3 @c Copyright (C) 1991-2020 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
47 @dircategory Software development
49 * Ld: (ld). The GNU linker.
54 This file documents the @sc{gnu} linker LD
55 @ifset VERSION_PACKAGE
56 @value{VERSION_PACKAGE}
58 version @value{VERSION}.
60 Copyright @copyright{} 1991-2020 Free Software Foundation, Inc.
62 Permission is granted to copy, distribute and/or modify this document
63 under the terms of the GNU Free Documentation License, Version 1.3
64 or any later version published by the Free Software Foundation;
65 with no Invariant Sections, with no Front-Cover Texts, and with no
66 Back-Cover Texts. A copy of the license is included in the
67 section entitled ``GNU Free Documentation License''.
71 @setchapternewpage odd
72 @settitle The GNU linker
77 @ifset VERSION_PACKAGE
78 @subtitle @value{VERSION_PACKAGE}
80 @subtitle Version @value{VERSION}
81 @author Steve Chamberlain
82 @author Ian Lance Taylor
87 \hfill Red Hat Inc\par
88 \hfill nickc\@credhat.com, doc\@redhat.com\par
89 \hfill {\it The GNU linker}\par
90 \hfill Edited by Jeffrey Osier (jeffrey\@cygnus.com)\par
92 \global\parindent=0pt % Steve likes it this way.
95 @vskip 0pt plus 1filll
96 @c man begin COPYRIGHT
97 Copyright @copyright{} 1991-2020 Free Software Foundation, Inc.
99 Permission is granted to copy, distribute and/or modify this document
100 under the terms of the GNU Free Documentation License, Version 1.3
101 or any later version published by the Free Software Foundation;
102 with no Invariant Sections, with no Front-Cover Texts, and with no
103 Back-Cover Texts. A copy of the license is included in the
104 section entitled ``GNU Free Documentation License''.
110 @c FIXME: Talk about importance of *order* of args, cmds to linker!
115 This file documents the @sc{gnu} linker ld
116 @ifset VERSION_PACKAGE
117 @value{VERSION_PACKAGE}
119 version @value{VERSION}.
121 This document is distributed under the terms of the GNU Free
122 Documentation License version 1.3. A copy of the license is included
123 in the section entitled ``GNU Free Documentation License''.
126 * Overview:: Overview
127 * Invocation:: Invocation
128 * Scripts:: Linker Scripts
129 * Plugins:: Linker Plugins
131 * Machine Dependent:: Machine Dependent Features
135 * H8/300:: ld and the H8/300
138 * Renesas:: ld and other Renesas micros
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 * S/390 ELF:: ld and S/390 ELF Support
165 * SPU ELF:: ld and SPU ELF Support
168 * TI COFF:: ld and the TI COFF
171 * Win32:: ld and WIN32 (cygwin/mingw)
174 * Xtensa:: ld and Xtensa Processors
177 @ifclear SingleFormat
180 @c Following blank line required for remaining bug in makeinfo conds/menus
182 * Reporting Bugs:: Reporting Bugs
183 * MRI:: MRI Compatible Script Files
184 * GNU Free Documentation License:: GNU Free Documentation License
185 * LD Index:: LD Index
192 @cindex @sc{gnu} linker
193 @cindex what is this?
196 @c man begin SYNOPSIS
197 ld [@b{options}] @var{objfile} @dots{}
201 ar(1), nm(1), objcopy(1), objdump(1), readelf(1) and
202 the Info entries for @file{binutils} and
207 @c man begin DESCRIPTION
209 @command{ld} combines a number of object and archive files, relocates
210 their data and ties up symbol references. Usually the last step in
211 compiling a program is to run @command{ld}.
213 @command{ld} accepts Linker Command Language files written in
214 a superset of AT&T's Link Editor Command Language syntax,
215 to provide explicit and total control over the linking process.
219 This man page does not describe the command language; see the
220 @command{ld} entry in @code{info} for full details on the command
221 language and on other aspects of the GNU linker.
224 @ifclear SingleFormat
225 This version of @command{ld} uses the general purpose BFD libraries
226 to operate on object files. This allows @command{ld} to read, combine, and
227 write object files in many different formats---for example, COFF or
228 @code{a.out}. Different formats may be linked together to produce any
229 available kind of object file. @xref{BFD}, for more information.
232 Aside from its flexibility, the @sc{gnu} linker is more helpful than other
233 linkers in providing diagnostic information. Many linkers abandon
234 execution immediately upon encountering an error; whenever possible,
235 @command{ld} continues executing, allowing you to identify other errors
236 (or, in some cases, to get an output file in spite of the error).
243 @c man begin DESCRIPTION
245 The @sc{gnu} linker @command{ld} is meant to cover a broad range of situations,
246 and to be as compatible as possible with other linkers. As a result,
247 you have many choices to control its behavior.
253 * Options:: Command-line Options
254 * Environment:: Environment Variables
258 @section Command-line Options
266 The linker supports a plethora of command-line options, but in actual
267 practice few of them are used in any particular context.
268 @cindex standard Unix system
269 For instance, a frequent use of @command{ld} is to link standard Unix
270 object files on a standard, supported Unix system. On such a system, to
271 link a file @code{hello.o}:
274 ld -o @var{output} /lib/crt0.o hello.o -lc
277 This tells @command{ld} to produce a file called @var{output} as the
278 result of linking the file @code{/lib/crt0.o} with @code{hello.o} and
279 the library @code{libc.a}, which will come from the standard search
280 directories. (See the discussion of the @samp{-l} option below.)
282 Some of the command-line options to @command{ld} may be specified at any
283 point in the command line. However, options which refer to files, such
284 as @samp{-l} or @samp{-T}, cause the file to be read at the point at
285 which the option appears in the command line, relative to the object
286 files and other file options. Repeating non-file options with a
287 different argument will either have no further effect, or override prior
288 occurrences (those further to the left on the command line) of that
289 option. Options which may be meaningfully specified more than once are
290 noted in the descriptions below.
293 Non-option arguments are object files or archives which are to be linked
294 together. They may follow, precede, or be mixed in with command-line
295 options, except that an object file argument may not be placed between
296 an option and its argument.
298 Usually the linker is invoked with at least one object file, but you can
299 specify other forms of binary input files using @samp{-l}, @samp{-R},
300 and the script command language. If @emph{no} binary input files at all
301 are specified, the linker does not produce any output, and issues the
302 message @samp{No input files}.
304 If the linker cannot recognize the format of an object file, it will
305 assume that it is a linker script. A script specified in this way
306 augments the main linker script used for the link (either the default
307 linker script or the one specified by using @samp{-T}). This feature
308 permits the linker to link against a file which appears to be an object
309 or an archive, but actually merely defines some symbol values, or uses
310 @code{INPUT} or @code{GROUP} to load other objects. Specifying a
311 script in this way merely augments the main linker script, with the
312 extra commands placed after the main script; use the @samp{-T} option
313 to replace the default linker script entirely, but note the effect of
314 the @code{INSERT} command. @xref{Scripts}.
