4 @include configdoc.texi
5 @c (configdoc.texi is generated by the Makefile)
13 * Ld: (ld). The GNU linker.
19 This file documents the @sc{gnu} linker LD version @value{VERSION}.
21 Copyright (C) 1991, 92, 93, 94, 95, 96, 97, 98, 99, 2000 Free Software Foundation, Inc.
25 Permission is granted to copy, distribute and/or modify this document
26 under the terms of the GNU Free Documentation License, Version 1.1
27 or any later version published by the Free Software Foundation;
28 with no Invariant Sections, with no Front-Cover Texts, and with no
29 Back-Cover Texts. A copy of the license is included in the
30 section entitled "GNU Free Documentation License".
32 Permission is granted to process this file through Tex and print the
33 results, provided the printed document carries copying permission
34 notice identical to this one except for the removal of this paragraph
35 (this paragraph not being relevant to the printed manual).
41 @setchapternewpage odd
42 @settitle Using LD, the GNU linker
45 @subtitle The GNU linker
47 @subtitle @code{ld} version 2
48 @subtitle Version @value{VERSION}
49 @author Steve Chamberlain
50 @author Ian Lance Taylor
55 \hfill Red Hat Inc\par
56 \hfill nickc\@credhat.com, doc\@redhat.com\par
57 \hfill {\it Using LD, the GNU linker}\par
58 \hfill Edited by Jeffrey Osier (jeffrey\@cygnus.com)\par
60 \global\parindent=0pt % Steve likes it this way.
63 @vskip 0pt plus 1filll
64 Copyright @copyright{} 1991, 92, 93, 94, 95, 96, 97, 98, 99, 2000 Free Software Foundation, Inc.
66 Permission is granted to copy, distribute and/or modify this document
67 under the terms of the GNU Free Documentation License, Version 1.1
68 or any later version published by the Free Software Foundation;
69 with no Invariant Sections, with no Front-Cover Texts, and with no
70 Back-Cover Texts. A copy of the license is included in the
71 section entitled "GNU Free Documentation License".
75 @c FIXME: Talk about importance of *order* of args, cmds to linker!
80 This file documents the @sc{gnu} linker ld version @value{VERSION}.
82 This document is distributed under the terms of the GNU Free
83 Documentation License. A copy of the license is included in the
84 section entitled "GNU Free Documentation License".
88 * Invocation:: Invocation
89 * Scripts:: Linker Scripts
91 * Machine Dependent:: Machine Dependent Features
95 * H8/300:: ld and the H8/300
98 * Hitachi:: ld and other Hitachi micros
101 * i960:: ld and the Intel 960 family
104 * TI COFF:: ld and the TI COFF
107 @ifclear SingleFormat
110 @c Following blank line required for remaining bug in makeinfo conds/menus
112 * Reporting Bugs:: Reporting Bugs
113 * MRI:: MRI Compatible Script Files
114 * GNU Free Documentation License:: GNU Free Documentation License
122 @cindex @sc{gnu} linker
123 @cindex what is this?
124 @code{ld} combines a number of object and archive files, relocates
125 their data and ties up symbol references. Usually the last step in
126 compiling a program is to run @code{ld}.
128 @code{ld} accepts Linker Command Language files written in
129 a superset of AT&T's Link Editor Command Language syntax,
130 to provide explicit and total control over the linking process.
132 @ifclear SingleFormat
133 This version of @code{ld} uses the general purpose BFD libraries
134 to operate on object files. This allows @code{ld} to read, combine, and
135 write object files in many different formats---for example, COFF or
136 @code{a.out}. Different formats may be linked together to produce any
137 available kind of object file. @xref{BFD}, for more information.
140 Aside from its flexibility, the @sc{gnu} linker is more helpful than other
141 linkers in providing diagnostic information. Many linkers abandon
142 execution immediately upon encountering an error; whenever possible,
143 @code{ld} continues executing, allowing you to identify other errors
144 (or, in some cases, to get an output file in spite of the error).
149 The @sc{gnu} linker @code{ld} is meant to cover a broad range of situations,
150 and to be as compatible as possible with other linkers. As a result,
151 you have many choices to control its behavior.
155 * Options:: Command Line Options
156 * Environment:: Environment Variables
160 @section Command Line Options
165 The linker supports a plethora of command-line options, but in actual
166 practice few of them are used in any particular context.
167 @cindex standard Unix system
168 For instance, a frequent use of @code{ld} is to link standard Unix
169 object files on a standard, supported Unix system. On such a system, to
170 link a file @code{hello.o}:
173 ld -o @var{output} /lib/crt0.o hello.o -lc
176 This tells @code{ld} to produce a file called @var{output} as the
177 result of linking the file @code{/lib/crt0.o} with @code{hello.o} and
178 the library @code{libc.a}, which will come from the standard search
179 directories. (See the discussion of the @samp{-l} option below.)
181 Some of the command-line options to @code{ld} may be specified at any
182 point in the command line. However, options which refer to files, such
183 as @samp{-l} or @samp{-T}, cause the file to be read at the point at
184 which the option appears in the command line, relative to the object
185 files and other file options. Repeating non-file options with a
186 different argument will either have no further effect, or override prior
187 occurrences (those further to the left on the command line) of that
188 option. Options which may be meaningfully specified more than once are
189 noted in the descriptions below.
192 Non-option arguments are object files or archives which are to be linked
193 together. They may follow, precede, or be mixed in with command-line
194 options, except that an object file argument may not be placed between
195 an option and its argument.
197 Usually the linker is invoked with at least one object file, but you can
198 specify other forms of binary input files using @samp{-l}, @samp{-R},
199 and the script command language. If @emph{no} binary input files at all
200 are specified, the linker does not produce any output, and issues the
201 message @samp{No input files}.
203 If the linker can not recognize the format of an object file, it will
204 assume that it is a linker script. A script specified in this way
205 augments the main linker script used for the link (either the default
206 linker script or the one specified by using @samp{-T}). This feature
207 permits the linker to link against a file which appears to be an object
208 or an archive, but actually merely defines some symbol values, or uses
209 @code{INPUT} or @code{GROUP} to load other objects. Note that
210 specifying a script in this way should only be used to augment the main
211 linker script; if you want to use some command that logically can only
212 appear once, such as the @code{SECTIONS} or @code{MEMORY} command, you
213 must replace the default linker script using the @samp{-T} option.
216 For options whose names are a single letter,
217 option arguments must either follow the option letter without intervening
218 whitespace, or be given as separate arguments immediately following the
219 option that requires them.
221 For options whose names are multiple letters, either one dash or two can
222 precede the option name; for example, @samp{-oformat} and
223 @samp{--oformat} are equivalent. Arguments to multiple-letter options
224 must either be separated from the option name by an equals sign, or be
225 given as separate arguments immediately following the option that
226 requires them. For example, @samp{--oformat srec} and
227 @samp{--oformat=srec} are equivalent. Unique abbreviations of the names
228 of multiple-letter options are accepted.
230 Note - if the linker is being invoked indirectly, via a compiler driver
231 (eg @samp{gcc}) then all the linker command line options should be
232 prefixed by @samp{-Wl,} (or whatever is appropriate for the particular
233 compiler driver) like this:
236 gcc -Wl,--startgroup foo.o bar.o -Wl,--endgroup
239 This is important, because otherwise the compiler driver program may
240 silently drop the linker options, resulting in a bad link.
242 Here is a table of the generic command line switches accepted by the GNU
246 @kindex -a@var{keyword}
247 @item -a@var{keyword}
248 This option is supported for HP/UX compatibility. The @var{keyword}
249 argument must be one of the strings @samp{archive}, @samp{shared}, or
250 @samp{default}. @samp{-aarchive} is functionally equivalent to
251 @samp{-Bstatic}, and the other two keywords are functionally equivalent
252 to @samp{-Bdynamic}. This option may be used any number of times.
255 @cindex architectures
257 @item -A@var{architecture}
258 @kindex --architecture=@var{arch}
259 @itemx --architecture=@var{architecture}
260 In the current release of @code{ld}, this option is useful only for the
261 Intel 960 family of architectures. In that @code{ld} configuration, the
262 @var{architecture} argument identifies the particular architecture in
263 the 960 family, enabling some safeguards and modifying the
264 archive-library search path. @xref{i960,,@code{ld} and the Intel 960
265 family}, for details.
267 Future releases of @code{ld} may support similar functionality for
268 other architecture families.
271 @ifclear SingleFormat
272 @cindex binary input format
273 @kindex -b @var{format}
274 @kindex --format=@var{format}
277 @item -b @var{input-format}
278 @itemx --format=@var{input-format}
279 @code{ld} may be configured to support more than one kind of object
280 file. If your @code{ld} is configured this way, you can use the
281 @samp{-b} option to specify the binary format for input object files
282 that follow this option on the command line. Even when @code{ld} is
283 configured to support alternative object formats, you don't usually need
284 to specify this, as @code{ld} should be configured to expect as a
285 default input format the most usual format on each machine.
286 @var{input-format} is a text string, the name of a particular format
287 supported by the BFD libraries. (You can list the available binary
288 formats with @samp{objdump -i}.)
291 You may want to use this option if you are linking files with an unusual
292 binary format. You can also use @samp{-b} to switch formats explicitly (when
293 linking object files of different formats), by including
294 @samp{-b @var{input-format}} before each group of object files in a
297 The default format is taken from the environment variable
302 You can also define the input format from a script, using the command
303 @code{TARGET}; see @ref{Format Commands}.
306 @kindex -c @var{MRI-cmdfile}
307 @kindex --mri-script=@var{MRI-cmdfile}
308 @cindex compatibility, MRI
309 @item -c @var{MRI-commandfile}
310 @itemx --mri-script=@var{MRI-commandfile}
311 For compatibility with linkers produced by MRI, @code{ld} accepts script
312 files written in an alternate, restricted command language, described in
313 @ref{MRI,,MRI Compatible Script Files}. Introduce MRI script files with
314 the option @samp{-c}; use the @samp{-T} option to run linker
315 scripts written in the general-purpose @code{ld} scripting language.
316 If @var{MRI-cmdfile} does not exist, @code{ld} looks for it in the directories
317 specified by any @samp{-L} options.
319 @cindex common allocation
326 These three options are equivalent; multiple forms are supported for
327 compatibility with other linkers. They assign space to common symbols
328 even if a relocatable output file is specified (with @samp{-r}). The
329 script command @code{FORCE_COMMON_ALLOCATION} has the same effect.
330 @xref{Miscellaneous Commands}.
332 @cindex entry point, from command line
333 @kindex -e @var{entry}
334 @kindex --entry=@var{entry}
336 @itemx --entry=@var{entry}
337 Use @var{entry} as the explicit symbol for beginning execution of your
338 program, rather than the default entry point. If there is no symbol
339 named @var{entry}, the linker will try to parse @var{entry} as a number,
340 and use that as the entry address (the number will be interpreted in
341 base 10; you may use a leading @samp{0x} for base 16, or a leading
342 @samp{0} for base 8). @xref{Entry Point}, for a discussion of defaults
343 and other ways of specifying the entry point.
345 @cindex dynamic symbol table
347 @kindex --export-dynamic
349 @itemx --export-dynamic
350 When creating a dynamically linked executable, add all symbols to the
351 dynamic symbol table. The dynamic symbol table is the set of symbols
352 which are visible from dynamic objects at run time.
354 If you do not use this option, the dynamic symbol table will normally
355 contain only those symbols which are referenced by some dynamic object
356 mentioned in the link.
358 If you use @code{dlopen} to load a dynamic object which needs to refer
359 back to the symbols defined by the program, rather than some other
360 dynamic object, then you will probably need to use this option when
361 linking the program itself.
363 @cindex big-endian objects
367 Link big-endian objects. This affects the default output format.
369 @cindex little-endian objects
372 Link little-endian objects. This affects the default output format.
377 @itemx --auxiliary @var{name}
378 When creating an ELF shared object, set the internal DT_AUXILIARY field
379 to the specified name. This tells the dynamic linker that the symbol
380 table of the shared object should be used as an auxiliary filter on the
381 symbol table of the shared object @var{name}.
383 If you later link a program against this filter object, then, when you
384 run the program, the dynamic linker will see the DT_AUXILIARY field. If
385 the dynamic linker resolves any symbols from the filter object, it will
386 first check whether there is a definition in the shared object
387 @var{name}. If there is one, it will be used instead of the definition
388 in the filter object. The shared object @var{name} need not exist.
389 Thus the shared object @var{name} may be used to provide an alternative
390 implementation of certain functions, perhaps for debugging or for
391 machine specific performance.
393 This option may be specified more than once. The DT_AUXILIARY entries
394 will be created in the order in which they appear on the command line.
399 @itemx --filter @var{name}
400 When creating an ELF shared object, set the internal DT_FILTER field to
401 the specified name. This tells the dynamic linker that the symbol table
402 of the shared object which is being created should be used as a filter
403 on the symbol table of the shared object @var{name}.
405 If you later link a program against this filter object, then, when you
406 run the program, the dynamic linker will see the DT_FILTER field. The
407 dynamic linker will resolve symbols according to the symbol table of the
408 filter object as usual, but it will actually link to the definitions
409 found in the shared object @var{name}. Thus the filter object can be
410 used to select a subset of the symbols provided by the object
413 Some older linkers used the @code{-F} option throughout a compilation
414 toolchain for specifying object-file format for both input and output
415 object files. The @sc{gnu} linker uses other mechanisms for this
416 purpose: the @code{-b}, @code{--format}, @code{--oformat} options, the
417 @code{TARGET} command in linker scripts, and the @code{GNUTARGET}
418 environment variable. The @sc{gnu} linker will ignore the @code{-F}
419 option when not creating an ELF shared object.
421 @cindex finalization function
423 @item -fini @var{name}
424 When creating an ELF executable or shared object, call NAME when the
425 executable or shared object is unloaded, by setting DT_FINI to the
426 address of the function. By default, the linker uses @code{_fini} as
427 the function to call.
431 Ignored. Provided for compatibility with other tools.
437 @itemx --gpsize=@var{value}
438 Set the maximum size of objects to be optimized using the GP register to
439 @var{size}. This is only meaningful for object file formats such as
440 MIPS ECOFF which supports putting large and small objects into different
441 sections. This is ignored for other object file formats.
443 @cindex runtime library name
445 @kindex -soname=@var{name}
447 @itemx -soname=@var{name}
448 When creating an ELF shared object, set the internal DT_SONAME field to
449 the specified name. When an executable is linked with a shared object
450 which has a DT_SONAME field, then when the executable is run the dynamic
451 linker will attempt to load the shared object specified by the DT_SONAME
452 field rather than the using the file name given to the linker.
455 @cindex incremental link
457 Perform an incremental link (same as option @samp{-r}).
459 @cindex initialization function
461 @item -init @var{name}
462 When creating an ELF executable or shared object, call NAME when the
463 executable or shared object is loaded, by setting DT_INIT to the address
464 of the function. By default, the linker uses @code{_init} as the
467 @cindex archive files, from cmd line
468 @kindex -l@var{archive}
469 @kindex --library=@var{archive}
470 @item -l@var{archive}
471 @itemx --library=@var{archive}
472 Add archive file @var{archive} to the list of files to link. This
473 option may be used any number of times. @code{ld} will search its
474 path-list for occurrences of @code{lib@var{archive}.a} for every
475 @var{archive} specified.
477 On systems which support shared libraries, @code{ld} may also search for
478 libraries with extensions other than @code{.a}. Specifically, on ELF
479 and SunOS systems, @code{ld} will search a directory for a library with
480 an extension of @code{.so} before searching for one with an extension of
481 @code{.a}. By convention, a @code{.so} extension indicates a shared
484 The linker will search an archive only once, at the location where it is
485 specified on the command line. If the archive defines a symbol which
486 was undefined in some object which appeared before the archive on the
487 command line, the linker will include the appropriate file(s) from the
488 archive. However, an undefined symbol in an object appearing later on
489 the command line will not cause the linker to search the archive again.
491 See the @code{-(} option for a way to force the linker to search
492 archives multiple times.
494 You may list the same archive multiple times on the command line.
497 This type of archive searching is standard for Unix linkers. However,
498 if you are using @code{ld} on AIX, note that it is different from the
499 behaviour of the AIX linker.
502 @cindex search directory, from cmd line
504 @kindex --library-path=@var{dir}
505 @item -L@var{searchdir}
506 @itemx --library-path=@var{searchdir}
507 Add path @var{searchdir} to the list of paths that @code{ld} will search
508 for archive libraries and @code{ld} control scripts. You may use this
509 option any number of times. The directories are searched in the order
510 in which they are specified on the command line. Directories specified
511 on the command line are searched before the default directories. All
512 @code{-L} options apply to all @code{-l} options, regardless of the
513 order in which the options appear.
516 The default set of paths searched (without being specified with
517 @samp{-L}) depends on which emulation mode @code{ld} is using, and in
518 some cases also on how it was configured. @xref{Environment}.
521 The paths can also be specified in a link script with the
522 @code{SEARCH_DIR} command. Directories specified this way are searched
523 at the point in which the linker script appears in the command line.
526 @kindex -m @var{emulation}
527 @item -m@var{emulation}
528 Emulate the @var{emulation} linker. You can list the available
529 emulations with the @samp{--verbose} or @samp{-V} options.
531 If the @samp{-m} option is not used, the emulation is taken from the
532 @code{LDEMULATION} environment variable, if that is defined.
534 Otherwise, the default emulation depends upon how the linker was
542 Print a link map to the standard output. A link map provides
543 information about the link, including the following:
547 Where object files and symbols are mapped into memory.
549 How common symbols are allocated.
551 All archive members included in the link, with a mention of the symbol
552 which caused the archive member to be brought in.
556 @cindex read-only text
561 Turn off page alignment of sections, and mark the output as
562 @code{NMAGIC} if possible.
566 @cindex read/write from cmd line
570 Set the text and data sections to be readable and writable. Also, do
571 not page-align the data segment. If the output format supports Unix
572 style magic numbers, mark the output as @code{OMAGIC}.
574 @kindex -o @var{output}
575 @kindex --output=@var{output}
576 @cindex naming the output file
577 @item -o @var{output}
578 @itemx --output=@var{output}
579 Use @var{output} as the name for the program produced by @code{ld}; if this
580 option is not specified, the name @file{a.out} is used by default. The
581 script command @code{OUTPUT} can also specify the output file name.
583 @kindex -O @var{level}
584 @cindex generating optimized output
586 If @var{level} is a numeric values greater than zero @code{ld} optimizes
587 the output. This might take significantly longer and therefore probably
588 should only be enabled for the final binary.
591 @kindex --emit-relocs
592 @cindex retain relocations in final executable
595 Leave relocation sections and contents in fully linked exececutables.
596 Post link analysis and optimization tools may need this information in
597 order to perform correct modifications of executables. This results
598 in larger executables.
601 @cindex relocatable output
603 @kindex --relocateable
605 @itemx --relocateable
606 Generate relocatable output---i.e., generate an output file that can in
607 turn serve as input to @code{ld}. This is often called @dfn{partial
608 linking}. As a side effect, in environments that support standard Unix
609 magic numbers, this option also sets the output file's magic number to
612 If this option is not specified, an absolute file is produced. When
613 linking C++ programs, this option @emph{will not} resolve references to
614 constructors; to do that, use @samp{-Ur}.
616 This option does the same thing as @samp{-i}.
618 @kindex -R @var{file}
619 @kindex --just-symbols=@var{file}
620 @cindex symbol-only input
621 @item -R @var{filename}
622 @itemx --just-symbols=@var{filename}
623 Read symbol names and their addresses from @var{filename}, but do not
624 relocate it or include it in the output. This allows your output file
625 to refer symbolically to absolute locations of memory defined in other
626 programs. You may use this option more than once.
628 For compatibility with other ELF linkers, if the @code{-R} option is
629 followed by a directory name, rather than a file name, it is treated as
630 the @code{-rpath} option.
634 @cindex strip all symbols
637 Omit all symbol information from the output file.
640 @kindex --strip-debug
641 @cindex strip debugger symbols
644 Omit debugger symbol information (but not all symbols) from the output file.
648 @cindex input files, displaying
651 Print the names of the input files as @code{ld} processes them.
653 @kindex -T @var{script}
654 @kindex --script=@var{script}
656 @item -T @var{scriptfile}
657 @itemx --script=@var{scriptfile}
658 Use @var{scriptfile} as the linker script. This script replaces
659 @code{ld}'s default linker script (rather than adding to it), so
660 @var{commandfile} must specify everything necessary to describe the
661 output file. You must use this option if you want to use a command
662 which can only appear once in a linker script, such as the
663 @code{SECTIONS} or @code{MEMORY} command. @xref{Scripts}. If
664 @var{scriptfile} does not exist in the current directory, @code{ld}
665 looks for it in the directories specified by any preceding @samp{-L}
666 options. Multiple @samp{-T} options accumulate.
