR} command adds @var{path} to the list of paths where
2110 @code{ld} looks for archive libraries. Using
2111 @code{SEARCH_DIR(@var{path})} is exactly like using @samp{-L @var{path}}
2112 on the command line (@pxref{Options,,Command Line Options}). If both
2113 are used, then the linker will search both paths. Paths specified using
2114 the command line option are searched first.
2116 @item STARTUP(@var{filename})
2117 @kindex STARTUP(@var{filename})
2118 @cindex first input file
2119 The @code{STARTUP} command is just like the @code{INPUT} command, except
2120 that @var{filename} will become the first input file to be linked, as
2121 though it were specified first on the command line. This may be useful
2122 when using a system in which the entry point is always the start of the
2126 @ifclear SingleFormat
2127 @node Format Commands
2128 @subsection Commands dealing with object file formats
2129 A couple of linker script commands deal with object file formats.
2132 @item OUTPUT_FORMAT(@var{bfdname})
2133 @itemx OUTPUT_FORMAT(@var{default}, @var{big}, @var{little})
2134 @kindex OUTPUT_FORMAT(@var{bfdname})
2135 @cindex output file format in linker script
2136 The @code{OUTPUT_FORMAT} command names the BFD format to use for the
2137 output file (@pxref{BFD}). Using @code{OUTPUT_FORMAT(@var{bfdname})} is
2138 exactly like using @samp{-oformat @var{bfdname}} on the command line
2139 (@pxref{Options,,Command Line Options}). If both are used, the command
2140 line option takes precedence.
2142 You can use @code{OUTPUT_FORMAT} with three arguments to use different
2143 formats based on the @samp{-EB} and @samp{-EL} command line options.
2144 This permits the linker script to set the output format based on the
2147 If neither @samp{-EB} nor @samp{-EL} are used, then the output format
2148 will be the first argument, @var{default}. If @samp{-EB} is used, the
2149 output format will be the second argument, @var{big}. If @samp{-EL} is
2150 used, the output format will be the third argument, @var{little}.
2152 For example, the default linker script for the MIPS ELF target uses this
2155 OUTPUT_FORMAT(elf32-bigmips, elf32-bigmips, elf32-littlemips)
2157 This says that the default format for the output file is
2158 @samp{elf32-bigmips}, but if the user uses the @samp{-EL} command line
2159 option, the output file will be created in the @samp{elf32-littlemips}
2162 @item TARGET(@var{bfdname})
2163 @kindex TARGET(@var{bfdname})
2164 @cindex input file format in linker script
2165 The @code{TARGET} command names the BFD format to use when reading input
2166 files. It affects subsequent @code{INPUT} and @code{GROUP} commands.
2167 This command is like using @samp{-b @var{bfdname}} on the command line
2168 (@pxref{Options,,Command Line Options}). If the @code{TARGET} command
2169 is used but @code{OUTPUT_FORMAT} is not, then the last @code{TARGET}
2170 command is also used to set the format for the output file. @xref{BFD}.
2174 @node Miscellaneous Commands
2175 @subsection Other linker script commands
2176 There are a few other linker scripts commands.
2179 @item ASSERT(@var{exp}, @var{message})
2181 @cindex assertion in linker script
2182 Ensure that @var{exp} is non-zero. If it is zero, then exit the linker
2183 with an error code, and print @var{message}.
2185 @item EXTERN(@var{symbol} @var{symbol} @dots{})
2187 @cindex undefined symbol in linker script
2188 Force @var{symbol} to be entered in the output file as an undefined
2189 symbol. Doing this may, for example, trigger linking of additional
2190 modules from standard libraries. You may list several @var{symbol}s for
2191 each @code{EXTERN}, and you may use @code{EXTERN} multiple times. This
2192 command has the same effect as the @samp{-u} command-line option.
2194 @item FORCE_COMMON_ALLOCATION
2195 @kindex FORCE_COMMON_ALLOCATION
2196 @cindex common allocation in linker script
2197 This command has the same effect as the @samp{-d} command-line option:
2198 to make @code{ld} assign space to common symbols even if a relocatable
2199 output file is specified (@samp{-r}).
2201 @item NOCROSSREFS(@var{section} @var{section} @dots{})
2202 @kindex NOCROSSREFS(@var{sections})
2203 @cindex cross references
2204 This command may be used to tell @code{ld} to issue an error about any
2205 references among certain output sections.
2207 In certain types of programs, particularly on embedded systems when
2208 using overlays, when one section is loaded into memory, another section
2209 will not be. Any direct references between the two sections would be
2210 errors. For example, it would be an error if code in one section called
2211 a function defined in the other section.
2213 The @code{NOCROSSREFS} command takes a list of output section names. If
2214 @code{ld} detects any cross references between the sections, it reports
2215 an error and returns a non-zero exit status. Note that the
2216 @code{NOCROSSREFS} command uses output section names, not input section
2219 @ifclear SingleFormat
2220 @item OUTPUT_ARCH(@var{bfdarch})
2221 @kindex OUTPUT_ARCH(@var{bfdarch})
2222 @cindex machine architecture
2223 @cindex architecture
2224 Specify a particular output machine architecture. The argument is one
2225 of the names used by the BFD library (@pxref{BFD}). You can see the
2226 architecture of an object file by using the @code{objdump} program with
2227 the @samp{-f} option.
2232 @section Assigning Values to Symbols
2233 @cindex assignment in scripts
2234 @cindex symbol definition, scripts
2235 @cindex variables, defining
2236 You may assign a value to a symbol in a linker script. This will define
2237 the symbol as a global symbol.
2240 * Simple Assignments:: Simple Assignments
2244 @node Simple Assignments
2245 @subsection Simple Assignments
2247 You may assign to a symbol using any of the C assignment operators:
2250 @item @var{symbol} = @var{expression} ;
2251 @itemx @var{symbol} += @var{expression} ;
2252 @itemx @var{symbol} -= @var{expression} ;
2253 @itemx @var{symbol} *= @var{expression} ;
2254 @itemx @var{symbol} /= @var{expression} ;
2255 @itemx @var{symbol} <<= @var{expression} ;
2256 @itemx @var{symbol} >>= @var{expression} ;
2257 @itemx @var{symbol} &= @var{expression} ;
2258 @itemx @var{symbol} |= @var{expression} ;
2261 The first case will define @var{symbol} to the value of
2262 @var{expression}. In the other cases, @var{symbol} must already be
2263 defined, and the value will be adjusted accordingly.
2265 The special symbol name @samp{.} indicates the location counter. You
2266 may only use this within a @code{SECTIONS} command.
2268 The semicolon after @var{expression} is required.
2270 Expressions are defined below; see @ref{Expressions}.
2272 You may write symbol assignments as commands in their own right, or as
2273 statements within a @code{SECTIONS} command, or as part of an output
2274 section description in a @code{SECTIONS} command.
2276 The section of the symbol will be set from the section of the
2277 expression; for more information, see @ref{Expression Section}.
2279 Here is an example showing the three different places that symbol
2280 assignments may be used:
2291 _bdata = (. + 3) & ~ 3;
2292 .data : @{ *(.data) @}
2296 In this example, the symbol @samp{floating_point} will be defined as
2297 zero. The symbol @samp{_etext} will be defined as the address following
2298 the last @samp{.text} input section. The symbol @samp{_bdata} will be
2299 defined as the address following the @samp{.text} output section aligned
2300 upward to a 4 byte boundary.
2305 In some cases, it is desirable for a linker script to define a symbol
2306 only if it is referenced and is not defined by any object included in
2307 the link. For example, traditional linkers defined the symbol
2308 @samp{etext}. However, ANSI C requires that the user be able to use
2309 @samp{etext} as a function name without encountering an error. The
2310 @code{PROVIDE} keyword may be used to define a symbol, such as
2311 @samp{etext}, only if it is referenced but not defined. The syntax is
2312 @code{PROVIDE(@var{symbol} = @var{expression})}.
2314 Here is an example of using @code{PROVIDE} to define @samp{etext}:
2327 In this example, if the program defines @samp{_etext} (with a leading
2328 underscore), the linker will give a multiple definition error. If, on
2329 the other hand, the program defines @samp{etext} (with no leading
2330 underscore), the linker will silently use the definition in the program.
2331 If the program references @samp{etext} but does not define it, the
2332 linker will use the definition in the linker script.
2335 @section SECTIONS command
2337 The @code{SECTIONS} command tells the linker how to map input sections
2338 into output sections, and how to place the output sections in memory.
2340 The format of the @code{SECTIONS} command is:
2344 @var{sections-command}
2345 @var{sections-command}
2350 Each @var{sections-command} may of be one of the following:
2354 an @code{ENTRY} command (@pxref{Entry Point,,Entry command})
2356 a symbol assignment (@pxref{Assignments})
2358 an output section description
2360 an overlay description
2363 The @code{ENTRY} command and symbol assignments are permitted inside the
2364 @code{SECTIONS} command for convenience in using the location counter in
2365 those commands. This can also make the linker script easier to
2366 understand because you can use those commands at meaningful points in
2367 the layout of the output file.
2369 Output section descriptions and overlay descriptions are described
2372 If you do not use a @code{SECTIONS} command in your linker script, the
2373 linker will place each input section into an identically named output
2374 section in the order that the sections are first encountered in the
2375 input files. If all input sections are present in the first file, for
2376 example, the order of sections in the output file will match the order
2377 in the first input file. The first section will be at address zero.
2380 * Output Section Description:: Output section description
2381 * Output Section Name:: Output section name
2382 * Output Section Address:: Output section address
2383 * Input Section:: Input section description
2384 * Output Section Data:: Output section data
2385 * Output Section Keywords:: Output section keywords
2386 * Output Section Discarding:: Output section discarding
2387 * Output Section Attributes:: Output section attributes
2388 * Overlay Description:: Overlay description
2391 @node Output Section Description
2392 @subsection Output section description
2393 The full description of an output section looks like this:
2396 @var{section} [@var{address}] [(@var{type})] : [AT(@var{lma})]
2398 @var{output-section-command}
2399 @var{output-section-command}
2401 @} [>@var{region}] [AT>@var{lma_region}] [:@var{phdr} :@var{phdr} @dots{}] [=@var{fillexp}]
2405 Most output sections do not use most of the optional section attributes.
2407 The whitespace around @var{section} is required, so that the section
2408 name is unambiguous. The colon and the curly braces are also required.
2409 The line breaks and other white space are optional.
2411 Each @var{output-section-command} may be one of the following:
2415 a symbol assignment (@pxref{Assignments})
2417 an input section description (@pxref{Input Section})
2419 data values to include directly (@pxref{Output Section Data})
2421 a special output section keyword (@pxref{Output Section Keywords})
2424 @node Output Section Name
2425 @subsection Output section name
2426 @cindex name, section
2427 @cindex section name
2428 The name of the output section is @var{section}. @var{section} must
2429 meet the constraints of your output format. In formats which only
2430 support a limited number of sections, such as @code{a.out}, the name
2431 must be one of the names supported by the format (@code{a.out}, for
2432 example, allows only @samp{.text}, @samp{.data} or @samp{.bss}). If the
2433 output format supports any number of sections, but with numbers and not
2434 names (as is the case for Oasys), the name should be supplied as a
2435 quoted numeric string. A section name may consist of any sequence of
2436 characters, but a name which contains any unusual characters such as
2437 commas must be quoted.
2439 The output section name @samp{/DISCARD/} is special; @ref{Output Section
2442 @node Output Section Address
2443 @subsection Output section address
2444 @cindex address, section
2445 @cindex section address
2446 The @var{address} is an expression for the VMA (the virtual memory
2447 address) of the output section. If you do not provide @var{address},
2448 the linker will set it based on @var{region} if present, or otherwise
2449 based on the current value of the location counter.
2451 If you provide @var{address}, the address of the output section will be
2452 set to precisely that. If you provide neither @var{address} nor
2453 @var{region}, then the address of the output section will be set to the
2454 current value of the location counter aligned to the alignment
2455 requirements of the output section. The alignment requirement of the
2456 output section is the strictest alignment of any input section contained
2457 within the output section.
2461 .text . : @{ *(.text) @}
2466 .text : @{ *(.text) @}
2469 are subtly different. The first will set the address of the
2470 @samp{.text} output section to the current value of the location
2471 counter. The second will set it to the current value of the location
2472 counter aligned to the strictest alignment of a @samp{.text} input
2475 The @var{address} may be an arbitrary expression; @ref{Expressions}.
2476 For example, if you want to align the section on a 0x10 byte boundary,
2477 so that the lowest four bits of the section address are zero, you could
2478 do something like this:
2480 .text ALIGN(0x10) : @{ *(.text) @}
2483 This works because @code{ALIGN} returns the current location counter
2484 aligned upward to the specified value.
