3 @c Copyright 1991-2013 Free Software Foundation, Inc.
6 @include configdoc.texi
7 @c (configdoc.texi is generated by the Makefile)
13 @macro gcctabopt{body}
19 @c Configure for the generation of man pages
45 @dircategory Software development
47 * Ld: (ld). The GNU linker.
52 This file documents the @sc{gnu} linker LD
53 @ifset VERSION_PACKAGE
54 @value{VERSION_PACKAGE}
56 version @value{VERSION}.
58 Copyright @copyright{} 1991-2013 Free Software Foundation, Inc.
60 Permission is granted to copy, distribute and/or modify this document
61 under the terms of the GNU Free Documentation License, Version 1.3
62 or any later version published by the Free Software Foundation;
63 with no Invariant Sections, with no Front-Cover Texts, and with no
64 Back-Cover Texts. A copy of the license is included in the
65 section entitled ``GNU Free Documentation License''.
69 @setchapternewpage odd
70 @settitle The GNU linker
75 @ifset VERSION_PACKAGE
76 @subtitle @value{VERSION_PACKAGE}
78 @subtitle Version @value{VERSION}
79 @author Steve Chamberlain
80 @author Ian Lance Taylor
85 \hfill Red Hat Inc\par
86 \hfill nickc\@credhat.com, doc\@redhat.com\par
87 \hfill {\it The GNU linker}\par
88 \hfill Edited by Jeffrey Osier (jeffrey\@cygnus.com)\par
90 \global\parindent=0pt % Steve likes it this way.
93 @vskip 0pt plus 1filll
94 @c man begin COPYRIGHT
95 Copyright @copyright{} 1991-2013 Free Software Foundation, Inc.
97 Permission is granted to copy, distribute and/or modify this document
98 under the terms of the GNU Free Documentation License, Version 1.3
99 or any later version published by the Free Software Foundation;
100 with no Invariant Sections, with no Front-Cover Texts, and with no
101 Back-Cover Texts. A copy of the license is included in the
102 section entitled ``GNU Free Documentation License''.
108 @c FIXME: Talk about importance of *order* of args, cmds to linker!
113 This file documents the @sc{gnu} linker ld
114 @ifset VERSION_PACKAGE
115 @value{VERSION_PACKAGE}
117 version @value{VERSION}.
119 This document is distributed under the terms of the GNU Free
120 Documentation License version 1.3. A copy of the license is included
121 in the section entitled ``GNU Free Documentation License''.
124 * Overview:: Overview
125 * Invocation:: Invocation
126 * Scripts:: Linker Scripts
128 * Machine Dependent:: Machine Dependent Features
132 * H8/300:: ld and the H8/300
135 * Renesas:: ld and other Renesas micros
138 * i960:: ld and the Intel 960 family
141 * ARM:: ld and the ARM family
144 * M68HC11/68HC12:: ld and the Motorola 68HC11 and 68HC12 families
147 * HPPA ELF32:: ld and HPPA 32-bit ELF
150 * M68K:: ld and Motorola 68K family
153 * MIPS:: ld and MIPS family
156 * PowerPC ELF32:: ld and PowerPC 32-bit ELF Support
159 * PowerPC64 ELF64:: ld and PowerPC64 64-bit ELF Support
162 * SPU ELF:: ld and SPU ELF Support
165 * TI COFF:: ld and the TI COFF
168 * Win32:: ld and WIN32 (cygwin/mingw)
171 * Xtensa:: ld and Xtensa Processors
174 @ifclear SingleFormat
177 @c Following blank line required for remaining bug in makeinfo conds/menus
179 * Reporting Bugs:: Reporting Bugs
180 * MRI:: MRI Compatible Script Files
181 * GNU Free Documentation License:: GNU Free Documentation License
182 * LD Index:: LD Index
189 @cindex @sc{gnu} linker
190 @cindex what is this?
193 @c man begin SYNOPSIS
194 ld [@b{options}] @var{objfile} @dots{}
198 ar(1), nm(1), objcopy(1), objdump(1), readelf(1) and
199 the Info entries for @file{binutils} and
204 @c man begin DESCRIPTION
206 @command{ld} combines a number of object and archive files, relocates
207 their data and ties up symbol references. Usually the last step in
208 compiling a program is to run @command{ld}.
210 @command{ld} accepts Linker Command Language files written in
211 a superset of AT&T's Link Editor Command Language syntax,
212 to provide explicit and total control over the linking process.
216 This man page does not describe the command language; see the
217 @command{ld} entry in @code{info} for full details on the command
218 language and on other aspects of the GNU linker.
221 @ifclear SingleFormat
222 This version of @command{ld} uses the general purpose BFD libraries
223 to operate on object files. This allows @command{ld} to read, combine, and
224 write object files in many different formats---for example, COFF or
225 @code{a.out}. Different formats may be linked together to produce any
226 available kind of object file. @xref{BFD}, for more information.
229 Aside from its flexibility, the @sc{gnu} linker is more helpful than other
230 linkers in providing diagnostic information. Many linkers abandon
231 execution immediately upon encountering an error; whenever possible,
232 @command{ld} continues executing, allowing you to identify other errors
233 (or, in some cases, to get an output file in spite of the error).
240 @c man begin DESCRIPTION
242 The @sc{gnu} linker @command{ld} is meant to cover a broad range of situations,
243 and to be as compatible as possible with other linkers. As a result,
244 you have many choices to control its behavior.
250 * Options:: Command Line Options
251 * Environment:: Environment Variables
255 @section Command Line Options
263 The linker supports a plethora of command-line options, but in actual
264 practice few of them are used in any particular context.
265 @cindex standard Unix system
266 For instance, a frequent use of @command{ld} is to link standard Unix
267 object files on a standard, supported Unix system. On such a system, to
268 link a file @code{hello.o}:
271 ld -o @var{output} /lib/crt0.o hello.o -lc
274 This tells @command{ld} to produce a file called @var{output} as the
275 result of linking the file @code{/lib/crt0.o} with @code{hello.o} and
276 the library @code{libc.a}, which will come from the standard search
277 directories. (See the discussion of the @samp{-l} option below.)
279 Some of the command-line options to @command{ld} may be specified at any
280 point in the command line. However, options which refer to files, such
281 as @samp{-l} or @samp{-T}, cause the file to be read at the point at
282 which the option appears in the command line, relative to the object
283 files and other file options. Repeating non-file options with a
284 different argument will either have no further effect, or override prior
285 occurrences (those further to the left on the command line) of that
286 option. Options which may be meaningfully specified more than once are
287 noted in the descriptions below.
290 Non-option arguments are object files or archives which are to be linked
291 together. They may follow, precede, or be mixed in with command-line
292 options, except that an object file argument may not be placed between
293 an option and its argument.
295 Usually the linker is invoked with at least one object file, but you can
296 specify other forms of binary input files using @samp{-l}, @samp{-R},
297 and the script command language. If @emph{no} binary input files at all
298 are specified, the linker does not produce any output, and issues the
299 message @samp{No input files}.
301 If the linker cannot recognize the format of an object file, it will
302 assume that it is a linker script. A script specified in this way
303 augments the main linker script used for the link (either the default
304 linker script or the one specified by using @samp{-T}). This feature
305 permits the linker to link against a file which appears to be an object
306 or an archive, but actually merely defines some symbol values, or uses
307 @code{INPUT} or @code{GROUP} to load other objects. Specifying a
308 script in this way merely augments the main linker script, with the
309 extra commands placed after the main script; use the @samp{-T} option
310 to replace the default linker script entirely, but note the effect of
311 the @code{INSERT} command. @xref{Scripts}.
313 For options whose names are a single letter,
314 option arguments must either follow the option letter without intervening
315 whitespace, or be given as separate arguments immediately following the
316 option that requires them.
318 For options whose names are multiple letters, either one dash or two can
319 precede the option name; for example, @samp{-trace-symbol} and
320 @samp{--trace-symbol} are equivalent. Note---there is one exception to
321 this rule. Multiple letter options that start with a lower case 'o' can
322 only be preceded by two dashes. This is to reduce confusion with the
323 @samp{-o} option. So for example @samp{-omagic} sets the output file
324 name to @samp{magic} whereas @samp{--omagic} sets the NMAGIC flag on the
327 Arguments to multiple-letter options must either be separated from the
328 option name by an equals sign, or be given as separate arguments
329 immediately following the option that requires them. For example,
330 @samp{--trace-symbol foo} and @samp{--trace-symbol=foo} are equivalent.
331 Unique abbreviations of the names of multiple-letter options are
334 Note---if the linker is being invoked indirectly, via a compiler driver
335 (e.g. @samp{gcc}) then all the linker command line options should be
336 prefixed by @samp{-Wl,} (or whatever is appropriate for the particular
337 compiler driver) like this:
340 gcc -Wl,--start-group foo.o bar.o -Wl,--end-group
343 This is important, because otherwise the compiler driver program may
344 silently drop the linker options, resulting in a bad link. Confusion
345 may also arise when passing options that require values through a
346 driver, as the use of a space between option and argument acts as
347 a separator, and causes the driver to pass only the option to the linker
348 and the argument to the compiler. In this case, it is simplest to use
349 the joined forms of both single- and multiple-letter options, such as:
352 gcc foo.o bar.o -Wl,-eENTRY -Wl,-Map=a.map
355 Here is a table of the generic command line switches accepted by the GNU
359 @include at-file.texi
361 @kindex -a @var{keyword}
362 @item -a @var{keyword}
363 This option is supported for HP/UX compatibility. The @var{keyword}
364 argument must be one of the strings @samp{archive}, @samp{shared}, or
365 @samp{default}. @samp{-aarchive} is functionally equivalent to
366 @samp{-Bstatic}, and the other two keywords are functionally equivalent
367 to @samp{-Bdynamic}. This option may be used any number of times.
369 @kindex --audit @var{AUDITLIB}
370 @item --audit @var{AUDITLIB}
371 Adds @var{AUDITLIB} to the @code{DT_AUDIT} entry of the dynamic section.
372 @var{AUDITLIB} is not checked for existence, nor will it use the DT_SONAME
373 specified in the library. If specified multiple times @code{DT_AUDIT}
374 will contain a colon separated list of audit interfaces to use. If the linker
375 finds an object with an audit entry while searching for shared libraries,
376 it will add a corresponding @code{DT_DEPAUDIT} entry in the output file.
377 This option is only meaningful on ELF platforms supporting the rtld-audit
381 @cindex architectures
382 @kindex -A @var{arch}
383 @item -A @var{architecture}
384 @kindex --architecture=@var{arch}
385 @itemx --architecture=@var{architecture}
386 In the current release of @command{ld}, this option is useful only for the
387 Intel 960 family of architectures. In that @command{ld} configuration, the
388 @var{architecture} argument identifies the particular architecture in
389 the 960 family, enabling some safeguards and modifying the
390 archive-library search path. @xref{i960,,@command{ld} and the Intel 960
391 family}, for details.
393 Future releases of @command{ld} may support similar functionality for
394 other architecture families.
397 @ifclear SingleFormat
398 @cindex binary input format
399 @kindex -b @var{format}
400 @kindex --format=@var{format}
403 @item -b @var{input-format}
404 @itemx --format=@var{input-format}
405 @command{ld} may be configured to support more than one kind of object
406 file. If your @command{ld} is configured this way, you can use the
407 @samp{-b} option to specify the binary format for input object files
408 that follow this option on the command line. Even when @command{ld} is
409 configured to support alternative object formats, you don't usually need
410 to specify this, as @command{ld} should be configured to expect as a
411 default input format the most usual format on each machine.
412 @var{input-format} is a text string, the name of a particular format
413 supported by the BFD libraries. (You can list the available binary
414 formats with @samp{objdump -i}.)
417 You may want to use this option if you are linking files with an unusual
418 binary format. You can also use @samp{-b} to switch formats explicitly (when
419 linking object files of different formats), by including
420 @samp{-b @var{input-format}} before each group of object files in a
423 The default format is taken from the environment variable
428 You can also define the input format from a script, using the command
431 see @ref{Format Commands}.
435 @kindex -c @var{MRI-cmdfile}
436 @kindex --mri-script=@var{MRI-cmdfile}
437 @cindex compatibility, MRI
438 @item -c @var{MRI-commandfile}
439 @itemx --mri-script=@var{MRI-commandfile}
440 For compatibility with linkers produced by MRI, @command{ld} accepts script
441 files written in an alternate, restricted command language, described in
443 @ref{MRI,,MRI Compatible Script Files}.
446 the MRI Compatible Script Files section of GNU ld documentation.
448 Introduce MRI script files with
449 the option @samp{-c}; use the @samp{-T} option to run linker
450 scripts written in the general-purpose @command{ld} scripting language.
451 If @var{MRI-cmdfile} does not exist, @command{ld} looks for it in the directories
452 specified by any @samp{-L} options.
454 @cindex common allocation
461 These three options are equivalent; multiple forms are supported for
462 compatibility with other linkers. They assign space to common symbols
463 even if a relocatable output file is specified (with @samp{-r}). The
464 script command @code{FORCE_COMMON_ALLOCATION} has the same effect.
465 @xref{Miscellaneous Commands}.
467 @kindex --depaudit @var{AUDITLIB}
468 @kindex -P @var{AUDITLIB}
469 @item --depaudit @var{AUDITLIB}
470 @itemx -P @var{AUDITLIB}
471 Adds @var{AUDITLIB} to the @code{DT_DEPAUDIT} entry of the dynamic section.
472 @var{AUDITLIB} is not checked for existence, nor will it use the DT_SONAME
473 specified in the library. If specified multiple times @code{DT_DEPAUDIT}
474 will contain a colon separated list of audit interfaces to use. This
475 option is only meaningful on ELF platforms supporting the rtld-audit interface.
476 The -P option is provided for Solaris compatibility.
478 @cindex entry point, from command line
479 @kindex -e @var{entry}
480 @kindex --entry=@var{entry}
482 @itemx --entry=@var{entry}
483 Use @var{entry} as the explicit symbol for beginning execution of your
484 program, rather than the default entry point. If there is no symbol
485 named @var{entry}, the linker will try to parse @var{entry} as a number,
486 and use that as the entry address (the number will be interpreted in
487 base 10; you may use a leading @samp{0x} for base 16, or a leading
488 @samp{0} for base 8). @xref{Entry Point}, for a discussion of defaults
489 and other ways of specifying the entry point.
491 @kindex --exclude-libs
492 @item --exclude-libs @var{lib},@var{lib},...
493 Specifies a list of archive libraries from which symbols should not be automatically
494 exported. The library names may be delimited by commas or colons. Specifying
495 @code{--exclude-libs ALL} excludes symbols in all archive libraries from
496 automatic export. This option is available only for the i386 PE targeted
497 port of the linker and for ELF targeted ports. For i386 PE, symbols
498 explicitly listed in a .def file are still exported, regardless of this
499 option. For ELF targeted ports, symbols affected by this option will
500 be treated as hidden.
502 @kindex --exclude-modules-for-implib
503 @item --exclude-modules-for-implib @var{module},@var{module},...
504 Specifies a list of object files or archive members, from which symbols
505 should not be automatically exported, but which should be copied wholesale
506 into the import library being generated during the link. The module names
507 may be delimited by commas or colons, and must match exactly the filenames
508 used by @command{ld} to open the files; for archive members, this is simply
509 the member name, but for object files the name listed must include and
510 match precisely any path used to specify the input file on the linker's
511 command-line. This option is available only for the i386 PE targeted port
512 of the linker. Symbols explicitly listed in a .def file are still exported,
513 regardless of this option.
515 @cindex dynamic symbol table
517 @kindex --export-dynamic
518 @kindex --no-export-dynamic
520 @itemx --export-dynamic
521 @itemx --no-export-dynamic
522 When creating a dynamically linked executable, using the @option{-E}
523 option or the @option{--export-dynamic} option causes the linker to add
524 all symbols to the dynamic symbol table. The dynamic symbol table is the
525 set of symbols which are visible from dynamic objects at run time.
527 If you do not use either of these options (or use the
528 @option{--no-export-dynamic} option to restore the default behavior), the
529 dynamic symbol table will normally contain only those symbols which are
530 referenced by some dynamic object mentioned in the link.
532 If you use @code{dlopen} to load a dynamic object which needs to refer
533 back to the symbols defined by the program, rather than some other
534 dynamic object, then you will probably need to use this option when
535 linking the program itself.
537 You can also use the dynamic list to control what symbols should
538 be added to the dynamic symbol table if the output format supports it.
539 See the description of @samp{--dynamic-list}.
541 Note that this option is specific to ELF targeted ports. PE targets
542 support a similar function to export all symbols from a DLL or EXE; see
543 the description of @samp{--export-all-symbols} below.
545 @ifclear SingleFormat
546 @cindex big-endian objects
550 Link big-endian objects. This affects the default output format.
552 @cindex little-endian objects
555 Link little-endian objects. This affects the default output format.
558 @kindex -f @var{name}
559 @kindex --auxiliary=@var{name}
561 @itemx --auxiliary=@var{name}
562 When creating an ELF shared object, set the internal DT_AUXILIARY field
563 to the specified name. This tells the dynamic linker that the symbol
564 table of the shared object should be used as an auxiliary filter on the
565 symbol table of the shared object @var{name}.
567 If you later link a program against this filter object, then, when you
568 run the program, the dynamic linker will see the DT_AUXILIARY field. If
569 the dynamic linker resolves any symbols from the filter object, it will
570 first check whether there is a definition in the shared object
571 @var{name}. If there is one, it will be used instead of the definition
572 in the filter object. The shared object @var{name} need not exist.
573 Thus the shared object @var{name} may be used to provide an alternative
574 implementation of certain functions, perhaps for debugging or for
575 machine specific performance.
577 This option may be specified more than once. The DT_AUXILIARY entries
578 will be created in the order in which they appear on the command line.
580 @kindex -F @var{name}
581 @kindex --filter=@var{name}
583 @itemx --filter=@var{name}
584 When creating an ELF shared object, set the internal DT_FILTER field to
585 the specified name. This tells the dynamic linker that the symbol table
586 of the shared object which is being created should be used as a filter
587 on the symbol table of the shared object @var{name}.
589 If you later link a program against this filter object, then, when you
590 run the program, the dynamic linker will see the DT_FILTER field. The
591 dynamic linker will resolve symbols according to the symbol table of the
592 filter object as usual, but it will actually link to the definitions
593 found in the shared object @var{name}. Thus the filter object can be
594 used to select a subset of the symbols provided by the object
597 Some older linkers used the @option{-F} option throughout a compilation
598 toolchain for specifying object-file format for both input and output
600 @ifclear SingleFormat
601 The @sc{gnu} linker uses other mechanisms for this purpose: the
602 @option{-b}, @option{--format}, @option{--oformat} options, the
603 @code{TARGET} command in linker scripts, and the @code{GNUTARGET}
604 environment variable.
606 The @sc{gnu} linker will ignore the @option{-F} option when not
607 creating an ELF shared object.
609 @cindex finalization function
610 @kindex -fini=@var{name}
611 @item -fini=@var{name}
612 When creating an ELF executable or shared object, call NAME when the
613 executable or shared object is unloaded, by setting DT_FINI to the
614 address of the function. By default, the linker uses @code{_fini} as
615 the function to call.
619 Ignored. Provided for compatibility with other tools.
621 @kindex -G @var{value}
622 @kindex --gpsize=@var{value}
625 @itemx --gpsize=@var{value}
626 Set the maximum size of objects to be optimized using the GP register to
627 @var{size}. This is only meaningful for object file formats such as
628 MIPS ELF that support putting large and small objects into different
629 sections. This is ignored for other object file formats.
631 @cindex runtime library name
632 @kindex -h @var{name}
633 @kindex -soname=@var{name}
635 @itemx -soname=@var{name}
636 When creating an ELF shared object, set the internal DT_SONAME field to
637 the specified name. When an executable is linked with a shared object
638 which has a DT_SONAME field, then when the executable is run the dynamic
639 linker will attempt to load the shared object specified by the DT_SONAME
640 field rather than the using the file name given to the linker.
643 @cindex incremental link
645 Perform an incremental link (same as option @samp{-r}).
647 @cindex initialization function
648 @kindex -init=@var{name}
649 @item -init=@var{name}
650 When creating an ELF executable or shared object, call NAME when the
651 executable or shared object is loaded, by setting DT_INIT to the address
652 of the function. By default, the linker uses @code{_init} as the
655 @cindex archive files, from cmd line
656 @kindex -l @var{namespec}
657 @kindex --library=@var{namespec}
658 @item -l @var{namespec}
659 @itemx --library=@var{namespec}
660 Add the archive or object file specified by @var{namespec} to the
661 list of files to link. This option may be used any number of times.
662 If @var{namespec} is of the form @file{:@var{filename}}, @command{ld}
663 will search the library path for a file called @var{filename}, otherwise it
664 will search the library path for a file called @file{lib@var{namespec}.a}.
666 On systems which support shared libraries, @command{ld} may also search for
667 files other than @file{lib@var{namespec}.a}. Specifically, on ELF
668 and SunOS systems, @command{ld} will search a directory for a library
669 called @file{lib@var{namespec}.so} before searching for one called
670 @file{lib@var{namespec}.a}. (By convention, a @code{.so} extension
671 indicates a shared library.) Note that this behavior does not apply
672 to @file{:@var{filename}}, which always specifies a file called
675 The linker will search an archive only once, at the location where it is
676 specified on the command line. If the archive defines a symbol which
677 was undefined in some object which appeared before the archive on the
678 command line, the linker will include the appropriate file(s) from the
679 archive. However, an undefined symbol in an object appearing later on
680 the command line will not cause the linker to search the archive again.
682 See the @option{-(} option for a way to force the linker to search
683 archives multiple times.
685 You may list the same archive multiple times on the command line.
688 This type of archive searching is standard for Unix linkers. However,
689 if you are using @command{ld} on AIX, note that it is different from the
690 behaviour of the AIX linker.
693 @cindex search directory, from cmd line
695 @kindex --library-path=@var{dir}
696 @item -L @var{searchdir}
697 @itemx --library-path=@var{searchdir}
698 Add path @var{searchdir} to the list of paths that @command{ld} will search
699 for archive libraries and @command{ld} control scripts. You may use this
700 option any number of times. The directories are searched in the order
701 in which they are specified on the command line. Directories specified
702 on the command line are searched before the default directories. All
703 @option{-L} options apply to all @option{-l} options, regardless of the
704 order in which the options appear. @option{-L} options do not affect
705 how @command{ld} searches for a linker script unless @option{-T}
708 If @var{searchdir} begins with @code{=}, then the @code{=} will be replaced
709 by the @dfn{sysroot prefix}, a path specified when the linker is configured.
712 The default set of paths searched (without being specified with
713 @samp{-L}) depends on which emulation mode @command{ld} is using, and in
714 some cases also on how it was configured. @xref{Environment}.
717 The paths can also be specified in a link script with the
718 @code{SEARCH_DIR} command. Directories specified this way are searched
719 at the point in which the linker script appears in the command line.
722 @kindex -m @var{emulation}
723 @item -m @var{emulation}
724 Emulate the @var{emulation} linker. You can list the available
725 emulations with the @samp{--verbose} or @samp{-V} options.
727 If the @samp{-m} option is not used, the emulation is taken from the
728 @code{LDEMULATION} environment variable, if that is defined.
730 Otherwise, the default emulation depends upon how the linker was
738 Print a link map to the standard output. A link map provides
739 information about the link, including the following:
743 Where object files are mapped into memory.
745 How common symbols are allocated.
747 All archive members included in the link, with a mention of the symbol
748 which caused the archive member to be brought in.
750 The values assigned to symbols.
752 Note - symbols whose values are computed by an expression which
753 involves a reference to a previous value of the same symbol may not
754 have correct result displayed in the link map. This is because the
755 linker discards intermediate results and only retains the final value
756 of an expression. Under such circumstances the linker will display
757 the final value enclosed by square brackets. Thus for example a
758 linker script containing:
766 will produce the following output in the link map if the @option{-M}
771 [0x0000000c] foo = (foo * 0x4)
772 [0x0000000c] foo = (foo + 0x8)
775 See @ref{Expressions} for more information about expressions in linker
780 @cindex read-only text
785 Turn off page alignment of sections, and disable linking against shared
786 libraries. If the output format supports Unix style magic numbers,
787 mark the output as @code{NMAGIC}.
791 @cindex read/write from cmd line
795 Set the text and data sections to be readable and writable. Also, do
796 not page-align the data segment, and disable linking against shared
797 libraries. If the output format supports Unix style magic numbers,
798 mark the output as @code{OMAGIC}. Note: Although a writable text section
799 is allowed for PE-COFF targets, it does not conform to the format
800 specification published by Microsoft.
805 This option negates most of the effects of the @option{-N} option. It
806 sets the text section to be read-only, and forces the data segment to
807 be page-aligned. Note - this option does not enable linking against
808 shared libraries. Use @option{-Bdynamic} for this.
810 @kindex -o @var{output}
811 @kindex --output=@var{output}
812 @cindex naming the output file
813 @item -o @var{output}
814 @itemx --output=@var{output}
815 Use @var{output} as the name for the program produced by @command{ld}; if this
816 option is not specified, the name @file{a.out} is used by default. The
817 script command @code{OUTPUT} can also specify the output file name.
819 @kindex -O @var{level}
820 @cindex generating optimized output
822 If @var{level} is a numeric values greater than zero @command{ld} optimizes
823 the output. This might take significantly longer and therefore probably
824 should only be enabled for the final binary. At the moment this
825 option only affects ELF shared library generation. Future releases of
826 the linker may make more use of this option. Also currently there is
827 no difference in the linker's behaviour for different non-zero values
828 of this option. Again this may change with future releases.
831 @kindex --emit-relocs
832 @cindex retain relocations in final executable
835 Leave relocation sections and contents in fully linked executables.
836 Post link analysis and optimization tools may need this information in
837 order to perform correct modifications of executables. This results
838 in larger executables.
840 This option is currently only supported on ELF platforms.
842 @kindex --force-dynamic
843 @cindex forcing the creation of dynamic sections
844 @item --force-dynamic
845 Force the output file to have dynamic sections. This option is specific
849 @cindex relocatable output
851 @kindex --relocatable
854 Generate relocatable output---i.e., generate an output file that can in
855 turn serve as input to @command{ld}. This is often called @dfn{partial
856 linking}. As a side effect, in environments that support standard Unix
857 magic numbers, this option also sets the output file's magic number to
859 @c ; see @option{-N}.
860 If this option is not specified, an absolute file is produced. When
861 linking C++ programs, this option @emph{will not} resolve references to
862 constructors; to do that, use @samp{-Ur}.
864 When an input file does not have the same format as the output file,
865 partial linking is only supported if that input file does not contain any
866 relocations. Different output formats can have further restrictions; for
867 example some @code{a.out}-based formats do not support partial linking
868 with input files in other formats at all.
870 This option does the same thing as @samp{-i}.
872 @kindex -R @var{file}
873 @kindex --just-symbols=@var{file}
874 @cindex symbol-only input
875 @item -R @var{filename}
876 @itemx --just-symbols=@var{filename}
877 Read symbol names and their addresses from @var{filename}, but do not
878 relocate it or include it in the output. This allows your output file
879 to refer symbolically to absolute locations of memory defined in other
880 programs. You may use this option more than once.
882 For compatibility with other ELF linkers, if the @option{-R} option is
883 followed by a directory name, rather than a file name, it is treated as
884 the @option{-rpath} option.
888 @cindex strip all symbols
891 Omit all symbol information from the output file.
894 @kindex --strip-debug
895 @cindex strip debugger symbols
898 Omit debugger symbol information (but not all symbols) from the output file.
902 @cindex input files, displaying
905 Print the names of the input files as @command{ld} processes them.
907 @kindex -T @var{script}
908 @kindex --script=@var{script}
910 @item -T @var{scriptfile}
911 @itemx --script=@var{scriptfile}
912 Use @var{scriptfile} as the linker script. This script replaces
913 @command{ld}'s default linker script (rather than adding to it), so
914 @var{commandfile} must specify everything necessary to describe the
915 output file. @xref{Scripts}. If @var{scriptfile} does not exist in
916 the current directory, @code{ld} looks for it in the directories
917 specified by any preceding @samp{-L} options. Multiple @samp{-T}
920 @kindex -dT @var{script}
921 @kindex --default-script=@var{script}
923 @item -dT @var{scriptfile}
924 @itemx --default-script=@var{scriptfile}
925 Use @var{scriptfile} as the default linker script. @xref{Scripts}.
927 This option is similar to the @option{--script} option except that
928 processing of the script is delayed until after the rest of the
929 command line has been processed. This allows options placed after the
930 @option{--default-script} option on the command line to affect the
931 behaviour of the linker script, which can be important when the linker
932 command line cannot be directly controlled by the user. (eg because
933 the command line is being constructed by another tool, such as
936 @kindex -u @var{symbol}
937 @kindex --undefined=@var{symbol}
938 @cindex undefined symbol
939 @item -u @var{symbol}
940 @itemx --undefined=@var{symbol}
941 Force @var{symbol} to be entered in the output file as an undefined
942 symbol. Doing this may, for example, trigger linking of additional
943 modules from standard libraries. @samp{-u} may be repeated with
944 different option arguments to enter additional undefined symbols. This
945 option is equivalent to the @code{EXTERN} linker script command.
950 For anything other than C++ programs, this option is equivalent to
951 @samp{-r}: it generates relocatable output---i.e., an output file that can in
952 turn serve as input to @command{ld}. When linking C++ programs, @samp{-Ur}
953 @emph{does} resolve references to constructors, unlike @samp{-r}.
954 It does not work to use @samp{-Ur} on files that were themselves linked
955 with @samp{-Ur}; once the constructor table has been built, it cannot
956 be added to. Use @samp{-Ur} only for the last partial link, and
957 @samp{-r} for the others.
959 @kindex --unique[=@var{SECTION}]
960 @item --unique[=@var{SECTION}]
961 Creates a separate output section for every input section matching
962 @var{SECTION}, or if the optional wildcard @var{SECTION} argument is
963 missing, for every orphan input section. An orphan section is one not
964 specifically mentioned in a linker script. You may use this option
965 multiple times on the command line; It prevents the normal merging of
966 input sections with the same name, overriding output section assignments
976 Display the version number for @command{ld}. The @option{-V} option also
977 lists the supported emulations.
980 @kindex --discard-all
981 @cindex deleting local symbols
984 Delete all local symbols.
987 @kindex --discard-locals
988 @cindex local symbols, deleting
990 @itemx --discard-locals
991 Delete all temporary local symbols. (These symbols start with
992 system-specific local label prefixes, typically @samp{.L} for ELF systems
993 or @samp{L} for traditional a.out systems.)
995 @kindex -y @var{symbol}
996 @kindex --trace-symbol=@var{symbol}
997 @cindex symbol tracing
998 @item -y @var{symbol}
999 @itemx --trace-symbol=@var{symbol}
1000 Print the name of each linked file in which @var{symbol} appears. This
1001 option may be given any number of times. On many systems it is necessary
1002 to prepend an underscore.
1004 This option is useful when you have an undefined symbol in your link but
1005 don't know where the reference is coming from.
1007 @kindex -Y @var{path}
1009 Add @var{path} to the default library search path. This option exists
1010 for Solaris compatibility.
1012 @kindex -z @var{keyword}
1013 @item -z @var{keyword}
1014 The recognized keywords are:
1018 Combines multiple reloc sections and sorts them to make dynamic symbol
1019 lookup caching possible.
1022 Disallows undefined symbols in object files. Undefined symbols in
1023 shared libraries are still allowed.
1026 Marks the object as requiring executable stack.
1029 This option is only meaningful when building a shared object. It makes
1030 the symbols defined by this shared object available for symbol resolution
1031 of subsequently loaded libraries.
1034 This option is only meaningful when building a shared object.
1035 It marks the object so that its runtime initialization will occur
1036 before the runtime initialization of any other objects brought into
1037 the process at the same time. Similarly the runtime finalization of
1038 the object will occur after the runtime finalization of any other
1042 Marks the object that its symbol table interposes before all symbols
1043 but the primary executable.
1046 When generating an executable or shared library, mark it to tell the
1047 dynamic linker to defer function call resolution to the point when
1048 the function is called (lazy binding), rather than at load time.
1049 Lazy binding is the default.
1052 Marks the object that its filters be processed immediately at
1056 Allows multiple definitions.
1059 Disables multiple reloc sections combining.
1062 Disables production of copy relocs.
1065 Marks the object that the search for dependencies of this object will
1066 ignore any default library search paths.
1069 Marks the object shouldn't be unloaded at runtime.
1072 Marks the object not available to @code{dlopen}.
1075 Marks the object can not be dumped by @code{dldump}.
1078 Marks the object as not requiring executable stack.
1081 Don't create an ELF @code{PT_GNU_RELRO} segment header in the object.
1084 When generating an executable or shared library, mark it to tell the
1085 dynamic linker to resolve all symbols when the program is started, or
1086 when the shared library is linked to using dlopen, instead of
1087 deferring function call resolution to the point when the function is
1091 Marks the object may contain $ORIGIN.
1094 Create an ELF @code{PT_GNU_RELRO} segment header in the object.
1096 @item max-page-size=@var{value}
1097 Set the emulation maximum page size to @var{value}.
1099 @item common-page-size=@var{value}
1100 Set the emulation common page size to @var{value}.
1102 @item stack-size=@var{value}
1103 Specify a stack size for in an ELF @code{PT_GNU_STACK} segment.
1104 Specifying zero will override any default non-zero sized
1105 @code{PT_GNU_STACK} segment creation.
1109 Other keywords are ignored for Solaris compatibility.
1112 @cindex groups of archives
1113 @item -( @var{archives} -)
1114 @itemx --start-group @var{archives} --end-group
1115 The @var{archives} should be a list of archive files. They may be
1116 either explicit file names, or @samp{-l} options.
1118 The specified archives are searched repeatedly until no new undefined
1119 references are created. Normally, an archive is searched only once in
1120 the order that it is specified on the command line. If a symbol in that
1121 archive is needed to resolve an undefined symbol referred to by an
1122 object in an archive that appears later on the command line, the linker
1123 would not be able to resolve that reference. By grouping the archives,
1124 they all be searched repeatedly until all possible references are
1127 Using this option has a significant performance cost. It is best to use
1128 it only when there are unavoidable circular references between two or
1131 @kindex --accept-unknown-input-arch
1132 @kindex --no-accept-unknown-input-arch
1133 @item --accept-unknown-input-arch
1134 @itemx --no-accept-unknown-input-arch
1135 Tells the linker to accept input files whose architecture cannot be
1136 recognised. The assumption is that the user knows what they are doing
1137 and deliberately wants to link in these unknown input files. This was
1138 the default behaviour of the linker, before release 2.14. The default
1139 behaviour from release 2.14 onwards is to reject such input files, and
1140 so the @samp{--accept-unknown-input-arch} option has been added to
1141 restore the old behaviour.