668 @kindex -u @var{symbol}
669 @kindex --undefined=@var{symbol}
670 @cindex undefined symbol
671 @item -u @var{symbol}
672 @itemx --undefined=@var{symbol}
673 Force @var{symbol} to be entered in the output file as an undefined
674 symbol. Doing this may, for example, trigger linking of additional
675 modules from standard libraries. @samp{-u} may be repeated with
676 different option arguments to enter additional undefined symbols. This
677 option is equivalent to the @code{EXTERN} linker script command.
682 For anything other than C++ programs, this option is equivalent to
683 @samp{-r}: it generates relocatable output---i.e., an output file that can in
684 turn serve as input to @code{ld}. When linking C++ programs, @samp{-Ur}
685 @emph{does} resolve references to constructors, unlike @samp{-r}.
686 It does not work to use @samp{-Ur} on files that were themselves linked
687 with @samp{-Ur}; once the constructor table has been built, it cannot
688 be added to. Use @samp{-Ur} only for the last partial link, and
689 @samp{-r} for the others.
693 Creates a separate output section for every orphan input section. This
694 option prevents the normal merging of orphan input sections with the same
695 name. An orphan section is one not specifically mentioned in a linker
696 script, so this option along with a custom linker script allows any
697 selection of input sections to be merged while others are kept separate.
706 Display the version number for @code{ld}. The @code{-V} option also
707 lists the supported emulations.
710 @kindex --discard-all
711 @cindex deleting local symbols
714 Delete all local symbols.
717 @kindex --discard-locals
718 @cindex local symbols, deleting
719 @cindex L, deleting symbols beginning
721 @itemx --discard-locals
722 Delete all temporary local symbols. For most targets, this is all local
723 symbols whose names begin with @samp{L}.
725 @kindex -y @var{symbol}
726 @kindex --trace-symbol=@var{symbol}
727 @cindex symbol tracing
728 @item -y @var{symbol}
729 @itemx --trace-symbol=@var{symbol}
730 Print the name of each linked file in which @var{symbol} appears. This
731 option may be given any number of times. On many systems it is necessary
732 to prepend an underscore.
734 This option is useful when you have an undefined symbol in your link but
735 don't know where the reference is coming from.
737 @kindex -Y @var{path}
739 Add @var{path} to the default library search path. This option exists
740 for Solaris compatibility.
742 @kindex -z @var{keyword}
743 @item -z @var{keyword}
744 The recognized keywords are @code{initfirst}, @code{interpose},
745 @code{loadfltr}, @code{nodefaultlib}, @code{nodelete}, @code{nodlopen},
746 @code{nodump}, @code{now} and @code{origin}. The other keywords are
747 ignored for Solaris compatibility. @code{initfirst} marks the object
748 to be initialized first at runtime before any other objects.
749 @code{interpose} marks the object that its symbol table interposes
750 before all symbols but the primary executable. @code{loadfltr} marks
751 the object that its filtees be processed immediately at runtime.
752 @code{nodefaultlib} marks the object that the search for dependencies
753 of this object will ignore any default library search paths.
754 @code{nodelete} marks the object shouldn't be unloaded at runtime.
755 @code{nodlopen} marks the object not available to @code{dlopen}.
756 @code{nodump} marks the object can not be dumped by @code{dldump}.
757 @code{now} marks the object with the non-lazy runtime binding.
758 @code{origin} marks the object may contain $ORIGIN.
761 @cindex groups of archives
762 @item -( @var{archives} -)
763 @itemx --start-group @var{archives} --end-group
764 The @var{archives} should be a list of archive files. They may be
765 either explicit file names, or @samp{-l} options.
767 The specified archives are searched repeatedly until no new undefined
768 references are created. Normally, an archive is searched only once in
769 the order that it is specified on the command line. If a symbol in that
770 archive is needed to resolve an undefined symbol referred to by an
771 object in an archive that appears later on the command line, the linker
772 would not be able to resolve that reference. By grouping the archives,
773 they all be searched repeatedly until all possible references are
776 Using this option has a significant performance cost. It is best to use
777 it only when there are unavoidable circular references between two or
780 @kindex -assert @var{keyword}
781 @item -assert @var{keyword}
782 This option is ignored for SunOS compatibility.
790 Link against dynamic libraries. This is only meaningful on platforms
791 for which shared libraries are supported. This option is normally the
792 default on such platforms. The different variants of this option are
793 for compatibility with various systems. You may use this option
794 multiple times on the command line: it affects library searching for
795 @code{-l} options which follow it.
805 Do not link against shared libraries. This is only meaningful on
806 platforms for which shared libraries are supported. The different
807 variants of this option are for compatibility with various systems. You
808 may use this option multiple times on the command line: it affects
809 library searching for @code{-l} options which follow it.
813 When creating a shared library, bind references to global symbols to the
814 definition within the shared library, if any. Normally, it is possible
815 for a program linked against a shared library to override the definition
816 within the shared library. This option is only meaningful on ELF
817 platforms which support shared libraries.
819 @kindex --check-sections
820 @kindex --no-check-sections
821 @item --check-sections
822 @itemx --no-check-sections
823 Asks the linker @emph{not} to check section addresses after they have
824 been assigned to see if there any overlaps. Normally the linker will
825 perform this check, and if it finds any overlaps it will produce
826 suitable error messages. The linker does know about, and does make
827 allowances for sections in overlays. The default behaviour can be
828 restored by using the command line switch @samp{--check-sections}.
830 @cindex cross reference table
833 Output a cross reference table. If a linker map file is being
834 generated, the cross reference table is printed to the map file.
835 Otherwise, it is printed on the standard output.
837 The format of the table is intentionally simple, so that it may be
838 easily processed by a script if necessary. The symbols are printed out,
839 sorted by name. For each symbol, a list of file names is given. If the
840 symbol is defined, the first file listed is the location of the
841 definition. The remaining files contain references to the symbol.
843 @cindex symbols, from command line
844 @kindex --defsym @var{symbol}=@var{exp}
845 @item --defsym @var{symbol}=@var{expression}
846 Create a global symbol in the output file, containing the absolute
847 address given by @var{expression}. You may use this option as many
848 times as necessary to define multiple symbols in the command line. A
849 limited form of arithmetic is supported for the @var{expression} in this
850 context: you may give a hexadecimal constant or the name of an existing
851 symbol, or use @code{+} and @code{-} to add or subtract hexadecimal
852 constants or symbols. If you need more elaborate expressions, consider
853 using the linker command language from a script (@pxref{Assignments,,
854 Assignment: Symbol Definitions}). @emph{Note:} there should be no white
855 space between @var{symbol}, the equals sign (``@key{=}''), and
858 @cindex demangling, from command line
859 @kindex --demangle[=@var{style}]
860 @kindex --no-demangle
861 @item --demangle[=@var{style}]
863 These options control whether to demangle symbol names in error messages
864 and other output. When the linker is told to demangle, it tries to
865 present symbol names in a readable fashion: it strips leading
866 underscores if they are used by the object file format, and converts C++
867 mangled symbol names into user readable names. Different compilers have
868 different mangling styles. The optional demangling style argument can be used
869 to choose an appropriate demangling style for your compiler. The linker will
870 demangle by default unless the environment variable @samp{COLLECT_NO_DEMANGLE}
871 is set. These options may be used to override the default.
873 @cindex dynamic linker, from command line
874 @kindex --dynamic-linker @var{file}
875 @item --dynamic-linker @var{file}
876 Set the name of the dynamic linker. This is only meaningful when
877 generating dynamically linked ELF executables. The default dynamic
878 linker is normally correct; don't use this unless you know what you are
881 @cindex MIPS embedded PIC code
882 @kindex --embedded-relocs
883 @item --embedded-relocs
884 This option is only meaningful when linking MIPS embedded PIC code,
885 generated by the -membedded-pic option to the @sc{gnu} compiler and
886 assembler. It causes the linker to create a table which may be used at
887 runtime to relocate any data which was statically initialized to pointer
888 values. See the code in testsuite/ld-empic for details.
890 @kindex --force-exe-suffix
891 @item --force-exe-suffix
892 Make sure that an output file has a .exe suffix.
894 If a successfully built fully linked output file does not have a
895 @code{.exe} or @code{.dll} suffix, this option forces the linker to copy
896 the output file to one of the same name with a @code{.exe} suffix. This
897 option is useful when using unmodified Unix makefiles on a Microsoft
898 Windows host, since some versions of Windows won't run an image unless
899 it ends in a @code{.exe} suffix.
901 @kindex --gc-sections
902 @kindex --no-gc-sections
903 @cindex garbage collection
904 @item --no-gc-sections
906 Enable garbage collection of unused input sections. It is ignored on
907 targets that do not support this option. This option is not compatible
908 with @samp{-r}, nor should it be used with dynamic linking. The default
909 behaviour (of not performing this garbage collection) can be restored by
910 specifying @samp{--no-gc-sections} on the command line.
916 Print a summary of the command-line options on the standard output and exit.
918 @kindex --target-help
920 Print a summary of all target specific options on the standard output and exit.
923 @item -Map @var{mapfile}
924 Print a link map to the file @var{mapfile}. See the description of the
925 @samp{-M} option, above.
928 @kindex --no-keep-memory
929 @item --no-keep-memory
930 @code{ld} normally optimizes for speed over memory usage by caching the
931 symbol tables of input files in memory. This option tells @code{ld} to
932 instead optimize for memory usage, by rereading the symbol tables as
933 necessary. This may be required if @code{ld} runs out of memory space
934 while linking a large executable.
936 @kindex --no-undefined
938 Normally when creating a non-symbolic shared library, undefined symbols
939 are allowed and left to be resolved by the runtime loader. This option
940 disallows such undefined symbols.
942 @kindex --allow-shlib-undefined
943 @item --allow-shlib-undefined
944 Allow undefined symbols in shared objects even when --no-undefined is
945 set. The net result will be that undefined symbols in regular objects
946 will still trigger an error, but undefined symbols in shared objects
947 will be ignored. The implementation of no_undefined makes the
948 assumption that the runtime linker will choke on undefined symbols.
949 However there is at least one system (BeOS) where undefined symbols in
950 shared libraries is normal since the kernel patches them at load time to
951 select which function is most appropriate for the current architecture.
952 I.E. dynamically select an appropriate memset function. Apparently it
953 is also normal for HPPA shared libraries to have undefined symbols.
955 @kindex --no-warn-mismatch
956 @item --no-warn-mismatch
957 Normally @code{ld} will give an error if you try to link together input
958 files that are mismatched for some reason, perhaps because they have
959 been compiled for different processors or for different endiannesses.
960 This option tells @code{ld} that it should silently permit such possible
961 errors. This option should only be used with care, in cases when you
962 have taken some special action that ensures that the linker errors are
965 @kindex --no-whole-archive
966 @item --no-whole-archive
967 Turn off the effect of the @code{--whole-archive} option for subsequent
970 @cindex output file after errors
971 @kindex --noinhibit-exec
972 @item --noinhibit-exec
973 Retain the executable output file whenever it is still usable.
974 Normally, the linker will not produce an output file if it encounters
975 errors during the link process; it exits without writing an output file
976 when it issues any error whatsoever.
978 @ifclear SingleFormat
980 @item --oformat @var{output-format}
981 @code{ld} may be configured to support more than one kind of object
982 file. If your @code{ld} is configured this way, you can use the
983 @samp{--oformat} option to specify the binary format for the output
984 object file. Even when @code{ld} is configured to support alternative
985 object formats, you don't usually need to specify this, as @code{ld}
986 should be configured to produce as a default output format the most
987 usual format on each machine. @var{output-format} is a text string, the
988 name of a particular format supported by the BFD libraries. (You can
989 list the available binary formats with @samp{objdump -i}.) The script
990 command @code{OUTPUT_FORMAT} can also specify the output format, but
991 this option overrides it. @xref{BFD}.
996 This option is ignored for Linux compatibility.
1000 This option is ignored for SVR4 compatibility.
1003 @cindex synthesizing linker
1004 @cindex relaxing addressing modes
1006 An option with machine dependent effects.
1008 This option is only supported on a few targets.
1011 @xref{H8/300,,@code{ld} and the H8/300}.
1014 @xref{i960,, @code{ld} and the Intel 960 family}.
1018 On some platforms, the @samp{--relax} option performs global
1019 optimizations that become possible when the linker resolves addressing
1020 in the program, such as relaxing address modes and synthesizing new
1021 instructions in the output object file.
1023 On some platforms these link time global optimizations may make symbolic
1024 debugging of the resulting executable impossible.
1027 the case for the Matsushita MN10200 and MN10300 family of processors.
1031 On platforms where this is not supported, @samp{--relax} is accepted,
1035 @cindex retaining specified symbols
1036 @cindex stripping all but some symbols
1037 @cindex symbols, retaining selectively
1038 @item --retain-symbols-file @var{filename}
1039 Retain @emph{only} the symbols listed in the file @var{filename},
1040 discarding all others. @var{filename} is simply a flat file, with one
1041 symbol name per line. This option is especially useful in environments
1045 where a large global symbol table is accumulated gradually, to conserve
1048 @samp{--retain-symbols-file} does @emph{not} discard undefined symbols,
1049 or symbols needed for relocations.
1051 You may only specify @samp{--retain-symbols-file} once in the command
1052 line. It overrides @samp{-s} and @samp{-S}.
1055 @item -rpath @var{dir}
1056 @cindex runtime library search path
1058 Add a directory to the runtime library search path. This is used when
1059 linking an ELF executable with shared objects. All @code{-rpath}
1060 arguments are concatenated and passed to the runtime linker, which uses
1061 them to locate shared objects at runtime. The @code{-rpath} option is
1062 also used when locating shared objects which are needed by shared
1063 objects explicitly included in the link; see the description of the
1064 @code{-rpath-link} option. If @code{-rpath} is not used when linking an
1065 ELF executable, the contents of the environment variable
1066 @code{LD_RUN_PATH} will be used if it is defined.
1068 The @code{-rpath} option may also be used on SunOS. By default, on
1069 SunOS, the linker will form a runtime search patch out of all the
1070 @code{-L} options it is given. If a @code{-rpath} option is used, the
1071 runtime search path will be formed exclusively using the @code{-rpath}
1072 options, ignoring the @code{-L} options. This can be useful when using
1073 gcc, which adds many @code{-L} options which may be on NFS mounted
1076 For compatibility with other ELF linkers, if the @code{-R} option is
1077 followed by a directory name, rather than a file name, it is treated as
1078 the @code{-rpath} option.
1082 @cindex link-time runtime library search path
1084 @item -rpath-link @var{DIR}
1085 When using ELF or SunOS, one shared library may require another. This
1086 happens when an @code{ld -shared} link includes a shared library as one
1089 When the linker encounters such a dependency when doing a non-shared,
1090 non-relocatable link, it will automatically try to locate the required
1091 shared library and include it in the link, if it is not included
1092 explicitly. In such a case, the @code{-rpath-link} option
1093 specifies the first set of directories to search. The
1094 @code{-rpath-link} option may specify a sequence of directory names
1095 either by specifying a list of names separated by colons, or by
1096 appearing multiple times.
1098 This option should be used with caution as it overrides the search path
1099 that may have been hard compiled into a shared library. In such a case it
1100 is possible to use unintentionally a different search path than the
1101 runtime linker would do.
1103 The linker uses the following search paths to locate required shared
1107 Any directories specified by @code{-rpath-link} options.
1109 Any directories specified by @code{-rpath} options. The difference
1110 between @code{-rpath} and @code{-rpath-link} is that directories
1111 specified by @code{-rpath} options are included in the executable and
1112 used at runtime, whereas the @code{-rpath-link} option is only effective
1113 at link time. It is for the native linker only.
1115 On an ELF system, if the @code{-rpath} and @code{rpath-link} options
1116 were not used, search the contents of the environment variable
1117 @code{LD_RUN_PATH}. It is for the native linker only.
1119 On SunOS, if the @code{-rpath} option was not used, search any
1120 directories specified using @code{-L} options.
1122 For a native linker, the contents of the environment variable
1123 @code{LD_LIBRARY_PATH}.
1125 For a native ELF linker, the directories in @code{DT_RUNPATH} or
1126 @code{DT_RPATH} of a shared library are searched for shared
1127 libraries needed by it. The @code{DT_RPATH} entries are ignored if
1128 @code{DT_RUNPATH} entries exist.
1130 The default directories, normally @file{/lib} and @file{/usr/lib}.
1132 For a native linker on an ELF system, if the file @file{/etc/ld.so.conf}
1133 exists, the list of directories found in that file.
1136 If the required shared library is not found, the linker will issue a
1137 warning and continue with the link.
1144 @cindex shared libraries
1145 Create a shared library. This is currently only supported on ELF, XCOFF
1146 and SunOS platforms. On SunOS, the linker will automatically create a
1147 shared library if the @code{-e} option is not used and there are
1148 undefined symbols in the link.
1151 @kindex --sort-common
1152 This option tells @code{ld} to sort the common symbols by size when it
1153 places them in the appropriate output sections. First come all the one
1154 byte symbols, then all the two bytes, then all the four bytes, and then
1155 everything else. This is to prevent gaps between symbols due to
1156 alignment constraints.
1158 @kindex --split-by-file
1159 @item --split-by-file [@var{size}]
1160 Similar to @code{--split-by-reloc} but creates a new output section for
1161 each input file when @var{size} is reached. @var{size} defaults to a
1162 size of 1 if not given.
1164 @kindex --split-by-reloc
1165 @item --split-by-reloc [@var{count}]
1166 Tries to creates extra sections in the output file so that no single
1167 output section in the file contains more than @var{count} relocations.
1168 This is useful when generating huge relocatable files for downloading into
1169 certain real time kernels with the COFF object file format; since COFF
1170 cannot represent more than 65535 relocations in a single section. Note
1171 that this will fail to work with object file formats which do not
1172 support arbitrary sections. The linker will not split up individual
1173 input sections for redistribution, so if a single input section contains
1174 more than @var{count} relocations one output section will contain that
1175 many relocations. @var{count} defaults to a value of 32768.
1179 Compute and display statistics about the operation of the linker, such
1180 as execution time and memory usage.
1182 @kindex --traditional-format
1183 @cindex traditional format
1184 @item --traditional-format
1185 For some targets, the output of @code{ld} is different in some ways from
1186 the output of some existing linker. This switch requests @code{ld} to
1187 use the traditional format instead.
1190 For example, on SunOS, @code{ld} combines duplicate entries in the
1191 symbol string table. This can reduce the size of an output file with
1192 full debugging information by over 30 percent. Unfortunately, the SunOS
1193 @code{dbx} program can not read the resulting program (@code{gdb} has no
1194 trouble). The @samp{--traditional-format} switch tells @code{ld} to not
1195 combine duplicate entries.
1197 @kindex --section-start @var{sectionname}=@var{org}
1198 @item --section-start @var{sectionname}=@var{org}
1199 Locate a section in the output file at the absolute
1200 address given by @var{org}. You may use this option as many
1201 times as necessary to locate multiple sections in the command
1203 @var{org} must be a single hexadecimal integer;
1204 for compatibility with other linkers, you may omit the leading
1205 @samp{0x} usually associated with hexadecimal values. @emph{Note:} there
1206 should be no white space between @var{sectionname}, the equals
1207 sign (``@key{=}''), and @var{org}.
1209 @kindex -Tbss @var{org}
1210 @kindex -Tdata @var{org}
1211 @kindex -Ttext @var{org}
1212 @cindex segment origins, cmd line
1213 @item -Tbss @var{org}
1214 @itemx -Tdata @var{org}
1215 @itemx -Ttext @var{org}
1216 Use @var{org} as the starting address for---respectively---the
1217 @code{bss}, @code{data}, or the @code{text} segment of the output file.
1218 @var{org} must be a single hexadecimal integer;
1219 for compatibility with other linkers, you may omit the leading
1220 @samp{0x} usually associated with hexadecimal values.
1226 Display the version number for @code{ld} and list the linker emulations
1227 supported. Display which input files can and cannot be opened. Display
1228 the linker script if using a default builtin script.
1230 @kindex --version-script=@var{version-scriptfile}
1231 @cindex version script, symbol versions
1232 @itemx --version-script=@var{version-scriptfile}
1233 Specify the name of a version script to the linker. This is typically
1234 used when creating shared libraries to specify additional information
1235 about the version heirarchy for the library being created. This option
1236 is only meaningful on ELF platforms which support shared libraries.
1239 @kindex --warn-comon
1240 @cindex warnings, on combining symbols
1241 @cindex combining symbols, warnings on
1243 Warn when a common symbol is combined with another common symbol or with
1244 a symbol definition. Unix linkers allow this somewhat sloppy practice,
1245 but linkers on some other operating systems do not. This option allows
1246 you to find potential problems from combining global symbols.
1247 Unfortunately, some C libraries use this practice, so you may get some
1248 warnings about symbols in the libraries as well as in your programs.
1250 There are three kinds of global symbols, illustrated here by C examples:
1254 A definition, which goes in the initialized data section of the output
1258 An undefined reference, which does not allocate space.
1259 There must be either a definition or a common symbol for the
1263 A common symbol. If there are only (one or more) common symbols for a
1264 variable, it goes in the uninitialized data area of the output file.
