3 @c Copyright 1991, 1992, 1993, 1994, 1995, 1996, 1997, 1998, 1999, 2000,
4 @c 2001, 2002, 2003, 2004, 2005, 2006 Free Software Foundation, Inc.
7 @include configdoc.texi
8 @c (configdoc.texi is generated by the Makefile)
14 @macro gcctabopt{body}
20 @c Configure for the generation of man pages
58 * Ld: (ld). The GNU linker.
64 This file documents the @sc{gnu} linker LD version @value{VERSION}.
66 Copyright (C) 1991, 92, 93, 94, 95, 96, 97, 98, 99, 2000,
67 2001, 2002, 2003, 2004, 2005, 2006 Free Software Foundation, Inc.
71 Permission is granted to copy, distribute and/or modify this document
72 under the terms of the GNU Free Documentation License, Version 1.1
73 or any later version published by the Free Software Foundation;
74 with no Invariant Sections, with no Front-Cover Texts, and with no
75 Back-Cover Texts. A copy of the license is included in the
76 section entitled ``GNU Free Documentation License''.
78 Permission is granted to process this file through Tex and print the
79 results, provided the printed document carries copying permission
80 notice identical to this one except for the removal of this paragraph
81 (this paragraph not being relevant to the printed manual).
87 @setchapternewpage odd
88 @settitle The GNU linker
92 @subtitle @code{ld} version 2
93 @subtitle Version @value{VERSION}
94 @author Steve Chamberlain
95 @author Ian Lance Taylor
100 \hfill Red Hat Inc\par
101 \hfill nickc\@credhat.com, doc\@redhat.com\par
102 \hfill {\it The GNU linker}\par
103 \hfill Edited by Jeffrey Osier (jeffrey\@cygnus.com)\par
105 \global\parindent=0pt % Steve likes it this way.
108 @vskip 0pt plus 1filll
109 @c man begin COPYRIGHT
110 Copyright @copyright{} 1991, 92, 93, 94, 95, 96, 97, 98, 99, 2000, 2001,
111 2002, 2003, 2004, 2005, 2006 Free Software Foundation, Inc.
113 Permission is granted to copy, distribute and/or modify this document
114 under the terms of the GNU Free Documentation License, Version 1.1
115 or any later version published by the Free Software Foundation;
116 with no Invariant Sections, with no Front-Cover Texts, and with no
117 Back-Cover Texts. A copy of the license is included in the
118 section entitled ``GNU Free Documentation License''.
123 @c FIXME: Talk about importance of *order* of args, cmds to linker!
128 This file documents the @sc{gnu} linker ld version @value{VERSION}.
130 This document is distributed under the terms of the GNU Free
131 Documentation License. A copy of the license is included in the
132 section entitled ``GNU Free Documentation License''.
135 * Overview:: Overview
136 * Invocation:: Invocation
137 * Scripts:: Linker Scripts
139 * Machine Dependent:: Machine Dependent Features
143 * H8/300:: ld and the H8/300
146 * Renesas:: ld and other Renesas micros
149 * i960:: ld and the Intel 960 family
152 * ARM:: ld and the ARM family
155 * HPPA ELF32:: ld and HPPA 32-bit ELF
158 * M68HC11/68HC12:: ld and the Motorola 68HC11 and 68HC12 families
161 * PowerPC ELF32:: ld and PowerPC 32-bit ELF Support
164 * PowerPC64 ELF64:: ld and PowerPC64 64-bit ELF Support
167 * TI COFF:: ld and the TI COFF
170 * Win32:: ld and WIN32 (cygwin/mingw)
173 * Xtensa:: ld and Xtensa Processors
176 @ifclear SingleFormat
179 @c Following blank line required for remaining bug in makeinfo conds/menus
181 * Reporting Bugs:: Reporting Bugs
182 * MRI:: MRI Compatible Script Files
183 * GNU Free Documentation License:: GNU Free Documentation License
184 * LD Index:: LD Index
191 @cindex @sc{gnu} linker
192 @cindex what is this?
195 @c man begin SYNOPSIS
196 ld [@b{options}] @var{objfile} @dots{}
200 ar(1), nm(1), objcopy(1), objdump(1), readelf(1) and
201 the Info entries for @file{binutils} and
206 @c man begin DESCRIPTION
208 @command{ld} combines a number of object and archive files, relocates
209 their data and ties up symbol references. Usually the last step in
210 compiling a program is to run @command{ld}.
212 @command{ld} accepts Linker Command Language files written in
213 a superset of AT&T's Link Editor Command Language syntax,
214 to provide explicit and total control over the linking process.
218 This man page does not describe the command language; see the
219 @command{ld} entry in @code{info} for full details on the command
220 language and on other aspects of the GNU linker.
223 @ifclear SingleFormat
224 This version of @command{ld} uses the general purpose BFD libraries
225 to operate on object files. This allows @command{ld} to read, combine, and
226 write object files in many different formats---for example, COFF or
227 @code{a.out}. Different formats may be linked together to produce any
228 available kind of object file. @xref{BFD}, for more information.
231 Aside from its flexibility, the @sc{gnu} linker is more helpful than other
232 linkers in providing diagnostic information. Many linkers abandon
233 execution immediately upon encountering an error; whenever possible,
234 @command{ld} continues executing, allowing you to identify other errors
235 (or, in some cases, to get an output file in spite of the error).
242 @c man begin DESCRIPTION
244 The @sc{gnu} linker @command{ld} is meant to cover a broad range of situations,
245 and to be as compatible as possible with other linkers. As a result,
246 you have many choices to control its behavior.
252 * Options:: Command Line Options
253 * Environment:: Environment Variables
257 @section Command Line Options
265 The linker supports a plethora of command-line options, but in actual
266 practice few of them are used in any particular context.
267 @cindex standard Unix system
268 For instance, a frequent use of @command{ld} is to link standard Unix
269 object files on a standard, supported Unix system. On such a system, to
270 link a file @code{hello.o}:
273 ld -o @var{output} /lib/crt0.o hello.o -lc
276 This tells @command{ld} to produce a file called @var{output} as the
277 result of linking the file @code{/lib/crt0.o} with @code{hello.o} and
278 the library @code{libc.a}, which will come from the standard search
279 directories. (See the discussion of the @samp{-l} option below.)
281 Some of the command-line options to @command{ld} may be specified at any
282 point in the command line. However, options which refer to files, such
283 as @samp{-l} or @samp{-T}, cause the file to be read at the point at
284 which the option appears in the command line, relative to the object
285 files and other file options. Repeating non-file options with a
286 different argument will either have no further effect, or override prior
287 occurrences (those further to the left on the command line) of that
288 option. Options which may be meaningfully specified more than once are
289 noted in the descriptions below.
292 Non-option arguments are object files or archives which are to be linked
293 together. They may follow, precede, or be mixed in with command-line
294 options, except that an object file argument may not be placed between
295 an option and its argument.
297 Usually the linker is invoked with at least one object file, but you can
298 specify other forms of binary input files using @samp{-l}, @samp{-R},
299 and the script command language. If @emph{no} binary input files at all
300 are specified, the linker does not produce any output, and issues the
301 message @samp{No input files}.
303 If the linker cannot recognize the format of an object file, it will
304 assume that it is a linker script. A script specified in this way
305 augments the main linker script used for the link (either the default
306 linker script or the one specified by using @samp{-T}). This feature
307 permits the linker to link against a file which appears to be an object
308 or an archive, but actually merely defines some symbol values, or uses
309 @code{INPUT} or @code{GROUP} to load other objects. Note that
310 specifying a script in this way merely augments the main linker script;
311 use the @samp{-T} option to replace the default linker script entirely.
314 For options whose names are a single letter,
315 option arguments must either follow the option letter without intervening
316 whitespace, or be given as separate arguments immediately following the
317 option that requires them.
319 For options whose names are multiple letters, either one dash or two can
320 precede the option name; for example, @samp{-trace-symbol} and
321 @samp{--trace-symbol} are equivalent. Note---there is one exception to
322 this rule. Multiple letter options that start with a lower case 'o' can
323 only be preceded by two dashes. This is to reduce confusion with the
324 @samp{-o} option. So for example @samp{-omagic} sets the output file
325 name to @samp{magic} whereas @samp{--omagic} sets the NMAGIC flag on the
328 Arguments to multiple-letter options must either be separated from the
329 option name by an equals sign, or be given as separate arguments
330 immediately following the option that requires them. For example,
331 @samp{--trace-symbol foo} and @samp{--trace-symbol=foo} are equivalent.
332 Unique abbreviations of the names of multiple-letter options are
335 Note---if the linker is being invoked indirectly, via a compiler driver
336 (e.g. @samp{gcc}) then all the linker command line options should be
337 prefixed by @samp{-Wl,} (or whatever is appropriate for the particular
338 compiler driver) like this:
341 gcc -Wl,--startgroup foo.o bar.o -Wl,--endgroup
344 This is important, because otherwise the compiler driver program may
345 silently drop the linker options, resulting in a bad link.
347 Here is a table of the generic command line switches accepted by the GNU
351 @include at-file.texi
353 @kindex -a@var{keyword}
354 @item -a@var{keyword}
355 This option is supported for HP/UX compatibility. The @var{keyword}
356 argument must be one of the strings @samp{archive}, @samp{shared}, or
357 @samp{default}. @samp{-aarchive} is functionally equivalent to
358 @samp{-Bstatic}, and the other two keywords are functionally equivalent
359 to @samp{-Bdynamic}. This option may be used any number of times.
362 @cindex architectures
364 @item -A@var{architecture}
365 @kindex --architecture=@var{arch}
366 @itemx --architecture=@var{architecture}
367 In the current release of @command{ld}, this option is useful only for the
368 Intel 960 family of architectures. In that @command{ld} configuration, the
369 @var{architecture} argument identifies the particular architecture in
370 the 960 family, enabling some safeguards and modifying the
371 archive-library search path. @xref{i960,,@command{ld} and the Intel 960
372 family}, for details.
374 Future releases of @command{ld} may support similar functionality for
375 other architecture families.
378 @ifclear SingleFormat
379 @cindex binary input format
380 @kindex -b @var{format}
381 @kindex --format=@var{format}
384 @item -b @var{input-format}
385 @itemx --format=@var{input-format}
386 @command{ld} may be configured to support more than one kind of object
387 file. If your @command{ld} is configured this way, you can use the
388 @samp{-b} option to specify the binary format for input object files
389 that follow this option on the command line. Even when @command{ld} is
390 configured to support alternative object formats, you don't usually need
391 to specify this, as @command{ld} should be configured to expect as a
392 default input format the most usual format on each machine.
393 @var{input-format} is a text string, the name of a particular format
394 supported by the BFD libraries. (You can list the available binary
395 formats with @samp{objdump -i}.)
398 You may want to use this option if you are linking files with an unusual
399 binary format. You can also use @samp{-b} to switch formats explicitly (when
400 linking object files of different formats), by including
401 @samp{-b @var{input-format}} before each group of object files in a
404 The default format is taken from the environment variable
409 You can also define the input format from a script, using the command
412 see @ref{Format Commands}.
416 @kindex -c @var{MRI-cmdfile}
417 @kindex --mri-script=@var{MRI-cmdfile}
418 @cindex compatibility, MRI
419 @item -c @var{MRI-commandfile}
420 @itemx --mri-script=@var{MRI-commandfile}
421 For compatibility with linkers produced by MRI, @command{ld} accepts script
422 files written in an alternate, restricted command language, described in
424 @ref{MRI,,MRI Compatible Script Files}.
427 the MRI Compatible Script Files section of GNU ld documentation.
429 Introduce MRI script files with
430 the option @samp{-c}; use the @samp{-T} option to run linker
431 scripts written in the general-purpose @command{ld} scripting language.
432 If @var{MRI-cmdfile} does not exist, @command{ld} looks for it in the directories
433 specified by any @samp{-L} options.
435 @cindex common allocation
442 These three options are equivalent; multiple forms are supported for
443 compatibility with other linkers. They assign space to common symbols
444 even if a relocatable output file is specified (with @samp{-r}). The
445 script command @code{FORCE_COMMON_ALLOCATION} has the same effect.
446 @xref{Miscellaneous Commands}.
448 @cindex entry point, from command line
449 @kindex -e @var{entry}
450 @kindex --entry=@var{entry}
452 @itemx --entry=@var{entry}
453 Use @var{entry} as the explicit symbol for beginning execution of your
454 program, rather than the default entry point. If there is no symbol
455 named @var{entry}, the linker will try to parse @var{entry} as a number,
456 and use that as the entry address (the number will be interpreted in
457 base 10; you may use a leading @samp{0x} for base 16, or a leading
458 @samp{0} for base 8). @xref{Entry Point}, for a discussion of defaults
459 and other ways of specifying the entry point.
461 @kindex --exclude-libs
462 @item --exclude-libs @var{lib},@var{lib},...
463 Specifies a list of archive libraries from which symbols should not be automatically
464 exported. The library names may be delimited by commas or colons. Specifying
465 @code{--exclude-libs ALL} excludes symbols in all archive libraries from
466 automatic export. This option is available only for the i386 PE targeted
467 port of the linker and for ELF targeted ports. For i386 PE, symbols
468 explicitly listed in a .def file are still exported, regardless of this
469 option. For ELF targeted ports, symbols affected by this option will
470 be treated as hidden.
472 @cindex dynamic symbol table
474 @kindex --export-dynamic
476 @itemx --export-dynamic
477 When creating a dynamically linked executable, add all symbols to the
478 dynamic symbol table. The dynamic symbol table is the set of symbols
479 which are visible from dynamic objects at run time.
481 If you do not use this option, the dynamic symbol table will normally
482 contain only those symbols which are referenced by some dynamic object
483 mentioned in the link.
485 If you use @code{dlopen} to load a dynamic object which needs to refer
486 back to the symbols defined by the program, rather than some other
487 dynamic object, then you will probably need to use this option when
488 linking the program itself.
490 You can also use the dynamic list to control what symbols should
491 be added to the dynamic symbol table if the output format supports it.
492 See the description of @samp{--dynamic-list}.
494 @ifclear SingleFormat
495 @cindex big-endian objects
499 Link big-endian objects. This affects the default output format.
501 @cindex little-endian objects
504 Link little-endian objects. This affects the default output format.
510 @itemx --auxiliary @var{name}
511 When creating an ELF shared object, set the internal DT_AUXILIARY field
512 to the specified name. This tells the dynamic linker that the symbol
513 table of the shared object should be used as an auxiliary filter on the
514 symbol table of the shared object @var{name}.
516 If you later link a program against this filter object, then, when you
517 run the program, the dynamic linker will see the DT_AUXILIARY field. If
518 the dynamic linker resolves any symbols from the filter object, it will
519 first check whether there is a definition in the shared object
520 @var{name}. If there is one, it will be used instead of the definition
521 in the filter object. The shared object @var{name} need not exist.
522 Thus the shared object @var{name} may be used to provide an alternative
523 implementation of certain functions, perhaps for debugging or for
524 machine specific performance.
526 This option may be specified more than once. The DT_AUXILIARY entries
527 will be created in the order in which they appear on the command line.
532 @itemx --filter @var{name}
533 When creating an ELF shared object, set the internal DT_FILTER field to
534 the specified name. This tells the dynamic linker that the symbol table
535 of the shared object which is being created should be used as a filter
536 on the symbol table of the shared object @var{name}.
538 If you later link a program against this filter object, then, when you
539 run the program, the dynamic linker will see the DT_FILTER field. The
540 dynamic linker will resolve symbols according to the symbol table of the
541 filter object as usual, but it will actually link to the definitions
542 found in the shared object @var{name}. Thus the filter object can be
543 used to select a subset of the symbols provided by the object
546 Some older linkers used the @option{-F} option throughout a compilation
547 toolchain for specifying object-file format for both input and output
549 @ifclear SingleFormat
550 The @sc{gnu} linker uses other mechanisms for this purpose: the
551 @option{-b}, @option{--format}, @option{--oformat} options, the
552 @code{TARGET} command in linker scripts, and the @code{GNUTARGET}
553 environment variable.
555 The @sc{gnu} linker will ignore the @option{-F} option when not
556 creating an ELF shared object.
558 @cindex finalization function
560 @item -fini @var{name}
561 When creating an ELF executable or shared object, call NAME when the
562 executable or shared object is unloaded, by setting DT_FINI to the
563 address of the function. By default, the linker uses @code{_fini} as
564 the function to call.
568 Ignored. Provided for compatibility with other tools.
574 @itemx --gpsize=@var{value}
575 Set the maximum size of objects to be optimized using the GP register to
576 @var{size}. This is only meaningful for object file formats such as
577 MIPS ECOFF which supports putting large and small objects into different
578 sections. This is ignored for other object file formats.
580 @cindex runtime library name
582 @kindex -soname=@var{name}
584 @itemx -soname=@var{name}
585 When creating an ELF shared object, set the internal DT_SONAME field to
586 the specified name. When an executable is linked with a shared object
587 which has a DT_SONAME field, then when the executable is run the dynamic
588 linker will attempt to load the shared object specified by the DT_SONAME
589 field rather than the using the file name given to the linker.
592 @cindex incremental link
594 Perform an incremental link (same as option @samp{-r}).
596 @cindex initialization function
598 @item -init @var{name}
599 When creating an ELF executable or shared object, call NAME when the
600 executable or shared object is loaded, by setting DT_INIT to the address
601 of the function. By default, the linker uses @code{_init} as the
604 @cindex archive files, from cmd line
605 @kindex -l@var{archive}
606 @kindex --library=@var{archive}
607 @item -l@var{archive}
608 @itemx --library=@var{archive}
609 Add archive file @var{archive} to the list of files to link. This
610 option may be used any number of times. @command{ld} will search its
611 path-list for occurrences of @code{lib@var{archive}.a} for every
612 @var{archive} specified.
614 On systems which support shared libraries, @command{ld} may also search for
615 libraries with extensions other than @code{.a}. Specifically, on ELF
616 and SunOS systems, @command{ld} will search a directory for a library with
617 an extension of @code{.so} before searching for one with an extension of
618 @code{.a}. By convention, a @code{.so} extension indicates a shared
621 The linker will search an archive only once, at the location where it is
622 specified on the command line. If the archive defines a symbol which
623 was undefined in some object which appeared before the archive on the
624 command line, the linker will include the appropriate file(s) from the
625 archive. However, an undefined symbol in an object appearing later on
626 the command line will not cause the linker to search the archive again.
628 See the @option{-(} option for a way to force the linker to search
629 archives multiple times.
631 You may list the same archive multiple times on the command line.
634 This type of archive searching is standard for Unix linkers. However,
635 if you are using @command{ld} on AIX, note that it is different from the
636 behaviour of the AIX linker.
639 @cindex search directory, from cmd line
641 @kindex --library-path=@var{dir}
642 @item -L@var{searchdir}
643 @itemx --library-path=@var{searchdir}
644 Add path @var{searchdir} to the list of paths that @command{ld} will search
645 for archive libraries and @command{ld} control scripts. You may use this
646 option any number of times. The directories are searched in the order
647 in which they are specified on the command line. Directories specified
648 on the command line are searched before the default directories. All
649 @option{-L} options apply to all @option{-l} options, regardless of the
650 order in which the options appear.
652 If @var{searchdir} begins with @code{=}, then the @code{=} will be replaced
653 by the @dfn{sysroot prefix}, a path specified when the linker is configured.
656 The default set of paths searched (without being specified with
657 @samp{-L}) depends on which emulation mode @command{ld} is using, and in
658 some cases also on how it was configured. @xref{Environment}.
661 The paths can also be specified in a link script with the
662 @code{SEARCH_DIR} command. Directories specified this way are searched
663 at the point in which the linker script appears in the command line.
666 @kindex -m @var{emulation}
667 @item -m@var{emulation}
668 Emulate the @var{emulation} linker. You can list the available
669 emulations with the @samp{--verbose} or @samp{-V} options.
671 If the @samp{-m} option is not used, the emulation is taken from the
672 @code{LDEMULATION} environment variable, if that is defined.
674 Otherwise, the default emulation depends upon how the linker was
682 Print a link map to the standard output. A link map provides
683 information about the link, including the following:
687 Where object files are mapped into memory.
689 How common symbols are allocated.
691 All archive members included in the link, with a mention of the symbol
692 which caused the archive member to be brought in.
694 The values assigned to symbols.
696 Note - symbols whose values are computed by an expression which
697 involves a reference to a previous value of the same symbol may not
698 have correct result displayed in the link map. This is because the
699 linker discards intermediate results and only retains the final value
700 of an expression. Under such circumstances the linker will display
701 the final value enclosed by square brackets. Thus for example a
702 linker script containing:
710 will produce the following output in the link map if the @option{-M}
715 [0x0000000c] foo = (foo * 0x4)
716 [0x0000000c] foo = (foo + 0x8)
719 See @ref{Expressions} for more information about expressions in linker
724 @cindex read-only text
729 Turn off page alignment of sections, and mark the output as
730 @code{NMAGIC} if possible.
734 @cindex read/write from cmd line
738 Set the text and data sections to be readable and writable. Also, do
739 not page-align the data segment, and disable linking against shared
740 libraries. If the output format supports Unix style magic numbers,
741 mark the output as @code{OMAGIC}. Note: Although a writable text section
742 is allowed for PE-COFF targets, it does not conform to the format
743 specification published by Microsoft.
748 This option negates most of the effects of the @option{-N} option. It
749 sets the text section to be read-only, and forces the data segment to
750 be page-aligned. Note - this option does not enable linking against
751 shared libraries. Use @option{-Bdynamic} for this.
753 @kindex -o @var{output}
754 @kindex --output=@var{output}
755 @cindex naming the output file
756 @item -o @var{output}
757 @itemx --output=@var{output}
758 Use @var{output} as the name for the program produced by @command{ld}; if this
759 option is not specified, the name @file{a.out} is used by default. The
760 script command @code{OUTPUT} can also specify the output file name.
762 @kindex -O @var{level}
763 @cindex generating optimized output
765 If @var{level} is a numeric values greater than zero @command{ld} optimizes
766 the output. This might take significantly longer and therefore probably
767 should only be enabled for the final binary.
770 @kindex --emit-relocs
771 @cindex retain relocations in final executable
774 Leave relocation sections and contents in fully linked executables.
775 Post link analysis and optimization tools may need this information in
776 order to perform correct modifications of executables. This results
777 in larger executables.
779 This option is currently only supported on ELF platforms.
781 @kindex --force-dynamic
782 @cindex forcing the creation of dynamic sections
783 @item --force-dynamic
784 Force the output file to have dynamic sections. This option is specific
788 @cindex relocatable output
790 @kindex --relocatable
793 Generate relocatable output---i.e., generate an output file that can in
794 turn serve as input to @command{ld}. This is often called @dfn{partial
795 linking}. As a side effect, in environments that support standard Unix
796 magic numbers, this option also sets the output file's magic number to
798 @c ; see @option{-N}.
799 If this option is not specified, an absolute file is produced. When
800 linking C++ programs, this option @emph{will not} resolve references to
801 constructors; to do that, use @samp{-Ur}.
803 When an input file does not have the same format as the output file,
804 partial linking is only supported if that input file does not contain any
805 relocations. Different output formats can have further restrictions; for
806 example some @code{a.out}-based formats do not support partial linking
807 with input files in other formats at all.
809 This option does the same thing as @samp{-i}.
811 @kindex -R @var{file}
812 @kindex --just-symbols=@var{file}
813 @cindex symbol-only input
814 @item -R @var{filename}
815 @itemx --just-symbols=@var{filename}
816 Read symbol names and their addresses from @var{filename}, but do not
817 relocate it or include it in the output. This allows your output file
818 to refer symbolically to absolute locations of memory defined in other
819 programs. You may use this option more than once.
821 For compatibility with other ELF linkers, if the @option{-R} option is
822 followed by a directory name, rather than a file name, it is treated as
823 the @option{-rpath} option.
827 @cindex strip all symbols
830 Omit all symbol information from the output file.
833 @kindex --strip-debug
834 @cindex strip debugger symbols
837 Omit debugger symbol information (but not all symbols) from the output file.
841 @cindex input files, displaying
844 Print the names of the input files as @command{ld} processes them.
846 @kindex -T @var{script}
847 @kindex --script=@var{script}
849 @item -T @var{scriptfile}
850 @itemx --script=@var{scriptfile}
851 Use @var{scriptfile} as the linker script. This script replaces
852 @command{ld}'s default linker script (rather than adding to it), so
853 @var{commandfile} must specify everything necessary to describe the
854 output file. @xref{Scripts}. If @var{scriptfile} does not exist in
855 the current directory, @code{ld} looks for it in the directories
856 specified by any preceding @samp{-L} options. Multiple @samp{-T}
859 @kindex -u @var{symbol}
860 @kindex --undefined=@var{symbol}
861 @cindex undefined symbol
862 @item -u @var{symbol}
863 @itemx --undefined=@var{symbol}
864 Force @var{symbol} to be entered in the output file as an undefined
865 symbol. Doing this may, for example, trigger linking of additional
866 modules from standard libraries. @samp{-u} may be repeated with
867 different option arguments to enter additional undefined symbols. This
868 option is equivalent to the @code{EXTERN} linker script command.
873 For anything other than C++ programs, this option is equivalent to
874 @samp{-r}: it generates relocatable output---i.e., an output file that can in
875 turn serve as input to @command{ld}. When linking C++ programs, @samp{-Ur}
876 @emph{does} resolve references to constructors, unlike @samp{-r}.
877 It does not work to use @samp{-Ur} on files that were themselves linked
878 with @samp{-Ur}; once the constructor table has been built, it cannot
879 be added to. Use @samp{-Ur} only for the last partial link, and
880 @samp{-r} for the others.
882 @kindex --unique[=@var{SECTION}]
883 @item --unique[=@var{SECTION}]
884 Creates a separate output section for every input section matching
885 @var{SECTION}, or if the optional wildcard @var{SECTION} argument is
886 missing, for every orphan input section. An orphan section is one not
887 specifically mentioned in a linker script. You may use this option
888 multiple times on the command line; It prevents the normal merging of
889 input sections with the same name, overriding output section assignments
899 Display the version number for @command{ld}. The @option{-V} option also
900 lists the supported emulations.
903 @kindex --discard-all
904 @cindex deleting local symbols
907 Delete all local symbols.
910 @kindex --discard-locals
911 @cindex local symbols, deleting
912 @cindex L, deleting symbols beginning
914 @itemx --discard-locals
915 Delete all temporary local symbols. For most targets, this is all local
916 symbols whose names begin with @samp{L}.
918 @kindex -y @var{symbol}
919 @kindex --trace-symbol=@var{symbol}
920 @cindex symbol tracing
921 @item -y @var{symbol}
922 @itemx --trace-symbol=@var{symbol}
923 Print the name of each linked file in which @var{symbol} appears. This
924 option may be given any number of times. On many systems it is necessary
925 to prepend an underscore.
927 This option is useful when you have an undefined symbol in your link but
928 don't know where the reference is coming from.
930 @kindex -Y @var{path}
932 Add @var{path} to the default library search path. This option exists
933 for Solaris compatibility.
935 @kindex -z @var{keyword}
936 @item -z @var{keyword}
937 The recognized keywords are:
941 Combines multiple reloc sections and sorts them to make dynamic symbol
942 lookup caching possible.
945 Disallows undefined symbols in object files. Undefined symbols in
946 shared libraries are still allowed.
949 Marks the object as requiring executable stack.
952 This option is only meaningful when building a shared object.
953 It marks the object so that its runtime initialization will occur
954 before the runtime initialization of any other objects brought into
955 the process at the same time. Similarly the runtime finalization of
956 the object will occur after the runtime finalization of any other
960 Marks the object that its symbol table interposes before all symbols
961 but the primary executable.
964 When generating an executable or shared library, mark it to tell the
965 dynamic linker to defer function call resolution to the point when
966 the function is called (lazy binding), rather than at load time.
967 Lazy binding is the default.
970 Marks the object that its filters be processed immediately at
974 Allows multiple definitions.
977 Disables multiple reloc sections combining.
980 Disables production of copy relocs.
983 Marks the object that the search for dependencies of this object will
984 ignore any default library search paths.