2486 Specifying @var{address} for a section will change the value of the
2490 @subsection Input section description
2491 @cindex input sections
2492 @cindex mapping input sections to output sections
2493 The most common output section command is an input section description.
2495 The input section description is the most basic linker script operation.
2496 You use output sections to tell the linker how to lay out your program
2497 in memory. You use input section descriptions to tell the linker how to
2498 map the input files into your memory layout.
2501 * Input Section Basics:: Input section basics
2502 * Input Section Wildcards:: Input section wildcard patterns
2503 * Input Section Common:: Input section for common symbols
2504 * Input Section Keep:: Input section and garbage collection
2505 * Input Section Example:: Input section example
2508 @node Input Section Basics
2509 @subsubsection Input section basics
2510 @cindex input section basics
2511 An input section description consists of a file name optionally followed
2512 by a list of section names in parentheses.
2514 The file name and the section name may be wildcard patterns, which we
2515 describe further below (@pxref{Input Section Wildcards}).
2517 The most common input section description is to include all input
2518 sections with a particular name in the output section. For example, to
2519 include all input @samp{.text} sections, you would write:
2524 Here the @samp{*} is a wildcard which matches any file name. To exclude a list
2525 of files from matching the file name wildcard, EXCLUDE_FILE may be used to
2526 match all files except the ones specified in the EXCLUDE_FILE list. For
2529 (*(EXCLUDE_FILE (*crtend.o *otherfile.o) .ctors))
2531 will cause all .ctors sections from all files except @file{crtend.o} and
2532 @file{otherfile.o} to be included.
2534 There are two ways to include more than one section:
2540 The difference between these is the order in which the @samp{.text} and
2541 @samp{.rdata} input sections will appear in the output section. In the
2542 first example, they will be intermingled. In the second example, all
2543 @samp{.text} input sections will appear first, followed by all
2544 @samp{.rdata} input sections.
2546 You can specify a file name to include sections from a particular file.
2547 You would do this if one or more of your files contain special data that
2548 needs to be at a particular location in memory. For example:
2553 If you use a file name without a list of sections, then all sections in
2554 the input file will be included in the output section. This is not
2555 commonly done, but it may by useful on occasion. For example:
2560 When you use a file name which does not contain any wild card
2561 characters, the linker will first see if you also specified the file
2562 name on the linker command line or in an @code{INPUT} command. If you
2563 did not, the linker will attempt to open the file as an input file, as
2564 though it appeared on the command line. Note that this differs from an
2565 @code{INPUT} command, because the linker will not search for the file in
2566 the archive search path.
2568 @node Input Section Wildcards
2569 @subsubsection Input section wildcard patterns
2570 @cindex input section wildcards
2571 @cindex wildcard file name patterns
2572 @cindex file name wildcard patterns
2573 @cindex section name wildcard patterns
2574 In an input section description, either the file name or the section
2575 name or both may be wildcard patterns.
2577 The file name of @samp{*} seen in many examples is a simple wildcard
2578 pattern for the file name.
2580 The wildcard patterns are like those used by the Unix shell.
2584 matches any number of characters
2586 matches any single character
2588 matches a single instance of any of the @var{chars}; the @samp{-}
2589 character may be used to specify a range of characters, as in
2590 @samp{[a-z]} to match any lower case letter
2592 quotes the following character
2595 When a file name is matched with a wildcard, the wildcard characters
2596 will not match a @samp{/} character (used to separate directory names on
2597 Unix). A pattern consisting of a single @samp{*} character is an
2598 exception; it will always match any file name, whether it contains a
2599 @samp{/} or not. In a section name, the wildcard characters will match
2600 a @samp{/} character.
2602 File name wildcard patterns only match files which are explicitly
2603 specified on the command line or in an @code{INPUT} command. The linker
2604 does not search directories to expand wildcards.
2606 If a file name matches more than one wildcard pattern, or if a file name
2607 appears explicitly and is also matched by a wildcard pattern, the linker
2608 will use the first match in the linker script. For example, this
2609 sequence of input section descriptions is probably in error, because the
2610 @file{data.o} rule will not be used:
2612 .data : @{ *(.data) @}
2613 .data1 : @{ data.o(.data) @}
2617 Normally, the linker will place files and sections matched by wildcards
2618 in the order in which they are seen during the link. You can change
2619 this by using the @code{SORT} keyword, which appears before a wildcard
2620 pattern in parentheses (e.g., @code{SORT(.text*)}). When the
2621 @code{SORT} keyword is used, the linker will sort the files or sections
2622 into ascending order by name before placing them in the output file.
2624 If you ever get confused about where input sections are going, use the
2625 @samp{-M} linker option to generate a map file. The map file shows
2626 precisely how input sections are mapped to output sections.
2628 This example shows how wildcard patterns might be used to partition
2629 files. This linker script directs the linker to place all @samp{.text}
2630 sections in @samp{.text} and all @samp{.bss} sections in @samp{.bss}.
2631 The linker will place the @samp{.data} section from all files beginning
2632 with an upper case character in @samp{.DATA}; for all other files, the
2633 linker will place the @samp{.data} section in @samp{.data}.
2637 .text : @{ *(.text) @}
2638 .DATA : @{ [A-Z]*(.data) @}
2639 .data : @{ *(.data) @}
2640 .bss : @{ *(.bss) @}
2645 @node Input Section Common
2646 @subsubsection Input section for common symbols
2647 @cindex common symbol placement
2648 @cindex uninitialized data placement
2649 A special notation is needed for common symbols, because in many object
2650 file formats common symbols do not have a particular input section. The
2651 linker treats common symbols as though they are in an input section
2652 named @samp{COMMON}.
2654 You may use file names with the @samp{COMMON} section just as with any
2655 other input sections. You can use this to place common symbols from a
2656 particular input file in one section while common symbols from other
2657 input files are placed in another section.
2659 In most cases, common symbols in input files will be placed in the
2660 @samp{.bss} section in the output file. For example:
2662 .bss @{ *(.bss) *(COMMON) @}
2665 @cindex scommon section
2666 @cindex small common symbols
2667 Some object file formats have more than one type of common symbol. For
2668 example, the MIPS ELF object file format distinguishes standard common
2669 symbols and small common symbols. In this case, the linker will use a
2670 different special section name for other types of common symbols. In
2671 the case of MIPS ELF, the linker uses @samp{COMMON} for standard common
2672 symbols and @samp{.scommon} for small common symbols. This permits you
2673 to map the different types of common symbols into memory at different
2677 You will sometimes see @samp{[COMMON]} in old linker scripts. This
2678 notation is now considered obsolete. It is equivalent to
2681 @node Input Section Keep
2682 @subsubsection Input section and garbage collection
2684 @cindex garbage collection
2685 When link-time garbage collection is in use (@samp{--gc-sections}),
2686 it is often useful to mark sections that should not be eliminated.
2687 This is accomplished by surrounding an input section's wildcard entry
2688 with @code{KEEP()}, as in @code{KEEP(*(.init))} or
2689 @code{KEEP(SORT(*)(.ctors))}.
2691 @node Input Section Example
2692 @subsubsection Input section example
2693 The following example is a complete linker script. It tells the linker
2694 to read all of the sections from file @file{all.o} and place them at the
2695 start of output section @samp{outputa} which starts at location
2696 @samp{0x10000}. All of section @samp{.input1} from file @file{foo.o}
2697 follows immediately, in the same output section. All of section
2698 @samp{.input2} from @file{foo.o} goes into output section
2699 @samp{outputb}, followed by section @samp{.input1} from @file{foo1.o}.
2700 All of the remaining @samp{.input1} and @samp{.input2} sections from any
2701 files are written to output section @samp{outputc}.
2725 @node Output Section Data
2726 @subsection Output section data
2728 @cindex section data
2729 @cindex output section data
2730 @kindex BYTE(@var{expression})
2731 @kindex SHORT(@var{expression})
2732 @kindex LONG(@var{expression})
2733 @kindex QUAD(@var{expression})
2734 @kindex SQUAD(@var{expression})
2735 You can include explicit bytes of data in an output section by using
2736 @code{BYTE}, @code{SHORT}, @code{LONG}, @code{QUAD}, or @code{SQUAD} as
2737 an output section command. Each keyword is followed by an expression in
2738 parentheses providing the value to store (@pxref{Expressions}). The
2739 value of the expression is stored at the current value of the location
2742 The @code{BYTE}, @code{SHORT}, @code{LONG}, and @code{QUAD} commands
2743 store one, two, four, and eight bytes (respectively). After storing the
2744 bytes, the location counter is incremented by the number of bytes
2747 For example, this will store the byte 1 followed by the four byte value
2748 of the symbol @samp{addr}:
2754 When using a 64 bit host or target, @code{QUAD} and @code{SQUAD} are the
2755 same; they both store an 8 byte, or 64 bit, value. When both host and
2756 target are 32 bits, an expression is computed as 32 bits. In this case
2757 @code{QUAD} stores a 32 bit value zero extended to 64 bits, and
2758 @code{SQUAD} stores a 32 bit value sign extended to 64 bits.
2760 If the object file format of the output file has an explicit endianness,
2761 which is the normal case, the value will be stored in that endianness.
2762 When the object file format does not have an explicit endianness, as is
2763 true of, for example, S-records, the value will be stored in the
2764 endianness of the first input object file.
2766 Note - these commands only work inside a section description and not
2767 between them, so the following will produce an error from the linker:
2769 SECTIONS @{@ .text : @{@ *(.text) @}@ LONG(1) .data : @{@ *(.data) @}@ @}@
2771 whereas this will work:
2773 SECTIONS @{@ .text : @{@ *(.text) ; LONG(1) @}@ .data : @{@ *(.data) @}@ @}@
2776 @kindex FILL(@var{expression})
2777 @cindex holes, filling
2778 @cindex unspecified memory
2779 You may use the @code{FILL} command to set the fill pattern for the
2780 current section. It is followed by an expression in parentheses. Any
2781 otherwise unspecified regions of memory within the section (for example,
2782 gaps left due to the required alignment of input sections) are filled
2783 with the two least significant bytes of the expression, repeated as
2784 necessary. A @code{FILL} statement covers memory locations after the
2785 point at which it occurs in the section definition; by including more
2786 than one @code{FILL} statement, you can have different fill patterns in
2787 different parts of an output section.
2789 This example shows how to fill unspecified regions of memory with the
2790 value @samp{0x9090}:
2795 The @code{FILL} command is similar to the @samp{=@var{fillexp}} output
2796 section attribute (@pxref{Output Section Fill}), but it only affects the
2797 part of the section following the @code{FILL} command, rather than the
2798 entire section. If both are used, the @code{FILL} command takes
2801 @node Output Section Keywords
2802 @subsection Output section keywords
2803 There are a couple of keywords which can appear as output section
2807 @kindex CREATE_OBJECT_SYMBOLS
2808 @cindex input filename symbols
2809 @cindex filename symbols
2810 @item CREATE_OBJECT_SYMBOLS
2811 The command tells the linker to create a symbol for each input file.
2812 The name of each symbol will be the name of the corresponding input
2813 file. The section of each symbol will be the output section in which
2814 the @code{CREATE_OBJECT_SYMBOLS} command appears.
2816 This is conventional for the a.out object file format. It is not
2817 normally used for any other object file format.
2819 @kindex CONSTRUCTORS
2820 @cindex C++ constructors, arranging in link
2821 @cindex constructors, arranging in link
2823 When linking using the a.out object file format, the linker uses an
2824 unusual set construct to support C++ global constructors and
2825 destructors. When linking object file formats which do not support
2826 arbitrary sections, such as ECOFF and XCOFF, the linker will
2827 automatically recognize C++ global constructors and destructors by name.
2828 For these object file formats, the @code{CONSTRUCTORS} command tells the
2829 linker to place constructor information in the output section where the
2830 @code{CONSTRUCTORS} command appears. The @code{CONSTRUCTORS} command is
2831 ignored for other object file formats.
2833 The symbol @w{@code{__CTOR_LIST__}} marks the start of the global
2834 constructors, and the symbol @w{@code{__DTOR_LIST}} marks the end. The
2835 first word in the list is the number of entries, followed by the address
2836 of each constructor or destructor, followed by a zero word. The
2837 compiler must arrange to actually run the code. For these object file
2838 formats @sc{gnu} C++ normally calls constructors from a subroutine
2839 @code{__main}; a call to @code{__main} is automatically inserted into
2840 the startup code for @code{main}. @sc{gnu} C++ normally runs
2841 destructors either by using @code{atexit}, or directly from the function
2844 For object file formats such as @code{COFF} or @code{ELF} which support
2845 arbitrary section names, @sc{gnu} C++ will normally arrange to put the
2846 addresses of global constructors and destructors into the @code{.ctors}
2847 and @code{.dtors} sections. Placing the following sequence into your
2848 linker script will build the sort of table which the @sc{gnu} C++
2849 runtime code expects to see.