1265 The linker merges multiple common symbols for the same variable into a
1266 single symbol. If they are of different sizes, it picks the largest
1267 size. The linker turns a common symbol into a declaration, if there is
1268 a definition of the same variable.
1271 The @samp{--warn-common} option can produce five kinds of warnings.
1272 Each warning consists of a pair of lines: the first describes the symbol
1273 just encountered, and the second describes the previous symbol
1274 encountered with the same name. One or both of the two symbols will be
1279 Turning a common symbol into a reference, because there is already a
1280 definition for the symbol.
1282 @var{file}(@var{section}): warning: common of `@var{symbol}'
1283 overridden by definition
1284 @var{file}(@var{section}): warning: defined here
1288 Turning a common symbol into a reference, because a later definition for
1289 the symbol is encountered. This is the same as the previous case,
1290 except that the symbols are encountered in a different order.
1292 @var{file}(@var{section}): warning: definition of `@var{symbol}'
1294 @var{file}(@var{section}): warning: common is here
1298 Merging a common symbol with a previous same-sized common symbol.
1300 @var{file}(@var{section}): warning: multiple common
1302 @var{file}(@var{section}): warning: previous common is here
1306 Merging a common symbol with a previous larger common symbol.
1308 @var{file}(@var{section}): warning: common of `@var{symbol}'
1309 overridden by larger common
1310 @var{file}(@var{section}): warning: larger common is here
1314 Merging a common symbol with a previous smaller common symbol. This is
1315 the same as the previous case, except that the symbols are
1316 encountered in a different order.
1318 @var{file}(@var{section}): warning: common of `@var{symbol}'
1319 overriding smaller common
1320 @var{file}(@var{section}): warning: smaller common is here
1324 @kindex --warn-constructors
1325 @item --warn-constructors
1326 Warn if any global constructors are used. This is only useful for a few
1327 object file formats. For formats like COFF or ELF, the linker can not
1328 detect the use of global constructors.
1330 @kindex --warn-multiple-gp
1331 @item --warn-multiple-gp
1332 Warn if multiple global pointer values are required in the output file.
1333 This is only meaningful for certain processors, such as the Alpha.
1334 Specifically, some processors put large-valued constants in a special
1335 section. A special register (the global pointer) points into the middle
1336 of this section, so that constants can be loaded efficiently via a
1337 base-register relative addressing mode. Since the offset in
1338 base-register relative mode is fixed and relatively small (e.g., 16
1339 bits), this limits the maximum size of the constant pool. Thus, in
1340 large programs, it is often necessary to use multiple global pointer
1341 values in order to be able to address all possible constants. This
1342 option causes a warning to be issued whenever this case occurs.
1345 @cindex warnings, on undefined symbols
1346 @cindex undefined symbols, warnings on
1348 Only warn once for each undefined symbol, rather than once per module
1351 @kindex --warn-section-align
1352 @cindex warnings, on section alignment
1353 @cindex section alignment, warnings on
1354 @item --warn-section-align
1355 Warn if the address of an output section is changed because of
1356 alignment. Typically, the alignment will be set by an input section.
1357 The address will only be changed if it not explicitly specified; that
1358 is, if the @code{SECTIONS} command does not specify a start address for
1359 the section (@pxref{SECTIONS}).
1361 @kindex --whole-archive
1362 @cindex including an entire archive
1363 @item --whole-archive
1364 For each archive mentioned on the command line after the
1365 @code{--whole-archive} option, include every object file in the archive
1366 in the link, rather than searching the archive for the required object
1367 files. This is normally used to turn an archive file into a shared
1368 library, forcing every object to be included in the resulting shared
1369 library. This option may be used more than once.
1371 Two notes when using this option from gcc: First, gcc doesn't know
1372 about this option, so you have to use @code{-Wl,-whole-archive}.
1373 Second, don't forget to use @code{-Wl,-no-whole-archive} after your
1374 list of archives, because gcc will add its own list of archives to
1375 your link and you may not want this flag to affect those as well.
1378 @item --wrap @var{symbol}
1379 Use a wrapper function for @var{symbol}. Any undefined reference to
1380 @var{symbol} will be resolved to @code{__wrap_@var{symbol}}. Any
1381 undefined reference to @code{__real_@var{symbol}} will be resolved to
1384 This can be used to provide a wrapper for a system function. The
1385 wrapper function should be called @code{__wrap_@var{symbol}}. If it
1386 wishes to call the system function, it should call
1387 @code{__real_@var{symbol}}.
1389 Here is a trivial example:
1393 __wrap_malloc (int c)
1395 printf ("malloc called with %ld\n", c);
1396 return __real_malloc (c);
1400 If you link other code with this file using @code{--wrap malloc}, then
1401 all calls to @code{malloc} will call the function @code{__wrap_malloc}
1402 instead. The call to @code{__real_malloc} in @code{__wrap_malloc} will
1403 call the real @code{malloc} function.
1405 You may wish to provide a @code{__real_malloc} function as well, so that
1406 links without the @code{--wrap} option will succeed. If you do this,
1407 you should not put the definition of @code{__real_malloc} in the same
1408 file as @code{__wrap_malloc}; if you do, the assembler may resolve the
1409 call before the linker has a chance to wrap it to @code{malloc}.
1411 @kindex --enable-new-dtags
1412 @kindex --disable-new-dtags
1413 @item --enable-new-dtags
1414 @itemx --disable-new-dtags
1415 This linker can create the new dynamic tags in ELF. But the older ELF
1416 systems may not understand them. If you specify
1417 @code{--enable-new-dtags}, the dynamic tags will be created as needed.
1418 If you specify @code{--disable-new-dtags}, no new dynamic tags will be
1419 created. By default, the new dynamic tags are not created. Note that
1420 those options are only available for ELF systems.
1424 @subsection Options specific to i386 PE targets
1426 The i386 PE linker supports the @code{-shared} option, which causes
1427 the output to be a dynamically linked library (DLL) instead of a
1428 normal executable. You should name the output @code{*.dll} when you
1429 use this option. In addition, the linker fully supports the standard
1430 @code{*.def} files, which may be specified on the linker command line
1431 like an object file (in fact, it should precede archives it exports
1432 symbols from, to ensure that they get linked in, just like a normal
1435 In addition to the options common to all targets, the i386 PE linker
1436 support additional command line options that are specific to the i386
1437 PE target. Options that take values may be separated from their
1438 values by either a space or an equals sign.
1442 @kindex --add-stdcall-alias
1443 @item --add-stdcall-alias
1444 If given, symbols with a stdcall suffix (@@@var{nn}) will be exported
1445 as-is and also with the suffix stripped.
1448 @item --base-file @var{file}
1449 Use @var{file} as the name of a file in which to save the base
1450 addresses of all the relocations needed for generating DLLs with
1455 Create a DLL instead of a regular executable. You may also use
1456 @code{-shared} or specify a @code{LIBRARY} in a given @code{.def}
1459 @kindex --enable-stdcall-fixup
1460 @kindex --disable-stdcall-fixup
1461 @item --enable-stdcall-fixup
1462 @itemx --disable-stdcall-fixup
1463 If the link finds a symbol that it cannot resolve, it will attempt to
1464 do "fuzzy linking" by looking for another defined symbol that differs
1465 only in the format of the symbol name (cdecl vs stdcall) and will
1466 resolve that symbol by linking to the match. For example, the
1467 undefined symbol @code{_foo} might be linked to the function
1468 @code{_foo@@12}, or the undefined symbol @code{_bar@@16} might be linked
1469 to the function @code{_bar}. When the linker does this, it prints a
1470 warning, since it normally should have failed to link, but sometimes
1471 import libraries generated from third-party dlls may need this feature
1472 to be usable. If you specify @code{--enable-stdcall-fixup}, this
1473 feature is fully enabled and warnings are not printed. If you specify
1474 @code{--disable-stdcall-fixup}, this feature is disabled and such
1475 mismatches are considered to be errors.
1477 @cindex DLLs, creating
1478 @kindex --export-all-symbols
1479 @item --export-all-symbols
1480 If given, all global symbols in the objects used to build a DLL will
1481 be exported by the DLL. Note that this is the default if there
1482 otherwise wouldn't be any exported symbols. When symbols are
1483 explicitly exported via DEF files or implicitly exported via function
1484 attributes, the default is to not export anything else unless this
1485 option is given. Note that the symbols @code{DllMain@@12},
1486 @code{DllEntryPoint@@0}, and @code{impure_ptr} will not be automatically
1489 @kindex --exclude-symbols
1490 @item --exclude-symbols @var{symbol},@var{symbol},...
1491 Specifies a list of symbols which should not be automatically
1492 exported. The symbol names may be delimited by commas or colons.
1494 @kindex --file-alignment
1495 @item --file-alignment
1496 Specify the file alignment. Sections in the file will always begin at
1497 file offsets which are multiples of this number. This defaults to
1502 @item --heap @var{reserve}
1503 @itemx --heap @var{reserve},@var{commit}
1504 Specify the amount of memory to reserve (and optionally commit) to be
1505 used as heap for this program. The default is 1Mb reserved, 4K
1509 @kindex --image-base
1510 @item --image-base @var{value}
1511 Use @var{value} as the base address of your program or dll. This is
1512 the lowest memory location that will be used when your program or dll
1513 is loaded. To reduce the need to relocate and improve performance of
1514 your dlls, each should have a unique base address and not overlap any
1515 other dlls. The default is 0x400000 for executables, and 0x10000000
1520 If given, the stdcall suffixes (@@@var{nn}) will be stripped from
1521 symbols before they are exported.
1523 @kindex --major-image-version
1524 @item --major-image-version @var{value}
1525 Sets the major number of the "image version". Defaults to 1.
1527 @kindex --major-os-version
1528 @item --major-os-version @var{value}
1529 Sets the major number of the "os version". Defaults to 4.
1531 @kindex --major-subsystem-version
1532 @item --major-subsystem-version @var{value}
1533 Sets the major number of the "subsystem version". Defaults to 4.
1535 @kindex --minor-image-version
1536 @item --minor-image-version @var{value}
1537 Sets the minor number of the "image version". Defaults to 0.
1539 @kindex --minor-os-version
1540 @item --minor-os-version @var{value}
1541 Sets the minor number of the "os version". Defaults to 0.
1543 @kindex --minor-subsystem-version
1544 @item --minor-subsystem-version @var{value}
1545 Sets the minor number of the "subsystem version". Defaults to 0.
1547 @cindex DEF files, creating
1548 @cindex DLLs, creating
1549 @kindex --output-def
1550 @item --output-def @var{file}
1551 The linker will create the file @var{file} which will contain a DEF
1552 file corresponding to the DLL the linker is generating. This DEF file
1553 (which should be called @code{*.def}) may be used to create an import
1554 library with @code{dlltool} or may be used as a reference to
1555 automatically or implicitly exported symbols.
1557 @kindex --section-alignment
1558 @item --section-alignment
1559 Sets the section alignment. Sections in memory will always begin at
1560 addresses which are a multiple of this number. Defaults to 0x1000.
1564 @item --stack @var{reserve}
1565 @itemx --stack @var{reserve},@var{commit}
1566 Specify the amount of memory to reserve (and optionally commit) to be
1567 used as stack for this program. The default is 32Mb reserved, 4K
1571 @item --subsystem @var{which}
1572 @itemx --subsystem @var{which}:@var{major}
1573 @itemx --subsystem @var{which}:@var{major}.@var{minor}
1574 Specifies the subsystem under which your program will execute. The
1575 legal values for @var{which} are @code{native}, @code{windows},
1576 @code{console}, and @code{posix}. You may optionally set the
1577 subsystem version also.
1583 @section Environment Variables
1585 You can change the behavior of @code{ld} with the environment variables
1586 @code{GNUTARGET}, @code{LDEMULATION}, and @code{COLLECT_NO_DEMANGLE}.
1589 @cindex default input format
1590 @code{GNUTARGET} determines the input-file object format if you don't
1591 use @samp{-b} (or its synonym @samp{--format}). Its value should be one
1592 of the BFD names for an input format (@pxref{BFD}). If there is no
1593 @code{GNUTARGET} in the environment, @code{ld} uses the natural format
1594 of the target. If @code{GNUTARGET} is set to @code{default} then BFD
1595 attempts to discover the input format by examining binary input files;
1596 this method often succeeds, but there are potential ambiguities, since
1597 there is no method of ensuring that the magic number used to specify
1598 object-file formats is unique. However, the configuration procedure for
1599 BFD on each system places the conventional format for that system first
1600 in the search-list, so ambiguities are resolved in favor of convention.
1603 @cindex default emulation
1604 @cindex emulation, default
1605 @code{LDEMULATION} determines the default emulation if you don't use the
1606 @samp{-m} option. The emulation can affect various aspects of linker
1607 behaviour, particularly the default linker script. You can list the
1608 available emulations with the @samp{--verbose} or @samp{-V} options. If
1609 the @samp{-m} option is not used, and the @code{LDEMULATION} environment
1610 variable is not defined, the default emulation depends upon how the
1611 linker was configured.
1614 @kindex COLLECT_NO_DEMANGLE
1615 @cindex demangling, default
1616 Normally, the linker will default to demangling symbols. However, if
1617 @code{COLLECT_NO_DEMANGLE} is set in the environment, then it will
1618 default to not demangling symbols. This environment variable is used in
1619 a similar fashion by the @code{gcc} linker wrapper program. The default
1620 may be overridden by the @samp{--demangle} and @samp{--no-demangle}
1624 @chapter Linker Scripts
1627 @cindex linker scripts
1628 @cindex command files
1629 Every link is controlled by a @dfn{linker script}. This script is
1630 written in the linker command language.
1632 The main purpose of the linker script is to describe how the sections in
1633 the input files should be mapped into the output file, and to control
1634 the memory layout of the output file. Most linker scripts do nothing
1635 more than this. However, when necessary, the linker script can also
1636 direct the linker to perform many other operations, using the commands
1639 The linker always uses a linker script. If you do not supply one
1640 yourself, the linker will use a default script that is compiled into the
1641 linker executable. You can use the @samp{--verbose} command line option
1642 to display the default linker script. Certain command line options,
1643 such as @samp{-r} or @samp{-N}, will affect the default linker script.
1645 You may supply your own linker script by using the @samp{-T} command
1646 line option. When you do this, your linker script will replace the
1647 default linker script.
1649 You may also use linker scripts implicitly by naming them as input files
1650 to the linker, as though they were files to be linked. @xref{Implicit
1654 * Basic Script Concepts:: Basic Linker Script Concepts
1655 * Script Format:: Linker Script Format
1656 * Simple Example:: Simple Linker Script Example
1657 * Simple Commands:: Simple Linker Script Commands
1658 * Assignments:: Assigning Values to Symbols
1659 * SECTIONS:: SECTIONS Command
1660 * MEMORY:: MEMORY Command
1661 * PHDRS:: PHDRS Command
1662 * VERSION:: VERSION Command
1663 * Expressions:: Expressions in Linker Scripts
1664 * Implicit Linker Scripts:: Implicit Linker Scripts
1667 @node Basic Script Concepts
1668 @section Basic Linker Script Concepts
1669 @cindex linker script concepts
1670 We need to define some basic concepts and vocabulary in order to
1671 describe the linker script language.
1673 The linker combines input files into a single output file. The output
1674 file and each input file are in a special data format known as an
1675 @dfn{object file format}. Each file is called an @dfn{object file}.
1676 The output file is often called an @dfn{executable}, but for our
1677 purposes we will also call it an object file. Each object file has,
1678 among other things, a list of @dfn{sections}. We sometimes refer to a
1679 section in an input file as an @dfn{input section}; similarly, a section
1680 in the output file is an @dfn{output section}.
1682 Each section in an object file has a name and a size. Most sections
1683 also have an associated block of data, known as the @dfn{section
1684 contents}. A section may be marked as @dfn{loadable}, which mean that
1685 the contents should be loaded into memory when the output file is run.
1686 A section with no contents may be @dfn{allocatable}, which means that an
1687 area in memory should be set aside, but nothing in particular should be
1688 loaded there (in some cases this memory must be zeroed out). A section
1689 which is neither loadable nor allocatable typically contains some sort
1690 of debugging information.
1692 Every loadable or allocatable output section has two addresses. The
1693 first is the @dfn{VMA}, or virtual memory address. This is the address
1694 the section will have when the output file is run. The second is the
1695 @dfn{LMA}, or load memory address. This is the address at which the
1696 section will be loaded. In most cases the two addresses will be the
1697 same. An example of when they might be different is when a data section
1698 is loaded into ROM, and then copied into RAM when the program starts up
1699 (this technique is often used to initialize global variables in a ROM
1700 based system). In this case the ROM address would be the LMA, and the
1701 RAM address would be the VMA.
1703 You can see the sections in an object file by using the @code{objdump}
1704 program with the @samp{-h} option.
1706 Every object file also has a list of @dfn{symbols}, known as the
1707 @dfn{symbol table}. A symbol may be defined or undefined. Each symbol
1708 has a name, and each defined symbol has an address, among other
1709 information. If you compile a C or C++ program into an object file, you
1710 will get a defined symbol for every defined function and global or
1711 static variable. Every undefined function or global variable which is
1712 referenced in the input file will become an undefined symbol.
1714 You can see the symbols in an object file by using the @code{nm}
1715 program, or by using the @code{objdump} program with the @samp{-t}
1719 @section Linker Script Format
1720 @cindex linker script format
1721 Linker scripts are text files.
1723 You write a linker script as a series of commands. Each command is
1724 either a keyword, possibly followed by arguments, or an assignment to a
1725 symbol. You may separate commands using semicolons. Whitespace is
1728 Strings such as file or format names can normally be entered directly.
1729 If the file name contains a character such as a comma which would
1730 otherwise serve to separate file names, you may put the file name in
1731 double quotes. There is no way to use a double quote character in a
1734 You may include comments in linker scripts just as in C, delimited by
1735 @samp{/*} and @samp{*/}. As in C, comments are syntactically equivalent
1738 @node Simple Example
1739 @section Simple Linker Script Example
1740 @cindex linker script example
1741 @cindex example of linker script
1742 Many linker scripts are fairly simple.
1744 The simplest possible linker script has just one command:
1745 @samp{SECTIONS}. You use the @samp{SECTIONS} command to describe the
1746 memory layout of the output file.
1748 The @samp{SECTIONS} command is a powerful command. Here we will
1749 describe a simple use of it. Let's assume your program consists only of
1750 code, initialized data, and uninitialized data. These will be in the
1751 @samp{.text}, @samp{.data}, and @samp{.bss} sections, respectively.
1752 Let's assume further that these are the only sections which appear in
1755 For this example, let's say that the code should be loaded at address
1756 0x10000, and that the data should start at address 0x8000000. Here is a
1757 linker script which will do that:
1762 .text : @{ *(.text) @}
1764 .data : @{ *(.data) @}
1765 .bss : @{ *(.bss) @}
1769 You write the @samp{SECTIONS} command as the keyword @samp{SECTIONS},
1770 followed by a series of symbol assignments and output section
1771 descriptions enclosed in curly braces.
1773 The first line inside the @samp{SECTIONS} command of the above example
1774 sets the value of the special symbol @samp{.}, which is the location
1775 counter. If you do not specify the address of an output section in some
1776 other way (other ways are described later), the address is set from the
1777 current value of the location counter. The location counter is then
1778 incremented by the size of the output section. At the start of the
1779 @samp{SECTIONS} command, the location counter has the value @samp{0}.
1781 The second line defines an output section, @samp{.text}. The colon is
1782 required syntax which may be ignored for now. Within the curly braces
1783 after the output section name, you list the names of the input sections
1784 which should be placed into this output section. The @samp{*} is a
1785 wildcard which matches any file name. The expression @samp{*(.text)}
1786 means all @samp{.text} input sections in all input files.
1788 Since the location counter is @samp{0x10000} when the output section
1789 @samp{.text} is defined, the linker will set the address of the
1790 @samp{.text} section in the output file to be @samp{0x10000}.