987 Marks the object shouldn't be unloaded at runtime.
990 Marks the object not available to @code{dlopen}.
993 Marks the object can not be dumped by @code{dldump}.
996 Marks the object as not requiring executable stack.
999 Don't create an ELF @code{PT_GNU_RELRO} segment header in the object.
1002 When generating an executable or shared library, mark it to tell the
1003 dynamic linker to resolve all symbols when the program is started, or
1004 when the shared library is linked to using dlopen, instead of
1005 deferring function call resolution to the point when the function is
1009 Marks the object may contain $ORIGIN.
1012 Create an ELF @code{PT_GNU_RELRO} segment header in the object.
1014 @item max-page-size=@var{value}
1015 Set the emulation maximum page size to @var{value}.
1017 @item common-page-size=@var{value}
1018 Set the emulation common page size to @var{value}.
1022 Other keywords are ignored for Solaris compatibility.
1025 @cindex groups of archives
1026 @item -( @var{archives} -)
1027 @itemx --start-group @var{archives} --end-group
1028 The @var{archives} should be a list of archive files. They may be
1029 either explicit file names, or @samp{-l} options.
1031 The specified archives are searched repeatedly until no new undefined
1032 references are created. Normally, an archive is searched only once in
1033 the order that it is specified on the command line. If a symbol in that
1034 archive is needed to resolve an undefined symbol referred to by an
1035 object in an archive that appears later on the command line, the linker
1036 would not be able to resolve that reference. By grouping the archives,
1037 they all be searched repeatedly until all possible references are
1040 Using this option has a significant performance cost. It is best to use
1041 it only when there are unavoidable circular references between two or
1044 @kindex --accept-unknown-input-arch
1045 @kindex --no-accept-unknown-input-arch
1046 @item --accept-unknown-input-arch
1047 @itemx --no-accept-unknown-input-arch
1048 Tells the linker to accept input files whose architecture cannot be
1049 recognised. The assumption is that the user knows what they are doing
1050 and deliberately wants to link in these unknown input files. This was
1051 the default behaviour of the linker, before release 2.14. The default
1052 behaviour from release 2.14 onwards is to reject such input files, and
1053 so the @samp{--accept-unknown-input-arch} option has been added to
1054 restore the old behaviour.
1057 @kindex --no-as-needed
1059 @itemx --no-as-needed
1060 This option affects ELF DT_NEEDED tags for dynamic libraries mentioned
1061 on the command line after the @option{--as-needed} option. Normally,
1062 the linker will add a DT_NEEDED tag for each dynamic library mentioned
1063 on the command line, regardless of whether the library is actually
1064 needed. @option{--as-needed} causes DT_NEEDED tags to only be emitted
1065 for libraries that satisfy some symbol reference from regular objects
1066 which is undefined at the point that the library was linked.
1067 @option{--no-as-needed} restores the default behaviour.
1069 @kindex --add-needed
1070 @kindex --no-add-needed
1072 @itemx --no-add-needed
1073 This option affects the treatment of dynamic libraries from ELF
1074 DT_NEEDED tags in dynamic libraries mentioned on the command line after
1075 the @option{--no-add-needed} option. Normally, the linker will add
1076 a DT_NEEDED tag for each dynamic library from DT_NEEDED tags.
1077 @option{--no-add-needed} causes DT_NEEDED tags will never be emitted
1078 for those libraries from DT_NEEDED tags. @option{--add-needed} restores
1079 the default behaviour.
1081 @kindex -assert @var{keyword}
1082 @item -assert @var{keyword}
1083 This option is ignored for SunOS compatibility.
1087 @kindex -call_shared
1091 Link against dynamic libraries. This is only meaningful on platforms
1092 for which shared libraries are supported. This option is normally the
1093 default on such platforms. The different variants of this option are
1094 for compatibility with various systems. You may use this option
1095 multiple times on the command line: it affects library searching for
1096 @option{-l} options which follow it.
1100 Set the @code{DF_1_GROUP} flag in the @code{DT_FLAGS_1} entry in the dynamic
1101 section. This causes the runtime linker to handle lookups in this
1102 object and its dependencies to be performed only inside the group.
1103 @option{--unresolved-symbols=report-all} is implied. This option is
1104 only meaningful on ELF platforms which support shared libraries.
1114 Do not link against shared libraries. This is only meaningful on
1115 platforms for which shared libraries are supported. The different
1116 variants of this option are for compatibility with various systems. You
1117 may use this option multiple times on the command line: it affects
1118 library searching for @option{-l} options which follow it. This
1119 option also implies @option{--unresolved-symbols=report-all}. This
1120 option can be used with @option{-shared}. Doing so means that a
1121 shared library is being created but that all of the library's external
1122 references must be resolved by pulling in entries from static
1127 When creating a shared library, bind references to global symbols to the
1128 definition within the shared library, if any. Normally, it is possible
1129 for a program linked against a shared library to override the definition
1130 within the shared library. This option is only meaningful on ELF
1131 platforms which support shared libraries.
1133 @kindex --dynamic-list=@var{dynamic-list-file}
1134 @item --dynamic-list=@var{dynamic-list-file}
1135 Specify the name of a dynamic list file to the linker. This is
1136 typically used when creating shared libraries to specify a list of
1137 global symbols whose references shouldn't be bound to the definition
1138 within the shared library, or creating dynamically linked executables
1139 to specify a list of symbols which should be added to the symbol table
1140 in the executable. This option is only meaningful on ELF platforms
1141 which support shared libraries.
1143 The format of the dynamic list is the same as the version node without
1144 scope and node name. See @ref{VERSION} for more information.
1146 @kindex --dynamic-list-cpp-typeinfo
1147 @item --dynamic-list-cpp-typeinfo
1148 Provide the builtin dynamic list for C++ runtime type identification.
1150 @kindex --check-sections
1151 @kindex --no-check-sections
1152 @item --check-sections
1153 @itemx --no-check-sections
1154 Asks the linker @emph{not} to check section addresses after they have
1155 been assigned to see if there are any overlaps. Normally the linker will
1156 perform this check, and if it finds any overlaps it will produce
1157 suitable error messages. The linker does know about, and does make
1158 allowances for sections in overlays. The default behaviour can be
1159 restored by using the command line switch @option{--check-sections}.
1161 @cindex cross reference table
1164 Output a cross reference table. If a linker map file is being
1165 generated, the cross reference table is printed to the map file.
1166 Otherwise, it is printed on the standard output.
1168 The format of the table is intentionally simple, so that it may be
1169 easily processed by a script if necessary. The symbols are printed out,
1170 sorted by name. For each symbol, a list of file names is given. If the
1171 symbol is defined, the first file listed is the location of the
1172 definition. The remaining files contain references to the symbol.
1174 @cindex common allocation
1175 @kindex --no-define-common
1176 @item --no-define-common
1177 This option inhibits the assignment of addresses to common symbols.
1178 The script command @code{INHIBIT_COMMON_ALLOCATION} has the same effect.
1179 @xref{Miscellaneous Commands}.
1181 The @samp{--no-define-common} option allows decoupling
1182 the decision to assign addresses to Common symbols from the choice
1183 of the output file type; otherwise a non-Relocatable output type
1184 forces assigning addresses to Common symbols.
1185 Using @samp{--no-define-common} allows Common symbols that are referenced
1186 from a shared library to be assigned addresses only in the main program.
1187 This eliminates the unused duplicate space in the shared library,
1188 and also prevents any possible confusion over resolving to the wrong
1189 duplicate when there are many dynamic modules with specialized search
1190 paths for runtime symbol resolution.
1192 @cindex symbols, from command line
1193 @kindex --defsym @var{symbol}=@var{exp}
1194 @item --defsym @var{symbol}=@var{expression}
1195 Create a global symbol in the output file, containing the absolute
1196 address given by @var{expression}. You may use this option as many
1197 times as necessary to define multiple symbols in the command line. A
1198 limited form of arithmetic is supported for the @var{expression} in this
1199 context: you may give a hexadecimal constant or the name of an existing
1200 symbol, or use @code{+} and @code{-} to add or subtract hexadecimal
1201 constants or symbols. If you need more elaborate expressions, consider
1202 using the linker command language from a script (@pxref{Assignments,,
1203 Assignment: Symbol Definitions}). @emph{Note:} there should be no white
1204 space between @var{symbol}, the equals sign (``@key{=}''), and
1207 @cindex demangling, from command line
1208 @kindex --demangle[=@var{style}]
1209 @kindex --no-demangle
1210 @item --demangle[=@var{style}]
1211 @itemx --no-demangle
1212 These options control whether to demangle symbol names in error messages
1213 and other output. When the linker is told to demangle, it tries to
1214 present symbol names in a readable fashion: it strips leading
1215 underscores if they are used by the object file format, and converts C++
1216 mangled symbol names into user readable names. Different compilers have
1217 different mangling styles. The optional demangling style argument can be used
1218 to choose an appropriate demangling style for your compiler. The linker will
1219 demangle by default unless the environment variable @samp{COLLECT_NO_DEMANGLE}
1220 is set. These options may be used to override the default.
1222 @cindex dynamic linker, from command line
1223 @kindex -I@var{file}
1224 @kindex --dynamic-linker @var{file}
1225 @item --dynamic-linker @var{file}
1226 Set the name of the dynamic linker. This is only meaningful when
1227 generating dynamically linked ELF executables. The default dynamic
1228 linker is normally correct; don't use this unless you know what you are
1232 @kindex --fatal-warnings
1233 @item --fatal-warnings
1234 Treat all warnings as errors.
1236 @kindex --force-exe-suffix
1237 @item --force-exe-suffix
1238 Make sure that an output file has a .exe suffix.
1240 If a successfully built fully linked output file does not have a
1241 @code{.exe} or @code{.dll} suffix, this option forces the linker to copy
1242 the output file to one of the same name with a @code{.exe} suffix. This
1243 option is useful when using unmodified Unix makefiles on a Microsoft
1244 Windows host, since some versions of Windows won't run an image unless
1245 it ends in a @code{.exe} suffix.
1247 @kindex --gc-sections
1248 @kindex --no-gc-sections
1249 @cindex garbage collection
1251 @itemx --no-gc-sections
1252 Enable garbage collection of unused input sections. It is ignored on
1253 targets that do not support this option. This option is not compatible
1254 with @samp{-r}. The default behaviour (of not performing this garbage
1255 collection) can be restored by specifying @samp{--no-gc-sections} on
1258 @kindex --print-gc-sections
1259 @kindex --no-print-gc-sections
1260 @cindex garbage collection
1261 @item --print-gc-sections
1262 @itemx --no-print-gc-sections
1263 List all sections removed by garbage collection. The listing is
1264 printed on stderr. This option is only effective if garbage
1265 collection has been enabled via the @samp{--gc-sections}) option. The
1266 default behaviour (of not listing the sections that are removed) can
1267 be restored by specifying @samp{--no-print-gc-sections} on the command
1274 Print a summary of the command-line options on the standard output and exit.
1276 @kindex --target-help
1278 Print a summary of all target specific options on the standard output and exit.
1281 @item -Map @var{mapfile}
1282 Print a link map to the file @var{mapfile}. See the description of the
1283 @option{-M} option, above.
1285 @cindex memory usage
1286 @kindex --no-keep-memory
1287 @item --no-keep-memory
1288 @command{ld} normally optimizes for speed over memory usage by caching the
1289 symbol tables of input files in memory. This option tells @command{ld} to
1290 instead optimize for memory usage, by rereading the symbol tables as
1291 necessary. This may be required if @command{ld} runs out of memory space
1292 while linking a large executable.
1294 @kindex --no-undefined
1296 @item --no-undefined
1298 Report unresolved symbol references from regular object files. This
1299 is done even if the linker is creating a non-symbolic shared library.
1300 The switch @option{--[no-]allow-shlib-undefined} controls the
1301 behaviour for reporting unresolved references found in shared
1302 libraries being linked in.
1304 @kindex --allow-multiple-definition
1306 @item --allow-multiple-definition
1308 Normally when a symbol is defined multiple times, the linker will
1309 report a fatal error. These options allow multiple definitions and the
1310 first definition will be used.
1312 @kindex --allow-shlib-undefined
1313 @kindex --no-allow-shlib-undefined
1314 @item --allow-shlib-undefined
1315 @itemx --no-allow-shlib-undefined
1316 Allows (the default) or disallows undefined symbols in shared libraries.
1317 This switch is similar to @option{--no-undefined} except that it
1318 determines the behaviour when the undefined symbols are in a
1319 shared library rather than a regular object file. It does not affect
1320 how undefined symbols in regular object files are handled.
1322 The reason that @option{--allow-shlib-undefined} is the default is that
1323 the shared library being specified at link time may not be the same as
1324 the one that is available at load time, so the symbols might actually be
1325 resolvable at load time. Plus there are some systems, (eg BeOS) where
1326 undefined symbols in shared libraries is normal. (The kernel patches
1327 them at load time to select which function is most appropriate
1328 for the current architecture. This is used for example to dynamically
1329 select an appropriate memset function). Apparently it is also normal
1330 for HPPA shared libraries to have undefined symbols.
1332 @kindex --no-undefined-version
1333 @item --no-undefined-version
1334 Normally when a symbol has an undefined version, the linker will ignore
1335 it. This option disallows symbols with undefined version and a fatal error
1336 will be issued instead.
1338 @kindex --default-symver
1339 @item --default-symver
1340 Create and use a default symbol version (the soname) for unversioned
1343 @kindex --default-imported-symver
1344 @item --default-imported-symver
1345 Create and use a default symbol version (the soname) for unversioned
1348 @kindex --no-warn-mismatch
1349 @item --no-warn-mismatch
1350 Normally @command{ld} will give an error if you try to link together input
1351 files that are mismatched for some reason, perhaps because they have
1352 been compiled for different processors or for different endiannesses.
1353 This option tells @command{ld} that it should silently permit such possible
1354 errors. This option should only be used with care, in cases when you
1355 have taken some special action that ensures that the linker errors are
1358 @kindex --no-whole-archive
1359 @item --no-whole-archive
1360 Turn off the effect of the @option{--whole-archive} option for subsequent
1363 @cindex output file after errors
1364 @kindex --noinhibit-exec
1365 @item --noinhibit-exec
1366 Retain the executable output file whenever it is still usable.
1367 Normally, the linker will not produce an output file if it encounters
1368 errors during the link process; it exits without writing an output file
1369 when it issues any error whatsoever.
1373 Only search library directories explicitly specified on the
1374 command line. Library directories specified in linker scripts
1375 (including linker scripts specified on the command line) are ignored.
1377 @ifclear SingleFormat
1379 @item --oformat @var{output-format}
1380 @command{ld} may be configured to support more than one kind of object
1381 file. If your @command{ld} is configured this way, you can use the
1382 @samp{--oformat} option to specify the binary format for the output
1383 object file. Even when @command{ld} is configured to support alternative
1384 object formats, you don't usually need to specify this, as @command{ld}
1385 should be configured to produce as a default output format the most
1386 usual format on each machine. @var{output-format} is a text string, the
1387 name of a particular format supported by the BFD libraries. (You can
1388 list the available binary formats with @samp{objdump -i}.) The script
1389 command @code{OUTPUT_FORMAT} can also specify the output format, but
1390 this option overrides it. @xref{BFD}.
1394 @kindex --pic-executable
1396 @itemx --pic-executable
1397 @cindex position independent executables
1398 Create a position independent executable. This is currently only supported on
1399 ELF platforms. Position independent executables are similar to shared
1400 libraries in that they are relocated by the dynamic linker to the virtual
1401 address the OS chooses for them (which can vary between invocations). Like
1402 normal dynamically linked executables they can be executed and symbols
1403 defined in the executable cannot be overridden by shared libraries.
1407 This option is ignored for Linux compatibility.
1411 This option is ignored for SVR4 compatibility.
1414 @cindex synthesizing linker
1415 @cindex relaxing addressing modes
1417 An option with machine dependent effects.
1419 This option is only supported on a few targets.
1422 @xref{H8/300,,@command{ld} and the H8/300}.
1425 @xref{i960,, @command{ld} and the Intel 960 family}.
1428 @xref{Xtensa,, @command{ld} and Xtensa Processors}.
1431 @xref{M68HC11/68HC12,,@command{ld} and the 68HC11 and 68HC12}.
1434 @xref{PowerPC ELF32,,@command{ld} and PowerPC 32-bit ELF Support}.
1437 On some platforms, the @samp{--relax} option performs global
1438 optimizations that become possible when the linker resolves addressing
1439 in the program, such as relaxing address modes and synthesizing new
1440 instructions in the output object file.
1442 On some platforms these link time global optimizations may make symbolic
1443 debugging of the resulting executable impossible.
1446 the case for the Matsushita MN10200 and MN10300 family of processors.
1450 On platforms where this is not supported, @samp{--relax} is accepted,
1454 @cindex retaining specified symbols
1455 @cindex stripping all but some symbols
1456 @cindex symbols, retaining selectively
1457 @item --retain-symbols-file @var{filename}
1458 Retain @emph{only} the symbols listed in the file @var{filename},
1459 discarding all others. @var{filename} is simply a flat file, with one
1460 symbol name per line. This option is especially useful in environments
1464 where a large global symbol table is accumulated gradually, to conserve
1467 @samp{--retain-symbols-file} does @emph{not} discard undefined symbols,
1468 or symbols needed for relocations.
1470 You may only specify @samp{--retain-symbols-file} once in the command
1471 line. It overrides @samp{-s} and @samp{-S}.
1474 @item -rpath @var{dir}
1475 @cindex runtime library search path
1477 Add a directory to the runtime library search path. This is used when
1478 linking an ELF executable with shared objects. All @option{-rpath}
1479 arguments are concatenated and passed to the runtime linker, which uses
1480 them to locate shared objects at runtime. The @option{-rpath} option is
1481 also used when locating shared objects which are needed by shared
1482 objects explicitly included in the link; see the description of the
1483 @option{-rpath-link} option. If @option{-rpath} is not used when linking an
1484 ELF executable, the contents of the environment variable
1485 @code{LD_RUN_PATH} will be used if it is defined.
1487 The @option{-rpath} option may also be used on SunOS. By default, on
1488 SunOS, the linker will form a runtime search patch out of all the
1489 @option{-L} options it is given. If a @option{-rpath} option is used, the
1490 runtime search path will be formed exclusively using the @option{-rpath}
1491 options, ignoring the @option{-L} options. This can be useful when using
1492 gcc, which adds many @option{-L} options which may be on NFS mounted
1495 For compatibility with other ELF linkers, if the @option{-R} option is
1496 followed by a directory name, rather than a file name, it is treated as
1497 the @option{-rpath} option.
1501 @cindex link-time runtime library search path
1503 @item -rpath-link @var{DIR}
1504 When using ELF or SunOS, one shared library may require another. This
1505 happens when an @code{ld -shared} link includes a shared library as one
1508 When the linker encounters such a dependency when doing a non-shared,
1509 non-relocatable link, it will automatically try to locate the required
1510 shared library and include it in the link, if it is not included
1511 explicitly. In such a case, the @option{-rpath-link} option
1512 specifies the first set of directories to search. The
1513 @option{-rpath-link} option may specify a sequence of directory names
1514 either by specifying a list of names separated by colons, or by
1515 appearing multiple times.
1517 This option should be used with caution as it overrides the search path
1518 that may have been hard compiled into a shared library. In such a case it
1519 is possible to use unintentionally a different search path than the
1520 runtime linker would do.
1522 The linker uses the following search paths to locate required shared
1526 Any directories specified by @option{-rpath-link} options.
1528 Any directories specified by @option{-rpath} options. The difference
1529 between @option{-rpath} and @option{-rpath-link} is that directories
1530 specified by @option{-rpath} options are included in the executable and
1531 used at runtime, whereas the @option{-rpath-link} option is only effective
1532 at link time. Searching @option{-rpath} in this way is only supported
1533 by native linkers and cross linkers which have been configured with
1534 the @option{--with-sysroot} option.
1536 On an ELF system, if the @option{-rpath} and @code{rpath-link} options
1537 were not used, search the contents of the environment variable
1538 @code{LD_RUN_PATH}. It is for the native linker only.
1540 On SunOS, if the @option{-rpath} option was not used, search any
1541 directories specified using @option{-L} options.
1543 For a native linker, the contents of the environment variable
1544 @code{LD_LIBRARY_PATH}.
1546 For a native ELF linker, the directories in @code{DT_RUNPATH} or
1547 @code{DT_RPATH} of a shared library are searched for shared
1548 libraries needed by it. The @code{DT_RPATH} entries are ignored if
1549 @code{DT_RUNPATH} entries exist.
1551 The default directories, normally @file{/lib} and @file{/usr/lib}.
1553 For a native linker on an ELF system, if the file @file{/etc/ld.so.conf}
1554 exists, the list of directories found in that file.
1557 If the required shared library is not found, the linker will issue a
1558 warning and continue with the link.
1565 @cindex shared libraries
1566 Create a shared library. This is currently only supported on ELF, XCOFF
1567 and SunOS platforms. On SunOS, the linker will automatically create a
1568 shared library if the @option{-e} option is not used and there are
1569 undefined symbols in the link.
1572 @kindex --sort-common
1573 This option tells @command{ld} to sort the common symbols by size when it
1574 places them in the appropriate output sections. First come all the one
1575 byte symbols, then all the two byte, then all the four byte, and then
1576 everything else. This is to prevent gaps between symbols due to
1577 alignment constraints.
1579 @kindex --sort-section name
1580 @item --sort-section name
1581 This option will apply @code{SORT_BY_NAME} to all wildcard section
1582 patterns in the linker script.
1584 @kindex --sort-section alignment
1585 @item --sort-section alignment
1586 This option will apply @code{SORT_BY_ALIGNMENT} to all wildcard section
1587 patterns in the linker script.
1589 @kindex --split-by-file
1590 @item --split-by-file [@var{size}]
1591 Similar to @option{--split-by-reloc} but creates a new output section for
1592 each input file when @var{size} is reached. @var{size} defaults to a
1593 size of 1 if not given.
1595 @kindex --split-by-reloc
1596 @item --split-by-reloc [@var{count}]
1597 Tries to creates extra sections in the output file so that no single
1598 output section in the file contains more than @var{count} relocations.
1599 This is useful when generating huge relocatable files for downloading into
1600 certain real time kernels with the COFF object file format; since COFF
1601 cannot represent more than 65535 relocations in a single section. Note
1602 that this will fail to work with object file formats which do not
1603 support arbitrary sections. The linker will not split up individual
1604 input sections for redistribution, so if a single input section contains
1605 more than @var{count} relocations one output section will contain that
1606 many relocations. @var{count} defaults to a value of 32768.
1610 Compute and display statistics about the operation of the linker, such
1611 as execution time and memory usage.
1614 @item --sysroot=@var{directory}
1615 Use @var{directory} as the location of the sysroot, overriding the
1616 configure-time default. This option is only supported by linkers
1617 that were configured using @option{--with-sysroot}.
1619 @kindex --traditional-format
1620 @cindex traditional format
1621 @item --traditional-format
1622 For some targets, the output of @command{ld} is different in some ways from
1623 the output of some existing linker. This switch requests @command{ld} to
1624 use the traditional format instead.
1627 For example, on SunOS, @command{ld} combines duplicate entries in the
1628 symbol string table. This can reduce the size of an output file with
1629 full debugging information by over 30 percent. Unfortunately, the SunOS
1630 @code{dbx} program can not read the resulting program (@code{gdb} has no
1631 trouble). The @samp{--traditional-format} switch tells @command{ld} to not
1632 combine duplicate entries.
1634 @kindex --section-start @var{sectionname}=@var{org}
1635 @item --section-start @var{sectionname}=@var{org}
1636 Locate a section in the output file at the absolute
1637 address given by @var{org}. You may use this option as many
1638 times as necessary to locate multiple sections in the command
1640 @var{org} must be a single hexadecimal integer;
1641 for compatibility with other linkers, you may omit the leading
1642 @samp{0x} usually associated with hexadecimal values. @emph{Note:} there
1643 should be no white space between @var{sectionname}, the equals
1644 sign (``@key{=}''), and @var{org}.
1646 @kindex -Tbss @var{org}
1647 @kindex -Tdata @var{org}
1648 @kindex -Ttext @var{org}
1649 @cindex segment origins, cmd line
1650 @item -Tbss @var{org}
1651 @itemx -Tdata @var{org}
1652 @itemx -Ttext @var{org}
1653 Same as --section-start, with @code{.bss}, @code{.data} or
1654 @code{.text} as the @var{sectionname}.
1656 @kindex --unresolved-symbols
1657 @item --unresolved-symbols=@var{method}
1658 Determine how to handle unresolved symbols. There are four possible
1659 values for @samp{method}:
1663 Do not report any unresolved symbols.
1666 Report all unresolved symbols. This is the default.
1668 @item ignore-in-object-files
1669 Report unresolved symbols that are contained in shared libraries, but
1670 ignore them if they come from regular object files.
1672 @item ignore-in-shared-libs
1673 Report unresolved symbols that come from regular object files, but
1674 ignore them if they come from shared libraries. This can be useful
1675 when creating a dynamic binary and it is known that all the shared
1676 libraries that it should be referencing are included on the linker's
1680 The behaviour for shared libraries on their own can also be controlled
1681 by the @option{--[no-]allow-shlib-undefined} option.
1683 Normally the linker will generate an error message for each reported
1684 unresolved symbol but the option @option{--warn-unresolved-symbols}
1685 can change this to a warning.
1691 Display the version number for @command{ld} and list the linker emulations
1692 supported. Display which input files can and cannot be opened. Display
1693 the linker script being used by the linker.
1695 @kindex --version-script=@var{version-scriptfile}
1696 @cindex version script, symbol versions
1697 @itemx --version-script=@var{version-scriptfile}
1698 Specify the name of a version script to the linker. This is typically
1699 used when creating shared libraries to specify additional information
1700 about the version hierarchy for the library being created. This option
1701 is only meaningful on ELF platforms which support shared libraries.
1704 @kindex --warn-common
1705 @cindex warnings, on combining symbols
1706 @cindex combining symbols, warnings on
1708 Warn when a common symbol is combined with another common symbol or with
1709 a symbol definition. Unix linkers allow this somewhat sloppy practise,
1710 but linkers on some other operating systems do not. This option allows
1711 you to find potential problems from combining global symbols.
1712 Unfortunately, some C libraries use this practise, so you may get some
1713 warnings about symbols in the libraries as well as in your programs.
1715 There are three kinds of global symbols, illustrated here by C examples:
1719 A definition, which goes in the initialized data section of the output
1723 An undefined reference, which does not allocate space.
1724 There must be either a definition or a common symbol for the
1728 A common symbol. If there are only (one or more) common symbols for a
1729 variable, it goes in the uninitialized data area of the output file.
1730 The linker merges multiple common symbols for the same variable into a
1731 single symbol. If they are of different sizes, it picks the largest
1732 size. The linker turns a common symbol into a declaration, if there is
1733 a definition of the same variable.
1736 The @samp{--warn-common} option can produce five kinds of warnings.
1737 Each warning consists of a pair of lines: the first describes the symbol
1738 just encountered, and the second describes the previous symbol
1739 encountered with the same name. One or both of the two symbols will be
1744 Turning a common symbol into a reference, because there is already a
1745 definition for the symbol.
1747 @var{file}(@var{section}): warning: common of `@var{symbol}'
1748 overridden by definition
1749 @var{file}(@var{section}): warning: defined here
1753 Turning a common symbol into a reference, because a later definition for
1754 the symbol is encountered. This is the same as the previous case,
1755 except that the symbols are encountered in a different order.
1757 @var{file}(@var{section}): warning: definition of `@var{symbol}'
1759 @var{file}(@var{section}): warning: common is here
1763 Merging a common symbol with a previous same-sized common symbol.
1765 @var{file}(@var{section}): warning: multiple common
1767 @var{file}(@var{section}): warning: previous common is here
1771 Merging a common symbol with a previous larger common symbol.
1773 @var{file}(@var{section}): warning: common of `@var{symbol}'
1774 overridden by larger common
1775 @var{file}(@var{section}): warning: larger common is here
1779 Merging a common symbol with a previous smaller common symbol. This is
1780 the same as the previous case, except that the symbols are
1781 encountered in a different order.
1783 @var{file}(@var{section}): warning: common of `@var{symbol}'
1784 overriding smaller common
1785 @var{file}(@var{section}): warning: smaller common is here
1789 @kindex --warn-constructors
1790 @item --warn-constructors
1791 Warn if any global constructors are used. This is only useful for a few
1792 object file formats. For formats like COFF or ELF, the linker can not
1793 detect the use of global constructors.