316 For options whose names are a single letter,
317 option arguments must either follow the option letter without intervening
318 whitespace, or be given as separate arguments immediately following the
319 option that requires them.
321 For options whose names are multiple letters, either one dash or two can
322 precede the option name; for example, @samp{-trace-symbol} and
323 @samp{--trace-symbol} are equivalent. Note---there is one exception to
324 this rule. Multiple letter options that start with a lower case 'o' can
325 only be preceded by two dashes. This is to reduce confusion with the
326 @samp{-o} option. So for example @samp{-omagic} sets the output file
327 name to @samp{magic} whereas @samp{--omagic} sets the NMAGIC flag on the
330 Arguments to multiple-letter options must either be separated from the
331 option name by an equals sign, or be given as separate arguments
332 immediately following the option that requires them. For example,
333 @samp{--trace-symbol foo} and @samp{--trace-symbol=foo} are equivalent.
334 Unique abbreviations of the names of multiple-letter options are
337 Note---if the linker is being invoked indirectly, via a compiler driver
338 (e.g. @samp{gcc}) then all the linker command-line options should be
339 prefixed by @samp{-Wl,} (or whatever is appropriate for the particular
340 compiler driver) like this:
343 gcc -Wl,--start-group foo.o bar.o -Wl,--end-group
346 This is important, because otherwise the compiler driver program may
347 silently drop the linker options, resulting in a bad link. Confusion
348 may also arise when passing options that require values through a
349 driver, as the use of a space between option and argument acts as
350 a separator, and causes the driver to pass only the option to the linker
351 and the argument to the compiler. In this case, it is simplest to use
352 the joined forms of both single- and multiple-letter options, such as:
355 gcc foo.o bar.o -Wl,-eENTRY -Wl,-Map=a.map
358 Here is a table of the generic command-line switches accepted by the GNU
362 @include at-file.texi
364 @kindex -a @var{keyword}
365 @item -a @var{keyword}
366 This option is supported for HP/UX compatibility. The @var{keyword}
367 argument must be one of the strings @samp{archive}, @samp{shared}, or
368 @samp{default}. @samp{-aarchive} is functionally equivalent to
369 @samp{-Bstatic}, and the other two keywords are functionally equivalent
370 to @samp{-Bdynamic}. This option may be used any number of times.
372 @kindex --audit @var{AUDITLIB}
373 @item --audit @var{AUDITLIB}
374 Adds @var{AUDITLIB} to the @code{DT_AUDIT} entry of the dynamic section.
375 @var{AUDITLIB} is not checked for existence, nor will it use the DT_SONAME
376 specified in the library. If specified multiple times @code{DT_AUDIT}
377 will contain a colon separated list of audit interfaces to use. If the linker
378 finds an object with an audit entry while searching for shared libraries,
379 it will add a corresponding @code{DT_DEPAUDIT} entry in the output file.
380 This option is only meaningful on ELF platforms supporting the rtld-audit
383 @ifclear SingleFormat
384 @cindex binary input format
385 @kindex -b @var{format}
386 @kindex --format=@var{format}
389 @item -b @var{input-format}
390 @itemx --format=@var{input-format}
391 @command{ld} may be configured to support more than one kind of object
392 file. If your @command{ld} is configured this way, you can use the
393 @samp{-b} option to specify the binary format for input object files
394 that follow this option on the command line. Even when @command{ld} is
395 configured to support alternative object formats, you don't usually need
396 to specify this, as @command{ld} should be configured to expect as a
397 default input format the most usual format on each machine.
398 @var{input-format} is a text string, the name of a particular format
399 supported by the BFD libraries. (You can list the available binary
400 formats with @samp{objdump -i}.)
403 You may want to use this option if you are linking files with an unusual
404 binary format. You can also use @samp{-b} to switch formats explicitly (when
405 linking object files of different formats), by including
406 @samp{-b @var{input-format}} before each group of object files in a
409 The default format is taken from the environment variable
414 You can also define the input format from a script, using the command
417 see @ref{Format Commands}.
421 @kindex -c @var{MRI-cmdfile}
422 @kindex --mri-script=@var{MRI-cmdfile}
423 @cindex compatibility, MRI
424 @item -c @var{MRI-commandfile}
425 @itemx --mri-script=@var{MRI-commandfile}
426 For compatibility with linkers produced by MRI, @command{ld} accepts script
427 files written in an alternate, restricted command language, described in
429 @ref{MRI,,MRI Compatible Script Files}.
432 the MRI Compatible Script Files section of GNU ld documentation.
434 Introduce MRI script files with
435 the option @samp{-c}; use the @samp{-T} option to run linker
436 scripts written in the general-purpose @command{ld} scripting language.
437 If @var{MRI-cmdfile} does not exist, @command{ld} looks for it in the directories
438 specified by any @samp{-L} options.
440 @cindex common allocation
447 These three options are equivalent; multiple forms are supported for
448 compatibility with other linkers. They assign space to common symbols
449 even if a relocatable output file is specified (with @samp{-r}). The
450 script command @code{FORCE_COMMON_ALLOCATION} has the same effect.
451 @xref{Miscellaneous Commands}.
453 @kindex --depaudit @var{AUDITLIB}
454 @kindex -P @var{AUDITLIB}
455 @item --depaudit @var{AUDITLIB}
456 @itemx -P @var{AUDITLIB}
457 Adds @var{AUDITLIB} to the @code{DT_DEPAUDIT} entry of the dynamic section.
458 @var{AUDITLIB} is not checked for existence, nor will it use the DT_SONAME
459 specified in the library. If specified multiple times @code{DT_DEPAUDIT}
460 will contain a colon separated list of audit interfaces to use. This
461 option is only meaningful on ELF platforms supporting the rtld-audit interface.
462 The -P option is provided for Solaris compatibility.
464 @kindex --enable-non-contiguous-regions
465 @item --enable-non-contiguous-regions
466 This option avoids generating an error if an input section does not
467 fit a matching output section. The linker tries to allocate the input
468 section to subseque nt matching output sections, and generates an
469 error only if no output section is large enough. This is useful when
470 several non-contiguous memory regions are available and the input
471 section does not require a particular one. The order in which input
472 sections are evaluated does not change, for instance:
476 MEM1 (rwx) : ORIGIN : 0x1000, LENGTH = 0x14
477 MEM2 (rwx) : ORIGIN : 0x1000, LENGTH = 0x40
478 MEM3 (rwx) : ORIGIN : 0x2000, LENGTH = 0x40
481 mem1 : @{ *(.data.*); @} > MEM1
482 mem2 : @{ *(.data.*); @} > MEM2
483 mem3 : @{ *(.data.*); @} > MEM2
491 results in .data.1 affected to mem1, and .data.2 and .data.3
492 affected to mem2, even though .data.3 would fit in mem3.