2853 LONG((__CTOR_END__ - __CTOR_LIST__) / 4 - 2)
2858 LONG((__DTOR_END__ - __DTOR_LIST__) / 4 - 2)
2864 If you are using the @sc{gnu} C++ support for initialization priority,
2865 which provides some control over the order in which global constructors
2866 are run, you must sort the constructors at link time to ensure that they
2867 are executed in the correct order. When using the @code{CONSTRUCTORS}
2868 command, use @samp{SORT(CONSTRUCTORS)} instead. When using the
2869 @code{.ctors} and @code{.dtors} sections, use @samp{*(SORT(.ctors))} and
2870 @samp{*(SORT(.dtors))} instead of just @samp{*(.ctors)} and
2873 Normally the compiler and linker will handle these issues automatically,
2874 and you will not need to concern yourself with them. However, you may
2875 need to consider this if you are using C++ and writing your own linker
2880 @node Output Section Discarding
2881 @subsection Output section discarding
2882 @cindex discarding sections
2883 @cindex sections, discarding
2884 @cindex removing sections
2885 The linker will not create output section which do not have any
2886 contents. This is for convenience when referring to input sections that
2887 may or may not be present in any of the input files. For example:
2892 will only create a @samp{.foo} section in the output file if there is a
2893 @samp{.foo} section in at least one input file.
2895 If you use anything other than an input section description as an output
2896 section command, such as a symbol assignment, then the output section
2897 will always be created, even if there are no matching input sections.
2900 The special output section name @samp{/DISCARD/} may be used to discard
2901 input sections. Any input sections which are assigned to an output
2902 section named @samp{/DISCARD/} are not included in the output file.
2904 @node Output Section Attributes
2905 @subsection Output section attributes
2906 @cindex output section attributes
2907 We showed above that the full description of an output section looked
2911 @var{section} [@var{address}] [(@var{type})] : [AT(@var{lma})]
2913 @var{output-section-command}
2914 @var{output-section-command}
2916 @} [>@var{region}] [AT>@var{lma_region}] [:@var{phdr} :@var{phdr} @dots{}] [=@var{fillexp}]
2919 We've already described @var{section}, @var{address}, and
2920 @var{output-section-command}. In this section we will describe the
2921 remaining section attributes.
2924 * Output Section Type:: Output section type
2925 * Output Section LMA:: Output section LMA
2926 * Output Section Region:: Output section region
2927 * Output Section Phdr:: Output section phdr
2928 * Output Section Fill:: Output section fill
2931 @node Output Section Type
2932 @subsubsection Output section type
2933 Each output section may have a type. The type is a keyword in
2934 parentheses. The following types are defined:
2938 The section should be marked as not loadable, so that it will not be
2939 loaded into memory when the program is run.
2944 These type names are supported for backward compatibility, and are
2945 rarely used. They all have the same effect: the section should be
2946 marked as not allocatable, so that no memory is allocated for the
2947 section when the program is run.
2951 @cindex prevent unnecessary loading
2952 @cindex loading, preventing
2953 The linker normally sets the attributes of an output section based on
2954 the input sections which map into it. You can override this by using
2955 the section type. For example, in the script sample below, the
2956 @samp{ROM} section is addressed at memory location @samp{0} and does not
2957 need to be loaded when the program is run. The contents of the
2958 @samp{ROM} section will appear in the linker output file as usual.
2962 ROM 0 (NOLOAD) : @{ @dots{} @}
2968 @node Output Section LMA
2969 @subsubsection Output section LMA
2970 @kindex AT>@var{lma_region}
2971 @kindex AT(@var{lma})
2972 @cindex load address
2973 @cindex section load address
2974 Every section has a virtual address (VMA) and a load address (LMA); see
2975 @ref{Basic Script Concepts}. The address expression which may appear in
2976 an output section description sets the VMA (@pxref{Output Section
2979 The linker will normally set the LMA equal to the VMA. You can change
2980 that by using the @code{AT} keyword. The expression @var{lma} that
2981 follows the @code{AT} keyword specifies the load address of the
2982 section. Alternatively, with @samp{AT>@var{lma_region}} expression,
2983 you may specify a memory region for the section's load address. @xref{MEMORY}.
2985 @cindex ROM initialized data
2986 @cindex initialized data in ROM
2987 This feature is designed to make it easy to build a ROM image. For
2988 example, the following linker script creates three output sections: one
2989 called @samp{.text}, which starts at @code{0x1000}, one called
2990 @samp{.mdata}, which is loaded at the end of the @samp{.text} section
2991 even though its VMA is @code{0x2000}, and one called @samp{.bss} to hold
2992 uninitialized data at address @code{0x3000}. The symbol @code{_data} is
2993 defined with the value @code{0x2000}, which shows that the location
2994 counter holds the VMA value, not the LMA value.
3000 .text 0x1000 : @{ *(.text) _etext = . ; @}
3002 AT ( ADDR (.text) + SIZEOF (.text) )
3003 @{ _data = . ; *(.data); _edata = . ; @}
3005 @{ _bstart = . ; *(.bss) *(COMMON) ; _bend = . ;@}
3010 The run-time initialization code for use with a program generated with
3011 this linker script would include something like the following, to copy
3012 the initialized data from the ROM image to its runtime address. Notice
3013 how this code takes advantage of the symbols defined by the linker
3018 extern char _etext, _data, _edata, _bstart, _bend;
3019 char *src = &_etext;
3022 /* ROM has data at end of text; copy it. */
3023 while (dst < &_edata) @{
3028 for (dst = &_bstart; dst< &_bend; dst++)
3033 @node Output Section Region
3034 @subsubsection Output section region
3035 @kindex >@var{region}
3036 @cindex section, assigning to memory region
3037 @cindex memory regions and sections
3038 You can assign a section to a previously defined region of memory by
3039 using @samp{>@var{region}}. @xref{MEMORY}.
3041 Here is a simple example:
3044 MEMORY @{ rom : ORIGIN = 0x1000, LENGTH = 0x1000 @}
3045 SECTIONS @{ ROM : @{ *(.text) @} >rom @}
3049 @node Output Section Phdr
3050 @subsubsection Output section phdr
3052 @cindex section, assigning to program header
3053 @cindex program headers and sections
3054 You can assign a section to a previously defined program segment by
3055 using @samp{:@var{phdr}}. @xref{PHDRS}. If a section is assigned to
3056 one or more segments, then all subsequent allocated sections will be
3057 assigned to those segments as well, unless they use an explicitly
3058 @code{:@var{phdr}} modifier. You can use @code{:NONE} to tell the
3059 linker to not put the section in any segment at all.
3061 Here is a simple example:
3064 PHDRS @{ text PT_LOAD ; @}
3065 SECTIONS @{ .text : @{ *(.text) @} :text @}
3069 @node Output Section Fill
3070 @subsubsection Output section fill
3071 @kindex =@var{fillexp}
3072 @cindex section fill pattern
3073 @cindex fill pattern, entire section
3074 You can set the fill pattern for an entire section by using
3075 @samp{=@var{fillexp}}. @var{fillexp} is an expression
3076 (@pxref{Expressions}). Any otherwise unspecified regions of memory
3077 within the output section (for example, gaps left due to the required
3078 alignment of input sections) will be filled with the two least
3079 significant bytes of the value, repeated as necessary.
3081 You can also change the fill value with a @code{FILL} command in the
3082 output section commands; see @ref{Output Section Data}.
3084 Here is a simple example:
3087 SECTIONS @{ .text : @{ *(.text) @} =0x9090 @}
3091 @node Overlay Description
3092 @subsection Overlay description
3095 An overlay description provides an easy way to describe sections which
3096 are to be loaded as part of a single memory image but are to be run at
3097 the same memory address. At run time, some sort of overlay manager will
3098 copy the overlaid sections in and out of the runtime memory address as
3099 required, perhaps by simply manipulating addressing bits. This approach
3100 can be useful, for example, when a certain region of memory is faster
3103 Overlays are described using the @code{OVERLAY} command. The
3104 @code{OVERLAY} command is used within a @code{SECTIONS} command, like an
3105 output section description. The full syntax of the @code{OVERLAY}
3106 command is as follows:
3109 OVERLAY [@var{start}] : [NOCROSSREFS] [AT ( @var{ldaddr} )]
3113 @var{output-section-command}
3114 @var{output-section-command}
3116 @} [:@var{phdr}@dots{}] [=@var{fill}]
3119 @var{output-section-command}
3120 @var{output-section-command}
3122 @} [:@var{phdr}@dots{}] [=@var{fill}]
3124 @} [>@var{region}] [:@var{phdr}@dots{}] [=@var{fill}]
3128 Everything is optional except @code{OVERLAY} (a keyword), and each
3129 section must have a name (@var{secname1} and @var{secname2} above). The
3130 section definitions within the @code{OVERLAY} construct are identical to
3131 those within the general @code{SECTIONS} contruct (@pxref{SECTIONS}),
3132 except that no addresses and no memory regions may be defined for
3133 sections within an @code{OVERLAY}.
3135 The sections are all defined with the same starting address. The load
3136 addresses of the sections are arranged such that they are consecutive in
3137 memory starting at the load address used for the @code{OVERLAY} as a
3138 whole (as with normal section definitions, the load address is optional,
3139 and defaults to the start address; the start address is also optional,
3140 and defaults to the current value of the location counter).
3142 If the @code{NOCROSSREFS} keyword is used, and there any references
3143 among the sections, the linker will report an error. Since the sections
3144 all run at the same address, it normally does not make sense for one
3145 section to refer directly to another. @xref{Miscellaneous Commands,
3148 For each section within the @code{OVERLAY}, the linker automatically
3149 defines two symbols. The symbol @code{__load_start_@var{secname}} is
3150 defined as the starting load address of the section. The symbol
3151 @code{__load_stop_@var{secname}} is defined as the final load address of
3152 the section. Any characters within @var{secname} which are not legal
3153 within C identifiers are removed. C (or assembler) code may use these
3154 symbols to move the overlaid sections around as necessary.
3156 At the end of the overlay, the value of the location counter is set to
3157 the start address of the overlay plus the size of the largest section.
3159 Here is an example. Remember that this would appear inside a
3160 @code{SECTIONS} construct.
3163 OVERLAY 0x1000 : AT (0x4000)
3165 .text0 @{ o1/*.o(.text) @}
3166 .text1 @{ o2/*.o(.text) @}
3171 This will define both @samp{.text0} and @samp{.text1} to start at
3172 address 0x1000. @samp{.text0} will be loaded at address 0x4000, and
3173 @samp{.text1} will be loaded immediately after @samp{.text0}. The
3174 following symbols will be defined: @code{__load_start_text0},
3175 @code{__load_stop_text0}, @code{__load_start_text1},
3176 @code{__load_stop_text1}.
3178 C code to copy overlay @code{.text1} into the overlay area might look
3183 extern char __load_start_text1, __load_stop_text1;
3184 memcpy ((char *) 0x1000, &__load_start_text1,
3185 &__load_stop_text1 - &__load_start_text1);
3189 Note that the @code{OVERLAY} command is just syntactic sugar, since
3190 everything it does can be done using the more basic commands. The above
3191 example could have been written identically as follows.
3195 .text0 0x1000 : AT (0x4000) @{ o1/*.o(.text) @}
3196 __load_start_text0 = LOADADDR (.text0);
3197 __load_stop_text0 = LOADADDR (.text0) + SIZEOF (.text0);
3198 .text1 0x1000 : AT (0x4000 + SIZEOF (.text0)) @{ o2/*.o(.text) @}
3199 __load_start_text1 = LOADADDR (.text1);
3200 __load_stop_text1 = LOADADDR (.text1) + SIZEOF (.text1);
3201 . = 0x1000 + MAX (SIZEOF (.text0), SIZEOF (.text1));
3206 @section MEMORY command
3208 @cindex memory regions
3209 @cindex regions of memory
3210 @cindex allocating memory
3211 @cindex discontinuous memory
3212 The linker's default configuration permits allocation of all available
3213 memory. You can override this by using the @code{MEMORY} command.
3215 The @code{MEMORY} command describes the location and size of blocks of
3216 memory in the target. You can use it to describe which memory regions
3217 may be used by the linker, and which memory regions it must avoid. You
3218 can then assign sections to particular memory regions. The linker will
3219 set section addresses based on the memory regions, and will warn about
3220 regions that become too full. The linker will not shuffle sections
3221 around to fit into the available regions.
3223 A linker script may contain at most one use of the @code{MEMORY}
3224 command. However, you can define as many blocks of memory within it as
3225 you wish. The syntax is:
3230 @var{name} [(@var{attr})] : ORIGIN = @var{origin}, LENGTH = @var{len}
3236 The @var{name} is a name used in the linker script to refer to the
3237 region. The region name has no meaning outside of the linker script.
3238 Region names are stored in a separate name space, and will not conflict
3239 with symbol names, file names, or section names. Each memory region
3240 must have a distinct name.