1144 @kindex --no-as-needed
1146 @itemx --no-as-needed
1147 This option affects ELF DT_NEEDED tags for dynamic libraries mentioned
1148 on the command line after the @option{--as-needed} option. Normally
1149 the linker will add a DT_NEEDED tag for each dynamic library mentioned
1150 on the command line, regardless of whether the library is actually
1151 needed or not. @option{--as-needed} causes a DT_NEEDED tag to only be
1152 emitted for a library that @emph{at that point in the link} satisfies a
1153 non-weak undefined symbol reference from a regular object file or, if
1154 the library is not found in the DT_NEEDED lists of other libraries, a
1155 non-weak undefined symbol reference from another dynamic library.
1156 Object files or libraries appearing on the command line @emph{after}
1157 the library in question do not affect whether the library is seen as
1158 needed. This is similar to the rules for extraction of object files
1159 from archives. @option{--no-as-needed} restores the default behaviour.
1161 @kindex --add-needed
1162 @kindex --no-add-needed
1164 @itemx --no-add-needed
1165 These two options have been deprecated because of the similarity of
1166 their names to the @option{--as-needed} and @option{--no-as-needed}
1167 options. They have been replaced by @option{--copy-dt-needed-entries}
1168 and @option{--no-copy-dt-needed-entries}.
1170 @kindex -assert @var{keyword}
1171 @item -assert @var{keyword}
1172 This option is ignored for SunOS compatibility.
1176 @kindex -call_shared
1180 Link against dynamic libraries. This is only meaningful on platforms
1181 for which shared libraries are supported. This option is normally the
1182 default on such platforms. The different variants of this option are
1183 for compatibility with various systems. You may use this option
1184 multiple times on the command line: it affects library searching for
1185 @option{-l} options which follow it.
1189 Set the @code{DF_1_GROUP} flag in the @code{DT_FLAGS_1} entry in the dynamic
1190 section. This causes the runtime linker to handle lookups in this
1191 object and its dependencies to be performed only inside the group.
1192 @option{--unresolved-symbols=report-all} is implied. This option is
1193 only meaningful on ELF platforms which support shared libraries.
1203 Do not link against shared libraries. This is only meaningful on
1204 platforms for which shared libraries are supported. The different
1205 variants of this option are for compatibility with various systems. You
1206 may use this option multiple times on the command line: it affects
1207 library searching for @option{-l} options which follow it. This
1208 option also implies @option{--unresolved-symbols=report-all}. This
1209 option can be used with @option{-shared}. Doing so means that a
1210 shared library is being created but that all of the library's external
1211 references must be resolved by pulling in entries from static
1216 When creating a shared library, bind references to global symbols to the
1217 definition within the shared library, if any. Normally, it is possible
1218 for a program linked against a shared library to override the definition
1219 within the shared library. This option is only meaningful on ELF
1220 platforms which support shared libraries.
1222 @kindex -Bsymbolic-functions
1223 @item -Bsymbolic-functions
1224 When creating a shared library, bind references to global function
1225 symbols to the definition within the shared library, if any.
1226 This option is only meaningful on ELF platforms which support shared
1229 @kindex --dynamic-list=@var{dynamic-list-file}
1230 @item --dynamic-list=@var{dynamic-list-file}
1231 Specify the name of a dynamic list file to the linker. This is
1232 typically used when creating shared libraries to specify a list of
1233 global symbols whose references shouldn't be bound to the definition
1234 within the shared library, or creating dynamically linked executables
1235 to specify a list of symbols which should be added to the symbol table
1236 in the executable. This option is only meaningful on ELF platforms
1237 which support shared libraries.
1239 The format of the dynamic list is the same as the version node without
1240 scope and node name. See @ref{VERSION} for more information.
1242 @kindex --dynamic-list-data
1243 @item --dynamic-list-data
1244 Include all global data symbols to the dynamic list.
1246 @kindex --dynamic-list-cpp-new
1247 @item --dynamic-list-cpp-new
1248 Provide the builtin dynamic list for C++ operator new and delete. It
1249 is mainly useful for building shared libstdc++.
1251 @kindex --dynamic-list-cpp-typeinfo
1252 @item --dynamic-list-cpp-typeinfo
1253 Provide the builtin dynamic list for C++ runtime type identification.
1255 @kindex --check-sections
1256 @kindex --no-check-sections
1257 @item --check-sections
1258 @itemx --no-check-sections
1259 Asks the linker @emph{not} to check section addresses after they have
1260 been assigned to see if there are any overlaps. Normally the linker will
1261 perform this check, and if it finds any overlaps it will produce
1262 suitable error messages. The linker does know about, and does make
1263 allowances for sections in overlays. The default behaviour can be
1264 restored by using the command line switch @option{--check-sections}.
1265 Section overlap is not usually checked for relocatable links. You can
1266 force checking in that case by using the @option{--check-sections}
1269 @kindex --copy-dt-needed-entries
1270 @kindex --no-copy-dt-needed-entries
1271 @item --copy-dt-needed-entries
1272 @itemx --no-copy-dt-needed-entries
1273 This option affects the treatment of dynamic libraries referred to
1274 by DT_NEEDED tags @emph{inside} ELF dynamic libraries mentioned on the
1275 command line. Normally the linker won't add a DT_NEEDED tag to the
1276 output binary for each library mentioned in a DT_NEEDED tag in an
1277 input dynamic library. With @option{--copy-dt-needed-entries}
1278 specified on the command line however any dynamic libraries that
1279 follow it will have their DT_NEEDED entries added. The default
1280 behaviour can be restored with @option{--no-copy-dt-needed-entries}.
1282 This option also has an effect on the resolution of symbols in dynamic
1283 libraries. With @option{--copy-dt-needed-entries} dynamic libraries
1284 mentioned on the command line will be recursively searched, following
1285 their DT_NEEDED tags to other libraries, in order to resolve symbols
1286 required by the output binary. With the default setting however
1287 the searching of dynamic libraries that follow it will stop with the
1288 dynamic library itself. No DT_NEEDED links will be traversed to resolve
1291 @cindex cross reference table
1294 Output a cross reference table. If a linker map file is being
1295 generated, the cross reference table is printed to the map file.
1296 Otherwise, it is printed on the standard output.
1298 The format of the table is intentionally simple, so that it may be
1299 easily processed by a script if necessary. The symbols are printed out,
1300 sorted by name. For each symbol, a list of file names is given. If the
1301 symbol is defined, the first file listed is the location of the
1302 definition. If the symbol is defined as a common value then any files
1303 where this happens appear next. Finally any files that reference the
1306 @cindex common allocation
1307 @kindex --no-define-common
1308 @item --no-define-common
1309 This option inhibits the assignment of addresses to common symbols.
1310 The script command @code{INHIBIT_COMMON_ALLOCATION} has the same effect.
1311 @xref{Miscellaneous Commands}.
1313 The @samp{--no-define-common} option allows decoupling
1314 the decision to assign addresses to Common symbols from the choice
1315 of the output file type; otherwise a non-Relocatable output type
1316 forces assigning addresses to Common symbols.
1317 Using @samp{--no-define-common} allows Common symbols that are referenced
1318 from a shared library to be assigned addresses only in the main program.
1319 This eliminates the unused duplicate space in the shared library,
1320 and also prevents any possible confusion over resolving to the wrong
1321 duplicate when there are many dynamic modules with specialized search
1322 paths for runtime symbol resolution.
1324 @cindex symbols, from command line
1325 @kindex --defsym=@var{symbol}=@var{exp}
1326 @item --defsym=@var{symbol}=@var{expression}
1327 Create a global symbol in the output file, containing the absolute
1328 address given by @var{expression}. You may use this option as many
1329 times as necessary to define multiple symbols in the command line. A
1330 limited form of arithmetic is supported for the @var{expression} in this
1331 context: you may give a hexadecimal constant or the name of an existing
1332 symbol, or use @code{+} and @code{-} to add or subtract hexadecimal
1333 constants or symbols. If you need more elaborate expressions, consider
1334 using the linker command language from a script (@pxref{Assignments,,
1335 Assignment: Symbol Definitions}). @emph{Note:} there should be no white
1336 space between @var{symbol}, the equals sign (``@key{=}''), and
1339 @cindex demangling, from command line
1340 @kindex --demangle[=@var{style}]
1341 @kindex --no-demangle
1342 @item --demangle[=@var{style}]
1343 @itemx --no-demangle
1344 These options control whether to demangle symbol names in error messages
1345 and other output. When the linker is told to demangle, it tries to
1346 present symbol names in a readable fashion: it strips leading
1347 underscores if they are used by the object file format, and converts C++
1348 mangled symbol names into user readable names. Different compilers have
1349 different mangling styles. The optional demangling style argument can be used
1350 to choose an appropriate demangling style for your compiler. The linker will
1351 demangle by default unless the environment variable @samp{COLLECT_NO_DEMANGLE}
1352 is set. These options may be used to override the default.
1354 @cindex dynamic linker, from command line
1355 @kindex -I@var{file}
1356 @kindex --dynamic-linker=@var{file}
1358 @itemx --dynamic-linker=@var{file}
1359 Set the name of the dynamic linker. This is only meaningful when
1360 generating dynamically linked ELF executables. The default dynamic
1361 linker is normally correct; don't use this unless you know what you are
1364 @kindex --fatal-warnings
1365 @kindex --no-fatal-warnings
1366 @item --fatal-warnings
1367 @itemx --no-fatal-warnings
1368 Treat all warnings as errors. The default behaviour can be restored
1369 with the option @option{--no-fatal-warnings}.
1371 @kindex --force-exe-suffix
1372 @item --force-exe-suffix
1373 Make sure that an output file has a .exe suffix.
1375 If a successfully built fully linked output file does not have a
1376 @code{.exe} or @code{.dll} suffix, this option forces the linker to copy
1377 the output file to one of the same name with a @code{.exe} suffix. This
1378 option is useful when using unmodified Unix makefiles on a Microsoft
1379 Windows host, since some versions of Windows won't run an image unless
1380 it ends in a @code{.exe} suffix.
1382 @kindex --gc-sections
1383 @kindex --no-gc-sections
1384 @cindex garbage collection
1386 @itemx --no-gc-sections
1387 Enable garbage collection of unused input sections. It is ignored on
1388 targets that do not support this option. The default behaviour (of not
1389 performing this garbage collection) can be restored by specifying
1390 @samp{--no-gc-sections} on the command line.
1392 @samp{--gc-sections} decides which input sections are used by
1393 examining symbols and relocations. The section containing the entry
1394 symbol and all sections containing symbols undefined on the
1395 command-line will be kept, as will sections containing symbols
1396 referenced by dynamic objects. Note that when building shared
1397 libraries, the linker must assume that any visible symbol is
1398 referenced. Once this initial set of sections has been determined,
1399 the linker recursively marks as used any section referenced by their
1400 relocations. See @samp{--entry} and @samp{--undefined}.
1402 This option can be set when doing a partial link (enabled with option
1403 @samp{-r}). In this case the root of symbols kept must be explicitly
1404 specified either by an @samp{--entry} or @samp{--undefined} option or by
1405 a @code{ENTRY} command in the linker script.
1407 @kindex --print-gc-sections
1408 @kindex --no-print-gc-sections
1409 @cindex garbage collection
1410 @item --print-gc-sections
1411 @itemx --no-print-gc-sections
1412 List all sections removed by garbage collection. The listing is
1413 printed on stderr. This option is only effective if garbage
1414 collection has been enabled via the @samp{--gc-sections}) option. The
1415 default behaviour (of not listing the sections that are removed) can
1416 be restored by specifying @samp{--no-print-gc-sections} on the command
1419 @kindex --print-output-format
1420 @cindex output format
1421 @item --print-output-format
1422 Print the name of the default output format (perhaps influenced by
1423 other command-line options). This is the string that would appear
1424 in an @code{OUTPUT_FORMAT} linker script command (@pxref{File Commands}).
1430 Print a summary of the command-line options on the standard output and exit.
1432 @kindex --target-help
1434 Print a summary of all target specific options on the standard output and exit.
1436 @kindex -Map=@var{mapfile}
1437 @item -Map=@var{mapfile}
1438 Print a link map to the file @var{mapfile}. See the description of the
1439 @option{-M} option, above.
1441 @cindex memory usage
1442 @kindex --no-keep-memory
1443 @item --no-keep-memory
1444 @command{ld} normally optimizes for speed over memory usage by caching the
1445 symbol tables of input files in memory. This option tells @command{ld} to
1446 instead optimize for memory usage, by rereading the symbol tables as
1447 necessary. This may be required if @command{ld} runs out of memory space
1448 while linking a large executable.
1450 @kindex --no-undefined
1452 @item --no-undefined
1454 Report unresolved symbol references from regular object files. This
1455 is done even if the linker is creating a non-symbolic shared library.
1456 The switch @option{--[no-]allow-shlib-undefined} controls the
1457 behaviour for reporting unresolved references found in shared
1458 libraries being linked in.
1460 @kindex --allow-multiple-definition
1462 @item --allow-multiple-definition
1464 Normally when a symbol is defined multiple times, the linker will
1465 report a fatal error. These options allow multiple definitions and the
1466 first definition will be used.
1468 @kindex --allow-shlib-undefined
1469 @kindex --no-allow-shlib-undefined
1470 @item --allow-shlib-undefined
1471 @itemx --no-allow-shlib-undefined
1472 Allows or disallows undefined symbols in shared libraries.
1473 This switch is similar to @option{--no-undefined} except that it
1474 determines the behaviour when the undefined symbols are in a
1475 shared library rather than a regular object file. It does not affect
1476 how undefined symbols in regular object files are handled.
1478 The default behaviour is to report errors for any undefined symbols
1479 referenced in shared libraries if the linker is being used to create
1480 an executable, but to allow them if the linker is being used to create
1483 The reasons for allowing undefined symbol references in shared
1484 libraries specified at link time are that:
1488 A shared library specified at link time may not be the same as the one
1489 that is available at load time, so the symbol might actually be
1490 resolvable at load time.
1492 There are some operating systems, eg BeOS and HPPA, where undefined
1493 symbols in shared libraries are normal.
1495 The BeOS kernel for example patches shared libraries at load time to
1496 select whichever function is most appropriate for the current
1497 architecture. This is used, for example, to dynamically select an
1498 appropriate memset function.
1501 @kindex --no-undefined-version
1502 @item --no-undefined-version
1503 Normally when a symbol has an undefined version, the linker will ignore
1504 it. This option disallows symbols with undefined version and a fatal error
1505 will be issued instead.
1507 @kindex --default-symver
1508 @item --default-symver
1509 Create and use a default symbol version (the soname) for unversioned
1512 @kindex --default-imported-symver
1513 @item --default-imported-symver
1514 Create and use a default symbol version (the soname) for unversioned
1517 @kindex --no-warn-mismatch
1518 @item --no-warn-mismatch
1519 Normally @command{ld} will give an error if you try to link together input
1520 files that are mismatched for some reason, perhaps because they have
1521 been compiled for different processors or for different endiannesses.
1522 This option tells @command{ld} that it should silently permit such possible
1523 errors. This option should only be used with care, in cases when you
1524 have taken some special action that ensures that the linker errors are
1527 @kindex --no-warn-search-mismatch
1528 @item --no-warn-search-mismatch
1529 Normally @command{ld} will give a warning if it finds an incompatible
1530 library during a library search. This option silences the warning.
1532 @kindex --no-whole-archive
1533 @item --no-whole-archive
1534 Turn off the effect of the @option{--whole-archive} option for subsequent
1537 @cindex output file after errors
1538 @kindex --noinhibit-exec
1539 @item --noinhibit-exec
1540 Retain the executable output file whenever it is still usable.
1541 Normally, the linker will not produce an output file if it encounters
1542 errors during the link process; it exits without writing an output file
1543 when it issues any error whatsoever.
1547 Only search library directories explicitly specified on the
1548 command line. Library directories specified in linker scripts
1549 (including linker scripts specified on the command line) are ignored.
1551 @ifclear SingleFormat
1552 @kindex --oformat=@var{output-format}
1553 @item --oformat=@var{output-format}
1554 @command{ld} may be configured to support more than one kind of object
1555 file. If your @command{ld} is configured this way, you can use the
1556 @samp{--oformat} option to specify the binary format for the output
1557 object file. Even when @command{ld} is configured to support alternative
1558 object formats, you don't usually need to specify this, as @command{ld}
1559 should be configured to produce as a default output format the most
1560 usual format on each machine. @var{output-format} is a text string, the
1561 name of a particular format supported by the BFD libraries. (You can
1562 list the available binary formats with @samp{objdump -i}.) The script
1563 command @code{OUTPUT_FORMAT} can also specify the output format, but
1564 this option overrides it. @xref{BFD}.
1568 @kindex --pic-executable
1570 @itemx --pic-executable
1571 @cindex position independent executables
1572 Create a position independent executable. This is currently only supported on
1573 ELF platforms. Position independent executables are similar to shared
1574 libraries in that they are relocated by the dynamic linker to the virtual
1575 address the OS chooses for them (which can vary between invocations). Like
1576 normal dynamically linked executables they can be executed and symbols
1577 defined in the executable cannot be overridden by shared libraries.
1581 This option is ignored for Linux compatibility.
1585 This option is ignored for SVR4 compatibility.
1588 @cindex synthesizing linker
1589 @cindex relaxing addressing modes
1593 An option with machine dependent effects.
1595 This option is only supported on a few targets.
1598 @xref{H8/300,,@command{ld} and the H8/300}.
1601 @xref{i960,, @command{ld} and the Intel 960 family}.
1604 @xref{Xtensa,, @command{ld} and Xtensa Processors}.
1607 @xref{M68HC11/68HC12,,@command{ld} and the 68HC11 and 68HC12}.
1610 @xref{Nios II,,@command{ld} and the Altera Nios II}.
1613 @xref{PowerPC ELF32,,@command{ld} and PowerPC 32-bit ELF Support}.
1616 On some platforms the @samp{--relax} option performs target specific,
1617 global optimizations that become possible when the linker resolves
1618 addressing in the program, such as relaxing address modes,
1619 synthesizing new instructions, selecting shorter version of current
1620 instructions, and combining constant values.
1622 On some platforms these link time global optimizations may make symbolic
1623 debugging of the resulting executable impossible.
1625 This is known to be the case for the Matsushita MN10200 and MN10300
1626 family of processors.
1630 On platforms where this is not supported, @samp{--relax} is accepted,
1634 On platforms where @samp{--relax} is accepted the option
1635 @samp{--no-relax} can be used to disable the feature.
1637 @cindex retaining specified symbols
1638 @cindex stripping all but some symbols
1639 @cindex symbols, retaining selectively
1640 @kindex --retain-symbols-file=@var{filename}
1641 @item --retain-symbols-file=@var{filename}
1642 Retain @emph{only} the symbols listed in the file @var{filename},
1643 discarding all others. @var{filename} is simply a flat file, with one
1644 symbol name per line. This option is especially useful in environments
1648 where a large global symbol table is accumulated gradually, to conserve
1651 @samp{--retain-symbols-file} does @emph{not} discard undefined symbols,
1652 or symbols needed for relocations.
1654 You may only specify @samp{--retain-symbols-file} once in the command
1655 line. It overrides @samp{-s} and @samp{-S}.
1658 @item -rpath=@var{dir}
1659 @cindex runtime library search path
1660 @kindex -rpath=@var{dir}
1661 Add a directory to the runtime library search path. This is used when
1662 linking an ELF executable with shared objects. All @option{-rpath}
1663 arguments are concatenated and passed to the runtime linker, which uses
1664 them to locate shared objects at runtime. The @option{-rpath} option is
1665 also used when locating shared objects which are needed by shared
1666 objects explicitly included in the link; see the description of the
1667 @option{-rpath-link} option. If @option{-rpath} is not used when linking an
1668 ELF executable, the contents of the environment variable
1669 @code{LD_RUN_PATH} will be used if it is defined.
1671 The @option{-rpath} option may also be used on SunOS. By default, on
1672 SunOS, the linker will form a runtime search patch out of all the
1673 @option{-L} options it is given. If a @option{-rpath} option is used, the
1674 runtime search path will be formed exclusively using the @option{-rpath}
1675 options, ignoring the @option{-L} options. This can be useful when using
1676 gcc, which adds many @option{-L} options which may be on NFS mounted
1679 For compatibility with other ELF linkers, if the @option{-R} option is
1680 followed by a directory name, rather than a file name, it is treated as
1681 the @option{-rpath} option.
1685 @cindex link-time runtime library search path
1686 @kindex -rpath-link=@var{dir}
1687 @item -rpath-link=@var{dir}
1688 When using ELF or SunOS, one shared library may require another. This
1689 happens when an @code{ld -shared} link includes a shared library as one
1692 When the linker encounters such a dependency when doing a non-shared,
1693 non-relocatable link, it will automatically try to locate the required
1694 shared library and include it in the link, if it is not included
1695 explicitly. In such a case, the @option{-rpath-link} option
1696 specifies the first set of directories to search. The
1697 @option{-rpath-link} option may specify a sequence of directory names
1698 either by specifying a list of names separated by colons, or by
1699 appearing multiple times.
1701 This option should be used with caution as it overrides the search path
1702 that may have been hard compiled into a shared library. In such a case it
1703 is possible to use unintentionally a different search path than the
1704 runtime linker would do.
1706 The linker uses the following search paths to locate required shared
1710 Any directories specified by @option{-rpath-link} options.
1712 Any directories specified by @option{-rpath} options. The difference
1713 between @option{-rpath} and @option{-rpath-link} is that directories
1714 specified by @option{-rpath} options are included in the executable and
1715 used at runtime, whereas the @option{-rpath-link} option is only effective
1716 at link time. Searching @option{-rpath} in this way is only supported
1717 by native linkers and cross linkers which have been configured with
1718 the @option{--with-sysroot} option.
1720 On an ELF system, for native linkers, if the @option{-rpath} and
1721 @option{-rpath-link} options were not used, search the contents of the
1722 environment variable @code{LD_RUN_PATH}.
1724 On SunOS, if the @option{-rpath} option was not used, search any
1725 directories specified using @option{-L} options.
1727 For a native linker, search the contents of the environment
1728 variable @code{LD_LIBRARY_PATH}.
1730 For a native ELF linker, the directories in @code{DT_RUNPATH} or
1731 @code{DT_RPATH} of a shared library are searched for shared
1732 libraries needed by it. The @code{DT_RPATH} entries are ignored if
1733 @code{DT_RUNPATH} entries exist.
1735 The default directories, normally @file{/lib} and @file{/usr/lib}.
1737 For a native linker on an ELF system, if the file @file{/etc/ld.so.conf}
1738 exists, the list of directories found in that file.
1741 If the required shared library is not found, the linker will issue a
1742 warning and continue with the link.
1749 @cindex shared libraries
1750 Create a shared library. This is currently only supported on ELF, XCOFF
1751 and SunOS platforms. On SunOS, the linker will automatically create a
1752 shared library if the @option{-e} option is not used and there are
1753 undefined symbols in the link.
1755 @kindex --sort-common
1757 @itemx --sort-common=ascending
1758 @itemx --sort-common=descending
1759 This option tells @command{ld} to sort the common symbols by alignment in
1760 ascending or descending order when it places them in the appropriate output
1761 sections. The symbol alignments considered are sixteen-byte or larger,
1762 eight-byte, four-byte, two-byte, and one-byte. This is to prevent gaps
1763 between symbols due to alignment constraints. If no sorting order is
1764 specified, then descending order is assumed.
1766 @kindex --sort-section=name
1767 @item --sort-section=name
1768 This option will apply @code{SORT_BY_NAME} to all wildcard section
1769 patterns in the linker script.
1771 @kindex --sort-section=alignment
1772 @item --sort-section=alignment
1773 This option will apply @code{SORT_BY_ALIGNMENT} to all wildcard section
1774 patterns in the linker script.
1776 @kindex --split-by-file
1777 @item --split-by-file[=@var{size}]
1778 Similar to @option{--split-by-reloc} but creates a new output section for
1779 each input file when @var{size} is reached. @var{size} defaults to a
1780 size of 1 if not given.
1782 @kindex --split-by-reloc
1783 @item --split-by-reloc[=@var{count}]
1784 Tries to creates extra sections in the output file so that no single
1785 output section in the file contains more than @var{count} relocations.
1786 This is useful when generating huge relocatable files for downloading into
1787 certain real time kernels with the COFF object file format; since COFF
1788 cannot represent more than 65535 relocations in a single section. Note
1789 that this will fail to work with object file formats which do not
1790 support arbitrary sections. The linker will not split up individual
1791 input sections for redistribution, so if a single input section contains
1792 more than @var{count} relocations one output section will contain that
1793 many relocations. @var{count} defaults to a value of 32768.
1797 Compute and display statistics about the operation of the linker, such
1798 as execution time and memory usage.
1800 @kindex --sysroot=@var{directory}
1801 @item --sysroot=@var{directory}
1802 Use @var{directory} as the location of the sysroot, overriding the
1803 configure-time default. This option is only supported by linkers
1804 that were configured using @option{--with-sysroot}.
1806 @kindex --traditional-format
1807 @cindex traditional format
1808 @item --traditional-format
1809 For some targets, the output of @command{ld} is different in some ways from
1810 the output of some existing linker. This switch requests @command{ld} to
1811 use the traditional format instead.
1814 For example, on SunOS, @command{ld} combines duplicate entries in the
1815 symbol string table. This can reduce the size of an output file with
1816 full debugging information by over 30 percent. Unfortunately, the SunOS
1817 @code{dbx} program can not read the resulting program (@code{gdb} has no
1818 trouble). The @samp{--traditional-format} switch tells @command{ld} to not
1819 combine duplicate entries.
1821 @kindex --section-start=@var{sectionname}=@var{org}
1822 @item --section-start=@var{sectionname}=@var{org}
1823 Locate a section in the output file at the absolute
1824 address given by @var{org}. You may use this option as many
1825 times as necessary to locate multiple sections in the command
1827 @var{org} must be a single hexadecimal integer;
1828 for compatibility with other linkers, you may omit the leading
1829 @samp{0x} usually associated with hexadecimal values. @emph{Note:} there
1830 should be no white space between @var{sectionname}, the equals
1831 sign (``@key{=}''), and @var{org}.
1833 @kindex -Tbss=@var{org}
1834 @kindex -Tdata=@var{org}
1835 @kindex -Ttext=@var{org}
1836 @cindex segment origins, cmd line
1837 @item -Tbss=@var{org}
1838 @itemx -Tdata=@var{org}
1839 @itemx -Ttext=@var{org}
1840 Same as @option{--section-start}, with @code{.bss}, @code{.data} or
1841 @code{.text} as the @var{sectionname}.
1843 @kindex -Ttext-segment=@var{org}
1844 @item -Ttext-segment=@var{org}
1845 @cindex text segment origin, cmd line
1846 When creating an ELF executable, it will set the address of the first
1847 byte of the text segment.
1849 @kindex -Trodata-segment=@var{org}
1850 @item -Trodata-segment=@var{org}
1851 @cindex rodata segment origin, cmd line
1852 When creating an ELF executable or shared object for a target where
1853 the read-only data is in its own segment separate from the executable
1854 text, it will set the address of the first byte of the read-only data segment.
1856 @kindex -Tldata-segment=@var{org}
1857 @item -Tldata-segment=@var{org}
1858 @cindex ldata segment origin, cmd line
1859 When creating an ELF executable or shared object for x86-64 medium memory
1860 model, it will set the address of the first byte of the ldata segment.
1862 @kindex --unresolved-symbols
1863 @item --unresolved-symbols=@var{method}
1864 Determine how to handle unresolved symbols. There are four possible
1865 values for @samp{method}:
1869 Do not report any unresolved symbols.
1872 Report all unresolved symbols. This is the default.
1874 @item ignore-in-object-files
1875 Report unresolved symbols that are contained in shared libraries, but
1876 ignore them if they come from regular object files.
1878 @item ignore-in-shared-libs
1879 Report unresolved symbols that come from regular object files, but
1880 ignore them if they come from shared libraries. This can be useful
1881 when creating a dynamic binary and it is known that all the shared
1882 libraries that it should be referencing are included on the linker's
1886 The behaviour for shared libraries on their own can also be controlled
1887 by the @option{--[no-]allow-shlib-undefined} option.
1889 Normally the linker will generate an error message for each reported
1890 unresolved symbol but the option @option{--warn-unresolved-symbols}
1891 can change this to a warning.
1893 @kindex --verbose[=@var{NUMBER}]
1894 @cindex verbose[=@var{NUMBER}]
1896 @itemx --verbose[=@var{NUMBER}]
1897 Display the version number for @command{ld} and list the linker emulations
1898 supported. Display which input files can and cannot be opened. Display
1899 the linker script being used by the linker. If the optional @var{NUMBER}
1900 argument > 1, plugin symbol status will also be displayed.
1902 @kindex --version-script=@var{version-scriptfile}
1903 @cindex version script, symbol versions
1904 @item --version-script=@var{version-scriptfile}
1905 Specify the name of a version script to the linker. This is typically
1906 used when creating shared libraries to specify additional information
1907 about the version hierarchy for the library being created. This option
1908 is only fully supported on ELF platforms which support shared libraries;
1909 see @ref{VERSION}. It is partially supported on PE platforms, which can
1910 use version scripts to filter symbol visibility in auto-export mode: any
1911 symbols marked @samp{local} in the version script will not be exported.
1914 @kindex --warn-common
1915 @cindex warnings, on combining symbols
1916 @cindex combining symbols, warnings on
1918 Warn when a common symbol is combined with another common symbol or with
1919 a symbol definition. Unix linkers allow this somewhat sloppy practice,
1920 but linkers on some other operating systems do not. This option allows
1921 you to find potential problems from combining global symbols.
1922 Unfortunately, some C libraries use this practice, so you may get some
1923 warnings about symbols in the libraries as well as in your programs.
1925 There are three kinds of global symbols, illustrated here by C examples:
1929 A definition, which goes in the initialized data section of the output
1933 An undefined reference, which does not allocate space.
1934 There must be either a definition or a common symbol for the
1938 A common symbol. If there are only (one or more) common symbols for a
1939 variable, it goes in the uninitialized data area of the output file.
1940 The linker merges multiple common symbols for the same variable into a
1941 single symbol. If they are of different sizes, it picks the largest
1942 size. The linker turns a common symbol into a declaration, if there is
1943 a definition of the same variable.
1946 The @samp{--warn-common} option can produce five kinds of warnings.
1947 Each warning consists of a pair of lines: the first describes the symbol
1948 just encountered, and the second describes the previous symbol
1949 encountered with the same name. One or both of the two symbols will be
1954 Turning a common symbol into a reference, because there is already a
1955 definition for the symbol.
1957 @var{file}(@var{section}): warning: common of `@var{symbol}'
1958 overridden by definition
1959 @var{file}(@var{section}): warning: defined here
1963 Turning a common symbol into a reference, because a later definition for
1964 the symbol is encountered. This is the same as the previous case,
1965 except that the symbols are encountered in a different order.
1967 @var{file}(@var{section}): warning: definition of `@var{symbol}'
1969 @var{file}(@var{section}): warning: common is here
1973 Merging a common symbol with a previous same-sized common symbol.
1975 @var{file}(@var{section}): warning: multiple common
1977 @var{file}(@var{section}): warning: previous common is here
1981 Merging a common symbol with a previous larger common symbol.
1983 @var{file}(@var{section}): warning: common of `@var{symbol}'
1984 overridden by larger common
1985 @var{file}(@var{section}): warning: larger common is here
1989 Merging a common symbol with a previous smaller common symbol. This is
1990 the same as the previous case, except that the symbols are
1991 encountered in a different order.
1993 @var{file}(@var{section}): warning: common of `@var{symbol}'
1994 overriding smaller common
1995 @var{file}(@var{section}): warning: smaller common is here
1999 @kindex --warn-constructors
2000 @item --warn-constructors
2001 Warn if any global constructors are used. This is only useful for a few
2002 object file formats. For formats like COFF or ELF, the linker can not
2003 detect the use of global constructors.
2005 @kindex --warn-multiple-gp
2006 @item --warn-multiple-gp
2007 Warn if multiple global pointer values are required in the output file.
2008 This is only meaningful for certain processors, such as the Alpha.
2009 Specifically, some processors put large-valued constants in a special
2010 section. A special register (the global pointer) points into the middle
2011 of this section, so that constants can be loaded efficiently via a
2012 base-register relative addressing mode. Since the offset in
2013 base-register relative mode is fixed and relatively small (e.g., 16
2014 bits), this limits the maximum size of the constant pool. Thus, in
2015 large programs, it is often necessary to use multiple global pointer
2016 values in order to be able to address all possible constants. This
2017 option causes a warning to be issued whenever this case occurs.
2020 @cindex warnings, on undefined symbols
2021 @cindex undefined symbols, warnings on
2023 Only warn once for each undefined symbol, rather than once per module
2026 @kindex --warn-section-align
2027 @cindex warnings, on section alignment
2028 @cindex section alignment, warnings on
2029 @item --warn-section-align
2030 Warn if the address of an output section is changed because of
2031 alignment. Typically, the alignment will be set by an input section.
2032 The address will only be changed if it not explicitly specified; that
2033 is, if the @code{SECTIONS} command does not specify a start address for
2034 the section (@pxref{SECTIONS}).
2036 @kindex --warn-shared-textrel
2037 @item --warn-shared-textrel
2038 Warn if the linker adds a DT_TEXTREL to a shared object.
2040 @kindex --warn-alternate-em
2041 @item --warn-alternate-em
2042 Warn if an object has alternate ELF machine code.
2044 @kindex --warn-unresolved-symbols
2045 @item --warn-unresolved-symbols
2046 If the linker is going to report an unresolved symbol (see the option
2047 @option{--unresolved-symbols}) it will normally generate an error.
2048 This option makes it generate a warning instead.
2050 @kindex --error-unresolved-symbols
2051 @item --error-unresolved-symbols
2052 This restores the linker's default behaviour of generating errors when
2053 it is reporting unresolved symbols.
2055 @kindex --whole-archive
2056 @cindex including an entire archive
2057 @item --whole-archive
2058 For each archive mentioned on the command line after the
2059 @option{--whole-archive} option, include every object file in the archive
2060 in the link, rather than searching the archive for the required object
2061 files. This is normally used to turn an archive file into a shared
2062 library, forcing every object to be included in the resulting shared
2063 library. This option may be used more than once.