495 This option is incompatible with INSERT statements because it changes
496 the way input sections are mapped to output sections.
498 @kindex --enable-non-contiguous-regions-warnings
499 @item --enable-non-contiguous-regions-warnings
500 This option enables warnings when
501 @code{--enable-non-contiguous-regions} allows possibly unexpected
502 matches in sections mapping, potentially leading to silently
503 discarding a section instead of failing because it does not fit any
506 @cindex entry point, from command line
507 @kindex -e @var{entry}
508 @kindex --entry=@var{entry}
510 @itemx --entry=@var{entry}
511 Use @var{entry} as the explicit symbol for beginning execution of your
512 program, rather than the default entry point. If there is no symbol
513 named @var{entry}, the linker will try to parse @var{entry} as a number,
514 and use that as the entry address (the number will be interpreted in
515 base 10; you may use a leading @samp{0x} for base 16, or a leading
516 @samp{0} for base 8). @xref{Entry Point}, for a discussion of defaults
517 and other ways of specifying the entry point.
519 @kindex --exclude-libs
520 @item --exclude-libs @var{lib},@var{lib},...
521 Specifies a list of archive libraries from which symbols should not be automatically
522 exported. The library names may be delimited by commas or colons. Specifying
523 @code{--exclude-libs ALL} excludes symbols in all archive libraries from
524 automatic export. This option is available only for the i386 PE targeted
525 port of the linker and for ELF targeted ports. For i386 PE, symbols
526 explicitly listed in a .def file are still exported, regardless of this
527 option. For ELF targeted ports, symbols affected by this option will
528 be treated as hidden.
530 @kindex --exclude-modules-for-implib
531 @item --exclude-modules-for-implib @var{module},@var{module},...
532 Specifies a list of object files or archive members, from which symbols
533 should not be automatically exported, but which should be copied wholesale
534 into the import library being generated during the link. The module names
535 may be delimited by commas or colons, and must match exactly the filenames
536 used by @command{ld} to open the files; for archive members, this is simply
537 the member name, but for object files the name listed must include and
538 match precisely any path used to specify the input file on the linker's
539 command-line. This option is available only for the i386 PE targeted port
540 of the linker. Symbols explicitly listed in a .def file are still exported,
541 regardless of this option.
543 @cindex dynamic symbol table
545 @kindex --export-dynamic
546 @kindex --no-export-dynamic
548 @itemx --export-dynamic
549 @itemx --no-export-dynamic
550 When creating a dynamically linked executable, using the @option{-E}
551 option or the @option{--export-dynamic} option causes the linker to add
552 all symbols to the dynamic symbol table. The dynamic symbol table is the
553 set of symbols which are visible from dynamic objects at run time.
555 If you do not use either of these options (or use the
556 @option{--no-export-dynamic} option to restore the default behavior), the
557 dynamic symbol table will normally contain only those symbols which are
558 referenced by some dynamic object mentioned in the link.
560 If you use @code{dlopen} to load a dynamic object which needs to refer
561 back to the symbols defined by the program, rather than some other
562 dynamic object, then you will probably need to use this option when
563 linking the program itself.
565 You can also use the dynamic list to control what symbols should
566 be added to the dynamic symbol table if the output format supports it.
567 See the description of @samp{--dynamic-list}.
569 Note that this option is specific to ELF targeted ports. PE targets
570 support a similar function to export all symbols from a DLL or EXE; see
571 the description of @samp{--export-all-symbols} below.
573 @kindex --export-dynamic-symbol=@var{glob}
574 @cindex export dynamic symbol
575 @item --export-dynamic-symbol=@var{glob}
576 When creating a dynamically linked executable, symbols matching
577 @var{glob} will be added to the dynamic symbol table. When creating a
578 shared library, references to symbols matching @var{glob} will not be
579 bound to the definitions within the shared library. This option is a
580 no-op when creating a shared library and @samp{-Bsymbolic} or
581 @samp{--dynamic-list} are not specified. This option is only meaningful
582 on ELF platforms which support shared libraries.
584 @kindex --export-dynamic-symbol-list=@var{file}
585 @cindex export dynamic symbol list
586 @item --export-dynamic-symbol-list=@var{file}
587 Specify a @samp{--export-dynamic-symbol} for each pattern in the file.
588 The format of the file is the same as the version node without
589 scope and node name. See @ref{VERSION} for more information.
591 @ifclear SingleFormat
592 @cindex big-endian objects
596 Link big-endian objects. This affects the default output format.
598 @cindex little-endian objects
601 Link little-endian objects. This affects the default output format.
604 @kindex -f @var{name}
605 @kindex --auxiliary=@var{name}
607 @itemx --auxiliary=@var{name}
608 When creating an ELF shared object, set the internal DT_AUXILIARY field
609 to the specified name. This tells the dynamic linker that the symbol
610 table of the shared object should be used as an auxiliary filter on the
611 symbol table of the shared object @var{name}.
613 If you later link a program against this filter object, then, when you
614 run the program, the dynamic linker will see the DT_AUXILIARY field. If
615 the dynamic linker resolves any symbols from the filter object, it will
616 first check whether there is a definition in the shared object
617 @var{name}. If there is one, it will be used instead of the definition
618 in the filter object. The shared object @var{name} need not exist.
619 Thus the shared object @var{name} may be used to provide an alternative
620 implementation of certain functions, perhaps for debugging or for
621 machine-specific performance.
623 This option may be specified more than once. The DT_AUXILIARY entries
624 will be created in the order in which they appear on the command line.
626 @kindex -F @var{name}
627 @kindex --filter=@var{name}
629 @itemx --filter=@var{name}
630 When creating an ELF shared object, set the internal DT_FILTER field to
631 the specified name. This tells the dynamic linker that the symbol table
632 of the shared object which is being created should be used as a filter
633 on the symbol table of the shared object @var{name}.
635 If you later link a program against this filter object, then, when you
636 run the program, the dynamic linker will see the DT_FILTER field. The
637 dynamic linker will resolve symbols according to the symbol table of the
638 filter object as usual, but it will actually link to the definitions
639 found in the shared object @var{name}. Thus the filter object can be
640 used to select a subset of the symbols provided by the object
643 Some older linkers used the @option{-F} option throughout a compilation
644 toolchain for specifying object-file format for both input and output
646 @ifclear SingleFormat
647 The @sc{gnu} linker uses other mechanisms for this purpose: the
648 @option{-b}, @option{--format}, @option{--oformat} options, the
649 @code{TARGET} command in linker scripts, and the @code{GNUTARGET}
650 environment variable.
652 The @sc{gnu} linker will ignore the @option{-F} option when not
653 creating an ELF shared object.