1795 @kindex --warn-multiple-gp
1796 @item --warn-multiple-gp
1797 Warn if multiple global pointer values are required in the output file.
1798 This is only meaningful for certain processors, such as the Alpha.
1799 Specifically, some processors put large-valued constants in a special
1800 section. A special register (the global pointer) points into the middle
1801 of this section, so that constants can be loaded efficiently via a
1802 base-register relative addressing mode. Since the offset in
1803 base-register relative mode is fixed and relatively small (e.g., 16
1804 bits), this limits the maximum size of the constant pool. Thus, in
1805 large programs, it is often necessary to use multiple global pointer
1806 values in order to be able to address all possible constants. This
1807 option causes a warning to be issued whenever this case occurs.
1810 @cindex warnings, on undefined symbols
1811 @cindex undefined symbols, warnings on
1813 Only warn once for each undefined symbol, rather than once per module
1816 @kindex --warn-section-align
1817 @cindex warnings, on section alignment
1818 @cindex section alignment, warnings on
1819 @item --warn-section-align
1820 Warn if the address of an output section is changed because of
1821 alignment. Typically, the alignment will be set by an input section.
1822 The address will only be changed if it not explicitly specified; that
1823 is, if the @code{SECTIONS} command does not specify a start address for
1824 the section (@pxref{SECTIONS}).
1826 @kindex --warn-shared-textrel
1827 @item --warn-shared-textrel
1828 Warn if the linker adds a DT_TEXTREL to a shared object.
1830 @kindex --warn-unresolved-symbols
1831 @item --warn-unresolved-symbols
1832 If the linker is going to report an unresolved symbol (see the option
1833 @option{--unresolved-symbols}) it will normally generate an error.
1834 This option makes it generate a warning instead.
1836 @kindex --error-unresolved-symbols
1837 @item --error-unresolved-symbols
1838 This restores the linker's default behaviour of generating errors when
1839 it is reporting unresolved symbols.
1841 @kindex --whole-archive
1842 @cindex including an entire archive
1843 @item --whole-archive
1844 For each archive mentioned on the command line after the
1845 @option{--whole-archive} option, include every object file in the archive
1846 in the link, rather than searching the archive for the required object
1847 files. This is normally used to turn an archive file into a shared
1848 library, forcing every object to be included in the resulting shared
1849 library. This option may be used more than once.
1851 Two notes when using this option from gcc: First, gcc doesn't know
1852 about this option, so you have to use @option{-Wl,-whole-archive}.
1853 Second, don't forget to use @option{-Wl,-no-whole-archive} after your
1854 list of archives, because gcc will add its own list of archives to
1855 your link and you may not want this flag to affect those as well.
1858 @item --wrap @var{symbol}
1859 Use a wrapper function for @var{symbol}. Any undefined reference to
1860 @var{symbol} will be resolved to @code{__wrap_@var{symbol}}. Any
1861 undefined reference to @code{__real_@var{symbol}} will be resolved to
1864 This can be used to provide a wrapper for a system function. The
1865 wrapper function should be called @code{__wrap_@var{symbol}}. If it
1866 wishes to call the system function, it should call
1867 @code{__real_@var{symbol}}.
1869 Here is a trivial example:
1873 __wrap_malloc (size_t c)
1875 printf ("malloc called with %zu\n", c);
1876 return __real_malloc (c);
1880 If you link other code with this file using @option{--wrap malloc}, then
1881 all calls to @code{malloc} will call the function @code{__wrap_malloc}
1882 instead. The call to @code{__real_malloc} in @code{__wrap_malloc} will
1883 call the real @code{malloc} function.
1885 You may wish to provide a @code{__real_malloc} function as well, so that
1886 links without the @option{--wrap} option will succeed. If you do this,
1887 you should not put the definition of @code{__real_malloc} in the same
1888 file as @code{__wrap_malloc}; if you do, the assembler may resolve the
1889 call before the linker has a chance to wrap it to @code{malloc}.
1891 @kindex --eh-frame-hdr
1892 @item --eh-frame-hdr
1893 Request creation of @code{.eh_frame_hdr} section and ELF
1894 @code{PT_GNU_EH_FRAME} segment header.
1896 @kindex --enable-new-dtags
1897 @kindex --disable-new-dtags
1898 @item --enable-new-dtags
1899 @itemx --disable-new-dtags
1900 This linker can create the new dynamic tags in ELF. But the older ELF
1901 systems may not understand them. If you specify
1902 @option{--enable-new-dtags}, the dynamic tags will be created as needed.
1903 If you specify @option{--disable-new-dtags}, no new dynamic tags will be
1904 created. By default, the new dynamic tags are not created. Note that
1905 those options are only available for ELF systems.
1907 @kindex --hash-size=@var{number}
1908 @item --hash-size=@var{number}
1909 Set the default size of the linker's hash tables to a prime number
1910 close to @var{number}. Increasing this value can reduce the length of
1911 time it takes the linker to perform its tasks, at the expense of
1912 increasing the linker's memory requirements. Similarly reducing this
1913 value can reduce the memory requirements at the expense of speed.
1915 @kindex --hash-style=@var{style}
1916 @item --hash-style=@var{style}
1917 Set the type of linker's hash table(s). @var{style} can be either
1918 @code{sysv} for classic ELF @code{.hash} section, @code{gnu} for
1919 new style GNU @code{.gnu.hash} section or @code{both} for both
1920 the classic ELF @code{.hash} and new style GNU @code{.gnu.hash}
1921 hash tables. The default is @code{sysv}.
1923 @kindex --reduce-memory-overheads
1924 @item --reduce-memory-overheads
1925 This option reduces memory requirements at ld runtime, at the expense of
1926 linking speed. This was introduced to select the old O(n^2) algorithm
1927 for link map file generation, rather than the new O(n) algorithm which uses
1928 about 40% more memory for symbol storage.
1930 Another effect of the switch is to set the default hash table size to
1931 1021, which again saves memory at the cost of lengthening the linker's
1932 run time. This is not done however if the @option{--hash-size} switch
1935 The @option{--reduce-memory-overheads} switch may be also be used to
1936 enable other tradeoffs in future versions of the linker.
1942 @subsection Options Specific to i386 PE Targets
1944 @c man begin OPTIONS
1946 The i386 PE linker supports the @option{-shared} option, which causes
1947 the output to be a dynamically linked library (DLL) instead of a
1948 normal executable. You should name the output @code{*.dll} when you
1949 use this option. In addition, the linker fully supports the standard
1950 @code{*.def} files, which may be specified on the linker command line
1951 like an object file (in fact, it should precede archives it exports
1952 symbols from, to ensure that they get linked in, just like a normal
1955 In addition to the options common to all targets, the i386 PE linker
1956 support additional command line options that are specific to the i386
1957 PE target. Options that take values may be separated from their
1958 values by either a space or an equals sign.
1962 @kindex --add-stdcall-alias
1963 @item --add-stdcall-alias
1964 If given, symbols with a stdcall suffix (@@@var{nn}) will be exported
1965 as-is and also with the suffix stripped.
1966 [This option is specific to the i386 PE targeted port of the linker]
1969 @item --base-file @var{file}
1970 Use @var{file} as the name of a file in which to save the base
1971 addresses of all the relocations needed for generating DLLs with
1973 [This is an i386 PE specific option]
1977 Create a DLL instead of a regular executable. You may also use
1978 @option{-shared} or specify a @code{LIBRARY} in a given @code{.def}
1980 [This option is specific to the i386 PE targeted port of the linker]
1982 @kindex --enable-stdcall-fixup
1983 @kindex --disable-stdcall-fixup
1984 @item --enable-stdcall-fixup
1985 @itemx --disable-stdcall-fixup
1986 If the link finds a symbol that it cannot resolve, it will attempt to
1987 do ``fuzzy linking'' by looking for another defined symbol that differs
1988 only in the format of the symbol name (cdecl vs stdcall) and will
1989 resolve that symbol by linking to the match. For example, the
1990 undefined symbol @code{_foo} might be linked to the function
1991 @code{_foo@@12}, or the undefined symbol @code{_bar@@16} might be linked
1992 to the function @code{_bar}. When the linker does this, it prints a
1993 warning, since it normally should have failed to link, but sometimes
1994 import libraries generated from third-party dlls may need this feature
1995 to be usable. If you specify @option{--enable-stdcall-fixup}, this
1996 feature is fully enabled and warnings are not printed. If you specify
1997 @option{--disable-stdcall-fixup}, this feature is disabled and such
1998 mismatches are considered to be errors.
1999 [This option is specific to the i386 PE targeted port of the linker]
2001 @cindex DLLs, creating
2002 @kindex --export-all-symbols
2003 @item --export-all-symbols
2004 If given, all global symbols in the objects used to build a DLL will
2005 be exported by the DLL. Note that this is the default if there
2006 otherwise wouldn't be any exported symbols. When symbols are
2007 explicitly exported via DEF files or implicitly exported via function
2008 attributes, the default is to not export anything else unless this
2009 option is given. Note that the symbols @code{DllMain@@12},
2010 @code{DllEntryPoint@@0}, @code{DllMainCRTStartup@@12}, and
2011 @code{impure_ptr} will not be automatically
2012 exported. Also, symbols imported from other DLLs will not be
2013 re-exported, nor will symbols specifying the DLL's internal layout
2014 such as those beginning with @code{_head_} or ending with
2015 @code{_iname}. In addition, no symbols from @code{libgcc},
2016 @code{libstd++}, @code{libmingw32}, or @code{crtX.o} will be exported.
2017 Symbols whose names begin with @code{__rtti_} or @code{__builtin_} will
2018 not be exported, to help with C++ DLLs. Finally, there is an
2019 extensive list of cygwin-private symbols that are not exported
2020 (obviously, this applies on when building DLLs for cygwin targets).
2021 These cygwin-excludes are: @code{_cygwin_dll_entry@@12},
2022 @code{_cygwin_crt0_common@@8}, @code{_cygwin_noncygwin_dll_entry@@12},
2023 @code{_fmode}, @code{_impure_ptr}, @code{cygwin_attach_dll},
2024 @code{cygwin_premain0}, @code{cygwin_premain1}, @code{cygwin_premain2},
2025 @code{cygwin_premain3}, and @code{environ}.
2026 [This option is specific to the i386 PE targeted port of the linker]
2028 @kindex --exclude-symbols
2029 @item --exclude-symbols @var{symbol},@var{symbol},...
2030 Specifies a list of symbols which should not be automatically
2031 exported. The symbol names may be delimited by commas or colons.
2032 [This option is specific to the i386 PE targeted port of the linker]
2034 @kindex --file-alignment
2035 @item --file-alignment
2036 Specify the file alignment. Sections in the file will always begin at
2037 file offsets which are multiples of this number. This defaults to
2039 [This option is specific to the i386 PE targeted port of the linker]
2043 @item --heap @var{reserve}
2044 @itemx --heap @var{reserve},@var{commit}
2045 Specify the amount of memory to reserve (and optionally commit) to be
2046 used as heap for this program. The default is 1Mb reserved, 4K
2048 [This option is specific to the i386 PE targeted port of the linker]
2051 @kindex --image-base
2052 @item --image-base @var{value}
2053 Use @var{value} as the base address of your program or dll. This is
2054 the lowest memory location that will be used when your program or dll
2055 is loaded. To reduce the need to relocate and improve performance of
2056 your dlls, each should have a unique base address and not overlap any
2057 other dlls. The default is 0x400000 for executables, and 0x10000000
2059 [This option is specific to the i386 PE targeted port of the linker]
2063 If given, the stdcall suffixes (@@@var{nn}) will be stripped from
2064 symbols before they are exported.
2065 [This option is specific to the i386 PE targeted port of the linker]
2067 @kindex --large-address-aware
2068 @item --large-address-aware
2069 If given, the appropriate bit in the ``Characteristics'' field of the COFF
2070 header is set to indicate that this executable supports virtual addresses
2071 greater than 2 gigabytes. This should be used in conjunction with the /3GB
2072 or /USERVA=@var{value} megabytes switch in the ``[operating systems]''
2073 section of the BOOT.INI. Otherwise, this bit has no effect.
2074 [This option is specific to PE targeted ports of the linker]
2076 @kindex --major-image-version
2077 @item --major-image-version @var{value}
2078 Sets the major number of the ``image version''. Defaults to 1.
2079 [This option is specific to the i386 PE targeted port of the linker]
2081 @kindex --major-os-version
2082 @item --major-os-version @var{value}
2083 Sets the major number of the ``os version''. Defaults to 4.
2084 [This option is specific to the i386 PE targeted port of the linker]
2086 @kindex --major-subsystem-version
2087 @item --major-subsystem-version @var{value}
2088 Sets the major number of the ``subsystem version''. Defaults to 4.
2089 [This option is specific to the i386 PE targeted port of the linker]
2091 @kindex --minor-image-version
2092 @item --minor-image-version @var{value}
2093 Sets the minor number of the ``image version''. Defaults to 0.
2094 [This option is specific to the i386 PE targeted port of the linker]
2096 @kindex --minor-os-version
2097 @item --minor-os-version @var{value}
2098 Sets the minor number of the ``os version''. Defaults to 0.
2099 [This option is specific to the i386 PE targeted port of the linker]
2101 @kindex --minor-subsystem-version
2102 @item --minor-subsystem-version @var{value}
2103 Sets the minor number of the ``subsystem version''. Defaults to 0.
2104 [This option is specific to the i386 PE targeted port of the linker]
2106 @cindex DEF files, creating
2107 @cindex DLLs, creating
2108 @kindex --output-def
2109 @item --output-def @var{file}
2110 The linker will create the file @var{file} which will contain a DEF
2111 file corresponding to the DLL the linker is generating. This DEF file
2112 (which should be called @code{*.def}) may be used to create an import
2113 library with @code{dlltool} or may be used as a reference to
2114 automatically or implicitly exported symbols.
2115 [This option is specific to the i386 PE targeted port of the linker]
2117 @cindex DLLs, creating
2118 @kindex --out-implib
2119 @item --out-implib @var{file}
2120 The linker will create the file @var{file} which will contain an
2121 import lib corresponding to the DLL the linker is generating. This
2122 import lib (which should be called @code{*.dll.a} or @code{*.a}
2123 may be used to link clients against the generated DLL; this behaviour
2124 makes it possible to skip a separate @code{dlltool} import library
2126 [This option is specific to the i386 PE targeted port of the linker]
2128 @kindex --enable-auto-image-base
2129 @item --enable-auto-image-base
2130 Automatically choose the image base for DLLs, unless one is specified
2131 using the @code{--image-base} argument. By using a hash generated
2132 from the dllname to create unique image bases for each DLL, in-memory
2133 collisions and relocations which can delay program execution are
2135 [This option is specific to the i386 PE targeted port of the linker]
2137 @kindex --disable-auto-image-base
2138 @item --disable-auto-image-base
2139 Do not automatically generate a unique image base. If there is no
2140 user-specified image base (@code{--image-base}) then use the platform
2142 [This option is specific to the i386 PE targeted port of the linker]
2144 @cindex DLLs, linking to
2145 @kindex --dll-search-prefix
2146 @item --dll-search-prefix @var{string}
2147 When linking dynamically to a dll without an import library,
2148 search for @code{<string><basename>.dll} in preference to
2149 @code{lib<basename>.dll}. This behaviour allows easy distinction
2150 between DLLs built for the various "subplatforms": native, cygwin,
2151 uwin, pw, etc. For instance, cygwin DLLs typically use
2152 @code{--dll-search-prefix=cyg}.
2153 [This option is specific to the i386 PE targeted port of the linker]
2155 @kindex --enable-auto-import
2156 @item --enable-auto-import
2157 Do sophisticated linking of @code{_symbol} to @code{__imp__symbol} for
2158 DATA imports from DLLs, and create the necessary thunking symbols when
2159 building the import libraries with those DATA exports. Note: Use of the
2160 'auto-import' extension will cause the text section of the image file
2161 to be made writable. This does not conform to the PE-COFF format
2162 specification published by Microsoft.
2164 Using 'auto-import' generally will 'just work' -- but sometimes you may
2167 "variable '<var>' can't be auto-imported. Please read the
2168 documentation for ld's @code{--enable-auto-import} for details."
2170 This message occurs when some (sub)expression accesses an address
2171 ultimately given by the sum of two constants (Win32 import tables only
2172 allow one). Instances where this may occur include accesses to member
2173 fields of struct variables imported from a DLL, as well as using a
2174 constant index into an array variable imported from a DLL. Any
2175 multiword variable (arrays, structs, long long, etc) may trigger
2176 this error condition. However, regardless of the exact data type
2177 of the offending exported variable, ld will always detect it, issue
2178 the warning, and exit.
2180 There are several ways to address this difficulty, regardless of the
2181 data type of the exported variable:
2183 One way is to use --enable-runtime-pseudo-reloc switch. This leaves the task
2184 of adjusting references in your client code for runtime environment, so
2185 this method works only when runtime environment supports this feature.
2187 A second solution is to force one of the 'constants' to be a variable --
2188 that is, unknown and un-optimizable at compile time. For arrays,
2189 there are two possibilities: a) make the indexee (the array's address)
2190 a variable, or b) make the 'constant' index a variable. Thus:
2193 extern type extern_array[];
2195 @{ volatile type *t=extern_array; t[1] @}
2201 extern type extern_array[];
2203 @{ volatile int t=1; extern_array[t] @}
2206 For structs (and most other multiword data types) the only option
2207 is to make the struct itself (or the long long, or the ...) variable:
2210 extern struct s extern_struct;
2211 extern_struct.field -->
2212 @{ volatile struct s *t=&extern_struct; t->field @}
2218 extern long long extern_ll;
2220 @{ volatile long long * local_ll=&extern_ll; *local_ll @}
2223 A third method of dealing with this difficulty is to abandon
2224 'auto-import' for the offending symbol and mark it with
2225 @code{__declspec(dllimport)}. However, in practise that
2226 requires using compile-time #defines to indicate whether you are
2227 building a DLL, building client code that will link to the DLL, or
2228 merely building/linking to a static library. In making the choice
2229 between the various methods of resolving the 'direct address with
2230 constant offset' problem, you should consider typical real-world usage:
2238 void main(int argc, char **argv)@{
2239 printf("%d\n",arr[1]);
2249 void main(int argc, char **argv)@{
2250 /* This workaround is for win32 and cygwin; do not "optimize" */
2251 volatile int *parr = arr;
2252 printf("%d\n",parr[1]);
2259 /* Note: auto-export is assumed (no __declspec(dllexport)) */
2260 #if (defined(_WIN32) || defined(__CYGWIN__)) && \
2261 !(defined(FOO_BUILD_DLL) || defined(FOO_STATIC))
2262 #define FOO_IMPORT __declspec(dllimport)
2266 extern FOO_IMPORT int arr[];
2269 void main(int argc, char **argv)@{
2270 printf("%d\n",arr[1]);
2274 A fourth way to avoid this problem is to re-code your
2275 library to use a functional interface rather than a data interface
2276 for the offending variables (e.g. set_foo() and get_foo() accessor
2278 [This option is specific to the i386 PE targeted port of the linker]
2280 @kindex --disable-auto-import
2281 @item --disable-auto-import
2282 Do not attempt to do sophisticated linking of @code{_symbol} to
2283 @code{__imp__symbol} for DATA imports from DLLs.
2284 [This option is specific to the i386 PE targeted port of the linker]
2286 @kindex --enable-runtime-pseudo-reloc
2287 @item --enable-runtime-pseudo-reloc
2288 If your code contains expressions described in --enable-auto-import section,
2289 that is, DATA imports from DLL with non-zero offset, this switch will create
2290 a vector of 'runtime pseudo relocations' which can be used by runtime
2291 environment to adjust references to such data in your client code.
2292 [This option is specific to the i386 PE targeted port of the linker]
2294 @kindex --disable-runtime-pseudo-reloc
2295 @item --disable-runtime-pseudo-reloc
2296 Do not create pseudo relocations for non-zero offset DATA imports from
2297 DLLs. This is the default.
2298 [This option is specific to the i386 PE targeted port of the linker]
2300 @kindex --enable-extra-pe-debug
2301 @item --enable-extra-pe-debug
2302 Show additional debug info related to auto-import symbol thunking.
2303 [This option is specific to the i386 PE targeted port of the linker]
2305 @kindex --section-alignment
2306 @item --section-alignment
2307 Sets the section alignment. Sections in memory will always begin at
2308 addresses which are a multiple of this number. Defaults to 0x1000.
2309 [This option is specific to the i386 PE targeted port of the linker]
2313 @item --stack @var{reserve}
2314 @itemx --stack @var{reserve},@var{commit}
2315 Specify the amount of memory to reserve (and optionally commit) to be
2316 used as stack for this program. The default is 2Mb reserved, 4K
2318 [This option is specific to the i386 PE targeted port of the linker]
2321 @item --subsystem @var{which}
2322 @itemx --subsystem @var{which}:@var{major}
2323 @itemx --subsystem @var{which}:@var{major}.@var{minor}
2324 Specifies the subsystem under which your program will execute. The
2325 legal values for @var{which} are @code{native}, @code{windows},
2326 @code{console}, @code{posix}, and @code{xbox}. You may optionally set
2327 the subsystem version also. Numeric values are also accepted for
2329 [This option is specific to the i386 PE targeted port of the linker]
2336 @subsection Options specific to Motorola 68HC11 and 68HC12 targets
2338 @c man begin OPTIONS
2340 The 68HC11 and 68HC12 linkers support specific options to control the
2341 memory bank switching mapping and trampoline code generation.
2345 @kindex --no-trampoline
2346 @item --no-trampoline
2347 This option disables the generation of trampoline. By default a trampoline
2348 is generated for each far function which is called using a @code{jsr}
2349 instruction (this happens when a pointer to a far function is taken).
2351 @kindex --bank-window
2352 @item --bank-window @var{name}
2353 This option indicates to the linker the name of the memory region in
2354 the @samp{MEMORY} specification that describes the memory bank window.
2355 The definition of such region is then used by the linker to compute
2356 paging and addresses within the memory window.
2365 @section Environment Variables
2367 @c man begin ENVIRONMENT
2369 You can change the behaviour of @command{ld} with the environment variables
2370 @ifclear SingleFormat
2373 @code{LDEMULATION} and @code{COLLECT_NO_DEMANGLE}.
2375 @ifclear SingleFormat
2377 @cindex default input format
2378 @code{GNUTARGET} determines the input-file object format if you don't
2379 use @samp{-b} (or its synonym @samp{--format}). Its value should be one
2380 of the BFD names for an input format (@pxref{BFD}). If there is no
2381 @code{GNUTARGET} in the environment, @command{ld} uses the natural format
2382 of the target. If @code{GNUTARGET} is set to @code{default} then BFD
2383 attempts to discover the input format by examining binary input files;
2384 this method often succeeds, but there are potential ambiguities, since
2385 there is no method of ensuring that the magic number used to specify
2386 object-file formats is unique. However, the configuration procedure for
2387 BFD on each system places the conventional format for that system first
2388 in the search-list, so ambiguities are resolved in favor of convention.
2392 @cindex default emulation
2393 @cindex emulation, default
2394 @code{LDEMULATION} determines the default emulation if you don't use the
2395 @samp{-m} option. The emulation can affect various aspects of linker
2396 behaviour, particularly the default linker script. You can list the
2397 available emulations with the @samp{--verbose} or @samp{-V} options. If
2398 the @samp{-m} option is not used, and the @code{LDEMULATION} environment
2399 variable is not defined, the default emulation depends upon how the
2400 linker was configured.
2402 @kindex COLLECT_NO_DEMANGLE
2403 @cindex demangling, default
2404 Normally, the linker will default to demangling symbols. However, if
2405 @code{COLLECT_NO_DEMANGLE} is set in the environment, then it will
2406 default to not demangling symbols. This environment variable is used in
2407 a similar fashion by the @code{gcc} linker wrapper program. The default
2408 may be overridden by the @samp{--demangle} and @samp{--no-demangle}
2415 @chapter Linker Scripts
2418 @cindex linker scripts
2419 @cindex command files
2420 Every link is controlled by a @dfn{linker script}. This script is
2421 written in the linker command language.
2423 The main purpose of the linker script is to describe how the sections in
2424 the input files should be mapped into the output file, and to control
2425 the memory layout of the output file. Most linker scripts do nothing
2426 more than this. However, when necessary, the linker script can also
2427 direct the linker to perform many other operations, using the commands
2430 The linker always uses a linker script. If you do not supply one
2431 yourself, the linker will use a default script that is compiled into the
2432 linker executable. You can use the @samp{--verbose} command line option
2433 to display the default linker script. Certain command line options,
2434 such as @samp{-r} or @samp{-N}, will affect the default linker script.
2436 You may supply your own linker script by using the @samp{-T} command
2437 line option. When you do this, your linker script will replace the
2438 default linker script.
2440 You may also use linker scripts implicitly by naming them as input files
2441 to the linker, as though they were files to be linked. @xref{Implicit
2445 * Basic Script Concepts:: Basic Linker Script Concepts
2446 * Script Format:: Linker Script Format
2447 * Simple Example:: Simple Linker Script Example
2448 * Simple Commands:: Simple Linker Script Commands
2449 * Assignments:: Assigning Values to Symbols
2450 * SECTIONS:: SECTIONS Command
2451 * MEMORY:: MEMORY Command
2452 * PHDRS:: PHDRS Command
2453 * VERSION:: VERSION Command
2454 * Expressions:: Expressions in Linker Scripts
2455 * Implicit Linker Scripts:: Implicit Linker Scripts
2458 @node Basic Script Concepts
2459 @section Basic Linker Script Concepts
2460 @cindex linker script concepts
2461 We need to define some basic concepts and vocabulary in order to
2462 describe the linker script language.
2464 The linker combines input files into a single output file. The output
2465 file and each input file are in a special data format known as an
2466 @dfn{object file format}. Each file is called an @dfn{object file}.
2467 The output file is often called an @dfn{executable}, but for our
2468 purposes we will also call it an object file. Each object file has,
2469 among other things, a list of @dfn{sections}. We sometimes refer to a
2470 section in an input file as an @dfn{input section}; similarly, a section
2471 in the output file is an @dfn{output section}.
2473 Each section in an object file has a name and a size. Most sections
2474 also have an associated block of data, known as the @dfn{section
2475 contents}. A section may be marked as @dfn{loadable}, which mean that
2476 the contents should be loaded into memory when the output file is run.
2477 A section with no contents may be @dfn{allocatable}, which means that an
2478 area in memory should be set aside, but nothing in particular should be
2479 loaded there (in some cases this memory must be zeroed out). A section
2480 which is neither loadable nor allocatable typically contains some sort
2481 of debugging information.
2483 Every loadable or allocatable output section has two addresses. The
2484 first is the @dfn{VMA}, or virtual memory address. This is the address
2485 the section will have when the output file is run. The second is the
2486 @dfn{LMA}, or load memory address. This is the address at which the
2487 section will be loaded. In most cases the two addresses will be the
2488 same. An example of when they might be different is when a data section
2489 is loaded into ROM, and then copied into RAM when the program starts up
2490 (this technique is often used to initialize global variables in a ROM
2491 based system). In this case the ROM address would be the LMA, and the
2492 RAM address would be the VMA.
2494 You can see the sections in an object file by using the @code{objdump}
2495 program with the @samp{-h} option.
2497 Every object file also has a list of @dfn{symbols}, known as the
2498 @dfn{symbol table}. A symbol may be defined or undefined. Each symbol
2499 has a name, and each defined symbol has an address, among other
2500 information. If you compile a C or C++ program into an object file, you
2501 will get a defined symbol for every defined function and global or
2502 static variable. Every undefined function or global variable which is
2503 referenced in the input file will become an undefined symbol.
2505 You can see the symbols in an object file by using the @code{nm}
2506 program, or by using the @code{objdump} program with the @samp{-t}
2510 @section Linker Script Format
2511 @cindex linker script format
2512 Linker scripts are text files.
2514 You write a linker script as a series of commands. Each command is
2515 either a keyword, possibly followed by arguments, or an assignment to a
2516 symbol. You may separate commands using semicolons. Whitespace is
2519 Strings such as file or format names can normally be entered directly.
2520 If the file name contains a character such as a comma which would
2521 otherwise serve to separate file names, you may put the file name in
2522 double quotes. There is no way to use a double quote character in a
2525 You may include comments in linker scripts just as in C, delimited by
2526 @samp{/*} and @samp{*/}. As in C, comments are syntactically equivalent
2529 @node Simple Example
2530 @section Simple Linker Script Example
2531 @cindex linker script example
2532 @cindex example of linker script
2533 Many linker scripts are fairly simple.
2535 The simplest possible linker script has just one command:
2536 @samp{SECTIONS}. You use the @samp{SECTIONS} command to describe the
2537 memory layout of the output file.