2065 Two notes when using this option from gcc: First, gcc doesn't know
2066 about this option, so you have to use @option{-Wl,-whole-archive}.
2067 Second, don't forget to use @option{-Wl,-no-whole-archive} after your
2068 list of archives, because gcc will add its own list of archives to
2069 your link and you may not want this flag to affect those as well.
2071 @kindex --wrap=@var{symbol}
2072 @item --wrap=@var{symbol}
2073 Use a wrapper function for @var{symbol}. Any undefined reference to
2074 @var{symbol} will be resolved to @code{__wrap_@var{symbol}}. Any
2075 undefined reference to @code{__real_@var{symbol}} will be resolved to
2078 This can be used to provide a wrapper for a system function. The
2079 wrapper function should be called @code{__wrap_@var{symbol}}. If it
2080 wishes to call the system function, it should call
2081 @code{__real_@var{symbol}}.
2083 Here is a trivial example:
2087 __wrap_malloc (size_t c)
2089 printf ("malloc called with %zu\n", c);
2090 return __real_malloc (c);
2094 If you link other code with this file using @option{--wrap malloc}, then
2095 all calls to @code{malloc} will call the function @code{__wrap_malloc}
2096 instead. The call to @code{__real_malloc} in @code{__wrap_malloc} will
2097 call the real @code{malloc} function.
2099 You may wish to provide a @code{__real_malloc} function as well, so that
2100 links without the @option{--wrap} option will succeed. If you do this,
2101 you should not put the definition of @code{__real_malloc} in the same
2102 file as @code{__wrap_malloc}; if you do, the assembler may resolve the
2103 call before the linker has a chance to wrap it to @code{malloc}.
2105 @kindex --eh-frame-hdr
2106 @item --eh-frame-hdr
2107 Request creation of @code{.eh_frame_hdr} section and ELF
2108 @code{PT_GNU_EH_FRAME} segment header.
2110 @kindex --ld-generated-unwind-info
2111 @item --no-ld-generated-unwind-info
2112 Request creation of @code{.eh_frame} unwind info for linker
2113 generated code sections like PLT. This option is on by default
2114 if linker generated unwind info is supported.
2116 @kindex --enable-new-dtags
2117 @kindex --disable-new-dtags
2118 @item --enable-new-dtags
2119 @itemx --disable-new-dtags
2120 This linker can create the new dynamic tags in ELF. But the older ELF
2121 systems may not understand them. If you specify
2122 @option{--enable-new-dtags}, the new dynamic tags will be created as needed
2123 and older dynamic tags will be omitted.
2124 If you specify @option{--disable-new-dtags}, no new dynamic tags will be
2125 created. By default, the new dynamic tags are not created. Note that
2126 those options are only available for ELF systems.
2128 @kindex --hash-size=@var{number}
2129 @item --hash-size=@var{number}
2130 Set the default size of the linker's hash tables to a prime number
2131 close to @var{number}. Increasing this value can reduce the length of
2132 time it takes the linker to perform its tasks, at the expense of
2133 increasing the linker's memory requirements. Similarly reducing this
2134 value can reduce the memory requirements at the expense of speed.
2136 @kindex --hash-style=@var{style}
2137 @item --hash-style=@var{style}
2138 Set the type of linker's hash table(s). @var{style} can be either
2139 @code{sysv} for classic ELF @code{.hash} section, @code{gnu} for
2140 new style GNU @code{.gnu.hash} section or @code{both} for both
2141 the classic ELF @code{.hash} and new style GNU @code{.gnu.hash}
2142 hash tables. The default is @code{sysv}.
2144 @kindex --reduce-memory-overheads
2145 @item --reduce-memory-overheads
2146 This option reduces memory requirements at ld runtime, at the expense of
2147 linking speed. This was introduced to select the old O(n^2) algorithm
2148 for link map file generation, rather than the new O(n) algorithm which uses
2149 about 40% more memory for symbol storage.
2151 Another effect of the switch is to set the default hash table size to
2152 1021, which again saves memory at the cost of lengthening the linker's
2153 run time. This is not done however if the @option{--hash-size} switch
2156 The @option{--reduce-memory-overheads} switch may be also be used to
2157 enable other tradeoffs in future versions of the linker.
2160 @kindex --build-id=@var{style}
2162 @itemx --build-id=@var{style}
2163 Request creation of @code{.note.gnu.build-id} ELF note section.
2164 The contents of the note are unique bits identifying this linked
2165 file. @var{style} can be @code{uuid} to use 128 random bits,
2166 @code{sha1} to use a 160-bit @sc{SHA1} hash on the normative
2167 parts of the output contents, @code{md5} to use a 128-bit
2168 @sc{MD5} hash on the normative parts of the output contents, or
2169 @code{0x@var{hexstring}} to use a chosen bit string specified as
2170 an even number of hexadecimal digits (@code{-} and @code{:}
2171 characters between digit pairs are ignored). If @var{style} is
2172 omitted, @code{sha1} is used.
2174 The @code{md5} and @code{sha1} styles produces an identifier
2175 that is always the same in an identical output file, but will be
2176 unique among all nonidentical output files. It is not intended
2177 to be compared as a checksum for the file's contents. A linked
2178 file may be changed later by other tools, but the build ID bit
2179 string identifying the original linked file does not change.
2181 Passing @code{none} for @var{style} disables the setting from any
2182 @code{--build-id} options earlier on the command line.
2187 @subsection Options Specific to i386 PE Targets
2189 @c man begin OPTIONS
2191 The i386 PE linker supports the @option{-shared} option, which causes
2192 the output to be a dynamically linked library (DLL) instead of a
2193 normal executable. You should name the output @code{*.dll} when you
2194 use this option. In addition, the linker fully supports the standard
2195 @code{*.def} files, which may be specified on the linker command line
2196 like an object file (in fact, it should precede archives it exports
2197 symbols from, to ensure that they get linked in, just like a normal
2200 In addition to the options common to all targets, the i386 PE linker
2201 support additional command line options that are specific to the i386
2202 PE target. Options that take values may be separated from their
2203 values by either a space or an equals sign.
2207 @kindex --add-stdcall-alias
2208 @item --add-stdcall-alias
2209 If given, symbols with a stdcall suffix (@@@var{nn}) will be exported
2210 as-is and also with the suffix stripped.
2211 [This option is specific to the i386 PE targeted port of the linker]
2214 @item --base-file @var{file}
2215 Use @var{file} as the name of a file in which to save the base
2216 addresses of all the relocations needed for generating DLLs with
2218 [This is an i386 PE specific option]
2222 Create a DLL instead of a regular executable. You may also use
2223 @option{-shared} or specify a @code{LIBRARY} in a given @code{.def}
2225 [This option is specific to the i386 PE targeted port of the linker]
2227 @kindex --enable-long-section-names
2228 @kindex --disable-long-section-names
2229 @item --enable-long-section-names
2230 @itemx --disable-long-section-names
2231 The PE variants of the Coff object format add an extension that permits
2232 the use of section names longer than eight characters, the normal limit
2233 for Coff. By default, these names are only allowed in object files, as
2234 fully-linked executable images do not carry the Coff string table required
2235 to support the longer names. As a GNU extension, it is possible to
2236 allow their use in executable images as well, or to (probably pointlessly!)
2237 disallow it in object files, by using these two options. Executable images
2238 generated with these long section names are slightly non-standard, carrying
2239 as they do a string table, and may generate confusing output when examined
2240 with non-GNU PE-aware tools, such as file viewers and dumpers. However,
2241 GDB relies on the use of PE long section names to find Dwarf-2 debug
2242 information sections in an executable image at runtime, and so if neither
2243 option is specified on the command-line, @command{ld} will enable long
2244 section names, overriding the default and technically correct behaviour,
2245 when it finds the presence of debug information while linking an executable
2246 image and not stripping symbols.
2247 [This option is valid for all PE targeted ports of the linker]
2249 @kindex --enable-stdcall-fixup
2250 @kindex --disable-stdcall-fixup
2251 @item --enable-stdcall-fixup
2252 @itemx --disable-stdcall-fixup
2253 If the link finds a symbol that it cannot resolve, it will attempt to
2254 do ``fuzzy linking'' by looking for another defined symbol that differs
2255 only in the format of the symbol name (cdecl vs stdcall) and will
2256 resolve that symbol by linking to the match. For example, the
2257 undefined symbol @code{_foo} might be linked to the function
2258 @code{_foo@@12}, or the undefined symbol @code{_bar@@16} might be linked
2259 to the function @code{_bar}. When the linker does this, it prints a
2260 warning, since it normally should have failed to link, but sometimes
2261 import libraries generated from third-party dlls may need this feature
2262 to be usable. If you specify @option{--enable-stdcall-fixup}, this
2263 feature is fully enabled and warnings are not printed. If you specify
2264 @option{--disable-stdcall-fixup}, this feature is disabled and such
2265 mismatches are considered to be errors.
2266 [This option is specific to the i386 PE targeted port of the linker]
2268 @kindex --leading-underscore
2269 @kindex --no-leading-underscore
2270 @item --leading-underscore
2271 @itemx --no-leading-underscore
2272 For most targets default symbol-prefix is an underscore and is defined
2273 in target's description. By this option it is possible to
2274 disable/enable the default underscore symbol-prefix.
2276 @cindex DLLs, creating
2277 @kindex --export-all-symbols
2278 @item --export-all-symbols
2279 If given, all global symbols in the objects used to build a DLL will
2280 be exported by the DLL. Note that this is the default if there
2281 otherwise wouldn't be any exported symbols. When symbols are
2282 explicitly exported via DEF files or implicitly exported via function
2283 attributes, the default is to not export anything else unless this
2284 option is given. Note that the symbols @code{DllMain@@12},
2285 @code{DllEntryPoint@@0}, @code{DllMainCRTStartup@@12}, and
2286 @code{impure_ptr} will not be automatically
2287 exported. Also, symbols imported from other DLLs will not be
2288 re-exported, nor will symbols specifying the DLL's internal layout
2289 such as those beginning with @code{_head_} or ending with
2290 @code{_iname}. In addition, no symbols from @code{libgcc},
2291 @code{libstd++}, @code{libmingw32}, or @code{crtX.o} will be exported.
2292 Symbols whose names begin with @code{__rtti_} or @code{__builtin_} will
2293 not be exported, to help with C++ DLLs. Finally, there is an
2294 extensive list of cygwin-private symbols that are not exported
2295 (obviously, this applies on when building DLLs for cygwin targets).
2296 These cygwin-excludes are: @code{_cygwin_dll_entry@@12},
2297 @code{_cygwin_crt0_common@@8}, @code{_cygwin_noncygwin_dll_entry@@12},
2298 @code{_fmode}, @code{_impure_ptr}, @code{cygwin_attach_dll},
2299 @code{cygwin_premain0}, @code{cygwin_premain1}, @code{cygwin_premain2},
2300 @code{cygwin_premain3}, and @code{environ}.
2301 [This option is specific to the i386 PE targeted port of the linker]
2303 @kindex --exclude-symbols
2304 @item --exclude-symbols @var{symbol},@var{symbol},...
2305 Specifies a list of symbols which should not be automatically
2306 exported. The symbol names may be delimited by commas or colons.
2307 [This option is specific to the i386 PE targeted port of the linker]
2309 @kindex --exclude-all-symbols
2310 @item --exclude-all-symbols
2311 Specifies no symbols should be automatically exported.
2312 [This option is specific to the i386 PE targeted port of the linker]
2314 @kindex --file-alignment
2315 @item --file-alignment
2316 Specify the file alignment. Sections in the file will always begin at
2317 file offsets which are multiples of this number. This defaults to
2319 [This option is specific to the i386 PE targeted port of the linker]
2323 @item --heap @var{reserve}
2324 @itemx --heap @var{reserve},@var{commit}
2325 Specify the number of bytes of memory to reserve (and optionally commit)
2326 to be used as heap for this program. The default is 1MB reserved, 4K
2328 [This option is specific to the i386 PE targeted port of the linker]
2331 @kindex --image-base
2332 @item --image-base @var{value}
2333 Use @var{value} as the base address of your program or dll. This is
2334 the lowest memory location that will be used when your program or dll
2335 is loaded. To reduce the need to relocate and improve performance of
2336 your dlls, each should have a unique base address and not overlap any
2337 other dlls. The default is 0x400000 for executables, and 0x10000000
2339 [This option is specific to the i386 PE targeted port of the linker]
2343 If given, the stdcall suffixes (@@@var{nn}) will be stripped from
2344 symbols before they are exported.
2345 [This option is specific to the i386 PE targeted port of the linker]
2347 @kindex --large-address-aware
2348 @item --large-address-aware
2349 If given, the appropriate bit in the ``Characteristics'' field of the COFF
2350 header is set to indicate that this executable supports virtual addresses
2351 greater than 2 gigabytes. This should be used in conjunction with the /3GB
2352 or /USERVA=@var{value} megabytes switch in the ``[operating systems]''
2353 section of the BOOT.INI. Otherwise, this bit has no effect.
2354 [This option is specific to PE targeted ports of the linker]
2356 @kindex --disable-large-address-aware
2357 @item --disable-large-address-aware
2358 Reverts the effect of a previous @samp{--large-address-aware} option.
2359 This is useful if @samp{--large-address-aware} is always set by the compiler
2360 driver (e.g. Cygwin gcc) and the executable does not support virtual
2361 addresses greater than 2 gigabytes.
2362 [This option is specific to PE targeted ports of the linker]
2364 @kindex --major-image-version
2365 @item --major-image-version @var{value}
2366 Sets the major number of the ``image version''. Defaults to 1.
2367 [This option is specific to the i386 PE targeted port of the linker]
2369 @kindex --major-os-version
2370 @item --major-os-version @var{value}
2371 Sets the major number of the ``os version''. Defaults to 4.
2372 [This option is specific to the i386 PE targeted port of the linker]
2374 @kindex --major-subsystem-version
2375 @item --major-subsystem-version @var{value}
2376 Sets the major number of the ``subsystem version''. Defaults to 4.
2377 [This option is specific to the i386 PE targeted port of the linker]
2379 @kindex --minor-image-version
2380 @item --minor-image-version @var{value}
2381 Sets the minor number of the ``image version''. Defaults to 0.
2382 [This option is specific to the i386 PE targeted port of the linker]
2384 @kindex --minor-os-version
2385 @item --minor-os-version @var{value}
2386 Sets the minor number of the ``os version''. Defaults to 0.
2387 [This option is specific to the i386 PE targeted port of the linker]
2389 @kindex --minor-subsystem-version
2390 @item --minor-subsystem-version @var{value}
2391 Sets the minor number of the ``subsystem version''. Defaults to 0.
2392 [This option is specific to the i386 PE targeted port of the linker]
2394 @cindex DEF files, creating
2395 @cindex DLLs, creating
2396 @kindex --output-def
2397 @item --output-def @var{file}
2398 The linker will create the file @var{file} which will contain a DEF
2399 file corresponding to the DLL the linker is generating. This DEF file
2400 (which should be called @code{*.def}) may be used to create an import
2401 library with @code{dlltool} or may be used as a reference to
2402 automatically or implicitly exported symbols.
2403 [This option is specific to the i386 PE targeted port of the linker]
2405 @cindex DLLs, creating
2406 @kindex --out-implib
2407 @item --out-implib @var{file}
2408 The linker will create the file @var{file} which will contain an
2409 import lib corresponding to the DLL the linker is generating. This
2410 import lib (which should be called @code{*.dll.a} or @code{*.a}
2411 may be used to link clients against the generated DLL; this behaviour
2412 makes it possible to skip a separate @code{dlltool} import library
2414 [This option is specific to the i386 PE targeted port of the linker]
2416 @kindex --enable-auto-image-base
2417 @item --enable-auto-image-base
2418 Automatically choose the image base for DLLs, unless one is specified
2419 using the @code{--image-base} argument. By using a hash generated
2420 from the dllname to create unique image bases for each DLL, in-memory
2421 collisions and relocations which can delay program execution are
2423 [This option is specific to the i386 PE targeted port of the linker]
2425 @kindex --disable-auto-image-base
2426 @item --disable-auto-image-base
2427 Do not automatically generate a unique image base. If there is no
2428 user-specified image base (@code{--image-base}) then use the platform
2430 [This option is specific to the i386 PE targeted port of the linker]
2432 @cindex DLLs, linking to
2433 @kindex --dll-search-prefix
2434 @item --dll-search-prefix @var{string}
2435 When linking dynamically to a dll without an import library,
2436 search for @code{<string><basename>.dll} in preference to
2437 @code{lib<basename>.dll}. This behaviour allows easy distinction
2438 between DLLs built for the various "subplatforms": native, cygwin,
2439 uwin, pw, etc. For instance, cygwin DLLs typically use
2440 @code{--dll-search-prefix=cyg}.
2441 [This option is specific to the i386 PE targeted port of the linker]
2443 @kindex --enable-auto-import
2444 @item --enable-auto-import
2445 Do sophisticated linking of @code{_symbol} to @code{__imp__symbol} for
2446 DATA imports from DLLs, and create the necessary thunking symbols when
2447 building the import libraries with those DATA exports. Note: Use of the
2448 'auto-import' extension will cause the text section of the image file
2449 to be made writable. This does not conform to the PE-COFF format
2450 specification published by Microsoft.
2452 Note - use of the 'auto-import' extension will also cause read only
2453 data which would normally be placed into the .rdata section to be
2454 placed into the .data section instead. This is in order to work
2455 around a problem with consts that is described here:
2456 http://www.cygwin.com/ml/cygwin/2004-09/msg01101.html
2458 Using 'auto-import' generally will 'just work' -- but sometimes you may
2461 "variable '<var>' can't be auto-imported. Please read the
2462 documentation for ld's @code{--enable-auto-import} for details."
2464 This message occurs when some (sub)expression accesses an address
2465 ultimately given by the sum of two constants (Win32 import tables only
2466 allow one). Instances where this may occur include accesses to member
2467 fields of struct variables imported from a DLL, as well as using a
2468 constant index into an array variable imported from a DLL. Any
2469 multiword variable (arrays, structs, long long, etc) may trigger
2470 this error condition. However, regardless of the exact data type
2471 of the offending exported variable, ld will always detect it, issue
2472 the warning, and exit.
2474 There are several ways to address this difficulty, regardless of the
2475 data type of the exported variable:
2477 One way is to use --enable-runtime-pseudo-reloc switch. This leaves the task
2478 of adjusting references in your client code for runtime environment, so
2479 this method works only when runtime environment supports this feature.
2481 A second solution is to force one of the 'constants' to be a variable --
2482 that is, unknown and un-optimizable at compile time. For arrays,
2483 there are two possibilities: a) make the indexee (the array's address)
2484 a variable, or b) make the 'constant' index a variable. Thus:
2487 extern type extern_array[];
2489 @{ volatile type *t=extern_array; t[1] @}
2495 extern type extern_array[];
2497 @{ volatile int t=1; extern_array[t] @}
2500 For structs (and most other multiword data types) the only option
2501 is to make the struct itself (or the long long, or the ...) variable:
2504 extern struct s extern_struct;
2505 extern_struct.field -->
2506 @{ volatile struct s *t=&extern_struct; t->field @}
2512 extern long long extern_ll;
2514 @{ volatile long long * local_ll=&extern_ll; *local_ll @}
2517 A third method of dealing with this difficulty is to abandon
2518 'auto-import' for the offending symbol and mark it with
2519 @code{__declspec(dllimport)}. However, in practice that
2520 requires using compile-time #defines to indicate whether you are
2521 building a DLL, building client code that will link to the DLL, or
2522 merely building/linking to a static library. In making the choice
2523 between the various methods of resolving the 'direct address with
2524 constant offset' problem, you should consider typical real-world usage:
2532 void main(int argc, char **argv)@{
2533 printf("%d\n",arr[1]);
2543 void main(int argc, char **argv)@{
2544 /* This workaround is for win32 and cygwin; do not "optimize" */
2545 volatile int *parr = arr;
2546 printf("%d\n",parr[1]);
2553 /* Note: auto-export is assumed (no __declspec(dllexport)) */
2554 #if (defined(_WIN32) || defined(__CYGWIN__)) && \
2555 !(defined(FOO_BUILD_DLL) || defined(FOO_STATIC))
2556 #define FOO_IMPORT __declspec(dllimport)
2560 extern FOO_IMPORT int arr[];
2563 void main(int argc, char **argv)@{
2564 printf("%d\n",arr[1]);
2568 A fourth way to avoid this problem is to re-code your
2569 library to use a functional interface rather than a data interface
2570 for the offending variables (e.g. set_foo() and get_foo() accessor
2572 [This option is specific to the i386 PE targeted port of the linker]
2574 @kindex --disable-auto-import
2575 @item --disable-auto-import
2576 Do not attempt to do sophisticated linking of @code{_symbol} to
2577 @code{__imp__symbol} for DATA imports from DLLs.
2578 [This option is specific to the i386 PE targeted port of the linker]
2580 @kindex --enable-runtime-pseudo-reloc
2581 @item --enable-runtime-pseudo-reloc
2582 If your code contains expressions described in --enable-auto-import section,
2583 that is, DATA imports from DLL with non-zero offset, this switch will create
2584 a vector of 'runtime pseudo relocations' which can be used by runtime
2585 environment to adjust references to such data in your client code.
2586 [This option is specific to the i386 PE targeted port of the linker]
2588 @kindex --disable-runtime-pseudo-reloc
2589 @item --disable-runtime-pseudo-reloc
2590 Do not create pseudo relocations for non-zero offset DATA imports from
2592 [This option is specific to the i386 PE targeted port of the linker]
2594 @kindex --enable-extra-pe-debug
2595 @item --enable-extra-pe-debug
2596 Show additional debug info related to auto-import symbol thunking.
2597 [This option is specific to the i386 PE targeted port of the linker]
2599 @kindex --section-alignment
2600 @item --section-alignment
2601 Sets the section alignment. Sections in memory will always begin at
2602 addresses which are a multiple of this number. Defaults to 0x1000.
2603 [This option is specific to the i386 PE targeted port of the linker]
2607 @item --stack @var{reserve}
2608 @itemx --stack @var{reserve},@var{commit}
2609 Specify the number of bytes of memory to reserve (and optionally commit)
2610 to be used as stack for this program. The default is 2MB reserved, 4K
2612 [This option is specific to the i386 PE targeted port of the linker]
2615 @item --subsystem @var{which}
2616 @itemx --subsystem @var{which}:@var{major}
2617 @itemx --subsystem @var{which}:@var{major}.@var{minor}
2618 Specifies the subsystem under which your program will execute. The
2619 legal values for @var{which} are @code{native}, @code{windows},
2620 @code{console}, @code{posix}, and @code{xbox}. You may optionally set
2621 the subsystem version also. Numeric values are also accepted for
2623 [This option is specific to the i386 PE targeted port of the linker]
2625 The following options set flags in the @code{DllCharacteristics} field
2626 of the PE file header:
2627 [These options are specific to PE targeted ports of the linker]
2629 @kindex --dynamicbase
2631 The image base address may be relocated using address space layout
2632 randomization (ASLR). This feature was introduced with MS Windows
2633 Vista for i386 PE targets.
2635 @kindex --forceinteg
2637 Code integrity checks are enforced.
2641 The image is compatible with the Data Execution Prevention.
2642 This feature was introduced with MS Windows XP SP2 for i386 PE targets.
2644 @kindex --no-isolation
2645 @item --no-isolation
2646 Although the image understands isolation, do not isolate the image.
2650 The image does not use SEH. No SE handler may be called from
2655 Do not bind this image.
2659 The driver uses the MS Windows Driver Model.
2663 The image is Terminal Server aware.
2665 @kindex --insert-timestamp
2666 @item --insert-timestamp
2667 Insert a real timestamp into the image, rather than the default value
2668 of zero. This will result in a slightly different results with each
2669 invocation, which could be helpful for distributing unique images.
2675 @subsection Options specific to C6X uClinux targets
2677 @c man begin OPTIONS
2679 The C6X uClinux target uses a binary format called DSBT to support shared
2680 libraries. Each shared library in the system needs to have a unique index;
2681 all executables use an index of 0.
2686 @item --dsbt-size @var{size}
2687 This option sets the number of entires in the DSBT of the current executable
2688 or shared library to @var{size}. The default is to create a table with 64
2691 @kindex --dsbt-index
2692 @item --dsbt-index @var{index}
2693 This option sets the DSBT index of the current executable or shared library
2694 to @var{index}. The default is 0, which is appropriate for generating
2695 executables. If a shared library is generated with a DSBT index of 0, the
2696 @code{R_C6000_DSBT_INDEX} relocs are copied into the output file.
2698 @kindex --no-merge-exidx-entries
2699 The @samp{--no-merge-exidx-entries} switch disables the merging of adjacent
2700 exidx entries in frame unwind info.
2708 @subsection Options specific to Motorola 68HC11 and 68HC12 targets
2710 @c man begin OPTIONS
2712 The 68HC11 and 68HC12 linkers support specific options to control the
2713 memory bank switching mapping and trampoline code generation.
2717 @kindex --no-trampoline
2718 @item --no-trampoline
2719 This option disables the generation of trampoline. By default a trampoline
2720 is generated for each far function which is called using a @code{jsr}
2721 instruction (this happens when a pointer to a far function is taken).
2723 @kindex --bank-window
2724 @item --bank-window @var{name}
2725 This option indicates to the linker the name of the memory region in
2726 the @samp{MEMORY} specification that describes the memory bank window.
2727 The definition of such region is then used by the linker to compute
2728 paging and addresses within the memory window.
2736 @subsection Options specific to Motorola 68K target
2738 @c man begin OPTIONS
2740 The following options are supported to control handling of GOT generation
2741 when linking for 68K targets.
2746 @item --got=@var{type}
2747 This option tells the linker which GOT generation scheme to use.
2748 @var{type} should be one of @samp{single}, @samp{negative},
2749 @samp{multigot} or @samp{target}. For more information refer to the
2750 Info entry for @file{ld}.
2758 @subsection Options specific to MIPS targets
2760 @c man begin OPTIONS
2762 The following options are supported to control microMIPS instruction
2763 generation when linking for MIPS targets.
2771 These options control the choice of microMIPS instructions used in code
2772 generated by the linker, such as that in the PLT or lazy binding stubs,
2773 or in relaxation. If @samp{--insn32} is used, then the linker only uses
2774 32-bit instruction encodings. By default or if @samp{--no-insn32} is
2775 used, all instruction encodings are used, including 16-bit ones where
2785 @section Environment Variables
2787 @c man begin ENVIRONMENT
2789 You can change the behaviour of @command{ld} with the environment variables
2790 @ifclear SingleFormat
2793 @code{LDEMULATION} and @code{COLLECT_NO_DEMANGLE}.
2795 @ifclear SingleFormat
2797 @cindex default input format
2798 @code{GNUTARGET} determines the input-file object format if you don't
2799 use @samp{-b} (or its synonym @samp{--format}). Its value should be one
2800 of the BFD names for an input format (@pxref{BFD}). If there is no
2801 @code{GNUTARGET} in the environment, @command{ld} uses the natural format
2802 of the target. If @code{GNUTARGET} is set to @code{default} then BFD
2803 attempts to discover the input format by examining binary input files;
2804 this method often succeeds, but there are potential ambiguities, since
2805 there is no method of ensuring that the magic number used to specify
2806 object-file formats is unique. However, the configuration procedure for
2807 BFD on each system places the conventional format for that system first
2808 in the search-list, so ambiguities are resolved in favor of convention.
2812 @cindex default emulation
2813 @cindex emulation, default
2814 @code{LDEMULATION} determines the default emulation if you don't use the
2815 @samp{-m} option. The emulation can affect various aspects of linker
2816 behaviour, particularly the default linker script. You can list the
2817 available emulations with the @samp{--verbose} or @samp{-V} options. If
2818 the @samp{-m} option is not used, and the @code{LDEMULATION} environment
2819 variable is not defined, the default emulation depends upon how the
2820 linker was configured.
2822 @kindex COLLECT_NO_DEMANGLE
2823 @cindex demangling, default
2824 Normally, the linker will default to demangling symbols. However, if
2825 @code{COLLECT_NO_DEMANGLE} is set in the environment, then it will
2826 default to not demangling symbols. This environment variable is used in
2827 a similar fashion by the @code{gcc} linker wrapper program. The default
2828 may be overridden by the @samp{--demangle} and @samp{--no-demangle}
2835 @chapter Linker Scripts
2838 @cindex linker scripts
2839 @cindex command files
2840 Every link is controlled by a @dfn{linker script}. This script is
2841 written in the linker command language.
2843 The main purpose of the linker script is to describe how the sections in
2844 the input files should be mapped into the output file, and to control
2845 the memory layout of the output file. Most linker scripts do nothing
2846 more than this. However, when necessary, the linker script can also
2847 direct the linker to perform many other operations, using the commands
2850 The linker always uses a linker script. If you do not supply one
2851 yourself, the linker will use a default script that is compiled into the
2852 linker executable. You can use the @samp{--verbose} command line option
2853 to display the default linker script. Certain command line options,
2854 such as @samp{-r} or @samp{-N}, will affect the default linker script.
2856 You may supply your own linker script by using the @samp{-T} command
2857 line option. When you do this, your linker script will replace the
2858 default linker script.
2860 You may also use linker scripts implicitly by naming them as input files
2861 to the linker, as though they were files to be linked. @xref{Implicit
2865 * Basic Script Concepts:: Basic Linker Script Concepts
2866 * Script Format:: Linker Script Format
2867 * Simple Example:: Simple Linker Script Example
2868 * Simple Commands:: Simple Linker Script Commands
2869 * Assignments:: Assigning Values to Symbols
2870 * SECTIONS:: SECTIONS Command
2871 * MEMORY:: MEMORY Command
2872 * PHDRS:: PHDRS Command
2873 * VERSION:: VERSION Command
2874 * Expressions:: Expressions in Linker Scripts
2875 * Implicit Linker Scripts:: Implicit Linker Scripts
2878 @node Basic Script Concepts
2879 @section Basic Linker Script Concepts
2880 @cindex linker script concepts
2881 We need to define some basic concepts and vocabulary in order to
2882 describe the linker script language.
2884 The linker combines input files into a single output file. The output
2885 file and each input file are in a special data format known as an
2886 @dfn{object file format}. Each file is called an @dfn{object file}.
2887 The output file is often called an @dfn{executable}, but for our
2888 purposes we will also call it an object file. Each object file has,
2889 among other things, a list of @dfn{sections}. We sometimes refer to a
2890 section in an input file as an @dfn{input section}; similarly, a section
2891 in the output file is an @dfn{output section}.
2893 Each section in an object file has a name and a size. Most sections
2894 also have an associated block of data, known as the @dfn{section
2895 contents}. A section may be marked as @dfn{loadable}, which means that
2896 the contents should be loaded into memory when the output file is run.
2897 A section with no contents may be @dfn{allocatable}, which means that an
2898 area in memory should be set aside, but nothing in particular should be
2899 loaded there (in some cases this memory must be zeroed out). A section
2900 which is neither loadable nor allocatable typically contains some sort
2901 of debugging information.
2903 Every loadable or allocatable output section has two addresses. The
2904 first is the @dfn{VMA}, or virtual memory address. This is the address
2905 the section will have when the output file is run. The second is the
2906 @dfn{LMA}, or load memory address. This is the address at which the
2907 section will be loaded. In most cases the two addresses will be the
2908 same. An example of when they might be different is when a data section
2909 is loaded into ROM, and then copied into RAM when the program starts up
2910 (this technique is often used to initialize global variables in a ROM
2911 based system). In this case the ROM address would be the LMA, and the
2912 RAM address would be the VMA.
2914 You can see the sections in an object file by using the @code{objdump}
2915 program with the @samp{-h} option.
2917 Every object file also has a list of @dfn{symbols}, known as the
2918 @dfn{symbol table}. A symbol may be defined or undefined. Each symbol
2919 has a name, and each defined symbol has an address, among other
2920 information. If you compile a C or C++ program into an object file, you
2921 will get a defined symbol for every defined function and global or
2922 static variable. Every undefined function or global variable which is
2923 referenced in the input file will become an undefined symbol.
2925 You can see the symbols in an object file by using the @code{nm}
2926 program, or by using the @code{objdump} program with the @samp{-t}
2930 @section Linker Script Format
2931 @cindex linker script format
2932 Linker scripts are text files.
2934 You write a linker script as a series of commands. Each command is
2935 either a keyword, possibly followed by arguments, or an assignment to a
2936 symbol. You may separate commands using semicolons. Whitespace is
2939 Strings such as file or format names can normally be entered directly.
2940 If the file name contains a character such as a comma which would
2941 otherwise serve to separate file names, you may put the file name in
2942 double quotes. There is no way to use a double quote character in a
2945 You may include comments in linker scripts just as in C, delimited by
2946 @samp{/*} and @samp{*/}. As in C, comments are syntactically equivalent
2949 @node Simple Example
2950 @section Simple Linker Script Example
2951 @cindex linker script example
2952 @cindex example of linker script
2953 Many linker scripts are fairly simple.
2955 The simplest possible linker script has just one command:
2956 @samp{SECTIONS}. You use the @samp{SECTIONS} command to describe the
2957 memory layout of the output file.
2959 The @samp{SECTIONS} command is a powerful command. Here we will
2960 describe a simple use of it. Let's assume your program consists only of
2961 code, initialized data, and uninitialized data. These will be in the
2962 @samp{.text}, @samp{.data}, and @samp{.bss} sections, respectively.
2963 Let's assume further that these are the only sections which appear in
2966 For this example, let's say that the code should be loaded at address
2967 0x10000, and that the data should start at address 0x8000000. Here is a
2968 linker script which will do that:
2973 .text : @{ *(.text) @}
2975 .data : @{ *(.data) @}
2976 .bss : @{ *(.bss) @}
2980 You write the @samp{SECTIONS} command as the keyword @samp{SECTIONS},
2981 followed by a series of symbol assignments and output section
2982 descriptions enclosed in curly braces.
2984 The first line inside the @samp{SECTIONS} command of the above example
2985 sets the value of the special symbol @samp{.}, which is the location
2986 counter. If you do not specify the address of an output section in some
2987 other way (other ways are described later), the address is set from the
2988 current value of the location counter. The location counter is then
2989 incremented by the size of the output section. At the start of the
2990 @samp{SECTIONS} command, the location counter has the value @samp{0}.
2992 The second line defines an output section, @samp{.text}. The colon is
2993 required syntax which may be ignored for now. Within the curly braces
2994 after the output section name, you list the names of the input sections
2995 which should be placed into this output section. The @samp{*} is a
2996 wildcard which matches any file name. The expression @samp{*(.text)}
2997 means all @samp{.text} input sections in all input files.