655 @cindex finalization function
656 @kindex -fini=@var{name}
657 @item -fini=@var{name}
658 When creating an ELF executable or shared object, call NAME when the
659 executable or shared object is unloaded, by setting DT_FINI to the
660 address of the function. By default, the linker uses @code{_fini} as
661 the function to call.
665 Ignored. Provided for compatibility with other tools.
667 @kindex -G @var{value}
668 @kindex --gpsize=@var{value}
671 @itemx --gpsize=@var{value}
672 Set the maximum size of objects to be optimized using the GP register to
673 @var{size}. This is only meaningful for object file formats such as
674 MIPS ELF that support putting large and small objects into different
675 sections. This is ignored for other object file formats.
677 @cindex runtime library name
678 @kindex -h @var{name}
679 @kindex -soname=@var{name}
681 @itemx -soname=@var{name}
682 When creating an ELF shared object, set the internal DT_SONAME field to
683 the specified name. When an executable is linked with a shared object
684 which has a DT_SONAME field, then when the executable is run the dynamic
685 linker will attempt to load the shared object specified by the DT_SONAME
686 field rather than the using the file name given to the linker.
689 @cindex incremental link
691 Perform an incremental link (same as option @samp{-r}).
693 @cindex initialization function
694 @kindex -init=@var{name}
695 @item -init=@var{name}
696 When creating an ELF executable or shared object, call NAME when the
697 executable or shared object is loaded, by setting DT_INIT to the address
698 of the function. By default, the linker uses @code{_init} as the
701 @cindex archive files, from cmd line
702 @kindex -l @var{namespec}
703 @kindex --library=@var{namespec}
704 @item -l @var{namespec}
705 @itemx --library=@var{namespec}
706 Add the archive or object file specified by @var{namespec} to the
707 list of files to link. This option may be used any number of times.
708 If @var{namespec} is of the form @file{:@var{filename}}, @command{ld}
709 will search the library path for a file called @var{filename}, otherwise it
710 will search the library path for a file called @file{lib@var{namespec}.a}.
712 On systems which support shared libraries, @command{ld} may also search for
713 files other than @file{lib@var{namespec}.a}. Specifically, on ELF
714 and SunOS systems, @command{ld} will search a directory for a library
715 called @file{lib@var{namespec}.so} before searching for one called
716 @file{lib@var{namespec}.a}. (By convention, a @code{.so} extension
717 indicates a shared library.) Note that this behavior does not apply
718 to @file{:@var{filename}}, which always specifies a file called
721 The linker will search an archive only once, at the location where it is
722 specified on the command line. If the archive defines a symbol which
723 was undefined in some object which appeared before the archive on the
724 command line, the linker will include the appropriate file(s) from the
725 archive. However, an undefined symbol in an object appearing later on
726 the command line will not cause the linker to search the archive again.
728 See the @option{-(} option for a way to force the linker to search
729 archives multiple times.
731 You may list the same archive multiple times on the command line.
734 This type of archive searching is standard for Unix linkers. However,
735 if you are using @command{ld} on AIX, note that it is different from the
736 behaviour of the AIX linker.
739 @cindex search directory, from cmd line
741 @kindex --library-path=@var{dir}
742 @item -L @var{searchdir}
743 @itemx --library-path=@var{searchdir}
744 Add path @var{searchdir} to the list of paths that @command{ld} will search
745 for archive libraries and @command{ld} control scripts. You may use this
746 option any number of times. The directories are searched in the order
747 in which they are specified on the command line. Directories specified
748 on the command line are searched before the default directories. All
749 @option{-L} options apply to all @option{-l} options, regardless of the
750 order in which the options appear. @option{-L} options do not affect
751 how @command{ld} searches for a linker script unless @option{-T}
754 If @var{searchdir} begins with @code{=} or @code{$SYSROOT}, then this
755 prefix will be replaced by the @dfn{sysroot prefix}, controlled by the
756 @samp{--sysroot} option, or specified when the linker is configured.
759 The default set of paths searched (without being specified with
760 @samp{-L}) depends on which emulation mode @command{ld} is using, and in
761 some cases also on how it was configured. @xref{Environment}.
764 The paths can also be specified in a link script with the
765 @code{SEARCH_DIR} command. Directories specified this way are searched
766 at the point in which the linker script appears in the command line.
769 @kindex -m @var{emulation}
770 @item -m @var{emulation}
771 Emulate the @var{emulation} linker. You can list the available
772 emulations with the @samp{--verbose} or @samp{-V} options.
774 If the @samp{-m} option is not used, the emulation is taken from the
775 @code{LDEMULATION} environment variable, if that is defined.
777 Otherwise, the default emulation depends upon how the linker was
785 Print a link map to the standard output. A link map provides
786 information about the link, including the following:
790 Where object files are mapped into memory.
792 How common symbols are allocated.
794 All archive members included in the link, with a mention of the symbol
795 which caused the archive member to be brought in.
797 The values assigned to symbols.
799 Note - symbols whose values are computed by an expression which
800 involves a reference to a previous value of the same symbol may not
801 have correct result displayed in the link map. This is because the
802 linker discards intermediate results and only retains the final value
803 of an expression. Under such circumstances the linker will display
804 the final value enclosed by square brackets. Thus for example a
805 linker script containing:
813 will produce the following output in the link map if the @option{-M}
818 [0x0000000c] foo = (foo * 0x4)
819 [0x0000000c] foo = (foo + 0x8)
822 See @ref{Expressions} for more information about expressions in linker
826 How GNU properties are merged.
828 When the linker merges input .note.gnu.property sections into one output
829 .note.gnu.property section, some properties are removed or updated.
830 These actions are reported in the link map. For example:
833 Removed property 0xc0000002 to merge foo.o (0x1) and bar.o (not found)
836 This indicates that property 0xc0000002 is removed from output when
837 merging properties in @file{foo.o}, whose property 0xc0000002 value
838 is 0x1, and @file{bar.o}, which doesn't have property 0xc0000002.
841 Updated property 0xc0010001 (0x1) to merge foo.o (0x1) and bar.o (0x1)
844 This indicates that property 0xc0010001 value is updated to 0x1 in output
845 when merging properties in @file{foo.o}, whose 0xc0010001 property value
846 is 0x1, and @file{bar.o}, whose 0xc0010001 property value is 0x1.
849 @cindex link map discarded
850 @kindex --print-map-discarded
851 @kindex --no-print-map-discarded
852 @item --print-map-discarded
853 @itemx --no-print-map-discarded
854 Print (or do not print) the list of discarded and garbage collected sections
855 in the link map. Enabled by default.
858 @cindex read-only text
863 Turn off page alignment of sections, and disable linking against shared
864 libraries. If the output format supports Unix style magic numbers,
865 mark the output as @code{NMAGIC}.