2539 The @samp{SECTIONS} command is a powerful command. Here we will
2540 describe a simple use of it. Let's assume your program consists only of
2541 code, initialized data, and uninitialized data. These will be in the
2542 @samp{.text}, @samp{.data}, and @samp{.bss} sections, respectively.
2543 Let's assume further that these are the only sections which appear in
2546 For this example, let's say that the code should be loaded at address
2547 0x10000, and that the data should start at address 0x8000000. Here is a
2548 linker script which will do that:
2553 .text : @{ *(.text) @}
2555 .data : @{ *(.data) @}
2556 .bss : @{ *(.bss) @}
2560 You write the @samp{SECTIONS} command as the keyword @samp{SECTIONS},
2561 followed by a series of symbol assignments and output section
2562 descriptions enclosed in curly braces.
2564 The first line inside the @samp{SECTIONS} command of the above example
2565 sets the value of the special symbol @samp{.}, which is the location
2566 counter. If you do not specify the address of an output section in some
2567 other way (other ways are described later), the address is set from the
2568 current value of the location counter. The location counter is then
2569 incremented by the size of the output section. At the start of the
2570 @samp{SECTIONS} command, the location counter has the value @samp{0}.
2572 The second line defines an output section, @samp{.text}. The colon is
2573 required syntax which may be ignored for now. Within the curly braces
2574 after the output section name, you list the names of the input sections
2575 which should be placed into this output section. The @samp{*} is a
2576 wildcard which matches any file name. The expression @samp{*(.text)}
2577 means all @samp{.text} input sections in all input files.
2579 Since the location counter is @samp{0x10000} when the output section
2580 @samp{.text} is defined, the linker will set the address of the
2581 @samp{.text} section in the output file to be @samp{0x10000}.
2583 The remaining lines define the @samp{.data} and @samp{.bss} sections in
2584 the output file. The linker will place the @samp{.data} output section
2585 at address @samp{0x8000000}. After the linker places the @samp{.data}
2586 output section, the value of the location counter will be
2587 @samp{0x8000000} plus the size of the @samp{.data} output section. The
2588 effect is that the linker will place the @samp{.bss} output section
2589 immediately after the @samp{.data} output section in memory.
2591 The linker will ensure that each output section has the required
2592 alignment, by increasing the location counter if necessary. In this
2593 example, the specified addresses for the @samp{.text} and @samp{.data}
2594 sections will probably satisfy any alignment constraints, but the linker
2595 may have to create a small gap between the @samp{.data} and @samp{.bss}
2598 That's it! That's a simple and complete linker script.
2600 @node Simple Commands
2601 @section Simple Linker Script Commands
2602 @cindex linker script simple commands
2603 In this section we describe the simple linker script commands.
2606 * Entry Point:: Setting the entry point
2607 * File Commands:: Commands dealing with files
2608 @ifclear SingleFormat
2609 * Format Commands:: Commands dealing with object file formats
2612 * Miscellaneous Commands:: Other linker script commands
2616 @subsection Setting the Entry Point
2617 @kindex ENTRY(@var{symbol})
2618 @cindex start of execution
2619 @cindex first instruction
2621 The first instruction to execute in a program is called the @dfn{entry
2622 point}. You can use the @code{ENTRY} linker script command to set the
2623 entry point. The argument is a symbol name:
2628 There are several ways to set the entry point. The linker will set the
2629 entry point by trying each of the following methods in order, and
2630 stopping when one of them succeeds:
2633 the @samp{-e} @var{entry} command-line option;
2635 the @code{ENTRY(@var{symbol})} command in a linker script;
2637 the value of the symbol @code{start}, if defined;
2639 the address of the first byte of the @samp{.text} section, if present;
2641 The address @code{0}.
2645 @subsection Commands Dealing with Files
2646 @cindex linker script file commands
2647 Several linker script commands deal with files.
2650 @item INCLUDE @var{filename}
2651 @kindex INCLUDE @var{filename}
2652 @cindex including a linker script
2653 Include the linker script @var{filename} at this point. The file will
2654 be searched for in the current directory, and in any directory specified
2655 with the @option{-L} option. You can nest calls to @code{INCLUDE} up to
2658 @item INPUT(@var{file}, @var{file}, @dots{})
2659 @itemx INPUT(@var{file} @var{file} @dots{})
2660 @kindex INPUT(@var{files})
2661 @cindex input files in linker scripts
2662 @cindex input object files in linker scripts
2663 @cindex linker script input object files
2664 The @code{INPUT} command directs the linker to include the named files
2665 in the link, as though they were named on the command line.
2667 For example, if you always want to include @file{subr.o} any time you do
2668 a link, but you can't be bothered to put it on every link command line,
2669 then you can put @samp{INPUT (subr.o)} in your linker script.
2671 In fact, if you like, you can list all of your input files in the linker
2672 script, and then invoke the linker with nothing but a @samp{-T} option.
2674 In case a @dfn{sysroot prefix} is configured, and the filename starts
2675 with the @samp{/} character, and the script being processed was
2676 located inside the @dfn{sysroot prefix}, the filename will be looked
2677 for in the @dfn{sysroot prefix}. Otherwise, the linker will try to
2678 open the file in the current directory. If it is not found, the
2679 linker will search through the archive library search path. See the
2680 description of @samp{-L} in @ref{Options,,Command Line Options}.
2682 If you use @samp{INPUT (-l@var{file})}, @command{ld} will transform the
2683 name to @code{lib@var{file}.a}, as with the command line argument
2686 When you use the @code{INPUT} command in an implicit linker script, the
2687 files will be included in the link at the point at which the linker
2688 script file is included. This can affect archive searching.
2690 @item GROUP(@var{file}, @var{file}, @dots{})
2691 @itemx GROUP(@var{file} @var{file} @dots{})
2692 @kindex GROUP(@var{files})
2693 @cindex grouping input files
2694 The @code{GROUP} command is like @code{INPUT}, except that the named
2695 files should all be archives, and they are searched repeatedly until no
2696 new undefined references are created. See the description of @samp{-(}
2697 in @ref{Options,,Command Line Options}.
2699 @item AS_NEEDED(@var{file}, @var{file}, @dots{})
2700 @itemx AS_NEEDED(@var{file} @var{file} @dots{})
2701 @kindex AS_NEEDED(@var{files})
2702 This construct can appear only inside of the @code{INPUT} or @code{GROUP}
2703 commands, among other filenames. The files listed will be handled
2704 as if they appear directly in the @code{INPUT} or @code{GROUP} commands,
2705 with the exception of ELF shared libraries, that will be added only
2706 when they are actually needed. This construct essentially enables
2707 @option{--as-needed} option for all the files listed inside of it
2708 and restores previous @option{--as-needed} resp. @option{--no-as-needed}
2711 @item OUTPUT(@var{filename})
2712 @kindex OUTPUT(@var{filename})
2713 @cindex output file name in linker script
2714 The @code{OUTPUT} command names the output file. Using
2715 @code{OUTPUT(@var{filename})} in the linker script is exactly like using
2716 @samp{-o @var{filename}} on the command line (@pxref{Options,,Command
2717 Line Options}). If both are used, the command line option takes
2720 You can use the @code{OUTPUT} command to define a default name for the
2721 output file other than the usual default of @file{a.out}.
2723 @item SEARCH_DIR(@var{path})
2724 @kindex SEARCH_DIR(@var{path})
2725 @cindex library search path in linker script
2726 @cindex archive search path in linker script
2727 @cindex search path in linker script
2728 The @code{SEARCH_DIR} command adds @var{path} to the list of paths where
2729 @command{ld} looks for archive libraries. Using
2730 @code{SEARCH_DIR(@var{path})} is exactly like using @samp{-L @var{path}}
2731 on the command line (@pxref{Options,,Command Line Options}). If both
2732 are used, then the linker will search both paths. Paths specified using
2733 the command line option are searched first.
2735 @item STARTUP(@var{filename})
2736 @kindex STARTUP(@var{filename})
2737 @cindex first input file
2738 The @code{STARTUP} command is just like the @code{INPUT} command, except
2739 that @var{filename} will become the first input file to be linked, as
2740 though it were specified first on the command line. This may be useful
2741 when using a system in which the entry point is always the start of the
2745 @ifclear SingleFormat
2746 @node Format Commands
2747 @subsection Commands Dealing with Object File Formats
2748 A couple of linker script commands deal with object file formats.
2751 @item OUTPUT_FORMAT(@var{bfdname})
2752 @itemx OUTPUT_FORMAT(@var{default}, @var{big}, @var{little})
2753 @kindex OUTPUT_FORMAT(@var{bfdname})
2754 @cindex output file format in linker script
2755 The @code{OUTPUT_FORMAT} command names the BFD format to use for the
2756 output file (@pxref{BFD}). Using @code{OUTPUT_FORMAT(@var{bfdname})} is
2757 exactly like using @samp{--oformat @var{bfdname}} on the command line
2758 (@pxref{Options,,Command Line Options}). If both are used, the command
2759 line option takes precedence.
2761 You can use @code{OUTPUT_FORMAT} with three arguments to use different
2762 formats based on the @samp{-EB} and @samp{-EL} command line options.
2763 This permits the linker script to set the output format based on the
2766 If neither @samp{-EB} nor @samp{-EL} are used, then the output format
2767 will be the first argument, @var{default}. If @samp{-EB} is used, the
2768 output format will be the second argument, @var{big}. If @samp{-EL} is
2769 used, the output format will be the third argument, @var{little}.
2771 For example, the default linker script for the MIPS ELF target uses this
2774 OUTPUT_FORMAT(elf32-bigmips, elf32-bigmips, elf32-littlemips)
2776 This says that the default format for the output file is
2777 @samp{elf32-bigmips}, but if the user uses the @samp{-EL} command line
2778 option, the output file will be created in the @samp{elf32-littlemips}
2781 @item TARGET(@var{bfdname})
2782 @kindex TARGET(@var{bfdname})
2783 @cindex input file format in linker script
2784 The @code{TARGET} command names the BFD format to use when reading input
2785 files. It affects subsequent @code{INPUT} and @code{GROUP} commands.
2786 This command is like using @samp{-b @var{bfdname}} on the command line
2787 (@pxref{Options,,Command Line Options}). If the @code{TARGET} command
2788 is used but @code{OUTPUT_FORMAT} is not, then the last @code{TARGET}
2789 command is also used to set the format for the output file. @xref{BFD}.
2793 @node Miscellaneous Commands
2794 @subsection Other Linker Script Commands
2795 There are a few other linker scripts commands.
2798 @item ASSERT(@var{exp}, @var{message})
2800 @cindex assertion in linker script
2801 Ensure that @var{exp} is non-zero. If it is zero, then exit the linker
2802 with an error code, and print @var{message}.
2804 @item EXTERN(@var{symbol} @var{symbol} @dots{})
2806 @cindex undefined symbol in linker script
2807 Force @var{symbol} to be entered in the output file as an undefined
2808 symbol. Doing this may, for example, trigger linking of additional
2809 modules from standard libraries. You may list several @var{symbol}s for
2810 each @code{EXTERN}, and you may use @code{EXTERN} multiple times. This
2811 command has the same effect as the @samp{-u} command-line option.
2813 @item FORCE_COMMON_ALLOCATION
2814 @kindex FORCE_COMMON_ALLOCATION
2815 @cindex common allocation in linker script
2816 This command has the same effect as the @samp{-d} command-line option:
2817 to make @command{ld} assign space to common symbols even if a relocatable
2818 output file is specified (@samp{-r}).
2820 @item INHIBIT_COMMON_ALLOCATION
2821 @kindex INHIBIT_COMMON_ALLOCATION
2822 @cindex common allocation in linker script
2823 This command has the same effect as the @samp{--no-define-common}
2824 command-line option: to make @code{ld} omit the assignment of addresses
2825 to common symbols even for a non-relocatable output file.
2827 @item NOCROSSREFS(@var{section} @var{section} @dots{})
2828 @kindex NOCROSSREFS(@var{sections})
2829 @cindex cross references
2830 This command may be used to tell @command{ld} to issue an error about any
2831 references among certain output sections.
2833 In certain types of programs, particularly on embedded systems when
2834 using overlays, when one section is loaded into memory, another section
2835 will not be. Any direct references between the two sections would be
2836 errors. For example, it would be an error if code in one section called
2837 a function defined in the other section.
2839 The @code{NOCROSSREFS} command takes a list of output section names. If
2840 @command{ld} detects any cross references between the sections, it reports
2841 an error and returns a non-zero exit status. Note that the
2842 @code{NOCROSSREFS} command uses output section names, not input section
2845 @ifclear SingleFormat
2846 @item OUTPUT_ARCH(@var{bfdarch})
2847 @kindex OUTPUT_ARCH(@var{bfdarch})
2848 @cindex machine architecture
2849 @cindex architecture
2850 Specify a particular output machine architecture. The argument is one
2851 of the names used by the BFD library (@pxref{BFD}). You can see the
2852 architecture of an object file by using the @code{objdump} program with
2853 the @samp{-f} option.
2858 @section Assigning Values to Symbols
2859 @cindex assignment in scripts
2860 @cindex symbol definition, scripts
2861 @cindex variables, defining
2862 You may assign a value to a symbol in a linker script. This will define
2863 the symbol and place it into the symbol table with a global scope.
2866 * Simple Assignments:: Simple Assignments
2868 * PROVIDE_HIDDEN:: PROVIDE_HIDDEN
2869 * Source Code Reference:: How to use a linker script defined symbol in source code
2872 @node Simple Assignments
2873 @subsection Simple Assignments
2875 You may assign to a symbol using any of the C assignment operators:
2878 @item @var{symbol} = @var{expression} ;
2879 @itemx @var{symbol} += @var{expression} ;
2880 @itemx @var{symbol} -= @var{expression} ;
2881 @itemx @var{symbol} *= @var{expression} ;
2882 @itemx @var{symbol} /= @var{expression} ;
2883 @itemx @var{symbol} <<= @var{expression} ;
2884 @itemx @var{symbol} >>= @var{expression} ;
2885 @itemx @var{symbol} &= @var{expression} ;
2886 @itemx @var{symbol} |= @var{expression} ;
2889 The first case will define @var{symbol} to the value of
2890 @var{expression}. In the other cases, @var{symbol} must already be
2891 defined, and the value will be adjusted accordingly.
2893 The special symbol name @samp{.} indicates the location counter. You
2894 may only use this within a @code{SECTIONS} command. @xref{Location Counter}.
2896 The semicolon after @var{expression} is required.
2898 Expressions are defined below; see @ref{Expressions}.
2900 You may write symbol assignments as commands in their own right, or as
2901 statements within a @code{SECTIONS} command, or as part of an output
2902 section description in a @code{SECTIONS} command.
2904 The section of the symbol will be set from the section of the
2905 expression; for more information, see @ref{Expression Section}.
2907 Here is an example showing the three different places that symbol
2908 assignments may be used:
2919 _bdata = (. + 3) & ~ 3;
2920 .data : @{ *(.data) @}
2924 In this example, the symbol @samp{floating_point} will be defined as
2925 zero. The symbol @samp{_etext} will be defined as the address following
2926 the last @samp{.text} input section. The symbol @samp{_bdata} will be
2927 defined as the address following the @samp{.text} output section aligned
2928 upward to a 4 byte boundary.
2933 In some cases, it is desirable for a linker script to define a symbol
2934 only if it is referenced and is not defined by any object included in
2935 the link. For example, traditional linkers defined the symbol
2936 @samp{etext}. However, ANSI C requires that the user be able to use
2937 @samp{etext} as a function name without encountering an error. The
2938 @code{PROVIDE} keyword may be used to define a symbol, such as
2939 @samp{etext}, only if it is referenced but not defined. The syntax is
2940 @code{PROVIDE(@var{symbol} = @var{expression})}.
2942 Here is an example of using @code{PROVIDE} to define @samp{etext}:
2955 In this example, if the program defines @samp{_etext} (with a leading
2956 underscore), the linker will give a multiple definition error. If, on
2957 the other hand, the program defines @samp{etext} (with no leading
2958 underscore), the linker will silently use the definition in the program.
2959 If the program references @samp{etext} but does not define it, the
2960 linker will use the definition in the linker script.
2962 @node PROVIDE_HIDDEN
2963 @subsection PROVIDE_HIDDEN
2964 @cindex PROVIDE_HIDDEN
2965 Similar to @code{PROVIDE}. For ELF targeted ports, the symbol will be
2966 hidden and won't be exported.
2968 @node Source Code Reference
2969 @subsection Source Code Reference
2971 Accessing a linker script defined variable from source code is not
2972 intuitive. In particular a linker script symbol is not equivalent to
2973 a variable declaration in a high level language, it is instead a
2974 symbol that does not have a value.
2976 Before going further, it is important to note that compilers often
2977 transform names in the source code into different names when they are
2978 stored in the symbol table. For example, Fortran compilers commonly
2979 prepend or append an underscore, and C++ performs extensive @samp{name
2980 mangling}. Therefore there might be a discrepancy between the name
2981 of a variable as it is used in source code and the name of the same
2982 variable as it is defined in a linker script. For example in C a
2983 linker script variable might be referred to as:
2989 But in the linker script it might be defined as:
2995 In the remaining examples however it is assumed that no name
2996 transformation has taken place.
2998 When a symbol is declared in a high level language such as C, two
2999 things happen. The first is that the compiler reserves enough space
3000 in the program's memory to hold the @emph{value} of the symbol. The
3001 second is that the compiler creates an entry in the program's symbol
3002 table which holds the symbol's @emph{address}. ie the symbol table
3003 contains the address of the block of memory holding the symbol's
3004 value. So for example the following C declaration, at file scope:
3010 creates a entry called @samp{foo} in the symbol table. This entry
3011 holds the address of an @samp{int} sized block of memory where the
3012 number 1000 is initially stored.
3014 When a program references a symbol the compiler generates code that
3015 first accesses the symbol table to find the address of the symbol's
3016 memory block and then code to read the value from that memory block.
3023 looks up the symbol @samp{foo} in the symbol table, gets the address
3024 associated with this symbol and then writes the value 1 into that
3031 looks up the symbol @samp{foo} in the symbol table, gets it address
3032 and then copies this address into the block of memory associated with
3033 the variable @samp{a}.
3035 Linker scripts symbol declarations, by contrast, create an entry in
3036 the symbol table but do not assign any memory to them. Thus they are
3037 an address without a value. So for example the linker script definition:
3043 creates an entry in the symbol table called @samp{foo} which holds
3044 the address of memory location 1000, but nothing special is stored at
3045 address 1000. This means that you cannot access the @emph{value} of a
3046 linker script defined symbol - it has no value - all you can do is
3047 access the @emph{address} of a linker script defined symbol.
3049 Hence when you are using a linker script defined symbol in source code
3050 you should always take the address of the symbol, and never attempt to
3051 use its value. For example suppose you want to copy the contents of a
3052 section of memory called .ROM into a section called .FLASH and the
3053 linker script contains these declarations:
3057 start_of_ROM = .ROM;
3058 end_of_ROM = .ROM + sizeof (.ROM) - 1;
3059 start_of_FLASH = .FLASH;
3063 Then the C source code to perform the copy would be:
3067 extern char start_of_ROM, end_of_ROM, start_of_FLASH;
3069 memcpy (& start_of_FLASH, & start_of_ROM, & end_of_ROM - & start_of_ROM);
3073 Note the use of the @samp{&} operators. These are correct.
3076 @section SECTIONS Command
3078 The @code{SECTIONS} command tells the linker how to map input sections
3079 into output sections, and how to place the output sections in memory.
3081 The format of the @code{SECTIONS} command is:
3085 @var{sections-command}
3086 @var{sections-command}
3091 Each @var{sections-command} may of be one of the following:
3095 an @code{ENTRY} command (@pxref{Entry Point,,Entry command})
3097 a symbol assignment (@pxref{Assignments})
3099 an output section description
3101 an overlay description
3104 The @code{ENTRY} command and symbol assignments are permitted inside the
3105 @code{SECTIONS} command for convenience in using the location counter in
3106 those commands. This can also make the linker script easier to
3107 understand because you can use those commands at meaningful points in
3108 the layout of the output file.
3110 Output section descriptions and overlay descriptions are described
3113 If you do not use a @code{SECTIONS} command in your linker script, the
3114 linker will place each input section into an identically named output
3115 section in the order that the sections are first encountered in the
3116 input files. If all input sections are present in the first file, for
3117 example, the order of sections in the output file will match the order
3118 in the first input file. The first section will be at address zero.
3121 * Output Section Description:: Output section description
3122 * Output Section Name:: Output section name
3123 * Output Section Address:: Output section address
3124 * Input Section:: Input section description
3125 * Output Section Data:: Output section data
3126 * Output Section Keywords:: Output section keywords
3127 * Output Section Discarding:: Output section discarding
3128 * Output Section Attributes:: Output section attributes
3129 * Overlay Description:: Overlay description
3132 @node Output Section Description
3133 @subsection Output Section Description
3134 The full description of an output section looks like this:
3137 @var{section} [@var{address}] [(@var{type})] :
3138 [AT(@var{lma})] [ALIGN(@var{section_align})] [SUBALIGN(@var{subsection_align})]
3140 @var{output-section-command}
3141 @var{output-section-command}
3143 @} [>@var{region}] [AT>@var{lma_region}] [:@var{phdr} :@var{phdr} @dots{}] [=@var{fillexp}]
3147 Most output sections do not use most of the optional section attributes.
3149 The whitespace around @var{section} is required, so that the section
3150 name is unambiguous. The colon and the curly braces are also required.
3151 The line breaks and other white space are optional.
3153 Each @var{output-section-command} may be one of the following:
3157 a symbol assignment (@pxref{Assignments})
3159 an input section description (@pxref{Input Section})
3161 data values to include directly (@pxref{Output Section Data})
3163 a special output section keyword (@pxref{Output Section Keywords})
3166 @node Output Section Name
3167 @subsection Output Section Name
3168 @cindex name, section
3169 @cindex section name
3170 The name of the output section is @var{section}. @var{section} must
3171 meet the constraints of your output format. In formats which only
3172 support a limited number of sections, such as @code{a.out}, the name
3173 must be one of the names supported by the format (@code{a.out}, for
3174 example, allows only @samp{.text}, @samp{.data} or @samp{.bss}). If the
3175 output format supports any number of sections, but with numbers and not
3176 names (as is the case for Oasys), the name should be supplied as a
3177 quoted numeric string. A section name may consist of any sequence of
3178 characters, but a name which contains any unusual characters such as
3179 commas must be quoted.
3181 The output section name @samp{/DISCARD/} is special; @ref{Output Section
3184 @node Output Section Address
3185 @subsection Output Section Address
3186 @cindex address, section
3187 @cindex section address
3188 The @var{address} is an expression for the VMA (the virtual memory
3189 address) of the output section. If you do not provide @var{address},
3190 the linker will set it based on @var{region} if present, or otherwise
3191 based on the current value of the location counter.
3193 If you provide @var{address}, the address of the output section will be
3194 set to precisely that. If you provide neither @var{address} nor
3195 @var{region}, then the address of the output section will be set to the
3196 current value of the location counter aligned to the alignment
3197 requirements of the output section. The alignment requirement of the
3198 output section is the strictest alignment of any input section contained
3199 within the output section.
3203 .text . : @{ *(.text) @}
3208 .text : @{ *(.text) @}
3211 are subtly different. The first will set the address of the
3212 @samp{.text} output section to the current value of the location
3213 counter. The second will set it to the current value of the location
3214 counter aligned to the strictest alignment of a @samp{.text} input
3217 The @var{address} may be an arbitrary expression; @ref{Expressions}.
3218 For example, if you want to align the section on a 0x10 byte boundary,
3219 so that the lowest four bits of the section address are zero, you could
3220 do something like this:
3222 .text ALIGN(0x10) : @{ *(.text) @}
3225 This works because @code{ALIGN} returns the current location counter
3226 aligned upward to the specified value.
3228 Specifying @var{address} for a section will change the value of the
3232 @subsection Input Section Description
3233 @cindex input sections
3234 @cindex mapping input sections to output sections
3235 The most common output section command is an input section description.
3237 The input section description is the most basic linker script operation.
3238 You use output sections to tell the linker how to lay out your program
3239 in memory. You use input section descriptions to tell the linker how to
3240 map the input files into your memory layout.
3243 * Input Section Basics:: Input section basics
3244 * Input Section Wildcards:: Input section wildcard patterns
3245 * Input Section Common:: Input section for common symbols
3246 * Input Section Keep:: Input section and garbage collection
3247 * Input Section Example:: Input section example
3250 @node Input Section Basics
3251 @subsubsection Input Section Basics
3252 @cindex input section basics
3253 An input section description consists of a file name optionally followed
3254 by a list of section names in parentheses.
3256 The file name and the section name may be wildcard patterns, which we
3257 describe further below (@pxref{Input Section Wildcards}).
3259 The most common input section description is to include all input
3260 sections with a particular name in the output section. For example, to
3261 include all input @samp{.text} sections, you would write:
3266 Here the @samp{*} is a wildcard which matches any file name. To exclude a list
3267 of files from matching the file name wildcard, EXCLUDE_FILE may be used to
3268 match all files except the ones specified in the EXCLUDE_FILE list. For
3271 (*(EXCLUDE_FILE (*crtend.o *otherfile.o) .ctors))
3273 will cause all .ctors sections from all files except @file{crtend.o} and
3274 @file{otherfile.o} to be included.
3276 There are two ways to include more than one section:
3282 The difference between these is the order in which the @samp{.text} and
3283 @samp{.rdata} input sections will appear in the output section. In the
3284 first example, they will be intermingled, appearing in the same order as
3285 they are found in the linker input. In the second example, all
3286 @samp{.text} input sections will appear first, followed by all
3287 @samp{.rdata} input sections.
3289 You can specify a file name to include sections from a particular file.
3290 You would do this if one or more of your files contain special data that
3291 needs to be at a particular location in memory. For example:
3296 If you use a file name without a list of sections, then all sections in
3297 the input file will be included in the output section. This is not
3298 commonly done, but it may by useful on occasion. For example:
3303 When you use a file name which does not contain any wild card
3304 characters, the linker will first see if you also specified the file
3305 name on the linker command line or in an @code{INPUT} command. If you
3306 did not, the linker will attempt to open the file as an input file, as
3307 though it appeared on the command line. Note that this differs from an
3308 @code{INPUT} command, because the linker will not search for the file in
3309 the archive search path.
3311 @node Input Section Wildcards
3312 @subsubsection Input Section Wildcard Patterns
3313 @cindex input section wildcards
3314 @cindex wildcard file name patterns
3315 @cindex file name wildcard patterns
3316 @cindex section name wildcard patterns
3317 In an input section description, either the file name or the section
3318 name or both may be wildcard patterns.
3320 The file name of @samp{*} seen in many examples is a simple wildcard
3321 pattern for the file name.
3323 The wildcard patterns are like those used by the Unix shell.
3327 matches any number of characters
3329 matches any single character
3331 matches a single instance of any of the @var{chars}; the @samp{-}
3332 character may be used to specify a range of characters, as in
3333 @samp{[a-z]} to match any lower case letter
3335 quotes the following character
3338 When a file name is matched with a wildcard, the wildcard characters
3339 will not match a @samp{/} character (used to separate directory names on
3340 Unix). A pattern consisting of a single @samp{*} character is an
3341 exception; it will always match any file name, whether it contains a
3342 @samp{/} or not. In a section name, the wildcard characters will match
3343 a @samp{/} character.
3345 File name wildcard patterns only match files which are explicitly
3346 specified on the command line or in an @code{INPUT} command. The linker
3347 does not search directories to expand wildcards.
3349 If a file name matches more than one wildcard pattern, or if a file name
3350 appears explicitly and is also matched by a wildcard pattern, the linker
3351 will use the first match in the linker script. For example, this
3352 sequence of input section descriptions is probably in error, because the
3353 @file{data.o} rule will not be used:
3355 .data : @{ *(.data) @}
3356 .data1 : @{ data.o(.data) @}
3359 @cindex SORT_BY_NAME
3360 Normally, the linker will place files and sections matched by wildcards
3361 in the order in which they are seen during the link. You can change
3362 this by using the @code{SORT_BY_NAME} keyword, which appears before a wildcard
3363 pattern in parentheses (e.g., @code{SORT_BY_NAME(.text*)}). When the
3364 @code{SORT_BY_NAME} keyword is used, the linker will sort the files or sections
3365 into ascending order by name before placing them in the output file.