3242 @cindex memory region attributes
3243 The @var{attr} string is an optional list of attributes that specify
3244 whether to use a particular memory region for an input section which is
3245 not explicitly mapped in the linker script. As described in
3246 @ref{SECTIONS}, if you do not specify an output section for some input
3247 section, the linker will create an output section with the same name as
3248 the input section. If you define region attributes, the linker will use
3249 them to select the memory region for the output section that it creates.
3251 The @var{attr} string must consist only of the following characters:
3266 Invert the sense of any of the preceding attributes
3269 If a unmapped section matches any of the listed attributes other than
3270 @samp{!}, it will be placed in the memory region. The @samp{!}
3271 attribute reverses this test, so that an unmapped section will be placed
3272 in the memory region only if it does not match any of the listed
3278 The @var{origin} is an expression for the start address of the memory
3279 region. The expression must evaluate to a constant before memory
3280 allocation is performed, which means that you may not use any section
3281 relative symbols. The keyword @code{ORIGIN} may be abbreviated to
3282 @code{org} or @code{o} (but not, for example, @code{ORG}).
3287 The @var{len} is an expression for the size in bytes of the memory
3288 region. As with the @var{origin} expression, the expression must
3289 evaluate to a constant before memory allocation is performed. The
3290 keyword @code{LENGTH} may be abbreviated to @code{len} or @code{l}.
3292 In the following example, we specify that there are two memory regions
3293 available for allocation: one starting at @samp{0} for 256 kilobytes,
3294 and the other starting at @samp{0x40000000} for four megabytes. The
3295 linker will place into the @samp{rom} memory region every section which
3296 is not explicitly mapped into a memory region, and is either read-only
3297 or executable. The linker will place other sections which are not
3298 explicitly mapped into a memory region into the @samp{ram} memory
3305 rom (rx) : ORIGIN = 0, LENGTH = 256K
3306 ram (!rx) : org = 0x40000000, l = 4M
3311 Once you define a memory region, you can direct the linker to place
3312 specific output sections into that memory region by using the
3313 @samp{>@var{region}} output section attribute. For example, if you have
3314 a memory region named @samp{mem}, you would use @samp{>mem} in the
3315 output section definition. @xref{Output Section Region}. If no address
3316 was specified for the output section, the linker will set the address to
3317 the next available address within the memory region. If the combined
3318 output sections directed to a memory region are too large for the
3319 region, the linker will issue an error message.
3322 @section PHDRS Command
3324 @cindex program headers
3325 @cindex ELF program headers
3326 @cindex program segments
3327 @cindex segments, ELF
3328 The ELF object file format uses @dfn{program headers}, also knows as
3329 @dfn{segments}. The program headers describe how the program should be
3330 loaded into memory. You can print them out by using the @code{objdump}
3331 program with the @samp{-p} option.
3333 When you run an ELF program on a native ELF system, the system loader
3334 reads the program headers in order to figure out how to load the
3335 program. This will only work if the program headers are set correctly.
3336 This manual does not describe the details of how the system loader
3337 interprets program headers; for more information, see the ELF ABI.
3339 The linker will create reasonable program headers by default. However,
3340 in some cases, you may need to specify the program headers more
3341 precisely. You may use the @code{PHDRS} command for this purpose. When
3342 the linker sees the @code{PHDRS} command in the linker script, it will
3343 not create any program headers other than the ones specified.
3345 The linker only pays attention to the @code{PHDRS} command when
3346 generating an ELF output file. In other cases, the linker will simply
3347 ignore @code{PHDRS}.
3349 This is the syntax of the @code{PHDRS} command. The words @code{PHDRS},
3350 @code{FILEHDR}, @code{AT}, and @code{FLAGS} are keywords.
3356 @var{name} @var{type} [ FILEHDR ] [ PHDRS ] [ AT ( @var{address} ) ]
3357 [ FLAGS ( @var{flags} ) ] ;
3362 The @var{name} is used only for reference in the @code{SECTIONS} command
3363 of the linker script. It is not put into the output file. Program
3364 header names are stored in a separate name space, and will not conflict
3365 with symbol names, file names, or section names. Each program header
3366 must have a distinct name.
3368 Certain program header types describe segments of memory which the
3369 system loader will load from the file. In the linker script, you
3370 specify the contents of these segments by placing allocatable output
3371 sections in the segments. You use the @samp{:@var{phdr}} output section
3372 attribute to place a section in a particular segment. @xref{Output
3375 It is normal to put certain sections in more than one segment. This
3376 merely implies that one segment of memory contains another. You may
3377 repeat @samp{:@var{phdr}}, using it once for each segment which should
3378 contain the section.
3380 If you place a section in one or more segments using @samp{:@var{phdr}},
3381 then the linker will place all subsequent allocatable sections which do
3382 not specify @samp{:@var{phdr}} in the same segments. This is for
3383 convenience, since generally a whole set of contiguous sections will be
3384 placed in a single segment. You can use @code{:NONE} to override the
3385 default segment and tell the linker to not put the section in any
3390 You may use the @code{FILEHDR} and @code{PHDRS} keywords appear after
3391 the program header type to further describe the contents of the segment.
3392 The @code{FILEHDR} keyword means that the segment should include the ELF
3393 file header. The @code{PHDRS} keyword means that the segment should
3394 include the ELF program headers themselves.
3396 The @var{type} may be one of the following. The numbers indicate the
3397 value of the keyword.
3400 @item @code{PT_NULL} (0)
3401 Indicates an unused program header.
3403 @item @code{PT_LOAD} (1)
3404 Indicates that this program header describes a segment to be loaded from
3407 @item @code{PT_DYNAMIC} (2)
3408 Indicates a segment where dynamic linking information can be found.
3410 @item @code{PT_INTERP} (3)
3411 Indicates a segment where the name of the program interpreter may be
3414 @item @code{PT_NOTE} (4)
3415 Indicates a segment holding note information.
3417 @item @code{PT_SHLIB} (5)
3418 A reserved program header type, defined but not specified by the ELF
3421 @item @code{PT_PHDR} (6)
3422 Indicates a segment where the program headers may be found.
3424 @item @var{expression}
3425 An expression giving the numeric type of the program header. This may
3426 be used for types not defined above.
3429 You can specify that a segment should be loaded at a particular address
3430 in memory by using an @code{AT} expression. This is identical to the
3431 @code{AT} command used as an output section attribute (@pxref{Output
3432 Section LMA}). The @code{AT} command for a program header overrides the
3433 output section attribute.
3435 The linker will normally set the segment flags based on the sections
3436 which comprise the segment. You may use the @code{FLAGS} keyword to
3437 explicitly specify the segment flags. The value of @var{flags} must be
3438 an integer. It is used to set the @code{p_flags} field of the program
3441 Here is an example of @code{PHDRS}. This shows a typical set of program
3442 headers used on a native ELF system.
3448 headers PT_PHDR PHDRS ;
3450 text PT_LOAD FILEHDR PHDRS ;
3452 dynamic PT_DYNAMIC ;
3458 .interp : @{ *(.interp) @} :text :interp
3459 .text : @{ *(.text) @} :text
3460 .rodata : @{ *(.rodata) @} /* defaults to :text */
3462 . = . + 0x1000; /* move to a new page in memory */
3463 .data : @{ *(.data) @} :data
3464 .dynamic : @{ *(.dynamic) @} :data :dynamic
3471 @section VERSION Command
3472 @kindex VERSION @{script text@}
3473 @cindex symbol versions
3474 @cindex version script
3475 @cindex versions of symbols
3476 The linker supports symbol versions when using ELF. Symbol versions are
3477 only useful when using shared libraries. The dynamic linker can use
3478 symbol versions to select a specific version of a function when it runs
3479 a program that may have been linked against an earlier version of the
3482 You can include a version script directly in the main linker script, or
3483 you can supply the version script as an implicit linker script. You can
3484 also use the @samp{--version-script} linker option.
3486 The syntax of the @code{VERSION} command is simply
3488 VERSION @{ version-script-commands @}
3491 The format of the version script commands is identical to that used by
3492 Sun's linker in Solaris 2.5. The version script defines a tree of
3493 version nodes. You specify the node names and interdependencies in the
3494 version script. You can specify which symbols are bound to which
3495 version nodes, and you can reduce a specified set of symbols to local
3496 scope so that they are not globally visible outside of the shared
3499 The easiest way to demonstrate the version script language is with a few
3521 This example version script defines three version nodes. The first
3522 version node defined is @samp{VERS_1.1}; it has no other dependencies.
3523 The script binds the symbol @samp{foo1} to @samp{VERS_1.1}. It reduces
3524 a number of symbols to local scope so that they are not visible outside
3525 of the shared library.
3527 Next, the version script defines node @samp{VERS_1.2}. This node
3528 depends upon @samp{VERS_1.1}. The script binds the symbol @samp{foo2}
3529 to the version node @samp{VERS_1.2}.
3531 Finally, the version script defines node @samp{VERS_2.0}. This node
3532 depends upon @samp{VERS_1.2}. The scripts binds the symbols @samp{bar1}
3533 and @samp{bar2} are bound to the version node @samp{VERS_2.0}.
3535 When the linker finds a symbol defined in a library which is not
3536 specifically bound to a version node, it will effectively bind it to an
3537 unspecified base version of the library. You can bind all otherwise
3538 unspecified symbols to a given version node by using @samp{global: *}
3539 somewhere in the version script.
3541 The names of the version nodes have no specific meaning other than what
3542 they might suggest to the person reading them. The @samp{2.0} version
3543 could just as well have appeared in between @samp{1.1} and @samp{1.2}.
3544 However, this would be a confusing way to write a version script.
3546 When you link an application against a shared library that has versioned
3547 symbols, the application itself knows which version of each symbol it
3548 requires, and it also knows which version nodes it needs from each
3549 shared library it is linked against. Thus at runtime, the dynamic
3550 loader can make a quick check to make sure that the libraries you have
3551 linked against do in fact supply all of the version nodes that the
3552 application will need to resolve all of the dynamic symbols. In this
3553 way it is possible for the dynamic linker to know with certainty that
3554 all external symbols that it needs will be resolvable without having to
3555 search for each symbol reference.
3557 The symbol versioning is in effect a much more sophisticated way of
3558 doing minor version checking that SunOS does. The fundamental problem
3559 that is being addressed here is that typically references to external
3560 functions are bound on an as-needed basis, and are not all bound when
3561 the application starts up. If a shared library is out of date, a
3562 required interface may be missing; when the application tries to use
3563 that interface, it may suddenly and unexpectedly fail. With symbol
3564 versioning, the user will get a warning when they start their program if
3565 the libraries being used with the application are too old.
3567 There are several GNU extensions to Sun's versioning approach. The
3568 first of these is the ability to bind a symbol to a version node in the
3569 source file where the symbol is defined instead of in the versioning
3570 script. This was done mainly to reduce the burden on the library
3571 maintainer. You can do this by putting something like:
3573 __asm__(".symver original_foo,foo@@VERS_1.1");
3576 in the C source file. This renames the function @samp{original_foo} to
3577 be an alias for @samp{foo} bound to the version node @samp{VERS_1.1}.
3578 The @samp{local:} directive can be used to prevent the symbol
3579 @samp{original_foo} from being exported.
3581 The second GNU extension is to allow multiple versions of the same
3582 function to appear in a given shared library. In this way you can make
3583 an incompatible change to an interface without increasing the major
3584 version number of the shared library, while still allowing applications
3585 linked against the old interface to continue to function.
3587 To do this, you must use multiple @samp{.symver} directives in the
3588 source file. Here is an example:
3591 __asm__(".symver original_foo,foo@@");
3592 __asm__(".symver old_foo,foo@@VERS_1.1");
3593 __asm__(".symver old_foo1,foo@@VERS_1.2");
3594 __asm__(".symver new_foo,foo@@@@VERS_2.0");
3597 In this example, @samp{foo@@} represents the symbol @samp{foo} bound to the
3598 unspecified base version of the symbol. The source file that contains this
3599 example would define 4 C functions: @samp{original_foo}, @samp{old_foo},
3600 @samp{old_foo1}, and @samp{new_foo}.
3602 When you have multiple definitions of a given symbol, there needs to be
3603 some way to specify a default version to which external references to
3604 this symbol will be bound. You can do this with the
3605 @samp{foo@@@@VERS_2.0} type of @samp{.symver} directive. You can only
3606 declare one version of a symbol as the default in this manner; otherwise
3607 you would effectively have multiple definitions of the same symbol.
3609 If you wish to bind a reference to a specific version of the symbol
3610 within the shared library, you can use the aliases of convenience
3611 (i.e. @samp{old_foo}), or you can use the @samp{.symver} directive to
3612 specifically bind to an external version of the function in question.