869 @cindex read/write from cmd line
873 Set the text and data sections to be readable and writable. Also, do
874 not page-align the data segment, and disable linking against shared
875 libraries. If the output format supports Unix style magic numbers,
876 mark the output as @code{OMAGIC}. Note: Although a writable text section
877 is allowed for PE-COFF targets, it does not conform to the format
878 specification published by Microsoft.
883 This option negates most of the effects of the @option{-N} option. It
884 sets the text section to be read-only, and forces the data segment to
885 be page-aligned. Note - this option does not enable linking against
886 shared libraries. Use @option{-Bdynamic} for this.
888 @kindex -o @var{output}
889 @kindex --output=@var{output}
890 @cindex naming the output file
891 @item -o @var{output}
892 @itemx --output=@var{output}
893 Use @var{output} as the name for the program produced by @command{ld}; if this
894 option is not specified, the name @file{a.out} is used by default. The
895 script command @code{OUTPUT} can also specify the output file name.
897 @kindex --dependency-file=@var{depfile}
898 @cindex dependency file
899 @item --dependency-file=@var{depfile}
900 Write a @dfn{dependency file} to @var{depfile}. This file contains a rule
901 suitable for @code{make} describing the output file and all the input files
902 that were read to produce it. The output is similar to the compiler's
903 output with @samp{-M -MP} (@pxref{Preprocessor Options,, Options
904 Controlling the Preprocessor, gcc.info, Using the GNU Compiler
905 Collection}). Note that there is no option like the compiler's @samp{-MM},
906 to exclude ``system files'' (which is not a well-specified concept in the
907 linker, unlike ``system headers'' in the compiler). So the output from
908 @samp{--dependency-file} is always specific to the exact state of the
909 installation where it was produced, and should not be copied into
910 distributed makefiles without careful editing.
912 @kindex -O @var{level}
913 @cindex generating optimized output
915 If @var{level} is a numeric values greater than zero @command{ld} optimizes
916 the output. This might take significantly longer and therefore probably
917 should only be enabled for the final binary. At the moment this
918 option only affects ELF shared library generation. Future releases of
919 the linker may make more use of this option. Also currently there is
920 no difference in the linker's behaviour for different non-zero values
921 of this option. Again this may change with future releases.
923 @kindex -plugin @var{name}
924 @item -plugin @var{name}
925 Involve a plugin in the linking process. The @var{name} parameter is
926 the absolute filename of the plugin. Usually this parameter is
927 automatically added by the complier, when using link time
928 optimization, but users can also add their own plugins if they so
931 Note that the location of the compiler originated plugins is different
932 from the place where the @command{ar}, @command{nm} and
933 @command{ranlib} programs search for their plugins. In order for
934 those commands to make use of a compiler based plugin it must first be
935 copied into the @file{$@{libdir@}/bfd-plugins} directory. All gcc
936 based linker plugins are backward compatible, so it is sufficient to
937 just copy in the newest one.
940 @cindex push state governing input file handling
942 The @option{--push-state} allows to preserve the current state of the
943 flags which govern the input file handling so that they can all be
944 restored with one corresponding @option{--pop-state} option.
946 The option which are covered are: @option{-Bdynamic}, @option{-Bstatic},
947 @option{-dn}, @option{-dy}, @option{-call_shared}, @option{-non_shared},
948 @option{-static}, @option{-N}, @option{-n}, @option{--whole-archive},
949 @option{--no-whole-archive}, @option{-r}, @option{-Ur},
950 @option{--copy-dt-needed-entries}, @option{--no-copy-dt-needed-entries},
951 @option{--as-needed}, @option{--no-as-needed}, and @option{-a}.
953 One target for this option are specifications for @file{pkg-config}. When
954 used with the @option{--libs} option all possibly needed libraries are
955 listed and then possibly linked with all the time. It is better to return
956 something as follows:
959 -Wl,--push-state,--as-needed -libone -libtwo -Wl,--pop-state
963 @cindex pop state governing input file handling
965 Undoes the effect of --push-state, restores the previous values of the
966 flags governing input file handling.
969 @kindex --emit-relocs
970 @cindex retain relocations in final executable
973 Leave relocation sections and contents in fully linked executables.
974 Post link analysis and optimization tools may need this information in
975 order to perform correct modifications of executables. This results
976 in larger executables.
978 This option is currently only supported on ELF platforms.
980 @kindex --force-dynamic
981 @cindex forcing the creation of dynamic sections
982 @item --force-dynamic
983 Force the output file to have dynamic sections. This option is specific
987 @cindex relocatable output
989 @kindex --relocatable
992 Generate relocatable output---i.e., generate an output file that can in
993 turn serve as input to @command{ld}. This is often called @dfn{partial
994 linking}. As a side effect, in environments that support standard Unix
995 magic numbers, this option also sets the output file's magic number to
997 @c ; see @option{-N}.
998 If this option is not specified, an absolute file is produced. When
999 linking C++ programs, this option @emph{will not} resolve references to
1000 constructors; to do that, use @samp{-Ur}.
1002 When an input file does not have the same format as the output file,
1003 partial linking is only supported if that input file does not contain any
1004 relocations. Different output formats can have further restrictions; for
1005 example some @code{a.out}-based formats do not support partial linking
1006 with input files in other formats at all.
1008 This option does the same thing as @samp{-i}.
1010 @kindex -R @var{file}
1011 @kindex --just-symbols=@var{file}
1012 @cindex symbol-only input
1013 @item -R @var{filename}
1014 @itemx --just-symbols=@var{filename}
1015 Read symbol names and their addresses from @var{filename}, but do not
1016 relocate it or include it in the output. This allows your output file
1017 to refer symbolically to absolute locations of memory defined in other
1018 programs. You may use this option more than once.
1020 For compatibility with other ELF linkers, if the @option{-R} option is
1021 followed by a directory name, rather than a file name, it is treated as
1022 the @option{-rpath} option.
1026 @cindex strip all symbols
1029 Omit all symbol information from the output file.
1032 @kindex --strip-debug
1033 @cindex strip debugger symbols
1035 @itemx --strip-debug
1036 Omit debugger symbol information (but not all symbols) from the output file.
1038 @kindex --strip-discarded
1039 @kindex --no-strip-discarded
1040 @item --strip-discarded
1041 @itemx --no-strip-discarded
1042 Omit (or do not omit) global symbols defined in discarded sections.
1047 @cindex input files, displaying
1050 Print the names of the input files as @command{ld} processes them. If
1051 @samp{-t} is given twice then members within archives are also printed.
1052 @samp{-t} output is useful to generate a list of all the object files
1053 and scripts involved in linking, for example, when packaging files for
1054 a linker bug report.