3367 @cindex SORT_BY_ALIGNMENT
3368 @code{SORT_BY_ALIGNMENT} is very similar to @code{SORT_BY_NAME}. The
3369 difference is @code{SORT_BY_ALIGNMENT} will sort sections into
3370 ascending order by alignment before placing them in the output file.
3373 @code{SORT} is an alias for @code{SORT_BY_NAME}.
3375 When there are nested section sorting commands in linker script, there
3376 can be at most 1 level of nesting for section sorting commands.
3380 @code{SORT_BY_NAME} (@code{SORT_BY_ALIGNMENT} (wildcard section pattern)).
3381 It will sort the input sections by name first, then by alignment if 2
3382 sections have the same name.
3384 @code{SORT_BY_ALIGNMENT} (@code{SORT_BY_NAME} (wildcard section pattern)).
3385 It will sort the input sections by alignment first, then by name if 2
3386 sections have the same alignment.
3388 @code{SORT_BY_NAME} (@code{SORT_BY_NAME} (wildcard section pattern)) is
3389 treated the same as @code{SORT_BY_NAME} (wildcard section pattern).
3391 @code{SORT_BY_ALIGNMENT} (@code{SORT_BY_ALIGNMENT} (wildcard section pattern))
3392 is treated the same as @code{SORT_BY_ALIGNMENT} (wildcard section pattern).
3394 All other nested section sorting commands are invalid.
3397 When both command line section sorting option and linker script
3398 section sorting command are used, section sorting command always
3399 takes precedence over the command line option.
3401 If the section sorting command in linker script isn't nested, the
3402 command line option will make the section sorting command to be
3403 treated as nested sorting command.
3407 @code{SORT_BY_NAME} (wildcard section pattern ) with
3408 @option{--sort-sections alignment} is equivalent to
3409 @code{SORT_BY_NAME} (@code{SORT_BY_ALIGNMENT} (wildcard section pattern)).
3411 @code{SORT_BY_ALIGNMENT} (wildcard section pattern) with
3412 @option{--sort-section name} is equivalent to
3413 @code{SORT_BY_ALIGNMENT} (@code{SORT_BY_NAME} (wildcard section pattern)).
3416 If the section sorting command in linker script is nested, the
3417 command line option will be ignored.
3419 If you ever get confused about where input sections are going, use the
3420 @samp{-M} linker option to generate a map file. The map file shows
3421 precisely how input sections are mapped to output sections.
3423 This example shows how wildcard patterns might be used to partition
3424 files. This linker script directs the linker to place all @samp{.text}
3425 sections in @samp{.text} and all @samp{.bss} sections in @samp{.bss}.
3426 The linker will place the @samp{.data} section from all files beginning
3427 with an upper case character in @samp{.DATA}; for all other files, the
3428 linker will place the @samp{.data} section in @samp{.data}.
3432 .text : @{ *(.text) @}
3433 .DATA : @{ [A-Z]*(.data) @}
3434 .data : @{ *(.data) @}
3435 .bss : @{ *(.bss) @}
3440 @node Input Section Common
3441 @subsubsection Input Section for Common Symbols
3442 @cindex common symbol placement
3443 @cindex uninitialized data placement
3444 A special notation is needed for common symbols, because in many object
3445 file formats common symbols do not have a particular input section. The
3446 linker treats common symbols as though they are in an input section
3447 named @samp{COMMON}.
3449 You may use file names with the @samp{COMMON} section just as with any
3450 other input sections. You can use this to place common symbols from a
3451 particular input file in one section while common symbols from other
3452 input files are placed in another section.
3454 In most cases, common symbols in input files will be placed in the
3455 @samp{.bss} section in the output file. For example:
3457 .bss @{ *(.bss) *(COMMON) @}
3460 @cindex scommon section
3461 @cindex small common symbols
3462 Some object file formats have more than one type of common symbol. For
3463 example, the MIPS ELF object file format distinguishes standard common
3464 symbols and small common symbols. In this case, the linker will use a
3465 different special section name for other types of common symbols. In
3466 the case of MIPS ELF, the linker uses @samp{COMMON} for standard common
3467 symbols and @samp{.scommon} for small common symbols. This permits you
3468 to map the different types of common symbols into memory at different
3472 You will sometimes see @samp{[COMMON]} in old linker scripts. This
3473 notation is now considered obsolete. It is equivalent to
3476 @node Input Section Keep
3477 @subsubsection Input Section and Garbage Collection
3479 @cindex garbage collection
3480 When link-time garbage collection is in use (@samp{--gc-sections}),
3481 it is often useful to mark sections that should not be eliminated.
3482 This is accomplished by surrounding an input section's wildcard entry
3483 with @code{KEEP()}, as in @code{KEEP(*(.init))} or
3484 @code{KEEP(SORT_BY_NAME(*)(.ctors))}.
3486 @node Input Section Example
3487 @subsubsection Input Section Example
3488 The following example is a complete linker script. It tells the linker
3489 to read all of the sections from file @file{all.o} and place them at the
3490 start of output section @samp{outputa} which starts at location
3491 @samp{0x10000}. All of section @samp{.input1} from file @file{foo.o}
3492 follows immediately, in the same output section. All of section
3493 @samp{.input2} from @file{foo.o} goes into output section
3494 @samp{outputb}, followed by section @samp{.input1} from @file{foo1.o}.
3495 All of the remaining @samp{.input1} and @samp{.input2} sections from any
3496 files are written to output section @samp{outputc}.
3524 @node Output Section Data
3525 @subsection Output Section Data
3527 @cindex section data
3528 @cindex output section data
3529 @kindex BYTE(@var{expression})
3530 @kindex SHORT(@var{expression})
3531 @kindex LONG(@var{expression})
3532 @kindex QUAD(@var{expression})
3533 @kindex SQUAD(@var{expression})
3534 You can include explicit bytes of data in an output section by using
3535 @code{BYTE}, @code{SHORT}, @code{LONG}, @code{QUAD}, or @code{SQUAD} as
3536 an output section command. Each keyword is followed by an expression in
3537 parentheses providing the value to store (@pxref{Expressions}). The
3538 value of the expression is stored at the current value of the location
3541 The @code{BYTE}, @code{SHORT}, @code{LONG}, and @code{QUAD} commands
3542 store one, two, four, and eight bytes (respectively). After storing the
3543 bytes, the location counter is incremented by the number of bytes
3546 For example, this will store the byte 1 followed by the four byte value
3547 of the symbol @samp{addr}:
3553 When using a 64 bit host or target, @code{QUAD} and @code{SQUAD} are the
3554 same; they both store an 8 byte, or 64 bit, value. When both host and
3555 target are 32 bits, an expression is computed as 32 bits. In this case
3556 @code{QUAD} stores a 32 bit value zero extended to 64 bits, and
3557 @code{SQUAD} stores a 32 bit value sign extended to 64 bits.
3559 If the object file format of the output file has an explicit endianness,
3560 which is the normal case, the value will be stored in that endianness.
3561 When the object file format does not have an explicit endianness, as is
3562 true of, for example, S-records, the value will be stored in the
3563 endianness of the first input object file.
3565 Note---these commands only work inside a section description and not
3566 between them, so the following will produce an error from the linker:
3568 SECTIONS @{@ .text : @{@ *(.text) @}@ LONG(1) .data : @{@ *(.data) @}@ @}@
3570 whereas this will work:
3572 SECTIONS @{@ .text : @{@ *(.text) ; LONG(1) @}@ .data : @{@ *(.data) @}@ @}@
3575 @kindex FILL(@var{expression})
3576 @cindex holes, filling
3577 @cindex unspecified memory
3578 You may use the @code{FILL} command to set the fill pattern for the
3579 current section. It is followed by an expression in parentheses. Any
3580 otherwise unspecified regions of memory within the section (for example,
3581 gaps left due to the required alignment of input sections) are filled
3582 with the value of the expression, repeated as
3583 necessary. A @code{FILL} statement covers memory locations after the
3584 point at which it occurs in the section definition; by including more
3585 than one @code{FILL} statement, you can have different fill patterns in
3586 different parts of an output section.
3588 This example shows how to fill unspecified regions of memory with the
3594 The @code{FILL} command is similar to the @samp{=@var{fillexp}} output
3595 section attribute, but it only affects the
3596 part of the section following the @code{FILL} command, rather than the
3597 entire section. If both are used, the @code{FILL} command takes
3598 precedence. @xref{Output Section Fill}, for details on the fill
3601 @node Output Section Keywords
3602 @subsection Output Section Keywords
3603 There are a couple of keywords which can appear as output section
3607 @kindex CREATE_OBJECT_SYMBOLS
3608 @cindex input filename symbols
3609 @cindex filename symbols
3610 @item CREATE_OBJECT_SYMBOLS
3611 The command tells the linker to create a symbol for each input file.
3612 The name of each symbol will be the name of the corresponding input
3613 file. The section of each symbol will be the output section in which
3614 the @code{CREATE_OBJECT_SYMBOLS} command appears.
3616 This is conventional for the a.out object file format. It is not
3617 normally used for any other object file format.
3619 @kindex CONSTRUCTORS
3620 @cindex C++ constructors, arranging in link
3621 @cindex constructors, arranging in link
3623 When linking using the a.out object file format, the linker uses an
3624 unusual set construct to support C++ global constructors and
3625 destructors. When linking object file formats which do not support
3626 arbitrary sections, such as ECOFF and XCOFF, the linker will
3627 automatically recognize C++ global constructors and destructors by name.
3628 For these object file formats, the @code{CONSTRUCTORS} command tells the
3629 linker to place constructor information in the output section where the
3630 @code{CONSTRUCTORS} command appears. The @code{CONSTRUCTORS} command is
3631 ignored for other object file formats.
3633 The symbol @w{@code{__CTOR_LIST__}} marks the start of the global
3634 constructors, and the symbol @w{@code{__CTOR_END__}} marks the end.
3635 Similarly, @w{@code{__DTOR_LIST__}} and @w{@code{__DTOR_END__}} mark
3636 the start and end of the global destructors. The
3637 first word in the list is the number of entries, followed by the address
3638 of each constructor or destructor, followed by a zero word. The
3639 compiler must arrange to actually run the code. For these object file
3640 formats @sc{gnu} C++ normally calls constructors from a subroutine
3641 @code{__main}; a call to @code{__main} is automatically inserted into
3642 the startup code for @code{main}. @sc{gnu} C++ normally runs
3643 destructors either by using @code{atexit}, or directly from the function
3646 For object file formats such as @code{COFF} or @code{ELF} which support
3647 arbitrary section names, @sc{gnu} C++ will normally arrange to put the
3648 addresses of global constructors and destructors into the @code{.ctors}
3649 and @code{.dtors} sections. Placing the following sequence into your
3650 linker script will build the sort of table which the @sc{gnu} C++
3651 runtime code expects to see.
3655 LONG((__CTOR_END__ - __CTOR_LIST__) / 4 - 2)
3660 LONG((__DTOR_END__ - __DTOR_LIST__) / 4 - 2)
3666 If you are using the @sc{gnu} C++ support for initialization priority,
3667 which provides some control over the order in which global constructors
3668 are run, you must sort the constructors at link time to ensure that they
3669 are executed in the correct order. When using the @code{CONSTRUCTORS}
3670 command, use @samp{SORT_BY_NAME(CONSTRUCTORS)} instead. When using the
3671 @code{.ctors} and @code{.dtors} sections, use @samp{*(SORT_BY_NAME(.ctors))} and
3672 @samp{*(SORT_BY_NAME(.dtors))} instead of just @samp{*(.ctors)} and
3675 Normally the compiler and linker will handle these issues automatically,
3676 and you will not need to concern yourself with them. However, you may
3677 need to consider this if you are using C++ and writing your own linker
3682 @node Output Section Discarding
3683 @subsection Output Section Discarding
3684 @cindex discarding sections
3685 @cindex sections, discarding
3686 @cindex removing sections
3687 The linker will not create output section which do not have any
3688 contents. This is for convenience when referring to input sections that
3689 may or may not be present in any of the input files. For example:
3694 will only create a @samp{.foo} section in the output file if there is a
3695 @samp{.foo} section in at least one input file.
3697 If you use anything other than an input section description as an output
3698 section command, such as a symbol assignment, then the output section
3699 will always be created, even if there are no matching input sections.
3702 The special output section name @samp{/DISCARD/} may be used to discard
3703 input sections. Any input sections which are assigned to an output
3704 section named @samp{/DISCARD/} are not included in the output file.
3706 @node Output Section Attributes
3707 @subsection Output Section Attributes
3708 @cindex output section attributes
3709 We showed above that the full description of an output section looked
3713 @var{section} [@var{address}] [(@var{type})] :
3714 [AT(@var{lma})] [ALIGN(@var{section_align})] [SUBALIGN(@var{subsection_align})]
3716 @var{output-section-command}
3717 @var{output-section-command}
3719 @} [>@var{region}] [AT>@var{lma_region}] [:@var{phdr} :@var{phdr} @dots{}] [=@var{fillexp}]
3722 We've already described @var{section}, @var{address}, and
3723 @var{output-section-command}. In this section we will describe the
3724 remaining section attributes.
3727 * Output Section Type:: Output section type
3728 * Output Section LMA:: Output section LMA
3729 * Forced Output Alignment:: Forced Output Alignment
3730 * Forced Input Alignment:: Forced Input Alignment
3731 * Output Section Region:: Output section region
3732 * Output Section Phdr:: Output section phdr
3733 * Output Section Fill:: Output section fill
3736 @node Output Section Type
3737 @subsubsection Output Section Type
3738 Each output section may have a type. The type is a keyword in
3739 parentheses. The following types are defined:
3743 The section should be marked as not loadable, so that it will not be
3744 loaded into memory when the program is run.
3749 These type names are supported for backward compatibility, and are
3750 rarely used. They all have the same effect: the section should be
3751 marked as not allocatable, so that no memory is allocated for the
3752 section when the program is run.
3756 @cindex prevent unnecessary loading
3757 @cindex loading, preventing
3758 The linker normally sets the attributes of an output section based on
3759 the input sections which map into it. You can override this by using
3760 the section type. For example, in the script sample below, the
3761 @samp{ROM} section is addressed at memory location @samp{0} and does not
3762 need to be loaded when the program is run. The contents of the
3763 @samp{ROM} section will appear in the linker output file as usual.
3767 ROM 0 (NOLOAD) : @{ @dots{} @}
3773 @node Output Section LMA
3774 @subsubsection Output Section LMA
3775 @kindex AT>@var{lma_region}
3776 @kindex AT(@var{lma})
3777 @cindex load address
3778 @cindex section load address
3779 Every section has a virtual address (VMA) and a load address (LMA); see
3780 @ref{Basic Script Concepts}. The address expression which may appear in
3781 an output section description sets the VMA (@pxref{Output Section
3784 The expression @var{lma} that follows the @code{AT} keyword specifies
3785 the load address of the section.
3787 Alternatively, with @samp{AT>@var{lma_region}} expression, you may
3788 specify a memory region for the section's load address. @xref{MEMORY}.
3789 Note that if the section has not had a VMA assigned to it then the
3790 linker will use the @var{lma_region} as the VMA region as well.
3792 If neither @code{AT} nor @code{AT>} is specified for an allocatable
3793 section, the linker will set the LMA such that the difference between
3794 VMA and LMA for the section is the same as the preceding output
3795 section in the same region. If there is no preceding output section
3796 or the section is not allocatable, the linker will set the LMA equal
3798 @xref{Output Section Region}.
3800 @cindex ROM initialized data
3801 @cindex initialized data in ROM
3802 This feature is designed to make it easy to build a ROM image. For
3803 example, the following linker script creates three output sections: one
3804 called @samp{.text}, which starts at @code{0x1000}, one called
3805 @samp{.mdata}, which is loaded at the end of the @samp{.text} section
3806 even though its VMA is @code{0x2000}, and one called @samp{.bss} to hold
3807 uninitialized data at address @code{0x3000}. The symbol @code{_data} is
3808 defined with the value @code{0x2000}, which shows that the location
3809 counter holds the VMA value, not the LMA value.
3815 .text 0x1000 : @{ *(.text) _etext = . ; @}
3817 AT ( ADDR (.text) + SIZEOF (.text) )
3818 @{ _data = . ; *(.data); _edata = . ; @}
3820 @{ _bstart = . ; *(.bss) *(COMMON) ; _bend = . ;@}
3825 The run-time initialization code for use with a program generated with
3826 this linker script would include something like the following, to copy
3827 the initialized data from the ROM image to its runtime address. Notice
3828 how this code takes advantage of the symbols defined by the linker
3833 extern char _etext, _data, _edata, _bstart, _bend;
3834 char *src = &_etext;
3837 /* ROM has data at end of text; copy it. */
3838 while (dst < &_edata) @{
3843 for (dst = &_bstart; dst< &_bend; dst++)
3848 @node Forced Output Alignment
3849 @subsubsection Forced Output Alignment
3850 @kindex ALIGN(@var{section_align})
3851 @cindex forcing output section alignment
3852 @cindex output section alignment
3853 You can increase an output section's alignment by using ALIGN.
3855 @node Forced Input Alignment
3856 @subsubsection Forced Input Alignment
3857 @kindex SUBALIGN(@var{subsection_align})
3858 @cindex forcing input section alignment
3859 @cindex input section alignment
3860 You can force input section alignment within an output section by using
3861 SUBALIGN. The value specified overrides any alignment given by input
3862 sections, whether larger or smaller.
3864 @node Output Section Region
3865 @subsubsection Output Section Region
3866 @kindex >@var{region}
3867 @cindex section, assigning to memory region
3868 @cindex memory regions and sections
3869 You can assign a section to a previously defined region of memory by
3870 using @samp{>@var{region}}. @xref{MEMORY}.
3872 Here is a simple example:
3875 MEMORY @{ rom : ORIGIN = 0x1000, LENGTH = 0x1000 @}
3876 SECTIONS @{ ROM : @{ *(.text) @} >rom @}
3880 @node Output Section Phdr
3881 @subsubsection Output Section Phdr
3883 @cindex section, assigning to program header
3884 @cindex program headers and sections
3885 You can assign a section to a previously defined program segment by
3886 using @samp{:@var{phdr}}. @xref{PHDRS}. If a section is assigned to
3887 one or more segments, then all subsequent allocated sections will be
3888 assigned to those segments as well, unless they use an explicitly
3889 @code{:@var{phdr}} modifier. You can use @code{:NONE} to tell the
3890 linker to not put the section in any segment at all.
3892 Here is a simple example:
3895 PHDRS @{ text PT_LOAD ; @}
3896 SECTIONS @{ .text : @{ *(.text) @} :text @}
3900 @node Output Section Fill
3901 @subsubsection Output Section Fill
3902 @kindex =@var{fillexp}
3903 @cindex section fill pattern
3904 @cindex fill pattern, entire section
3905 You can set the fill pattern for an entire section by using
3906 @samp{=@var{fillexp}}. @var{fillexp} is an expression
3907 (@pxref{Expressions}). Any otherwise unspecified regions of memory
3908 within the output section (for example, gaps left due to the required
3909 alignment of input sections) will be filled with the value, repeated as
3910 necessary. If the fill expression is a simple hex number, ie. a string
3911 of hex digit starting with @samp{0x} and without a trailing @samp{k} or @samp{M}, then
3912 an arbitrarily long sequence of hex digits can be used to specify the
3913 fill pattern; Leading zeros become part of the pattern too. For all
3914 other cases, including extra parentheses or a unary @code{+}, the fill
3915 pattern is the four least significant bytes of the value of the
3916 expression. In all cases, the number is big-endian.
3918 You can also change the fill value with a @code{FILL} command in the
3919 output section commands; (@pxref{Output Section Data}).
3921 Here is a simple example:
3924 SECTIONS @{ .text : @{ *(.text) @} =0x90909090 @}
3928 @node Overlay Description
3929 @subsection Overlay Description
3932 An overlay description provides an easy way to describe sections which
3933 are to be loaded as part of a single memory image but are to be run at
3934 the same memory address. At run time, some sort of overlay manager will
3935 copy the overlaid sections in and out of the runtime memory address as
3936 required, perhaps by simply manipulating addressing bits. This approach
3937 can be useful, for example, when a certain region of memory is faster
3940 Overlays are described using the @code{OVERLAY} command. The
3941 @code{OVERLAY} command is used within a @code{SECTIONS} command, like an
3942 output section description. The full syntax of the @code{OVERLAY}
3943 command is as follows:
3946 OVERLAY [@var{start}] : [NOCROSSREFS] [AT ( @var{ldaddr} )]
3950 @var{output-section-command}
3951 @var{output-section-command}
3953 @} [:@var{phdr}@dots{}] [=@var{fill}]
3956 @var{output-section-command}
3957 @var{output-section-command}
3959 @} [:@var{phdr}@dots{}] [=@var{fill}]
3961 @} [>@var{region}] [:@var{phdr}@dots{}] [=@var{fill}]
3965 Everything is optional except @code{OVERLAY} (a keyword), and each
3966 section must have a name (@var{secname1} and @var{secname2} above). The
3967 section definitions within the @code{OVERLAY} construct are identical to
3968 those within the general @code{SECTIONS} contruct (@pxref{SECTIONS}),
3969 except that no addresses and no memory regions may be defined for
3970 sections within an @code{OVERLAY}.
3972 The sections are all defined with the same starting address. The load
3973 addresses of the sections are arranged such that they are consecutive in
3974 memory starting at the load address used for the @code{OVERLAY} as a
3975 whole (as with normal section definitions, the load address is optional,
3976 and defaults to the start address; the start address is also optional,
3977 and defaults to the current value of the location counter).
3979 If the @code{NOCROSSREFS} keyword is used, and there any references
3980 among the sections, the linker will report an error. Since the sections
3981 all run at the same address, it normally does not make sense for one
3982 section to refer directly to another. @xref{Miscellaneous Commands,
3985 For each section within the @code{OVERLAY}, the linker automatically
3986 defines two symbols. The symbol @code{__load_start_@var{secname}} is
3987 defined as the starting load address of the section. The symbol
3988 @code{__load_stop_@var{secname}} is defined as the final load address of
3989 the section. Any characters within @var{secname} which are not legal
3990 within C identifiers are removed. C (or assembler) code may use these
3991 symbols to move the overlaid sections around as necessary.
3993 At the end of the overlay, the value of the location counter is set to
3994 the start address of the overlay plus the size of the largest section.
3996 Here is an example. Remember that this would appear inside a
3997 @code{SECTIONS} construct.
4000 OVERLAY 0x1000 : AT (0x4000)
4002 .text0 @{ o1/*.o(.text) @}
4003 .text1 @{ o2/*.o(.text) @}
4008 This will define both @samp{.text0} and @samp{.text1} to start at
4009 address 0x1000. @samp{.text0} will be loaded at address 0x4000, and
4010 @samp{.text1} will be loaded immediately after @samp{.text0}. The
4011 following symbols will be defined: @code{__load_start_text0},
4012 @code{__load_stop_text0}, @code{__load_start_text1},
4013 @code{__load_stop_text1}.
4015 C code to copy overlay @code{.text1} into the overlay area might look
4020 extern char __load_start_text1, __load_stop_text1;
4021 memcpy ((char *) 0x1000, &__load_start_text1,
4022 &__load_stop_text1 - &__load_start_text1);
4026 Note that the @code{OVERLAY} command is just syntactic sugar, since
4027 everything it does can be done using the more basic commands. The above
4028 example could have been written identically as follows.
4032 .text0 0x1000 : AT (0x4000) @{ o1/*.o(.text) @}
4033 __load_start_text0 = LOADADDR (.text0);
4034 __load_stop_text0 = LOADADDR (.text0) + SIZEOF (.text0);
4035 .text1 0x1000 : AT (0x4000 + SIZEOF (.text0)) @{ o2/*.o(.text) @}
4036 __load_start_text1 = LOADADDR (.text1);
4037 __load_stop_text1 = LOADADDR (.text1) + SIZEOF (.text1);
4038 . = 0x1000 + MAX (SIZEOF (.text0), SIZEOF (.text1));
4043 @section MEMORY Command
4045 @cindex memory regions
4046 @cindex regions of memory
4047 @cindex allocating memory
4048 @cindex discontinuous memory
4049 The linker's default configuration permits allocation of all available
4050 memory. You can override this by using the @code{MEMORY} command.
4052 The @code{MEMORY} command describes the location and size of blocks of
4053 memory in the target. You can use it to describe which memory regions
4054 may be used by the linker, and which memory regions it must avoid. You
4055 can then assign sections to particular memory regions. The linker will
4056 set section addresses based on the memory regions, and will warn about
4057 regions that become too full. The linker will not shuffle sections
4058 around to fit into the available regions.
4060 A linker script may contain at most one use of the @code{MEMORY}
4061 command. However, you can define as many blocks of memory within it as
4062 you wish. The syntax is:
4067 @var{name} [(@var{attr})] : ORIGIN = @var{origin}, LENGTH = @var{len}
4073 The @var{name} is a name used in the linker script to refer to the
4074 region. The region name has no meaning outside of the linker script.
4075 Region names are stored in a separate name space, and will not conflict
4076 with symbol names, file names, or section names. Each memory region
4077 must have a distinct name.
4079 @cindex memory region attributes
4080 The @var{attr} string is an optional list of attributes that specify
4081 whether to use a particular memory region for an input section which is
4082 not explicitly mapped in the linker script. As described in
4083 @ref{SECTIONS}, if you do not specify an output section for some input
4084 section, the linker will create an output section with the same name as
4085 the input section. If you define region attributes, the linker will use
4086 them to select the memory region for the output section that it creates.
4088 The @var{attr} string must consist only of the following characters:
4103 Invert the sense of any of the preceding attributes
4106 If a unmapped section matches any of the listed attributes other than
4107 @samp{!}, it will be placed in the memory region. The @samp{!}
4108 attribute reverses this test, so that an unmapped section will be placed
4109 in the memory region only if it does not match any of the listed
4115 The @var{origin} is an numerical expression for the start address of
4116 the memory region. The expression must evaluate to a constant and it
4117 cannot involve any symbols. The keyword @code{ORIGIN} may be
4118 abbreviated to @code{org} or @code{o} (but not, for example,
4124 The @var{len} is an expression for the size in bytes of the memory
4125 region. As with the @var{origin} expression, the expression must
4126 be numerical only and must evaluate to a constant. The keyword
4127 @code{LENGTH} may be abbreviated to @code{len} or @code{l}.
4129 In the following example, we specify that there are two memory regions
4130 available for allocation: one starting at @samp{0} for 256 kilobytes,
4131 and the other starting at @samp{0x40000000} for four megabytes. The
4132 linker will place into the @samp{rom} memory region every section which
4133 is not explicitly mapped into a memory region, and is either read-only
4134 or executable. The linker will place other sections which are not
4135 explicitly mapped into a memory region into the @samp{ram} memory
4142 rom (rx) : ORIGIN = 0, LENGTH = 256K
4143 ram (!rx) : org = 0x40000000, l = 4M
4148 Once you define a memory region, you can direct the linker to place
4149 specific output sections into that memory region by using the
4150 @samp{>@var{region}} output section attribute. For example, if you have
4151 a memory region named @samp{mem}, you would use @samp{>mem} in the
4152 output section definition. @xref{Output Section Region}. If no address
4153 was specified for the output section, the linker will set the address to
4154 the next available address within the memory region. If the combined
4155 output sections directed to a memory region are too large for the
4156 region, the linker will issue an error message.
4158 It is possible to access the origin and length of a memory in an
4159 expression via the @code{ORIGIN(@var{memory})} and
4160 @code{LENGTH(@var{memory})} functions:
4164 _fstack = ORIGIN(ram) + LENGTH(ram) - 4;
4169 @section PHDRS Command
4171 @cindex program headers
4172 @cindex ELF program headers
4173 @cindex program segments
4174 @cindex segments, ELF
4175 The ELF object file format uses @dfn{program headers}, also knows as
4176 @dfn{segments}. The program headers describe how the program should be
4177 loaded into memory. You can print them out by using the @code{objdump}
4178 program with the @samp{-p} option.
4180 When you run an ELF program on a native ELF system, the system loader
4181 reads the program headers in order to figure out how to load the
4182 program. This will only work if the program headers are set correctly.
4183 This manual does not describe the details of how the system loader
4184 interprets program headers; for more information, see the ELF ABI.