2999 Since the location counter is @samp{0x10000} when the output section
3000 @samp{.text} is defined, the linker will set the address of the
3001 @samp{.text} section in the output file to be @samp{0x10000}.
3003 The remaining lines define the @samp{.data} and @samp{.bss} sections in
3004 the output file. The linker will place the @samp{.data} output section
3005 at address @samp{0x8000000}. After the linker places the @samp{.data}
3006 output section, the value of the location counter will be
3007 @samp{0x8000000} plus the size of the @samp{.data} output section. The
3008 effect is that the linker will place the @samp{.bss} output section
3009 immediately after the @samp{.data} output section in memory.
3011 The linker will ensure that each output section has the required
3012 alignment, by increasing the location counter if necessary. In this
3013 example, the specified addresses for the @samp{.text} and @samp{.data}
3014 sections will probably satisfy any alignment constraints, but the linker
3015 may have to create a small gap between the @samp{.data} and @samp{.bss}
3018 That's it! That's a simple and complete linker script.
3020 @node Simple Commands
3021 @section Simple Linker Script Commands
3022 @cindex linker script simple commands
3023 In this section we describe the simple linker script commands.
3026 * Entry Point:: Setting the entry point
3027 * File Commands:: Commands dealing with files
3028 @ifclear SingleFormat
3029 * Format Commands:: Commands dealing with object file formats
3032 * REGION_ALIAS:: Assign alias names to memory regions
3033 * Miscellaneous Commands:: Other linker script commands
3037 @subsection Setting the Entry Point
3038 @kindex ENTRY(@var{symbol})
3039 @cindex start of execution
3040 @cindex first instruction
3042 The first instruction to execute in a program is called the @dfn{entry
3043 point}. You can use the @code{ENTRY} linker script command to set the
3044 entry point. The argument is a symbol name:
3049 There are several ways to set the entry point. The linker will set the
3050 entry point by trying each of the following methods in order, and
3051 stopping when one of them succeeds:
3054 the @samp{-e} @var{entry} command-line option;
3056 the @code{ENTRY(@var{symbol})} command in a linker script;
3058 the value of a target specific symbol, if it is defined; For many
3059 targets this is @code{start}, but PE and BeOS based systems for example
3060 check a list of possible entry symbols, matching the first one found.
3062 the address of the first byte of the @samp{.text} section, if present;
3064 The address @code{0}.
3068 @subsection Commands Dealing with Files
3069 @cindex linker script file commands
3070 Several linker script commands deal with files.
3073 @item INCLUDE @var{filename}
3074 @kindex INCLUDE @var{filename}
3075 @cindex including a linker script
3076 Include the linker script @var{filename} at this point. The file will
3077 be searched for in the current directory, and in any directory specified
3078 with the @option{-L} option. You can nest calls to @code{INCLUDE} up to
3081 You can place @code{INCLUDE} directives at the top level, in @code{MEMORY} or
3082 @code{SECTIONS} commands, or in output section descriptions.
3084 @item INPUT(@var{file}, @var{file}, @dots{})
3085 @itemx INPUT(@var{file} @var{file} @dots{})
3086 @kindex INPUT(@var{files})
3087 @cindex input files in linker scripts
3088 @cindex input object files in linker scripts
3089 @cindex linker script input object files
3090 The @code{INPUT} command directs the linker to include the named files
3091 in the link, as though they were named on the command line.
3093 For example, if you always want to include @file{subr.o} any time you do
3094 a link, but you can't be bothered to put it on every link command line,
3095 then you can put @samp{INPUT (subr.o)} in your linker script.
3097 In fact, if you like, you can list all of your input files in the linker
3098 script, and then invoke the linker with nothing but a @samp{-T} option.
3100 In case a @dfn{sysroot prefix} is configured, and the filename starts
3101 with the @samp{/} character, and the script being processed was
3102 located inside the @dfn{sysroot prefix}, the filename will be looked
3103 for in the @dfn{sysroot prefix}. Otherwise, the linker will try to
3104 open the file in the current directory. If it is not found, the
3105 linker will search through the archive library search path. See the
3106 description of @samp{-L} in @ref{Options,,Command Line Options}.
3108 If you use @samp{INPUT (-l@var{file})}, @command{ld} will transform the
3109 name to @code{lib@var{file}.a}, as with the command line argument
3112 When you use the @code{INPUT} command in an implicit linker script, the
3113 files will be included in the link at the point at which the linker
3114 script file is included. This can affect archive searching.
3116 @item GROUP(@var{file}, @var{file}, @dots{})
3117 @itemx GROUP(@var{file} @var{file} @dots{})
3118 @kindex GROUP(@var{files})
3119 @cindex grouping input files
3120 The @code{GROUP} command is like @code{INPUT}, except that the named
3121 files should all be archives, and they are searched repeatedly until no
3122 new undefined references are created. See the description of @samp{-(}
3123 in @ref{Options,,Command Line Options}.
3125 @item AS_NEEDED(@var{file}, @var{file}, @dots{})
3126 @itemx AS_NEEDED(@var{file} @var{file} @dots{})
3127 @kindex AS_NEEDED(@var{files})
3128 This construct can appear only inside of the @code{INPUT} or @code{GROUP}
3129 commands, among other filenames. The files listed will be handled
3130 as if they appear directly in the @code{INPUT} or @code{GROUP} commands,
3131 with the exception of ELF shared libraries, that will be added only
3132 when they are actually needed. This construct essentially enables
3133 @option{--as-needed} option for all the files listed inside of it
3134 and restores previous @option{--as-needed} resp. @option{--no-as-needed}
3137 @item OUTPUT(@var{filename})
3138 @kindex OUTPUT(@var{filename})
3139 @cindex output file name in linker script
3140 The @code{OUTPUT} command names the output file. Using
3141 @code{OUTPUT(@var{filename})} in the linker script is exactly like using
3142 @samp{-o @var{filename}} on the command line (@pxref{Options,,Command
3143 Line Options}). If both are used, the command line option takes
3146 You can use the @code{OUTPUT} command to define a default name for the
3147 output file other than the usual default of @file{a.out}.
3149 @item SEARCH_DIR(@var{path})
3150 @kindex SEARCH_DIR(@var{path})
3151 @cindex library search path in linker script
3152 @cindex archive search path in linker script
3153 @cindex search path in linker script
3154 The @code{SEARCH_DIR} command adds @var{path} to the list of paths where
3155 @command{ld} looks for archive libraries. Using
3156 @code{SEARCH_DIR(@var{path})} is exactly like using @samp{-L @var{path}}
3157 on the command line (@pxref{Options,,Command Line Options}). If both
3158 are used, then the linker will search both paths. Paths specified using
3159 the command line option are searched first.
3161 @item STARTUP(@var{filename})
3162 @kindex STARTUP(@var{filename})
3163 @cindex first input file
3164 The @code{STARTUP} command is just like the @code{INPUT} command, except
3165 that @var{filename} will become the first input file to be linked, as
3166 though it were specified first on the command line. This may be useful
3167 when using a system in which the entry point is always the start of the
3171 @ifclear SingleFormat
3172 @node Format Commands
3173 @subsection Commands Dealing with Object File Formats
3174 A couple of linker script commands deal with object file formats.
3177 @item OUTPUT_FORMAT(@var{bfdname})
3178 @itemx OUTPUT_FORMAT(@var{default}, @var{big}, @var{little})
3179 @kindex OUTPUT_FORMAT(@var{bfdname})
3180 @cindex output file format in linker script
3181 The @code{OUTPUT_FORMAT} command names the BFD format to use for the
3182 output file (@pxref{BFD}). Using @code{OUTPUT_FORMAT(@var{bfdname})} is
3183 exactly like using @samp{--oformat @var{bfdname}} on the command line
3184 (@pxref{Options,,Command Line Options}). If both are used, the command
3185 line option takes precedence.
3187 You can use @code{OUTPUT_FORMAT} with three arguments to use different
3188 formats based on the @samp{-EB} and @samp{-EL} command line options.
3189 This permits the linker script to set the output format based on the
3192 If neither @samp{-EB} nor @samp{-EL} are used, then the output format
3193 will be the first argument, @var{default}. If @samp{-EB} is used, the
3194 output format will be the second argument, @var{big}. If @samp{-EL} is
3195 used, the output format will be the third argument, @var{little}.
3197 For example, the default linker script for the MIPS ELF target uses this
3200 OUTPUT_FORMAT(elf32-bigmips, elf32-bigmips, elf32-littlemips)
3202 This says that the default format for the output file is
3203 @samp{elf32-bigmips}, but if the user uses the @samp{-EL} command line
3204 option, the output file will be created in the @samp{elf32-littlemips}
3207 @item TARGET(@var{bfdname})
3208 @kindex TARGET(@var{bfdname})
3209 @cindex input file format in linker script
3210 The @code{TARGET} command names the BFD format to use when reading input
3211 files. It affects subsequent @code{INPUT} and @code{GROUP} commands.
3212 This command is like using @samp{-b @var{bfdname}} on the command line
3213 (@pxref{Options,,Command Line Options}). If the @code{TARGET} command
3214 is used but @code{OUTPUT_FORMAT} is not, then the last @code{TARGET}
3215 command is also used to set the format for the output file. @xref{BFD}.
3220 @subsection Assign alias names to memory regions
3221 @kindex REGION_ALIAS(@var{alias}, @var{region})
3222 @cindex region alias
3223 @cindex region names
3225 Alias names can be added to existing memory regions created with the
3226 @ref{MEMORY} command. Each name corresponds to at most one memory region.
3229 REGION_ALIAS(@var{alias}, @var{region})
3232 The @code{REGION_ALIAS} function creates an alias name @var{alias} for the
3233 memory region @var{region}. This allows a flexible mapping of output sections
3234 to memory regions. An example follows.
3236 Suppose we have an application for embedded systems which come with various
3237 memory storage devices. All have a general purpose, volatile memory @code{RAM}
3238 that allows code execution or data storage. Some may have a read-only,
3239 non-volatile memory @code{ROM} that allows code execution and read-only data
3240 access. The last variant is a read-only, non-volatile memory @code{ROM2} with
3241 read-only data access and no code execution capability. We have four output
3246 @code{.text} program code;
3248 @code{.rodata} read-only data;
3250 @code{.data} read-write initialized data;
3252 @code{.bss} read-write zero initialized data.
3255 The goal is to provide a linker command file that contains a system independent
3256 part defining the output sections and a system dependent part mapping the
3257 output sections to the memory regions available on the system. Our embedded
3258 systems come with three different memory setups @code{A}, @code{B} and
3260 @multitable @columnfractions .25 .25 .25 .25
3261 @item Section @tab Variant A @tab Variant B @tab Variant C
3262 @item .text @tab RAM @tab ROM @tab ROM
3263 @item .rodata @tab RAM @tab ROM @tab ROM2
3264 @item .data @tab RAM @tab RAM/ROM @tab RAM/ROM2
3265 @item .bss @tab RAM @tab RAM @tab RAM
3267 The notation @code{RAM/ROM} or @code{RAM/ROM2} means that this section is
3268 loaded into region @code{ROM} or @code{ROM2} respectively. Please note that
3269 the load address of the @code{.data} section starts in all three variants at
3270 the end of the @code{.rodata} section.
3272 The base linker script that deals with the output sections follows. It
3273 includes the system dependent @code{linkcmds.memory} file that describes the
3276 INCLUDE linkcmds.memory
3289 .data : AT (rodata_end)
3294 data_size = SIZEOF(.data);
3295 data_load_start = LOADADDR(.data);
3303 Now we need three different @code{linkcmds.memory} files to define memory
3304 regions and alias names. The content of @code{linkcmds.memory} for the three
3305 variants @code{A}, @code{B} and @code{C}:
3308 Here everything goes into the @code{RAM}.
3312 RAM : ORIGIN = 0, LENGTH = 4M
3315 REGION_ALIAS("REGION_TEXT", RAM);
3316 REGION_ALIAS("REGION_RODATA", RAM);
3317 REGION_ALIAS("REGION_DATA", RAM);
3318 REGION_ALIAS("REGION_BSS", RAM);
3321 Program code and read-only data go into the @code{ROM}. Read-write data goes
3322 into the @code{RAM}. An image of the initialized data is loaded into the
3323 @code{ROM} and will be copied during system start into the @code{RAM}.
3327 ROM : ORIGIN = 0, LENGTH = 3M
3328 RAM : ORIGIN = 0x10000000, LENGTH = 1M
3331 REGION_ALIAS("REGION_TEXT", ROM);
3332 REGION_ALIAS("REGION_RODATA", ROM);
3333 REGION_ALIAS("REGION_DATA", RAM);
3334 REGION_ALIAS("REGION_BSS", RAM);
3337 Program code goes into the @code{ROM}. Read-only data goes into the
3338 @code{ROM2}. Read-write data goes into the @code{RAM}. An image of the
3339 initialized data is loaded into the @code{ROM2} and will be copied during
3340 system start into the @code{RAM}.
3344 ROM : ORIGIN = 0, LENGTH = 2M
3345 ROM2 : ORIGIN = 0x10000000, LENGTH = 1M
3346 RAM : ORIGIN = 0x20000000, LENGTH = 1M
3349 REGION_ALIAS("REGION_TEXT", ROM);
3350 REGION_ALIAS("REGION_RODATA", ROM2);
3351 REGION_ALIAS("REGION_DATA", RAM);
3352 REGION_ALIAS("REGION_BSS", RAM);
3356 It is possible to write a common system initialization routine to copy the
3357 @code{.data} section from @code{ROM} or @code{ROM2} into the @code{RAM} if
3362 extern char data_start [];
3363 extern char data_size [];
3364 extern char data_load_start [];
3366 void copy_data(void)
3368 if (data_start != data_load_start)
3370 memcpy(data_start, data_load_start, (size_t) data_size);
3375 @node Miscellaneous Commands
3376 @subsection Other Linker Script Commands
3377 There are a few other linker scripts commands.
3380 @item ASSERT(@var{exp}, @var{message})
3382 @cindex assertion in linker script
3383 Ensure that @var{exp} is non-zero. If it is zero, then exit the linker
3384 with an error code, and print @var{message}.
3386 @item EXTERN(@var{symbol} @var{symbol} @dots{})
3388 @cindex undefined symbol in linker script
3389 Force @var{symbol} to be entered in the output file as an undefined
3390 symbol. Doing this may, for example, trigger linking of additional
3391 modules from standard libraries. You may list several @var{symbol}s for
3392 each @code{EXTERN}, and you may use @code{EXTERN} multiple times. This
3393 command has the same effect as the @samp{-u} command-line option.
3395 @item FORCE_COMMON_ALLOCATION
3396 @kindex FORCE_COMMON_ALLOCATION
3397 @cindex common allocation in linker script
3398 This command has the same effect as the @samp{-d} command-line option:
3399 to make @command{ld} assign space to common symbols even if a relocatable
3400 output file is specified (@samp{-r}).
3402 @item INHIBIT_COMMON_ALLOCATION
3403 @kindex INHIBIT_COMMON_ALLOCATION
3404 @cindex common allocation in linker script
3405 This command has the same effect as the @samp{--no-define-common}
3406 command-line option: to make @code{ld} omit the assignment of addresses
3407 to common symbols even for a non-relocatable output file.
3409 @item INSERT [ AFTER | BEFORE ] @var{output_section}
3411 @cindex insert user script into default script
3412 This command is typically used in a script specified by @samp{-T} to
3413 augment the default @code{SECTIONS} with, for example, overlays. It
3414 inserts all prior linker script statements after (or before)
3415 @var{output_section}, and also causes @samp{-T} to not override the
3416 default linker script. The exact insertion point is as for orphan
3417 sections. @xref{Location Counter}. The insertion happens after the
3418 linker has mapped input sections to output sections. Prior to the
3419 insertion, since @samp{-T} scripts are parsed before the default
3420 linker script, statements in the @samp{-T} script occur before the
3421 default linker script statements in the internal linker representation
3422 of the script. In particular, input section assignments will be made
3423 to @samp{-T} output sections before those in the default script. Here
3424 is an example of how a @samp{-T} script using @code{INSERT} might look:
3431 .ov1 @{ ov1*(.text) @}
3432 .ov2 @{ ov2*(.text) @}
3438 @item NOCROSSREFS(@var{section} @var{section} @dots{})
3439 @kindex NOCROSSREFS(@var{sections})
3440 @cindex cross references
3441 This command may be used to tell @command{ld} to issue an error about any
3442 references among certain output sections.
3444 In certain types of programs, particularly on embedded systems when
3445 using overlays, when one section is loaded into memory, another section
3446 will not be. Any direct references between the two sections would be
3447 errors. For example, it would be an error if code in one section called
3448 a function defined in the other section.
3450 The @code{NOCROSSREFS} command takes a list of output section names. If
3451 @command{ld} detects any cross references between the sections, it reports
3452 an error and returns a non-zero exit status. Note that the
3453 @code{NOCROSSREFS} command uses output section names, not input section
3456 @ifclear SingleFormat
3457 @item OUTPUT_ARCH(@var{bfdarch})
3458 @kindex OUTPUT_ARCH(@var{bfdarch})
3459 @cindex machine architecture
3460 @cindex architecture
3461 Specify a particular output machine architecture. The argument is one
3462 of the names used by the BFD library (@pxref{BFD}). You can see the
3463 architecture of an object file by using the @code{objdump} program with
3464 the @samp{-f} option.
3467 @item LD_FEATURE(@var{string})
3468 @kindex LD_FEATURE(@var{string})
3469 This command may be used to modify @command{ld} behavior. If
3470 @var{string} is @code{"SANE_EXPR"} then absolute symbols and numbers
3471 in a script are simply treated as numbers everywhere.
3472 @xref{Expression Section}.
3476 @section Assigning Values to Symbols
3477 @cindex assignment in scripts
3478 @cindex symbol definition, scripts
3479 @cindex variables, defining
3480 You may assign a value to a symbol in a linker script. This will define
3481 the symbol and place it into the symbol table with a global scope.
3484 * Simple Assignments:: Simple Assignments
3487 * PROVIDE_HIDDEN:: PROVIDE_HIDDEN
3488 * Source Code Reference:: How to use a linker script defined symbol in source code
3491 @node Simple Assignments
3492 @subsection Simple Assignments
3494 You may assign to a symbol using any of the C assignment operators:
3497 @item @var{symbol} = @var{expression} ;
3498 @itemx @var{symbol} += @var{expression} ;
3499 @itemx @var{symbol} -= @var{expression} ;
3500 @itemx @var{symbol} *= @var{expression} ;
3501 @itemx @var{symbol} /= @var{expression} ;
3502 @itemx @var{symbol} <<= @var{expression} ;
3503 @itemx @var{symbol} >>= @var{expression} ;
3504 @itemx @var{symbol} &= @var{expression} ;
3505 @itemx @var{symbol} |= @var{expression} ;
3508 The first case will define @var{symbol} to the value of
3509 @var{expression}. In the other cases, @var{symbol} must already be
3510 defined, and the value will be adjusted accordingly.
3512 The special symbol name @samp{.} indicates the location counter. You
3513 may only use this within a @code{SECTIONS} command. @xref{Location Counter}.
3515 The semicolon after @var{expression} is required.
3517 Expressions are defined below; see @ref{Expressions}.
3519 You may write symbol assignments as commands in their own right, or as
3520 statements within a @code{SECTIONS} command, or as part of an output
3521 section description in a @code{SECTIONS} command.
3523 The section of the symbol will be set from the section of the
3524 expression; for more information, see @ref{Expression Section}.
3526 Here is an example showing the three different places that symbol
3527 assignments may be used:
3538 _bdata = (. + 3) & ~ 3;
3539 .data : @{ *(.data) @}
3543 In this example, the symbol @samp{floating_point} will be defined as
3544 zero. The symbol @samp{_etext} will be defined as the address following
3545 the last @samp{.text} input section. The symbol @samp{_bdata} will be
3546 defined as the address following the @samp{.text} output section aligned
3547 upward to a 4 byte boundary.
3552 For ELF targeted ports, define a symbol that will be hidden and won't be
3553 exported. The syntax is @code{HIDDEN(@var{symbol} = @var{expression})}.
3555 Here is the example from @ref{Simple Assignments}, rewritten to use
3559 HIDDEN(floating_point = 0);
3567 HIDDEN(_bdata = (. + 3) & ~ 3);
3568 .data : @{ *(.data) @}
3572 In this case none of the three symbols will be visible outside this module.
3577 In some cases, it is desirable for a linker script to define a symbol
3578 only if it is referenced and is not defined by any object included in
3579 the link. For example, traditional linkers defined the symbol
3580 @samp{etext}. However, ANSI C requires that the user be able to use
3581 @samp{etext} as a function name without encountering an error. The
3582 @code{PROVIDE} keyword may be used to define a symbol, such as
3583 @samp{etext}, only if it is referenced but not defined. The syntax is
3584 @code{PROVIDE(@var{symbol} = @var{expression})}.
3586 Here is an example of using @code{PROVIDE} to define @samp{etext}:
3599 In this example, if the program defines @samp{_etext} (with a leading
3600 underscore), the linker will give a multiple definition error. If, on
3601 the other hand, the program defines @samp{etext} (with no leading
3602 underscore), the linker will silently use the definition in the program.
3603 If the program references @samp{etext} but does not define it, the
3604 linker will use the definition in the linker script.
3606 @node PROVIDE_HIDDEN
3607 @subsection PROVIDE_HIDDEN
3608 @cindex PROVIDE_HIDDEN
3609 Similar to @code{PROVIDE}. For ELF targeted ports, the symbol will be
3610 hidden and won't be exported.
3612 @node Source Code Reference
3613 @subsection Source Code Reference
3615 Accessing a linker script defined variable from source code is not
3616 intuitive. In particular a linker script symbol is not equivalent to
3617 a variable declaration in a high level language, it is instead a
3618 symbol that does not have a value.
3620 Before going further, it is important to note that compilers often
3621 transform names in the source code into different names when they are
3622 stored in the symbol table. For example, Fortran compilers commonly
3623 prepend or append an underscore, and C++ performs extensive @samp{name
3624 mangling}. Therefore there might be a discrepancy between the name
3625 of a variable as it is used in source code and the name of the same
3626 variable as it is defined in a linker script. For example in C a
3627 linker script variable might be referred to as:
3633 But in the linker script it might be defined as:
3639 In the remaining examples however it is assumed that no name
3640 transformation has taken place.
3642 When a symbol is declared in a high level language such as C, two
3643 things happen. The first is that the compiler reserves enough space
3644 in the program's memory to hold the @emph{value} of the symbol. The
3645 second is that the compiler creates an entry in the program's symbol
3646 table which holds the symbol's @emph{address}. ie the symbol table
3647 contains the address of the block of memory holding the symbol's
3648 value. So for example the following C declaration, at file scope:
3654 creates an entry called @samp{foo} in the symbol table. This entry
3655 holds the address of an @samp{int} sized block of memory where the
3656 number 1000 is initially stored.
3658 When a program references a symbol the compiler generates code that
3659 first accesses the symbol table to find the address of the symbol's
3660 memory block and then code to read the value from that memory block.
3667 looks up the symbol @samp{foo} in the symbol table, gets the address
3668 associated with this symbol and then writes the value 1 into that
3675 looks up the symbol @samp{foo} in the symbol table, gets its address
3676 and then copies this address into the block of memory associated with
3677 the variable @samp{a}.
3679 Linker scripts symbol declarations, by contrast, create an entry in
3680 the symbol table but do not assign any memory to them. Thus they are
3681 an address without a value. So for example the linker script definition:
3687 creates an entry in the symbol table called @samp{foo} which holds
3688 the address of memory location 1000, but nothing special is stored at
3689 address 1000. This means that you cannot access the @emph{value} of a
3690 linker script defined symbol - it has no value - all you can do is
3691 access the @emph{address} of a linker script defined symbol.
3693 Hence when you are using a linker script defined symbol in source code
3694 you should always take the address of the symbol, and never attempt to
3695 use its value. For example suppose you want to copy the contents of a
3696 section of memory called .ROM into a section called .FLASH and the
3697 linker script contains these declarations:
3701 start_of_ROM = .ROM;
3702 end_of_ROM = .ROM + sizeof (.ROM) - 1;
3703 start_of_FLASH = .FLASH;
3707 Then the C source code to perform the copy would be:
3711 extern char start_of_ROM, end_of_ROM, start_of_FLASH;
3713 memcpy (& start_of_FLASH, & start_of_ROM, & end_of_ROM - & start_of_ROM);
3717 Note the use of the @samp{&} operators. These are correct.
3720 @section SECTIONS Command
3722 The @code{SECTIONS} command tells the linker how to map input sections
3723 into output sections, and how to place the output sections in memory.
3725 The format of the @code{SECTIONS} command is:
3729 @var{sections-command}
3730 @var{sections-command}
3735 Each @var{sections-command} may of be one of the following:
3739 an @code{ENTRY} command (@pxref{Entry Point,,Entry command})
3741 a symbol assignment (@pxref{Assignments})
3743 an output section description
3745 an overlay description
3748 The @code{ENTRY} command and symbol assignments are permitted inside the
3749 @code{SECTIONS} command for convenience in using the location counter in
3750 those commands. This can also make the linker script easier to
3751 understand because you can use those commands at meaningful points in
3752 the layout of the output file.
3754 Output section descriptions and overlay descriptions are described
3757 If you do not use a @code{SECTIONS} command in your linker script, the
3758 linker will place each input section into an identically named output
3759 section in the order that the sections are first encountered in the
3760 input files. If all input sections are present in the first file, for
3761 example, the order of sections in the output file will match the order
3762 in the first input file. The first section will be at address zero.
3765 * Output Section Description:: Output section description
3766 * Output Section Name:: Output section name
3767 * Output Section Address:: Output section address
3768 * Input Section:: Input section description
3769 * Output Section Data:: Output section data
3770 * Output Section Keywords:: Output section keywords
3771 * Output Section Discarding:: Output section discarding
3772 * Output Section Attributes:: Output section attributes
3773 * Overlay Description:: Overlay description
3776 @node Output Section Description
3777 @subsection Output Section Description
3778 The full description of an output section looks like this:
3781 @var{section} [@var{address}] [(@var{type})] :
3783 [ALIGN(@var{section_align}) | ALIGN_WITH_INPUT]
3784 [SUBALIGN(@var{subsection_align})]
3787 @var{output-section-command}
3788 @var{output-section-command}
3790 @} [>@var{region}] [AT>@var{lma_region}] [:@var{phdr} :@var{phdr} @dots{}] [=@var{fillexp}]
3794 Most output sections do not use most of the optional section attributes.
3796 The whitespace around @var{section} is required, so that the section
3797 name is unambiguous. The colon and the curly braces are also required.
3798 The line breaks and other white space are optional.
3800 Each @var{output-section-command} may be one of the following:
3804 a symbol assignment (@pxref{Assignments})
3806 an input section description (@pxref{Input Section})
3808 data values to include directly (@pxref{Output Section Data})
3810 a special output section keyword (@pxref{Output Section Keywords})
3813 @node Output Section Name
3814 @subsection Output Section Name
3815 @cindex name, section
3816 @cindex section name
3817 The name of the output section is @var{section}. @var{section} must
3818 meet the constraints of your output format. In formats which only
3819 support a limited number of sections, such as @code{a.out}, the name
3820 must be one of the names supported by the format (@code{a.out}, for
3821 example, allows only @samp{.text}, @samp{.data} or @samp{.bss}). If the
3822 output format supports any number of sections, but with numbers and not
3823 names (as is the case for Oasys), the name should be supplied as a
3824 quoted numeric string. A section name may consist of any sequence of
3825 characters, but a name which contains any unusual characters such as
3826 commas must be quoted.
3828 The output section name @samp{/DISCARD/} is special; @ref{Output Section
3831 @node Output Section Address
3832 @subsection Output Section Address
3833 @cindex address, section
3834 @cindex section address
3835 The @var{address} is an expression for the VMA (the virtual memory
3836 address) of the output section. This address is optional, but if it
3837 is provided then the output address will be set exactly as specified.
3839 If the output address is not specified then one will be chosen for the
3840 section, based on the heuristic below. This address will be adjusted
3841 to fit the alignment requirement of the output section. The
3842 alignment requirement is the strictest alignment of any input section
3843 contained within the output section.
3845 The output section address heuristic is as follows:
3849 If an output memory @var{region} is set for the section then it
3850 is added to this region and its address will be the next free address
3854 If the MEMORY command has been used to create a list of memory
3855 regions then the first region which has attributes compatible with the
3856 section is selected to contain it. The section's output address will
3857 be the next free address in that region; @ref{MEMORY}.
3860 If no memory regions were specified, or none match the section then
3861 the output address will be based on the current value of the location
3869 .text . : @{ *(.text) @}
3876 .text : @{ *(.text) @}
3880 are subtly different. The first will set the address of the
3881 @samp{.text} output section to the current value of the location
3882 counter. The second will set it to the current value of the location
3883 counter aligned to the strictest alignment of any of the @samp{.text}
3886 The @var{address} may be an arbitrary expression; @ref{Expressions}.
3887 For example, if you want to align the section on a 0x10 byte boundary,
3888 so that the lowest four bits of the section address are zero, you could
3889 do something like this:
3891 .text ALIGN(0x10) : @{ *(.text) @}
3894 This works because @code{ALIGN} returns the current location counter
3895 aligned upward to the specified value.
3897 Specifying @var{address} for a section will change the value of the
3898 location counter, provided that the section is non-empty. (Empty
3899 sections are ignored).
3902 @subsection Input Section Description
3903 @cindex input sections
3904 @cindex mapping input sections to output sections
3905 The most common output section command is an input section description.
3907 The input section description is the most basic linker script operation.
3908 You use output sections to tell the linker how to lay out your program
3909 in memory. You use input section descriptions to tell the linker how to
3910 map the input files into your memory layout.
3913 * Input Section Basics:: Input section basics
3914 * Input Section Wildcards:: Input section wildcard patterns
3915 * Input Section Common:: Input section for common symbols
3916 * Input Section Keep:: Input section and garbage collection
3917 * Input Section Example:: Input section example
3920 @node Input Section Basics
3921 @subsubsection Input Section Basics
3922 @cindex input section basics
3923 An input section description consists of a file name optionally followed
3924 by a list of section names in parentheses.
3926 The file name and the section name may be wildcard patterns, which we
3927 describe further below (@pxref{Input Section Wildcards}).
3929 The most common input section description is to include all input
3930 sections with a particular name in the output section. For example, to
3931 include all input @samp{.text} sections, you would write:
3936 Here the @samp{*} is a wildcard which matches any file name. To exclude a list
3937 of files from matching the file name wildcard, EXCLUDE_FILE may be used to
3938 match all files except the ones specified in the EXCLUDE_FILE list. For
3941 *(EXCLUDE_FILE (*crtend.o *otherfile.o) .ctors)
3943 will cause all .ctors sections from all files except @file{crtend.o} and
3944 @file{otherfile.o} to be included.
3946 There are two ways to include more than one section:
3952 The difference between these is the order in which the @samp{.text} and
3953 @samp{.rdata} input sections will appear in the output section. In the
3954 first example, they will be intermingled, appearing in the same order as
3955 they are found in the linker input. In the second example, all
3956 @samp{.text} input sections will appear first, followed by all
3957 @samp{.rdata} input sections.
3959 You can specify a file name to include sections from a particular file.
3960 You would do this if one or more of your files contain special data that
3961 needs to be at a particular location in memory. For example:
3966 To refine the sections that are included based on the section flags
3967 of an input section, INPUT_SECTION_FLAGS may be used.
3969 Here is a simple example for using Section header flags for ELF sections:
3974 .text : @{ INPUT_SECTION_FLAGS (SHF_MERGE & SHF_STRINGS) *(.text) @}
3975 .text2 : @{ INPUT_SECTION_FLAGS (!SHF_WRITE) *(.text) @}
3980 In this example, the output section @samp{.text} will be comprised of any
3981 input section matching the name *(.text) whose section header flags
3982 @code{SHF_MERGE} and @code{SHF_STRINGS} are set. The output section
3983 @samp{.text2} will be comprised of any input section matching the name *(.text)
3984 whose section header flag @code{SHF_WRITE} is clear.
3986 You can also specify files within archives by writing a pattern
3987 matching the archive, a colon, then the pattern matching the file,
3988 with no whitespace around the colon.
3992 matches file within archive
3994 matches the whole archive
3996 matches file but not one in an archive
3999 Either one or both of @samp{archive} and @samp{file} can contain shell
4000 wildcards. On DOS based file systems, the linker will assume that a
4001 single letter followed by a colon is a drive specifier, so
4002 @samp{c:myfile.o} is a simple file specification, not @samp{myfile.o}
4003 within an archive called @samp{c}. @samp{archive:file} filespecs may
4004 also be used within an @code{EXCLUDE_FILE} list, but may not appear in
4005 other linker script contexts. For instance, you cannot extract a file
4006 from an archive by using @samp{archive:file} in an @code{INPUT}
4009 If you use a file name without a list of sections, then all sections in
4010 the input file will be included in the output section. This is not
4011 commonly done, but it may by useful on occasion. For example:
4016 When you use a file name which is not an @samp{archive:file} specifier
4017 and does not contain any wild card
4018 characters, the linker will first see if you also specified the file
4019 name on the linker command line or in an @code{INPUT} command. If you
4020 did not, the linker will attempt to open the file as an input file, as
4021 though it appeared on the command line. Note that this differs from an
4022 @code{INPUT} command, because the linker will not search for the file in
4023 the archive search path.
4025 @node Input Section Wildcards
4026 @subsubsection Input Section Wildcard Patterns
4027 @cindex input section wildcards
4028 @cindex wildcard file name patterns
4029 @cindex file name wildcard patterns
4030 @cindex section name wildcard patterns
4031 In an input section description, either the file name or the section
4032 name or both may be wildcard patterns.
4034 The file name of @samp{*} seen in many examples is a simple wildcard
4035 pattern for the file name.
4037 The wildcard patterns are like those used by the Unix shell.
4041 matches any number of characters
4043 matches any single character
4045 matches a single instance of any of the @var{chars}; the @samp{-}
4046 character may be used to specify a range of characters, as in
4047 @samp{[a-z]} to match any lower case letter
4049 quotes the following character
4052 When a file name is matched with a wildcard, the wildcard characters
4053 will not match a @samp{/} character (used to separate directory names on
4054 Unix). A pattern consisting of a single @samp{*} character is an
4055 exception; it will always match any file name, whether it contains a
4056 @samp{/} or not. In a section name, the wildcard characters will match
4057 a @samp{/} character.
4059 File name wildcard patterns only match files which are explicitly
4060 specified on the command line or in an @code{INPUT} command. The linker
4061 does not search directories to expand wildcards.