3614 You can also specify the language in the version script:
3617 VERSION extern "lang" @{ version-script-commands @}
3620 The supported @samp{lang}s are @samp{C}, @samp{C++}, and @samp{Java}.
3621 The linker will iterate over the list of symbols at the link time and
3622 demangle them according to @samp{lang} before matching them to the
3623 patterns specified in @samp{version-script-commands}.
3626 @section Expressions in Linker Scripts
3629 The syntax for expressions in the linker script language is identical to
3630 that of C expressions. All expressions are evaluated as integers. All
3631 expressions are evaluated in the same size, which is 32 bits if both the
3632 host and target are 32 bits, and is otherwise 64 bits.
3634 You can use and set symbol values in expressions.
3636 The linker defines several special purpose builtin functions for use in
3640 * Constants:: Constants
3641 * Symbols:: Symbol Names
3642 * Location Counter:: The Location Counter
3643 * Operators:: Operators
3644 * Evaluation:: Evaluation
3645 * Expression Section:: The Section of an Expression
3646 * Builtin Functions:: Builtin Functions
3650 @subsection Constants
3651 @cindex integer notation
3652 @cindex constants in linker scripts
3653 All constants are integers.
3655 As in C, the linker considers an integer beginning with @samp{0} to be
3656 octal, and an integer beginning with @samp{0x} or @samp{0X} to be
3657 hexadecimal. The linker considers other integers to be decimal.
3659 @cindex scaled integers
3660 @cindex K and M integer suffixes
3661 @cindex M and K integer suffixes
3662 @cindex suffixes for integers
3663 @cindex integer suffixes
3664 In addition, you can use the suffixes @code{K} and @code{M} to scale a
3668 @c END TEXI2ROFF-KILL
3669 @code{1024} or @code{1024*1024}
3673 ${\rm 1024}$ or ${\rm 1024}^2$
3675 @c END TEXI2ROFF-KILL
3676 respectively. For example, the following all refer to the same quantity:
3684 @subsection Symbol Names
3685 @cindex symbol names
3687 @cindex quoted symbol names
3689 Unless quoted, symbol names start with a letter, underscore, or period
3690 and may include letters, digits, underscores, periods, and hyphens.
3691 Unquoted symbol names must not conflict with any keywords. You can
3692 specify a symbol which contains odd characters or has the same name as a
3693 keyword by surrounding the symbol name in double quotes:
3696 "with a space" = "also with a space" + 10;
3699 Since symbols can contain many non-alphabetic characters, it is safest
3700 to delimit symbols with spaces. For example, @samp{A-B} is one symbol,
3701 whereas @samp{A - B} is an expression involving subtraction.
3703 @node Location Counter
3704 @subsection The Location Counter
3707 @cindex location counter
3708 @cindex current output location
3709 The special linker variable @dfn{dot} @samp{.} always contains the
3710 current output location counter. Since the @code{.} always refers to a
3711 location in an output section, it may only appear in an expression
3712 within a @code{SECTIONS} command. The @code{.} symbol may appear
3713 anywhere that an ordinary symbol is allowed in an expression.
3716 Assigning a value to @code{.} will cause the location counter to be
3717 moved. This may be used to create holes in the output section. The
3718 location counter may never be moved backwards.
3734 In the previous example, the @samp{.text} section from @file{file1} is
3735 located at the beginning of the output section @samp{output}. It is
3736 followed by a 1000 byte gap. Then the @samp{.text} section from
3737 @file{file2} appears, also with a 1000 byte gap following before the
3738 @samp{.text} section from @file{file3}. The notation @samp{= 0x1234}
3739 specifies what data to write in the gaps (@pxref{Output Section Fill}).
3741 @cindex dot inside sections
3742 Note: @code{.} actually refers to the byte offset from the start of the
3743 current containing object. Normally this is the @code{SECTIONS}
3744 statement, whoes start address is 0, hence @code{.} can be used as an
3745 absolute address. If @code{.} is used inside a section description
3746 however, it refers to the byte offset from the start of that section,
3747 not an absolute address. Thus in a script like this:
3765 The @samp{.text} section will be assigned a starting address of 0x100
3766 and a size of exactly 0x200 bytes, even if there is not enough data in
3767 the @samp{.text} input sections to fill this area. (If there is too
3768 much data, an error will be produced because this would be an attempt to
3769 move @code{.} backwards). The @samp{.data} section will start at 0x500
3770 and it will have an extra 0x600 bytes worth of space after the end of
3771 the values from the @samp{.data} input sections and before the end of
3772 the @samp{.data} output section itself.
3776 @subsection Operators
3777 @cindex operators for arithmetic
3778 @cindex arithmetic operators
3779 @cindex precedence in expressions
3780 The linker recognizes the standard C set of arithmetic operators, with
3781 the standard bindings and precedence levels:
3784 @c END TEXI2ROFF-KILL
3786 precedence associativity Operators Notes
3792 5 left == != > < <= >=
3798 11 right &= += -= *= /= (2)
3802 (1) Prefix operators
3803 (2) @xref{Assignments}.
3807 \vskip \baselineskip
3808 %"lispnarrowing" is the extra indent used generally for smallexample
3809 \hskip\lispnarrowing\vbox{\offinterlineskip
3812 {\vrule#&\strut\hfil\ #\ \hfil&\vrule#&\strut\hfil\ #\ \hfil&\vrule#&\strut\hfil\ {\tt #}\ \hfil&\vrule#\cr
3813 height2pt&\omit&&\omit&&\omit&\cr
3814 &Precedence&& Associativity &&{\rm Operators}&\cr
3815 height2pt&\omit&&\omit&&\omit&\cr
3817 height2pt&\omit&&\omit&&\omit&\cr
3819 % '176 is tilde, '~' in tt font
3820 &1&&left&&\qquad- \char'176\ !\qquad\dag&\cr
3821 &2&&left&&* / \%&\cr
3824 &5&&left&&== != > < <= >=&\cr
3827 &8&&left&&{\&\&}&\cr
3830 &11&&right&&\qquad\&= += -= *= /=\qquad\ddag&\cr
3832 height2pt&\omit&&\omit&&\omit&\cr}
3837 @obeylines@parskip=0pt@parindent=0pt
3838 @dag@quad Prefix operators.
3839 @ddag@quad @xref{Assignments}.
3842 @c END TEXI2ROFF-KILL
3845 @subsection Evaluation
3846 @cindex lazy evaluation
3847 @cindex expression evaluation order
3848 The linker evaluates expressions lazily. It only computes the value of
3849 an expression when absolutely necessary.
3851 The linker needs some information, such as the value of the start
3852 address of the first section, and the origins and lengths of memory
3853 regions, in order to do any linking at all. These values are computed
3854 as soon as possible when the linker reads in the linker script.
3856 However, other values (such as symbol values) are not known or needed
3857 until after storage allocation. Such values are evaluated later, when
3858 other information (such as the sizes of output sections) is available
3859 for use in the symbol assignment expression.
3861 The sizes of sections cannot be known until after allocation, so
3862 assignments dependent upon these are not performed until after
3865 Some expressions, such as those depending upon the location counter
3866 @samp{.}, must be evaluated during section allocation.
3868 If the result of an expression is required, but the value is not
3869 available, then an error results. For example, a script like the
3875 .text 9+this_isnt_constant :
3881 will cause the error message @samp{non constant expression for initial
3884 @node Expression Section
3885 @subsection The Section of an Expression
3886 @cindex expression sections
3887 @cindex absolute expressions
3888 @cindex relative expressions
3889 @cindex absolute and relocatable symbols
3890 @cindex relocatable and absolute symbols
3891 @cindex symbols, relocatable and absolute
3892 When the linker evaluates an expression, the result is either absolute
3893 or relative to some section. A relative expression is expressed as a
3894 fixed offset from the base of a section.
3896 The position of the expression within the linker script determines
3897 whether it is absolute or relative. An expression which appears within
3898 an output section definition is relative to the base of the output
3899 section. An expression which appears elsewhere will be absolute.
3901 A symbol set to a relative expression will be relocatable if you request
3902 relocatable output using the @samp{-r} option. That means that a
3903 further link operation may change the value of the symbol. The symbol's
3904 section will be the section of the relative expression.
3906 A symbol set to an absolute expression will retain the same value
3907 through any further link operation. The symbol will be absolute, and
3908 will not have any particular associated section.
3910 You can use the builtin function @code{ABSOLUTE} to force an expression
3911 to be absolute when it would otherwise be relative. For example, to
3912 create an absolute symbol set to the address of the end of the output
3913 section @samp{.data}:
3917 .data : @{ *(.data) _edata = ABSOLUTE(.); @}
3921 If @samp{ABSOLUTE} were not used, @samp{_edata} would be relative to the
3922 @samp{.data} section.
3924 @node Builtin Functions
3925 @subsection Builtin Functions
3926 @cindex functions in expressions
3927 The linker script language includes a number of builtin functions for
3928 use in linker script expressions.
3931 @item ABSOLUTE(@var{exp})
3932 @kindex ABSOLUTE(@var{exp})
3933 @cindex expression, absolute
3934 Return the absolute (non-relocatable, as opposed to non-negative) value
3935 of the expression @var{exp}. Primarily useful to assign an absolute
3936 value to a symbol within a section definition, where symbol values are
3937 normally section relative. @xref{Expression Section}.
3939 @item ADDR(@var{section})
3940 @kindex ADDR(@var{section})
3941 @cindex section address in expression
3942 Return the absolute address (the VMA) of the named @var{section}. Your
3943 script must previously have defined the location of that section. In
3944 the following example, @code{symbol_1} and @code{symbol_2} are assigned
3951 start_of_output_1 = ABSOLUTE(.);
3956 symbol_1 = ADDR(.output1);
3957 symbol_2 = start_of_output_1;
3963 @item ALIGN(@var{exp})
3964 @kindex ALIGN(@var{exp})
3965 @cindex round up location counter
3966 @cindex align location counter
3967 Return the location counter (@code{.}) aligned to the next @var{exp}
3968 boundary. @var{exp} must be an expression whose value is a power of
3969 two. This is equivalent to
3971 (. + @var{exp} - 1) & ~(@var{exp} - 1)
3974 @code{ALIGN} doesn't change the value of the location counter---it just
3975 does arithmetic on it. Here is an example which aligns the output
3976 @code{.data} section to the next @code{0x2000} byte boundary after the
3977 preceding section and sets a variable within the section to the next
3978 @code{0x8000} boundary after the input sections:
3982 .data ALIGN(0x2000): @{
3984 variable = ALIGN(0x8000);
3990 The first use of @code{ALIGN} in this example specifies the location of
3991 a section because it is used as the optional @var{address} attribute of
3992 a section definition (@pxref{Output Section Address}). The second use
3993 of @code{ALIGN} is used to defines the value of a symbol.
3995 The builtin function @code{NEXT} is closely related to @code{ALIGN}.
3997 @item BLOCK(@var{exp})
3998 @kindex BLOCK(@var{exp})
3999 This is a synonym for @code{ALIGN}, for compatibility with older linker
4000 scripts. It is most often seen when setting the address of an output
4003 @item DEFINED(@var{symbol})
4004 @kindex DEFINED(@var{symbol})
4005 @cindex symbol defaults
4006 Return 1 if @var{symbol} is in the linker global symbol table and is
4007 defined, otherwise return 0. You can use this function to provide
4008 default values for symbols. For example, the following script fragment
4009 shows how to set a global symbol @samp{begin} to the first location in
4010 the @samp{.text} section---but if a symbol called @samp{begin} already
4011 existed, its value is preserved:
4017 begin = DEFINED(begin) ? begin : . ;
4025 @item LOADADDR(@var{section})
4026 @kindex LOADADDR(@var{section})
4027 @cindex section load address in expression
4028 Return the absolute LMA of the named @var{section}. This is normally
4029 the same as @code{ADDR}, but it may be different if the @code{AT}
4030 attribute is used in the output section definition (@pxref{Output
4034 @item MAX(@var{exp1}, @var{exp2})
4035 Returns the maximum of @var{exp1} and @var{exp2}.
4038 @item MIN(@var{exp1}, @var{exp2})
4039 Returns the minimum of @var{exp1} and @var{exp2}.
4041 @item NEXT(@var{exp})
4042 @kindex NEXT(@var{exp})
4043 @cindex unallocated address, next
4044 Return the next unallocated address that is a multiple of @var{exp}.
4045 This function is closely related to @code{ALIGN(@var{exp})}; unless you
4046 use the @code{MEMORY} command to define discontinuous memory for the
4047 output file, the two functions are equivalent.