1792 The remaining lines define the @samp{.data} and @samp{.bss} sections in
1793 the output file. The linker will place the @samp{.data} output section
1794 at address @samp{0x8000000}. After the linker places the @samp{.data}
1795 output section, the value of the location counter will be
1796 @samp{0x8000000} plus the size of the @samp{.data} output section. The
1797 effect is that the linker will place the @samp{.bss} output section
1798 immediately after the @samp{.data} output section in memory
1800 The linker will ensure that each output section has the required
1801 alignment, by increasing the location counter if necessary. In this
1802 example, the specified addresses for the @samp{.text} and @samp{.data}
1803 sections will probably satisfy any alignment constraints, but the linker
1804 may have to create a small gap between the @samp{.data} and @samp{.bss}
1807 That's it! That's a simple and complete linker script.
1809 @node Simple Commands
1810 @section Simple Linker Script Commands
1811 @cindex linker script simple commands
1812 In this section we describe the simple linker script commands.
1815 * Entry Point:: Setting the entry point
1816 * File Commands:: Commands dealing with files
1817 @ifclear SingleFormat
1818 * Format Commands:: Commands dealing with object file formats
1821 * Miscellaneous Commands:: Other linker script commands
1825 @subsection Setting the entry point
1826 @kindex ENTRY(@var{symbol})
1827 @cindex start of execution
1828 @cindex first instruction
1830 The first instruction to execute in a program is called the @dfn{entry
1831 point}. You can use the @code{ENTRY} linker script command to set the
1832 entry point. The argument is a symbol name:
1837 There are several ways to set the entry point. The linker will set the
1838 entry point by trying each of the following methods in order, and
1839 stopping when one of them succeeds:
1842 the @samp{-e} @var{entry} command-line option;
1844 the @code{ENTRY(@var{symbol})} command in a linker script;
1846 the value of the symbol @code{start}, if defined;
1848 the address of the first byte of the @samp{.text} section, if present;
1850 The address @code{0}.
1854 @subsection Commands dealing with files
1855 @cindex linker script file commands
1856 Several linker script commands deal with files.
1859 @item INCLUDE @var{filename}
1860 @kindex INCLUDE @var{filename}
1861 @cindex including a linker script
1862 Include the linker script @var{filename} at this point. The file will
1863 be searched for in the current directory, and in any directory specified
1864 with the @code{-L} option. You can nest calls to @code{INCLUDE} up to
1867 @item INPUT(@var{file}, @var{file}, @dots{})
1868 @itemx INPUT(@var{file} @var{file} @dots{})
1869 @kindex INPUT(@var{files})
1870 @cindex input files in linker scripts
1871 @cindex input object files in linker scripts
1872 @cindex linker script input object files
1873 The @code{INPUT} command directs the linker to include the named files
1874 in the link, as though they were named on the command line.
1876 For example, if you always want to include @file{subr.o} any time you do
1877 a link, but you can't be bothered to put it on every link command line,
1878 then you can put @samp{INPUT (subr.o)} in your linker script.
1880 In fact, if you like, you can list all of your input files in the linker
1881 script, and then invoke the linker with nothing but a @samp{-T} option.
1883 The linker will first try to open the file in the current directory. If
1884 it is not found, the linker will search through the archive library
1885 search path. See the description of @samp{-L} in @ref{Options,,Command
1888 If you use @samp{INPUT (-l@var{file})}, @code{ld} will transform the
1889 name to @code{lib@var{file}.a}, as with the command line argument
1892 When you use the @code{INPUT} command in an implicit linker script, the
1893 files will be included in the link at the point at which the linker
1894 script file is included. This can affect archive searching.
1896 @item GROUP(@var{file}, @var{file}, @dots{})
1897 @itemx GROUP(@var{file} @var{file} @dots{})
1898 @kindex GROUP(@var{files})
1899 @cindex grouping input files
1900 The @code{GROUP} command is like @code{INPUT}, except that the named
1901 files should all be archives, and they are searched repeatedly until no
1902 new undefined references are created. See the description of @samp{-(}
1903 in @ref{Options,,Command Line Options}.
1905 @item OUTPUT(@var{filename})
1906 @kindex OUTPUT(@var{filename})
1907 @cindex output file name in linker scripot
1908 The @code{OUTPUT} command names the output file. Using
1909 @code{OUTPUT(@var{filename})} in the linker script is exactly like using
1910 @samp{-o @var{filename}} on the command line (@pxref{Options,,Command
1911 Line Options}). If both are used, the command line option takes
1914 You can use the @code{OUTPUT} command to define a default name for the
1915 output file other than the usual default of @file{a.out}.
1917 @item SEARCH_DIR(@var{path})
1918 @kindex SEARCH_DIR(@var{path})
1919 @cindex library search path in linker script
1920 @cindex archive search path in linker script
1921 @cindex search path in linker script
1922 The @code{SEARCH_DIR} command adds @var{path} to the list of paths where
1923 @code{ld} looks for archive libraries. Using
1924 @code{SEARCH_DIR(@var{path})} is exactly like using @samp{-L @var{path}}
1925 on the command line (@pxref{Options,,Command Line Options}). If both
1926 are used, then the linker will search both paths. Paths specified using
1927 the command line option are searched first.
1929 @item STARTUP(@var{filename})
1930 @kindex STARTUP(@var{filename})
1931 @cindex first input file
1932 The @code{STARTUP} command is just like the @code{INPUT} command, except
1933 that @var{filename} will become the first input file to be linked, as
1934 though it were specified first on the command line. This may be useful
1935 when using a system in which the entry point is always the start of the
1939 @ifclear SingleFormat
1940 @node Format Commands
1941 @subsection Commands dealing with object file formats
1942 A couple of linker script commands deal with object file formats.
1945 @item OUTPUT_FORMAT(@var{bfdname})
1946 @itemx OUTPUT_FORMAT(@var{default}, @var{big}, @var{little})
1947 @kindex OUTPUT_FORMAT(@var{bfdname})
1948 @cindex output file format in linker script
1949 The @code{OUTPUT_FORMAT} command names the BFD format to use for the
1950 output file (@pxref{BFD}). Using @code{OUTPUT_FORMAT(@var{bfdname})} is
1951 exactly like using @samp{-oformat @var{bfdname}} on the command line
1952 (@pxref{Options,,Command Line Options}). If both are used, the command
1953 line option takes precedence.
1955 You can use @code{OUTPUT_FORMAT} with three arguments to use different
1956 formats based on the @samp{-EB} and @samp{-EL} command line options.
1957 This permits the linker script to set the output format based on the
1960 If neither @samp{-EB} nor @samp{-EL} are used, then the output format
1961 will be the first argument, @var{default}. If @samp{-EB} is used, the
1962 output format will be the second argument, @var{big}. If @samp{-EL} is
1963 used, the output format will be the third argument, @var{little}.
1965 For example, the default linker script for the MIPS ELF target uses this
1968 OUTPUT_FORMAT(elf32-bigmips, elf32-bigmips, elf32-littlemips)
1970 This says that the default format for the output file is
1971 @samp{elf32-bigmips}, but if the user uses the @samp{-EL} command line
1972 option, the output file will be created in the @samp{elf32-littlemips}
1975 @item TARGET(@var{bfdname})
1976 @kindex TARGET(@var{bfdname})
1977 @cindex input file format in linker script
1978 The @code{TARGET} command names the BFD format to use when reading input
1979 files. It affects subsequent @code{INPUT} and @code{GROUP} commands.
1980 This command is like using @samp{-b @var{bfdname}} on the command line
1981 (@pxref{Options,,Command Line Options}). If the @code{TARGET} command
1982 is used but @code{OUTPUT_FORMAT} is not, then the last @code{TARGET}
1983 command is also used to set the format for the output file. @xref{BFD}.
1987 @node Miscellaneous Commands
1988 @subsection Other linker script commands
1989 There are a few other linker scripts commands.
1992 @item ASSERT(@var{exp}, @var{message})
1994 @cindex assertion in linker script
1995 Ensure that @var{exp} is non-zero. If it is zero, then exit the linker
1996 with an error code, and print @var{message}.
1998 @item EXTERN(@var{symbol} @var{symbol} @dots{})
2000 @cindex undefined symbol in linker script
2001 Force @var{symbol} to be entered in the output file as an undefined
2002 symbol. Doing this may, for example, trigger linking of additional
2003 modules from standard libraries. You may list several @var{symbol}s for
2004 each @code{EXTERN}, and you may use @code{EXTERN} multiple times. This
2005 command has the same effect as the @samp{-u} command-line option.
2007 @item FORCE_COMMON_ALLOCATION
2008 @kindex FORCE_COMMON_ALLOCATION
2009 @cindex common allocation in linker script
2010 This command has the same effect as the @samp{-d} command-line option:
2011 to make @code{ld} assign space to common symbols even if a relocatable
2012 output file is specified (@samp{-r}).
2014 @item NOCROSSREFS(@var{section} @var{section} @dots{})
2015 @kindex NOCROSSREFS(@var{sections})
2016 @cindex cross references
2017 This command may be used to tell @code{ld} to issue an error about any
2018 references among certain output sections.
2020 In certain types of programs, particularly on embedded systems when
2021 using overlays, when one section is loaded into memory, another section
2022 will not be. Any direct references between the two sections would be
2023 errors. For example, it would be an error if code in one section called
2024 a function defined in the other section.
2026 The @code{NOCROSSREFS} command takes a list of output section names. If
2027 @code{ld} detects any cross references between the sections, it reports
2028 an error and returns a non-zero exit status. Note that the
2029 @code{NOCROSSREFS} command uses output section names, not input section
2032 @ifclear SingleFormat
2033 @item OUTPUT_ARCH(@var{bfdarch})
2034 @kindex OUTPUT_ARCH(@var{bfdarch})
2035 @cindex machine architecture
2036 @cindex architecture
2037 Specify a particular output machine architecture. The argument is one
2038 of the names used by the BFD library (@pxref{BFD}). You can see the
2039 architecture of an object file by using the @code{objdump} program with
2040 the @samp{-f} option.
2045 @section Assigning Values to Symbols
2046 @cindex assignment in scripts
2047 @cindex symbol definition, scripts
2048 @cindex variables, defining
2049 You may assign a value to a symbol in a linker script. This will define
2050 the symbol as a global symbol.
2053 * Simple Assignments:: Simple Assignments
2057 @node Simple Assignments
2058 @subsection Simple Assignments
2060 You may assign to a symbol using any of the C assignment operators:
2063 @item @var{symbol} = @var{expression} ;
2064 @itemx @var{symbol} += @var{expression} ;
2065 @itemx @var{symbol} -= @var{expression} ;
2066 @itemx @var{symbol} *= @var{expression} ;
2067 @itemx @var{symbol} /= @var{expression} ;
2068 @itemx @var{symbol} <<= @var{expression} ;
2069 @itemx @var{symbol} >>= @var{expression} ;
2070 @itemx @var{symbol} &= @var{expression} ;
2071 @itemx @var{symbol} |= @var{expression} ;
2074 The first case will define @var{symbol} to the value of
2075 @var{expression}. In the other cases, @var{symbol} must already be
2076 defined, and the value will be adjusted accordingly.
2078 The special symbol name @samp{.} indicates the location counter. You
2079 may only use this within a @code{SECTIONS} command.
2081 The semicolon after @var{expression} is required.
2083 Expressions are defined below; see @ref{Expressions}.
2085 You may write symbol assignments as commands in their own right, or as
2086 statements within a @code{SECTIONS} command, or as part of an output
2087 section description in a @code{SECTIONS} command.
2089 The section of the symbol will be set from the section of the
2090 expression; for more information, see @ref{Expression Section}.
2092 Here is an example showing the three different places that symbol
2093 assignments may be used:
2104 _bdata = (. + 3) & ~ 4;
2105 .data : @{ *(.data) @}
2109 In this example, the symbol @samp{floating_point} will be defined as
2110 zero. The symbol @samp{_etext} will be defined as the address following
2111 the last @samp{.text} input section. The symbol @samp{_bdata} will be
2112 defined as the address following the @samp{.text} output section aligned
2113 upward to a 4 byte boundary.
2118 In some cases, it is desirable for a linker script to define a symbol
2119 only if it is referenced and is not defined by any object included in
2120 the link. For example, traditional linkers defined the symbol
2121 @samp{etext}. However, ANSI C requires that the user be able to use
2122 @samp{etext} as a function name without encountering an error. The
2123 @code{PROVIDE} keyword may be used to define a symbol, such as
2124 @samp{etext}, only if it is referenced but not defined. The syntax is
2125 @code{PROVIDE(@var{symbol} = @var{expression})}.
2127 Here is an example of using @code{PROVIDE} to define @samp{etext}:
2140 In this example, if the program defines @samp{_etext} (with a leading
2141 underscore), the linker will give a multiple definition error. If, on
2142 the other hand, the program defines @samp{etext} (with no leading
2143 underscore), the linker will silently use the definition in the program.
2144 If the program references @samp{etext} but does not define it, the
2145 linker will use the definition in the linker script.
2148 @section SECTIONS command
2150 The @code{SECTIONS} command tells the linker how to map input sections
2151 into output sections, and how to place the output sections in memory.
2153 The format of the @code{SECTIONS} command is:
2157 @var{sections-command}
2158 @var{sections-command}
2163 Each @var{sections-command} may of be one of the following:
2167 an @code{ENTRY} command (@pxref{Entry Point,,Entry command})
2169 a symbol assignment (@pxref{Assignments})
2171 an output section description
2173 an overlay description
2176 The @code{ENTRY} command and symbol assignments are permitted inside the
2177 @code{SECTIONS} command for convenience in using the location counter in
2178 those commands. This can also make the linker script easier to
2179 understand because you can use those commands at meaningful points in
2180 the layout of the output file.
2182 Output section descriptions and overlay descriptions are described
2185 If you do not use a @code{SECTIONS} command in your linker script, the
2186 linker will place each input section into an identically named output
2187 section in the order that the sections are first encountered in the
2188 input files. If all input sections are present in the first file, for
2189 example, the order of sections in the output file will match the order
2190 in the first input file. The first section will be at address zero.
2193 * Output Section Description:: Output section description
2194 * Output Section Name:: Output section name
2195 * Output Section Address:: Output section address
2196 * Input Section:: Input section description
2197 * Output Section Data:: Output section data
2198 * Output Section Keywords:: Output section keywords
2199 * Output Section Discarding:: Output section discarding
2200 * Output Section Attributes:: Output section attributes
2201 * Overlay Description:: Overlay description
2204 @node Output Section Description
2205 @subsection Output section description
2206 The full description of an output section looks like this:
2209 @var{section} [@var{address}] [(@var{type})] : [AT(@var{lma})]
2211 @var{output-section-command}
2212 @var{output-section-command}
2214 @} [>@var{region}] [AT>@var{lma_region}] [:@var{phdr} :@var{phdr} @dots{}] [=@var{fillexp}]
2218 Most output sections do not use most of the optional section attributes.
2220 The whitespace around @var{section} is required, so that the section
2221 name is unambiguous. The colon and the curly braces are also required.
2222 The line breaks and other white space are optional.
2224 Each @var{output-section-command} may be one of the following:
2228 a symbol assignment (@pxref{Assignments})
2230 an input section description (@pxref{Input Section})
2232 data values to include directly (@pxref{Output Section Data})
2234 a special output section keyword (@pxref{Output Section Keywords})
2237 @node Output Section Name
2238 @subsection Output section name
2239 @cindex name, section
2240 @cindex section name
2241 The name of the output section is @var{section}. @var{section} must
2242 meet the constraints of your output format. In formats which only
2243 support a limited number of sections, such as @code{a.out}, the name
2244 must be one of the names supported by the format (@code{a.out}, for
2245 example, allows only @samp{.text}, @samp{.data} or @samp{.bss}). If the
2246 output format supports any number of sections, but with numbers and not
2247 names (as is the case for Oasys), the name should be supplied as a
2248 quoted numeric string. A section name may consist of any sequence of
2249 characters, but a name which contains any unusual characters such as
2250 commas must be quoted.
2252 The output section name @samp{/DISCARD/} is special; @ref{Output Section
2255 @node Output Section Address
2256 @subsection Output section address
2257 @cindex address, section
2258 @cindex section address
2259 The @var{address} is an expression for the VMA (the virtual memory
2260 address) of the output section. If you do not provide @var{address},
2261 the linker will set it based on @var{region} if present, or otherwise
2262 based on the current value of the location counter.
2264 If you provide @var{address}, the address of the output section will be
2265 set to precisely that. If you provide neither @var{address} nor
2266 @var{region}, then the address of the output section will be set to the
2267 current value of the location counter aligned to the alignment
2268 requirements of the output section. The alignment requirement of the
2269 output section is the strictest alignment of any input section contained
2270 within the output section.
2274 .text . : @{ *(.text) @}
2279 .text : @{ *(.text) @}
2282 are subtly different. The first will set the address of the
2283 @samp{.text} output section to the current value of the location
2284 counter. The second will set it to the current value of the location
2285 counter aligned to the strictest alignment of a @samp{.text} input
2288 The @var{address} may be an arbitrary expression; @ref{Expressions}.
2289 For example, if you want to align the section on a 0x10 byte boundary,
2290 so that the lowest four bits of the section address are zero, you could
2291 do something like this:
2293 .text ALIGN(0x10) : @{ *(.text) @}
2296 This works because @code{ALIGN} returns the current location counter
2297 aligned upward to the specified value.
2299 Specifying @var{address} for a section will change the value of the
2303 @subsection Input section description
2304 @cindex input sections
2305 @cindex mapping input sections to output sections
2306 The most common output section command is an input section description.
2308 The input section description is the most basic linker script operation.
2309 You use output sections to tell the linker how to lay out your program
2310 in memory. You use input section descriptions to tell the linker how to
2311 map the input files into your memory layout.
2314 * Input Section Basics:: Input section basics
2315 * Input Section Wildcards:: Input section wildcard patterns
2316 * Input Section Common:: Input section for common symbols
2317 * Input Section Keep:: Input section and garbage collection
2318 * Input Section Example:: Input section example
2321 @node Input Section Basics
2322 @subsubsection Input section basics
2323 @cindex input section basics
2324 An input section description consists of a file name optionally followed
2325 by a list of section names in parentheses.
2327 The file name and the section name may be wildcard patterns, which we
2328 describe further below (@pxref{Input Section Wildcards}).
2330 The most common input section description is to include all input
2331 sections with a particular name in the output section. For example, to
2332 include all input @samp{.text} sections, you would write:
2337 Here the @samp{*} is a wildcard which matches any file name. To exclude a list
2338 of files from matching the file name wildcard, EXCLUDE_FILE may be used to
2339 match all files except the ones specified in the EXCLUDE_FILE list. For
2342 (*(EXCLUDE_FILE (*crtend.o *otherfile.o) .ctors))
2344 will cause all .ctors sections from all files except @file{crtend.o} and
2345 @file{otherfile.o} to be included.
2347 There are two ways to include more than one section:
2353 The difference between these is the order in which the @samp{.text} and
2354 @samp{.rdata} input sections will appear in the output section. In the
2355 first example, they will be intermingled. In the second example, all
2356 @samp{.text} input sections will appear first, followed by all
2357 @samp{.rdata} input sections.
2359 You can specify a file name to include sections from a particular file.
2360 You would do this if one or more of your files contain special data that
2361 needs to be at a particular location in memory. For example:
2366 If you use a file name without a list of sections, then all sections in
2367 the input file will be included in the output section. This is not
2368 commonly done, but it may by useful on occasion. For example:
2373 When you use a file name which does not contain any wild card
2374 characters, the linker will first see if you also specified the file
2375 name on the linker command line or in an @code{INPUT} command. If you
2376 did not, the linker will attempt to open the file as an input file, as
2377 though it appeared on the command line. Note that this differs from an
2378 @code{INPUT} command, because the linker will not search for the file in
2379 the archive search path.
2381 @node Input Section Wildcards
2382 @subsubsection Input section wildcard patterns
2383 @cindex input section wildcards
2384 @cindex wildcard file name patterns
2385 @cindex file name wildcard patterns
2386 @cindex section name wildcard patterns
2387 In an input section description, either the file name or the section
2388 name or both may be wildcard patterns.
2390 The file name of @samp{*} seen in many examples is a simple wildcard
2391 pattern for the file name.
2393 The wildcard patterns are like those used by the Unix shell.
2397 matches any number of characters
2399 matches any single character
2401 matches a single instance of any of the @var{chars}; the @samp{-}
2402 character may be used to specify a range of characters, as in
2403 @samp{[a-z]} to match any lower case letter
2405 quotes the following character
2408 When a file name is matched with a wildcard, the wildcard characters
2409 will not match a @samp{/} character (used to separate directory names on
2410 Unix). A pattern consisting of a single @samp{*} character is an
2411 exception; it will always match any file name, whether it contains a
2412 @samp{/} or not. In a section name, the wildcard characters will match
2413 a @samp{/} character.
2415 File name wildcard patterns only match files which are explicitly
2416 specified on the command line or in an @code{INPUT} command. The linker
2417 does not search directories to expand wildcards.
2419 If a file name matches more than one wildcard pattern, or if a file name
2420 appears explicitly and is also matched by a wildcard pattern, the linker
2421 will use the first match in the linker script. For example, this
2422 sequence of input section descriptions is probably in error, because the
2423 @file{data.o} rule will not be used:
2425 .data : @{ *(.data) @}
2426 .data1 : @{ data.o(.data) @}
2430 Normally, the linker will place files and sections matched by wildcards
2431 in the order in which they are seen during the link. You can change
2432 this by using the @code{SORT} keyword, which appears before a wildcard
2433 pattern in parentheses (e.g., @code{SORT(.text*)}). When the
2434 @code{SORT} keyword is used, the linker will sort the files or sections
2435 into ascending order by name before placing them in the output file.
2437 If you ever get confused about where input sections are going, use the
2438 @samp{-M} linker option to generate a map file. The map file shows
2439 precisely how input sections are mapped to output sections.
2441 This example shows how wildcard patterns might be used to partition
2442 files. This linker script directs the linker to place all @samp{.text}
2443 sections in @samp{.text} and all @samp{.bss} sections in @samp{.bss}.