4063 If a file name matches more than one wildcard pattern, or if a file name
4064 appears explicitly and is also matched by a wildcard pattern, the linker
4065 will use the first match in the linker script. For example, this
4066 sequence of input section descriptions is probably in error, because the
4067 @file{data.o} rule will not be used:
4069 .data : @{ *(.data) @}
4070 .data1 : @{ data.o(.data) @}
4073 @cindex SORT_BY_NAME
4074 Normally, the linker will place files and sections matched by wildcards
4075 in the order in which they are seen during the link. You can change
4076 this by using the @code{SORT_BY_NAME} keyword, which appears before a wildcard
4077 pattern in parentheses (e.g., @code{SORT_BY_NAME(.text*)}). When the
4078 @code{SORT_BY_NAME} keyword is used, the linker will sort the files or sections
4079 into ascending order by name before placing them in the output file.
4081 @cindex SORT_BY_ALIGNMENT
4082 @code{SORT_BY_ALIGNMENT} is very similar to @code{SORT_BY_NAME}. The
4083 difference is @code{SORT_BY_ALIGNMENT} will sort sections into
4084 descending order by alignment before placing them in the output file.
4085 Larger alignments are placed before smaller alignments in order to
4086 reduce the amount of padding necessary.
4088 @cindex SORT_BY_INIT_PRIORITY
4089 @code{SORT_BY_INIT_PRIORITY} is very similar to @code{SORT_BY_NAME}. The
4090 difference is @code{SORT_BY_INIT_PRIORITY} will sort sections into
4091 ascending order by numerical value of the GCC init_priority attribute
4092 encoded in the section name before placing them in the output file.
4095 @code{SORT} is an alias for @code{SORT_BY_NAME}.
4097 When there are nested section sorting commands in linker script, there
4098 can be at most 1 level of nesting for section sorting commands.
4102 @code{SORT_BY_NAME} (@code{SORT_BY_ALIGNMENT} (wildcard section pattern)).
4103 It will sort the input sections by name first, then by alignment if two
4104 sections have the same name.
4106 @code{SORT_BY_ALIGNMENT} (@code{SORT_BY_NAME} (wildcard section pattern)).
4107 It will sort the input sections by alignment first, then by name if two
4108 sections have the same alignment.
4110 @code{SORT_BY_NAME} (@code{SORT_BY_NAME} (wildcard section pattern)) is
4111 treated the same as @code{SORT_BY_NAME} (wildcard section pattern).
4113 @code{SORT_BY_ALIGNMENT} (@code{SORT_BY_ALIGNMENT} (wildcard section pattern))
4114 is treated the same as @code{SORT_BY_ALIGNMENT} (wildcard section pattern).
4116 All other nested section sorting commands are invalid.
4119 When both command line section sorting option and linker script
4120 section sorting command are used, section sorting command always
4121 takes precedence over the command line option.
4123 If the section sorting command in linker script isn't nested, the
4124 command line option will make the section sorting command to be
4125 treated as nested sorting command.
4129 @code{SORT_BY_NAME} (wildcard section pattern ) with
4130 @option{--sort-sections alignment} is equivalent to
4131 @code{SORT_BY_NAME} (@code{SORT_BY_ALIGNMENT} (wildcard section pattern)).
4133 @code{SORT_BY_ALIGNMENT} (wildcard section pattern) with
4134 @option{--sort-section name} is equivalent to
4135 @code{SORT_BY_ALIGNMENT} (@code{SORT_BY_NAME} (wildcard section pattern)).
4138 If the section sorting command in linker script is nested, the
4139 command line option will be ignored.
4142 @code{SORT_NONE} disables section sorting by ignoring the command line
4143 section sorting option.
4145 If you ever get confused about where input sections are going, use the
4146 @samp{-M} linker option to generate a map file. The map file shows
4147 precisely how input sections are mapped to output sections.
4149 This example shows how wildcard patterns might be used to partition
4150 files. This linker script directs the linker to place all @samp{.text}
4151 sections in @samp{.text} and all @samp{.bss} sections in @samp{.bss}.
4152 The linker will place the @samp{.data} section from all files beginning
4153 with an upper case character in @samp{.DATA}; for all other files, the
4154 linker will place the @samp{.data} section in @samp{.data}.
4158 .text : @{ *(.text) @}
4159 .DATA : @{ [A-Z]*(.data) @}
4160 .data : @{ *(.data) @}
4161 .bss : @{ *(.bss) @}
4166 @node Input Section Common
4167 @subsubsection Input Section for Common Symbols
4168 @cindex common symbol placement
4169 @cindex uninitialized data placement
4170 A special notation is needed for common symbols, because in many object
4171 file formats common symbols do not have a particular input section. The
4172 linker treats common symbols as though they are in an input section
4173 named @samp{COMMON}.
4175 You may use file names with the @samp{COMMON} section just as with any
4176 other input sections. You can use this to place common symbols from a
4177 particular input file in one section while common symbols from other
4178 input files are placed in another section.
4180 In most cases, common symbols in input files will be placed in the
4181 @samp{.bss} section in the output file. For example:
4183 .bss @{ *(.bss) *(COMMON) @}
4186 @cindex scommon section
4187 @cindex small common symbols
4188 Some object file formats have more than one type of common symbol. For
4189 example, the MIPS ELF object file format distinguishes standard common
4190 symbols and small common symbols. In this case, the linker will use a
4191 different special section name for other types of common symbols. In
4192 the case of MIPS ELF, the linker uses @samp{COMMON} for standard common
4193 symbols and @samp{.scommon} for small common symbols. This permits you
4194 to map the different types of common symbols into memory at different
4198 You will sometimes see @samp{[COMMON]} in old linker scripts. This
4199 notation is now considered obsolete. It is equivalent to
4202 @node Input Section Keep
4203 @subsubsection Input Section and Garbage Collection
4205 @cindex garbage collection
4206 When link-time garbage collection is in use (@samp{--gc-sections}),
4207 it is often useful to mark sections that should not be eliminated.
4208 This is accomplished by surrounding an input section's wildcard entry
4209 with @code{KEEP()}, as in @code{KEEP(*(.init))} or
4210 @code{KEEP(SORT_BY_NAME(*)(.ctors))}.
4212 @node Input Section Example
4213 @subsubsection Input Section Example
4214 The following example is a complete linker script. It tells the linker
4215 to read all of the sections from file @file{all.o} and place them at the
4216 start of output section @samp{outputa} which starts at location
4217 @samp{0x10000}. All of section @samp{.input1} from file @file{foo.o}
4218 follows immediately, in the same output section. All of section
4219 @samp{.input2} from @file{foo.o} goes into output section
4220 @samp{outputb}, followed by section @samp{.input1} from @file{foo1.o}.
4221 All of the remaining @samp{.input1} and @samp{.input2} sections from any
4222 files are written to output section @samp{outputc}.
4250 @node Output Section Data
4251 @subsection Output Section Data
4253 @cindex section data
4254 @cindex output section data
4255 @kindex BYTE(@var{expression})
4256 @kindex SHORT(@var{expression})
4257 @kindex LONG(@var{expression})
4258 @kindex QUAD(@var{expression})
4259 @kindex SQUAD(@var{expression})
4260 You can include explicit bytes of data in an output section by using
4261 @code{BYTE}, @code{SHORT}, @code{LONG}, @code{QUAD}, or @code{SQUAD} as
4262 an output section command. Each keyword is followed by an expression in
4263 parentheses providing the value to store (@pxref{Expressions}). The
4264 value of the expression is stored at the current value of the location
4267 The @code{BYTE}, @code{SHORT}, @code{LONG}, and @code{QUAD} commands
4268 store one, two, four, and eight bytes (respectively). After storing the
4269 bytes, the location counter is incremented by the number of bytes
4272 For example, this will store the byte 1 followed by the four byte value
4273 of the symbol @samp{addr}:
4279 When using a 64 bit host or target, @code{QUAD} and @code{SQUAD} are the
4280 same; they both store an 8 byte, or 64 bit, value. When both host and
4281 target are 32 bits, an expression is computed as 32 bits. In this case
4282 @code{QUAD} stores a 32 bit value zero extended to 64 bits, and
4283 @code{SQUAD} stores a 32 bit value sign extended to 64 bits.
4285 If the object file format of the output file has an explicit endianness,
4286 which is the normal case, the value will be stored in that endianness.
4287 When the object file format does not have an explicit endianness, as is
4288 true of, for example, S-records, the value will be stored in the
4289 endianness of the first input object file.
4291 Note---these commands only work inside a section description and not
4292 between them, so the following will produce an error from the linker:
4294 SECTIONS @{@ .text : @{@ *(.text) @}@ LONG(1) .data : @{@ *(.data) @}@ @}@
4296 whereas this will work:
4298 SECTIONS @{@ .text : @{@ *(.text) ; LONG(1) @}@ .data : @{@ *(.data) @}@ @}@
4301 @kindex FILL(@var{expression})
4302 @cindex holes, filling
4303 @cindex unspecified memory
4304 You may use the @code{FILL} command to set the fill pattern for the
4305 current section. It is followed by an expression in parentheses. Any
4306 otherwise unspecified regions of memory within the section (for example,
4307 gaps left due to the required alignment of input sections) are filled
4308 with the value of the expression, repeated as
4309 necessary. A @code{FILL} statement covers memory locations after the
4310 point at which it occurs in the section definition; by including more
4311 than one @code{FILL} statement, you can have different fill patterns in
4312 different parts of an output section.
4314 This example shows how to fill unspecified regions of memory with the
4320 The @code{FILL} command is similar to the @samp{=@var{fillexp}} output
4321 section attribute, but it only affects the
4322 part of the section following the @code{FILL} command, rather than the
4323 entire section. If both are used, the @code{FILL} command takes
4324 precedence. @xref{Output Section Fill}, for details on the fill
4327 @node Output Section Keywords
4328 @subsection Output Section Keywords
4329 There are a couple of keywords which can appear as output section
4333 @kindex CREATE_OBJECT_SYMBOLS
4334 @cindex input filename symbols
4335 @cindex filename symbols
4336 @item CREATE_OBJECT_SYMBOLS
4337 The command tells the linker to create a symbol for each input file.
4338 The name of each symbol will be the name of the corresponding input
4339 file. The section of each symbol will be the output section in which
4340 the @code{CREATE_OBJECT_SYMBOLS} command appears.
4342 This is conventional for the a.out object file format. It is not
4343 normally used for any other object file format.
4345 @kindex CONSTRUCTORS
4346 @cindex C++ constructors, arranging in link
4347 @cindex constructors, arranging in link
4349 When linking using the a.out object file format, the linker uses an
4350 unusual set construct to support C++ global constructors and
4351 destructors. When linking object file formats which do not support
4352 arbitrary sections, such as ECOFF and XCOFF, the linker will
4353 automatically recognize C++ global constructors and destructors by name.
4354 For these object file formats, the @code{CONSTRUCTORS} command tells the
4355 linker to place constructor information in the output section where the
4356 @code{CONSTRUCTORS} command appears. The @code{CONSTRUCTORS} command is
4357 ignored for other object file formats.
4359 The symbol @w{@code{__CTOR_LIST__}} marks the start of the global
4360 constructors, and the symbol @w{@code{__CTOR_END__}} marks the end.
4361 Similarly, @w{@code{__DTOR_LIST__}} and @w{@code{__DTOR_END__}} mark
4362 the start and end of the global destructors. The
4363 first word in the list is the number of entries, followed by the address
4364 of each constructor or destructor, followed by a zero word. The
4365 compiler must arrange to actually run the code. For these object file
4366 formats @sc{gnu} C++ normally calls constructors from a subroutine
4367 @code{__main}; a call to @code{__main} is automatically inserted into
4368 the startup code for @code{main}. @sc{gnu} C++ normally runs
4369 destructors either by using @code{atexit}, or directly from the function
4372 For object file formats such as @code{COFF} or @code{ELF} which support
4373 arbitrary section names, @sc{gnu} C++ will normally arrange to put the
4374 addresses of global constructors and destructors into the @code{.ctors}
4375 and @code{.dtors} sections. Placing the following sequence into your
4376 linker script will build the sort of table which the @sc{gnu} C++
4377 runtime code expects to see.
4381 LONG((__CTOR_END__ - __CTOR_LIST__) / 4 - 2)
4386 LONG((__DTOR_END__ - __DTOR_LIST__) / 4 - 2)
4392 If you are using the @sc{gnu} C++ support for initialization priority,
4393 which provides some control over the order in which global constructors
4394 are run, you must sort the constructors at link time to ensure that they
4395 are executed in the correct order. When using the @code{CONSTRUCTORS}
4396 command, use @samp{SORT_BY_NAME(CONSTRUCTORS)} instead. When using the
4397 @code{.ctors} and @code{.dtors} sections, use @samp{*(SORT_BY_NAME(.ctors))} and
4398 @samp{*(SORT_BY_NAME(.dtors))} instead of just @samp{*(.ctors)} and
4401 Normally the compiler and linker will handle these issues automatically,
4402 and you will not need to concern yourself with them. However, you may
4403 need to consider this if you are using C++ and writing your own linker
4408 @node Output Section Discarding
4409 @subsection Output Section Discarding
4410 @cindex discarding sections
4411 @cindex sections, discarding
4412 @cindex removing sections
4413 The linker will not normally create output sections with no contents.
4414 This is for convenience when referring to input sections that may or
4415 may not be present in any of the input files. For example:
4417 .foo : @{ *(.foo) @}
4420 will only create a @samp{.foo} section in the output file if there is a
4421 @samp{.foo} section in at least one input file, and if the input
4422 sections are not all empty. Other link script directives that allocate
4423 space in an output section will also create the output section. So
4424 too will assignments to dot even if the assignment does not create
4425 space, except for @samp{. = 0}, @samp{. = . + 0}, @samp{. = sym},
4426 @samp{. = . + sym} and @samp{. = ALIGN (. != 0, expr, 1)} when
4427 @samp{sym} is an absolute symbol of value 0 defined in the script.
4428 This allows you to force output of an empty section with @samp{. = .}.
4430 The linker will ignore address assignments (@pxref{Output Section Address})
4431 on discarded output sections, except when the linker script defines
4432 symbols in the output section. In that case the linker will obey
4433 the address assignments, possibly advancing dot even though the
4434 section is discarded.
4437 The special output section name @samp{/DISCARD/} may be used to discard
4438 input sections. Any input sections which are assigned to an output
4439 section named @samp{/DISCARD/} are not included in the output file.
4441 @node Output Section Attributes
4442 @subsection Output Section Attributes
4443 @cindex output section attributes
4444 We showed above that the full description of an output section looked
4449 @var{section} [@var{address}] [(@var{type})] :
4451 [ALIGN(@var{section_align})]
4452 [SUBALIGN(@var{subsection_align})]
4455 @var{output-section-command}
4456 @var{output-section-command}
4458 @} [>@var{region}] [AT>@var{lma_region}] [:@var{phdr} :@var{phdr} @dots{}] [=@var{fillexp}]
4462 We've already described @var{section}, @var{address}, and
4463 @var{output-section-command}. In this section we will describe the
4464 remaining section attributes.
4467 * Output Section Type:: Output section type
4468 * Output Section LMA:: Output section LMA
4469 * Forced Output Alignment:: Forced Output Alignment
4470 * Forced Input Alignment:: Forced Input Alignment
4471 * Output Section Constraint:: Output section constraint
4472 * Output Section Region:: Output section region
4473 * Output Section Phdr:: Output section phdr
4474 * Output Section Fill:: Output section fill
4477 @node Output Section Type
4478 @subsubsection Output Section Type
4479 Each output section may have a type. The type is a keyword in
4480 parentheses. The following types are defined:
4484 The section should be marked as not loadable, so that it will not be
4485 loaded into memory when the program is run.
4490 These type names are supported for backward compatibility, and are
4491 rarely used. They all have the same effect: the section should be
4492 marked as not allocatable, so that no memory is allocated for the
4493 section when the program is run.
4497 @cindex prevent unnecessary loading
4498 @cindex loading, preventing
4499 The linker normally sets the attributes of an output section based on
4500 the input sections which map into it. You can override this by using
4501 the section type. For example, in the script sample below, the
4502 @samp{ROM} section is addressed at memory location @samp{0} and does not
4503 need to be loaded when the program is run.
4507 ROM 0 (NOLOAD) : @{ @dots{} @}
4513 @node Output Section LMA
4514 @subsubsection Output Section LMA
4515 @kindex AT>@var{lma_region}
4516 @kindex AT(@var{lma})
4517 @cindex load address
4518 @cindex section load address
4519 Every section has a virtual address (VMA) and a load address (LMA); see
4520 @ref{Basic Script Concepts}. The virtual address is specified by the
4521 @pxref{Output Section Address} described earlier. The load address is
4522 specified by the @code{AT} or @code{AT>} keywords. Specifying a load
4523 address is optional.
4525 The @code{AT} keyword takes an expression as an argument. This
4526 specifies the exact load address of the section. The @code{AT>} keyword
4527 takes the name of a memory region as an argument. @xref{MEMORY}. The
4528 load address of the section is set to the next free address in the
4529 region, aligned to the section's alignment requirements.
4531 If neither @code{AT} nor @code{AT>} is specified for an allocatable
4532 section, the linker will use the following heuristic to determine the
4537 If the section has a specific VMA address, then this is used as
4538 the LMA address as well.
4541 If the section is not allocatable then its LMA is set to its VMA.
4544 Otherwise if a memory region can be found that is compatible
4545 with the current section, and this region contains at least one
4546 section, then the LMA is set so the difference between the
4547 VMA and LMA is the same as the difference between the VMA and LMA of
4548 the last section in the located region.
4551 If no memory regions have been declared then a default region
4552 that covers the entire address space is used in the previous step.
4555 If no suitable region could be found, or there was no previous
4556 section then the LMA is set equal to the VMA.
4559 @cindex ROM initialized data
4560 @cindex initialized data in ROM
4561 This feature is designed to make it easy to build a ROM image. For
4562 example, the following linker script creates three output sections: one
4563 called @samp{.text}, which starts at @code{0x1000}, one called
4564 @samp{.mdata}, which is loaded at the end of the @samp{.text} section
4565 even though its VMA is @code{0x2000}, and one called @samp{.bss} to hold
4566 uninitialized data at address @code{0x3000}. The symbol @code{_data} is
4567 defined with the value @code{0x2000}, which shows that the location
4568 counter holds the VMA value, not the LMA value.
4574 .text 0x1000 : @{ *(.text) _etext = . ; @}
4576 AT ( ADDR (.text) + SIZEOF (.text) )
4577 @{ _data = . ; *(.data); _edata = . ; @}
4579 @{ _bstart = . ; *(.bss) *(COMMON) ; _bend = . ;@}
4584 The run-time initialization code for use with a program generated with
4585 this linker script would include something like the following, to copy
4586 the initialized data from the ROM image to its runtime address. Notice
4587 how this code takes advantage of the symbols defined by the linker
4592 extern char _etext, _data, _edata, _bstart, _bend;
4593 char *src = &_etext;
4596 /* ROM has data at end of text; copy it. */
4597 while (dst < &_edata)
4601 for (dst = &_bstart; dst< &_bend; dst++)
4606 @node Forced Output Alignment
4607 @subsubsection Forced Output Alignment
4608 @kindex ALIGN(@var{section_align})
4609 @cindex forcing output section alignment
4610 @cindex output section alignment
4611 You can increase an output section's alignment by using ALIGN. As an
4612 alternative you can force the output section alignment to the maximum alignment
4613 of all its input sections with ALIGN_WITH_INPUT. The alignment forced by
4614 ALIGN_WITH_INPUT is used even in case the load and virtual memory regions are
4617 @node Forced Input Alignment
4618 @subsubsection Forced Input Alignment
4619 @kindex SUBALIGN(@var{subsection_align})
4620 @cindex forcing input section alignment
4621 @cindex input section alignment
4622 You can force input section alignment within an output section by using
4623 SUBALIGN. The value specified overrides any alignment given by input
4624 sections, whether larger or smaller.
4626 @node Output Section Constraint
4627 @subsubsection Output Section Constraint
4630 @cindex constraints on output sections
4631 You can specify that an output section should only be created if all
4632 of its input sections are read-only or all of its input sections are
4633 read-write by using the keyword @code{ONLY_IF_RO} and
4634 @code{ONLY_IF_RW} respectively.
4636 @node Output Section Region
4637 @subsubsection Output Section Region
4638 @kindex >@var{region}
4639 @cindex section, assigning to memory region
4640 @cindex memory regions and sections
4641 You can assign a section to a previously defined region of memory by
4642 using @samp{>@var{region}}. @xref{MEMORY}.
4644 Here is a simple example:
4647 MEMORY @{ rom : ORIGIN = 0x1000, LENGTH = 0x1000 @}
4648 SECTIONS @{ ROM : @{ *(.text) @} >rom @}
4652 @node Output Section Phdr
4653 @subsubsection Output Section Phdr
4655 @cindex section, assigning to program header
4656 @cindex program headers and sections
4657 You can assign a section to a previously defined program segment by
4658 using @samp{:@var{phdr}}. @xref{PHDRS}. If a section is assigned to
4659 one or more segments, then all subsequent allocated sections will be
4660 assigned to those segments as well, unless they use an explicitly
4661 @code{:@var{phdr}} modifier. You can use @code{:NONE} to tell the
4662 linker to not put the section in any segment at all.
4664 Here is a simple example:
4667 PHDRS @{ text PT_LOAD ; @}
4668 SECTIONS @{ .text : @{ *(.text) @} :text @}
4672 @node Output Section Fill
4673 @subsubsection Output Section Fill
4674 @kindex =@var{fillexp}
4675 @cindex section fill pattern
4676 @cindex fill pattern, entire section
4677 You can set the fill pattern for an entire section by using
4678 @samp{=@var{fillexp}}. @var{fillexp} is an expression
4679 (@pxref{Expressions}). Any otherwise unspecified regions of memory
4680 within the output section (for example, gaps left due to the required
4681 alignment of input sections) will be filled with the value, repeated as
4682 necessary. If the fill expression is a simple hex number, ie. a string
4683 of hex digit starting with @samp{0x} and without a trailing @samp{k} or @samp{M}, then
4684 an arbitrarily long sequence of hex digits can be used to specify the
4685 fill pattern; Leading zeros become part of the pattern too. For all
4686 other cases, including extra parentheses or a unary @code{+}, the fill
4687 pattern is the four least significant bytes of the value of the
4688 expression. In all cases, the number is big-endian.
4690 You can also change the fill value with a @code{FILL} command in the
4691 output section commands; (@pxref{Output Section Data}).
4693 Here is a simple example:
4696 SECTIONS @{ .text : @{ *(.text) @} =0x90909090 @}
4700 @node Overlay Description
4701 @subsection Overlay Description
4704 An overlay description provides an easy way to describe sections which
4705 are to be loaded as part of a single memory image but are to be run at
4706 the same memory address. At run time, some sort of overlay manager will
4707 copy the overlaid sections in and out of the runtime memory address as
4708 required, perhaps by simply manipulating addressing bits. This approach
4709 can be useful, for example, when a certain region of memory is faster
4712 Overlays are described using the @code{OVERLAY} command. The
4713 @code{OVERLAY} command is used within a @code{SECTIONS} command, like an
4714 output section description. The full syntax of the @code{OVERLAY}
4715 command is as follows:
4718 OVERLAY [@var{start}] : [NOCROSSREFS] [AT ( @var{ldaddr} )]
4722 @var{output-section-command}
4723 @var{output-section-command}
4725 @} [:@var{phdr}@dots{}] [=@var{fill}]
4728 @var{output-section-command}
4729 @var{output-section-command}
4731 @} [:@var{phdr}@dots{}] [=@var{fill}]
4733 @} [>@var{region}] [:@var{phdr}@dots{}] [=@var{fill}]
4737 Everything is optional except @code{OVERLAY} (a keyword), and each
4738 section must have a name (@var{secname1} and @var{secname2} above). The
4739 section definitions within the @code{OVERLAY} construct are identical to
4740 those within the general @code{SECTIONS} construct (@pxref{SECTIONS}),
4741 except that no addresses and no memory regions may be defined for
4742 sections within an @code{OVERLAY}.
4744 The sections are all defined with the same starting address. The load
4745 addresses of the sections are arranged such that they are consecutive in
4746 memory starting at the load address used for the @code{OVERLAY} as a
4747 whole (as with normal section definitions, the load address is optional,
4748 and defaults to the start address; the start address is also optional,
4749 and defaults to the current value of the location counter).
4751 If the @code{NOCROSSREFS} keyword is used, and there are any
4752 references among the sections, the linker will report an error. Since
4753 the sections all run at the same address, it normally does not make
4754 sense for one section to refer directly to another.
4755 @xref{Miscellaneous Commands, NOCROSSREFS}.
4757 For each section within the @code{OVERLAY}, the linker automatically
4758 provides two symbols. The symbol @code{__load_start_@var{secname}} is
4759 defined as the starting load address of the section. The symbol
4760 @code{__load_stop_@var{secname}} is defined as the final load address of
4761 the section. Any characters within @var{secname} which are not legal
4762 within C identifiers are removed. C (or assembler) code may use these
4763 symbols to move the overlaid sections around as necessary.
4765 At the end of the overlay, the value of the location counter is set to
4766 the start address of the overlay plus the size of the largest section.
4768 Here is an example. Remember that this would appear inside a
4769 @code{SECTIONS} construct.
4772 OVERLAY 0x1000 : AT (0x4000)
4774 .text0 @{ o1/*.o(.text) @}
4775 .text1 @{ o2/*.o(.text) @}
4780 This will define both @samp{.text0} and @samp{.text1} to start at
4781 address 0x1000. @samp{.text0} will be loaded at address 0x4000, and
4782 @samp{.text1} will be loaded immediately after @samp{.text0}. The
4783 following symbols will be defined if referenced: @code{__load_start_text0},
4784 @code{__load_stop_text0}, @code{__load_start_text1},
4785 @code{__load_stop_text1}.
4787 C code to copy overlay @code{.text1} into the overlay area might look
4792 extern char __load_start_text1, __load_stop_text1;
4793 memcpy ((char *) 0x1000, &__load_start_text1,
4794 &__load_stop_text1 - &__load_start_text1);
4798 Note that the @code{OVERLAY} command is just syntactic sugar, since
4799 everything it does can be done using the more basic commands. The above
4800 example could have been written identically as follows.
4804 .text0 0x1000 : AT (0x4000) @{ o1/*.o(.text) @}
4805 PROVIDE (__load_start_text0 = LOADADDR (.text0));
4806 PROVIDE (__load_stop_text0 = LOADADDR (.text0) + SIZEOF (.text0));
4807 .text1 0x1000 : AT (0x4000 + SIZEOF (.text0)) @{ o2/*.o(.text) @}
4808 PROVIDE (__load_start_text1 = LOADADDR (.text1));
4809 PROVIDE (__load_stop_text1 = LOADADDR (.text1) + SIZEOF (.text1));
4810 . = 0x1000 + MAX (SIZEOF (.text0), SIZEOF (.text1));
4815 @section MEMORY Command
4817 @cindex memory regions
4818 @cindex regions of memory
4819 @cindex allocating memory
4820 @cindex discontinuous memory
4821 The linker's default configuration permits allocation of all available
4822 memory. You can override this by using the @code{MEMORY} command.
4824 The @code{MEMORY} command describes the location and size of blocks of
4825 memory in the target. You can use it to describe which memory regions
4826 may be used by the linker, and which memory regions it must avoid. You
4827 can then assign sections to particular memory regions. The linker will
4828 set section addresses based on the memory regions, and will warn about
4829 regions that become too full. The linker will not shuffle sections
4830 around to fit into the available regions.
4832 A linker script may contain at most one use of the @code{MEMORY}
4833 command. However, you can define as many blocks of memory within it as
4834 you wish. The syntax is:
4839 @var{name} [(@var{attr})] : ORIGIN = @var{origin}, LENGTH = @var{len}
4845 The @var{name} is a name used in the linker script to refer to the
4846 region. The region name has no meaning outside of the linker script.
4847 Region names are stored in a separate name space, and will not conflict
4848 with symbol names, file names, or section names. Each memory region
4849 must have a distinct name within the @code{MEMORY} command. However you can
4850 add later alias names to existing memory regions with the @ref{REGION_ALIAS}
4853 @cindex memory region attributes
4854 The @var{attr} string is an optional list of attributes that specify
4855 whether to use a particular memory region for an input section which is
4856 not explicitly mapped in the linker script. As described in
4857 @ref{SECTIONS}, if you do not specify an output section for some input
4858 section, the linker will create an output section with the same name as
4859 the input section. If you define region attributes, the linker will use
4860 them to select the memory region for the output section that it creates.
4862 The @var{attr} string must consist only of the following characters:
4877 Invert the sense of any of the attributes that follow
4880 If a unmapped section matches any of the listed attributes other than
4881 @samp{!}, it will be placed in the memory region. The @samp{!}
4882 attribute reverses this test, so that an unmapped section will be placed
4883 in the memory region only if it does not match any of the listed
4889 The @var{origin} is an numerical expression for the start address of
4890 the memory region. The expression must evaluate to a constant and it
4891 cannot involve any symbols. The keyword @code{ORIGIN} may be
4892 abbreviated to @code{org} or @code{o} (but not, for example,
4898 The @var{len} is an expression for the size in bytes of the memory
4899 region. As with the @var{origin} expression, the expression must
4900 be numerical only and must evaluate to a constant. The keyword
4901 @code{LENGTH} may be abbreviated to @code{len} or @code{l}.
4903 In the following example, we specify that there are two memory regions
4904 available for allocation: one starting at @samp{0} for 256 kilobytes,
4905 and the other starting at @samp{0x40000000} for four megabytes. The
4906 linker will place into the @samp{rom} memory region every section which
4907 is not explicitly mapped into a memory region, and is either read-only
4908 or executable. The linker will place other sections which are not
4909 explicitly mapped into a memory region into the @samp{ram} memory
4916 rom (rx) : ORIGIN = 0, LENGTH = 256K
4917 ram (!rx) : org = 0x40000000, l = 4M
4922 Once you define a memory region, you can direct the linker to place
4923 specific output sections into that memory region by using the
4924 @samp{>@var{region}} output section attribute. For example, if you have
4925 a memory region named @samp{mem}, you would use @samp{>mem} in the
4926 output section definition. @xref{Output Section Region}. If no address
4927 was specified for the output section, the linker will set the address to
4928 the next available address within the memory region. If the combined
4929 output sections directed to a memory region are too large for the
4930 region, the linker will issue an error message.
4932 It is possible to access the origin and length of a memory in an
4933 expression via the @code{ORIGIN(@var{memory})} and
4934 @code{LENGTH(@var{memory})} functions:
4938 _fstack = ORIGIN(ram) + LENGTH(ram) - 4;
4943 @section PHDRS Command
4945 @cindex program headers
4946 @cindex ELF program headers
4947 @cindex program segments
4948 @cindex segments, ELF
4949 The ELF object file format uses @dfn{program headers}, also knows as
4950 @dfn{segments}. The program headers describe how the program should be
4951 loaded into memory. You can print them out by using the @code{objdump}
4952 program with the @samp{-p} option.
4954 When you run an ELF program on a native ELF system, the system loader
4955 reads the program headers in order to figure out how to load the
4956 program. This will only work if the program headers are set correctly.
4957 This manual does not describe the details of how the system loader
4958 interprets program headers; for more information, see the ELF ABI.
4960 The linker will create reasonable program headers by default. However,
4961 in some cases, you may need to specify the program headers more
4962 precisely. You may use the @code{PHDRS} command for this purpose. When
4963 the linker sees the @code{PHDRS} command in the linker script, it will
4964 not create any program headers other than the ones specified.
4966 The linker only pays attention to the @code{PHDRS} command when
4967 generating an ELF output file. In other cases, the linker will simply
4968 ignore @code{PHDRS}.
4970 This is the syntax of the @code{PHDRS} command. The words @code{PHDRS},
4971 @code{FILEHDR}, @code{AT}, and @code{FLAGS} are keywords.
4977 @var{name} @var{type} [ FILEHDR ] [ PHDRS ] [ AT ( @var{address} ) ]
4978 [ FLAGS ( @var{flags} ) ] ;
4983 The @var{name} is used only for reference in the @code{SECTIONS} command
4984 of the linker script. It is not put into the output file. Program
4985 header names are stored in a separate name space, and will not conflict
4986 with symbol names, file names, or section names. Each program header
4987 must have a distinct name. The headers are processed in order and it
4988 is usual for them to map to sections in ascending load address order.
4990 Certain program header types describe segments of memory which the
4991 system loader will load from the file. In the linker script, you
4992 specify the contents of these segments by placing allocatable output
4993 sections in the segments. You use the @samp{:@var{phdr}} output section
4994 attribute to place a section in a particular segment. @xref{Output
4997 It is normal to put certain sections in more than one segment. This
4998 merely implies that one segment of memory contains another. You may
4999 repeat @samp{:@var{phdr}}, using it once for each segment which should
5000 contain the section.
5002 If you place a section in one or more segments using @samp{:@var{phdr}},
5003 then the linker will place all subsequent allocatable sections which do
5004 not specify @samp{:@var{phdr}} in the same segments. This is for
5005 convenience, since generally a whole set of contiguous sections will be
5006 placed in a single segment. You can use @code{:NONE} to override the
5007 default segment and tell the linker to not put the section in any
5012 You may use the @code{FILEHDR} and @code{PHDRS} keywords after
5013 the program header type to further describe the contents of the segment.
5014 The @code{FILEHDR} keyword means that the segment should include the ELF
5015 file header. The @code{PHDRS} keyword means that the segment should
5016 include the ELF program headers themselves. If applied to a loadable
5017 segment (@code{PT_LOAD}), all prior loadable segments must have one of
5020 The @var{type} may be one of the following. The numbers indicate the
5021 value of the keyword.
5024 @item @code{PT_NULL} (0)
5025 Indicates an unused program header.
5027 @item @code{PT_LOAD} (1)
5028 Indicates that this program header describes a segment to be loaded from
5031 @item @code{PT_DYNAMIC} (2)
5032 Indicates a segment where dynamic linking information can be found.
5034 @item @code{PT_INTERP} (3)
5035 Indicates a segment where the name of the program interpreter may be
5038 @item @code{PT_NOTE} (4)
5039 Indicates a segment holding note information.
5041 @item @code{PT_SHLIB} (5)
5042 A reserved program header type, defined but not specified by the ELF
5045 @item @code{PT_PHDR} (6)
5046 Indicates a segment where the program headers may be found.
5048 @item @var{expression}
5049 An expression giving the numeric type of the program header. This may
5050 be used for types not defined above.