4049 @item SIZEOF(@var{section})
4050 @kindex SIZEOF(@var{section})
4051 @cindex section size
4052 Return the size in bytes of the named @var{section}, if that section has
4053 been allocated. If the section has not been allocated when this is
4054 evaluated, the linker will report an error. In the following example,
4055 @code{symbol_1} and @code{symbol_2} are assigned identical values:
4064 symbol_1 = .end - .start ;
4065 symbol_2 = SIZEOF(.output);
4070 @item SIZEOF_HEADERS
4071 @itemx sizeof_headers
4072 @kindex SIZEOF_HEADERS
4074 Return the size in bytes of the output file's headers. This is
4075 information which appears at the start of the output file. You can use
4076 this number when setting the start address of the first section, if you
4077 choose, to facilitate paging.
4079 @cindex not enough room for program headers
4080 @cindex program headers, not enough room
4081 When producing an ELF output file, if the linker script uses the
4082 @code{SIZEOF_HEADERS} builtin function, the linker must compute the
4083 number of program headers before it has determined all the section
4084 addresses and sizes. If the linker later discovers that it needs
4085 additional program headers, it will report an error @samp{not enough
4086 room for program headers}. To avoid this error, you must avoid using
4087 the @code{SIZEOF_HEADERS} function, or you must rework your linker
4088 script to avoid forcing the linker to use additional program headers, or
4089 you must define the program headers yourself using the @code{PHDRS}
4090 command (@pxref{PHDRS}).
4093 @node Implicit Linker Scripts
4094 @section Implicit Linker Scripts
4095 @cindex implicit linker scripts
4096 If you specify a linker input file which the linker can not recognize as
4097 an object file or an archive file, it will try to read the file as a
4098 linker script. If the file can not be parsed as a linker script, the
4099 linker will report an error.
4101 An implicit linker script will not replace the default linker script.
4103 Typically an implicit linker script would contain only symbol
4104 assignments, or the @code{INPUT}, @code{GROUP}, or @code{VERSION}
4107 Any input files read because of an implicit linker script will be read
4108 at the position in the command line where the implicit linker script was
4109 read. This can affect archive searching.
4112 @node Machine Dependent
4113 @chapter Machine Dependent Features
4115 @cindex machine dependencies
4116 @code{ld} has additional features on some platforms; the following
4117 sections describe them. Machines where @code{ld} has no additional
4118 functionality are not listed.
4121 * H8/300:: @code{ld} and the H8/300
4122 * i960:: @code{ld} and the Intel 960 family
4123 * ARM:: @code{ld} and the ARM family
4124 * HPPA ELF32:: @code{ld} and HPPA 32-bit ELF
4126 * TI COFF:: @code{ld} and TI COFF
4131 @c FIXME! This could use @raisesections/@lowersections, but there seems to be a conflict
4132 @c between those and node-defaulting.
4139 @section @code{ld} and the H8/300
4141 @cindex H8/300 support
4142 For the H8/300, @code{ld} can perform these global optimizations when
4143 you specify the @samp{--relax} command-line option.
4146 @cindex relaxing on H8/300
4147 @item relaxing address modes
4148 @code{ld} finds all @code{jsr} and @code{jmp} instructions whose
4149 targets are within eight bits, and turns them into eight-bit
4150 program-counter relative @code{bsr} and @code{bra} instructions,
4153 @cindex synthesizing on H8/300
4154 @item synthesizing instructions
4155 @c FIXME: specifically mov.b, or any mov instructions really?
4156 @code{ld} finds all @code{mov.b} instructions which use the
4157 sixteen-bit absolute address form, but refer to the top
4158 page of memory, and changes them to use the eight-bit address form.
4159 (That is: the linker turns @samp{mov.b @code{@@}@var{aa}:16} into
4160 @samp{mov.b @code{@@}@var{aa}:8} whenever the address @var{aa} is in the
4161 top page of memory).
4171 @c This stuff is pointless to say unless you're especially concerned
4172 @c with Hitachi chips; don't enable it for generic case, please.
4174 @chapter @code{ld} and other Hitachi chips
4176 @code{ld} also supports the H8/300H, the H8/500, and the Hitachi SH. No
4177 special features, commands, or command-line options are required for
4188 @section @code{ld} and the Intel 960 family
4190 @cindex i960 support
4192 You can use the @samp{-A@var{architecture}} command line option to
4193 specify one of the two-letter names identifying members of the 960
4194 family; the option specifies the desired output target, and warns of any
4195 incompatible instructions in the input files. It also modifies the
4196 linker's search strategy for archive libraries, to support the use of
4197 libraries specific to each particular architecture, by including in the
4198 search loop names suffixed with the string identifying the architecture.
4200 For example, if your @code{ld} command line included @w{@samp{-ACA}} as
4201 well as @w{@samp{-ltry}}, the linker would look (in its built-in search
4202 paths, and in any paths you specify with @samp{-L}) for a library with
4215 The first two possibilities would be considered in any event; the last
4216 two are due to the use of @w{@samp{-ACA}}.
4218 You can meaningfully use @samp{-A} more than once on a command line, since
4219 the 960 architecture family allows combination of target architectures; each
4220 use will add another pair of name variants to search for when @w{@samp{-l}}
4221 specifies a library.
4223 @cindex @code{--relax} on i960
4224 @cindex relaxing on i960
4225 @code{ld} supports the @samp{--relax} option for the i960 family. If
4226 you specify @samp{--relax}, @code{ld} finds all @code{balx} and
4227 @code{calx} instructions whose targets are within 24 bits, and turns
4228 them into 24-bit program-counter relative @code{bal} and @code{cal}
4229 instructions, respectively. @code{ld} also turns @code{cal}
4230 instructions into @code{bal} instructions when it determines that the
4231 target subroutine is a leaf routine (that is, the target subroutine does
4232 not itself call any subroutines).
4244 @section @code{ld}'s support for interworking between ARM and Thumb code
4246 @cindex ARM interworking support
4247 @kindex --support-old-code
4248 For the ARM, @code{ld} will generate code stubs to allow functions calls
4249 betweem ARM and Thumb code. These stubs only work with code that has
4250 been compiled and assembled with the @samp{-mthumb-interwork} command
4251 line option. If it is necessary to link with old ARM object files or
4252 libraries, which have not been compiled with the -mthumb-interwork
4253 option then the @samp{--support-old-code} command line switch should be
4254 given to the linker. This will make it generate larger stub functions
4255 which will work with non-interworking aware ARM code. Note, however,
4256 the linker does not support generating stubs for function calls to
4257 non-interworking aware Thumb code.
4259 @cindex thumb entry point
4260 @cindex entry point, thumb
4261 @kindex --thumb-entry=@var{entry}
4262 The @samp{--thumb-entry} switch is a duplicate of the generic
4263 @samp{--entry} switch, in that it sets the program's starting address.
4264 But it also sets the bottom bit of the address, so that it can be
4265 branched to using a BX instruction, and the program will start
4266 executing in Thumb mode straight away.
4269 @section @code{ld} and HPPA 32-bit ELF support
4270 @cindex HPPA multiple sub-space stubs
4271 @kindex --multi-subspace
4272 When generating a shared library, @code{ld} will by default generate
4273 import stubs suitable for use with a single sub-space application.
4274 The @samp{--multi-subspace} switch causes @code{ld} to generate export
4275 stubs, and different (larger) import stubs suitable for use with
4276 multiple sub-spaces.
4278 @cindex HPPA stub grouping
4279 @kindex --stub-group-size=@var{N}
4280 Long branch stubs and import/export stubs are placed by @code{ld} in
4281 stub sections located between groups of input sections.
4282 @samp{--stub-group-size} specifies the maximum size of a group of input
4283 sections handled by one stub section. Since branch offsets are signed,
4284 a stub section may serve two groups of input sections, one group before
4285 the stub section, and one group after it. However, when using
4286 conditional branches that require stubs, it may be better (for branch
4287 prediction) that stub sections only serve one group of input sections.
4288 A negative value for @samp{N} chooses this scheme, ensuring that
4289 branches to stubs always use a negative offset. Two special values of
4290 @samp{N} are recognized, @samp{1} and @samp{-1}. These both instruct
4291 @code{ld} to automatically size input section groups for the branch types
4292 detected, with the same behaviour regarding stub placement as other
4293 positive or negative values of @samp{N} respectively.
4295 Note that @samp{--stub-group-size} does not split input sections. A
4296 single input section larger than the group size specified will of course
4297 create a larger group (of one section). If input sections are too
4298 large, it may not be possible for a branch to reach its stub.
4302 @section @code{ld}'s support for various TI COFF versions
4303 @cindex TI COFF versions
4304 @kindex --format=@var{version}
4305 The @samp{--format} switch allows selection of one of the various
4306 TI COFF versions. The latest of this writing is 2; versions 0 and 1 are
4307 also supported. The TI COFF versions also vary in header byte-order
4308 format; @code{ld} will read any version or byte order, but the output
4309 header format depends on the default specified by the specific target.
4316 @ifclear SingleFormat
4321 @cindex object file management
4322 @cindex object formats available
4324 The linker accesses object and archive files using the BFD libraries.
4325 These libraries allow the linker to use the same routines to operate on
4326 object files whatever the object file format. A different object file
4327 format can be supported simply by creating a new BFD back end and adding
4328 it to the library. To conserve runtime memory, however, the linker and
4329 associated tools are usually configured to support only a subset of the
4330 object file formats available. You can use @code{objdump -i}
4331 (@pxref{objdump,,objdump,binutils.info,The GNU Binary Utilities}) to
4332 list all the formats available for your configuration.
4334 @cindex BFD requirements
4335 @cindex requirements for BFD
4336 As with most implementations, BFD is a compromise between
4337 several conflicting requirements. The major factor influencing
4338 BFD design was efficiency: any time used converting between
4339 formats is time which would not have been spent had BFD not
4340 been involved. This is partly offset by abstraction payback; since
4341 BFD simplifies applications and back ends, more time and care
4342 may be spent optimizing algorithms for a greater speed.
4344 One minor artifact of the BFD solution which you should bear in
4345 mind is the potential for information loss. There are two places where
4346 useful information can be lost using the BFD mechanism: during
4347 conversion and during output. @xref{BFD information loss}.
4350 * BFD outline:: How it works: an outline of BFD
4354 @section How it works: an outline of BFD
4355 @cindex opening object files
4356 @include bfdsumm.texi
4359 @node Reporting Bugs
4360 @chapter Reporting Bugs
4361 @cindex bugs in @code{ld}
4362 @cindex reporting bugs in @code{ld}
4364 Your bug reports play an essential role in making @code{ld} reliable.
4366 Reporting a bug may help you by bringing a solution to your problem, or
4367 it may not. But in any case the principal function of a bug report is
4368 to help the entire community by making the next version of @code{ld}
4369 work better. Bug reports are your contribution to the maintenance of
4372 In order for a bug report to serve its purpose, you must include the
4373 information that enables us to fix the bug.
4376 * Bug Criteria:: Have you found a bug?
4377 * Bug Reporting:: How to report bugs
4381 @section Have you found a bug?
4382 @cindex bug criteria
4384 If you are not sure whether you have found a bug, here are some guidelines:
4387 @cindex fatal signal
4388 @cindex linker crash
4389 @cindex crash of linker
4391 If the linker gets a fatal signal, for any input whatever, that is a
4392 @code{ld} bug. Reliable linkers never crash.
4394 @cindex error on valid input
4396 If @code{ld} produces an error message for valid input, that is a bug.
4398 @cindex invalid input
4400 If @code{ld} does not produce an error message for invalid input, that
4401 may be a bug. In the general case, the linker can not verify that
4402 object files are correct.
4405 If you are an experienced user of linkers, your suggestions for
4406 improvement of @code{ld} are welcome in any case.
4410 @section How to report bugs
4412 @cindex @code{ld} bugs, reporting
4414 A number of companies and individuals offer support for @sc{gnu}
4415 products. If you obtained @code{ld} from a support organization, we
4416 recommend you contact that organization first.
4418 You can find contact information for many support companies and
4419 individuals in the file @file{etc/SERVICE} in the @sc{gnu} Emacs
4422 Otherwise, send bug reports for @code{ld} to
4423 @samp{bug-binutils@@gnu.org}.
4425 The fundamental principle of reporting bugs usefully is this:
4426 @strong{report all the facts}. If you are not sure whether to state a
4427 fact or leave it out, state it!
4429 Often people omit facts because they think they know what causes the
4430 problem and assume that some details do not matter. Thus, you might
4431 assume that the name of a symbol you use in an example does not matter.
4432 Well, probably it does not, but one cannot be sure. Perhaps the bug is
4433 a stray memory reference which happens to fetch from the location where
4434 that name is stored in memory; perhaps, if the name were different, the
4435 contents of that location would fool the linker into doing the right
4436 thing despite the bug. Play it safe and give a specific, complete
4437 example. That is the easiest thing for you to do, and the most helpful.
4439 Keep in mind that the purpose of a bug report is to enable us to fix the bug if
4440 it is new to us. Therefore, always write your bug reports on the assumption
4441 that the bug has not been reported previously.