4186 The linker will create reasonable program headers by default. However,
4187 in some cases, you may need to specify the program headers more
4188 precisely. You may use the @code{PHDRS} command for this purpose. When
4189 the linker sees the @code{PHDRS} command in the linker script, it will
4190 not create any program headers other than the ones specified.
4192 The linker only pays attention to the @code{PHDRS} command when
4193 generating an ELF output file. In other cases, the linker will simply
4194 ignore @code{PHDRS}.
4196 This is the syntax of the @code{PHDRS} command. The words @code{PHDRS},
4197 @code{FILEHDR}, @code{AT}, and @code{FLAGS} are keywords.
4203 @var{name} @var{type} [ FILEHDR ] [ PHDRS ] [ AT ( @var{address} ) ]
4204 [ FLAGS ( @var{flags} ) ] ;
4209 The @var{name} is used only for reference in the @code{SECTIONS} command
4210 of the linker script. It is not put into the output file. Program
4211 header names are stored in a separate name space, and will not conflict
4212 with symbol names, file names, or section names. Each program header
4213 must have a distinct name.
4215 Certain program header types describe segments of memory which the
4216 system loader will load from the file. In the linker script, you
4217 specify the contents of these segments by placing allocatable output
4218 sections in the segments. You use the @samp{:@var{phdr}} output section
4219 attribute to place a section in a particular segment. @xref{Output
4222 It is normal to put certain sections in more than one segment. This
4223 merely implies that one segment of memory contains another. You may
4224 repeat @samp{:@var{phdr}}, using it once for each segment which should
4225 contain the section.
4227 If you place a section in one or more segments using @samp{:@var{phdr}},
4228 then the linker will place all subsequent allocatable sections which do
4229 not specify @samp{:@var{phdr}} in the same segments. This is for
4230 convenience, since generally a whole set of contiguous sections will be
4231 placed in a single segment. You can use @code{:NONE} to override the
4232 default segment and tell the linker to not put the section in any
4237 You may use the @code{FILEHDR} and @code{PHDRS} keywords appear after
4238 the program header type to further describe the contents of the segment.
4239 The @code{FILEHDR} keyword means that the segment should include the ELF
4240 file header. The @code{PHDRS} keyword means that the segment should
4241 include the ELF program headers themselves.
4243 The @var{type} may be one of the following. The numbers indicate the
4244 value of the keyword.
4247 @item @code{PT_NULL} (0)
4248 Indicates an unused program header.
4250 @item @code{PT_LOAD} (1)
4251 Indicates that this program header describes a segment to be loaded from
4254 @item @code{PT_DYNAMIC} (2)
4255 Indicates a segment where dynamic linking information can be found.
4257 @item @code{PT_INTERP} (3)
4258 Indicates a segment where the name of the program interpreter may be
4261 @item @code{PT_NOTE} (4)
4262 Indicates a segment holding note information.
4264 @item @code{PT_SHLIB} (5)
4265 A reserved program header type, defined but not specified by the ELF
4268 @item @code{PT_PHDR} (6)
4269 Indicates a segment where the program headers may be found.
4271 @item @var{expression}
4272 An expression giving the numeric type of the program header. This may
4273 be used for types not defined above.
4276 You can specify that a segment should be loaded at a particular address
4277 in memory by using an @code{AT} expression. This is identical to the
4278 @code{AT} command used as an output section attribute (@pxref{Output
4279 Section LMA}). The @code{AT} command for a program header overrides the
4280 output section attribute.
4282 The linker will normally set the segment flags based on the sections
4283 which comprise the segment. You may use the @code{FLAGS} keyword to
4284 explicitly specify the segment flags. The value of @var{flags} must be
4285 an integer. It is used to set the @code{p_flags} field of the program
4288 Here is an example of @code{PHDRS}. This shows a typical set of program
4289 headers used on a native ELF system.
4295 headers PT_PHDR PHDRS ;
4297 text PT_LOAD FILEHDR PHDRS ;
4299 dynamic PT_DYNAMIC ;
4305 .interp : @{ *(.interp) @} :text :interp
4306 .text : @{ *(.text) @} :text
4307 .rodata : @{ *(.rodata) @} /* defaults to :text */
4309 . = . + 0x1000; /* move to a new page in memory */
4310 .data : @{ *(.data) @} :data
4311 .dynamic : @{ *(.dynamic) @} :data :dynamic
4318 @section VERSION Command
4319 @kindex VERSION @{script text@}
4320 @cindex symbol versions
4321 @cindex version script
4322 @cindex versions of symbols
4323 The linker supports symbol versions when using ELF. Symbol versions are
4324 only useful when using shared libraries. The dynamic linker can use
4325 symbol versions to select a specific version of a function when it runs
4326 a program that may have been linked against an earlier version of the
4329 You can include a version script directly in the main linker script, or
4330 you can supply the version script as an implicit linker script. You can
4331 also use the @samp{--version-script} linker option.
4333 The syntax of the @code{VERSION} command is simply
4335 VERSION @{ version-script-commands @}
4338 The format of the version script commands is identical to that used by
4339 Sun's linker in Solaris 2.5. The version script defines a tree of
4340 version nodes. You specify the node names and interdependencies in the
4341 version script. You can specify which symbols are bound to which
4342 version nodes, and you can reduce a specified set of symbols to local
4343 scope so that they are not globally visible outside of the shared
4346 The easiest way to demonstrate the version script language is with a few
4367 "int f(int, double)";
4372 This example version script defines three version nodes. The first
4373 version node defined is @samp{VERS_1.1}; it has no other dependencies.
4374 The script binds the symbol @samp{foo1} to @samp{VERS_1.1}. It reduces
4375 a number of symbols to local scope so that they are not visible outside
4376 of the shared library; this is done using wildcard patterns, so that any
4377 symbol whose name begins with @samp{old}, @samp{original}, or @samp{new}
4378 is matched. The wildcard patterns available are the same as those used
4379 in the shell when matching filenames (also known as ``globbing'').
4380 However, if you specify the symbol name inside double quotes, then the
4381 name is treated as literal, rather than as a glob pattern.
4383 Next, the version script defines node @samp{VERS_1.2}. This node
4384 depends upon @samp{VERS_1.1}. The script binds the symbol @samp{foo2}
4385 to the version node @samp{VERS_1.2}.
4387 Finally, the version script defines node @samp{VERS_2.0}. This node
4388 depends upon @samp{VERS_1.2}. The scripts binds the symbols @samp{bar1}
4389 and @samp{bar2} are bound to the version node @samp{VERS_2.0}.
4391 When the linker finds a symbol defined in a library which is not
4392 specifically bound to a version node, it will effectively bind it to an
4393 unspecified base version of the library. You can bind all otherwise
4394 unspecified symbols to a given version node by using @samp{global: *;}
4395 somewhere in the version script.
4397 The names of the version nodes have no specific meaning other than what
4398 they might suggest to the person reading them. The @samp{2.0} version
4399 could just as well have appeared in between @samp{1.1} and @samp{1.2}.
4400 However, this would be a confusing way to write a version script.
4402 Node name can be omited, provided it is the only version node
4403 in the version script. Such version script doesn't assign any versions to
4404 symbols, only selects which symbols will be globally visible out and which
4408 @{ global: foo; bar; local: *; @};
4411 When you link an application against a shared library that has versioned
4412 symbols, the application itself knows which version of each symbol it
4413 requires, and it also knows which version nodes it needs from each
4414 shared library it is linked against. Thus at runtime, the dynamic
4415 loader can make a quick check to make sure that the libraries you have
4416 linked against do in fact supply all of the version nodes that the
4417 application will need to resolve all of the dynamic symbols. In this
4418 way it is possible for the dynamic linker to know with certainty that
4419 all external symbols that it needs will be resolvable without having to
4420 search for each symbol reference.
4422 The symbol versioning is in effect a much more sophisticated way of
4423 doing minor version checking that SunOS does. The fundamental problem
4424 that is being addressed here is that typically references to external
4425 functions are bound on an as-needed basis, and are not all bound when
4426 the application starts up. If a shared library is out of date, a
4427 required interface may be missing; when the application tries to use
4428 that interface, it may suddenly and unexpectedly fail. With symbol
4429 versioning, the user will get a warning when they start their program if
4430 the libraries being used with the application are too old.
4432 There are several GNU extensions to Sun's versioning approach. The
4433 first of these is the ability to bind a symbol to a version node in the
4434 source file where the symbol is defined instead of in the versioning
4435 script. This was done mainly to reduce the burden on the library
4436 maintainer. You can do this by putting something like:
4438 __asm__(".symver original_foo,foo@@VERS_1.1");
4441 in the C source file. This renames the function @samp{original_foo} to
4442 be an alias for @samp{foo} bound to the version node @samp{VERS_1.1}.
4443 The @samp{local:} directive can be used to prevent the symbol
4444 @samp{original_foo} from being exported. A @samp{.symver} directive
4445 takes precedence over a version script.
4447 The second GNU extension is to allow multiple versions of the same
4448 function to appear in a given shared library. In this way you can make
4449 an incompatible change to an interface without increasing the major
4450 version number of the shared library, while still allowing applications
4451 linked against the old interface to continue to function.
4453 To do this, you must use multiple @samp{.symver} directives in the
4454 source file. Here is an example:
4457 __asm__(".symver original_foo,foo@@");
4458 __asm__(".symver old_foo,foo@@VERS_1.1");
4459 __asm__(".symver old_foo1,foo@@VERS_1.2");
4460 __asm__(".symver new_foo,foo@@@@VERS_2.0");
4463 In this example, @samp{foo@@} represents the symbol @samp{foo} bound to the
4464 unspecified base version of the symbol. The source file that contains this
4465 example would define 4 C functions: @samp{original_foo}, @samp{old_foo},
4466 @samp{old_foo1}, and @samp{new_foo}.
4468 When you have multiple definitions of a given symbol, there needs to be
4469 some way to specify a default version to which external references to
4470 this symbol will be bound. You can do this with the
4471 @samp{foo@@@@VERS_2.0} type of @samp{.symver} directive. You can only
4472 declare one version of a symbol as the default in this manner; otherwise
4473 you would effectively have multiple definitions of the same symbol.
4475 If you wish to bind a reference to a specific version of the symbol
4476 within the shared library, you can use the aliases of convenience
4477 (i.e., @samp{old_foo}), or you can use the @samp{.symver} directive to
4478 specifically bind to an external version of the function in question.
4480 You can also specify the language in the version script:
4483 VERSION extern "lang" @{ version-script-commands @}
4486 The supported @samp{lang}s are @samp{C}, @samp{C++}, and @samp{Java}.
4487 The linker will iterate over the list of symbols at the link time and
4488 demangle them according to @samp{lang} before matching them to the
4489 patterns specified in @samp{version-script-commands}.
4491 Demangled names may contains spaces and other special characters. As
4492 described above, you can use a glob pattern to match demangled names,
4493 or you can use a double-quoted string to match the string exactly. In
4494 the latter case, be aware that minor differences (such as differing
4495 whitespace) between the version script and the demangler output will
4496 cause a mismatch. As the exact string generated by the demangler
4497 might change in the future, even if the mangled name does not, you
4498 should check that all of your version directives are behaving as you
4499 expect when you upgrade.
4502 @section Expressions in Linker Scripts
4505 The syntax for expressions in the linker script language is identical to
4506 that of C expressions. All expressions are evaluated as integers. All
4507 expressions are evaluated in the same size, which is 32 bits if both the
4508 host and target are 32 bits, and is otherwise 64 bits.
4510 You can use and set symbol values in expressions.
4512 The linker defines several special purpose builtin functions for use in
4516 * Constants:: Constants
4517 * Symbols:: Symbol Names
4518 * Orphan Sections:: Orphan Sections
4519 * Location Counter:: The Location Counter
4520 * Operators:: Operators
4521 * Evaluation:: Evaluation
4522 * Expression Section:: The Section of an Expression
4523 * Builtin Functions:: Builtin Functions
4527 @subsection Constants
4528 @cindex integer notation
4529 @cindex constants in linker scripts
4530 All constants are integers.
4532 As in C, the linker considers an integer beginning with @samp{0} to be
4533 octal, and an integer beginning with @samp{0x} or @samp{0X} to be
4534 hexadecimal. The linker considers other integers to be decimal.
4536 @cindex scaled integers
4537 @cindex K and M integer suffixes
4538 @cindex M and K integer suffixes
4539 @cindex suffixes for integers
4540 @cindex integer suffixes
4541 In addition, you can use the suffixes @code{K} and @code{M} to scale a
4545 @c END TEXI2ROFF-KILL
4546 @code{1024} or @code{1024*1024}
4550 ${\rm 1024}$ or ${\rm 1024}^2$
4552 @c END TEXI2ROFF-KILL
4553 respectively. For example, the following all refer to the same quantity:
4561 @subsection Symbol Names
4562 @cindex symbol names
4564 @cindex quoted symbol names
4566 Unless quoted, symbol names start with a letter, underscore, or period
4567 and may include letters, digits, underscores, periods, and hyphens.
4568 Unquoted symbol names must not conflict with any keywords. You can
4569 specify a symbol which contains odd characters or has the same name as a
4570 keyword by surrounding the symbol name in double quotes:
4573 "with a space" = "also with a space" + 10;
4576 Since symbols can contain many non-alphabetic characters, it is safest
4577 to delimit symbols with spaces. For example, @samp{A-B} is one symbol,
4578 whereas @samp{A - B} is an expression involving subtraction.
4580 @node Orphan Sections
4581 @subsection Orphan Sections
4583 Orphan sections are sections present in the input files which
4584 are not explicitly placed into the output file by the linker
4585 script. The linker will still copy these sections into the
4586 output file, but it has to guess as to where they should be
4587 placed. The linker uses a simple heuristic to do this. It
4588 attempts to place orphan sections after non-orphan sections of the
4589 same attribute, such as code vs data, loadable vs non-loadable, etc.
4590 If there is not enough room to do this then it places
4591 at the end of the file.
4593 For ELF targets, the attribute of the section includes section type as
4594 well as section flag.
4596 @node Location Counter
4597 @subsection The Location Counter
4600 @cindex location counter
4601 @cindex current output location
4602 The special linker variable @dfn{dot} @samp{.} always contains the
4603 current output location counter. Since the @code{.} always refers to a
4604 location in an output section, it may only appear in an expression
4605 within a @code{SECTIONS} command. The @code{.} symbol may appear
4606 anywhere that an ordinary symbol is allowed in an expression.
4609 Assigning a value to @code{.} will cause the location counter to be
4610 moved. This may be used to create holes in the output section. The
4611 location counter may not be moved backwards inside an output section,
4612 and may not be moved backwards outside of an output section if so
4613 doing creates areas with overlapping LMAs.
4629 In the previous example, the @samp{.text} section from @file{file1} is
4630 located at the beginning of the output section @samp{output}. It is
4631 followed by a 1000 byte gap. Then the @samp{.text} section from
4632 @file{file2} appears, also with a 1000 byte gap following before the
4633 @samp{.text} section from @file{file3}. The notation @samp{= 0x12345678}
4634 specifies what data to write in the gaps (@pxref{Output Section Fill}).
4636 @cindex dot inside sections
4637 Note: @code{.} actually refers to the byte offset from the start of the
4638 current containing object. Normally this is the @code{SECTIONS}
4639 statement, whose start address is 0, hence @code{.} can be used as an
4640 absolute address. If @code{.} is used inside a section description
4641 however, it refers to the byte offset from the start of that section,
4642 not an absolute address. Thus in a script like this:
4660 The @samp{.text} section will be assigned a starting address of 0x100
4661 and a size of exactly 0x200 bytes, even if there is not enough data in
4662 the @samp{.text} input sections to fill this area. (If there is too
4663 much data, an error will be produced because this would be an attempt to
4664 move @code{.} backwards). The @samp{.data} section will start at 0x500
4665 and it will have an extra 0x600 bytes worth of space after the end of
4666 the values from the @samp{.data} input sections and before the end of
4667 the @samp{.data} output section itself.
4669 @cindex dot outside sections
4670 Setting symbols to the value of the location counter outside of an
4671 output section statement can result in unexpected values if the linker
4672 needs to place orphan sections. For example, given the following:
4678 .text: @{ *(.text) @}
4682 .data: @{ *(.data) @}
4687 If the linker needs to place some input section, e.g. @code{.rodata},
4688 not mentioned in the script, it might choose to place that section
4689 between @code{.text} and @code{.data}. You might think the linker
4690 should place @code{.rodata} on the blank line in the above script, but
4691 blank lines are of no particular significance to the linker. As well,
4692 the linker doesn't associate the above symbol names with their
4693 sections. Instead, it assumes that all assignments or other
4694 statements belong to the previous output section, except for the
4695 special case of an assignment to @code{.}. I.e., the linker will
4696 place the orphan @code{.rodata} section as if the script was written
4703 .text: @{ *(.text) @}
4707 .rodata: @{ *(.rodata) @}
4708 .data: @{ *(.data) @}
4713 This may or may not be the script author's intention for the value of
4714 @code{start_of_data}. One way to influence the orphan section
4715 placement is to assign the location counter to itself, as the linker
4716 assumes that an assignment to @code{.} is setting the start address of
4717 a following output section and thus should be grouped with that
4718 section. So you could write:
4724 .text: @{ *(.text) @}
4729 .data: @{ *(.data) @}
4734 Now, the orphan @code{.rodata} section will be placed between
4735 @code{end_of_text} and @code{start_of_data}.
4739 @subsection Operators
4740 @cindex operators for arithmetic
4741 @cindex arithmetic operators
4742 @cindex precedence in expressions
4743 The linker recognizes the standard C set of arithmetic operators, with
4744 the standard bindings and precedence levels:
4747 @c END TEXI2ROFF-KILL
4749 precedence associativity Operators Notes
4755 5 left == != > < <= >=
4761 11 right &= += -= *= /= (2)
4765 (1) Prefix operators
4766 (2) @xref{Assignments}.
4770 \vskip \baselineskip
4771 %"lispnarrowing" is the extra indent used generally for smallexample
4772 \hskip\lispnarrowing\vbox{\offinterlineskip
4775 {\vrule#&\strut\hfil\ #\ \hfil&\vrule#&\strut\hfil\ #\ \hfil&\vrule#&\strut\hfil\ {\tt #}\ \hfil&\vrule#\cr
4776 height2pt&\omit&&\omit&&\omit&\cr
4777 &Precedence&& Associativity &&{\rm Operators}&\cr
4778 height2pt&\omit&&\omit&&\omit&\cr
4780 height2pt&\omit&&\omit&&\omit&\cr
4782 % '176 is tilde, '~' in tt font
4783 &1&&left&&\qquad- \char'176\ !\qquad\dag&\cr
4784 &2&&left&&* / \%&\cr
4787 &5&&left&&== != > < <= >=&\cr
4790 &8&&left&&{\&\&}&\cr
4793 &11&&right&&\qquad\&= += -= *= /=\qquad\ddag&\cr
4795 height2pt&\omit&&\omit&&\omit&\cr}
4800 @obeylines@parskip=0pt@parindent=0pt
4801 @dag@quad Prefix operators.
4802 @ddag@quad @xref{Assignments}.
4805 @c END TEXI2ROFF-KILL
4808 @subsection Evaluation
4809 @cindex lazy evaluation
4810 @cindex expression evaluation order
4811 The linker evaluates expressions lazily. It only computes the value of
4812 an expression when absolutely necessary.
4814 The linker needs some information, such as the value of the start
4815 address of the first section, and the origins and lengths of memory
4816 regions, in order to do any linking at all. These values are computed
4817 as soon as possible when the linker reads in the linker script.
4819 However, other values (such as symbol values) are not known or needed
4820 until after storage allocation. Such values are evaluated later, when
4821 other information (such as the sizes of output sections) is available
4822 for use in the symbol assignment expression.
4824 The sizes of sections cannot be known until after allocation, so
4825 assignments dependent upon these are not performed until after
4828 Some expressions, such as those depending upon the location counter
4829 @samp{.}, must be evaluated during section allocation.
4831 If the result of an expression is required, but the value is not
4832 available, then an error results. For example, a script like the
4838 .text 9+this_isnt_constant :
4844 will cause the error message @samp{non constant expression for initial
4847 @node Expression Section
4848 @subsection The Section of an Expression
4849 @cindex expression sections
4850 @cindex absolute expressions
4851 @cindex relative expressions
4852 @cindex absolute and relocatable symbols
4853 @cindex relocatable and absolute symbols
4854 @cindex symbols, relocatable and absolute
4855 When the linker evaluates an expression, the result is either absolute
4856 or relative to some section. A relative expression is expressed as a
4857 fixed offset from the base of a section.
4859 The position of the expression within the linker script determines
4860 whether it is absolute or relative. An expression which appears within
4861 an output section definition is relative to the base of the output
4862 section. An expression which appears elsewhere will be absolute.
4864 A symbol set to a relative expression will be relocatable if you request
4865 relocatable output using the @samp{-r} option. That means that a
4866 further link operation may change the value of the symbol. The symbol's
4867 section will be the section of the relative expression.
4869 A symbol set to an absolute expression will retain the same value
4870 through any further link operation. The symbol will be absolute, and
4871 will not have any particular associated section.
4873 You can use the builtin function @code{ABSOLUTE} to force an expression
4874 to be absolute when it would otherwise be relative. For example, to
4875 create an absolute symbol set to the address of the end of the output
4876 section @samp{.data}:
4880 .data : @{ *(.data) _edata = ABSOLUTE(.); @}
4884 If @samp{ABSOLUTE} were not used, @samp{_edata} would be relative to the
4885 @samp{.data} section.
4887 @node Builtin Functions
4888 @subsection Builtin Functions
4889 @cindex functions in expressions
4890 The linker script language includes a number of builtin functions for
4891 use in linker script expressions.
4894 @item ABSOLUTE(@var{exp})
4895 @kindex ABSOLUTE(@var{exp})
4896 @cindex expression, absolute
4897 Return the absolute (non-relocatable, as opposed to non-negative) value
4898 of the expression @var{exp}. Primarily useful to assign an absolute
4899 value to a symbol within a section definition, where symbol values are
4900 normally section relative. @xref{Expression Section}.
4902 @item ADDR(@var{section})
4903 @kindex ADDR(@var{section})
4904 @cindex section address in expression
4905 Return the absolute address (the VMA) of the named @var{section}. Your
4906 script must previously have defined the location of that section. In
4907 the following example, @code{symbol_1} and @code{symbol_2} are assigned
4914 start_of_output_1 = ABSOLUTE(.);
4919 symbol_1 = ADDR(.output1);
4920 symbol_2 = start_of_output_1;
4926 @item ALIGN(@var{align})
4927 @itemx ALIGN(@var{exp},@var{align})
4928 @kindex ALIGN(@var{align})
4929 @kindex ALIGN(@var{exp},@var{align})
4930 @cindex round up location counter
4931 @cindex align location counter
4932 @cindex round up expression
4933 @cindex align expression
4934 Return the location counter (@code{.}) or arbitrary expression aligned
4935 to the next @var{align} boundary. The single operand @code{ALIGN}
4936 doesn't change the value of the location counter---it just does
4937 arithmetic on it. The two operand @code{ALIGN} allows an arbitrary
4938 expression to be aligned upwards (@code{ALIGN(@var{align})} is
4939 equivalent to @code{ALIGN(., @var{align})}).
4941 Here is an example which aligns the output @code{.data} section to the
4942 next @code{0x2000} byte boundary after the preceding section and sets a
4943 variable within the section to the next @code{0x8000} boundary after the
4948 .data ALIGN(0x2000): @{
4950 variable = ALIGN(0x8000);
4956 The first use of @code{ALIGN} in this example specifies the location of
4957 a section because it is used as the optional @var{address} attribute of
4958 a section definition (@pxref{Output Section Address}). The second use
4959 of @code{ALIGN} is used to defines the value of a symbol.
4961 The builtin function @code{NEXT} is closely related to @code{ALIGN}.
4963 @item BLOCK(@var{exp})
4964 @kindex BLOCK(@var{exp})
4965 This is a synonym for @code{ALIGN}, for compatibility with older linker
4966 scripts. It is most often seen when setting the address of an output
4969 @item DATA_SEGMENT_ALIGN(@var{maxpagesize}, @var{commonpagesize})
4970 @kindex DATA_SEGMENT_ALIGN(@var{maxpagesize}, @var{commonpagesize})
4971 This is equivalent to either
4973 (ALIGN(@var{maxpagesize}) + (. & (@var{maxpagesize} - 1)))
4977 (ALIGN(@var{maxpagesize}) + (. & (@var{maxpagesize} - @var{commonpagesize})))
4980 depending on whether the latter uses fewer @var{commonpagesize} sized pages
4981 for the data segment (area between the result of this expression and
4982 @code{DATA_SEGMENT_END}) than the former or not.
4983 If the latter form is used, it means @var{commonpagesize} bytes of runtime
4984 memory will be saved at the expense of up to @var{commonpagesize} wasted
4985 bytes in the on-disk file.
4987 This expression can only be used directly in @code{SECTIONS} commands, not in
4988 any output section descriptions and only once in the linker script.
4989 @var{commonpagesize} should be less or equal to @var{maxpagesize} and should
4990 be the system page size the object wants to be optimized for (while still
4991 working on system page sizes up to @var{maxpagesize}).
4996 . = DATA_SEGMENT_ALIGN(0x10000, 0x2000);
4999 @item DATA_SEGMENT_END(@var{exp})
5000 @kindex DATA_SEGMENT_END(@var{exp})
5001 This defines the end of data segment for @code{DATA_SEGMENT_ALIGN}
5002 evaluation purposes.
5005 . = DATA_SEGMENT_END(.);
5008 @item DATA_SEGMENT_RELRO_END(@var{offset}, @var{exp})
5009 @kindex DATA_SEGMENT_RELRO_END(@var{offset}, @var{exp})
5010 This defines the end of the @code{PT_GNU_RELRO} segment when
5011 @samp{-z relro} option is used. Second argument is returned.
5012 When @samp{-z relro} option is not present, @code{DATA_SEGMENT_RELRO_END}
5013 does nothing, otherwise @code{DATA_SEGMENT_ALIGN} is padded so that
5014 @var{exp} + @var{offset} is aligned to the most commonly used page
5015 boundary for particular target. If present in the linker script,
5016 it must always come in between @code{DATA_SEGMENT_ALIGN} and
5017 @code{DATA_SEGMENT_END}.
5020 . = DATA_SEGMENT_RELRO_END(24, .);
5023 @item DEFINED(@var{symbol})
5024 @kindex DEFINED(@var{symbol})
5025 @cindex symbol defaults
5026 Return 1 if @var{symbol} is in the linker global symbol table and is
5027 defined before the statement using DEFINED in the script, otherwise
5028 return 0. You can use this function to provide
5029 default values for symbols. For example, the following script fragment
5030 shows how to set a global symbol @samp{begin} to the first location in
5031 the @samp{.text} section---but if a symbol called @samp{begin} already
5032 existed, its value is preserved:
5038 begin = DEFINED(begin) ? begin : . ;
5046 @item LENGTH(@var{memory})
5047 @kindex LENGTH(@var{memory})
5048 Return the length of the memory region named @var{memory}.
5050 @item LOADADDR(@var{section})
5051 @kindex LOADADDR(@var{section})
5052 @cindex section load address in expression
5053 Return the absolute LMA of the named @var{section}. This is normally
5054 the same as @code{ADDR}, but it may be different if the @code{AT}
5055 attribute is used in the output section definition (@pxref{Output
5059 @item MAX(@var{exp1}, @var{exp2})
5060 Returns the maximum of @var{exp1} and @var{exp2}.
5063 @item MIN(@var{exp1}, @var{exp2})
5064 Returns the minimum of @var{exp1} and @var{exp2}.
5066 @item NEXT(@var{exp})
5067 @kindex NEXT(@var{exp})
5068 @cindex unallocated address, next
5069 Return the next unallocated address that is a multiple of @var{exp}.
5070 This function is closely related to @code{ALIGN(@var{exp})}; unless you
5071 use the @code{MEMORY} command to define discontinuous memory for the
5072 output file, the two functions are equivalent.