5053 You can specify that a segment should be loaded at a particular address
5054 in memory by using an @code{AT} expression. This is identical to the
5055 @code{AT} command used as an output section attribute (@pxref{Output
5056 Section LMA}). The @code{AT} command for a program header overrides the
5057 output section attribute.
5059 The linker will normally set the segment flags based on the sections
5060 which comprise the segment. You may use the @code{FLAGS} keyword to
5061 explicitly specify the segment flags. The value of @var{flags} must be
5062 an integer. It is used to set the @code{p_flags} field of the program
5065 Here is an example of @code{PHDRS}. This shows a typical set of program
5066 headers used on a native ELF system.
5072 headers PT_PHDR PHDRS ;
5074 text PT_LOAD FILEHDR PHDRS ;
5076 dynamic PT_DYNAMIC ;
5082 .interp : @{ *(.interp) @} :text :interp
5083 .text : @{ *(.text) @} :text
5084 .rodata : @{ *(.rodata) @} /* defaults to :text */
5086 . = . + 0x1000; /* move to a new page in memory */
5087 .data : @{ *(.data) @} :data
5088 .dynamic : @{ *(.dynamic) @} :data :dynamic
5095 @section VERSION Command
5096 @kindex VERSION @{script text@}
5097 @cindex symbol versions
5098 @cindex version script
5099 @cindex versions of symbols
5100 The linker supports symbol versions when using ELF. Symbol versions are
5101 only useful when using shared libraries. The dynamic linker can use
5102 symbol versions to select a specific version of a function when it runs
5103 a program that may have been linked against an earlier version of the
5106 You can include a version script directly in the main linker script, or
5107 you can supply the version script as an implicit linker script. You can
5108 also use the @samp{--version-script} linker option.
5110 The syntax of the @code{VERSION} command is simply
5112 VERSION @{ version-script-commands @}
5115 The format of the version script commands is identical to that used by
5116 Sun's linker in Solaris 2.5. The version script defines a tree of
5117 version nodes. You specify the node names and interdependencies in the
5118 version script. You can specify which symbols are bound to which
5119 version nodes, and you can reduce a specified set of symbols to local
5120 scope so that they are not globally visible outside of the shared
5123 The easiest way to demonstrate the version script language is with a few
5149 This example version script defines three version nodes. The first
5150 version node defined is @samp{VERS_1.1}; it has no other dependencies.
5151 The script binds the symbol @samp{foo1} to @samp{VERS_1.1}. It reduces
5152 a number of symbols to local scope so that they are not visible outside
5153 of the shared library; this is done using wildcard patterns, so that any
5154 symbol whose name begins with @samp{old}, @samp{original}, or @samp{new}
5155 is matched. The wildcard patterns available are the same as those used
5156 in the shell when matching filenames (also known as ``globbing'').
5157 However, if you specify the symbol name inside double quotes, then the
5158 name is treated as literal, rather than as a glob pattern.
5160 Next, the version script defines node @samp{VERS_1.2}. This node
5161 depends upon @samp{VERS_1.1}. The script binds the symbol @samp{foo2}
5162 to the version node @samp{VERS_1.2}.
5164 Finally, the version script defines node @samp{VERS_2.0}. This node
5165 depends upon @samp{VERS_1.2}. The scripts binds the symbols @samp{bar1}
5166 and @samp{bar2} are bound to the version node @samp{VERS_2.0}.
5168 When the linker finds a symbol defined in a library which is not
5169 specifically bound to a version node, it will effectively bind it to an
5170 unspecified base version of the library. You can bind all otherwise
5171 unspecified symbols to a given version node by using @samp{global: *;}
5172 somewhere in the version script. Note that it's slightly crazy to use
5173 wildcards in a global spec except on the last version node. Global
5174 wildcards elsewhere run the risk of accidentally adding symbols to the
5175 set exported for an old version. That's wrong since older versions
5176 ought to have a fixed set of symbols.
5178 The names of the version nodes have no specific meaning other than what
5179 they might suggest to the person reading them. The @samp{2.0} version
5180 could just as well have appeared in between @samp{1.1} and @samp{1.2}.
5181 However, this would be a confusing way to write a version script.
5183 Node name can be omitted, provided it is the only version node
5184 in the version script. Such version script doesn't assign any versions to
5185 symbols, only selects which symbols will be globally visible out and which
5189 @{ global: foo; bar; local: *; @};
5192 When you link an application against a shared library that has versioned
5193 symbols, the application itself knows which version of each symbol it
5194 requires, and it also knows which version nodes it needs from each
5195 shared library it is linked against. Thus at runtime, the dynamic
5196 loader can make a quick check to make sure that the libraries you have
5197 linked against do in fact supply all of the version nodes that the
5198 application will need to resolve all of the dynamic symbols. In this
5199 way it is possible for the dynamic linker to know with certainty that
5200 all external symbols that it needs will be resolvable without having to
5201 search for each symbol reference.
5203 The symbol versioning is in effect a much more sophisticated way of
5204 doing minor version checking that SunOS does. The fundamental problem
5205 that is being addressed here is that typically references to external
5206 functions are bound on an as-needed basis, and are not all bound when
5207 the application starts up. If a shared library is out of date, a
5208 required interface may be missing; when the application tries to use
5209 that interface, it may suddenly and unexpectedly fail. With symbol
5210 versioning, the user will get a warning when they start their program if
5211 the libraries being used with the application are too old.
5213 There are several GNU extensions to Sun's versioning approach. The
5214 first of these is the ability to bind a symbol to a version node in the
5215 source file where the symbol is defined instead of in the versioning
5216 script. This was done mainly to reduce the burden on the library
5217 maintainer. You can do this by putting something like:
5219 __asm__(".symver original_foo,foo@@VERS_1.1");
5222 in the C source file. This renames the function @samp{original_foo} to
5223 be an alias for @samp{foo} bound to the version node @samp{VERS_1.1}.
5224 The @samp{local:} directive can be used to prevent the symbol
5225 @samp{original_foo} from being exported. A @samp{.symver} directive
5226 takes precedence over a version script.
5228 The second GNU extension is to allow multiple versions of the same
5229 function to appear in a given shared library. In this way you can make
5230 an incompatible change to an interface without increasing the major
5231 version number of the shared library, while still allowing applications
5232 linked against the old interface to continue to function.
5234 To do this, you must use multiple @samp{.symver} directives in the
5235 source file. Here is an example:
5238 __asm__(".symver original_foo,foo@@");
5239 __asm__(".symver old_foo,foo@@VERS_1.1");
5240 __asm__(".symver old_foo1,foo@@VERS_1.2");
5241 __asm__(".symver new_foo,foo@@@@VERS_2.0");
5244 In this example, @samp{foo@@} represents the symbol @samp{foo} bound to the
5245 unspecified base version of the symbol. The source file that contains this
5246 example would define 4 C functions: @samp{original_foo}, @samp{old_foo},
5247 @samp{old_foo1}, and @samp{new_foo}.
5249 When you have multiple definitions of a given symbol, there needs to be
5250 some way to specify a default version to which external references to
5251 this symbol will be bound. You can do this with the
5252 @samp{foo@@@@VERS_2.0} type of @samp{.symver} directive. You can only
5253 declare one version of a symbol as the default in this manner; otherwise
5254 you would effectively have multiple definitions of the same symbol.
5256 If you wish to bind a reference to a specific version of the symbol
5257 within the shared library, you can use the aliases of convenience
5258 (i.e., @samp{old_foo}), or you can use the @samp{.symver} directive to
5259 specifically bind to an external version of the function in question.
5261 You can also specify the language in the version script:
5264 VERSION extern "lang" @{ version-script-commands @}
5267 The supported @samp{lang}s are @samp{C}, @samp{C++}, and @samp{Java}.
5268 The linker will iterate over the list of symbols at the link time and
5269 demangle them according to @samp{lang} before matching them to the
5270 patterns specified in @samp{version-script-commands}. The default
5271 @samp{lang} is @samp{C}.
5273 Demangled names may contains spaces and other special characters. As
5274 described above, you can use a glob pattern to match demangled names,
5275 or you can use a double-quoted string to match the string exactly. In
5276 the latter case, be aware that minor differences (such as differing
5277 whitespace) between the version script and the demangler output will
5278 cause a mismatch. As the exact string generated by the demangler
5279 might change in the future, even if the mangled name does not, you
5280 should check that all of your version directives are behaving as you
5281 expect when you upgrade.
5284 @section Expressions in Linker Scripts
5287 The syntax for expressions in the linker script language is identical to
5288 that of C expressions. All expressions are evaluated as integers. All
5289 expressions are evaluated in the same size, which is 32 bits if both the
5290 host and target are 32 bits, and is otherwise 64 bits.
5292 You can use and set symbol values in expressions.
5294 The linker defines several special purpose builtin functions for use in
5298 * Constants:: Constants
5299 * Symbolic Constants:: Symbolic constants
5300 * Symbols:: Symbol Names
5301 * Orphan Sections:: Orphan Sections
5302 * Location Counter:: The Location Counter
5303 * Operators:: Operators
5304 * Evaluation:: Evaluation
5305 * Expression Section:: The Section of an Expression
5306 * Builtin Functions:: Builtin Functions
5310 @subsection Constants
5311 @cindex integer notation
5312 @cindex constants in linker scripts
5313 All constants are integers.
5315 As in C, the linker considers an integer beginning with @samp{0} to be
5316 octal, and an integer beginning with @samp{0x} or @samp{0X} to be
5317 hexadecimal. Alternatively the linker accepts suffixes of @samp{h} or
5318 @samp{H} for hexadecimal, @samp{o} or @samp{O} for octal, @samp{b} or
5319 @samp{B} for binary and @samp{d} or @samp{D} for decimal. Any integer
5320 value without a prefix or a suffix is considered to be decimal.
5322 @cindex scaled integers
5323 @cindex K and M integer suffixes
5324 @cindex M and K integer suffixes
5325 @cindex suffixes for integers
5326 @cindex integer suffixes
5327 In addition, you can use the suffixes @code{K} and @code{M} to scale a
5331 @c END TEXI2ROFF-KILL
5332 @code{1024} or @code{1024*1024}
5336 ${\rm 1024}$ or ${\rm 1024}^2$
5338 @c END TEXI2ROFF-KILL
5339 respectively. For example, the following
5340 all refer to the same quantity:
5349 Note - the @code{K} and @code{M} suffixes cannot be used in
5350 conjunction with the base suffixes mentioned above.
5352 @node Symbolic Constants
5353 @subsection Symbolic Constants
5354 @cindex symbolic constants
5356 It is possible to refer to target specific constants via the use of
5357 the @code{CONSTANT(@var{name})} operator, where @var{name} is one of:
5362 The target's maximum page size.
5364 @item COMMONPAGESIZE
5365 @kindex COMMONPAGESIZE
5366 The target's default page size.
5372 .text ALIGN (CONSTANT (MAXPAGESIZE)) : @{ *(.text) @}
5375 will create a text section aligned to the largest page boundary
5376 supported by the target.
5379 @subsection Symbol Names
5380 @cindex symbol names
5382 @cindex quoted symbol names
5384 Unless quoted, symbol names start with a letter, underscore, or period
5385 and may include letters, digits, underscores, periods, and hyphens.
5386 Unquoted symbol names must not conflict with any keywords. You can
5387 specify a symbol which contains odd characters or has the same name as a
5388 keyword by surrounding the symbol name in double quotes:
5391 "with a space" = "also with a space" + 10;
5394 Since symbols can contain many non-alphabetic characters, it is safest
5395 to delimit symbols with spaces. For example, @samp{A-B} is one symbol,
5396 whereas @samp{A - B} is an expression involving subtraction.
5398 @node Orphan Sections
5399 @subsection Orphan Sections
5401 Orphan sections are sections present in the input files which
5402 are not explicitly placed into the output file by the linker
5403 script. The linker will still copy these sections into the
5404 output file, but it has to guess as to where they should be
5405 placed. The linker uses a simple heuristic to do this. It
5406 attempts to place orphan sections after non-orphan sections of the
5407 same attribute, such as code vs data, loadable vs non-loadable, etc.
5408 If there is not enough room to do this then it places
5409 at the end of the file.
5411 For ELF targets, the attribute of the section includes section type as
5412 well as section flag.
5414 If an orphaned section's name is representable as a C identifier then
5415 the linker will automatically @pxref{PROVIDE} two symbols:
5416 __start_SECNAME and __stop_SECNAME, where SECNAME is the name of the
5417 section. These indicate the start address and end address of the
5418 orphaned section respectively. Note: most section names are not
5419 representable as C identifiers because they contain a @samp{.}
5422 @node Location Counter
5423 @subsection The Location Counter
5426 @cindex location counter
5427 @cindex current output location
5428 The special linker variable @dfn{dot} @samp{.} always contains the
5429 current output location counter. Since the @code{.} always refers to a
5430 location in an output section, it may only appear in an expression
5431 within a @code{SECTIONS} command. The @code{.} symbol may appear
5432 anywhere that an ordinary symbol is allowed in an expression.
5435 Assigning a value to @code{.} will cause the location counter to be
5436 moved. This may be used to create holes in the output section. The
5437 location counter may not be moved backwards inside an output section,
5438 and may not be moved backwards outside of an output section if so
5439 doing creates areas with overlapping LMAs.
5455 In the previous example, the @samp{.text} section from @file{file1} is
5456 located at the beginning of the output section @samp{output}. It is
5457 followed by a 1000 byte gap. Then the @samp{.text} section from
5458 @file{file2} appears, also with a 1000 byte gap following before the
5459 @samp{.text} section from @file{file3}. The notation @samp{= 0x12345678}
5460 specifies what data to write in the gaps (@pxref{Output Section Fill}).
5462 @cindex dot inside sections
5463 Note: @code{.} actually refers to the byte offset from the start of the
5464 current containing object. Normally this is the @code{SECTIONS}
5465 statement, whose start address is 0, hence @code{.} can be used as an
5466 absolute address. If @code{.} is used inside a section description
5467 however, it refers to the byte offset from the start of that section,
5468 not an absolute address. Thus in a script like this:
5486 The @samp{.text} section will be assigned a starting address of 0x100
5487 and a size of exactly 0x200 bytes, even if there is not enough data in
5488 the @samp{.text} input sections to fill this area. (If there is too
5489 much data, an error will be produced because this would be an attempt to
5490 move @code{.} backwards). The @samp{.data} section will start at 0x500
5491 and it will have an extra 0x600 bytes worth of space after the end of
5492 the values from the @samp{.data} input sections and before the end of
5493 the @samp{.data} output section itself.
5495 @cindex dot outside sections
5496 Setting symbols to the value of the location counter outside of an
5497 output section statement can result in unexpected values if the linker
5498 needs to place orphan sections. For example, given the following:
5504 .text: @{ *(.text) @}
5508 .data: @{ *(.data) @}
5513 If the linker needs to place some input section, e.g. @code{.rodata},
5514 not mentioned in the script, it might choose to place that section
5515 between @code{.text} and @code{.data}. You might think the linker
5516 should place @code{.rodata} on the blank line in the above script, but
5517 blank lines are of no particular significance to the linker. As well,
5518 the linker doesn't associate the above symbol names with their
5519 sections. Instead, it assumes that all assignments or other
5520 statements belong to the previous output section, except for the
5521 special case of an assignment to @code{.}. I.e., the linker will
5522 place the orphan @code{.rodata} section as if the script was written
5529 .text: @{ *(.text) @}
5533 .rodata: @{ *(.rodata) @}
5534 .data: @{ *(.data) @}
5539 This may or may not be the script author's intention for the value of
5540 @code{start_of_data}. One way to influence the orphan section
5541 placement is to assign the location counter to itself, as the linker
5542 assumes that an assignment to @code{.} is setting the start address of
5543 a following output section and thus should be grouped with that
5544 section. So you could write:
5550 .text: @{ *(.text) @}
5555 .data: @{ *(.data) @}
5560 Now, the orphan @code{.rodata} section will be placed between
5561 @code{end_of_text} and @code{start_of_data}.
5565 @subsection Operators
5566 @cindex operators for arithmetic
5567 @cindex arithmetic operators
5568 @cindex precedence in expressions
5569 The linker recognizes the standard C set of arithmetic operators, with
5570 the standard bindings and precedence levels:
5573 @c END TEXI2ROFF-KILL
5575 precedence associativity Operators Notes
5581 5 left == != > < <= >=
5587 11 right &= += -= *= /= (2)
5591 (1) Prefix operators
5592 (2) @xref{Assignments}.
5596 \vskip \baselineskip
5597 %"lispnarrowing" is the extra indent used generally for smallexample
5598 \hskip\lispnarrowing\vbox{\offinterlineskip
5601 {\vrule#&\strut\hfil\ #\ \hfil&\vrule#&\strut\hfil\ #\ \hfil&\vrule#&\strut\hfil\ {\tt #}\ \hfil&\vrule#\cr
5602 height2pt&\omit&&\omit&&\omit&\cr
5603 &Precedence&& Associativity &&{\rm Operators}&\cr
5604 height2pt&\omit&&\omit&&\omit&\cr
5606 height2pt&\omit&&\omit&&\omit&\cr
5608 % '176 is tilde, '~' in tt font
5609 &1&&left&&\qquad- \char'176\ !\qquad\dag&\cr
5610 &2&&left&&* / \%&\cr
5613 &5&&left&&== != > < <= >=&\cr
5616 &8&&left&&{\&\&}&\cr
5619 &11&&right&&\qquad\&= += -= *= /=\qquad\ddag&\cr
5621 height2pt&\omit&&\omit&&\omit&\cr}
5626 @obeylines@parskip=0pt@parindent=0pt
5627 @dag@quad Prefix operators.
5628 @ddag@quad @xref{Assignments}.
5631 @c END TEXI2ROFF-KILL
5634 @subsection Evaluation
5635 @cindex lazy evaluation
5636 @cindex expression evaluation order
5637 The linker evaluates expressions lazily. It only computes the value of
5638 an expression when absolutely necessary.
5640 The linker needs some information, such as the value of the start
5641 address of the first section, and the origins and lengths of memory
5642 regions, in order to do any linking at all. These values are computed
5643 as soon as possible when the linker reads in the linker script.
5645 However, other values (such as symbol values) are not known or needed
5646 until after storage allocation. Such values are evaluated later, when
5647 other information (such as the sizes of output sections) is available
5648 for use in the symbol assignment expression.
5650 The sizes of sections cannot be known until after allocation, so
5651 assignments dependent upon these are not performed until after
5654 Some expressions, such as those depending upon the location counter
5655 @samp{.}, must be evaluated during section allocation.
5657 If the result of an expression is required, but the value is not
5658 available, then an error results. For example, a script like the
5664 .text 9+this_isnt_constant :
5670 will cause the error message @samp{non constant expression for initial
5673 @node Expression Section
5674 @subsection The Section of an Expression
5675 @cindex expression sections
5676 @cindex absolute expressions
5677 @cindex relative expressions
5678 @cindex absolute and relocatable symbols
5679 @cindex relocatable and absolute symbols
5680 @cindex symbols, relocatable and absolute
5681 Addresses and symbols may be section relative, or absolute. A section
5682 relative symbol is relocatable. If you request relocatable output
5683 using the @samp{-r} option, a further link operation may change the
5684 value of a section relative symbol. On the other hand, an absolute
5685 symbol will retain the same value throughout any further link
5688 Some terms in linker expressions are addresses. This is true of
5689 section relative symbols and for builtin functions that return an
5690 address, such as @code{ADDR}, @code{LOADADDR}, @code{ORIGIN} and
5691 @code{SEGMENT_START}. Other terms are simply numbers, or are builtin
5692 functions that return a non-address value, such as @code{LENGTH}.
5693 One complication is that unless you set @code{LD_FEATURE ("SANE_EXPR")}
5694 (@pxref{Miscellaneous Commands}), numbers and absolute symbols are treated
5695 differently depending on their location, for compatibility with older
5696 versions of @code{ld}. Expressions appearing outside an output
5697 section definition treat all numbers as absolute addresses.
5698 Expressions appearing inside an output section definition treat
5699 absolute symbols as numbers. If @code{LD_FEATURE ("SANE_EXPR")} is
5700 given, then absolute symbols and numbers are simply treated as numbers
5703 In the following simple example,
5710 __executable_start = 0x100;
5714 __data_start = 0x10;
5722 both @code{.} and @code{__executable_start} are set to the absolute
5723 address 0x100 in the first two assignments, then both @code{.} and
5724 @code{__data_start} are set to 0x10 relative to the @code{.data}
5725 section in the second two assignments.
5727 For expressions involving numbers, relative addresses and absolute
5728 addresses, ld follows these rules to evaluate terms:
5732 Unary operations on an absolute address or number, and binary
5733 operations on two absolute addresses or two numbers, or between one
5734 absolute address and a number, apply the operator to the value(s).
5736 Unary operations on a relative address, and binary operations on two
5737 relative addresses in the same section or between one relative address
5738 and a number, apply the operator to the offset part of the address(es).
5740 Other binary operations, that is, between two relative addresses not
5741 in the same section, or between a relative address and an absolute
5742 address, first convert any non-absolute term to an absolute address
5743 before applying the operator.
5746 The result section of each sub-expression is as follows:
5750 An operation involving only numbers results in a number.
5752 The result of comparisons, @samp{&&} and @samp{||} is also a number.
5754 The result of other binary arithmetic and logical operations on two
5755 relative addresses in the same section or two absolute addresses
5756 (after above conversions) is also a number.
5758 The result of other operations on relative addresses or one
5759 relative address and a number, is a relative address in the same
5760 section as the relative operand(s).
5762 The result of other operations on absolute addresses (after above
5763 conversions) is an absolute address.
5766 You can use the builtin function @code{ABSOLUTE} to force an expression
5767 to be absolute when it would otherwise be relative. For example, to
5768 create an absolute symbol set to the address of the end of the output
5769 section @samp{.data}:
5773 .data : @{ *(.data) _edata = ABSOLUTE(.); @}
5777 If @samp{ABSOLUTE} were not used, @samp{_edata} would be relative to the
5778 @samp{.data} section.
5780 Using @code{LOADADDR} also forces an expression absolute, since this
5781 particular builtin function returns an absolute address.
5783 @node Builtin Functions
5784 @subsection Builtin Functions
5785 @cindex functions in expressions
5786 The linker script language includes a number of builtin functions for
5787 use in linker script expressions.
5790 @item ABSOLUTE(@var{exp})
5791 @kindex ABSOLUTE(@var{exp})
5792 @cindex expression, absolute
5793 Return the absolute (non-relocatable, as opposed to non-negative) value
5794 of the expression @var{exp}. Primarily useful to assign an absolute
5795 value to a symbol within a section definition, where symbol values are
5796 normally section relative. @xref{Expression Section}.
5798 @item ADDR(@var{section})
5799 @kindex ADDR(@var{section})
5800 @cindex section address in expression
5801 Return the address (VMA) of the named @var{section}. Your
5802 script must previously have defined the location of that section. In
5803 the following example, @code{start_of_output_1}, @code{symbol_1} and
5804 @code{symbol_2} are assigned equivalent values, except that
5805 @code{symbol_1} will be relative to the @code{.output1} section while
5806 the other two will be absolute:
5812 start_of_output_1 = ABSOLUTE(.);
5817 symbol_1 = ADDR(.output1);
5818 symbol_2 = start_of_output_1;
5824 @item ALIGN(@var{align})
5825 @itemx ALIGN(@var{exp},@var{align})
5826 @kindex ALIGN(@var{align})
5827 @kindex ALIGN(@var{exp},@var{align})
5828 @cindex round up location counter
5829 @cindex align location counter
5830 @cindex round up expression
5831 @cindex align expression
5832 Return the location counter (@code{.}) or arbitrary expression aligned
5833 to the next @var{align} boundary. The single operand @code{ALIGN}
5834 doesn't change the value of the location counter---it just does
5835 arithmetic on it. The two operand @code{ALIGN} allows an arbitrary
5836 expression to be aligned upwards (@code{ALIGN(@var{align})} is
5837 equivalent to @code{ALIGN(., @var{align})}).
5839 Here is an example which aligns the output @code{.data} section to the
5840 next @code{0x2000} byte boundary after the preceding section and sets a
5841 variable within the section to the next @code{0x8000} boundary after the
5846 .data ALIGN(0x2000): @{
5848 variable = ALIGN(0x8000);
5854 The first use of @code{ALIGN} in this example specifies the location of
5855 a section because it is used as the optional @var{address} attribute of
5856 a section definition (@pxref{Output Section Address}). The second use
5857 of @code{ALIGN} is used to defines the value of a symbol.
5859 The builtin function @code{NEXT} is closely related to @code{ALIGN}.
5861 @item ALIGNOF(@var{section})
5862 @kindex ALIGNOF(@var{section})
5863 @cindex section alignment
5864 Return the alignment in bytes of the named @var{section}, if that section has
5865 been allocated. If the section has not been allocated when this is
5866 evaluated, the linker will report an error. In the following example,
5867 the alignment of the @code{.output} section is stored as the first
5868 value in that section.
5873 LONG (ALIGNOF (.output))
5880 @item BLOCK(@var{exp})
5881 @kindex BLOCK(@var{exp})
5882 This is a synonym for @code{ALIGN}, for compatibility with older linker
5883 scripts. It is most often seen when setting the address of an output
5886 @item DATA_SEGMENT_ALIGN(@var{maxpagesize}, @var{commonpagesize})
5887 @kindex DATA_SEGMENT_ALIGN(@var{maxpagesize}, @var{commonpagesize})
5888 This is equivalent to either
5890 (ALIGN(@var{maxpagesize}) + (. & (@var{maxpagesize} - 1)))
5894 (ALIGN(@var{maxpagesize}) + (. & (@var{maxpagesize} - @var{commonpagesize})))
5897 depending on whether the latter uses fewer @var{commonpagesize} sized pages
5898 for the data segment (area between the result of this expression and
5899 @code{DATA_SEGMENT_END}) than the former or not.
5900 If the latter form is used, it means @var{commonpagesize} bytes of runtime
5901 memory will be saved at the expense of up to @var{commonpagesize} wasted
5902 bytes in the on-disk file.
5904 This expression can only be used directly in @code{SECTIONS} commands, not in
5905 any output section descriptions and only once in the linker script.
5906 @var{commonpagesize} should be less or equal to @var{maxpagesize} and should
5907 be the system page size the object wants to be optimized for (while still
5908 working on system page sizes up to @var{maxpagesize}).
5913 . = DATA_SEGMENT_ALIGN(0x10000, 0x2000);
5916 @item DATA_SEGMENT_END(@var{exp})
5917 @kindex DATA_SEGMENT_END(@var{exp})
5918 This defines the end of data segment for @code{DATA_SEGMENT_ALIGN}
5919 evaluation purposes.
5922 . = DATA_SEGMENT_END(.);
5925 @item DATA_SEGMENT_RELRO_END(@var{offset}, @var{exp})
5926 @kindex DATA_SEGMENT_RELRO_END(@var{offset}, @var{exp})
5927 This defines the end of the @code{PT_GNU_RELRO} segment when
5928 @samp{-z relro} option is used.
5929 When @samp{-z relro} option is not present, @code{DATA_SEGMENT_RELRO_END}
5930 does nothing, otherwise @code{DATA_SEGMENT_ALIGN} is padded so that
5931 @var{exp} + @var{offset} is aligned to the most commonly used page
5932 boundary for particular target. If present in the linker script,
5933 it must always come in between @code{DATA_SEGMENT_ALIGN} and
5934 @code{DATA_SEGMENT_END}. Evaluates to the second argument plus any
5935 padding needed at the end of the @code{PT_GNU_RELRO} segment due to
5939 . = DATA_SEGMENT_RELRO_END(24, .);
5942 @item DEFINED(@var{symbol})
5943 @kindex DEFINED(@var{symbol})
5944 @cindex symbol defaults
5945 Return 1 if @var{symbol} is in the linker global symbol table and is
5946 defined before the statement using DEFINED in the script, otherwise
5947 return 0. You can use this function to provide
5948 default values for symbols. For example, the following script fragment
5949 shows how to set a global symbol @samp{begin} to the first location in
5950 the @samp{.text} section---but if a symbol called @samp{begin} already
5951 existed, its value is preserved:
5957 begin = DEFINED(begin) ? begin : . ;
5965 @item LENGTH(@var{memory})
5966 @kindex LENGTH(@var{memory})
5967 Return the length of the memory region named @var{memory}.
5969 @item LOADADDR(@var{section})
5970 @kindex LOADADDR(@var{section})
5971 @cindex section load address in expression
5972 Return the absolute LMA of the named @var{section}. (@pxref{Output
5975 @item LOG2CEIL(@var{exp})
5976 @kindex LOG2CEIL(@var{exp})
5977 Return the binary logarithm of @var{exp} rounded towards infinity.
5978 @code{LOG2CEIL(0)} returns 0.
5981 @item MAX(@var{exp1}, @var{exp2})
5982 Returns the maximum of @var{exp1} and @var{exp2}.
5985 @item MIN(@var{exp1}, @var{exp2})
5986 Returns the minimum of @var{exp1} and @var{exp2}.
5988 @item NEXT(@var{exp})
5989 @kindex NEXT(@var{exp})
5990 @cindex unallocated address, next
5991 Return the next unallocated address that is a multiple of @var{exp}.
5992 This function is closely related to @code{ALIGN(@var{exp})}; unless you
5993 use the @code{MEMORY} command to define discontinuous memory for the
5994 output file, the two functions are equivalent.
5996 @item ORIGIN(@var{memory})
5997 @kindex ORIGIN(@var{memory})
5998 Return the origin of the memory region named @var{memory}.
6000 @item SEGMENT_START(@var{segment}, @var{default})
6001 @kindex SEGMENT_START(@var{segment}, @var{default})
6002 Return the base address of the named @var{segment}. If an explicit
6003 value has already been given for this segment (with a command-line
6004 @samp{-T} option) then that value will be returned otherwise the value
6005 will be @var{default}. At present, the @samp{-T} command-line option
6006 can only be used to set the base address for the ``text'', ``data'', and
6007 ``bss'' sections, but you can use @code{SEGMENT_START} with any segment
6010 @item SIZEOF(@var{section})
6011 @kindex SIZEOF(@var{section})
6012 @cindex section size
6013 Return the size in bytes of the named @var{section}, if that section has
6014 been allocated. If the section has not been allocated when this is
6015 evaluated, the linker will report an error. In the following example,
6016 @code{symbol_1} and @code{symbol_2} are assigned identical values:
6025 symbol_1 = .end - .start ;
6026 symbol_2 = SIZEOF(.output);
6031 @item SIZEOF_HEADERS
6032 @itemx sizeof_headers
6033 @kindex SIZEOF_HEADERS
6035 Return the size in bytes of the output file's headers. This is
6036 information which appears at the start of the output file. You can use
6037 this number when setting the start address of the first section, if you
6038 choose, to facilitate paging.
6040 @cindex not enough room for program headers
6041 @cindex program headers, not enough room
6042 When producing an ELF output file, if the linker script uses the
6043 @code{SIZEOF_HEADERS} builtin function, the linker must compute the
6044 number of program headers before it has determined all the section
6045 addresses and sizes. If the linker later discovers that it needs
6046 additional program headers, it will report an error @samp{not enough
6047 room for program headers}. To avoid this error, you must avoid using
6048 the @code{SIZEOF_HEADERS} function, or you must rework your linker
6049 script to avoid forcing the linker to use additional program headers, or
6050 you must define the program headers yourself using the @code{PHDRS}
6051 command (@pxref{PHDRS}).
6054 @node Implicit Linker Scripts
6055 @section Implicit Linker Scripts
6056 @cindex implicit linker scripts
6057 If you specify a linker input file which the linker can not recognize as
6058 an object file or an archive file, it will try to read the file as a
6059 linker script. If the file can not be parsed as a linker script, the
6060 linker will report an error.
6062 An implicit linker script will not replace the default linker script.
6064 Typically an implicit linker script would contain only symbol
6065 assignments, or the @code{INPUT}, @code{GROUP}, or @code{VERSION}
6068 Any input files read because of an implicit linker script will be read
6069 at the position in the command line where the implicit linker script was
6070 read. This can affect archive searching.
6073 @node Machine Dependent
6074 @chapter Machine Dependent Features
6076 @cindex machine dependencies
6077 @command{ld} has additional features on some platforms; the following
6078 sections describe them. Machines where @command{ld} has no additional
6079 functionality are not listed.
6083 * H8/300:: @command{ld} and the H8/300
6086 * i960:: @command{ld} and the Intel 960 family
6089 * M68HC11/68HC12:: @code{ld} and the Motorola 68HC11 and 68HC12 families
6092 * ARM:: @command{ld} and the ARM family
6095 * HPPA ELF32:: @command{ld} and HPPA 32-bit ELF
6098 * M68K:: @command{ld} and the Motorola 68K family
6101 * MIPS:: @command{ld} and the MIPS family
6104 * MMIX:: @command{ld} and MMIX
6107 * MSP430:: @command{ld} and MSP430
6110 * NDS32:: @command{ld} and NDS32
6113 * Nios II:: @command{ld} and the Altera Nios II
6116 * PowerPC ELF32:: @command{ld} and PowerPC 32-bit ELF Support
6119 * PowerPC64 ELF64:: @command{ld} and PowerPC64 64-bit ELF Support
6122 * SPU ELF:: @command{ld} and SPU ELF Support
6125 * TI COFF:: @command{ld} and TI COFF
6128 * WIN32:: @command{ld} and WIN32 (cygwin/mingw)
6131 * Xtensa:: @command{ld} and Xtensa Processors
6142 @section @command{ld} and the H8/300
6144 @cindex H8/300 support
6145 For the H8/300, @command{ld} can perform these global optimizations when
6146 you specify the @samp{--relax} command-line option.
6149 @cindex relaxing on H8/300
6150 @item relaxing address modes
6151 @command{ld} finds all @code{jsr} and @code{jmp} instructions whose
6152 targets are within eight bits, and turns them into eight-bit
6153 program-counter relative @code{bsr} and @code{bra} instructions,
6156 @cindex synthesizing on H8/300
6157 @item synthesizing instructions
6158 @c FIXME: specifically mov.b, or any mov instructions really? -> mov.b only, at least on H8, H8H, H8S
6159 @command{ld} finds all @code{mov.b} instructions which use the
6160 sixteen-bit absolute address form, but refer to the top
6161 page of memory, and changes them to use the eight-bit address form.
6162 (That is: the linker turns @samp{mov.b @code{@@}@var{aa}:16} into
6163 @samp{mov.b @code{@@}@var{aa}:8} whenever the address @var{aa} is in the
6164 top page of memory).