2444 The linker will place the @samp{.data} section from all files beginning
2445 with an upper case character in @samp{.DATA}; for all other files, the
2446 linker will place the @samp{.data} section in @samp{.data}.
2450 .text : @{ *(.text) @}
2451 .DATA : @{ [A-Z]*(.data) @}
2452 .data : @{ *(.data) @}
2453 .bss : @{ *(.bss) @}
2458 @node Input Section Common
2459 @subsubsection Input section for common symbols
2460 @cindex common symbol placement
2461 @cindex uninitialized data placement
2462 A special notation is needed for common symbols, because in many object
2463 file formats common symbols do not have a particular input section. The
2464 linker treats common symbols as though they are in an input section
2465 named @samp{COMMON}.
2467 You may use file names with the @samp{COMMON} section just as with any
2468 other input sections. You can use this to place common symbols from a
2469 particular input file in one section while common symbols from other
2470 input files are placed in another section.
2472 In most cases, common symbols in input files will be placed in the
2473 @samp{.bss} section in the output file. For example:
2475 .bss @{ *(.bss) *(COMMON) @}
2478 @cindex scommon section
2479 @cindex small common symbols
2480 Some object file formats have more than one type of common symbol. For
2481 example, the MIPS ELF object file format distinguishes standard common
2482 symbols and small common symbols. In this case, the linker will use a
2483 different special section name for other types of common symbols. In
2484 the case of MIPS ELF, the linker uses @samp{COMMON} for standard common
2485 symbols and @samp{.scommon} for small common symbols. This permits you
2486 to map the different types of common symbols into memory at different
2490 You will sometimes see @samp{[COMMON]} in old linker scripts. This
2491 notation is now considered obsolete. It is equivalent to
2494 @node Input Section Keep
2495 @subsubsection Input section and garbage collection
2497 @cindex garbage collection
2498 When link-time garbage collection is in use (@samp{--gc-sections}),
2499 it is often useful to mark sections that should not be eliminated.
2500 This is accomplished by surrounding an input section's wildcard entry
2501 with @code{KEEP()}, as in @code{KEEP(*(.init))} or
2502 @code{KEEP(SORT(*)(.ctors))}.
2504 @node Input Section Example
2505 @subsubsection Input section example
2506 The following example is a complete linker script. It tells the linker
2507 to read all of the sections from file @file{all.o} and place them at the
2508 start of output section @samp{outputa} which starts at location
2509 @samp{0x10000}. All of section @samp{.input1} from file @file{foo.o}
2510 follows immediately, in the same output section. All of section
2511 @samp{.input2} from @file{foo.o} goes into output section
2512 @samp{outputb}, followed by section @samp{.input1} from @file{foo1.o}.
2513 All of the remaining @samp{.input1} and @samp{.input2} sections from any
2514 files are written to output section @samp{outputc}.
2538 @node Output Section Data
2539 @subsection Output section data
2541 @cindex section data
2542 @cindex output section data
2543 @kindex BYTE(@var{expression})
2544 @kindex SHORT(@var{expression})
2545 @kindex LONG(@var{expression})
2546 @kindex QUAD(@var{expression})
2547 @kindex SQUAD(@var{expression})
2548 You can include explicit bytes of data in an output section by using
2549 @code{BYTE}, @code{SHORT}, @code{LONG}, @code{QUAD}, or @code{SQUAD} as
2550 an output section command. Each keyword is followed by an expression in
2551 parentheses providing the value to store (@pxref{Expressions}). The
2552 value of the expression is stored at the current value of the location
2555 The @code{BYTE}, @code{SHORT}, @code{LONG}, and @code{QUAD} commands
2556 store one, two, four, and eight bytes (respectively). After storing the
2557 bytes, the location counter is incremented by the number of bytes
2560 For example, this will store the byte 1 followed by the four byte value
2561 of the symbol @samp{addr}:
2567 When using a 64 bit host or target, @code{QUAD} and @code{SQUAD} are the
2568 same; they both store an 8 byte, or 64 bit, value. When both host and
2569 target are 32 bits, an expression is computed as 32 bits. In this case
2570 @code{QUAD} stores a 32 bit value zero extended to 64 bits, and
2571 @code{SQUAD} stores a 32 bit value sign extended to 64 bits.
2573 If the object file format of the output file has an explicit endianness,
2574 which is the normal case, the value will be stored in that endianness.
2575 When the object file format does not have an explicit endianness, as is
2576 true of, for example, S-records, the value will be stored in the
2577 endianness of the first input object file.
2579 Note - these commands only work inside a section description and not
2580 between them, so the following will produce an error from the linker:
2582 SECTIONS @{@ .text : @{@ *(.text) @}@ LONG(1) .data : @{@ *(.data) @}@ @}@
2584 whereas this will work:
2586 SECTIONS @{@ .text : @{@ *(.text) ; LONG(1) @}@ .data : @{@ *(.data) @}@ @}@
2589 @kindex FILL(@var{expression})
2590 @cindex holes, filling
2591 @cindex unspecified memory
2592 You may use the @code{FILL} command to set the fill pattern for the
2593 current section. It is followed by an expression in parentheses. Any
2594 otherwise unspecified regions of memory within the section (for example,
2595 gaps left due to the required alignment of input sections) are filled
2596 with the two least significant bytes of the expression, repeated as
2597 necessary. A @code{FILL} statement covers memory locations after the
2598 point at which it occurs in the section definition; by including more
2599 than one @code{FILL} statement, you can have different fill patterns in
2600 different parts of an output section.
2602 This example shows how to fill unspecified regions of memory with the
2603 value @samp{0x9090}:
2608 The @code{FILL} command is similar to the @samp{=@var{fillexp}} output
2609 section attribute (@pxref{Output Section Fill}), but it only affects the
2610 part of the section following the @code{FILL} command, rather than the
2611 entire section. If both are used, the @code{FILL} command takes
2614 @node Output Section Keywords
2615 @subsection Output section keywords
2616 There are a couple of keywords which can appear as output section
2620 @kindex CREATE_OBJECT_SYMBOLS
2621 @cindex input filename symbols
2622 @cindex filename symbols
2623 @item CREATE_OBJECT_SYMBOLS
2624 The command tells the linker to create a symbol for each input file.
2625 The name of each symbol will be the name of the corresponding input
2626 file. The section of each symbol will be the output section in which
2627 the @code{CREATE_OBJECT_SYMBOLS} command appears.
2629 This is conventional for the a.out object file format. It is not
2630 normally used for any other object file format.
2632 @kindex CONSTRUCTORS
2633 @cindex C++ constructors, arranging in link
2634 @cindex constructors, arranging in link
2636 When linking using the a.out object file format, the linker uses an
2637 unusual set construct to support C++ global constructors and
2638 destructors. When linking object file formats which do not support
2639 arbitrary sections, such as ECOFF and XCOFF, the linker will
2640 automatically recognize C++ global constructors and destructors by name.
2641 For these object file formats, the @code{CONSTRUCTORS} command tells the
2642 linker to place constructor information in the output section where the
2643 @code{CONSTRUCTORS} command appears. The @code{CONSTRUCTORS} command is
2644 ignored for other object file formats.
2646 The symbol @w{@code{__CTOR_LIST__}} marks the start of the global
2647 constructors, and the symbol @w{@code{__DTOR_LIST}} marks the end. The
2648 first word in the list is the number of entries, followed by the address
2649 of each constructor or destructor, followed by a zero word. The
2650 compiler must arrange to actually run the code. For these object file
2651 formats @sc{gnu} C++ normally calls constructors from a subroutine
2652 @code{__main}; a call to @code{__main} is automatically inserted into
2653 the startup code for @code{main}. @sc{gnu} C++ normally runs
2654 destructors either by using @code{atexit}, or directly from the function
2657 For object file formats such as @code{COFF} or @code{ELF} which support
2658 arbitrary section names, @sc{gnu} C++ will normally arrange to put the
2659 addresses of global constructors and destructors into the @code{.ctors}
2660 and @code{.dtors} sections. Placing the following sequence into your
2661 linker script will build the sort of table which the @sc{gnu} C++
2662 runtime code expects to see.
2666 LONG((__CTOR_END__ - __CTOR_LIST__) / 4 - 2)
2671 LONG((__DTOR_END__ - __DTOR_LIST__) / 4 - 2)
2677 If you are using the @sc{gnu} C++ support for initialization priority,
2678 which provides some control over the order in which global constructors
2679 are run, you must sort the constructors at link time to ensure that they
2680 are executed in the correct order. When using the @code{CONSTRUCTORS}
2681 command, use @samp{SORT(CONSTRUCTORS)} instead. When using the
2682 @code{.ctors} and @code{.dtors} sections, use @samp{*(SORT(.ctors))} and
2683 @samp{*(SORT(.dtors))} instead of just @samp{*(.ctors)} and
2686 Normally the compiler and linker will handle these issues automatically,
2687 and you will not need to concern yourself with them. However, you may
2688 need to consider this if you are using C++ and writing your own linker
2693 @node Output Section Discarding
2694 @subsection Output section discarding
2695 @cindex discarding sections
2696 @cindex sections, discarding
2697 @cindex removing sections
2698 The linker will not create output section which do not have any
2699 contents. This is for convenience when referring to input sections that
2700 may or may not be present in any of the input files. For example:
2705 will only create a @samp{.foo} section in the output file if there is a
2706 @samp{.foo} section in at least one input file.
2708 If you use anything other than an input section description as an output
2709 section command, such as a symbol assignment, then the output section
2710 will always be created, even if there are no matching input sections.
2713 The special output section name @samp{/DISCARD/} may be used to discard
2714 input sections. Any input sections which are assigned to an output
2715 section named @samp{/DISCARD/} are not included in the output file.
2717 @node Output Section Attributes
2718 @subsection Output section attributes
2719 @cindex output section attributes
2720 We showed above that the full description of an output section looked
2724 @var{section} [@var{address}] [(@var{type})] : [AT(@var{lma})]
2726 @var{output-section-command}
2727 @var{output-section-command}
2729 @} [>@var{region}] [AT>@var{lma_region}] [:@var{phdr} :@var{phdr} @dots{}] [=@var{fillexp}]
2732 We've already described @var{section}, @var{address}, and
2733 @var{output-section-command}. In this section we will describe the
2734 remaining section attributes.
2737 * Output Section Type:: Output section type
2738 * Output Section LMA:: Output section LMA
2739 * Output Section Region:: Output section region
2740 * Output Section Phdr:: Output section phdr
2741 * Output Section Fill:: Output section fill
2744 @node Output Section Type
2745 @subsubsection Output section type
2746 Each output section may have a type. The type is a keyword in
2747 parentheses. The following types are defined:
2751 The section should be marked as not loadable, so that it will not be
2752 loaded into memory when the program is run.
2757 These type names are supported for backward compatibility, and are
2758 rarely used. They all have the same effect: the section should be
2759 marked as not allocatable, so that no memory is allocated for the
2760 section when the program is run.
2764 @cindex prevent unnecessary loading
2765 @cindex loading, preventing
2766 The linker normally sets the attributes of an output section based on
2767 the input sections which map into it. You can override this by using
2768 the section type. For example, in the script sample below, the
2769 @samp{ROM} section is addressed at memory location @samp{0} and does not
2770 need to be loaded when the program is run. The contents of the
2771 @samp{ROM} section will appear in the linker output file as usual.
2775 ROM 0 (NOLOAD) : @{ @dots{} @}
2781 @node Output Section LMA
2782 @subsubsection Output section LMA
2783 @kindex AT>@var{lma_region}
2784 @kindex AT(@var{lma})
2785 @cindex load address
2786 @cindex section load address
2787 Every section has a virtual address (VMA) and a load address (LMA); see
2788 @ref{Basic Script Concepts}. The address expression which may appear in
2789 an output section description sets the VMA (@pxref{Output Section
2792 The linker will normally set the LMA equal to the VMA. You can change
2793 that by using the @code{AT} keyword. The expression @var{lma} that
2794 follows the @code{AT} keyword specifies the load address of the
2795 section. Alternatively, with @samp{AT>@var{lma_region}} expression,
2796 you may specify a memory region for the section's load address. @xref{MEMORY}.
2798 @cindex ROM initialized data
2799 @cindex initialized data in ROM
2800 This feature is designed to make it easy to build a ROM image. For
2801 example, the following linker script creates three output sections: one
2802 called @samp{.text}, which starts at @code{0x1000}, one called
2803 @samp{.mdata}, which is loaded at the end of the @samp{.text} section
2804 even though its VMA is @code{0x2000}, and one called @samp{.bss} to hold
2805 uninitialized data at address @code{0x3000}. The symbol @code{_data} is
2806 defined with the value @code{0x2000}, which shows that the location
2807 counter holds the VMA value, not the LMA value.
2813 .text 0x1000 : @{ *(.text) _etext = . ; @}
2815 AT ( ADDR (.text) + SIZEOF (.text) )
2816 @{ _data = . ; *(.data); _edata = . ; @}
2818 @{ _bstart = . ; *(.bss) *(COMMON) ; _bend = . ;@}
2823 The run-time initialization code for use with a program generated with
2824 this linker script would include something like the following, to copy
2825 the initialized data from the ROM image to its runtime address. Notice
2826 how this code takes advantage of the symbols defined by the linker
2831 extern char _etext, _data, _edata, _bstart, _bend;
2832 char *src = &_etext;
2835 /* ROM has data at end of text; copy it. */
2836 while (dst < &_edata) @{
2841 for (dst = &_bstart; dst< &_bend; dst++)
2846 @node Output Section Region
2847 @subsubsection Output section region
2848 @kindex >@var{region}
2849 @cindex section, assigning to memory region
2850 @cindex memory regions and sections
2851 You can assign a section to a previously defined region of memory by
2852 using @samp{>@var{region}}. @xref{MEMORY}.
2854 Here is a simple example:
2857 MEMORY @{ rom : ORIGIN = 0x1000, LENGTH = 0x1000 @}
2858 SECTIONS @{ ROM : @{ *(.text) @} >rom @}
2862 @node Output Section Phdr
2863 @subsubsection Output section phdr
2865 @cindex section, assigning to program header
2866 @cindex program headers and sections
2867 You can assign a section to a previously defined program segment by
2868 using @samp{:@var{phdr}}. @xref{PHDRS}. If a section is assigned to
2869 one or more segments, then all subsequent allocated sections will be
2870 assigned to those segments as well, unless they use an explicitly
2871 @code{:@var{phdr}} modifier. You can use @code{:NONE} to tell the
2872 linker to not put the section in any segment at all.
2874 Here is a simple example:
2877 PHDRS @{ text PT_LOAD ; @}
2878 SECTIONS @{ .text : @{ *(.text) @} :text @}
2882 @node Output Section Fill
2883 @subsubsection Output section fill
2884 @kindex =@var{fillexp}
2885 @cindex section fill pattern
2886 @cindex fill pattern, entire section
2887 You can set the fill pattern for an entire section by using
2888 @samp{=@var{fillexp}}. @var{fillexp} is an expression
2889 (@pxref{Expressions}). Any otherwise unspecified regions of memory
2890 within the output section (for example, gaps left due to the required
2891 alignment of input sections) will be filled with the two least
2892 significant bytes of the value, repeated as necessary.
2894 You can also change the fill value with a @code{FILL} command in the
2895 output section commands; see @ref{Output Section Data}.
2897 Here is a simple example:
2900 SECTIONS @{ .text : @{ *(.text) @} =0x9090 @}
2904 @node Overlay Description
2905 @subsection Overlay description
2908 An overlay description provides an easy way to describe sections which
2909 are to be loaded as part of a single memory image but are to be run at
2910 the same memory address. At run time, some sort of overlay manager will
2911 copy the overlaid sections in and out of the runtime memory address as
2912 required, perhaps by simply manipulating addressing bits. This approach
2913 can be useful, for example, when a certain region of memory is faster
2916 Overlays are described using the @code{OVERLAY} command. The
2917 @code{OVERLAY} command is used within a @code{SECTIONS} command, like an
2918 output section description. The full syntax of the @code{OVERLAY}
2919 command is as follows:
2922 OVERLAY [@var{start}] : [NOCROSSREFS] [AT ( @var{ldaddr} )]
2926 @var{output-section-command}
2927 @var{output-section-command}
2929 @} [:@var{phdr}@dots{}] [=@var{fill}]
2932 @var{output-section-command}
2933 @var{output-section-command}
2935 @} [:@var{phdr}@dots{}] [=@var{fill}]
2937 @} [>@var{region}] [:@var{phdr}@dots{}] [=@var{fill}]
2941 Everything is optional except @code{OVERLAY} (a keyword), and each
2942 section must have a name (@var{secname1} and @var{secname2} above). The
2943 section definitions within the @code{OVERLAY} construct are identical to
2944 those within the general @code{SECTIONS} contruct (@pxref{SECTIONS}),
2945 except that no addresses and no memory regions may be defined for
2946 sections within an @code{OVERLAY}.
2948 The sections are all defined with the same starting address. The load
2949 addresses of the sections are arranged such that they are consecutive in
2950 memory starting at the load address used for the @code{OVERLAY} as a
2951 whole (as with normal section definitions, the load address is optional,
2952 and defaults to the start address; the start address is also optional,
2953 and defaults to the current value of the location counter).
2955 If the @code{NOCROSSREFS} keyword is used, and there any references
2956 among the sections, the linker will report an error. Since the sections
2957 all run at the same address, it normally does not make sense for one
2958 section to refer directly to another. @xref{Miscellaneous Commands,
2961 For each section within the @code{OVERLAY}, the linker automatically
2962 defines two symbols. The symbol @code{__load_start_@var{secname}} is
2963 defined as the starting load address of the section. The symbol
2964 @code{__load_stop_@var{secname}} is defined as the final load address of
2965 the section. Any characters within @var{secname} which are not legal
2966 within C identifiers are removed. C (or assembler) code may use these
2967 symbols to move the overlaid sections around as necessary.
2969 At the end of the overlay, the value of the location counter is set to
2970 the start address of the overlay plus the size of the largest section.
2972 Here is an example. Remember that this would appear inside a
2973 @code{SECTIONS} construct.
2976 OVERLAY 0x1000 : AT (0x4000)
2978 .text0 @{ o1/*.o(.text) @}
2979 .text1 @{ o2/*.o(.text) @}
2984 This will define both @samp{.text0} and @samp{.text1} to start at
2985 address 0x1000. @samp{.text0} will be loaded at address 0x4000, and
2986 @samp{.text1} will be loaded immediately after @samp{.text0}. The
2987 following symbols will be defined: @code{__load_start_text0},
2988 @code{__load_stop_text0}, @code{__load_start_text1},
2989 @code{__load_stop_text1}.
2991 C code to copy overlay @code{.text1} into the overlay area might look
2996 extern char __load_start_text1, __load_stop_text1;
2997 memcpy ((char *) 0x1000, &__load_start_text1,
2998 &__load_stop_text1 - &__load_start_text1);
3002 Note that the @code{OVERLAY} command is just syntactic sugar, since
3003 everything it does can be done using the more basic commands. The above
3004 example could have been written identically as follows.
3008 .text0 0x1000 : AT (0x4000) @{ o1/*.o(.text) @}
3009 __load_start_text0 = LOADADDR (.text0);
3010 __load_stop_text0 = LOADADDR (.text0) + SIZEOF (.text0);
3011 .text1 0x1000 : AT (0x4000 + SIZEOF (.text0)) @{ o2/*.o(.text) @}
3012 __load_start_text1 = LOADADDR (.text1);
3013 __load_stop_text1 = LOADADDR (.text1) + SIZEOF (.text1);
3014 . = 0x1000 + MAX (SIZEOF (.text0), SIZEOF (.text1));
3019 @section MEMORY command
3021 @cindex memory regions
3022 @cindex regions of memory
3023 @cindex allocating memory
3024 @cindex discontinuous memory
3025 The linker's default configuration permits allocation of all available
3026 memory. You can override this by using the @code{MEMORY} command.
3028 The @code{MEMORY} command describes the location and size of blocks of
3029 memory in the target. You can use it to describe which memory regions
3030 may be used by the linker, and which memory regions it must avoid. You
3031 can then assign sections to particular memory regions. The linker will
3032 set section addresses based on the memory regions, and will warn about
3033 regions that become too full. The linker will not shuffle sections
3034 around to fit into the available regions.
3036 A linker script may contain at most one use of the @code{MEMORY}
3037 command. However, you can define as many blocks of memory within it as
3038 you wish. The syntax is:
3043 @var{name} [(@var{attr})] : ORIGIN = @var{origin}, LENGTH = @var{len}
3049 The @var{name} is a name used in the linker script to refer to the
3050 region. The region name has no meaning outside of the linker script.
3051 Region names are stored in a separate name space, and will not conflict
3052 with symbol names, file names, or section names. Each memory region
3053 must have a distinct name.