5074 @item ORIGIN(@var{memory})
5075 @kindex ORIGIN(@var{memory})
5076 Return the origin of the memory region named @var{memory}.
5078 @item SEGMENT_START(@var{segment}, @var{default})
5079 @kindex SEGMENT_START(@var{segment}, @var{default})
5080 Return the base address of the named @var{segment}. If an explicit
5081 value has been given for this segment (with a command-line @samp{-T}
5082 option) that value will be returned; otherwise the value will be
5083 @var{default}. At present, the @samp{-T} command-line option can only
5084 be used to set the base address for the ``text'', ``data'', and
5085 ``bss'' sections, but you use @code{SEGMENT_START} with any segment
5088 @item SIZEOF(@var{section})
5089 @kindex SIZEOF(@var{section})
5090 @cindex section size
5091 Return the size in bytes of the named @var{section}, if that section has
5092 been allocated. If the section has not been allocated when this is
5093 evaluated, the linker will report an error. In the following example,
5094 @code{symbol_1} and @code{symbol_2} are assigned identical values:
5103 symbol_1 = .end - .start ;
5104 symbol_2 = SIZEOF(.output);
5109 @item SIZEOF_HEADERS
5110 @itemx sizeof_headers
5111 @kindex SIZEOF_HEADERS
5113 Return the size in bytes of the output file's headers. This is
5114 information which appears at the start of the output file. You can use
5115 this number when setting the start address of the first section, if you
5116 choose, to facilitate paging.
5118 @cindex not enough room for program headers
5119 @cindex program headers, not enough room
5120 When producing an ELF output file, if the linker script uses the
5121 @code{SIZEOF_HEADERS} builtin function, the linker must compute the
5122 number of program headers before it has determined all the section
5123 addresses and sizes. If the linker later discovers that it needs
5124 additional program headers, it will report an error @samp{not enough
5125 room for program headers}. To avoid this error, you must avoid using
5126 the @code{SIZEOF_HEADERS} function, or you must rework your linker
5127 script to avoid forcing the linker to use additional program headers, or
5128 you must define the program headers yourself using the @code{PHDRS}
5129 command (@pxref{PHDRS}).
5132 @node Implicit Linker Scripts
5133 @section Implicit Linker Scripts
5134 @cindex implicit linker scripts
5135 If you specify a linker input file which the linker can not recognize as
5136 an object file or an archive file, it will try to read the file as a
5137 linker script. If the file can not be parsed as a linker script, the
5138 linker will report an error.
5140 An implicit linker script will not replace the default linker script.
5142 Typically an implicit linker script would contain only symbol
5143 assignments, or the @code{INPUT}, @code{GROUP}, or @code{VERSION}
5146 Any input files read because of an implicit linker script will be read
5147 at the position in the command line where the implicit linker script was
5148 read. This can affect archive searching.
5151 @node Machine Dependent
5152 @chapter Machine Dependent Features
5154 @cindex machine dependencies
5155 @command{ld} has additional features on some platforms; the following
5156 sections describe them. Machines where @command{ld} has no additional
5157 functionality are not listed.
5161 * H8/300:: @command{ld} and the H8/300
5164 * i960:: @command{ld} and the Intel 960 family
5167 * ARM:: @command{ld} and the ARM family
5170 * HPPA ELF32:: @command{ld} and HPPA 32-bit ELF
5173 * MMIX:: @command{ld} and MMIX
5176 * MSP430:: @command{ld} and MSP430
5179 * M68HC11/68HC12:: @code{ld} and the Motorola 68HC11 and 68HC12 families
5182 * PowerPC ELF32:: @command{ld} and PowerPC 32-bit ELF Support
5185 * PowerPC64 ELF64:: @command{ld} and PowerPC64 64-bit ELF Support
5188 * TI COFF:: @command{ld} and TI COFF
5191 * WIN32:: @command{ld} and WIN32 (cygwin/mingw)
5194 * Xtensa:: @command{ld} and Xtensa Processors
5205 @section @command{ld} and the H8/300
5207 @cindex H8/300 support
5208 For the H8/300, @command{ld} can perform these global optimizations when
5209 you specify the @samp{--relax} command-line option.
5212 @cindex relaxing on H8/300
5213 @item relaxing address modes
5214 @command{ld} finds all @code{jsr} and @code{jmp} instructions whose
5215 targets are within eight bits, and turns them into eight-bit
5216 program-counter relative @code{bsr} and @code{bra} instructions,
5219 @cindex synthesizing on H8/300
5220 @item synthesizing instructions
5221 @c FIXME: specifically mov.b, or any mov instructions really?
5222 @command{ld} finds all @code{mov.b} instructions which use the
5223 sixteen-bit absolute address form, but refer to the top
5224 page of memory, and changes them to use the eight-bit address form.
5225 (That is: the linker turns @samp{mov.b @code{@@}@var{aa}:16} into
5226 @samp{mov.b @code{@@}@var{aa}:8} whenever the address @var{aa} is in the
5227 top page of memory).
5229 @item bit manipulation instructions
5230 @command{ld} finds all bit manipulation instructions like @code{band, bclr,
5231 biand, bild, bior, bist, bixor, bld, bnot, bor, bset, bst, btst, bxor}
5232 which use 32 bit and 16 bit absolute address form, but refer to the top
5233 page of memory, and changes them to use the 8 bit address form.
5234 (That is: the linker turns @samp{bset #xx:3,@code{@@}@var{aa}:32} into
5235 @samp{bset #xx:3,@code{@@}@var{aa}:8} whenever the address @var{aa} is in
5236 the top page of memory).
5238 @item system control instructions
5239 @command{ld} finds all @code{ldc.w, stc.w} instructions which use the
5240 32 bit absolute address form, but refer to the top page of memory, and
5241 changes them to use 16 bit address form.
5242 (That is: the linker turns @samp{ldc.w @code{@@}@var{aa}:32,ccr} into
5243 @samp{ldc.w @code{@@}@var{aa}:16,ccr} whenever the address @var{aa} is in
5244 the top page of memory).
5254 @c This stuff is pointless to say unless you're especially concerned
5255 @c with Renesas chips; don't enable it for generic case, please.
5257 @chapter @command{ld} and Other Renesas Chips
5259 @command{ld} also supports the Renesas (formerly Hitachi) H8/300H,
5260 H8/500, and SH chips. No special features, commands, or command-line
5261 options are required for these chips.
5271 @section @command{ld} and the Intel 960 Family
5273 @cindex i960 support
5275 You can use the @samp{-A@var{architecture}} command line option to
5276 specify one of the two-letter names identifying members of the 960
5277 family; the option specifies the desired output target, and warns of any
5278 incompatible instructions in the input files. It also modifies the
5279 linker's search strategy for archive libraries, to support the use of
5280 libraries specific to each particular architecture, by including in the
5281 search loop names suffixed with the string identifying the architecture.
5283 For example, if your @command{ld} command line included @w{@samp{-ACA}} as
5284 well as @w{@samp{-ltry}}, the linker would look (in its built-in search
5285 paths, and in any paths you specify with @samp{-L}) for a library with
5298 The first two possibilities would be considered in any event; the last
5299 two are due to the use of @w{@samp{-ACA}}.
5301 You can meaningfully use @samp{-A} more than once on a command line, since
5302 the 960 architecture family allows combination of target architectures; each
5303 use will add another pair of name variants to search for when @w{@samp{-l}}
5304 specifies a library.
5306 @cindex @option{--relax} on i960
5307 @cindex relaxing on i960
5308 @command{ld} supports the @samp{--relax} option for the i960 family. If
5309 you specify @samp{--relax}, @command{ld} finds all @code{balx} and
5310 @code{calx} instructions whose targets are within 24 bits, and turns
5311 them into 24-bit program-counter relative @code{bal} and @code{cal}
5312 instructions, respectively. @command{ld} also turns @code{cal}
5313 instructions into @code{bal} instructions when it determines that the
5314 target subroutine is a leaf routine (that is, the target subroutine does
5315 not itself call any subroutines).
5332 @node M68HC11/68HC12
5333 @section @command{ld} and the Motorola 68HC11 and 68HC12 families
5335 @cindex M68HC11 and 68HC12 support
5337 @subsection Linker Relaxation
5339 For the Motorola 68HC11, @command{ld} can perform these global
5340 optimizations when you specify the @samp{--relax} command-line option.
5343 @cindex relaxing on M68HC11
5344 @item relaxing address modes
5345 @command{ld} finds all @code{jsr} and @code{jmp} instructions whose
5346 targets are within eight bits, and turns them into eight-bit
5347 program-counter relative @code{bsr} and @code{bra} instructions,
5350 @command{ld} also looks at all 16-bit extended addressing modes and
5351 transforms them in a direct addressing mode when the address is in
5352 page 0 (between 0 and 0x0ff).
5354 @item relaxing gcc instruction group
5355 When @command{gcc} is called with @option{-mrelax}, it can emit group
5356 of instructions that the linker can optimize to use a 68HC11 direct
5357 addressing mode. These instructions consists of @code{bclr} or
5358 @code{bset} instructions.
5362 @subsection Trampoline Generation
5364 @cindex trampoline generation on M68HC11
5365 @cindex trampoline generation on M68HC12
5366 For 68HC11 and 68HC12, @command{ld} can generate trampoline code to
5367 call a far function using a normal @code{jsr} instruction. The linker
5368 will also change the relocation to some far function to use the
5369 trampoline address instead of the function address. This is typically the
5370 case when a pointer to a function is taken. The pointer will in fact
5371 point to the function trampoline.
5379 @section @command{ld} and the ARM family
5381 @cindex ARM interworking support
5382 @kindex --support-old-code
5383 For the ARM, @command{ld} will generate code stubs to allow functions calls
5384 between ARM and Thumb code. These stubs only work with code that has
5385 been compiled and assembled with the @samp{-mthumb-interwork} command
5386 line option. If it is necessary to link with old ARM object files or
5387 libraries, which have not been compiled with the -mthumb-interwork
5388 option then the @samp{--support-old-code} command line switch should be
5389 given to the linker. This will make it generate larger stub functions
5390 which will work with non-interworking aware ARM code. Note, however,
5391 the linker does not support generating stubs for function calls to
5392 non-interworking aware Thumb code.
5394 @cindex thumb entry point
5395 @cindex entry point, thumb
5396 @kindex --thumb-entry=@var{entry}
5397 The @samp{--thumb-entry} switch is a duplicate of the generic
5398 @samp{--entry} switch, in that it sets the program's starting address.
5399 But it also sets the bottom bit of the address, so that it can be
5400 branched to using a BX instruction, and the program will start
5401 executing in Thumb mode straight away.
5405 The @samp{--be8} switch instructs @command{ld} to generate BE8 format
5406 executables. This option is only valid when linking big-endian objects.
5407 The resulting image will contain big-endian data and little-endian code.
5410 @kindex --target1-rel
5411 @kindex --target1-abs
5412 The @samp{R_ARM_TARGET1} relocation is typically used for entries in the
5413 @samp{.init_array} section. It is interpreted as either @samp{R_ARM_REL32}
5414 or @samp{R_ARM_ABS32}, depending on the target. The @samp{--target1-rel}
5415 and @samp{--target1-abs} switches override the default.
5418 @kindex --target2=@var{type}
5419 The @samp{--target2=type} switch overrides the default definition of the
5420 @samp{R_ARM_TARGET2} relocation. Valid values for @samp{type}, their
5421 meanings, and target defaults are as follows:
5424 @samp{R_ARM_REL32} (arm*-*-elf, arm*-*-eabi)
5426 @samp{R_ARM_ABS32} (arm*-*-symbianelf)
5428 @samp{R_ARM_GOT_PREL} (arm*-*-linux, arm*-*-*bsd)
5433 The @samp{R_ARM_V4BX} relocation (defined by the ARM AAELF
5434 specification) enables objects compiled for the ARMv4 architecture to be
5435 interworking-safe when linked with other objects compiled for ARMv4t, but
5436 also allows pure ARMv4 binaries to be built from the same ARMv4 objects.
5438 In the latter case, the switch @option{--fix-v4bx} must be passed to the
5439 linker, which causes v4t @code{BX rM} instructions to be rewritten as
5440 @code{MOV PC,rM}, since v4 processors do not have a @code{BX} instruction.
5442 In the former case, the switch should not be used, and @samp{R_ARM_V4BX}
5443 relocations are ignored.
5447 The @samp{--use-blx} switch enables the linker to use ARM/Thumb
5448 BLX instructions (available on ARMv5t and above) in various
5449 situations. Currently it is used to perform calls via the PLT from Thumb
5450 code using BLX rather than using BX and a mode-switching stub before
5451 each PLT entry. This should lead to such calls executing slightly faster.
5453 This option is enabled implicitly for SymbianOS, so there is no need to
5454 specify it if you are using that target.
5467 @section @command{ld} and HPPA 32-bit ELF Support
5468 @cindex HPPA multiple sub-space stubs
5469 @kindex --multi-subspace
5470 When generating a shared library, @command{ld} will by default generate
5471 import stubs suitable for use with a single sub-space application.
5472 The @samp{--multi-subspace} switch causes @command{ld} to generate export
5473 stubs, and different (larger) import stubs suitable for use with
5474 multiple sub-spaces.
5476 @cindex HPPA stub grouping
5477 @kindex --stub-group-size=@var{N}
5478 Long branch stubs and import/export stubs are placed by @command{ld} in
5479 stub sections located between groups of input sections.
5480 @samp{--stub-group-size} specifies the maximum size of a group of input
5481 sections handled by one stub section. Since branch offsets are signed,
5482 a stub section may serve two groups of input sections, one group before
5483 the stub section, and one group after it. However, when using
5484 conditional branches that require stubs, it may be better (for branch
5485 prediction) that stub sections only serve one group of input sections.
5486 A negative value for @samp{N} chooses this scheme, ensuring that
5487 branches to stubs always use a negative offset. Two special values of
5488 @samp{N} are recognized, @samp{1} and @samp{-1}. These both instruct
5489 @command{ld} to automatically size input section groups for the branch types
5490 detected, with the same behaviour regarding stub placement as other
5491 positive or negative values of @samp{N} respectively.
5493 Note that @samp{--stub-group-size} does not split input sections. A
5494 single input section larger than the group size specified will of course
5495 create a larger group (of one section). If input sections are too
5496 large, it may not be possible for a branch to reach its stub.
5509 @section @code{ld} and MMIX
5510 For MMIX, there is a choice of generating @code{ELF} object files or
5511 @code{mmo} object files when linking. The simulator @code{mmix}
5512 understands the @code{mmo} format. The binutils @code{objcopy} utility
5513 can translate between the two formats.
5515 There is one special section, the @samp{.MMIX.reg_contents} section.
5516 Contents in this section is assumed to correspond to that of global
5517 registers, and symbols referring to it are translated to special symbols,
5518 equal to registers. In a final link, the start address of the
5519 @samp{.MMIX.reg_contents} section corresponds to the first allocated
5520 global register multiplied by 8. Register @code{$255} is not included in
5521 this section; it is always set to the program entry, which is at the
5522 symbol @code{Main} for @code{mmo} files.
5524 Symbols with the prefix @code{__.MMIX.start.}, for example
5525 @code{__.MMIX.start..text} and @code{__.MMIX.start..data} are special;
5526 there must be only one each, even if they are local. The default linker
5527 script uses these to set the default start address of a section.
5529 Initial and trailing multiples of zero-valued 32-bit words in a section,
5530 are left out from an mmo file.
5543 @section @code{ld} and MSP430
5544 For the MSP430 it is possible to select the MPU architecture. The flag @samp{-m [mpu type]}
5545 will select an appropriate linker script for selected MPU type. (To get a list of known MPUs
5546 just pass @samp{-m help} option to the linker).
5548 @cindex MSP430 extra sections
5549 The linker will recognize some extra sections which are MSP430 specific:
5552 @item @samp{.vectors}
5553 Defines a portion of ROM where interrupt vectors located.
5555 @item @samp{.bootloader}
5556 Defines the bootloader portion of the ROM (if applicable). Any code
5557 in this section will be uploaded to the MPU.
5559 @item @samp{.infomem}
5560 Defines an information memory section (if applicable). Any code in
5561 this section will be uploaded to the MPU.
5563 @item @samp{.infomemnobits}
5564 This is the same as the @samp{.infomem} section except that any code
5565 in this section will not be uploaded to the MPU.
5567 @item @samp{.noinit}
5568 Denotes a portion of RAM located above @samp{.bss} section.
5570 The last two sections are used by gcc.
5584 @section @command{ld} and PowerPC 32-bit ELF Support
5585 @cindex PowerPC long branches
5586 @kindex --relax on PowerPC
5587 Branches on PowerPC processors are limited to a signed 26-bit
5588 displacement, which may result in @command{ld} giving
5589 @samp{relocation truncated to fit} errors with very large programs.
5590 @samp{--relax} enables the generation of trampolines that can access
5591 the entire 32-bit address space. These trampolines are inserted at
5592 section boundaries, so may not themselves be reachable if an input
5593 section exceeds 33M in size.
5595 @cindex PowerPC ELF32 options
5600 Current PowerPC GCC accepts a @samp{-msecure-plt} option that
5601 generates code capable of using a newer PLT and GOT layout that has
5602 the security advantage of no executable section ever needing to be
5603 writable and no writable section ever being executable. PowerPC
5604 @command{ld} will generate this layout, including stubs to access the
5605 PLT, if all input files (including startup and static libraries) were
5606 compiled with @samp{-msecure-plt}. @samp{--bss-plt} forces the old
5607 BSS PLT (and GOT layout) which can give slightly better performance.
5612 The new secure PLT and GOT are placed differently relative to other
5613 sections compared to older BSS PLT and GOT placement. The location of
5614 @code{.plt} must change because the new secure PLT is an initialized
5615 section while the old PLT is uninitialized. The reason for the
5616 @code{.got} change is more subtle: The new placement allows
5617 @code{.got} to be read-only in applications linked with
5618 @samp{-z relro -z now}. However, this placement means that
5619 @code{.sdata} cannot always be used in shared libraries, because the
5620 PowerPC ABI accesses @code{.sdata} in shared libraries from the GOT
5621 pointer. @samp{--sdata-got} forces the old GOT placement. PowerPC
5622 GCC doesn't use @code{.sdata} in shared libraries, so this option is
5623 really only useful for other compilers that may do so.
5625 @cindex PowerPC stub symbols
5626 @kindex --emit-stub-syms
5627 @item --emit-stub-syms
5628 This option causes @command{ld} to label linker stubs with a local
5629 symbol that encodes the stub type and destination.
5631 @cindex PowerPC TLS optimization
5632 @kindex --no-tls-optimize
5633 @item --no-tls-optimize
5634 PowerPC @command{ld} normally performs some optimization of code
5635 sequences used to access Thread-Local Storage. Use this option to
5636 disable the optimization.
5649 @node PowerPC64 ELF64
5650 @section @command{ld} and PowerPC64 64-bit ELF Support
5652 @cindex PowerPC64 ELF64 options
5654 @cindex PowerPC64 stub grouping
5655 @kindex --stub-group-size
5656 @item --stub-group-size
5657 Long branch stubs, PLT call stubs and TOC adjusting stubs are placed
5658 by @command{ld} in stub sections located between groups of input sections.
5659 @samp{--stub-group-size} specifies the maximum size of a group of input
5660 sections handled by one stub section. Since branch offsets are signed,
5661 a stub section may serve two groups of input sections, one group before
5662 the stub section, and one group after it. However, when using
5663 conditional branches that require stubs, it may be better (for branch
5664 prediction) that stub sections only serve one group of input sections.
5665 A negative value for @samp{N} chooses this scheme, ensuring that
5666 branches to stubs always use a negative offset. Two special values of
5667 @samp{N} are recognized, @samp{1} and @samp{-1}. These both instruct
5668 @command{ld} to automatically size input section groups for the branch types
5669 detected, with the same behaviour regarding stub placement as other
5670 positive or negative values of @samp{N} respectively.
5672 Note that @samp{--stub-group-size} does not split input sections. A
5673 single input section larger than the group size specified will of course
5674 create a larger group (of one section). If input sections are too
5675 large, it may not be possible for a branch to reach its stub.
5677 @cindex PowerPC64 stub symbols
5678 @kindex --emit-stub-syms
5679 @item --emit-stub-syms
5680 This option causes @command{ld} to label linker stubs with a local
5681 symbol that encodes the stub type and destination.
5683 @cindex PowerPC64 dot symbols
5685 @kindex --no-dotsyms
5686 @item --dotsyms, --no-dotsyms
5687 These two options control how @command{ld} interprets version patterns
5688 in a version script. Older PowerPC64 compilers emitted both a
5689 function descriptor symbol with the same name as the function, and a
5690 code entry symbol with the name prefixed by a dot (@samp{.}). To
5691 properly version a function @samp{foo}, the version script thus needs
5692 to control both @samp{foo} and @samp{.foo}. The option
5693 @samp{--dotsyms}, on by default, automatically adds the required
5694 dot-prefixed patterns. Use @samp{--no-dotsyms} to disable this
5697 @cindex PowerPC64 TLS optimization
5698 @kindex --no-tls-optimize
5699 @item --no-tls-optimize
5700 PowerPC64 @command{ld} normally performs some optimization of code
5701 sequences used to access Thread-Local Storage. Use this option to
5702 disable the optimization.
5704 @cindex PowerPC64 OPD optimization
5705 @kindex --no-opd-optimize
5706 @item --no-opd-optimize
5707 PowerPC64 @command{ld} normally removes @code{.opd} section entries
5708 corresponding to deleted link-once functions, or functions removed by
5709 the action of @samp{--gc-sections} or linker scrip @code{/DISCARD/}.
5710 Use this option to disable @code{.opd} optimization.
5712 @cindex PowerPC64 OPD spacing
5713 @kindex --non-overlapping-opd
5714 @item --non-overlapping-opd
5715 Some PowerPC64 compilers have an option to generate compressed
5716 @code{.opd} entries spaced 16 bytes apart, overlapping the third word,
5717 the static chain pointer (unused in C) with the first word of the next
5718 entry. This option expands such entries to the full 24 bytes.
5720 @cindex PowerPC64 TOC optimization
5721 @kindex --no-toc-optimize
5722 @item --no-toc-optimize
5723 PowerPC64 @command{ld} normally removes unused @code{.toc} section
5724 entries. Such entries are detected by examining relocations that
5725 reference the TOC in code sections. A reloc in a deleted code section
5726 marks a TOC word as unneeded, while a reloc in a kept code section
5727 marks a TOC word as needed. Since the TOC may reference itself, TOC
5728 relocs are also examined. TOC words marked as both needed and
5729 unneeded will of course be kept. TOC words without any referencing
5730 reloc are assumed to be part of a multi-word entry, and are kept or
5731 discarded as per the nearest marked preceding word. This works
5732 reliably for compiler generated code, but may be incorrect if assembly
5733 code is used to insert TOC entries. Use this option to disable the
5736 @cindex PowerPC64 multi-TOC
5737 @kindex --no-multi-toc
5738 @item --no-multi-toc
5739 By default, PowerPC64 GCC generates code for a TOC model where TOC
5740 entries are accessed with a 16-bit offset from r2. This limits the
5741 total TOC size to 64K. PowerPC64 @command{ld} extends this limit by
5742 grouping code sections such that each group uses less than 64K for its
5743 TOC entries, then inserts r2 adjusting stubs between inter-group
5744 calls. @command{ld} does not split apart input sections, so cannot
5745 help if a single input file has a @code{.toc} section that exceeds
5746 64K, most likely from linking multiple files with @command{ld -r}.
5747 Use this option to turn off this feature.
5761 @section @command{ld}'s Support for Various TI COFF Versions
5762 @cindex TI COFF versions
5763 @kindex --format=@var{version}
5764 The @samp{--format} switch allows selection of one of the various
5765 TI COFF versions. The latest of this writing is 2; versions 0 and 1 are
5766 also supported. The TI COFF versions also vary in header byte-order
5767 format; @command{ld} will read any version or byte order, but the output
5768 header format depends on the default specified by the specific target.
5781 @section @command{ld} and WIN32 (cygwin/mingw)
5783 This section describes some of the win32 specific @command{ld} issues.
5784 See @ref{Options,,Command Line Options} for detailed description of the
5785 command line options mentioned here.
5788 @cindex import libraries
5789 @item import libraries
5790 The standard Windows linker creates and uses so-called import
5791 libraries, which contains information for linking to dll's. They are
5792 regular static archives and are handled as any other static
5793 archive. The cygwin and mingw ports of @command{ld} have specific
5794 support for creating such libraries provided with the
5795 @samp{--out-implib} command line option.
5797 @item exporting DLL symbols
5798 @cindex exporting DLL symbols
5799 The cygwin/mingw @command{ld} has several ways to export symbols for dll's.
5802 @item using auto-export functionality
5803 @cindex using auto-export functionality
5804 By default @command{ld} exports symbols with the auto-export functionality,
5805 which is controlled by the following command line options:
5808 @item --export-all-symbols [This is the default]
5809 @item --exclude-symbols
5810 @item --exclude-libs
5813 If, however, @samp{--export-all-symbols} is not given explicitly on the
5814 command line, then the default auto-export behavior will be @emph{disabled}
5815 if either of the following are true:
5818 @item A DEF file is used.
5819 @item Any symbol in any object file was marked with the __declspec(dllexport) attribute.
5822 @item using a DEF file
5823 @cindex using a DEF file
5824 Another way of exporting symbols is using a DEF file. A DEF file is
5825 an ASCII file containing definitions of symbols which should be
5826 exported when a dll is created. Usually it is named @samp{<dll
5827 name>.def} and is added as any other object file to the linker's
5828 command line. The file's name must end in @samp{.def} or @samp{.DEF}.
5831 gcc -o <output> <objectfiles> <dll name>.def
5834 Using a DEF file turns off the normal auto-export behavior, unless the
5835 @samp{--export-all-symbols} option is also used.
5837 Here is an example of a DEF file for a shared library called @samp{xyz.dll}:
5840 LIBRARY "xyz.dll" BASE=0x20000000
5846 another_foo = abc.dll.afoo
5850 This example defines a DLL with a non-default base address and five
5851 symbols in the export table. The third exported symbol @code{_bar} is an
5852 alias for the second. The fourth symbol, @code{another_foo} is resolved
5853 by "forwarding" to another module and treating it as an alias for
5854 @code{afoo} exported from the DLL @samp{abc.dll}. The final symbol
5855 @code{var1} is declared to be a data object.
5857 The optional @code{LIBRARY <name>} command indicates the @emph{internal}
5858 name of the output DLL. If @samp{<name>} does not include a suffix,
5859 the default library suffix, @samp{.DLL} is appended.
5861 When the .DEF file is used to build an application, rather than a
5862 library, the @code{NAME <name>} command should be used instead of
5863 @code{LIBRARY}. If @samp{<name>} does not include a suffix, the default
5864 executable suffix, @samp{.EXE} is appended.
5866 With either @code{LIBRARY <name>} or @code{NAME <name>} the optional
5867 specification @code{BASE = <number>} may be used to specify a
5868 non-default base address for the image.
5870 If neither @code{LIBRARY <name>} nor @code{NAME <name>} is specified,
5871 or they specify an empty string, the internal name is the same as the
5872 filename specified on the command line.
5874 The complete specification of an export symbol is:
5878 ( ( ( <name1> [ = <name2> ] )
5879 | ( <name1> = <module-name> . <external-name>))
5880 [ @@ <integer> ] [NONAME] [DATA] [CONSTANT] [PRIVATE] ) *
5883 Declares @samp{<name1>} as an exported symbol from the DLL, or declares
5884 @samp{<name1>} as an exported alias for @samp{<name2>}; or declares
5885 @samp{<name1>} as a "forward" alias for the symbol
5886 @samp{<external-name>} in the DLL @samp{<module-name>}.
5887 Optionally, the symbol may be exported by the specified ordinal
5888 @samp{<integer>} alias.
5890 The optional keywords that follow the declaration indicate:
5892 @code{NONAME}: Do not put the symbol name in the DLL's export table. It
5893 will still be exported by its ordinal alias (either the value specified
5894 by the .def specification or, otherwise, the value assigned by the
5895 linker). The symbol name, however, does remain visible in the import
5896 library (if any), unless @code{PRIVATE} is also specified.
5898 @code{DATA}: The symbol is a variable or object, rather than a function.
5899 The import lib will export only an indirect reference to @code{foo} as
5900 the symbol @code{_imp__foo} (ie, @code{foo} must be resolved as
5903 @code{CONSTANT}: Like @code{DATA}, but put the undecorated @code{foo} as
5904 well as @code{_imp__foo} into the import library. Both refer to the
5905 read-only import address table's pointer to the variable, not to the
5906 variable itself. This can be dangerous. If the user code fails to add
5907 the @code{dllimport} attribute and also fails to explicitly add the
5908 extra indirection that the use of the attribute enforces, the
5909 application will behave unexpectedly.