6166 @command{ld} finds all @code{mov} instructions which use the register
6167 indirect with 32-bit displacement addressing mode, but use a small
6168 displacement inside 16-bit displacement range, and changes them to use
6169 the 16-bit displacement form. (That is: the linker turns @samp{mov.b
6170 @code{@@}@var{d}:32,ERx} into @samp{mov.b @code{@@}@var{d}:16,ERx}
6171 whenever the displacement @var{d} is in the 16 bit signed integer
6172 range. Only implemented in ELF-format ld).
6174 @item bit manipulation instructions
6175 @command{ld} finds all bit manipulation instructions like @code{band, bclr,
6176 biand, bild, bior, bist, bixor, bld, bnot, bor, bset, bst, btst, bxor}
6177 which use 32 bit and 16 bit absolute address form, but refer to the top
6178 page of memory, and changes them to use the 8 bit address form.
6179 (That is: the linker turns @samp{bset #xx:3,@code{@@}@var{aa}:32} into
6180 @samp{bset #xx:3,@code{@@}@var{aa}:8} whenever the address @var{aa} is in
6181 the top page of memory).
6183 @item system control instructions
6184 @command{ld} finds all @code{ldc.w, stc.w} instructions which use the
6185 32 bit absolute address form, but refer to the top page of memory, and
6186 changes them to use 16 bit address form.
6187 (That is: the linker turns @samp{ldc.w @code{@@}@var{aa}:32,ccr} into
6188 @samp{ldc.w @code{@@}@var{aa}:16,ccr} whenever the address @var{aa} is in
6189 the top page of memory).
6199 @c This stuff is pointless to say unless you're especially concerned
6200 @c with Renesas chips; don't enable it for generic case, please.
6202 @chapter @command{ld} and Other Renesas Chips
6204 @command{ld} also supports the Renesas (formerly Hitachi) H8/300H,
6205 H8/500, and SH chips. No special features, commands, or command-line
6206 options are required for these chips.
6216 @section @command{ld} and the Intel 960 Family
6218 @cindex i960 support
6220 You can use the @samp{-A@var{architecture}} command line option to
6221 specify one of the two-letter names identifying members of the 960
6222 family; the option specifies the desired output target, and warns of any
6223 incompatible instructions in the input files. It also modifies the
6224 linker's search strategy for archive libraries, to support the use of
6225 libraries specific to each particular architecture, by including in the
6226 search loop names suffixed with the string identifying the architecture.
6228 For example, if your @command{ld} command line included @w{@samp{-ACA}} as
6229 well as @w{@samp{-ltry}}, the linker would look (in its built-in search
6230 paths, and in any paths you specify with @samp{-L}) for a library with
6243 The first two possibilities would be considered in any event; the last
6244 two are due to the use of @w{@samp{-ACA}}.
6246 You can meaningfully use @samp{-A} more than once on a command line, since
6247 the 960 architecture family allows combination of target architectures; each
6248 use will add another pair of name variants to search for when @w{@samp{-l}}
6249 specifies a library.
6251 @cindex @option{--relax} on i960
6252 @cindex relaxing on i960
6253 @command{ld} supports the @samp{--relax} option for the i960 family. If
6254 you specify @samp{--relax}, @command{ld} finds all @code{balx} and
6255 @code{calx} instructions whose targets are within 24 bits, and turns
6256 them into 24-bit program-counter relative @code{bal} and @code{cal}
6257 instructions, respectively. @command{ld} also turns @code{cal}
6258 instructions into @code{bal} instructions when it determines that the
6259 target subroutine is a leaf routine (that is, the target subroutine does
6260 not itself call any subroutines).
6262 @cindex Cortex-A8 erratum workaround
6263 @kindex --fix-cortex-a8
6264 @kindex --no-fix-cortex-a8
6265 The @samp{--fix-cortex-a8} switch enables a link-time workaround for an erratum in certain Cortex-A8 processors. The workaround is enabled by default if you are targeting the ARM v7-A architecture profile. It can be enabled otherwise by specifying @samp{--fix-cortex-a8}, or disabled unconditionally by specifying @samp{--no-fix-cortex-a8}.
6267 The erratum only affects Thumb-2 code. Please contact ARM for further details.
6269 @kindex --merge-exidx-entries
6270 @kindex --no-merge-exidx-entries
6271 The @samp{--no-merge-exidx-entries} switch disables the merging of adjacent exidx entries in debuginfo.
6288 @node M68HC11/68HC12
6289 @section @command{ld} and the Motorola 68HC11 and 68HC12 families
6291 @cindex M68HC11 and 68HC12 support
6293 @subsection Linker Relaxation
6295 For the Motorola 68HC11, @command{ld} can perform these global
6296 optimizations when you specify the @samp{--relax} command-line option.
6299 @cindex relaxing on M68HC11
6300 @item relaxing address modes
6301 @command{ld} finds all @code{jsr} and @code{jmp} instructions whose
6302 targets are within eight bits, and turns them into eight-bit
6303 program-counter relative @code{bsr} and @code{bra} instructions,
6306 @command{ld} also looks at all 16-bit extended addressing modes and
6307 transforms them in a direct addressing mode when the address is in
6308 page 0 (between 0 and 0x0ff).
6310 @item relaxing gcc instruction group
6311 When @command{gcc} is called with @option{-mrelax}, it can emit group
6312 of instructions that the linker can optimize to use a 68HC11 direct
6313 addressing mode. These instructions consists of @code{bclr} or
6314 @code{bset} instructions.
6318 @subsection Trampoline Generation
6320 @cindex trampoline generation on M68HC11
6321 @cindex trampoline generation on M68HC12
6322 For 68HC11 and 68HC12, @command{ld} can generate trampoline code to
6323 call a far function using a normal @code{jsr} instruction. The linker
6324 will also change the relocation to some far function to use the
6325 trampoline address instead of the function address. This is typically the
6326 case when a pointer to a function is taken. The pointer will in fact
6327 point to the function trampoline.
6335 @section @command{ld} and the ARM family
6337 @cindex ARM interworking support
6338 @kindex --support-old-code
6339 For the ARM, @command{ld} will generate code stubs to allow functions calls
6340 between ARM and Thumb code. These stubs only work with code that has
6341 been compiled and assembled with the @samp{-mthumb-interwork} command
6342 line option. If it is necessary to link with old ARM object files or
6343 libraries, which have not been compiled with the -mthumb-interwork
6344 option then the @samp{--support-old-code} command line switch should be
6345 given to the linker. This will make it generate larger stub functions
6346 which will work with non-interworking aware ARM code. Note, however,
6347 the linker does not support generating stubs for function calls to
6348 non-interworking aware Thumb code.
6350 @cindex thumb entry point
6351 @cindex entry point, thumb
6352 @kindex --thumb-entry=@var{entry}
6353 The @samp{--thumb-entry} switch is a duplicate of the generic
6354 @samp{--entry} switch, in that it sets the program's starting address.
6355 But it also sets the bottom bit of the address, so that it can be
6356 branched to using a BX instruction, and the program will start
6357 executing in Thumb mode straight away.
6359 @cindex PE import table prefixing
6360 @kindex --use-nul-prefixed-import-tables
6361 The @samp{--use-nul-prefixed-import-tables} switch is specifying, that
6362 the import tables idata4 and idata5 have to be generated with a zero
6363 element prefix for import libraries. This is the old style to generate
6364 import tables. By default this option is turned off.
6368 The @samp{--be8} switch instructs @command{ld} to generate BE8 format
6369 executables. This option is only valid when linking big-endian objects.
6370 The resulting image will contain big-endian data and little-endian code.
6373 @kindex --target1-rel
6374 @kindex --target1-abs
6375 The @samp{R_ARM_TARGET1} relocation is typically used for entries in the
6376 @samp{.init_array} section. It is interpreted as either @samp{R_ARM_REL32}
6377 or @samp{R_ARM_ABS32}, depending on the target. The @samp{--target1-rel}
6378 and @samp{--target1-abs} switches override the default.
6381 @kindex --target2=@var{type}
6382 The @samp{--target2=type} switch overrides the default definition of the
6383 @samp{R_ARM_TARGET2} relocation. Valid values for @samp{type}, their
6384 meanings, and target defaults are as follows:
6387 @samp{R_ARM_REL32} (arm*-*-elf, arm*-*-eabi)
6389 @samp{R_ARM_ABS32} (arm*-*-symbianelf)
6391 @samp{R_ARM_GOT_PREL} (arm*-*-linux, arm*-*-*bsd)
6396 The @samp{R_ARM_V4BX} relocation (defined by the ARM AAELF
6397 specification) enables objects compiled for the ARMv4 architecture to be
6398 interworking-safe when linked with other objects compiled for ARMv4t, but
6399 also allows pure ARMv4 binaries to be built from the same ARMv4 objects.
6401 In the latter case, the switch @option{--fix-v4bx} must be passed to the
6402 linker, which causes v4t @code{BX rM} instructions to be rewritten as
6403 @code{MOV PC,rM}, since v4 processors do not have a @code{BX} instruction.
6405 In the former case, the switch should not be used, and @samp{R_ARM_V4BX}
6406 relocations are ignored.
6408 @cindex FIX_V4BX_INTERWORKING
6409 @kindex --fix-v4bx-interworking
6410 Replace @code{BX rM} instructions identified by @samp{R_ARM_V4BX}
6411 relocations with a branch to the following veneer:
6419 This allows generation of libraries/applications that work on ARMv4 cores
6420 and are still interworking safe. Note that the above veneer clobbers the
6421 condition flags, so may cause incorrect program behavior in rare cases.
6425 The @samp{--use-blx} switch enables the linker to use ARM/Thumb
6426 BLX instructions (available on ARMv5t and above) in various
6427 situations. Currently it is used to perform calls via the PLT from Thumb
6428 code using BLX rather than using BX and a mode-switching stub before
6429 each PLT entry. This should lead to such calls executing slightly faster.
6431 This option is enabled implicitly for SymbianOS, so there is no need to
6432 specify it if you are using that target.
6434 @cindex VFP11_DENORM_FIX
6435 @kindex --vfp11-denorm-fix
6436 The @samp{--vfp11-denorm-fix} switch enables a link-time workaround for a
6437 bug in certain VFP11 coprocessor hardware, which sometimes allows
6438 instructions with denorm operands (which must be handled by support code)
6439 to have those operands overwritten by subsequent instructions before
6440 the support code can read the intended values.
6442 The bug may be avoided in scalar mode if you allow at least one
6443 intervening instruction between a VFP11 instruction which uses a register
6444 and another instruction which writes to the same register, or at least two
6445 intervening instructions if vector mode is in use. The bug only affects
6446 full-compliance floating-point mode: you do not need this workaround if
6447 you are using "runfast" mode. Please contact ARM for further details.
6449 If you know you are using buggy VFP11 hardware, you can
6450 enable this workaround by specifying the linker option
6451 @samp{--vfp-denorm-fix=scalar} if you are using the VFP11 scalar
6452 mode only, or @samp{--vfp-denorm-fix=vector} if you are using
6453 vector mode (the latter also works for scalar code). The default is
6454 @samp{--vfp-denorm-fix=none}.
6456 If the workaround is enabled, instructions are scanned for
6457 potentially-troublesome sequences, and a veneer is created for each
6458 such sequence which may trigger the erratum. The veneer consists of the
6459 first instruction of the sequence and a branch back to the subsequent
6460 instruction. The original instruction is then replaced with a branch to
6461 the veneer. The extra cycles required to call and return from the veneer
6462 are sufficient to avoid the erratum in both the scalar and vector cases.
6464 @cindex ARM1176 erratum workaround
6465 @kindex --fix-arm1176
6466 @kindex --no-fix-arm1176
6467 The @samp{--fix-arm1176} switch enables a link-time workaround for an erratum
6468 in certain ARM1176 processors. The workaround is enabled by default if you
6469 are targeting ARM v6 (excluding ARM v6T2) or earlier. It can be disabled
6470 unconditionally by specifying @samp{--no-fix-arm1176}.
6472 Further information is available in the ``ARM1176JZ-S and ARM1176JZF-S
6473 Programmer Advice Notice'' available on the ARM documentation website at:
6474 http://infocenter.arm.com/.
6476 @cindex NO_ENUM_SIZE_WARNING
6477 @kindex --no-enum-size-warning
6478 The @option{--no-enum-size-warning} switch prevents the linker from
6479 warning when linking object files that specify incompatible EABI
6480 enumeration size attributes. For example, with this switch enabled,
6481 linking of an object file using 32-bit enumeration values with another
6482 using enumeration values fitted into the smallest possible space will
6485 @cindex NO_WCHAR_SIZE_WARNING
6486 @kindex --no-wchar-size-warning
6487 The @option{--no-wchar-size-warning} switch prevents the linker from
6488 warning when linking object files that specify incompatible EABI
6489 @code{wchar_t} size attributes. For example, with this switch enabled,
6490 linking of an object file using 32-bit @code{wchar_t} values with another
6491 using 16-bit @code{wchar_t} values will not be diagnosed.
6494 @kindex --pic-veneer
6495 The @samp{--pic-veneer} switch makes the linker use PIC sequences for
6496 ARM/Thumb interworking veneers, even if the rest of the binary
6497 is not PIC. This avoids problems on uClinux targets where
6498 @samp{--emit-relocs} is used to generate relocatable binaries.
6500 @cindex STUB_GROUP_SIZE
6501 @kindex --stub-group-size=@var{N}
6502 The linker will automatically generate and insert small sequences of
6503 code into a linked ARM ELF executable whenever an attempt is made to
6504 perform a function call to a symbol that is too far away. The
6505 placement of these sequences of instructions - called stubs - is
6506 controlled by the command line option @option{--stub-group-size=N}.
6507 The placement is important because a poor choice can create a need for
6508 duplicate stubs, increasing the code size. The linker will try to
6509 group stubs together in order to reduce interruptions to the flow of
6510 code, but it needs guidance as to how big these groups should be and
6511 where they should be placed.
6513 The value of @samp{N}, the parameter to the
6514 @option{--stub-group-size=} option controls where the stub groups are
6515 placed. If it is negative then all stubs are placed after the first
6516 branch that needs them. If it is positive then the stubs can be
6517 placed either before or after the branches that need them. If the
6518 value of @samp{N} is 1 (either +1 or -1) then the linker will choose
6519 exactly where to place groups of stubs, using its built in heuristics.
6520 A value of @samp{N} greater than 1 (or smaller than -1) tells the
6521 linker that a single group of stubs can service at most @samp{N} bytes
6522 from the input sections.
6524 The default, if @option{--stub-group-size=} is not specified, is
6527 Farcalls stubs insertion is fully supported for the ARM-EABI target
6528 only, because it relies on object files properties not present
6542 @section @command{ld} and HPPA 32-bit ELF Support
6543 @cindex HPPA multiple sub-space stubs
6544 @kindex --multi-subspace
6545 When generating a shared library, @command{ld} will by default generate
6546 import stubs suitable for use with a single sub-space application.
6547 The @samp{--multi-subspace} switch causes @command{ld} to generate export
6548 stubs, and different (larger) import stubs suitable for use with
6549 multiple sub-spaces.
6551 @cindex HPPA stub grouping
6552 @kindex --stub-group-size=@var{N}
6553 Long branch stubs and import/export stubs are placed by @command{ld} in
6554 stub sections located between groups of input sections.
6555 @samp{--stub-group-size} specifies the maximum size of a group of input
6556 sections handled by one stub section. Since branch offsets are signed,
6557 a stub section may serve two groups of input sections, one group before
6558 the stub section, and one group after it. However, when using
6559 conditional branches that require stubs, it may be better (for branch
6560 prediction) that stub sections only serve one group of input sections.
6561 A negative value for @samp{N} chooses this scheme, ensuring that
6562 branches to stubs always use a negative offset. Two special values of
6563 @samp{N} are recognized, @samp{1} and @samp{-1}. These both instruct
6564 @command{ld} to automatically size input section groups for the branch types
6565 detected, with the same behaviour regarding stub placement as other
6566 positive or negative values of @samp{N} respectively.
6568 Note that @samp{--stub-group-size} does not split input sections. A
6569 single input section larger than the group size specified will of course
6570 create a larger group (of one section). If input sections are too
6571 large, it may not be possible for a branch to reach its stub.
6584 @section @command{ld} and the Motorola 68K family
6586 @cindex Motorola 68K GOT generation
6587 @kindex --got=@var{type}
6588 The @samp{--got=@var{type}} option lets you choose the GOT generation scheme.
6589 The choices are @samp{single}, @samp{negative}, @samp{multigot} and
6590 @samp{target}. When @samp{target} is selected the linker chooses
6591 the default GOT generation scheme for the current target.
6592 @samp{single} tells the linker to generate a single GOT with
6593 entries only at non-negative offsets.
6594 @samp{negative} instructs the linker to generate a single GOT with
6595 entries at both negative and positive offsets. Not all environments
6597 @samp{multigot} allows the linker to generate several GOTs in the
6598 output file. All GOT references from a single input object
6599 file access the same GOT, but references from different input object
6600 files might access different GOTs. Not all environments support such GOTs.
6613 @section @command{ld} and the MIPS family
6615 @cindex MIPS microMIPS instruction choice selection
6618 The @samp{--insn32} and @samp{--no-insn32} options control the choice of
6619 microMIPS instructions used in code generated by the linker, such as that
6620 in the PLT or lazy binding stubs, or in relaxation. If @samp{--insn32} is
6621 used, then the linker only uses 32-bit instruction encodings. By default
6622 or if @samp{--no-insn32} is used, all instruction encodings are used,
6623 including 16-bit ones where possible.
6636 @section @code{ld} and MMIX
6637 For MMIX, there is a choice of generating @code{ELF} object files or
6638 @code{mmo} object files when linking. The simulator @code{mmix}
6639 understands the @code{mmo} format. The binutils @code{objcopy} utility
6640 can translate between the two formats.
6642 There is one special section, the @samp{.MMIX.reg_contents} section.
6643 Contents in this section is assumed to correspond to that of global
6644 registers, and symbols referring to it are translated to special symbols,
6645 equal to registers. In a final link, the start address of the
6646 @samp{.MMIX.reg_contents} section corresponds to the first allocated
6647 global register multiplied by 8. Register @code{$255} is not included in
6648 this section; it is always set to the program entry, which is at the
6649 symbol @code{Main} for @code{mmo} files.
6651 Global symbols with the prefix @code{__.MMIX.start.}, for example
6652 @code{__.MMIX.start..text} and @code{__.MMIX.start..data} are special.
6653 The default linker script uses these to set the default start address
6656 Initial and trailing multiples of zero-valued 32-bit words in a section,
6657 are left out from an mmo file.
6670 @section @code{ld} and MSP430
6671 For the MSP430 it is possible to select the MPU architecture. The flag @samp{-m [mpu type]}
6672 will select an appropriate linker script for selected MPU type. (To get a list of known MPUs
6673 just pass @samp{-m help} option to the linker).
6675 @cindex MSP430 extra sections
6676 The linker will recognize some extra sections which are MSP430 specific:
6679 @item @samp{.vectors}
6680 Defines a portion of ROM where interrupt vectors located.
6682 @item @samp{.bootloader}
6683 Defines the bootloader portion of the ROM (if applicable). Any code
6684 in this section will be uploaded to the MPU.
6686 @item @samp{.infomem}
6687 Defines an information memory section (if applicable). Any code in
6688 this section will be uploaded to the MPU.
6690 @item @samp{.infomemnobits}
6691 This is the same as the @samp{.infomem} section except that any code
6692 in this section will not be uploaded to the MPU.
6694 @item @samp{.noinit}
6695 Denotes a portion of RAM located above @samp{.bss} section.
6697 The last two sections are used by gcc.
6711 @section @code{ld} and NDS32
6712 @kindex relaxing on NDS32
6713 For NDS32, there are some options to select relaxation behavior. The linker
6714 relaxes objects according to these options.
6717 @item @samp{--m[no-]fp-as-gp}
6718 Disable/enable fp-as-gp relaxation.
6720 @item @samp{--mexport-symbols=FILE}
6721 Exporting symbols and their address into FILE as linker script.
6723 @item @samp{--m[no-]ex9}
6724 Disable/enable link-time EX9 relaxation.
6726 @item @samp{--mexport-ex9=FILE}
6727 Export the EX9 table after linking.
6729 @item @samp{--mimport-ex9=FILE}
6730 Import the Ex9 table for EX9 relaxation.
6732 @item @samp{--mupdate-ex9}
6733 Update the existing EX9 table.
6735 @item @samp{--mex9-limit=NUM}
6736 Maximum number of entries in the ex9 table.
6738 @item @samp{--mex9-loop-aware}
6739 Avoid generating the EX9 instruction inside the loop.
6741 @item @samp{--m[no-]ifc}
6742 Disable/enable the link-time IFC optimization.
6744 @item @samp{--mifc-loop-aware}
6745 Avoid generating the IFC instruction inside the loop.
6759 @section @command{ld} and the Altera Nios II
6760 @cindex Nios II call relaxation
6761 @kindex --relax on Nios II
6763 Call and immediate jump instructions on Nios II processors are limited to
6764 transferring control to addresses in the same 256MB memory segment,
6765 which may result in @command{ld} giving
6766 @samp{relocation truncated to fit} errors with very large programs.
6767 The command-line option @option{--relax} enables the generation of
6768 trampolines that can access the entire 32-bit address space for calls
6769 outside the normal @code{call} and @code{jmpi} address range. These
6770 trampolines are inserted at section boundaries, so may not themselves
6771 be reachable if an input section and its associated call trampolines are
6774 The @option{--relax} option is enabled by default unless @option{-r}
6775 is also specified. You can disable trampoline generation by using the
6776 @option{--no-relax} linker option. You can also disable this optimization
6777 locally by using the @samp{set .noat} directive in assembly-language
6778 source files, as the linker-inserted trampolines use the @code{at}
6779 register as a temporary.
6781 Note that the linker @option{--relax} option is independent of assembler
6782 relaxation options, and that using the GNU assembler's @option{-relax-all}
6783 option interferes with the linker's more selective call instruction relaxation.
6796 @section @command{ld} and PowerPC 32-bit ELF Support
6797 @cindex PowerPC long branches
6798 @kindex --relax on PowerPC
6799 Branches on PowerPC processors are limited to a signed 26-bit
6800 displacement, which may result in @command{ld} giving
6801 @samp{relocation truncated to fit} errors with very large programs.
6802 @samp{--relax} enables the generation of trampolines that can access
6803 the entire 32-bit address space. These trampolines are inserted at
6804 section boundaries, so may not themselves be reachable if an input
6805 section exceeds 33M in size. You may combine @samp{-r} and
6806 @samp{--relax} to add trampolines in a partial link. In that case
6807 both branches to undefined symbols and inter-section branches are also
6808 considered potentially out of range, and trampolines inserted.
6810 @cindex PowerPC ELF32 options
6815 Current PowerPC GCC accepts a @samp{-msecure-plt} option that
6816 generates code capable of using a newer PLT and GOT layout that has
6817 the security advantage of no executable section ever needing to be
6818 writable and no writable section ever being executable. PowerPC
6819 @command{ld} will generate this layout, including stubs to access the
6820 PLT, if all input files (including startup and static libraries) were
6821 compiled with @samp{-msecure-plt}. @samp{--bss-plt} forces the old
6822 BSS PLT (and GOT layout) which can give slightly better performance.
6824 @kindex --secure-plt
6826 @command{ld} will use the new PLT and GOT layout if it is linking new
6827 @samp{-fpic} or @samp{-fPIC} code, but does not do so automatically
6828 when linking non-PIC code. This option requests the new PLT and GOT
6829 layout. A warning will be given if some object file requires the old
6835 The new secure PLT and GOT are placed differently relative to other
6836 sections compared to older BSS PLT and GOT placement. The location of
6837 @code{.plt} must change because the new secure PLT is an initialized
6838 section while the old PLT is uninitialized. The reason for the
6839 @code{.got} change is more subtle: The new placement allows
6840 @code{.got} to be read-only in applications linked with
6841 @samp{-z relro -z now}. However, this placement means that
6842 @code{.sdata} cannot always be used in shared libraries, because the
6843 PowerPC ABI accesses @code{.sdata} in shared libraries from the GOT
6844 pointer. @samp{--sdata-got} forces the old GOT placement. PowerPC
6845 GCC doesn't use @code{.sdata} in shared libraries, so this option is
6846 really only useful for other compilers that may do so.
6848 @cindex PowerPC stub symbols
6849 @kindex --emit-stub-syms
6850 @item --emit-stub-syms
6851 This option causes @command{ld} to label linker stubs with a local
6852 symbol that encodes the stub type and destination.
6854 @cindex PowerPC TLS optimization
6855 @kindex --no-tls-optimize
6856 @item --no-tls-optimize
6857 PowerPC @command{ld} normally performs some optimization of code
6858 sequences used to access Thread-Local Storage. Use this option to
6859 disable the optimization.
6872 @node PowerPC64 ELF64
6873 @section @command{ld} and PowerPC64 64-bit ELF Support
6875 @cindex PowerPC64 ELF64 options
6877 @cindex PowerPC64 stub grouping
6878 @kindex --stub-group-size
6879 @item --stub-group-size
6880 Long branch stubs, PLT call stubs and TOC adjusting stubs are placed
6881 by @command{ld} in stub sections located between groups of input sections.
6882 @samp{--stub-group-size} specifies the maximum size of a group of input
6883 sections handled by one stub section. Since branch offsets are signed,
6884 a stub section may serve two groups of input sections, one group before
6885 the stub section, and one group after it. However, when using
6886 conditional branches that require stubs, it may be better (for branch
6887 prediction) that stub sections only serve one group of input sections.
6888 A negative value for @samp{N} chooses this scheme, ensuring that
6889 branches to stubs always use a negative offset. Two special values of
6890 @samp{N} are recognized, @samp{1} and @samp{-1}. These both instruct
6891 @command{ld} to automatically size input section groups for the branch types
6892 detected, with the same behaviour regarding stub placement as other
6893 positive or negative values of @samp{N} respectively.
6895 Note that @samp{--stub-group-size} does not split input sections. A
6896 single input section larger than the group size specified will of course
6897 create a larger group (of one section). If input sections are too
6898 large, it may not be possible for a branch to reach its stub.
6900 @cindex PowerPC64 stub symbols
6901 @kindex --emit-stub-syms
6902 @item --emit-stub-syms
6903 This option causes @command{ld} to label linker stubs with a local
6904 symbol that encodes the stub type and destination.
6906 @cindex PowerPC64 dot symbols
6908 @kindex --no-dotsyms
6909 @item --dotsyms, --no-dotsyms
6910 These two options control how @command{ld} interprets version patterns
6911 in a version script. Older PowerPC64 compilers emitted both a
6912 function descriptor symbol with the same name as the function, and a
6913 code entry symbol with the name prefixed by a dot (@samp{.}). To
6914 properly version a function @samp{foo}, the version script thus needs
6915 to control both @samp{foo} and @samp{.foo}. The option
6916 @samp{--dotsyms}, on by default, automatically adds the required
6917 dot-prefixed patterns. Use @samp{--no-dotsyms} to disable this
6920 @cindex PowerPC64 TLS optimization
6921 @kindex --no-tls-optimize
6922 @item --no-tls-optimize
6923 PowerPC64 @command{ld} normally performs some optimization of code
6924 sequences used to access Thread-Local Storage. Use this option to
6925 disable the optimization.
6927 @cindex PowerPC64 OPD optimization
6928 @kindex --no-opd-optimize
6929 @item --no-opd-optimize
6930 PowerPC64 @command{ld} normally removes @code{.opd} section entries
6931 corresponding to deleted link-once functions, or functions removed by
6932 the action of @samp{--gc-sections} or linker script @code{/DISCARD/}.
6933 Use this option to disable @code{.opd} optimization.
6935 @cindex PowerPC64 OPD spacing
6936 @kindex --non-overlapping-opd
6937 @item --non-overlapping-opd
6938 Some PowerPC64 compilers have an option to generate compressed
6939 @code{.opd} entries spaced 16 bytes apart, overlapping the third word,
6940 the static chain pointer (unused in C) with the first word of the next
6941 entry. This option expands such entries to the full 24 bytes.
6943 @cindex PowerPC64 TOC optimization
6944 @kindex --no-toc-optimize
6945 @item --no-toc-optimize
6946 PowerPC64 @command{ld} normally removes unused @code{.toc} section
6947 entries. Such entries are detected by examining relocations that
6948 reference the TOC in code sections. A reloc in a deleted code section
6949 marks a TOC word as unneeded, while a reloc in a kept code section
6950 marks a TOC word as needed. Since the TOC may reference itself, TOC
6951 relocs are also examined. TOC words marked as both needed and
6952 unneeded will of course be kept. TOC words without any referencing
6953 reloc are assumed to be part of a multi-word entry, and are kept or
6954 discarded as per the nearest marked preceding word. This works
6955 reliably for compiler generated code, but may be incorrect if assembly
6956 code is used to insert TOC entries. Use this option to disable the
6959 @cindex PowerPC64 multi-TOC
6960 @kindex --no-multi-toc
6961 @item --no-multi-toc
6962 If given any toc option besides @code{-mcmodel=medium} or
6963 @code{-mcmodel=large}, PowerPC64 GCC generates code for a TOC model
6965 entries are accessed with a 16-bit offset from r2. This limits the
6966 total TOC size to 64K. PowerPC64 @command{ld} extends this limit by
6967 grouping code sections such that each group uses less than 64K for its
6968 TOC entries, then inserts r2 adjusting stubs between inter-group
6969 calls. @command{ld} does not split apart input sections, so cannot
6970 help if a single input file has a @code{.toc} section that exceeds
6971 64K, most likely from linking multiple files with @command{ld -r}.
6972 Use this option to turn off this feature.
6974 @cindex PowerPC64 TOC sorting
6975 @kindex --no-toc-sort
6977 By default, @command{ld} sorts TOC sections so that those whose file
6978 happens to have a section called @code{.init} or @code{.fini} are
6979 placed first, followed by TOC sections referenced by code generated
6980 with PowerPC64 gcc's @code{-mcmodel=small}, and lastly TOC sections
6981 referenced only by code generated with PowerPC64 gcc's
6982 @code{-mcmodel=medium} or @code{-mcmodel=large} options. Doing this
6983 results in better TOC grouping for multi-TOC. Use this option to turn
6986 @cindex PowerPC64 PLT stub alignment
6988 @kindex --no-plt-align
6990 @itemx --no-plt-align
6991 Use these options to control whether individual PLT call stubs are
6992 aligned to a 32-byte boundary, or to the specified power of two
6993 boundary when using @code{--plt-align=}. By default PLT call stubs
6996 @cindex PowerPC64 PLT call stub static chain
6997 @kindex --plt-static-chain
6998 @kindex --no-plt-static-chain
6999 @item --plt-static-chain
7000 @itemx --no-plt-static-chain
7001 Use these options to control whether PLT call stubs load the static
7002 chain pointer (r11). @code{ld} defaults to not loading the static
7003 chain since there is never any need to do so on a PLT call.
7005 @cindex PowerPC64 PLT call stub thread safety
7006 @kindex --plt-thread-safe
7007 @kindex --no-plt-thread-safe
7008 @item --plt-thread-safe
7009 @itemx --no-thread-safe
7010 With power7's weakly ordered memory model, it is possible when using
7011 lazy binding for ld.so to update a plt entry in one thread and have
7012 another thread see the individual plt entry words update in the wrong
7013 order, despite ld.so carefully writing in the correct order and using
7014 memory write barriers. To avoid this we need some sort of read
7015 barrier in the call stub, or use LD_BIND_NOW=1. By default, @code{ld}
7016 looks for calls to commonly used functions that create threads, and if
7017 seen, adds the necessary barriers. Use these options to change the
7032 @section @command{ld} and SPU ELF Support
7034 @cindex SPU ELF options
7040 This option marks an executable as a PIC plugin module.
7042 @cindex SPU overlays
7043 @kindex --no-overlays
7045 Normally, @command{ld} recognizes calls to functions within overlay
7046 regions, and redirects such calls to an overlay manager via a stub.
7047 @command{ld} also provides a built-in overlay manager. This option
7048 turns off all this special overlay handling.
7050 @cindex SPU overlay stub symbols
7051 @kindex --emit-stub-syms
7052 @item --emit-stub-syms
7053 This option causes @command{ld} to label overlay stubs with a local
7054 symbol that encodes the stub type and destination.
7056 @cindex SPU extra overlay stubs
7057 @kindex --extra-overlay-stubs
7058 @item --extra-overlay-stubs
7059 This option causes @command{ld} to add overlay call stubs on all
7060 function calls out of overlay regions. Normally stubs are not added
7061 on calls to non-overlay regions.
7063 @cindex SPU local store size
7064 @kindex --local-store=lo:hi
7065 @item --local-store=lo:hi
7066 @command{ld} usually checks that a final executable for SPU fits in
7067 the address range 0 to 256k. This option may be used to change the
7068 range. Disable the check entirely with @option{--local-store=0:0}.
7071 @kindex --stack-analysis
7072 @item --stack-analysis
7073 SPU local store space is limited. Over-allocation of stack space
7074 unnecessarily limits space available for code and data, while
7075 under-allocation results in runtime failures. If given this option,
7076 @command{ld} will provide an estimate of maximum stack usage.
7077 @command{ld} does this by examining symbols in code sections to
7078 determine the extents of functions, and looking at function prologues
7079 for stack adjusting instructions. A call-graph is created by looking
7080 for relocations on branch instructions. The graph is then searched
7081 for the maximum stack usage path. Note that this analysis does not
7082 find calls made via function pointers, and does not handle recursion
7083 and other cycles in the call graph. Stack usage may be
7084 under-estimated if your code makes such calls. Also, stack usage for
7085 dynamic allocation, e.g. alloca, will not be detected. If a link map
7086 is requested, detailed information about each function's stack usage
7087 and calls will be given.
7090 @kindex --emit-stack-syms
7091 @item --emit-stack-syms
7092 This option, if given along with @option{--stack-analysis} will result
7093 in @command{ld} emitting stack sizing symbols for each function.
7094 These take the form @code{__stack_<function_name>} for global
7095 functions, and @code{__stack_<number>_<function_name>} for static
7096 functions. @code{<number>} is the section id in hex. The value of
7097 such symbols is the stack requirement for the corresponding function.
7098 The symbol size will be zero, type @code{STT_NOTYPE}, binding
7099 @code{STB_LOCAL}, and section @code{SHN_ABS}.
7113 @section @command{ld}'s Support for Various TI COFF Versions
7114 @cindex TI COFF versions
7115 @kindex --format=@var{version}
7116 The @samp{--format} switch allows selection of one of the various
7117 TI COFF versions. The latest of this writing is 2; versions 0 and 1 are
7118 also supported. The TI COFF versions also vary in header byte-order
7119 format; @command{ld} will read any version or byte order, but the output
7120 header format depends on the default specified by the specific target.