1056 @kindex -T @var{script}
1057 @kindex --script=@var{script}
1058 @cindex script files
1059 @item -T @var{scriptfile}
1060 @itemx --script=@var{scriptfile}
1061 Use @var{scriptfile} as the linker script. This script replaces
1062 @command{ld}'s default linker script (rather than adding to it), so
1063 @var{commandfile} must specify everything necessary to describe the
1064 output file. @xref{Scripts}. If @var{scriptfile} does not exist in
1065 the current directory, @code{ld} looks for it in the directories
1066 specified by any preceding @samp{-L} options. Multiple @samp{-T}
1069 @kindex -dT @var{script}
1070 @kindex --default-script=@var{script}
1071 @cindex script files
1072 @item -dT @var{scriptfile}
1073 @itemx --default-script=@var{scriptfile}
1074 Use @var{scriptfile} as the default linker script. @xref{Scripts}.
1076 This option is similar to the @option{--script} option except that
1077 processing of the script is delayed until after the rest of the
1078 command line has been processed. This allows options placed after the
1079 @option{--default-script} option on the command line to affect the
1080 behaviour of the linker script, which can be important when the linker
1081 command line cannot be directly controlled by the user. (eg because
1082 the command line is being constructed by another tool, such as
1085 @kindex -u @var{symbol}
1086 @kindex --undefined=@var{symbol}
1087 @cindex undefined symbol
1088 @item -u @var{symbol}
1089 @itemx --undefined=@var{symbol}
1090 Force @var{symbol} to be entered in the output file as an undefined
1091 symbol. Doing this may, for example, trigger linking of additional
1092 modules from standard libraries. @samp{-u} may be repeated with
1093 different option arguments to enter additional undefined symbols. This
1094 option is equivalent to the @code{EXTERN} linker script command.
1096 If this option is being used to force additional modules to be pulled
1097 into the link, and if it is an error for the symbol to remain
1098 undefined, then the option @option{--require-defined} should be used
1101 @kindex --require-defined=@var{symbol}
1102 @cindex symbols, require defined
1103 @cindex defined symbol
1104 @item --require-defined=@var{symbol}
1105 Require that @var{symbol} is defined in the output file. This option
1106 is the same as option @option{--undefined} except that if @var{symbol}
1107 is not defined in the output file then the linker will issue an error
1108 and exit. The same effect can be achieved in a linker script by using
1109 @code{EXTERN}, @code{ASSERT} and @code{DEFINED} together. This option
1110 can be used multiple times to require additional symbols.
1113 @cindex constructors
1115 For anything other than C++ programs, this option is equivalent to
1116 @samp{-r}: it generates relocatable output---i.e., an output file that can in
1117 turn serve as input to @command{ld}. When linking C++ programs, @samp{-Ur}
1118 @emph{does} resolve references to constructors, unlike @samp{-r}.
1119 It does not work to use @samp{-Ur} on files that were themselves linked
1120 with @samp{-Ur}; once the constructor table has been built, it cannot
1121 be added to. Use @samp{-Ur} only for the last partial link, and
1122 @samp{-r} for the others.
1124 @kindex --orphan-handling=@var{MODE}
1125 @cindex orphan sections
1126 @cindex sections, orphan
1127 @item --orphan-handling=@var{MODE}
1128 Control how orphan sections are handled. An orphan section is one not
1129 specifically mentioned in a linker script. @xref{Orphan Sections}.
1131 @var{MODE} can have any of the following values:
1135 Orphan sections are placed into a suitable output section following
1136 the strategy described in @ref{Orphan Sections}. The option
1137 @samp{--unique} also affects how sections are placed.
1140 All orphan sections are discarded, by placing them in the
1141 @samp{/DISCARD/} section (@pxref{Output Section Discarding}).
1144 The linker will place the orphan section as for @code{place} and also
1148 The linker will exit with an error if any orphan section is found.
1151 The default if @samp{--orphan-handling} is not given is @code{place}.
1153 @kindex --unique[=@var{SECTION}]
1154 @item --unique[=@var{SECTION}]
1155 Creates a separate output section for every input section matching
1156 @var{SECTION}, or if the optional wildcard @var{SECTION} argument is
1157 missing, for every orphan input section. An orphan section is one not
1158 specifically mentioned in a linker script. You may use this option
1159 multiple times on the command line; It prevents the normal merging of
1160 input sections with the same name, overriding output section assignments
1170 Display the version number for @command{ld}. The @option{-V} option also
1171 lists the supported emulations.
1174 @kindex --discard-all
1175 @cindex deleting local symbols
1177 @itemx --discard-all
1178 Delete all local symbols.
1181 @kindex --discard-locals
1182 @cindex local symbols, deleting
1184 @itemx --discard-locals
1185 Delete all temporary local symbols. (These symbols start with
1186 system-specific local label prefixes, typically @samp{.L} for ELF systems
1187 or @samp{L} for traditional a.out systems.)
1189 @kindex -y @var{symbol}
1190 @kindex --trace-symbol=@var{symbol}
1191 @cindex symbol tracing
1192 @item -y @var{symbol}
1193 @itemx --trace-symbol=@var{symbol}
1194 Print the name of each linked file in which @var{symbol} appears. This
1195 option may be given any number of times. On many systems it is necessary
1196 to prepend an underscore.
1198 This option is useful when you have an undefined symbol in your link but
1199 don't know where the reference is coming from.
1201 @kindex -Y @var{path}
1203 Add @var{path} to the default library search path. This option exists
1204 for Solaris compatibility.
1206 @kindex -z @var{keyword}
1207 @item -z @var{keyword}
1208 The recognized keywords are:
1212 Always generate BND prefix in PLT entries. Supported for Linux/x86_64.
1214 @item call-nop=prefix-addr
1215 @itemx call-nop=suffix-nop
1216 @itemx call-nop=prefix-@var{byte}
1217 @itemx call-nop=suffix-@var{byte}
1218 Specify the 1-byte @code{NOP} padding when transforming indirect call
1219 to a locally defined function, foo, via its GOT slot.
1220 @option{call-nop=prefix-addr} generates @code{0x67 call foo}.
1221 @option{call-nop=suffix-nop} generates @code{call foo 0x90}.
1222 @option{call-nop=prefix-@var{byte}} generates @code{@var{byte} call foo}.
1223 @option{call-nop=suffix-@var{byte}} generates @code{call foo @var{byte}}.
1224 Supported for i386 and x86_64.
1226 @item cet-report=none
1227 @itemx cet-report=warning
1228 @itemx cet-report=error
1229 Specify how to report the missing GNU_PROPERTY_X86_FEATURE_1_IBT and
1230 GNU_PROPERTY_X86_FEATURE_1_SHSTK properties in input .note.gnu.property
1231 section. @option{cet-report=none}, which is the default, will make the
1232 linker not report missing properties in input files.
1233 @option{cet-report=warning} will make the linker issue a warning for
1234 missing properties in input files. @option{cet-report=error} will make
1235 the linker issue an error for missing properties in input files.
1236 Note that @option{ibt} will turn off the missing
1237 GNU_PROPERTY_X86_FEATURE_1_IBT property report and @option{shstk} will
1238 turn off the missing GNU_PROPERTY_X86_FEATURE_1_SHSTK property report.