5911 @code{PRIVATE}: Put the symbol in the DLL's export table, but do not put
5912 it into the static import library used to resolve imports at link time. The
5913 symbol can still be imported using the @code{LoadLibrary/GetProcAddress}
5914 API at runtime or by by using the GNU ld extension of linking directly to
5915 the DLL without an import library.
5917 See ld/deffilep.y in the binutils sources for the full specification of
5918 other DEF file statements
5920 @cindex creating a DEF file
5921 While linking a shared dll, @command{ld} is able to create a DEF file
5922 with the @samp{--output-def <file>} command line option.
5924 @item Using decorations
5925 @cindex Using decorations
5926 Another way of marking symbols for export is to modify the source code
5927 itself, so that when building the DLL each symbol to be exported is
5931 __declspec(dllexport) int a_variable
5932 __declspec(dllexport) void a_function(int with_args)
5935 All such symbols will be exported from the DLL. If, however,
5936 any of the object files in the DLL contain symbols decorated in
5937 this way, then the normal auto-export behavior is disabled, unless
5938 the @samp{--export-all-symbols} option is also used.
5940 Note that object files that wish to access these symbols must @emph{not}
5941 decorate them with dllexport. Instead, they should use dllimport,
5945 __declspec(dllimport) int a_variable
5946 __declspec(dllimport) void a_function(int with_args)
5949 This complicates the structure of library header files, because
5950 when included by the library itself the header must declare the
5951 variables and functions as dllexport, but when included by client
5952 code the header must declare them as dllimport. There are a number
5953 of idioms that are typically used to do this; often client code can
5954 omit the __declspec() declaration completely. See
5955 @samp{--enable-auto-import} and @samp{automatic data imports} for more
5959 @cindex automatic data imports
5960 @item automatic data imports
5961 The standard Windows dll format supports data imports from dlls only
5962 by adding special decorations (dllimport/dllexport), which let the
5963 compiler produce specific assembler instructions to deal with this
5964 issue. This increases the effort necessary to port existing Un*x
5965 code to these platforms, especially for large
5966 c++ libraries and applications. The auto-import feature, which was
5967 initially provided by Paul Sokolovsky, allows one to omit the
5968 decorations to achieve a behavior that conforms to that on POSIX/Un*x
5969 platforms. This feature is enabled with the @samp{--enable-auto-import}
5970 command-line option, although it is enabled by default on cygwin/mingw.
5971 The @samp{--enable-auto-import} option itself now serves mainly to
5972 suppress any warnings that are ordinarily emitted when linked objects
5973 trigger the feature's use.
5975 auto-import of variables does not always work flawlessly without
5976 additional assistance. Sometimes, you will see this message
5978 "variable '<var>' can't be auto-imported. Please read the
5979 documentation for ld's @code{--enable-auto-import} for details."
5981 The @samp{--enable-auto-import} documentation explains why this error
5982 occurs, and several methods that can be used to overcome this difficulty.
5983 One of these methods is the @emph{runtime pseudo-relocs} feature, described
5986 @cindex runtime pseudo-relocation
5987 For complex variables imported from DLLs (such as structs or classes),
5988 object files typically contain a base address for the variable and an
5989 offset (@emph{addend}) within the variable--to specify a particular
5990 field or public member, for instance. Unfortunately, the runtime loader used
5991 in win32 environments is incapable of fixing these references at runtime
5992 without the additional information supplied by dllimport/dllexport decorations.
5993 The standard auto-import feature described above is unable to resolve these
5996 The @samp{--enable-runtime-pseudo-relocs} switch allows these references to
5997 be resolved without error, while leaving the task of adjusting the references
5998 themselves (with their non-zero addends) to specialized code provided by the
5999 runtime environment. Recent versions of the cygwin and mingw environments and
6000 compilers provide this runtime support; older versions do not. However, the
6001 support is only necessary on the developer's platform; the compiled result will
6002 run without error on an older system.
6004 @samp{--enable-runtime-pseudo-relocs} is not the default; it must be explicitly
6007 @cindex direct linking to a dll
6008 @item direct linking to a dll
6009 The cygwin/mingw ports of @command{ld} support the direct linking,
6010 including data symbols, to a dll without the usage of any import
6011 libraries. This is much faster and uses much less memory than does the
6012 traditional import library method, especially when linking large
6013 libraries or applications. When @command{ld} creates an import lib, each
6014 function or variable exported from the dll is stored in its own bfd, even
6015 though a single bfd could contain many exports. The overhead involved in
6016 storing, loading, and processing so many bfd's is quite large, and explains the
6017 tremendous time, memory, and storage needed to link against particularly
6018 large or complex libraries when using import libs.
6020 Linking directly to a dll uses no extra command-line switches other than
6021 @samp{-L} and @samp{-l}, because @command{ld} already searches for a number
6022 of names to match each library. All that is needed from the developer's
6023 perspective is an understanding of this search, in order to force ld to
6024 select the dll instead of an import library.
6027 For instance, when ld is called with the argument @samp{-lxxx} it will attempt
6028 to find, in the first directory of its search path,
6040 before moving on to the next directory in the search path.
6042 (*) Actually, this is not @samp{cygxxx.dll} but in fact is @samp{<prefix>xxx.dll},
6043 where @samp{<prefix>} is set by the @command{ld} option
6044 @samp{--dll-search-prefix=<prefix>}. In the case of cygwin, the standard gcc spec
6045 file includes @samp{--dll-search-prefix=cyg}, so in effect we actually search for
6048 Other win32-based unix environments, such as mingw or pw32, may use other
6049 @samp{<prefix>}es, although at present only cygwin makes use of this feature. It
6050 was originally intended to help avoid name conflicts among dll's built for the
6051 various win32/un*x environments, so that (for example) two versions of a zlib dll
6052 could coexist on the same machine.
6054 The generic cygwin/mingw path layout uses a @samp{bin} directory for
6055 applications and dll's and a @samp{lib} directory for the import
6056 libraries (using cygwin nomenclature):
6062 libxxx.dll.a (in case of dll's)
6063 libxxx.a (in case of static archive)
6066 Linking directly to a dll without using the import library can be
6069 1. Use the dll directly by adding the @samp{bin} path to the link line
6071 gcc -Wl,-verbose -o a.exe -L../bin/ -lxxx
6074 However, as the dll's often have version numbers appended to their names
6075 (@samp{cygncurses-5.dll}) this will often fail, unless one specifies
6076 @samp{-L../bin -lncurses-5} to include the version. Import libs are generally
6077 not versioned, and do not have this difficulty.
6079 2. Create a symbolic link from the dll to a file in the @samp{lib}
6080 directory according to the above mentioned search pattern. This
6081 should be used to avoid unwanted changes in the tools needed for
6085 ln -s bin/cygxxx.dll lib/[cyg|lib|]xxx.dll[.a]
6088 Then you can link without any make environment changes.
6091 gcc -Wl,-verbose -o a.exe -L../lib/ -lxxx
6094 This technique also avoids the version number problems, because the following is
6101 libxxx.dll.a -> ../bin/cygxxx-5.dll
6104 Linking directly to a dll without using an import lib will work
6105 even when auto-import features are exercised, and even when
6106 @samp{--enable-runtime-pseudo-relocs} is used.
6108 Given the improvements in speed and memory usage, one might justifiably
6109 wonder why import libraries are used at all. There are three reasons:
6111 1. Until recently, the link-directly-to-dll functionality did @emph{not}
6112 work with auto-imported data.
6114 2. Sometimes it is necessary to include pure static objects within the
6115 import library (which otherwise contains only bfd's for indirection
6116 symbols that point to the exports of a dll). Again, the import lib
6117 for the cygwin kernel makes use of this ability, and it is not
6118 possible to do this without an import lib.
6120 3. Symbol aliases can only be resolved using an import lib. This is
6121 critical when linking against OS-supplied dll's (eg, the win32 API)
6122 in which symbols are usually exported as undecorated aliases of their
6123 stdcall-decorated assembly names.
6125 So, import libs are not going away. But the ability to replace
6126 true import libs with a simple symbolic link to (or a copy of)
6127 a dll, in many cases, is a useful addition to the suite of tools
6128 binutils makes available to the win32 developer. Given the
6129 massive improvements in memory requirements during linking, storage
6130 requirements, and linking speed, we expect that many developers
6131 will soon begin to use this feature whenever possible.
6133 @item symbol aliasing
6135 @item adding additional names
6136 Sometimes, it is useful to export symbols with additional names.
6137 A symbol @samp{foo} will be exported as @samp{foo}, but it can also be
6138 exported as @samp{_foo} by using special directives in the DEF file
6139 when creating the dll. This will affect also the optional created
6140 import library. Consider the following DEF file:
6143 LIBRARY "xyz.dll" BASE=0x61000000
6150 The line @samp{_foo = foo} maps the symbol @samp{foo} to @samp{_foo}.
6152 Another method for creating a symbol alias is to create it in the
6153 source code using the "weak" attribute:
6156 void foo () @{ /* Do something. */; @}
6157 void _foo () __attribute__ ((weak, alias ("foo")));
6160 See the gcc manual for more information about attributes and weak
6163 @item renaming symbols
6164 Sometimes it is useful to rename exports. For instance, the cygwin
6165 kernel does this regularly. A symbol @samp{_foo} can be exported as
6166 @samp{foo} but not as @samp{_foo} by using special directives in the
6167 DEF file. (This will also affect the import library, if it is
6168 created). In the following example:
6171 LIBRARY "xyz.dll" BASE=0x61000000
6177 The line @samp{_foo = foo} maps the exported symbol @samp{foo} to
6181 Note: using a DEF file disables the default auto-export behavior,
6182 unless the @samp{--export-all-symbols} command line option is used.
6183 If, however, you are trying to rename symbols, then you should list
6184 @emph{all} desired exports in the DEF file, including the symbols
6185 that are not being renamed, and do @emph{not} use the
6186 @samp{--export-all-symbols} option. If you list only the
6187 renamed symbols in the DEF file, and use @samp{--export-all-symbols}
6188 to handle the other symbols, then the both the new names @emph{and}
6189 the original names for the renamed symbols will be exported.
6190 In effect, you'd be aliasing those symbols, not renaming them,
6191 which is probably not what you wanted.
6193 @cindex weak externals
6194 @item weak externals
6195 The Windows object format, PE, specifies a form of weak symbols called
6196 weak externals. When a weak symbol is linked and the symbol is not
6197 defined, the weak symbol becomes an alias for some other symbol. There
6198 are three variants of weak externals:
6200 @item Definition is searched for in objects and libraries, historically
6201 called lazy externals.
6202 @item Definition is searched for only in other objects, not in libraries.
6203 This form is not presently implemented.
6204 @item No search; the symbol is an alias. This form is not presently
6207 As a GNU extension, weak symbols that do not specify an alternate symbol
6208 are supported. If the symbol is undefined when linking, the symbol
6209 uses a default value.
6223 @section @code{ld} and Xtensa Processors
6225 @cindex Xtensa processors
6226 The default @command{ld} behavior for Xtensa processors is to interpret
6227 @code{SECTIONS} commands so that lists of explicitly named sections in a
6228 specification with a wildcard file will be interleaved when necessary to
6229 keep literal pools within the range of PC-relative load offsets. For
6230 example, with the command:
6242 @command{ld} may interleave some of the @code{.literal}
6243 and @code{.text} sections from different object files to ensure that the
6244 literal pools are within the range of PC-relative load offsets. A valid
6245 interleaving might place the @code{.literal} sections from an initial
6246 group of files followed by the @code{.text} sections of that group of
6247 files. Then, the @code{.literal} sections from the rest of the files
6248 and the @code{.text} sections from the rest of the files would follow.
6250 @cindex @option{--relax} on Xtensa
6251 @cindex relaxing on Xtensa
6252 Relaxation is enabled by default for the Xtensa version of @command{ld} and
6253 provides two important link-time optimizations. The first optimization
6254 is to combine identical literal values to reduce code size. A redundant
6255 literal will be removed and all the @code{L32R} instructions that use it
6256 will be changed to reference an identical literal, as long as the
6257 location of the replacement literal is within the offset range of all
6258 the @code{L32R} instructions. The second optimization is to remove
6259 unnecessary overhead from assembler-generated ``longcall'' sequences of
6260 @code{L32R}/@code{CALLX@var{n}} when the target functions are within
6261 range of direct @code{CALL@var{n}} instructions.
6263 For each of these cases where an indirect call sequence can be optimized
6264 to a direct call, the linker will change the @code{CALLX@var{n}}
6265 instruction to a @code{CALL@var{n}} instruction, remove the @code{L32R}
6266 instruction, and remove the literal referenced by the @code{L32R}
6267 instruction if it is not used for anything else. Removing the
6268 @code{L32R} instruction always reduces code size but can potentially
6269 hurt performance by changing the alignment of subsequent branch targets.
6270 By default, the linker will always preserve alignments, either by
6271 switching some instructions between 24-bit encodings and the equivalent
6272 density instructions or by inserting a no-op in place of the @code{L32R}
6273 instruction that was removed. If code size is more important than
6274 performance, the @option{--size-opt} option can be used to prevent the
6275 linker from widening density instructions or inserting no-ops, except in
6276 a few cases where no-ops are required for correctness.
6278 The following Xtensa-specific command-line options can be used to
6281 @cindex Xtensa options
6285 Since the Xtensa version of @code{ld} enables the @option{--relax} option
6286 by default, the @option{--no-relax} option is provided to disable
6290 When optimizing indirect calls to direct calls, optimize for code size
6291 more than performance. With this option, the linker will not insert
6292 no-ops or widen density instructions to preserve branch target
6293 alignment. There may still be some cases where no-ops are required to
6294 preserve the correctness of the code.
6302 @ifclear SingleFormat
6307 @cindex object file management
6308 @cindex object formats available
6310 The linker accesses object and archive files using the BFD libraries.
6311 These libraries allow the linker to use the same routines to operate on
6312 object files whatever the object file format. A different object file
6313 format can be supported simply by creating a new BFD back end and adding
6314 it to the library. To conserve runtime memory, however, the linker and
6315 associated tools are usually configured to support only a subset of the
6316 object file formats available. You can use @code{objdump -i}
6317 (@pxref{objdump,,objdump,binutils.info,The GNU Binary Utilities}) to
6318 list all the formats available for your configuration.
6320 @cindex BFD requirements
6321 @cindex requirements for BFD
6322 As with most implementations, BFD is a compromise between
6323 several conflicting requirements. The major factor influencing
6324 BFD design was efficiency: any time used converting between
6325 formats is time which would not have been spent had BFD not
6326 been involved. This is partly offset by abstraction payback; since
6327 BFD simplifies applications and back ends, more time and care
6328 may be spent optimizing algorithms for a greater speed.
6330 One minor artifact of the BFD solution which you should bear in
6331 mind is the potential for information loss. There are two places where
6332 useful information can be lost using the BFD mechanism: during
6333 conversion and during output. @xref{BFD information loss}.
6336 * BFD outline:: How it works: an outline of BFD
6340 @section How It Works: An Outline of BFD
6341 @cindex opening object files
6342 @include bfdsumm.texi
6345 @node Reporting Bugs
6346 @chapter Reporting Bugs
6347 @cindex bugs in @command{ld}
6348 @cindex reporting bugs in @command{ld}
6350 Your bug reports play an essential role in making @command{ld} reliable.
6352 Reporting a bug may help you by bringing a solution to your problem, or
6353 it may not. But in any case the principal function of a bug report is
6354 to help the entire community by making the next version of @command{ld}
6355 work better. Bug reports are your contribution to the maintenance of
6358 In order for a bug report to serve its purpose, you must include the
6359 information that enables us to fix the bug.
6362 * Bug Criteria:: Have you found a bug?
6363 * Bug Reporting:: How to report bugs
6367 @section Have You Found a Bug?
6368 @cindex bug criteria
6370 If you are not sure whether you have found a bug, here are some guidelines:
6373 @cindex fatal signal
6374 @cindex linker crash
6375 @cindex crash of linker
6377 If the linker gets a fatal signal, for any input whatever, that is a
6378 @command{ld} bug. Reliable linkers never crash.
6380 @cindex error on valid input
6382 If @command{ld} produces an error message for valid input, that is a bug.
6384 @cindex invalid input
6386 If @command{ld} does not produce an error message for invalid input, that
6387 may be a bug. In the general case, the linker can not verify that
6388 object files are correct.
6391 If you are an experienced user of linkers, your suggestions for
6392 improvement of @command{ld} are welcome in any case.
6396 @section How to Report Bugs
6398 @cindex @command{ld} bugs, reporting
6400 A number of companies and individuals offer support for @sc{gnu}
6401 products. If you obtained @command{ld} from a support organization, we
6402 recommend you contact that organization first.
6404 You can find contact information for many support companies and
6405 individuals in the file @file{etc/SERVICE} in the @sc{gnu} Emacs
6408 Otherwise, send bug reports for @command{ld} to
6409 @samp{bug-binutils@@gnu.org}.
6411 The fundamental principle of reporting bugs usefully is this:
6412 @strong{report all the facts}. If you are not sure whether to state a
6413 fact or leave it out, state it!
6415 Often people omit facts because they think they know what causes the
6416 problem and assume that some details do not matter. Thus, you might
6417 assume that the name of a symbol you use in an example does not
6418 matter. Well, probably it does not, but one cannot be sure. Perhaps
6419 the bug is a stray memory reference which happens to fetch from the
6420 location where that name is stored in memory; perhaps, if the name
6421 were different, the contents of that location would fool the linker
6422 into doing the right thing despite the bug. Play it safe and give a
6423 specific, complete example. That is the easiest thing for you to do,
6424 and the most helpful.
6426 Keep in mind that the purpose of a bug report is to enable us to fix
6427 the bug if it is new to us. Therefore, always write your bug reports
6428 on the assumption that the bug has not been reported previously.
6430 Sometimes people give a few sketchy facts and ask, ``Does this ring a
6431 bell?'' This cannot help us fix a bug, so it is basically useless. We
6432 respond by asking for enough details to enable us to investigate.
6433 You might as well expedite matters by sending them to begin with.
6435 To enable us to fix the bug, you should include all these things:
6439 The version of @command{ld}. @command{ld} announces it if you start it with
6440 the @samp{--version} argument.
6442 Without this, we will not know whether there is any point in looking for
6443 the bug in the current version of @command{ld}.
6446 Any patches you may have applied to the @command{ld} source, including any
6447 patches made to the @code{BFD} library.
6450 The type of machine you are using, and the operating system name and
6454 What compiler (and its version) was used to compile @command{ld}---e.g.
6458 The command arguments you gave the linker to link your example and
6459 observe the bug. To guarantee you will not omit something important,
6460 list them all. A copy of the Makefile (or the output from make) is
6463 If we were to try to guess the arguments, we would probably guess wrong
6464 and then we might not encounter the bug.
6467 A complete input file, or set of input files, that will reproduce the
6468 bug. It is generally most helpful to send the actual object files
6469 provided that they are reasonably small. Say no more than 10K. For
6470 bigger files you can either make them available by FTP or HTTP or else
6471 state that you are willing to send the object file(s) to whomever
6472 requests them. (Note - your email will be going to a mailing list, so
6473 we do not want to clog it up with large attachments). But small
6474 attachments are best.
6476 If the source files were assembled using @code{gas} or compiled using
6477 @code{gcc}, then it may be OK to send the source files rather than the
6478 object files. In this case, be sure to say exactly what version of
6479 @code{gas} or @code{gcc} was used to produce the object files. Also say
6480 how @code{gas} or @code{gcc} were configured.
6483 A description of what behavior you observe that you believe is
6484 incorrect. For example, ``It gets a fatal signal.''
6486 Of course, if the bug is that @command{ld} gets a fatal signal, then we
6487 will certainly notice it. But if the bug is incorrect output, we might
6488 not notice unless it is glaringly wrong. You might as well not give us
6489 a chance to make a mistake.
6491 Even if the problem you experience is a fatal signal, you should still
6492 say so explicitly. Suppose something strange is going on, such as, your
6493 copy of @command{ld} is out of sync, or you have encountered a bug in the
6494 C library on your system. (This has happened!) Your copy might crash
6495 and ours would not. If you told us to expect a crash, then when ours
6496 fails to crash, we would know that the bug was not happening for us. If
6497 you had not told us to expect a crash, then we would not be able to draw
6498 any conclusion from our observations.
6501 If you wish to suggest changes to the @command{ld} source, send us context
6502 diffs, as generated by @code{diff} with the @samp{-u}, @samp{-c}, or
6503 @samp{-p} option. Always send diffs from the old file to the new file.
6504 If you even discuss something in the @command{ld} source, refer to it by
6505 context, not by line number.
6507 The line numbers in our development sources will not match those in your
6508 sources. Your line numbers would convey no useful information to us.
6511 Here are some things that are not necessary:
6515 A description of the envelope of the bug.
6517 Often people who encounter a bug spend a lot of time investigating
6518 which changes to the input file will make the bug go away and which
6519 changes will not affect it.
6521 This is often time consuming and not very useful, because the way we
6522 will find the bug is by running a single example under the debugger
6523 with breakpoints, not by pure deduction from a series of examples.
6524 We recommend that you save your time for something else.
6526 Of course, if you can find a simpler example to report @emph{instead}
6527 of the original one, that is a convenience for us. Errors in the
6528 output will be easier to spot, running under the debugger will take
6529 less time, and so on.
6531 However, simplification is not vital; if you do not want to do this,
6532 report the bug anyway and send us the entire test case you used.
6535 A patch for the bug.
6537 A patch for the bug does help us if it is a good one. But do not omit
6538 the necessary information, such as the test case, on the assumption that
6539 a patch is all we need. We might see problems with your patch and decide
6540 to fix the problem another way, or we might not understand it at all.
6542 Sometimes with a program as complicated as @command{ld} it is very hard to
6543 construct an example that will make the program follow a certain path
6544 through the code. If you do not send us the example, we will not be
6545 able to construct one, so we will not be able to verify that the bug is
6548 And if we cannot understand what bug you are trying to fix, or why your
6549 patch should be an improvement, we will not install it. A test case will
6550 help us to understand.
6553 A guess about what the bug is or what it depends on.
6555 Such guesses are usually wrong. Even we cannot guess right about such
6556 things without first using the debugger to find the facts.
6560 @appendix MRI Compatible Script Files
6561 @cindex MRI compatibility
6562 To aid users making the transition to @sc{gnu} @command{ld} from the MRI
6563 linker, @command{ld} can use MRI compatible linker scripts as an
6564 alternative to the more general-purpose linker scripting language
6565 described in @ref{Scripts}. MRI compatible linker scripts have a much
6566 simpler command set than the scripting language otherwise used with
6567 @command{ld}. @sc{gnu} @command{ld} supports the most commonly used MRI
6568 linker commands; these commands are described here.
6570 In general, MRI scripts aren't of much use with the @code{a.out} object
6571 file format, since it only has three sections and MRI scripts lack some
6572 features to make use of them.
6574 You can specify a file containing an MRI-compatible script using the
6575 @samp{-c} command-line option.
6577 Each command in an MRI-compatible script occupies its own line; each
6578 command line starts with the keyword that identifies the command (though
6579 blank lines are also allowed for punctuation). If a line of an
6580 MRI-compatible script begins with an unrecognized keyword, @command{ld}
6581 issues a warning message, but continues processing the script.
6583 Lines beginning with @samp{*} are comments.
6585 You can write these commands using all upper-case letters, or all
6586 lower case; for example, @samp{chip} is the same as @samp{CHIP}.
6587 The following list shows only the upper-case form of each command.
6590 @cindex @code{ABSOLUTE} (MRI)
6591 @item ABSOLUTE @var{secname}
6592 @itemx ABSOLUTE @var{secname}, @var{secname}, @dots{} @var{secname}
6593 Normally, @command{ld} includes in the output file all sections from all
6594 the input files. However, in an MRI-compatible script, you can use the
6595 @code{ABSOLUTE} command to restrict the sections that will be present in
6596 your output program. If the @code{ABSOLUTE} command is used at all in a
6597 script, then only the sections named explicitly in @code{ABSOLUTE}
6598 commands will appear in the linker output. You can still use other
6599 input sections (whatever you select on the command line, or using
6600 @code{LOAD}) to resolve addresses in the output file.
6602 @cindex @code{ALIAS} (MRI)
6603 @item ALIAS @var{out-secname}, @var{in-secname}
6604 Use this command to place the data from input section @var{in-secname}
6605 in a section called @var{out-secname} in the linker output file.
6607 @var{in-secname} may be an integer.
6609 @cindex @code{ALIGN} (MRI)
6610 @item ALIGN @var{secname} = @var{expression}
6611 Align the section called @var{secname} to @var{expression}. The
6612 @var{expression} should be a power of two.
6614 @cindex @code{BASE} (MRI)
6615 @item BASE @var{expression}
6616 Use the value of @var{expression} as the lowest address (other than
6617 absolute addresses) in the output file.
6619 @cindex @code{CHIP} (MRI)
6620 @item CHIP @var{expression}
6621 @itemx CHIP @var{expression}, @var{expression}
6622 This command does nothing; it is accepted only for compatibility.
6624 @cindex @code{END} (MRI)
6626 This command does nothing whatever; it's only accepted for compatibility.
6628 @cindex @code{FORMAT} (MRI)
6629 @item FORMAT @var{output-format}
6630 Similar to the @code{OUTPUT_FORMAT} command in the more general linker
6631 language, but restricted to one of these output formats:
6635 S-records, if @var{output-format} is @samp{S}
6638 IEEE, if @var{output-format} is @samp{IEEE}
6641 COFF (the @samp{coff-m68k} variant in BFD), if @var{output-format} is
6645 @cindex @code{LIST} (MRI)
6646 @item LIST @var{anything}@dots{}
6647 Print (to the standard output file) a link map, as produced by the
6648 @command{ld} command-line option @samp{-M}.
6650 The keyword @code{LIST} may be followed by anything on the
6651 same line, with no change in its effect.
6653 @cindex @code{LOAD} (MRI)
6654 @item LOAD @var{filename}
6655 @itemx LOAD @var{filename}, @var{filename}, @dots{} @var{filename}
6656 Include one or more object file @var{filename} in the link; this has the
6657 same effect as specifying @var{filename} directly on the @command{ld}
6660 @cindex @code{NAME} (MRI)
6661 @item NAME @var{output-name}
6662 @var{output-name} is the name for the program produced by @command{ld}; the
6663 MRI-compatible command @code{NAME} is equivalent to the command-line
6664 option @samp{-o} or the general script language command @code{OUTPUT}.
6666 @cindex @code{ORDER} (MRI)
6667 @item ORDER @var{secname}, @var{secname}, @dots{} @var{secname}
6668 @itemx ORDER @var{secname} @var{secname} @var{secname}
6669 Normally, @command{ld} orders the sections in its output file in the
6670 order in which they first appear in the input files. In an MRI-compatible
6671 script, you can override this ordering with the @code{ORDER} command. The
6672 sections you list with @code{ORDER} will appear first in your output
6673 file, in the order specified.
6675 @cindex @code{PUBLIC} (MRI)
6676 @item PUBLIC @var{name}=@var{expression}
6677 @itemx PUBLIC @var{name},@var{expression}
6678 @itemx PUBLIC @var{name} @var{expression}
6679 Supply a value (@var{expression}) for external symbol
6680 @var{name} used in the linker input files.
6682 @cindex @code{SECT} (MRI)
6683 @item SECT @var{secname}, @var{expression}
6684 @itemx SECT @var{secname}=@var{expression}
6685 @itemx SECT @var{secname} @var{expression}
6686 You can use any of these three forms of the @code{SECT} command to
6687 specify the start address (@var{expression}) for section @var{secname}.
6688 If you have more than one @code{SECT} statement for the same
6689 @var{secname}, only the @emph{first} sets the start address.
6695 @unnumbered LD Index
6700 % I think something like @colophon should be in texinfo. In the
6702 \long\def\colophon{\hbox to0pt{}\vfill
6703 \centerline{The body of this manual is set in}
6704 \centerline{\fontname\tenrm,}
6705 \centerline{with headings in {\bf\fontname\tenbf}}
6706 \centerline{and examples in {\tt\fontname\tentt}.}
6707 \centerline{{\it\fontname\tenit\/} and}
6708 \centerline{{\sl\fontname\tensl\/}}
6709 \centerline{are used for emphasis.}\vfill}
6711 % Blame: doc@cygnus.com, 28mar91.