4443 Sometimes people give a few sketchy facts and ask, ``Does this ring a
4444 bell?'' Those bug reports are useless, and we urge everyone to
4445 @emph{refuse to respond to them} except to chide the sender to report
4448 To enable us to fix the bug, you should include all these things:
4452 The version of @code{ld}. @code{ld} announces it if you start it with
4453 the @samp{--version} argument.
4455 Without this, we will not know whether there is any point in looking for
4456 the bug in the current version of @code{ld}.
4459 Any patches you may have applied to the @code{ld} source, including any
4460 patches made to the @code{BFD} library.
4463 The type of machine you are using, and the operating system name and
4467 What compiler (and its version) was used to compile @code{ld}---e.g.
4471 The command arguments you gave the linker to link your example and
4472 observe the bug. To guarantee you will not omit something important,
4473 list them all. A copy of the Makefile (or the output from make) is
4476 If we were to try to guess the arguments, we would probably guess wrong
4477 and then we might not encounter the bug.
4480 A complete input file, or set of input files, that will reproduce the
4481 bug. It is generally most helpful to send the actual object files,
4482 uuencoded if necessary to get them through the mail system. Making them
4483 available for anonymous FTP is not as good, but may be the only
4484 reasonable choice for large object files.
4486 If the source files were assembled using @code{gas} or compiled using
4487 @code{gcc}, then it may be OK to send the source files rather than the
4488 object files. In this case, be sure to say exactly what version of
4489 @code{gas} or @code{gcc} was used to produce the object files. Also say
4490 how @code{gas} or @code{gcc} were configured.
4493 A description of what behavior you observe that you believe is
4494 incorrect. For example, ``It gets a fatal signal.''
4496 Of course, if the bug is that @code{ld} gets a fatal signal, then we
4497 will certainly notice it. But if the bug is incorrect output, we might
4498 not notice unless it is glaringly wrong. You might as well not give us
4499 a chance to make a mistake.
4501 Even if the problem you experience is a fatal signal, you should still
4502 say so explicitly. Suppose something strange is going on, such as, your
4503 copy of @code{ld} is out of synch, or you have encountered a bug in the
4504 C library on your system. (This has happened!) Your copy might crash
4505 and ours would not. If you told us to expect a crash, then when ours
4506 fails to crash, we would know that the bug was not happening for us. If
4507 you had not told us to expect a crash, then we would not be able to draw
4508 any conclusion from our observations.
4511 If you wish to suggest changes to the @code{ld} source, send us context
4512 diffs, as generated by @code{diff} with the @samp{-u}, @samp{-c}, or
4513 @samp{-p} option. Always send diffs from the old file to the new file.
4514 If you even discuss something in the @code{ld} source, refer to it by
4515 context, not by line number.
4517 The line numbers in our development sources will not match those in your
4518 sources. Your line numbers would convey no useful information to us.
4521 Here are some things that are not necessary:
4525 A description of the envelope of the bug.
4527 Often people who encounter a bug spend a lot of time investigating
4528 which changes to the input file will make the bug go away and which
4529 changes will not affect it.
4531 This is often time consuming and not very useful, because the way we
4532 will find the bug is by running a single example under the debugger
4533 with breakpoints, not by pure deduction from a series of examples.
4534 We recommend that you save your time for something else.
4536 Of course, if you can find a simpler example to report @emph{instead}
4537 of the original one, that is a convenience for us. Errors in the
4538 output will be easier to spot, running under the debugger will take
4539 less time, and so on.
4541 However, simplification is not vital; if you do not want to do this,
4542 report the bug anyway and send us the entire test case you used.
4545 A patch for the bug.
4547 A patch for the bug does help us if it is a good one. But do not omit
4548 the necessary information, such as the test case, on the assumption that
4549 a patch is all we need. We might see problems with your patch and decide
4550 to fix the problem another way, or we might not understand it at all.
4552 Sometimes with a program as complicated as @code{ld} it is very hard to
4553 construct an example that will make the program follow a certain path
4554 through the code. If you do not send us the example, we will not be
4555 able to construct one, so we will not be able to verify that the bug is
4558 And if we cannot understand what bug you are trying to fix, or why your
4559 patch should be an improvement, we will not install it. A test case will
4560 help us to understand.
4563 A guess about what the bug is or what it depends on.
4565 Such guesses are usually wrong. Even we cannot guess right about such
4566 things without first using the debugger to find the facts.
4570 @appendix MRI Compatible Script Files
4571 @cindex MRI compatibility
4572 To aid users making the transition to @sc{gnu} @code{ld} from the MRI
4573 linker, @code{ld} can use MRI compatible linker scripts as an
4574 alternative to the more general-purpose linker scripting language
4575 described in @ref{Scripts}. MRI compatible linker scripts have a much
4576 simpler command set than the scripting language otherwise used with
4577 @code{ld}. @sc{gnu} @code{ld} supports the most commonly used MRI
4578 linker commands; these commands are described here.
4580 In general, MRI scripts aren't of much use with the @code{a.out} object
4581 file format, since it only has three sections and MRI scripts lack some
4582 features to make use of them.
4584 You can specify a file containing an MRI-compatible script using the
4585 @samp{-c} command-line option.
4587 Each command in an MRI-compatible script occupies its own line; each
4588 command line starts with the keyword that identifies the command (though
4589 blank lines are also allowed for punctuation). If a line of an
4590 MRI-compatible script begins with an unrecognized keyword, @code{ld}
4591 issues a warning message, but continues processing the script.
4593 Lines beginning with @samp{*} are comments.
4595 You can write these commands using all upper-case letters, or all
4596 lower case; for example, @samp{chip} is the same as @samp{CHIP}.
4597 The following list shows only the upper-case form of each command.
4600 @cindex @code{ABSOLUTE} (MRI)
4601 @item ABSOLUTE @var{secname}
4602 @itemx ABSOLUTE @var{secname}, @var{secname}, @dots{} @var{secname}
4603 Normally, @code{ld} includes in the output file all sections from all
4604 the input files. However, in an MRI-compatible script, you can use the
4605 @code{ABSOLUTE} command to restrict the sections that will be present in
4606 your output program. If the @code{ABSOLUTE} command is used at all in a
4607 script, then only the sections named explicitly in @code{ABSOLUTE}
4608 commands will appear in the linker output. You can still use other
4609 input sections (whatever you select on the command line, or using
4610 @code{LOAD}) to resolve addresses in the output file.
4612 @cindex @code{ALIAS} (MRI)
4613 @item ALIAS @var{out-secname}, @var{in-secname}
4614 Use this command to place the data from input section @var{in-secname}
4615 in a section called @var{out-secname} in the linker output file.
4617 @var{in-secname} may be an integer.
4619 @cindex @code{ALIGN} (MRI)
4620 @item ALIGN @var{secname} = @var{expression}
4621 Align the section called @var{secname} to @var{expression}. The
4622 @var{expression} should be a power of two.
4624 @cindex @code{BASE} (MRI)
4625 @item BASE @var{expression}
4626 Use the value of @var{expression} as the lowest address (other than
4627 absolute addresses) in the output file.
4629 @cindex @code{CHIP} (MRI)
4630 @item CHIP @var{expression}
4631 @itemx CHIP @var{expression}, @var{expression}
4632 This command does nothing; it is accepted only for compatibility.
4634 @cindex @code{END} (MRI)
4636 This command does nothing whatever; it's only accepted for compatibility.
4638 @cindex @code{FORMAT} (MRI)
4639 @item FORMAT @var{output-format}
4640 Similar to the @code{OUTPUT_FORMAT} command in the more general linker
4641 language, but restricted to one of these output formats:
4645 S-records, if @var{output-format} is @samp{S}
4648 IEEE, if @var{output-format} is @samp{IEEE}
4651 COFF (the @samp{coff-m68k} variant in BFD), if @var{output-format} is
4655 @cindex @code{LIST} (MRI)
4656 @item LIST @var{anything}@dots{}
4657 Print (to the standard output file) a link map, as produced by the
4658 @code{ld} command-line option @samp{-M}.
4660 The keyword @code{LIST} may be followed by anything on the
4661 same line, with no change in its effect.
4663 @cindex @code{LOAD} (MRI)
4664 @item LOAD @var{filename}
4665 @itemx LOAD @var{filename}, @var{filename}, @dots{} @var{filename}
4666 Include one or more object file @var{filename} in the link; this has the
4667 same effect as specifying @var{filename} directly on the @code{ld}
4670 @cindex @code{NAME} (MRI)
4671 @item NAME @var{output-name}
4672 @var{output-name} is the name for the program produced by @code{ld}; the
4673 MRI-compatible command @code{NAME} is equivalent to the command-line
4674 option @samp{-o} or the general script language command @code{OUTPUT}.
4676 @cindex @code{ORDER} (MRI)
4677 @item ORDER @var{secname}, @var{secname}, @dots{} @var{secname}
4678 @itemx ORDER @var{secname} @var{secname} @var{secname}
4679 Normally, @code{ld} orders the sections in its output file in the
4680 order in which they first appear in the input files. In an MRI-compatible
4681 script, you can override this ordering with the @code{ORDER} command. The
4682 sections you list with @code{ORDER} will appear first in your output
4683 file, in the order specified.
4685 @cindex @code{PUBLIC} (MRI)
4686 @item PUBLIC @var{name}=@var{expression}
4687 @itemx PUBLIC @var{name},@var{expression}
4688 @itemx PUBLIC @var{name} @var{expression}
4689 Supply a value (@var{expression}) for external symbol
4690 @var{name} used in the linker input files.
4692 @cindex @code{SECT} (MRI)
4693 @item SECT @var{secname}, @var{expression}
4694 @itemx SECT @var{secname}=@var{expression}
4695 @itemx SECT @var{secname} @var{expression}
4696 You can use any of these three forms of the @code{SECT} command to
4697 specify the start address (@var{expression}) for section @var{secname}.
4698 If you have more than one @code{SECT} statement for the same
4699 @var{secname}, only the @emph{first} sets the start address.
4702 @node GNU Free Documentation License
4703 @appendix GNU Free Documentation License
4704 @cindex GNU Free Documentation License
4706 GNU Free Documentation License
4708 Version 1.1, March 2000
4710 Copyright (C) 2000 Free Software Foundation, Inc.
4711 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
4713 Everyone is permitted to copy and distribute verbatim copies
4714 of this license document, but changing it is not allowed.
4719 The purpose of this License is to make a manual, textbook, or other
4720 written document "free" in the sense of freedom: to assure everyone
4721 the effective freedom to copy and redistribute it, with or without
4722 modifying it, either commercially or noncommercially. Secondarily,
4723 this License preserves for the author and publisher a way to get
4724 credit for their work, while not being considered responsible for
4725 modifications made by others.
4727 This License is a kind of "copyleft", which means that derivative
4728 works of the document must themselves be free in the same sense. It
4729 complements the GNU General Public License, which is a copyleft
4730 license designed for free software.
4732 We have designed this License in order to use it for manuals for free
4733 software, because free software needs free documentation: a free
4734 program should come with manuals providing the same freedoms that the
4735 software does. But this License is not limited to software manuals;
4736 it can be used for any textual work, regardless of subject matter or
4737 whether it is published as a printed book. We recommend this License
4738 principally for works whose purpose is instruction or reference.
4741 1. APPLICABILITY AND DEFINITIONS
4743 This License applies to any manual or other work that contains a
4744 notice placed by the copyright holder saying it can be distributed
4745 under the terms of this License. The "Document", below, refers to any
4746 such manual or work. Any member of the public is a licensee, and is
4749 A "Modified Version" of the Document means any work containing the
4750 Document or a portion of it, either copied verbatim, or with
4751 modifications and/or translated into another language.
4753 A "Secondary Section" is a named appendix or a front-matter section of
4754 the Document that deals exclusively with the relationship of the
4755 publishers or authors of the Document to the Document's overall subject
4756 (or to related matters) and contains nothing that could fall directly
4757 within that overall subject. (For example, if the Document is in part a
4758 textbook of mathematics, a Secondary Section may not explain any
4759 mathematics.) The relationship could be a matter of historical
4760 connection with the subject or with related matters, or of legal,
4761 commercial, philosophical, ethical or political position regarding
4764 The "Invariant Sections" are certain Secondary Sections whose titles
4765 are designated, as being those of Invariant Sections, in the notice
4766 that says that the Document is released under this License.
4768 The "Cover Texts" are certain short passages of text that are listed,
4769 as Front-Cover Texts or Back-Cover Texts, in the notice that says that
4770 the Document is released under this License.
4772 A "Transparent" copy of the Document means a machine-readable copy,
4773 represented in a format whose specification is available to the
4774 general public, whose contents can be viewed and edited directly and
4775 straightforwardly with generic text editors or (for images composed of
4776 pixels) generic paint programs or (for drawings) some widely available
4777 drawing editor, and that is suitable for input to text formatters or
4778 for automatic translation to a variety of formats suitable for input
4779 to text formatters. A copy made in an otherwise Transparent file
4780 format whose markup has been designed to thwart or discourage
4781 subsequent modification by readers is not Transparent. A copy that is
4782 not "Transparent" is called "Opaque".