1239 Supported for Linux/i386 and Linux/x86_64.
1243 Combine multiple dynamic relocation sections and sort to improve
1244 dynamic symbol lookup caching. Do not do this if @samp{nocombreloc}.
1248 Generate common symbols with STT_COMMON type during a relocatable
1249 link. Use STT_OBJECT type if @samp{nocommon}.
1251 @item common-page-size=@var{value}
1252 Set the page size most commonly used to @var{value}. Memory image
1253 layout will be optimized to minimize memory pages if the system is
1254 using pages of this size.
1257 Report unresolved symbol references from regular object files. This
1258 is done even if the linker is creating a non-symbolic shared library.
1259 This option is the inverse of @samp{-z undefs}.
1261 @item dynamic-undefined-weak
1262 @itemx nodynamic-undefined-weak
1263 Make undefined weak symbols dynamic when building a dynamic object,
1264 if they are referenced from a regular object file and not forced local
1265 by symbol visibility or versioning. Do not make them dynamic if
1266 @samp{nodynamic-undefined-weak}. If neither option is given, a target
1267 may default to either option being in force, or make some other
1268 selection of undefined weak symbols dynamic. Not all targets support
1272 Marks the object as requiring executable stack.
1275 This option is only meaningful when building a shared object. It makes
1276 the symbols defined by this shared object available for symbol resolution
1277 of subsequently loaded libraries.
1280 This option is only meaningful when building a dynamic executable.
1281 This option marks the executable as requiring global auditing by
1282 setting the @code{DF_1_GLOBAUDIT} bit in the @code{DT_FLAGS_1} dynamic
1283 tag. Global auditing requires that any auditing library defined via
1284 the @option{--depaudit} or @option{-P} command-line options be run for
1285 all dynamic objects loaded by the application.
1288 Generate Intel Indirect Branch Tracking (IBT) enabled PLT entries.
1289 Supported for Linux/i386 and Linux/x86_64.
1292 Generate GNU_PROPERTY_X86_FEATURE_1_IBT in .note.gnu.property section
1293 to indicate compatibility with IBT. This also implies @option{ibtplt}.
1294 Supported for Linux/i386 and Linux/x86_64.
1297 This option is only meaningful when building a shared object.
1298 It marks the object so that its runtime initialization will occur
1299 before the runtime initialization of any other objects brought into
1300 the process at the same time. Similarly the runtime finalization of
1301 the object will occur after the runtime finalization of any other
1305 Specify that the dynamic loader should modify its symbol search order
1306 so that symbols in this shared library interpose all other shared
1307 libraries not so marked.
1310 When generating an executable or shared library, mark it to tell the
1311 dynamic linker to defer function call resolution to the point when
1312 the function is called (lazy binding), rather than at load time.
1313 Lazy binding is the default.
1316 Specify that the object's filters be processed immediately at runtime.
1318 @item max-page-size=@var{value}
1319 Set the maximum memory page size supported to @var{value}.
1322 Allow multiple definitions.
1325 Disable linker generated .dynbss variables used in place of variables
1326 defined in shared libraries. May result in dynamic text relocations.
1329 Specify that the dynamic loader search for dependencies of this object
1330 should ignore any default library search paths.
1333 Specify that the object shouldn't be unloaded at runtime.
1336 Specify that the object is not available to @code{dlopen}.
1339 Specify that the object can not be dumped by @code{dldump}.
1342 Marks the object as not requiring executable stack.
1344 @item noextern-protected-data
1345 Don't treat protected data symbols as external when building a shared
1346 library. This option overrides the linker backend default. It can be
1347 used to work around incorrect relocations against protected data symbols
1348 generated by compiler. Updates on protected data symbols by another
1349 module aren't visible to the resulting shared library. Supported for
1352 @item noreloc-overflow
1353 Disable relocation overflow check. This can be used to disable
1354 relocation overflow check if there will be no dynamic relocation
1355 overflow at run-time. Supported for x86_64.
1358 When generating an executable or shared library, mark it to tell the
1359 dynamic linker to resolve all symbols when the program is started, or
1360 when the shared library is loaded by dlopen, instead of deferring
1361 function call resolution to the point when the function is first
1365 Specify that the object requires @samp{$ORIGIN} handling in paths.
1369 Create an ELF @code{PT_GNU_RELRO} segment header in the object. This
1370 specifies a memory segment that should be made read-only after
1371 relocation, if supported. Specifying @samp{common-page-size} smaller
1372 than the system page size will render this protection ineffective.
1373 Don't create an ELF @code{PT_GNU_RELRO} segment if @samp{norelro}.
1376 @itemx noseparate-code
1377 Create separate code @code{PT_LOAD} segment header in the object. This
1378 specifies a memory segment that should contain only instructions and must
1379 be in wholly disjoint pages from any other data. Don't create separate
1380 code @code{PT_LOAD} segment if @samp{noseparate-code} is used.
1383 @itemx nounique-symbol
1384 Avoid duplicated local symbol names in the symbol string table. Append
1385 ".@code{number}" to duplicated local symbol names if @samp{unique-symbol}
1386 is used. @option{nounique-symbol} is the default.
1389 Generate GNU_PROPERTY_X86_FEATURE_1_SHSTK in .note.gnu.property section
1390 to indicate compatibility with Intel Shadow Stack. Supported for
1391 Linux/i386 and Linux/x86_64.
1393 @item stack-size=@var{value}
1394 Specify a stack size for an ELF @code{PT_GNU_STACK} segment.
1395 Specifying zero will override any default non-zero sized
1396 @code{PT_GNU_STACK} segment creation.
1398 @item start-stop-visibility=@var{value}
1400 @cindex ELF symbol visibility
1401 Specify the ELF symbol visibility for synthesized
1402 @code{__start_SECNAME} and @code{__stop_SECNAME} symbols (@pxref{Input
1403 Section Example}). @var{value} must be exactly @samp{default},
1404 @samp{internal}, @samp{hidden}, or @samp{protected}. If no @samp{-z
1405 start-stop-visibility} option is given, @samp{protected} is used for
1406 compatibility with historical practice. However, it's highly
1407 recommended to use @samp{-z start-stop-visibility=hidden} in new
1408 programs and shared libraries so that these symbols are not exported
1409 between shared objects, which is not usually what's intended.
1414 Report an error if DT_TEXTREL is set, i.e., if the position-independent
1415 or shared object has dynamic relocations in read-only sections. Don't
1416 report an error if @samp{notext} or @samp{textoff}.
1419 Do not report unresolved symbol references from regular object files,
1420 either when creating an executable, or when creating a shared library.