7133 @section @command{ld} and WIN32 (cygwin/mingw)
7135 This section describes some of the win32 specific @command{ld} issues.
7136 See @ref{Options,,Command Line Options} for detailed description of the
7137 command line options mentioned here.
7140 @cindex import libraries
7141 @item import libraries
7142 The standard Windows linker creates and uses so-called import
7143 libraries, which contains information for linking to dll's. They are
7144 regular static archives and are handled as any other static
7145 archive. The cygwin and mingw ports of @command{ld} have specific
7146 support for creating such libraries provided with the
7147 @samp{--out-implib} command line option.
7149 @item exporting DLL symbols
7150 @cindex exporting DLL symbols
7151 The cygwin/mingw @command{ld} has several ways to export symbols for dll's.
7154 @item using auto-export functionality
7155 @cindex using auto-export functionality
7156 By default @command{ld} exports symbols with the auto-export functionality,
7157 which is controlled by the following command line options:
7160 @item --export-all-symbols [This is the default]
7161 @item --exclude-symbols
7162 @item --exclude-libs
7163 @item --exclude-modules-for-implib
7164 @item --version-script
7167 When auto-export is in operation, @command{ld} will export all the non-local
7168 (global and common) symbols it finds in a DLL, with the exception of a few
7169 symbols known to belong to the system's runtime and libraries. As it will
7170 often not be desirable to export all of a DLL's symbols, which may include
7171 private functions that are not part of any public interface, the command-line
7172 options listed above may be used to filter symbols out from the list for
7173 exporting. The @samp{--output-def} option can be used in order to see the
7174 final list of exported symbols with all exclusions taken into effect.
7176 If @samp{--export-all-symbols} is not given explicitly on the
7177 command line, then the default auto-export behavior will be @emph{disabled}
7178 if either of the following are true:
7181 @item A DEF file is used.
7182 @item Any symbol in any object file was marked with the __declspec(dllexport) attribute.
7185 @item using a DEF file
7186 @cindex using a DEF file
7187 Another way of exporting symbols is using a DEF file. A DEF file is
7188 an ASCII file containing definitions of symbols which should be
7189 exported when a dll is created. Usually it is named @samp{<dll
7190 name>.def} and is added as any other object file to the linker's
7191 command line. The file's name must end in @samp{.def} or @samp{.DEF}.
7194 gcc -o <output> <objectfiles> <dll name>.def
7197 Using a DEF file turns off the normal auto-export behavior, unless the
7198 @samp{--export-all-symbols} option is also used.
7200 Here is an example of a DEF file for a shared library called @samp{xyz.dll}:
7203 LIBRARY "xyz.dll" BASE=0x20000000
7209 another_foo = abc.dll.afoo
7215 This example defines a DLL with a non-default base address and seven
7216 symbols in the export table. The third exported symbol @code{_bar} is an
7217 alias for the second. The fourth symbol, @code{another_foo} is resolved
7218 by "forwarding" to another module and treating it as an alias for
7219 @code{afoo} exported from the DLL @samp{abc.dll}. The final symbol
7220 @code{var1} is declared to be a data object. The @samp{doo} symbol in
7221 export library is an alias of @samp{foo}, which gets the string name
7222 in export table @samp{foo2}. The @samp{eoo} symbol is an data export
7223 symbol, which gets in export table the name @samp{var1}.
7225 The optional @code{LIBRARY <name>} command indicates the @emph{internal}
7226 name of the output DLL. If @samp{<name>} does not include a suffix,
7227 the default library suffix, @samp{.DLL} is appended.
7229 When the .DEF file is used to build an application, rather than a
7230 library, the @code{NAME <name>} command should be used instead of
7231 @code{LIBRARY}. If @samp{<name>} does not include a suffix, the default
7232 executable suffix, @samp{.EXE} is appended.
7234 With either @code{LIBRARY <name>} or @code{NAME <name>} the optional
7235 specification @code{BASE = <number>} may be used to specify a
7236 non-default base address for the image.
7238 If neither @code{LIBRARY <name>} nor @code{NAME <name>} is specified,
7239 or they specify an empty string, the internal name is the same as the
7240 filename specified on the command line.
7242 The complete specification of an export symbol is:
7246 ( ( ( <name1> [ = <name2> ] )
7247 | ( <name1> = <module-name> . <external-name>))
7248 [ @@ <integer> ] [NONAME] [DATA] [CONSTANT] [PRIVATE] [== <name3>] ) *
7251 Declares @samp{<name1>} as an exported symbol from the DLL, or declares
7252 @samp{<name1>} as an exported alias for @samp{<name2>}; or declares
7253 @samp{<name1>} as a "forward" alias for the symbol
7254 @samp{<external-name>} in the DLL @samp{<module-name>}.
7255 Optionally, the symbol may be exported by the specified ordinal
7256 @samp{<integer>} alias. The optional @samp{<name3>} is the to be used
7257 string in import/export table for the symbol.
7259 The optional keywords that follow the declaration indicate:
7261 @code{NONAME}: Do not put the symbol name in the DLL's export table. It
7262 will still be exported by its ordinal alias (either the value specified
7263 by the .def specification or, otherwise, the value assigned by the
7264 linker). The symbol name, however, does remain visible in the import
7265 library (if any), unless @code{PRIVATE} is also specified.
7267 @code{DATA}: The symbol is a variable or object, rather than a function.
7268 The import lib will export only an indirect reference to @code{foo} as
7269 the symbol @code{_imp__foo} (ie, @code{foo} must be resolved as
7272 @code{CONSTANT}: Like @code{DATA}, but put the undecorated @code{foo} as
7273 well as @code{_imp__foo} into the import library. Both refer to the
7274 read-only import address table's pointer to the variable, not to the
7275 variable itself. This can be dangerous. If the user code fails to add
7276 the @code{dllimport} attribute and also fails to explicitly add the
7277 extra indirection that the use of the attribute enforces, the
7278 application will behave unexpectedly.
7280 @code{PRIVATE}: Put the symbol in the DLL's export table, but do not put
7281 it into the static import library used to resolve imports at link time. The
7282 symbol can still be imported using the @code{LoadLibrary/GetProcAddress}
7283 API at runtime or by by using the GNU ld extension of linking directly to
7284 the DLL without an import library.
7286 See ld/deffilep.y in the binutils sources for the full specification of
7287 other DEF file statements
7289 @cindex creating a DEF file
7290 While linking a shared dll, @command{ld} is able to create a DEF file
7291 with the @samp{--output-def <file>} command line option.
7293 @item Using decorations
7294 @cindex Using decorations
7295 Another way of marking symbols for export is to modify the source code
7296 itself, so that when building the DLL each symbol to be exported is
7300 __declspec(dllexport) int a_variable
7301 __declspec(dllexport) void a_function(int with_args)
7304 All such symbols will be exported from the DLL. If, however,
7305 any of the object files in the DLL contain symbols decorated in
7306 this way, then the normal auto-export behavior is disabled, unless
7307 the @samp{--export-all-symbols} option is also used.
7309 Note that object files that wish to access these symbols must @emph{not}
7310 decorate them with dllexport. Instead, they should use dllimport,
7314 __declspec(dllimport) int a_variable
7315 __declspec(dllimport) void a_function(int with_args)
7318 This complicates the structure of library header files, because
7319 when included by the library itself the header must declare the
7320 variables and functions as dllexport, but when included by client
7321 code the header must declare them as dllimport. There are a number
7322 of idioms that are typically used to do this; often client code can
7323 omit the __declspec() declaration completely. See
7324 @samp{--enable-auto-import} and @samp{automatic data imports} for more
7328 @cindex automatic data imports
7329 @item automatic data imports
7330 The standard Windows dll format supports data imports from dlls only
7331 by adding special decorations (dllimport/dllexport), which let the
7332 compiler produce specific assembler instructions to deal with this
7333 issue. This increases the effort necessary to port existing Un*x
7334 code to these platforms, especially for large
7335 c++ libraries and applications. The auto-import feature, which was
7336 initially provided by Paul Sokolovsky, allows one to omit the
7337 decorations to achieve a behavior that conforms to that on POSIX/Un*x
7338 platforms. This feature is enabled with the @samp{--enable-auto-import}
7339 command-line option, although it is enabled by default on cygwin/mingw.
7340 The @samp{--enable-auto-import} option itself now serves mainly to
7341 suppress any warnings that are ordinarily emitted when linked objects
7342 trigger the feature's use.
7344 auto-import of variables does not always work flawlessly without
7345 additional assistance. Sometimes, you will see this message
7347 "variable '<var>' can't be auto-imported. Please read the
7348 documentation for ld's @code{--enable-auto-import} for details."
7350 The @samp{--enable-auto-import} documentation explains why this error
7351 occurs, and several methods that can be used to overcome this difficulty.
7352 One of these methods is the @emph{runtime pseudo-relocs} feature, described
7355 @cindex runtime pseudo-relocation
7356 For complex variables imported from DLLs (such as structs or classes),
7357 object files typically contain a base address for the variable and an
7358 offset (@emph{addend}) within the variable--to specify a particular
7359 field or public member, for instance. Unfortunately, the runtime loader used
7360 in win32 environments is incapable of fixing these references at runtime
7361 without the additional information supplied by dllimport/dllexport decorations.
7362 The standard auto-import feature described above is unable to resolve these
7365 The @samp{--enable-runtime-pseudo-relocs} switch allows these references to
7366 be resolved without error, while leaving the task of adjusting the references
7367 themselves (with their non-zero addends) to specialized code provided by the
7368 runtime environment. Recent versions of the cygwin and mingw environments and
7369 compilers provide this runtime support; older versions do not. However, the
7370 support is only necessary on the developer's platform; the compiled result will
7371 run without error on an older system.
7373 @samp{--enable-runtime-pseudo-relocs} is not the default; it must be explicitly
7376 @cindex direct linking to a dll
7377 @item direct linking to a dll
7378 The cygwin/mingw ports of @command{ld} support the direct linking,
7379 including data symbols, to a dll without the usage of any import
7380 libraries. This is much faster and uses much less memory than does the
7381 traditional import library method, especially when linking large
7382 libraries or applications. When @command{ld} creates an import lib, each
7383 function or variable exported from the dll is stored in its own bfd, even
7384 though a single bfd could contain many exports. The overhead involved in
7385 storing, loading, and processing so many bfd's is quite large, and explains the
7386 tremendous time, memory, and storage needed to link against particularly
7387 large or complex libraries when using import libs.
7389 Linking directly to a dll uses no extra command-line switches other than
7390 @samp{-L} and @samp{-l}, because @command{ld} already searches for a number
7391 of names to match each library. All that is needed from the developer's
7392 perspective is an understanding of this search, in order to force ld to
7393 select the dll instead of an import library.
7396 For instance, when ld is called with the argument @samp{-lxxx} it will attempt
7397 to find, in the first directory of its search path,
7409 before moving on to the next directory in the search path.
7411 (*) Actually, this is not @samp{cygxxx.dll} but in fact is @samp{<prefix>xxx.dll},
7412 where @samp{<prefix>} is set by the @command{ld} option
7413 @samp{--dll-search-prefix=<prefix>}. In the case of cygwin, the standard gcc spec
7414 file includes @samp{--dll-search-prefix=cyg}, so in effect we actually search for
7417 Other win32-based unix environments, such as mingw or pw32, may use other
7418 @samp{<prefix>}es, although at present only cygwin makes use of this feature. It
7419 was originally intended to help avoid name conflicts among dll's built for the
7420 various win32/un*x environments, so that (for example) two versions of a zlib dll
7421 could coexist on the same machine.
7423 The generic cygwin/mingw path layout uses a @samp{bin} directory for
7424 applications and dll's and a @samp{lib} directory for the import
7425 libraries (using cygwin nomenclature):
7431 libxxx.dll.a (in case of dll's)
7432 libxxx.a (in case of static archive)
7435 Linking directly to a dll without using the import library can be
7438 1. Use the dll directly by adding the @samp{bin} path to the link line
7440 gcc -Wl,-verbose -o a.exe -L../bin/ -lxxx
7443 However, as the dll's often have version numbers appended to their names
7444 (@samp{cygncurses-5.dll}) this will often fail, unless one specifies
7445 @samp{-L../bin -lncurses-5} to include the version. Import libs are generally
7446 not versioned, and do not have this difficulty.
7448 2. Create a symbolic link from the dll to a file in the @samp{lib}
7449 directory according to the above mentioned search pattern. This
7450 should be used to avoid unwanted changes in the tools needed for
7454 ln -s bin/cygxxx.dll lib/[cyg|lib|]xxx.dll[.a]
7457 Then you can link without any make environment changes.
7460 gcc -Wl,-verbose -o a.exe -L../lib/ -lxxx
7463 This technique also avoids the version number problems, because the following is
7470 libxxx.dll.a -> ../bin/cygxxx-5.dll
7473 Linking directly to a dll without using an import lib will work
7474 even when auto-import features are exercised, and even when
7475 @samp{--enable-runtime-pseudo-relocs} is used.
7477 Given the improvements in speed and memory usage, one might justifiably
7478 wonder why import libraries are used at all. There are three reasons:
7480 1. Until recently, the link-directly-to-dll functionality did @emph{not}
7481 work with auto-imported data.
7483 2. Sometimes it is necessary to include pure static objects within the
7484 import library (which otherwise contains only bfd's for indirection
7485 symbols that point to the exports of a dll). Again, the import lib
7486 for the cygwin kernel makes use of this ability, and it is not
7487 possible to do this without an import lib.
7489 3. Symbol aliases can only be resolved using an import lib. This is
7490 critical when linking against OS-supplied dll's (eg, the win32 API)
7491 in which symbols are usually exported as undecorated aliases of their
7492 stdcall-decorated assembly names.
7494 So, import libs are not going away. But the ability to replace
7495 true import libs with a simple symbolic link to (or a copy of)
7496 a dll, in many cases, is a useful addition to the suite of tools
7497 binutils makes available to the win32 developer. Given the
7498 massive improvements in memory requirements during linking, storage
7499 requirements, and linking speed, we expect that many developers
7500 will soon begin to use this feature whenever possible.
7502 @item symbol aliasing
7504 @item adding additional names
7505 Sometimes, it is useful to export symbols with additional names.
7506 A symbol @samp{foo} will be exported as @samp{foo}, but it can also be
7507 exported as @samp{_foo} by using special directives in the DEF file
7508 when creating the dll. This will affect also the optional created
7509 import library. Consider the following DEF file:
7512 LIBRARY "xyz.dll" BASE=0x61000000
7519 The line @samp{_foo = foo} maps the symbol @samp{foo} to @samp{_foo}.
7521 Another method for creating a symbol alias is to create it in the
7522 source code using the "weak" attribute:
7525 void foo () @{ /* Do something. */; @}
7526 void _foo () __attribute__ ((weak, alias ("foo")));
7529 See the gcc manual for more information about attributes and weak
7532 @item renaming symbols
7533 Sometimes it is useful to rename exports. For instance, the cygwin
7534 kernel does this regularly. A symbol @samp{_foo} can be exported as
7535 @samp{foo} but not as @samp{_foo} by using special directives in the
7536 DEF file. (This will also affect the import library, if it is
7537 created). In the following example:
7540 LIBRARY "xyz.dll" BASE=0x61000000
7546 The line @samp{_foo = foo} maps the exported symbol @samp{foo} to
7550 Note: using a DEF file disables the default auto-export behavior,
7551 unless the @samp{--export-all-symbols} command line option is used.
7552 If, however, you are trying to rename symbols, then you should list
7553 @emph{all} desired exports in the DEF file, including the symbols
7554 that are not being renamed, and do @emph{not} use the
7555 @samp{--export-all-symbols} option. If you list only the
7556 renamed symbols in the DEF file, and use @samp{--export-all-symbols}
7557 to handle the other symbols, then the both the new names @emph{and}
7558 the original names for the renamed symbols will be exported.
7559 In effect, you'd be aliasing those symbols, not renaming them,
7560 which is probably not what you wanted.
7562 @cindex weak externals
7563 @item weak externals
7564 The Windows object format, PE, specifies a form of weak symbols called
7565 weak externals. When a weak symbol is linked and the symbol is not
7566 defined, the weak symbol becomes an alias for some other symbol. There
7567 are three variants of weak externals:
7569 @item Definition is searched for in objects and libraries, historically
7570 called lazy externals.
7571 @item Definition is searched for only in other objects, not in libraries.
7572 This form is not presently implemented.
7573 @item No search; the symbol is an alias. This form is not presently
7576 As a GNU extension, weak symbols that do not specify an alternate symbol
7577 are supported. If the symbol is undefined when linking, the symbol
7578 uses a default value.
7580 @cindex aligned common symbols
7581 @item aligned common symbols
7582 As a GNU extension to the PE file format, it is possible to specify the
7583 desired alignment for a common symbol. This information is conveyed from
7584 the assembler or compiler to the linker by means of GNU-specific commands
7585 carried in the object file's @samp{.drectve} section, which are recognized
7586 by @command{ld} and respected when laying out the common symbols. Native
7587 tools will be able to process object files employing this GNU extension,
7588 but will fail to respect the alignment instructions, and may issue noisy
7589 warnings about unknown linker directives.
7603 @section @code{ld} and Xtensa Processors
7605 @cindex Xtensa processors
7606 The default @command{ld} behavior for Xtensa processors is to interpret
7607 @code{SECTIONS} commands so that lists of explicitly named sections in a
7608 specification with a wildcard file will be interleaved when necessary to
7609 keep literal pools within the range of PC-relative load offsets. For
7610 example, with the command:
7622 @command{ld} may interleave some of the @code{.literal}
7623 and @code{.text} sections from different object files to ensure that the
7624 literal pools are within the range of PC-relative load offsets. A valid
7625 interleaving might place the @code{.literal} sections from an initial
7626 group of files followed by the @code{.text} sections of that group of
7627 files. Then, the @code{.literal} sections from the rest of the files
7628 and the @code{.text} sections from the rest of the files would follow.
7630 @cindex @option{--relax} on Xtensa
7631 @cindex relaxing on Xtensa
7632 Relaxation is enabled by default for the Xtensa version of @command{ld} and
7633 provides two important link-time optimizations. The first optimization
7634 is to combine identical literal values to reduce code size. A redundant
7635 literal will be removed and all the @code{L32R} instructions that use it
7636 will be changed to reference an identical literal, as long as the
7637 location of the replacement literal is within the offset range of all
7638 the @code{L32R} instructions. The second optimization is to remove
7639 unnecessary overhead from assembler-generated ``longcall'' sequences of
7640 @code{L32R}/@code{CALLX@var{n}} when the target functions are within
7641 range of direct @code{CALL@var{n}} instructions.
7643 For each of these cases where an indirect call sequence can be optimized
7644 to a direct call, the linker will change the @code{CALLX@var{n}}
7645 instruction to a @code{CALL@var{n}} instruction, remove the @code{L32R}
7646 instruction, and remove the literal referenced by the @code{L32R}
7647 instruction if it is not used for anything else. Removing the
7648 @code{L32R} instruction always reduces code size but can potentially
7649 hurt performance by changing the alignment of subsequent branch targets.
7650 By default, the linker will always preserve alignments, either by
7651 switching some instructions between 24-bit encodings and the equivalent
7652 density instructions or by inserting a no-op in place of the @code{L32R}
7653 instruction that was removed. If code size is more important than
7654 performance, the @option{--size-opt} option can be used to prevent the
7655 linker from widening density instructions or inserting no-ops, except in
7656 a few cases where no-ops are required for correctness.
7658 The following Xtensa-specific command-line options can be used to
7661 @cindex Xtensa options
7664 When optimizing indirect calls to direct calls, optimize for code size
7665 more than performance. With this option, the linker will not insert
7666 no-ops or widen density instructions to preserve branch target
7667 alignment. There may still be some cases where no-ops are required to
7668 preserve the correctness of the code.
7676 @ifclear SingleFormat
7681 @cindex object file management
7682 @cindex object formats available
7684 The linker accesses object and archive files using the BFD libraries.
7685 These libraries allow the linker to use the same routines to operate on
7686 object files whatever the object file format. A different object file
7687 format can be supported simply by creating a new BFD back end and adding
7688 it to the library. To conserve runtime memory, however, the linker and
7689 associated tools are usually configured to support only a subset of the
7690 object file formats available. You can use @code{objdump -i}
7691 (@pxref{objdump,,objdump,binutils.info,The GNU Binary Utilities}) to
7692 list all the formats available for your configuration.
7694 @cindex BFD requirements
7695 @cindex requirements for BFD
7696 As with most implementations, BFD is a compromise between
7697 several conflicting requirements. The major factor influencing
7698 BFD design was efficiency: any time used converting between
7699 formats is time which would not have been spent had BFD not
7700 been involved. This is partly offset by abstraction payback; since
7701 BFD simplifies applications and back ends, more time and care
7702 may be spent optimizing algorithms for a greater speed.
7704 One minor artifact of the BFD solution which you should bear in
7705 mind is the potential for information loss. There are two places where
7706 useful information can be lost using the BFD mechanism: during
7707 conversion and during output. @xref{BFD information loss}.
7710 * BFD outline:: How it works: an outline of BFD
7714 @section How It Works: An Outline of BFD
7715 @cindex opening object files
7716 @include bfdsumm.texi
7719 @node Reporting Bugs
7720 @chapter Reporting Bugs
7721 @cindex bugs in @command{ld}
7722 @cindex reporting bugs in @command{ld}
7724 Your bug reports play an essential role in making @command{ld} reliable.
7726 Reporting a bug may help you by bringing a solution to your problem, or
7727 it may not. But in any case the principal function of a bug report is
7728 to help the entire community by making the next version of @command{ld}
7729 work better. Bug reports are your contribution to the maintenance of
7732 In order for a bug report to serve its purpose, you must include the
7733 information that enables us to fix the bug.
7736 * Bug Criteria:: Have you found a bug?
7737 * Bug Reporting:: How to report bugs
7741 @section Have You Found a Bug?
7742 @cindex bug criteria
7744 If you are not sure whether you have found a bug, here are some guidelines:
7747 @cindex fatal signal
7748 @cindex linker crash
7749 @cindex crash of linker
7751 If the linker gets a fatal signal, for any input whatever, that is a
7752 @command{ld} bug. Reliable linkers never crash.
7754 @cindex error on valid input
7756 If @command{ld} produces an error message for valid input, that is a bug.
7758 @cindex invalid input
7760 If @command{ld} does not produce an error message for invalid input, that
7761 may be a bug. In the general case, the linker can not verify that
7762 object files are correct.
7765 If you are an experienced user of linkers, your suggestions for
7766 improvement of @command{ld} are welcome in any case.
7770 @section How to Report Bugs
7772 @cindex @command{ld} bugs, reporting
7774 A number of companies and individuals offer support for @sc{gnu}
7775 products. If you obtained @command{ld} from a support organization, we
7776 recommend you contact that organization first.
7778 You can find contact information for many support companies and
7779 individuals in the file @file{etc/SERVICE} in the @sc{gnu} Emacs
7783 Otherwise, send bug reports for @command{ld} to
7787 The fundamental principle of reporting bugs usefully is this:
7788 @strong{report all the facts}. If you are not sure whether to state a
7789 fact or leave it out, state it!
7791 Often people omit facts because they think they know what causes the
7792 problem and assume that some details do not matter. Thus, you might
7793 assume that the name of a symbol you use in an example does not
7794 matter. Well, probably it does not, but one cannot be sure. Perhaps
7795 the bug is a stray memory reference which happens to fetch from the
7796 location where that name is stored in memory; perhaps, if the name
7797 were different, the contents of that location would fool the linker
7798 into doing the right thing despite the bug. Play it safe and give a
7799 specific, complete example. That is the easiest thing for you to do,
7800 and the most helpful.
7802 Keep in mind that the purpose of a bug report is to enable us to fix
7803 the bug if it is new to us. Therefore, always write your bug reports
7804 on the assumption that the bug has not been reported previously.
7806 Sometimes people give a few sketchy facts and ask, ``Does this ring a
7807 bell?'' This cannot help us fix a bug, so it is basically useless. We
7808 respond by asking for enough details to enable us to investigate.
7809 You might as well expedite matters by sending them to begin with.
7811 To enable us to fix the bug, you should include all these things:
7815 The version of @command{ld}. @command{ld} announces it if you start it with
7816 the @samp{--version} argument.
7818 Without this, we will not know whether there is any point in looking for
7819 the bug in the current version of @command{ld}.
7822 Any patches you may have applied to the @command{ld} source, including any
7823 patches made to the @code{BFD} library.
7826 The type of machine you are using, and the operating system name and
7830 What compiler (and its version) was used to compile @command{ld}---e.g.
7834 The command arguments you gave the linker to link your example and
7835 observe the bug. To guarantee you will not omit something important,
7836 list them all. A copy of the Makefile (or the output from make) is
7839 If we were to try to guess the arguments, we would probably guess wrong
7840 and then we might not encounter the bug.
7843 A complete input file, or set of input files, that will reproduce the
7844 bug. It is generally most helpful to send the actual object files
7845 provided that they are reasonably small. Say no more than 10K. For
7846 bigger files you can either make them available by FTP or HTTP or else
7847 state that you are willing to send the object file(s) to whomever
7848 requests them. (Note - your email will be going to a mailing list, so
7849 we do not want to clog it up with large attachments). But small
7850 attachments are best.
7852 If the source files were assembled using @code{gas} or compiled using
7853 @code{gcc}, then it may be OK to send the source files rather than the
7854 object files. In this case, be sure to say exactly what version of
7855 @code{gas} or @code{gcc} was used to produce the object files. Also say
7856 how @code{gas} or @code{gcc} were configured.
7859 A description of what behavior you observe that you believe is
7860 incorrect. For example, ``It gets a fatal signal.''
7862 Of course, if the bug is that @command{ld} gets a fatal signal, then we
7863 will certainly notice it. But if the bug is incorrect output, we might
7864 not notice unless it is glaringly wrong. You might as well not give us
7865 a chance to make a mistake.
7867 Even if the problem you experience is a fatal signal, you should still
7868 say so explicitly. Suppose something strange is going on, such as, your
7869 copy of @command{ld} is out of sync, or you have encountered a bug in the
7870 C library on your system. (This has happened!) Your copy might crash
7871 and ours would not. If you told us to expect a crash, then when ours
7872 fails to crash, we would know that the bug was not happening for us. If
7873 you had not told us to expect a crash, then we would not be able to draw
7874 any conclusion from our observations.
7877 If you wish to suggest changes to the @command{ld} source, send us context
7878 diffs, as generated by @code{diff} with the @samp{-u}, @samp{-c}, or
7879 @samp{-p} option. Always send diffs from the old file to the new file.
7880 If you even discuss something in the @command{ld} source, refer to it by
7881 context, not by line number.
7883 The line numbers in our development sources will not match those in your
7884 sources. Your line numbers would convey no useful information to us.
7887 Here are some things that are not necessary:
7891 A description of the envelope of the bug.
7893 Often people who encounter a bug spend a lot of time investigating
7894 which changes to the input file will make the bug go away and which
7895 changes will not affect it.
7897 This is often time consuming and not very useful, because the way we
7898 will find the bug is by running a single example under the debugger
7899 with breakpoints, not by pure deduction from a series of examples.
7900 We recommend that you save your time for something else.
7902 Of course, if you can find a simpler example to report @emph{instead}
7903 of the original one, that is a convenience for us. Errors in the
7904 output will be easier to spot, running under the debugger will take
7905 less time, and so on.
7907 However, simplification is not vital; if you do not want to do this,
7908 report the bug anyway and send us the entire test case you used.
7911 A patch for the bug.
7913 A patch for the bug does help us if it is a good one. But do not omit
7914 the necessary information, such as the test case, on the assumption that
7915 a patch is all we need. We might see problems with your patch and decide
7916 to fix the problem another way, or we might not understand it at all.
7918 Sometimes with a program as complicated as @command{ld} it is very hard to
7919 construct an example that will make the program follow a certain path
7920 through the code. If you do not send us the example, we will not be
7921 able to construct one, so we will not be able to verify that the bug is
7924 And if we cannot understand what bug you are trying to fix, or why your
7925 patch should be an improvement, we will not install it. A test case will
7926 help us to understand.
7929 A guess about what the bug is or what it depends on.
7931 Such guesses are usually wrong. Even we cannot guess right about such
7932 things without first using the debugger to find the facts.
7936 @appendix MRI Compatible Script Files
7937 @cindex MRI compatibility
7938 To aid users making the transition to @sc{gnu} @command{ld} from the MRI
7939 linker, @command{ld} can use MRI compatible linker scripts as an
7940 alternative to the more general-purpose linker scripting language
7941 described in @ref{Scripts}. MRI compatible linker scripts have a much
7942 simpler command set than the scripting language otherwise used with
7943 @command{ld}. @sc{gnu} @command{ld} supports the most commonly used MRI
7944 linker commands; these commands are described here.
7946 In general, MRI scripts aren't of much use with the @code{a.out} object
7947 file format, since it only has three sections and MRI scripts lack some
7948 features to make use of them.
7950 You can specify a file containing an MRI-compatible script using the
7951 @samp{-c} command-line option.
7953 Each command in an MRI-compatible script occupies its own line; each
7954 command line starts with the keyword that identifies the command (though
7955 blank lines are also allowed for punctuation). If a line of an
7956 MRI-compatible script begins with an unrecognized keyword, @command{ld}
7957 issues a warning message, but continues processing the script.
7959 Lines beginning with @samp{*} are comments.
7961 You can write these commands using all upper-case letters, or all
7962 lower case; for example, @samp{chip} is the same as @samp{CHIP}.
7963 The following list shows only the upper-case form of each command.
7966 @cindex @code{ABSOLUTE} (MRI)
7967 @item ABSOLUTE @var{secname}
7968 @itemx ABSOLUTE @var{secname}, @var{secname}, @dots{} @var{secname}
7969 Normally, @command{ld} includes in the output file all sections from all
7970 the input files. However, in an MRI-compatible script, you can use the
7971 @code{ABSOLUTE} command to restrict the sections that will be present in
7972 your output program. If the @code{ABSOLUTE} command is used at all in a
7973 script, then only the sections named explicitly in @code{ABSOLUTE}
7974 commands will appear in the linker output. You can still use other
7975 input sections (whatever you select on the command line, or using
7976 @code{LOAD}) to resolve addresses in the output file.
7978 @cindex @code{ALIAS} (MRI)
7979 @item ALIAS @var{out-secname}, @var{in-secname}
7980 Use this command to place the data from input section @var{in-secname}
7981 in a section called @var{out-secname} in the linker output file.
7983 @var{in-secname} may be an integer.
7985 @cindex @code{ALIGN} (MRI)
7986 @item ALIGN @var{secname} = @var{expression}
7987 Align the section called @var{secname} to @var{expression}. The
7988 @var{expression} should be a power of two.
7990 @cindex @code{BASE} (MRI)
7991 @item BASE @var{expression}
7992 Use the value of @var{expression} as the lowest address (other than
7993 absolute addresses) in the output file.
7995 @cindex @code{CHIP} (MRI)
7996 @item CHIP @var{expression}
7997 @itemx CHIP @var{expression}, @var{expression}
7998 This command does nothing; it is accepted only for compatibility.
8000 @cindex @code{END} (MRI)
8002 This command does nothing whatever; it's only accepted for compatibility.
8004 @cindex @code{FORMAT} (MRI)
8005 @item FORMAT @var{output-format}
8006 Similar to the @code{OUTPUT_FORMAT} command in the more general linker
8007 language, but restricted to one of these output formats:
8011 S-records, if @var{output-format} is @samp{S}
8014 IEEE, if @var{output-format} is @samp{IEEE}
8017 COFF (the @samp{coff-m68k} variant in BFD), if @var{output-format} is
8021 @cindex @code{LIST} (MRI)
8022 @item LIST @var{anything}@dots{}
8023 Print (to the standard output file) a link map, as produced by the
8024 @command{ld} command-line option @samp{-M}.
8026 The keyword @code{LIST} may be followed by anything on the
8027 same line, with no change in its effect.
8029 @cindex @code{LOAD} (MRI)
8030 @item LOAD @var{filename}
8031 @itemx LOAD @var{filename}, @var{filename}, @dots{} @var{filename}
8032 Include one or more object file @var{filename} in the link; this has the
8033 same effect as specifying @var{filename} directly on the @command{ld}
8036 @cindex @code{NAME} (MRI)
8037 @item NAME @var{output-name}
8038 @var{output-name} is the name for the program produced by @command{ld}; the
8039 MRI-compatible command @code{NAME} is equivalent to the command-line
8040 option @samp{-o} or the general script language command @code{OUTPUT}.
8042 @cindex @code{ORDER} (MRI)
8043 @item ORDER @var{secname}, @var{secname}, @dots{} @var{secname}
8044 @itemx ORDER @var{secname} @var{secname} @var{secname}
8045 Normally, @command{ld} orders the sections in its output file in the
8046 order in which they first appear in the input files. In an MRI-compatible
8047 script, you can override this ordering with the @code{ORDER} command. The
8048 sections you list with @code{ORDER} will appear first in your output
8049 file, in the order specified.
8051 @cindex @code{PUBLIC} (MRI)
8052 @item PUBLIC @var{name}=@var{expression}
8053 @itemx PUBLIC @var{name},@var{expression}
8054 @itemx PUBLIC @var{name} @var{expression}
8055 Supply a value (@var{expression}) for external symbol
8056 @var{name} used in the linker input files.
8058 @cindex @code{SECT} (MRI)
8059 @item SECT @var{secname}, @var{expression}
8060 @itemx SECT @var{secname}=@var{expression}
8061 @itemx SECT @var{secname} @var{expression}
8062 You can use any of these three forms of the @code{SECT} command to
8063 specify the start address (@var{expression}) for section @var{secname}.
8064 If you have more than one @code{SECT} statement for the same
8065 @var{secname}, only the @emph{first} sets the start address.
8068 @node GNU Free Documentation License
8069 @appendix GNU Free Documentation License
8073 @unnumbered LD Index
8078 % I think something like @@colophon should be in texinfo. In the
8080 \long\def\colophon{\hbox to0pt{}\vfill
8081 \centerline{The body of this manual is set in}
8082 \centerline{\fontname\tenrm,}
8083 \centerline{with headings in {\bf\fontname\tenbf}}
8084 \centerline{and examples in {\tt\fontname\tentt}.}
8085 \centerline{{\it\fontname\tenit\/} and}
8086 \centerline{{\sl\fontname\tensl\/}}
8087 \centerline{are used for emphasis.}\vfill}
8089 % Blame: doc@@cygnus.com, 28mar91.