4784 Examples of suitable formats for Transparent copies include plain
4785 ASCII without markup, Texinfo input format, LaTeX input format, SGML
4786 or XML using a publicly available DTD, and standard-conforming simple
4787 HTML designed for human modification. Opaque formats include
4788 PostScript, PDF, proprietary formats that can be read and edited only
4789 by proprietary word processors, SGML or XML for which the DTD and/or
4790 processing tools are not generally available, and the
4791 machine-generated HTML produced by some word processors for output
4794 The "Title Page" means, for a printed book, the title page itself,
4795 plus such following pages as are needed to hold, legibly, the material
4796 this License requires to appear in the title page. For works in
4797 formats which do not have any title page as such, "Title Page" means
4798 the text near the most prominent appearance of the work's title,
4799 preceding the beginning of the body of the text.
4804 You may copy and distribute the Document in any medium, either
4805 commercially or noncommercially, provided that this License, the
4806 copyright notices, and the license notice saying this License applies
4807 to the Document are reproduced in all copies, and that you add no other
4808 conditions whatsoever to those of this License. You may not use
4809 technical measures to obstruct or control the reading or further
4810 copying of the copies you make or distribute. However, you may accept
4811 compensation in exchange for copies. If you distribute a large enough
4812 number of copies you must also follow the conditions in section 3.
4814 You may also lend copies, under the same conditions stated above, and
4815 you may publicly display copies.
4818 3. COPYING IN QUANTITY
4820 If you publish printed copies of the Document numbering more than 100,
4821 and the Document's license notice requires Cover Texts, you must enclose
4822 the copies in covers that carry, clearly and legibly, all these Cover
4823 Texts: Front-Cover Texts on the front cover, and Back-Cover Texts on
4824 the back cover. Both covers must also clearly and legibly identify
4825 you as the publisher of these copies. The front cover must present
4826 the full title with all words of the title equally prominent and
4827 visible. You may add other material on the covers in addition.
4828 Copying with changes limited to the covers, as long as they preserve
4829 the title of the Document and satisfy these conditions, can be treated
4830 as verbatim copying in other respects.
4832 If the required texts for either cover are too voluminous to fit
4833 legibly, you should put the first ones listed (as many as fit
4834 reasonably) on the actual cover, and continue the rest onto adjacent
4837 If you publish or distribute Opaque copies of the Document numbering
4838 more than 100, you must either include a machine-readable Transparent
4839 copy along with each Opaque copy, or state in or with each Opaque copy
4840 a publicly-accessible computer-network location containing a complete
4841 Transparent copy of the Document, free of added material, which the
4842 general network-using public has access to download anonymously at no
4843 charge using public-standard network protocols. If you use the latter
4844 option, you must take reasonably prudent steps, when you begin
4845 distribution of Opaque copies in quantity, to ensure that this
4846 Transparent copy will remain thus accessible at the stated location
4847 until at least one year after the last time you distribute an Opaque
4848 copy (directly or through your agents or retailers) of that edition to
4851 It is requested, but not required, that you contact the authors of the
4852 Document well before redistributing any large number of copies, to give
4853 them a chance to provide you with an updated version of the Document.
4858 You may copy and distribute a Modified Version of the Document under
4859 the conditions of sections 2 and 3 above, provided that you release
4860 the Modified Version under precisely this License, with the Modified
4861 Version filling the role of the Document, thus licensing distribution
4862 and modification of the Modified Version to whoever possesses a copy
4863 of it. In addition, you must do these things in the Modified Version:
4865 A. Use in the Title Page (and on the covers, if any) a title distinct
4866 from that of the Document, and from those of previous versions
4867 (which should, if there were any, be listed in the History section
4868 of the Document). You may use the same title as a previous version
4869 if the original publisher of that version gives permission.
4870 B. List on the Title Page, as authors, one or more persons or entities
4871 responsible for authorship of the modifications in the Modified
4872 Version, together with at least five of the principal authors of the
4873 Document (all of its principal authors, if it has less than five).
4874 C. State on the Title page the name of the publisher of the
4875 Modified Version, as the publisher.
4876 D. Preserve all the copyright notices of the Document.
4877 E. Add an appropriate copyright notice for your modifications
4878 adjacent to the other copyright notices.
4879 F. Include, immediately after the copyright notices, a license notice
4880 giving the public permission to use the Modified Version under the
4881 terms of this License, in the form shown in the Addendum below.
4882 G. Preserve in that license notice the full lists of Invariant Sections
4883 and required Cover Texts given in the Document's license notice.
4884 H. Include an unaltered copy of this License.
4885 I. Preserve the section entitled "History", and its title, and add to
4886 it an item stating at least the title, year, new authors, and
4887 publisher of the Modified Version as given on the Title Page. If
4888 there is no section entitled "History" in the Document, create one
4889 stating the title, year, authors, and publisher of the Document as
4890 given on its Title Page, then add an item describing the Modified
4891 Version as stated in the previous sentence.
4892 J. Preserve the network location, if any, given in the Document for
4893 public access to a Transparent copy of the Document, and likewise
4894 the network locations given in the Document for previous versions
4895 it was based on. These may be placed in the "History" section.
4896 You may omit a network location for a work that was published at
4897 least four years before the Document itself, or if the original
4898 publisher of the version it refers to gives permission.
4899 K. In any section entitled "Acknowledgements" or "Dedications",
4900 preserve the section's title, and preserve in the section all the
4901 substance and tone of each of the contributor acknowledgements
4902 and/or dedications given therein.
4903 L. Preserve all the Invariant Sections of the Document,
4904 unaltered in their text and in their titles. Section numbers
4905 or the equivalent are not considered part of the section titles.
4906 M. Delete any section entitled "Endorsements". Such a section
4907 may not be included in the Modified Version.
4908 N. Do not retitle any existing section as "Endorsements"
4909 or to conflict in title with any Invariant Section.
4911 If the Modified Version includes new front-matter sections or
4912 appendices that qualify as Secondary Sections and contain no material
4913 copied from the Document, you may at your option designate some or all
4914 of these sections as invariant. To do this, add their titles to the
4915 list of Invariant Sections in the Modified Version's license notice.
4916 These titles must be distinct from any other section titles.
4918 You may add a section entitled "Endorsements", provided it contains
4919 nothing but endorsements of your Modified Version by various
4920 parties--for example, statements of peer review or that the text has
4921 been approved by an organization as the authoritative definition of a
4924 You may add a passage of up to five words as a Front-Cover Text, and a
4925 passage of up to 25 words as a Back-Cover Text, to the end of the list
4926 of Cover Texts in the Modified Version. Only one passage of
4927 Front-Cover Text and one of Back-Cover Text may be added by (or
4928 through arrangements made by) any one entity. If the Document already
4929 includes a cover text for the same cover, previously added by you or
4930 by arrangement made by the same entity you are acting on behalf of,
4931 you may not add another; but you may replace the old one, on explicit
4932 permission from the previous publisher that added the old one.
4934 The author(s) and publisher(s) of the Document do not by this License
4935 give permission to use their names for publicity for or to assert or
4936 imply endorsement of any Modified Version.
4939 5. COMBINING DOCUMENTS
4941 You may combine the Document with other documents released under this
4942 License, under the terms defined in section 4 above for modified
4943 versions, provided that you include in the combination all of the
4944 Invariant Sections of all of the original documents, unmodified, and
4945 list them all as Invariant Sections of your combined work in its
4948 The combined work need only contain one copy of this License, and
4949 multiple identical Invariant Sections may be replaced with a single
4950 copy. If there are multiple Invariant Sections with the same name but
4951 different contents, make the title of each such section unique by
4952 adding at the end of it, in parentheses, the name of the original
4953 author or publisher of that section if known, or else a unique number.
4954 Make the same adjustment to the section titles in the list of
4955 Invariant Sections in the license notice of the combined work.
4957 In the combination, you must combine any sections entitled "History"
4958 in the various original documents, forming one section entitled
4959 "History"; likewise combine any sections entitled "Acknowledgements",
4960 and any sections entitled "Dedications". You must delete all sections
4961 entitled "Endorsements."
4964 6. COLLECTIONS OF DOCUMENTS
4966 You may make a collection consisting of the Document and other documents
4967 released under this License, and replace the individual copies of this
4968 License in the various documents with a single copy that is included in
4969 the collection, provided that you follow the rules of this License for
4970 verbatim copying of each of the documents in all other respects.
4972 You may extract a single document from such a collection, and distribute
4973 it individually under this License, provided you insert a copy of this
4974 License into the extracted document, and follow this License in all
4975 other respects regarding verbatim copying of that document.
4978 7. AGGREGATION WITH INDEPENDENT WORKS
4980 A compilation of the Document or its derivatives with other separate
4981 and independent documents or works, in or on a volume of a storage or
4982 distribution medium, does not as a whole count as a Modified Version
4983 of the Document, provided no compilation copyright is claimed for the
4984 compilation. Such a compilation is called an "aggregate", and this
4985 License does not apply to the other self-contained works thus compiled
4986 with the Document, on account of their being thus compiled, if they
4987 are not themselves derivative works of the Document.
4989 If the Cover Text requirement of section 3 is applicable to these
4990 copies of the Document, then if the Document is less than one quarter
4991 of the entire aggregate, the Document's Cover Texts may be placed on
4992 covers that surround only the Document within the aggregate.
4993 Otherwise they must appear on covers around the whole aggregate.
4998 Translation is considered a kind of modification, so you may
4999 distribute translations of the Document under the terms of section 4.
5000 Replacing Invariant Sections with translations requires special
5001 permission from their copyright holders, but you may include
5002 translations of some or all Invariant Sections in addition to the
5003 original versions of these Invariant Sections. You may include a
5004 translation of this License provided that you also include the
5005 original English version of this License. In case of a disagreement
5006 between the translation and the original English version of this
5007 License, the original English version will prevail.
5012 You may not copy, modify, sublicense, or distribute the Document except
5013 as expressly provided for under this License. Any other attempt to
5014 copy, modify, sublicense or distribute the Document is void, and will
5015 automatically terminate your rights under this License. However,
5016 parties who have received copies, or rights, from you under this
5017 License will not have their licenses terminated so long as such
5018 parties remain in full compliance.
5021 10. FUTURE REVISIONS OF THIS LICENSE
5023 The Free Software Foundation may publish new, revised versions
5024 of the GNU Free Documentation License from time to time. Such new
5025 versions will be similar in spirit to the present version, but may
5026 differ in detail to address new problems or concerns. See
5027 http://www.gnu.org/copyleft/.
5029 Each version of the License is given a distinguishing version number.
5030 If the Document specifies that a particular numbered version of this
5031 License "or any later version" applies to it, you have the option of
5032 following the terms and conditions either of that specified version or
5033 of any later version that has been published (not as a draft) by the
5034 Free Software Foundation. If the Document does not specify a version
5035 number of this License, you may choose any version ever published (not
5036 as a draft) by the Free Software Foundation.
5039 ADDENDUM: How to use this License for your documents
5041 To use this License in a document you have written, include a copy of
5042 the License in the document and put the following copyright and
5043 license notices just after the title page:
5046 Copyright (c) YEAR YOUR NAME.
5047 Permission is granted to copy, distribute and/or modify this document
5048 under the terms of the GNU Free Documentation License, Version 1.1
5049 or any later version published by the Free Software Foundation;
5050 with the Invariant Sections being LIST THEIR TITLES, with the
5051 Front-Cover Texts being LIST, and with the Back-Cover Texts being LIST.
5052 A copy of the license is included in the section entitled "GNU
5053 Free Documentation License".
5056 If you have no Invariant Sections, write "with no Invariant Sections"
5057 instead of saying which ones are invariant. If you have no
5058 Front-Cover Texts, write "no Front-Cover Texts" instead of
5059 "Front-Cover Texts being LIST"; likewise for Back-Cover Texts.
5061 If your document contains nontrivial examples of program code, we
5062 recommend releasing these examples in parallel under your choice of
5063 free software license, such as the GNU General Public License,
5064 to permit their use in free software.
5072 % I think something like @colophon should be in texinfo. In the
5074 \long\def\colophon{\hbox to0pt{}\vfill
5075 \centerline{The body of this manual is set in}
5076 \centerline{\fontname\tenrm,}
5077 \centerline{with headings in {\bf\fontname\tenbf}}
5078 \centerline{and examples in {\tt\fontname\tentt}.}
5079 \centerline{{\it\fontname\tenit\/} and}
5080 \centerline{{\sl\fontname\tensl\/}}
5081 \centerline{are used for emphasis.}\vfill}
5083 % Blame: doc@cygnus.com, 28mar91.