3055 @cindex memory region attributes
3056 The @var{attr} string is an optional list of attributes that specify
3057 whether to use a particular memory region for an input section which is
3058 not explicitly mapped in the linker script. As described in
3059 @ref{SECTIONS}, if you do not specify an output section for some input
3060 section, the linker will create an output section with the same name as
3061 the input section. If you define region attributes, the linker will use
3062 them to select the memory region for the output section that it creates.
3064 The @var{attr} string must consist only of the following characters:
3079 Invert the sense of any of the preceding attributes
3082 If a unmapped section matches any of the listed attributes other than
3083 @samp{!}, it will be placed in the memory region. The @samp{!}
3084 attribute reverses this test, so that an unmapped section will be placed
3085 in the memory region only if it does not match any of the listed
3091 The @var{origin} is an expression for the start address of the memory
3092 region. The expression must evaluate to a constant before memory
3093 allocation is performed, which means that you may not use any section
3094 relative symbols. The keyword @code{ORIGIN} may be abbreviated to
3095 @code{org} or @code{o} (but not, for example, @code{ORG}).
3100 The @var{len} is an expression for the size in bytes of the memory
3101 region. As with the @var{origin} expression, the expression must
3102 evaluate to a constant before memory allocation is performed. The
3103 keyword @code{LENGTH} may be abbreviated to @code{len} or @code{l}.
3105 In the following example, we specify that there are two memory regions
3106 available for allocation: one starting at @samp{0} for 256 kilobytes,
3107 and the other starting at @samp{0x40000000} for four megabytes. The
3108 linker will place into the @samp{rom} memory region every section which
3109 is not explicitly mapped into a memory region, and is either read-only
3110 or executable. The linker will place other sections which are not
3111 explicitly mapped into a memory region into the @samp{ram} memory
3118 rom (rx) : ORIGIN = 0, LENGTH = 256K
3119 ram (!rx) : org = 0x40000000, l = 4M
3124 Once you define a memory region, you can direct the linker to place
3125 specific output sections into that memory region by using the
3126 @samp{>@var{region}} output section attribute. For example, if you have
3127 a memory region named @samp{mem}, you would use @samp{>mem} in the
3128 output section definition. @xref{Output Section Region}. If no address
3129 was specified for the output section, the linker will set the address to
3130 the next available address within the memory region. If the combined
3131 output sections directed to a memory region are too large for the
3132 region, the linker will issue an error message.
3135 @section PHDRS Command
3137 @cindex program headers
3138 @cindex ELF program headers
3139 @cindex program segments
3140 @cindex segments, ELF
3141 The ELF object file format uses @dfn{program headers}, also knows as
3142 @dfn{segments}. The program headers describe how the program should be
3143 loaded into memory. You can print them out by using the @code{objdump}
3144 program with the @samp{-p} option.
3146 When you run an ELF program on a native ELF system, the system loader
3147 reads the program headers in order to figure out how to load the
3148 program. This will only work if the program headers are set correctly.
3149 This manual does not describe the details of how the system loader
3150 interprets program headers; for more information, see the ELF ABI.
3152 The linker will create reasonable program headers by default. However,
3153 in some cases, you may need to specify the program headers more
3154 precisely. You may use the @code{PHDRS} command for this purpose. When
3155 the linker sees the @code{PHDRS} command in the linker script, it will
3156 not create any program headers other than the ones specified.
3158 The linker only pays attention to the @code{PHDRS} command when
3159 generating an ELF output file. In other cases, the linker will simply
3160 ignore @code{PHDRS}.
3162 This is the syntax of the @code{PHDRS} command. The words @code{PHDRS},
3163 @code{FILEHDR}, @code{AT}, and @code{FLAGS} are keywords.
3169 @var{name} @var{type} [ FILEHDR ] [ PHDRS ] [ AT ( @var{address} ) ]
3170 [ FLAGS ( @var{flags} ) ] ;
3175 The @var{name} is used only for reference in the @code{SECTIONS} command
3176 of the linker script. It is not put into the output file. Program
3177 header names are stored in a separate name space, and will not conflict
3178 with symbol names, file names, or section names. Each program header
3179 must have a distinct name.
3181 Certain program header types describe segments of memory which the
3182 system loader will load from the file. In the linker script, you
3183 specify the contents of these segments by placing allocatable output
3184 sections in the segments. You use the @samp{:@var{phdr}} output section
3185 attribute to place a section in a particular segment. @xref{Output
3188 It is normal to put certain sections in more than one segment. This
3189 merely implies that one segment of memory contains another. You may
3190 repeat @samp{:@var{phdr}}, using it once for each segment which should
3191 contain the section.
3193 If you place a section in one or more segments using @samp{:@var{phdr}},
3194 then the linker will place all subsequent allocatable sections which do
3195 not specify @samp{:@var{phdr}} in the same segments. This is for
3196 convenience, since generally a whole set of contiguous sections will be
3197 placed in a single segment. You can use @code{:NONE} to override the
3198 default segment and tell the linker to not put the section in any
3203 You may use the @code{FILEHDR} and @code{PHDRS} keywords appear after
3204 the program header type to further describe the contents of the segment.
3205 The @code{FILEHDR} keyword means that the segment should include the ELF
3206 file header. The @code{PHDRS} keyword means that the segment should
3207 include the ELF program headers themselves.
3209 The @var{type} may be one of the following. The numbers indicate the
3210 value of the keyword.
3213 @item @code{PT_NULL} (0)
3214 Indicates an unused program header.
3216 @item @code{PT_LOAD} (1)
3217 Indicates that this program header describes a segment to be loaded from
3220 @item @code{PT_DYNAMIC} (2)
3221 Indicates a segment where dynamic linking information can be found.
3223 @item @code{PT_INTERP} (3)
3224 Indicates a segment where the name of the program interpreter may be
3227 @item @code{PT_NOTE} (4)
3228 Indicates a segment holding note information.
3230 @item @code{PT_SHLIB} (5)
3231 A reserved program header type, defined but not specified by the ELF
3234 @item @code{PT_PHDR} (6)
3235 Indicates a segment where the program headers may be found.
3237 @item @var{expression}
3238 An expression giving the numeric type of the program header. This may
3239 be used for types not defined above.
3242 You can specify that a segment should be loaded at a particular address
3243 in memory by using an @code{AT} expression. This is identical to the
3244 @code{AT} command used as an output section attribute (@pxref{Output
3245 Section LMA}). The @code{AT} command for a program header overrides the
3246 output section attribute.
3248 The linker will normally set the segment flags based on the sections
3249 which comprise the segment. You may use the @code{FLAGS} keyword to
3250 explicitly specify the segment flags. The value of @var{flags} must be
3251 an integer. It is used to set the @code{p_flags} field of the program
3254 Here is an example of @code{PHDRS}. This shows a typical set of program
3255 headers used on a native ELF system.
3261 headers PT_PHDR PHDRS ;
3263 text PT_LOAD FILEHDR PHDRS ;
3265 dynamic PT_DYNAMIC ;
3271 .interp : @{ *(.interp) @} :text :interp
3272 .text : @{ *(.text) @} :text
3273 .rodata : @{ *(.rodata) @} /* defaults to :text */
3275 . = . + 0x1000; /* move to a new page in memory */
3276 .data : @{ *(.data) @} :data
3277 .dynamic : @{ *(.dynamic) @} :data :dynamic
3284 @section VERSION Command
3285 @kindex VERSION @{script text@}
3286 @cindex symbol versions
3287 @cindex version script
3288 @cindex versions of symbols
3289 The linker supports symbol versions when using ELF. Symbol versions are
3290 only useful when using shared libraries. The dynamic linker can use
3291 symbol versions to select a specific version of a function when it runs
3292 a program that may have been linked against an earlier version of the
3295 You can include a version script directly in the main linker script, or
3296 you can supply the version script as an implicit linker script. You can
3297 also use the @samp{--version-script} linker option.
3299 The syntax of the @code{VERSION} command is simply
3301 VERSION @{ version-script-commands @}
3304 The format of the version script commands is identical to that used by
3305 Sun's linker in Solaris 2.5. The version script defines a tree of
3306 version nodes. You specify the node names and interdependencies in the
3307 version script. You can specify which symbols are bound to which
3308 version nodes, and you can reduce a specified set of symbols to local
3309 scope so that they are not globally visible outside of the shared
3312 The easiest way to demonstrate the version script language is with a few
3334 This example version script defines three version nodes. The first
3335 version node defined is @samp{VERS_1.1}; it has no other dependencies.
3336 The script binds the symbol @samp{foo1} to @samp{VERS_1.1}. It reduces
3337 a number of symbols to local scope so that they are not visible outside
3338 of the shared library.
3340 Next, the version script defines node @samp{VERS_1.2}. This node
3341 depends upon @samp{VERS_1.1}. The script binds the symbol @samp{foo2}
3342 to the version node @samp{VERS_1.2}.
3344 Finally, the version script defines node @samp{VERS_2.0}. This node
3345 depends upon @samp{VERS_1.2}. The scripts binds the symbols @samp{bar1}
3346 and @samp{bar2} are bound to the version node @samp{VERS_2.0}.
3348 When the linker finds a symbol defined in a library which is not
3349 specifically bound to a version node, it will effectively bind it to an
3350 unspecified base version of the library. You can bind all otherwise
3351 unspecified symbols to a given version node by using @samp{global: *}
3352 somewhere in the version script.
3354 The names of the version nodes have no specific meaning other than what
3355 they might suggest to the person reading them. The @samp{2.0} version
3356 could just as well have appeared in between @samp{1.1} and @samp{1.2}.
3357 However, this would be a confusing way to write a version script.
3359 When you link an application against a shared library that has versioned
3360 symbols, the application itself knows which version of each symbol it
3361 requires, and it also knows which version nodes it needs from each
3362 shared library it is linked against. Thus at runtime, the dynamic
3363 loader can make a quick check to make sure that the libraries you have
3364 linked against do in fact supply all of the version nodes that the
3365 application will need to resolve all of the dynamic symbols. In this
3366 way it is possible for the dynamic linker to know with certainty that
3367 all external symbols that it needs will be resolvable without having to
3368 search for each symbol reference.
3370 The symbol versioning is in effect a much more sophisticated way of
3371 doing minor version checking that SunOS does. The fundamental problem
3372 that is being addressed here is that typically references to external
3373 functions are bound on an as-needed basis, and are not all bound when
3374 the application starts up. If a shared library is out of date, a
3375 required interface may be missing; when the application tries to use
3376 that interface, it may suddenly and unexpectedly fail. With symbol
3377 versioning, the user will get a warning when they start their program if
3378 the libraries being used with the application are too old.
3380 There are several GNU extensions to Sun's versioning approach. The
3381 first of these is the ability to bind a symbol to a version node in the
3382 source file where the symbol is defined instead of in the versioning
3383 script. This was done mainly to reduce the burden on the library
3384 maintainer. You can do this by putting something like:
3386 __asm__(".symver original_foo,foo@@VERS_1.1");
3389 in the C source file. This renames the function @samp{original_foo} to
3390 be an alias for @samp{foo} bound to the version node @samp{VERS_1.1}.
3391 The @samp{local:} directive can be used to prevent the symbol
3392 @samp{original_foo} from being exported.
3394 The second GNU extension is to allow multiple versions of the same
3395 function to appear in a given shared library. In this way you can make
3396 an incompatible change to an interface without increasing the major
3397 version number of the shared library, while still allowing applications
3398 linked against the old interface to continue to function.
3400 To do this, you must use multiple @samp{.symver} directives in the
3401 source file. Here is an example:
3404 __asm__(".symver original_foo,foo@@");
3405 __asm__(".symver old_foo,foo@@VERS_1.1");
3406 __asm__(".symver old_foo1,foo@@VERS_1.2");
3407 __asm__(".symver new_foo,foo@@@@VERS_2.0");
3410 In this example, @samp{foo@@} represents the symbol @samp{foo} bound to the
3411 unspecified base version of the symbol. The source file that contains this
3412 example would define 4 C functions: @samp{original_foo}, @samp{old_foo},
3413 @samp{old_foo1}, and @samp{new_foo}.
3415 When you have multiple definitions of a given symbol, there needs to be
3416 some way to specify a default version to which external references to
3417 this symbol will be bound. You can do this with the
3418 @samp{foo@@@@VERS_2.0} type of @samp{.symver} directive. You can only
3419 declare one version of a symbol as the default in this manner; otherwise
3420 you would effectively have multiple definitions of the same symbol.
3422 If you wish to bind a reference to a specific version of the symbol
3423 within the shared library, you can use the aliases of convenience
3424 (i.e. @samp{old_foo}), or you can use the @samp{.symver} directive to
3425 specifically bind to an external version of the function in question.
3428 @section Expressions in Linker Scripts
3431 The syntax for expressions in the linker script language is identical to
3432 that of C expressions. All expressions are evaluated as integers. All
3433 expressions are evaluated in the same size, which is 32 bits if both the
3434 host and target are 32 bits, and is otherwise 64 bits.
3436 You can use and set symbol values in expressions.
3438 The linker defines several special purpose builtin functions for use in
3442 * Constants:: Constants
3443 * Symbols:: Symbol Names
3444 * Location Counter:: The Location Counter
3445 * Operators:: Operators
3446 * Evaluation:: Evaluation
3447 * Expression Section:: The Section of an Expression
3448 * Builtin Functions:: Builtin Functions
3452 @subsection Constants
3453 @cindex integer notation
3454 @cindex constants in linker scripts
3455 All constants are integers.
3457 As in C, the linker considers an integer beginning with @samp{0} to be
3458 octal, and an integer beginning with @samp{0x} or @samp{0X} to be
3459 hexadecimal. The linker considers other integers to be decimal.
3461 @cindex scaled integers
3462 @cindex K and M integer suffixes
3463 @cindex M and K integer suffixes
3464 @cindex suffixes for integers
3465 @cindex integer suffixes
3466 In addition, you can use the suffixes @code{K} and @code{M} to scale a
3470 @c END TEXI2ROFF-KILL
3471 @code{1024} or @code{1024*1024}
3475 ${\rm 1024}$ or ${\rm 1024}^2$
3477 @c END TEXI2ROFF-KILL
3478 respectively. For example, the following all refer to the same quantity:
3486 @subsection Symbol Names
3487 @cindex symbol names
3489 @cindex quoted symbol names
3491 Unless quoted, symbol names start with a letter, underscore, or period
3492 and may include letters, digits, underscores, periods, and hyphens.
3493 Unquoted symbol names must not conflict with any keywords. You can
3494 specify a symbol which contains odd characters or has the same name as a
3495 keyword by surrounding the symbol name in double quotes:
3498 "with a space" = "also with a space" + 10;
3501 Since symbols can contain many non-alphabetic characters, it is safest
3502 to delimit symbols with spaces. For example, @samp{A-B} is one symbol,
3503 whereas @samp{A - B} is an expression involving subtraction.
3505 @node Location Counter
3506 @subsection The Location Counter
3509 @cindex location counter
3510 @cindex current output location
3511 The special linker variable @dfn{dot} @samp{.} always contains the
3512 current output location counter. Since the @code{.} always refers to a
3513 location in an output section, it may only appear in an expression
3514 within a @code{SECTIONS} command. The @code{.} symbol may appear
3515 anywhere that an ordinary symbol is allowed in an expression.
3518 Assigning a value to @code{.} will cause the location counter to be
3519 moved. This may be used to create holes in the output section. The
3520 location counter may never be moved backwards.
3536 In the previous example, the @samp{.text} section from @file{file1} is
3537 located at the beginning of the output section @samp{output}. It is
3538 followed by a 1000 byte gap. Then the @samp{.text} section from
3539 @file{file2} appears, also with a 1000 byte gap following before the
3540 @samp{.text} section from @file{file3}. The notation @samp{= 0x1234}
3541 specifies what data to write in the gaps (@pxref{Output Section Fill}).
3543 @cindex dot inside sections
3544 Note: @code{.} actually refers to the byte offset from the start of the
3545 current containing object. Normally this is the @code{SECTIONS}
3546 statement, whoes start address is 0, hence @code{.} can be used as an
3547 absolute address. If @code{.} is used inside a section description
3548 however, it refers to the byte offset from the start of that section,
3549 not an absolute address. Thus in a script like this:
3567 The @samp{.text} section will be assigned a starting address of 0x100
3568 and a size of exactly 0x200 bytes, even if there is not enough data in
3569 the @samp{.text} input sections to fill this area. (If there is too
3570 much data, an error will be produced because this would be an attempt to
3571 move @code{.} backwards). The @samp{.data} section will start at 0x500
3572 and it will have an extra 0x600 bytes worth of space after the end of
3573 the values from the @samp{.data} input sections and before the end of
3574 the @samp{.data} output section itself.
3578 @subsection Operators
3579 @cindex operators for arithmetic
3580 @cindex arithmetic operators
3581 @cindex precedence in expressions
3582 The linker recognizes the standard C set of arithmetic operators, with
3583 the standard bindings and precedence levels:
3586 @c END TEXI2ROFF-KILL
3588 precedence associativity Operators Notes
3594 5 left == != > < <= >=
3600 11 right &= += -= *= /= (2)
3604 (1) Prefix operators
3605 (2) @xref{Assignments}.
3609 \vskip \baselineskip
3610 %"lispnarrowing" is the extra indent used generally for smallexample
3611 \hskip\lispnarrowing\vbox{\offinterlineskip
3614 {\vrule#&\strut\hfil\ #\ \hfil&\vrule#&\strut\hfil\ #\ \hfil&\vrule#&\strut\hfil\ {\tt #}\ \hfil&\vrule#\cr
3615 height2pt&\omit&&\omit&&\omit&\cr
3616 &Precedence&& Associativity &&{\rm Operators}&\cr
3617 height2pt&\omit&&\omit&&\omit&\cr
3619 height2pt&\omit&&\omit&&\omit&\cr
3621 % '176 is tilde, '~' in tt font
3622 &1&&left&&\qquad- \char'176\ !\qquad\dag&\cr
3623 &2&&left&&* / \%&\cr
3626 &5&&left&&== != > < <= >=&\cr
3629 &8&&left&&{\&\&}&\cr
3632 &11&&right&&\qquad\&= += -= *= /=\qquad\ddag&\cr
3634 height2pt&\omit&&\omit&&\omit&\cr}
3639 @obeylines@parskip=0pt@parindent=0pt
3640 @dag@quad Prefix operators.
3641 @ddag@quad @xref{Assignments}.
3644 @c END TEXI2ROFF-KILL
3647 @subsection Evaluation
3648 @cindex lazy evaluation
3649 @cindex expression evaluation order
3650 The linker evaluates expressions lazily. It only computes the value of
3651 an expression when absolutely necessary.
3653 The linker needs some information, such as the value of the start
3654 address of the first section, and the origins and lengths of memory
3655 regions, in order to do any linking at all. These values are computed
3656 as soon as possible when the linker reads in the linker script.
3658 However, other values (such as symbol values) are not known or needed
3659 until after storage allocation. Such values are evaluated later, when
3660 other information (such as the sizes of output sections) is available
3661 for use in the symbol assignment expression.
3663 The sizes of sections cannot be known until after allocation, so
3664 assignments dependent upon these are not performed until after
3667 Some expressions, such as those depending upon the location counter
3668 @samp{.}, must be evaluated during section allocation.
3670 If the result of an expression is required, but the value is not
3671 available, then an error results. For example, a script like the
3677 .text 9+this_isnt_constant :
3683 will cause the error message @samp{non constant expression for initial
3686 @node Expression Section
3687 @subsection The Section of an Expression
3688 @cindex expression sections
3689 @cindex absolute expressions
3690 @cindex relative expressions
3691 @cindex absolute and relocatable symbols
3692 @cindex relocatable and absolute symbols
3693 @cindex symbols, relocatable and absolute
3694 When the linker evaluates an expression, the result is either absolute
3695 or relative to some section. A relative expression is expressed as a
3696 fixed offset from the base of a section.
3698 The position of the expression within the linker script determines
3699 whether it is absolute or relative. An expression which appears within
3700 an output section definition is relative to the base of the output
3701 section. An expression which appears elsewhere will be absolute.
3703 A symbol set to a relative expression will be relocatable if you request
3704 relocatable output using the @samp{-r} option. That means that a
3705 further link operation may change the value of the symbol. The symbol's
3706 section will be the section of the relative expression.
3708 A symbol set to an absolute expression will retain the same value
3709 through any further link operation. The symbol will be absolute, and
3710 will not have any particular associated section.
3712 You can use the builtin function @code{ABSOLUTE} to force an expression
3713 to be absolute when it would otherwise be relative. For example, to
3714 create an absolute symbol set to the address of the end of the output
3715 section @samp{.data}:
3719 .data : @{ *(.data) _edata = ABSOLUTE(.); @}
3723 If @samp{ABSOLUTE} were not used, @samp{_edata} would be relative to the
3724 @samp{.data} section.
3726 @node Builtin Functions
3727 @subsection Builtin Functions
3728 @cindex functions in expressions
3729 The linker script language includes a number of builtin functions for
3730 use in linker script expressions.
3733 @item ABSOLUTE(@var{exp})
3734 @kindex ABSOLUTE(@var{exp})
3735 @cindex expression, absolute
3736 Return the absolute (non-relocatable, as opposed to non-negative) value
3737 of the expression @var{exp}. Primarily useful to assign an absolute
3738 value to a symbol within a section definition, where symbol values are
3739 normally section relative. @xref{Expression Section}.
3741 @item ADDR(@var{section})
3742 @kindex ADDR(@var{section})
3743 @cindex section address in expression
3744 Return the absolute address (the VMA) of the named @var{section}. Your
3745 script must previously have defined the location of that section. In
3746 the following example, @code{symbol_1} and @code{symbol_2} are assigned
3753 start_of_output_1 = ABSOLUTE(.);
3758 symbol_1 = ADDR(.output1);
3759 symbol_2 = start_of_output_1;
3765 @item ALIGN(@var{exp})
3766 @kindex ALIGN(@var{exp})
3767 @cindex round up location counter
3768 @cindex align location counter
3769 Return the location counter (@code{.}) aligned to the next @var{exp}
3770 boundary. @var{exp} must be an expression whose value is a power of
3771 two. This is equivalent to
3773 (. + @var{exp} - 1) & ~(@var{exp} - 1)
3776 @code{ALIGN} doesn't change the value of the location counter---it just
3777 does arithmetic on it. Here is an example which aligns the output
3778 @code{.data} section to the next @code{0x2000} byte boundary after the
3779 preceding section and sets a variable within the section to the next
3780 @code{0x8000} boundary after the input sections:
3784 .data ALIGN(0x2000): @{
3786 variable = ALIGN(0x8000);
3792 The first use of @code{ALIGN} in this example specifies the location of
3793 a section because it is used as the optional @var{address} attribute of
3794 a section definition (@pxref{Output Section Address}). The second use
3795 of @code{ALIGN} is used to defines the value of a symbol.
3797 The builtin function @code{NEXT} is closely related to @code{ALIGN}.
3799 @item BLOCK(@var{exp})
3800 @kindex BLOCK(@var{exp})
3801 This is a synonym for @code{ALIGN}, for compatibility with older linker
3802 scripts. It is most often seen when setting the address of an output
3805 @item DEFINED(@var{symbol})
3806 @kindex DEFINED(@var{symbol})
3807 @cindex symbol defaults
3808 Return 1 if @var{symbol} is in the linker global symbol table and is
3809 defined, otherwise return 0. You can use this function to provide
3810 default values for symbols. For example, the following script fragment
3811 shows how to set a global symbol @samp{begin} to the first location in
3812 the @samp{.text} section---but if a symbol called @samp{begin} already
3813 existed, its value is preserved:
3819 begin = DEFINED(begin) ? begin : . ;
3827 @item LOADADDR(@var{section})
3828 @kindex LOADADDR(@var{section})
3829 @cindex section load address in expression
3830 Return the absolute LMA of the named @var{section}. This is normally
3831 the same as @code{ADDR}, but it may be different if the @code{AT}
3832 attribute is used in the output section definition (@pxref{Output
3836 @item MAX(@var{exp1}, @var{exp2})
3837 Returns the maximum of @var{exp1} and @var{exp2}.
3840 @item MIN(@var{exp1}, @var{exp2})
3841 Returns the minimum of @var{exp1} and @var{exp2}.
3843 @item NEXT(@var{exp})
3844 @kindex NEXT(@var{exp})
3845 @cindex unallocated address, next
3846 Return the next unallocated address that is a multiple of @var{exp}.
3847 This function is closely related to @code{ALIGN(@var{exp})}; unless you
3848 use the @code{MEMORY} command to define discontinuous memory for the
3849 output file, the two functions are equivalent.
3851 @item SIZEOF(@var{section})
3852 @kindex SIZEOF(@var{section})
3853 @cindex section size
3854 Return the size in bytes of the named @var{section}, if that section has
3855 been allocated. If the section has not been allocated when this is
3856 evaluated, the linker will report an error. In the following example,
3857 @code{symbol_1} and @code{symbol_2} are assigned identical values:
3866 symbol_1 = .end - .start ;
3867 symbol_2 = SIZEOF(.output);
3872 @item SIZEOF_HEADERS
3873 @itemx sizeof_headers
3874 @kindex SIZEOF_HEADERS
3876 Return the size in bytes of the output file's headers. This is
3877 information which appears at the start of the output file. You can use
3878 this number when setting the start address of the first section, if you
3879 choose, to facilitate paging.
3881 @cindex not enough room for program headers
3882 @cindex program headers, not enough room
3883 When producing an ELF output file, if the linker script uses the
3884 @code{SIZEOF_HEADERS} builtin function, the linker must compute the
3885 number of program headers before it has determined all the section
3886 addresses and sizes. If the linker later discovers that it needs
3887 additional program headers, it will report an error @samp{not enough
3888 room for program headers}. To avoid this error, you must avoid using
3889 the @code{SIZEOF_HEADERS} function, or you must rework your linker
3890 script to avoid forcing the linker to use additional program headers, or
3891 you must define the program headers yourself using the @code{PHDRS}
3892 command (@pxref{PHDRS}).
3895 @node Implicit Linker Scripts
3896 @section Implicit Linker Scripts
3897 @cindex implicit linker scripts
3898 If you specify a linker input file which the linker can not recognize as
3899 an object file or an archive file, it will try to read the file as a
3900 linker script. If the file can not be parsed as a linker script, the
3901 linker will report an error.
3903 An implicit linker script will not replace the default linker script.
3905 Typically an implicit linker script would contain only symbol
3906 assignments, or the @code{INPUT}, @code{GROUP}, or @code{VERSION}
3909 Any input files read because of an implicit linker script will be read
3910 at the position in the command line where the implicit linker script was
3911 read. This can affect archive searching.
3914 @node Machine Dependent
3915 @chapter Machine Dependent Features
3917 @cindex machine dependencies
3918 @code{ld} has additional features on some platforms; the following
3919 sections describe them. Machines where @code{ld} has no additional
3920 functionality are not listed.
3923 * H8/300:: @code{ld} and the H8/300
3924 * i960:: @code{ld} and the Intel 960 family
3925 * ARM:: @code{ld} and the ARM family
3926 * HPPA ELF32:: @code{ld} and HPPA 32-bit ELF
3928 * TI COFF:: @code{ld} and TI COFF
3933 @c FIXME! This could use @raisesections/@lowersections, but there seems to be a conflict
3934 @c between those and node-defaulting.
3941 @section @code{ld} and the H8/300
3943 @cindex H8/300 support
3944 For the H8/300, @code{ld} can perform these global optimizations when
3945 you specify the @samp{--relax} command-line option.
3948 @cindex relaxing on H8/300
3949 @item relaxing address modes
3950 @code{ld} finds all @code{jsr} and @code{jmp} instructions whose
3951 targets are within eight bits, and turns them into eight-bit
3952 program-counter relative @code{bsr} and @code{bra} instructions,
3955 @cindex synthesizing on H8/300
3956 @item synthesizing instructions
3957 @c FIXME: specifically mov.b, or any mov instructions really?
3958 @code{ld} finds all @code{mov.b} instructions which use the
3959 sixteen-bit absolute address form, but refer to the top
3960 page of memory, and changes them to use the eight-bit address form.
3961 (That is: the linker turns @samp{mov.b @code{@@}@var{aa}:16} into
3962 @samp{mov.b @code{@@}@var{aa}:8} whenever the address @var{aa} is in the
3963 top page of memory).
3973 @c This stuff is pointless to say unless you're especially concerned
3974 @c with Hitachi chips; don't enable it for generic case, please.
3976 @chapter @code{ld} and other Hitachi chips
3978 @code{ld} also supports the H8/300H, the H8/500, and the Hitachi SH. No
3979 special features, commands, or command-line options are required for
3990 @section @code{ld} and the Intel 960 family
3992 @cindex i960 support
3994 You can use the @samp{-A@var{architecture}} command line option to
3995 specify one of the two-letter names identifying members of the 960
3996 family; the option specifies the desired output target, and warns of any
3997 incompatible instructions in the input files. It also modifies the
3998 linker's search strategy for archive libraries, to support the use of
3999 libraries specific to each particular architecture, by including in the
4000 search loop names suffixed with the string identifying the architecture.
4002 For example, if your @code{ld} command line included @w{@samp{-ACA}} as
4003 well as @w{@samp{-ltry}}, the linker would look (in its built-in search
4004 paths, and in any paths you specify with @samp{-L}) for a library with
4017 The first two possibilities would be considered in any event; the last
4018 two are due to the use of @w{@samp{-ACA}}.
4020 You can meaningfully use @samp{-A} more than once on a command line, since
4021 the 960 architecture family allows combination of target architectures; each
4022 use will add another pair of name variants to search for when @w{@samp{-l}}
4023 specifies a library.
4025 @cindex @code{--relax} on i960
4026 @cindex relaxing on i960
4027 @code{ld} supports the @samp{--relax} option for the i960 family. If
4028 you specify @samp{--relax}, @code{ld} finds all @code{balx} and
4029 @code{calx} instructions whose targets are within 24 bits, and turns
4030 them into 24-bit program-counter relative @code{bal} and @code{cal}
4031 instructions, respectively. @code{ld} also turns @code{cal}
4032 instructions into @code{bal} instructions when it determines that the
4033 target subroutine is a leaf routine (that is, the target subroutine does
4034 not itself call any subroutines).
4046 @section @code{ld}'s support for interworking between ARM and Thumb code
4048 @cindex ARM interworking support
4049 @kindex --support-old-code
4050 For the ARM, @code{ld} will generate code stubs to allow functions calls
4051 betweem ARM and Thumb code. These stubs only work with code that has
4052 been compiled and assembled with the @samp{-mthumb-interwork} command
4053 line option. If it is necessary to link with old ARM object files or
4054 libraries, which have not been compiled with the -mthumb-interwork
4055 option then the @samp{--support-old-code} command line switch should be
4056 given to the linker. This will make it generate larger stub functions
4057 which will work with non-interworking aware ARM code. Note, however,
4058 the linker does not support generating stubs for function calls to
4059 non-interworking aware Thumb code.
4061 @cindex thumb entry point
4062 @cindex entry point, thumb
4063 @kindex --thumb-entry=@var{entry}
4064 The @samp{--thumb-entry} switch is a duplicate of the generic
4065 @samp{--entry} switch, in that it sets the program's starting address.
4066 But it also sets the bottom bit of the address, so that it can be
4067 branched to using a BX instruction, and the program will start
4068 executing in Thumb mode straight away.
4071 @section @code{ld} and HPPA 32-bit ELF support
4072 @cindex HPPA multiple sub-space stubs
4073 @kindex --multi-subspace
4074 When generating a shared library, @code{ld} will by default generate
4075 import stubs suitable for use with a single sub-space application.
4076 The @samp{--multi-subspace} switch causes @code{ld} to generate export
4077 stubs, and different (larger) import stubs suitable for use with
4078 multiple sub-spaces.
4080 @cindex HPPA stub grouping
4081 @kindex --stub-group-size=@var{N}
4082 Long branch stubs and import/export stubs are placed by @code{ld} in
4083 stub sections located between groups of input sections.
4084 @samp{--stub-group-size} specifies the maximum size of a group of input
4085 sections handled by one stub section. Since branch offsets are signed,
4086 a stub section may serve two groups of input sections, one group before
4087 the stub section, and one group after it. However, when using
4088 conditional branches that require stubs, it may be better (for branch
4089 prediction) that stub sections only serve one group of input sections.
4090 A negative value for @samp{N} chooses this scheme, ensuring that
4091 branches to stubs always use a negative offset. Two special values of
4092 @samp{N} are recognized, @samp{1} and @samp{-1}. These both instruct
4093 @code{ld} to automatically size input section groups for the branch types
4094 detected, with the same behaviour regarding stub placement as other
4095 positive or negative values of @samp{N} respectively.
4097 Note that @samp{--stub-group-size} does not split input sections. A
4098 single input section larger than the group size specified will of course
4099 create a larger group (of one section). If input sections are too
4100 large, it may not be possible for a branch to reach its stub.
4104 @section @code{ld}'s support for various TI COFF versions
4105 @cindex TI COFF versions
4106 @kindex --format=@var{version}
4107 The @samp{--format} switch allows selection of one of the various
4108 TI COFF versions. The latest of this writing is 2; versions 0 and 1 are
4109 also supported. The TI COFF versions also vary in header byte-order
4110 format; @code{ld} will read any version or byte order, but the output
4111 header format depends on the default specified by the specific target.
4118 @ifclear SingleFormat
4123 @cindex object file management
4124 @cindex object formats available
4126 The linker accesses object and archive files using the BFD libraries.
4127 These libraries allow the linker to use the same routines to operate on
4128 object files whatever the object file format. A different object file
4129 format can be supported simply by creating a new BFD back end and adding
4130 it to the library. To conserve runtime memory, however, the linker and
4131 associated tools are usually configured to support only a subset of the
4132 object file formats available. You can use @code{objdump -i}
4133 (@pxref{objdump,,objdump,binutils.info,The GNU Binary Utilities}) to
4134 list all the formats available for your configuration.
4136 @cindex BFD requirements
4137 @cindex requirements for BFD
4138 As with most implementations, BFD is a compromise between
4139 several conflicting requirements. The major factor influencing
4140 BFD design was efficiency: any time used converting between
4141 formats is time which would not have been spent had BFD not
4142 been involved. This is partly offset by abstraction payback; since
4143 BFD simplifies applications and back ends, more time and care
4144 may be spent optimizing algorithms for a greater speed.
4146 One minor artifact of the BFD solution which you should bear in
4147 mind is the potential for information loss. There are two places where
4148 useful information can be lost using the BFD mechanism: during
4149 conversion and during output. @xref{BFD information loss}.
4152 * BFD outline:: How it works: an outline of BFD
4156 @section How it works: an outline of BFD
4157 @cindex opening object files
4158 @include bfdsumm.texi
4161 @node Reporting Bugs
4162 @chapter Reporting Bugs
4163 @cindex bugs in @code{ld}
4164 @cindex reporting bugs in @code{ld}
4166 Your bug reports play an essential role in making @code{ld} reliable.
4168 Reporting a bug may help you by bringing a solution to your problem, or
4169 it may not. But in any case the principal function of a bug report is
4170 to help the entire community by making the next version of @code{ld}
4171 work better. Bug reports are your contribution to the maintenance of
4174 In order for a bug report to serve its purpose, you must include the
4175 information that enables us to fix the bug.
4178 * Bug Criteria:: Have you found a bug?
4179 * Bug Reporting:: How to report bugs
4183 @section Have you found a bug?
4184 @cindex bug criteria
4186 If you are not sure whether you have found a bug, here are some guidelines:
4189 @cindex fatal signal
4190 @cindex linker crash
4191 @cindex crash of linker
4193 If the linker gets a fatal signal, for any input whatever, that is a
4194 @code{ld} bug. Reliable linkers never crash.
4196 @cindex error on valid input
4198 If @code{ld} produces an error message for valid input, that is a bug.
4200 @cindex invalid input
4202 If @code{ld} does not produce an error message for invalid input, that
4203 may be a bug. In the general case, the linker can not verify that
4204 object files are correct.
4207 If you are an experienced user of linkers, your suggestions for
4208 improvement of @code{ld} are welcome in any case.
4212 @section How to report bugs
4214 @cindex @code{ld} bugs, reporting
4216 A number of companies and individuals offer support for @sc{gnu}
4217 products. If you obtained @code{ld} from a support organization, we
4218 recommend you contact that organization first.
4220 You can find contact information for many support companies and
4221 individuals in the file @file{etc/SERVICE} in the @sc{gnu} Emacs
4224 Otherwise, send bug reports for @code{ld} to
4225 @samp{bug-gnu-utils@@gnu.org}.
4227 The fundamental principle of reporting bugs usefully is this:
4228 @strong{report all the facts}. If you are not sure whether to state a
4229 fact or leave it out, state it!
4231 Often people omit facts because they think they know what causes the
4232 problem and assume that some details do not matter. Thus, you might
4233 assume that the name of a symbol you use in an example does not matter.
4234 Well, probably it does not, but one cannot be sure. Perhaps the bug is
4235 a stray memory reference which happens to fetch from the location where
4236 that name is stored in memory; perhaps, if the name were different, the
4237 contents of that location would fool the linker into doing the right
4238 thing despite the bug. Play it safe and give a specific, complete
4239 example. That is the easiest thing for you to do, and the most helpful.
4241 Keep in mind that the purpose of a bug report is to enable us to fix the bug if
4242 it is new to us. Therefore, always write your bug reports on the assumption
4243 that the bug has not been reported previously.
4245 Sometimes people give a few sketchy facts and ask, ``Does this ring a
4246 bell?'' Those bug reports are useless, and we urge everyone to
4247 @emph{refuse to respond to them} except to chide the sender to report
4250 To enable us to fix the bug, you should include all these things:
4254 The version of @code{ld}. @code{ld} announces it if you start it with
4255 the @samp{--version} argument.
4257 Without this, we will not know whether there is any point in looking for
4258 the bug in the current version of @code{ld}.
4261 Any patches you may have applied to the @code{ld} source, including any
4262 patches made to the @code{BFD} library.
4265 The type of machine you are using, and the operating system name and
4269 What compiler (and its version) was used to compile @code{ld}---e.g.
4273 The command arguments you gave the linker to link your example and
4274 observe the bug. To guarantee you will not omit something important,
4275 list them all. A copy of the Makefile (or the output from make) is
4278 If we were to try to guess the arguments, we would probably guess wrong
4279 and then we might not encounter the bug.
4282 A complete input file, or set of input files, that will reproduce the
4283 bug. It is generally most helpful to send the actual object files,
4284 uuencoded if necessary to get them through the mail system. Making them
4285 available for anonymous FTP is not as good, but may be the only
4286 reasonable choice for large object files.
4288 If the source files were assembled using @code{gas} or compiled using
4289 @code{gcc}, then it may be OK to send the source files rather than the
4290 object files. In this case, be sure to say exactly what version of
4291 @code{gas} or @code{gcc} was used to produce the object files. Also say
4292 how @code{gas} or @code{gcc} were configured.
4295 A description of what behavior you observe that you believe is
4296 incorrect. For example, ``It gets a fatal signal.''
4298 Of course, if the bug is that @code{ld} gets a fatal signal, then we
4299 will certainly notice it. But if the bug is incorrect output, we might
4300 not notice unless it is glaringly wrong. You might as well not give us
4301 a chance to make a mistake.
4303 Even if the problem you experience is a fatal signal, you should still
4304 say so explicitly. Suppose something strange is going on, such as, your
4305 copy of @code{ld} is out of synch, or you have encountered a bug in the
4306 C library on your system. (This has happened!) Your copy might crash
4307 and ours would not. If you told us to expect a crash, then when ours
4308 fails to crash, we would know that the bug was not happening for us. If
4309 you had not told us to expect a crash, then we would not be able to draw
4310 any conclusion from our observations.
4313 If you wish to suggest changes to the @code{ld} source, send us context
4314 diffs, as generated by @code{diff} with the @samp{-u}, @samp{-c}, or
4315 @samp{-p} option. Always send diffs from the old file to the new file.
4316 If you even discuss something in the @code{ld} source, refer to it by
4317 context, not by line number.
4319 The line numbers in our development sources will not match those in your
4320 sources. Your line numbers would convey no useful information to us.
4323 Here are some things that are not necessary:
4327 A description of the envelope of the bug.
4329 Often people who encounter a bug spend a lot of time investigating
4330 which changes to the input file will make the bug go away and which
4331 changes will not affect it.
4333 This is often time consuming and not very useful, because the way we
4334 will find the bug is by running a single example under the debugger
4335 with breakpoints, not by pure deduction from a series of examples.
4336 We recommend that you save your time for something else.
4338 Of course, if you can find a simpler example to report @emph{instead}
4339 of the original one, that is a convenience for us. Errors in the
4340 output will be easier to spot, running under the debugger will take
4341 less time, and so on.
4343 However, simplification is not vital; if you do not want to do this,
4344 report the bug anyway and send us the entire test case you used.
4347 A patch for the bug.
4349 A patch for the bug does help us if it is a good one. But do not omit
4350 the necessary information, such as the test case, on the assumption that
4351 a patch is all we need. We might see problems with your patch and decide
4352 to fix the problem another way, or we might not understand it at all.
4354 Sometimes with a program as complicated as @code{ld} it is very hard to
4355 construct an example that will make the program follow a certain path
4356 through the code. If you do not send us the example, we will not be
4357 able to construct one, so we will not be able to verify that the bug is
4360 And if we cannot understand what bug you are trying to fix, or why your
4361 patch should be an improvement, we will not install it. A test case will
4362 help us to understand.
4365 A guess about what the bug is or what it depends on.
4367 Such guesses are usually wrong. Even we cannot guess right about such
4368 things without first using the debugger to find the facts.
4372 @appendix MRI Compatible Script Files
4373 @cindex MRI compatibility
4374 To aid users making the transition to @sc{gnu} @code{ld} from the MRI
4375 linker, @code{ld} can use MRI compatible linker scripts as an
4376 alternative to the more general-purpose linker scripting language
4377 described in @ref{Scripts}. MRI compatible linker scripts have a much
4378 simpler command set than the scripting language otherwise used with
4379 @code{ld}. @sc{gnu} @code{ld} supports the most commonly used MRI
4380 linker commands; these commands are described here.
4382 In general, MRI scripts aren't of much use with the @code{a.out} object
4383 file format, since it only has three sections and MRI scripts lack some
4384 features to make use of them.
4386 You can specify a file containing an MRI-compatible script using the
4387 @samp{-c} command-line option.
4389 Each command in an MRI-compatible script occupies its own line; each
4390 command line starts with the keyword that identifies the command (though
4391 blank lines are also allowed for punctuation). If a line of an
4392 MRI-compatible script begins with an unrecognized keyword, @code{ld}
4393 issues a warning message, but continues processing the script.
4395 Lines beginning with @samp{*} are comments.
4397 You can write these commands using all upper-case letters, or all
4398 lower case; for example, @samp{chip} is the same as @samp{CHIP}.
4399 The following list shows only the upper-case form of each command.
4402 @cindex @code{ABSOLUTE} (MRI)
4403 @item ABSOLUTE @var{secname}
4404 @itemx ABSOLUTE @var{secname}, @var{secname}, @dots{} @var{secname}
4405 Normally, @code{ld} includes in the output file all sections from all
4406 the input files. However, in an MRI-compatible script, you can use the
4407 @code{ABSOLUTE} command to restrict the sections that will be present in
4408 your output program. If the @code{ABSOLUTE} command is used at all in a
4409 script, then only the sections named explicitly in @code{ABSOLUTE}
4410 commands will appear in the linker output. You can still use other
4411 input sections (whatever you select on the command line, or using
4412 @code{LOAD}) to resolve addresses in the output file.
4414 @cindex @code{ALIAS} (MRI)
4415 @item ALIAS @var{out-secname}, @var{in-secname}
4416 Use this command to place the data from input section @var{in-secname}
4417 in a section called @var{out-secname} in the linker output file.
4419 @var{in-secname} may be an integer.
4421 @cindex @code{ALIGN} (MRI)
4422 @item ALIGN @var{secname} = @var{expression}
4423 Align the section called @var{secname} to @var{expression}. The
4424 @var{expression} should be a power of two.
4426 @cindex @code{BASE} (MRI)
4427 @item BASE @var{expression}
4428 Use the value of @var{expression} as the lowest address (other than
4429 absolute addresses) in the output file.
4431 @cindex @code{CHIP} (MRI)
4432 @item CHIP @var{expression}
4433 @itemx CHIP @var{expression}, @var{expression}
4434 This command does nothing; it is accepted only for compatibility.
4436 @cindex @code{END} (MRI)
4438 This command does nothing whatever; it's only accepted for compatibility.
4440 @cindex @code{FORMAT} (MRI)
4441 @item FORMAT @var{output-format}
4442 Similar to the @code{OUTPUT_FORMAT} command in the more general linker
4443 language, but restricted to one of these output formats:
4447 S-records, if @var{output-format} is @samp{S}
4450 IEEE, if @var{output-format} is @samp{IEEE}
4453 COFF (the @samp{coff-m68k} variant in BFD), if @var{output-format} is
4457 @cindex @code{LIST} (MRI)
4458 @item LIST @var{anything}@dots{}
4459 Print (to the standard output file) a link map, as produced by the
4460 @code{ld} command-line option @samp{-M}.
4462 The keyword @code{LIST} may be followed by anything on the
4463 same line, with no change in its effect.
4465 @cindex @code{LOAD} (MRI)
4466 @item LOAD @var{filename}
4467 @itemx LOAD @var{filename}, @var{filename}, @dots{} @var{filename}
4468 Include one or more object file @var{filename} in the link; this has the
4469 same effect as specifying @var{filename} directly on the @code{ld}
4472 @cindex @code{NAME} (MRI)
4473 @item NAME @var{output-name}
4474 @var{output-name} is the name for the program produced by @code{ld}; the
4475 MRI-compatible command @code{NAME} is equivalent to the command-line
4476 option @samp{-o} or the general script language command @code{OUTPUT}.
4478 @cindex @code{ORDER} (MRI)
4479 @item ORDER @var{secname}, @var{secname}, @dots{} @var{secname}
4480 @itemx ORDER @var{secname} @var{secname} @var{secname}
4481 Normally, @code{ld} orders the sections in its output file in the
4482 order in which they first appear in the input files. In an MRI-compatible
4483 script, you can override this ordering with the @code{ORDER} command. The
4484 sections you list with @code{ORDER} will appear first in your output
4485 file, in the order specified.
4487 @cindex @code{PUBLIC} (MRI)
4488 @item PUBLIC @var{name}=@var{expression}
4489 @itemx PUBLIC @var{name},@var{expression}
4490 @itemx PUBLIC @var{name} @var{expression}
4491 Supply a value (@var{expression}) for external symbol
4492 @var{name} used in the linker input files.
4494 @cindex @code{SECT} (MRI)
4495 @item SECT @var{secname}, @var{expression}
4496 @itemx SECT @var{secname}=@var{expression}
4497 @itemx SECT @var{secname} @var{expression}
4498 You can use any of these three forms of the @code{SECT} command to
4499 specify the start address (@var{expression}) for section @var{secname}.
4500 If you have more than one @code{SECT} statement for the same
4501 @var{secname}, only the @emph{first} sets the start address.
4504 @node GNU Free Documentation License
4505 @appendix GNU Free Documentation License
4506 @cindex GNU Free Documentation License
4508 GNU Free Documentation License
4510 Version 1.1, March 2000
4512 Copyright (C) 2000 Free Software Foundation, Inc.
4513 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
4515 Everyone is permitted to copy and distribute verbatim copies
4516 of this license document, but changing it is not allowed.
4521 The purpose of this License is to make a manual, textbook, or other
4522 written document "free" in the sense of freedom: to assure everyone
4523 the effective freedom to copy and redistribute it, with or without
4524 modifying it, either commercially or noncommercially. Secondarily,
4525 this License preserves for the author and publisher a way to get
4526 credit for their work, while not being considered responsible for
4527 modifications made by others.
4529 This License is a kind of "copyleft", which means that derivative
4530 works of the document must themselves be free in the same sense. It
4531 complements the GNU General Public License, which is a copyleft
4532 license designed for free software.
4534 We have designed this License in order to use it for manuals for free
4535 software, because free software needs free documentation: a free
4536 program should come with manuals providing the same freedoms that the
4537 software does. But this License is not limited to software manuals;
4538 it can be used for any textual work, regardless of subject matter or
4539 whether it is published as a printed book. We recommend this License
4540 principally for works whose purpose is instruction or reference.
4543 1. APPLICABILITY AND DEFINITIONS
4545 This License applies to any manual or other work that contains a
4546 notice placed by the copyright holder saying it can be distributed
4547 under the terms of this License. The "Document", below, refers to any
4548 such manual or work. Any member of the public is a licensee, and is
4551 A "Modified Version" of the Document means any work containing the
4552 Document or a portion of it, either copied verbatim, or with
4553 modifications and/or translated into another language.
4555 A "Secondary Section" is a named appendix or a front-matter section of
4556 the Document that deals exclusively with the relationship of the
4557 publishers or authors of the Document to the Document's overall subject
4558 (or to related matters) and contains nothing that could fall directly
4559 within that overall subject. (For example, if the Document is in part a
4560 textbook of mathematics, a Secondary Section may not explain any
4561 mathematics.) The relationship could be a matter of historical
4562 connection with the subject or with related matters, or of legal,
4563 commercial, philosophical, ethical or political position regarding
4566 The "Invariant Sections" are certain Secondary Sections whose titles
4567 are designated, as being those of Invariant Sections, in the notice
4568 that says that the Document is released under this License.
4570 The "Cover Texts" are certain short passages of text that are listed,
4571 as Front-Cover Texts or Back-Cover Texts, in the notice that says that
4572 the Document is released under this License.
4574 A "Transparent" copy of the Document means a machine-readable copy,
4575 represented in a format whose specification is available to the
4576 general public, whose contents can be viewed and edited directly and
4577 straightforwardly with generic text editors or (for images composed of
4578 pixels) generic paint programs or (for drawings) some widely available
4579 drawing editor, and that is suitable for input to text formatters or
4580 for automatic translation to a variety of formats suitable for input
4581 to text formatters. A copy made in an otherwise Transparent file
4582 format whose markup has been designed to thwart or discourage
4583 subsequent modification by readers is not Transparent. A copy that is
4584 not "Transparent" is called "Opaque".
4586 Examples of suitable formats for Transparent copies include plain
4587 ASCII without markup, Texinfo input format, LaTeX input format, SGML
4588 or XML using a publicly available DTD, and standard-conforming simple
4589 HTML designed for human modification. Opaque formats include
4590 PostScript, PDF, proprietary formats that can be read and edited only
4591 by proprietary word processors, SGML or XML for which the DTD and/or
4592 processing tools are not generally available, and the
4593 machine-generated HTML produced by some word processors for output
4596 The "Title Page" means, for a printed book, the title page itself,
4597 plus such following pages as are needed to hold, legibly, the material
4598 this License requires to appear in the title page. For works in
4599 formats which do not have any title page as such, "Title Page" means
4600 the text near the most prominent appearance of the work's title,
4601 preceding the beginning of the body of the text.
4606 You may copy and distribute the Document in any medium, either
4607 commercially or noncommercially, provided that this License, the
4608 copyright notices, and the license notice saying this License applies
4609 to the Document are reproduced in all copies, and that you add no other
4610 conditions whatsoever to those of this License. You may not use
4611 technical measures to obstruct or control the reading or further
4612 copying of the copies you make or distribute. However, you may accept
4613 compensation in exchange for copies. If you distribute a large enough
4614 number of copies you must also follow the conditions in section 3.
4616 You may also lend copies, under the same conditions stated above, and
4617 you may publicly display copies.
4620 3. COPYING IN QUANTITY
4622 If you publish printed copies of the Document numbering more than 100,
4623 and the Document's license notice requires Cover Texts, you must enclose
4624 the copies in covers that carry, clearly and legibly, all these Cover
4625 Texts: Front-Cover Texts on the front cover, and Back-Cover Texts on
4626 the back cover. Both covers must also clearly and legibly identify
4627 you as the publisher of these copies. The front cover must present
4628 the full title with all words of the title equally prominent and
4629 visible. You may add other material on the covers in addition.
4630 Copying with changes limited to the covers, as long as they preserve
4631 the title of the Document and satisfy these conditions, can be treated
4632 as verbatim copying in other respects.
4634 If the required texts for either cover are too voluminous to fit
4635 legibly, you should put the first ones listed (as many as fit
4636 reasonably) on the actual cover, and continue the rest onto adjacent
4639 If you publish or distribute Opaque copies of the Document numbering
4640 more than 100, you must either include a machine-readable Transparent
4641 copy along with each Opaque copy, or state in or with each Opaque copy
4642 a publicly-accessible computer-network location containing a complete
4643 Transparent copy of the Document, free of added material, which the
4644 general network-using public has access to download anonymously at no
4645 charge using public-standard network protocols. If you use the latter
4646 option, you must take reasonably prudent steps, when you begin
4647 distribution of Opaque copies in quantity, to ensure that this
4648 Transparent copy will remain thus accessible at the stated location
4649 until at least one year after the last time you distribute an Opaque
4650 copy (directly or through your agents or retailers) of that edition to
4653 It is requested, but not required, that you contact the authors of the
4654 Document well before redistributing any large number of copies, to give
4655 them a chance to provide you with an updated version of the Document.
4660 You may copy and distribute a Modified Version of the Document under
4661 the conditions of sections 2 and 3 above, provided that you release
4662 the Modified Version under precisely this License, with the Modified
4663 Version filling the role of the Document, thus licensing distribution
4664 and modification of the Modified Version to whoever possesses a copy
4665 of it. In addition, you must do these things in the Modified Version:
4667 A. Use in the Title Page (and on the covers, if any) a title distinct
4668 from that of the Document, and from those of previous versions
4669 (which should, if there were any, be listed in the History section
4670 of the Document). You may use the same title as a previous version
4671 if the original publisher of that version gives permission.
4672 B. List on the Title Page, as authors, one or more persons or entities
4673 responsible for authorship of the modifications in the Modified
4674 Version, together with at least five of the principal authors of the
4675 Document (all of its principal authors, if it has less than five).
4676 C. State on the Title page the name of the publisher of the
4677 Modified Version, as the publisher.
4678 D. Preserve all the copyright notices of the Document.
4679 E. Add an appropriate copyright notice for your modifications
4680 adjacent to the other copyright notices.
4681 F. Include, immediately after the copyright notices, a license notice
4682 giving the public permission to use the Modified Version under the
4683 terms of this License, in the form shown in the Addendum below.
4684 G. Preserve in that license notice the full lists of Invariant Sections
4685 and required Cover Texts given in the Document's license notice.
4686 H. Include an unaltered copy of this License.
4687 I. Preserve the section entitled "History", and its title, and add to
4688 it an item stating at least the title, year, new authors, and
4689 publisher of the Modified Version as given on the Title Page. If
4690 there is no section entitled "History" in the Document, create one
4691 stating the title, year, authors, and publisher of the Document as
4692 given on its Title Page, then add an item describing the Modified
4693 Version as stated in the previous sentence.
4694 J. Preserve the network location, if any, given in the Document for
4695 public access to a Transparent copy of the Document, and likewise
4696 the network locations given in the Document for previous versions
4697 it was based on. These may be placed in the "History" section.
4698 You may omit a network location for a work that was published at
4699 least four years before the Document itself, or if the original
4700 publisher of the version it refers to gives permission.
4701 K. In any section entitled "Acknowledgements" or "Dedications",
4702 preserve the section's title, and preserve in the section all the
4703 substance and tone of each of the contributor acknowledgements
4704 and/or dedications given therein.
4705 L. Preserve all the Invariant Sections of the Document,
4706 unaltered in their text and in their titles. Section numbers
4707 or the equivalent are not considered part of the section titles.
4708 M. Delete any section entitled "Endorsements". Such a section
4709 may not be included in the Modified Version.
4710 N. Do not retitle any existing section as "Endorsements"
4711 or to conflict in title with any Invariant Section.
4713 If the Modified Version includes new front-matter sections or
4714 appendices that qualify as Secondary Sections and contain no material
4715 copied from the Document, you may at your option designate some or all
4716 of these sections as invariant. To do this, add their titles to the
4717 list of Invariant Sections in the Modified Version's license notice.
4718 These titles must be distinct from any other section titles.
4720 You may add a section entitled "Endorsements", provided it contains
4721 nothing but endorsements of your Modified Version by various
4722 parties--for example, statements of peer review or that the text has
4723 been approved by an organization as the authoritative definition of a
4726 You may add a passage of up to five words as a Front-Cover Text, and a
4727 passage of up to 25 words as a Back-Cover Text, to the end of the list
4728 of Cover Texts in the Modified Version. Only one passage of
4729 Front-Cover Text and one of Back-Cover Text may be added by (or
4730 through arrangements made by) any one entity. If the Document already
4731 includes a cover text for the same cover, previously added by you or
4732 by arrangement made by the same entity you are acting on behalf of,
4733 you may not add another; but you may replace the old one, on explicit
4734 permission from the previous publisher that added the old one.
4736 The author(s) and publisher(s) of the Document do not by this License
4737 give permission to use their names for publicity for or to assert or
4738 imply endorsement of any Modified Version.
4741 5. COMBINING DOCUMENTS
4743 You may combine the Document with other documents released under this
4744 License, under the terms defined in section 4 above for modified
4745 versions, provided that you include in the combination all of the
4746 Invariant Sections of all of the original documents, unmodified, and
4747 list them all as Invariant Sections of your combined work in its
4750 The combined work need only contain one copy of this License, and
4751 multiple identical Invariant Sections may be replaced with a single
4752 copy. If there are multiple Invariant Sections with the same name but
4753 different contents, make the title of each such section unique by
4754 adding at the end of it, in parentheses, the name of the original
4755 author or publisher of that section if known, or else a unique number.
4756 Make the same adjustment to the section titles in the list of
4757 Invariant Sections in the license notice of the combined work.
4759 In the combination, you must combine any sections entitled "History"
4760 in the various original documents, forming one section entitled
4761 "History"; likewise combine any sections entitled "Acknowledgements",
4762 and any sections entitled "Dedications". You must delete all sections
4763 entitled "Endorsements."
4766 6. COLLECTIONS OF DOCUMENTS
4768 You may make a collection consisting of the Document and other documents
4769 released under this License, and replace the individual copies of this
4770 License in the various documents with a single copy that is included in
4771 the collection, provided that you follow the rules of this License for
4772 verbatim copying of each of the documents in all other respects.
4774 You may extract a single document from such a collection, and distribute
4775 it individually under this License, provided you insert a copy of this
4776 License into the extracted document, and follow this License in all
4777 other respects regarding verbatim copying of that document.
4780 7. AGGREGATION WITH INDEPENDENT WORKS
4782 A compilation of the Document or its derivatives with other separate
4783 and independent documents or works, in or on a volume of a storage or
4784 distribution medium, does not as a whole count as a Modified Version
4785 of the Document, provided no compilation copyright is claimed for the
4786 compilation. Such a compilation is called an "aggregate", and this
4787 License does not apply to the other self-contained works thus compiled
4788 with the Document, on account of their being thus compiled, if they
4789 are not themselves derivative works of the Document.
4791 If the Cover Text requirement of section 3 is applicable to these
4792 copies of the Document, then if the Document is less than one quarter
4793 of the entire aggregate, the Document's Cover Texts may be placed on
4794 covers that surround only the Document within the aggregate.
4795 Otherwise they must appear on covers around the whole aggregate.
4800 Translation is considered a kind of modification, so you may
4801 distribute translations of the Document under the terms of section 4.
4802 Replacing Invariant Sections with translations requires special
4803 permission from their copyright holders, but you may include
4804 translations of some or all Invariant Sections in addition to the
4805 original versions of these Invariant Sections. You may include a
4806 translation of this License provided that you also include the
4807 original English version of this License. In case of a disagreement
4808 between the translation and the original English version of this
4809 License, the original English version will prevail.
4814 You may not copy, modify, sublicense, or distribute the Document except
4815 as expressly provided for under this License. Any other attempt to
4816 copy, modify, sublicense or distribute the Document is void, and will
4817 automatically terminate your rights under this License. However,
4818 parties who have received copies, or rights, from you under this
4819 License will not have their licenses terminated so long as such
4820 parties remain in full compliance.
4823 10. FUTURE REVISIONS OF THIS LICENSE
4825 The Free Software Foundation may publish new, revised versions
4826 of the GNU Free Documentation License from time to time. Such new
4827 versions will be similar in spirit to the present version, but may
4828 differ in detail to address new problems or concerns. See
4829 http://www.gnu.org/copyleft/.
4831 Each version of the License is given a distinguishing version number.
4832 If the Document specifies that a particular numbered version of this
4833 License "or any later version" applies to it, you have the option of
4834 following the terms and conditions either of that specified version or
4835 of any later version that has been published (not as a draft) by the
4836 Free Software Foundation. If the Document does not specify a version
4837 number of this License, you may choose any version ever published (not
4838 as a draft) by the Free Software Foundation.
4841 ADDENDUM: How to use this License for your documents
4843 To use this License in a document you have written, include a copy of
4844 the License in the document and put the following copyright and
4845 license notices just after the title page:
4848 Copyright (c) YEAR YOUR NAME.
4849 Permission is granted to copy, distribute and/or modify this document
4850 under the terms of the GNU Free Documentation License, Version 1.1
4851 or any later version published by the Free Software Foundation;
4852 with the Invariant Sections being LIST THEIR TITLES, with the
4853 Front-Cover Texts being LIST, and with the Back-Cover Texts being LIST.
4854 A copy of the license is included in the section entitled "GNU
4855 Free Documentation License".
4858 If you have no Invariant Sections, write "with no Invariant Sections"
4859 instead of saying which ones are invariant. If you have no
4860 Front-Cover Texts, write "no Front-Cover Texts" instead of
4861 "Front-Cover Texts being LIST"; likewise for Back-Cover Texts.
4863 If your document contains nontrivial examples of program code, we
4864 recommend releasing these examples in parallel under your choice of
4865 free software license, such as the GNU General Public License,
4866 to permit their use in free software.
4874 % I think something like @colophon should be in texinfo. In the
4876 \long\def\colophon{\hbox to0pt{}\vfill
4877 \centerline{The body of this manual is set in}
4878 \centerline{\fontname\tenrm,}
4879 \centerline{with headings in {\bf\fontname\tenbf}}
4880 \centerline{and examples in {\tt\fontname\tentt}.}
4881 \centerline{{\it\fontname\tenit\/} and}
4882 \centerline{{\sl\fontname\tensl\/}}
4883 \centerline{are used for emphasis.}\vfill}
4885 % Blame: doc@cygnus.com, 28mar91.