1421 This option is the inverse of @samp{-z defs}.
1423 @item x86-64-baseline
1427 Specify the x86-64 ISA level needed in .note.gnu.property section.
1428 @option{x86-64-baseline} generates @code{GNU_PROPERTY_X86_ISA_1_BASELINE}.
1429 @option{x86-64-v2} generates @code{GNU_PROPERTY_X86_ISA_1_V2}.
1430 @option{x86-64-v3} generates @code{GNU_PROPERTY_X86_ISA_1_V3}.
1431 @option{x86-64-v4} generates @code{GNU_PROPERTY_X86_ISA_1_V4}.
1432 Supported for Linux/i386 and Linux/x86_64.
1436 Other keywords are ignored for Solaris compatibility.
1439 @cindex groups of archives
1440 @item -( @var{archives} -)
1441 @itemx --start-group @var{archives} --end-group
1442 The @var{archives} should be a list of archive files. They may be
1443 either explicit file names, or @samp{-l} options.
1445 The specified archives are searched repeatedly until no new undefined
1446 references are created. Normally, an archive is searched only once in
1447 the order that it is specified on the command line. If a symbol in that
1448 archive is needed to resolve an undefined symbol referred to by an
1449 object in an archive that appears later on the command line, the linker
1450 would not be able to resolve that reference. By grouping the archives,
1451 they will all be searched repeatedly until all possible references are
1454 Using this option has a significant performance cost. It is best to use
1455 it only when there are unavoidable circular references between two or
1458 @kindex --accept-unknown-input-arch
1459 @kindex --no-accept-unknown-input-arch
1460 @item --accept-unknown-input-arch
1461 @itemx --no-accept-unknown-input-arch
1462 Tells the linker to accept input files whose architecture cannot be
1463 recognised. The assumption is that the user knows what they are doing
1464 and deliberately wants to link in these unknown input files. This was
1465 the default behaviour of the linker, before release 2.14. The default
1466 behaviour from release 2.14 onwards is to reject such input files, and
1467 so the @samp{--accept-unknown-input-arch} option has been added to
1468 restore the old behaviour.
1471 @kindex --no-as-needed
1473 @itemx --no-as-needed
1474 This option affects ELF DT_NEEDED tags for dynamic libraries mentioned
1475 on the command line after the @option{--as-needed} option. Normally
1476 the linker will add a DT_NEEDED tag for each dynamic library mentioned
1477 on the command line, regardless of whether the library is actually
1478 needed or not. @option{--as-needed} causes a DT_NEEDED tag to only be
1479 emitted for a library that @emph{at that point in the link} satisfies a
1480 non-weak undefined symbol reference from a regular object file or, if
1481 the library is not found in the DT_NEEDED lists of other needed libraries, a
1482 non-weak undefined symbol reference from another needed dynamic library.
1483 Object files or libraries appearing on the command line @emph{after}
1484 the library in question do not affect whether the library is seen as
1485 needed. This is similar to the rules for extraction of object files
1486 from archives. @option{--no-as-needed} restores the default behaviour.
1488 @kindex --add-needed
1489 @kindex --no-add-needed
1491 @itemx --no-add-needed
1492 These two options have been deprecated because of the similarity of
1493 their names to the @option{--as-needed} and @option{--no-as-needed}
1494 options. They have been replaced by @option{--copy-dt-needed-entries}
1495 and @option{--no-copy-dt-needed-entries}.
1497 @kindex -assert @var{keyword}
1498 @item -assert @var{keyword}
1499 This option is ignored for SunOS compatibility.
1503 @kindex -call_shared
1507 Link against dynamic libraries. This is only meaningful on platforms
1508 for which shared libraries are supported. This option is normally the
1509 default on such platforms. The different variants of this option are
1510 for compatibility with various systems. You may use this option
1511 multiple times on the command line: it affects library searching for
1512 @option{-l} options which follow it.
1516 Set the @code{DF_1_GROUP} flag in the @code{DT_FLAGS_1} entry in the dynamic
1517 section. This causes the runtime linker to handle lookups in this
1518 object and its dependencies to be performed only inside the group.
1519 @option{--unresolved-symbols=report-all} is implied. This option is
1520 only meaningful on ELF platforms which support shared libraries.
1530 Do not link against shared libraries. This is only meaningful on
1531 platforms for which shared libraries are supported. The different
1532 variants of this option are for compatibility with various systems. You
1533 may use this option multiple times on the command line: it affects
1534 library searching for @option{-l} options which follow it. This
1535 option also implies @option{--unresolved-symbols=report-all}. This
1536 option can be used with @option{-shared}. Doing so means that a
1537 shared library is being created but that all of the library's external
1538 references must be resolved by pulling in entries from static
1543 When creating a shared library, bind references to global symbols to the
1544 definition within the shared library, if any. Normally, it is possible
1545 for a program linked against a shared library to override the definition
1546 within the shared library. This option is only meaningful on ELF
1547 platforms which support shared libraries.
1549 @kindex -Bsymbolic-functions
1550 @item -Bsymbolic-functions
1551 When creating a shared library, bind references to global function
1552 symbols to the definition within the shared library, if any.
1553 This option is only meaningful on ELF platforms which support shared
1556 @kindex --dynamic-list=@var{dynamic-list-file}
1557 @item --dynamic-list=@var{dynamic-list-file}
1558 Specify the name of a dynamic list file to the linker. This is
1559 typically used when creating shared libraries to specify a list of
1560 global symbols whose references shouldn't be bound to the definition
1561 within the shared library, or creating dynamically linked executables
1562 to specify a list of symbols which should be added to the symbol table
1563 in the executable. This option is only meaningful on ELF platforms
1564 which support shared libraries.
1566 The format of the dynamic list is the same as the version node without
1567 scope and node name. See @ref{VERSION} for more information.
1569 @kindex --dynamic-list-data
1570 @item --dynamic-list-data
1571 Include all global data symbols to the dynamic list.
1573 @kindex --dynamic-list-cpp-new
1574 @item --dynamic-list-cpp-new
1575 Provide the builtin dynamic list for C++ operator new and delete. It
1576 is mainly useful for building shared libstdc++.
1578 @kindex --dynamic-list-cpp-typeinfo
1579 @item --dynamic-list-cpp-typeinfo
1580 Provide the builtin dynamic list for C++ runtime type identification.
1582 @kindex --check-sections
1583 @kindex --no-check-sections
1584 @item --check-sections
1585 @itemx --no-check-sections
1586 Asks the linker @emph{not} to check section addresses after they have
1587 been assigned to see if there are any overlaps. Normally the linker will
1588 perform this check, and if it finds any overlaps it will produce
1589 suitable error messages. The linker does know about, and does make
1590 allowances for sections in overlays. The default behaviour can be
1591 restored by using the command-line switch @option{--check-sections}.
1592 Section overlap is not usually checked for relocatable links. You can
1593 force checking in that case by using the @option{--check-sections}
1596 @kindex --copy-dt-needed-entries
1597 @kindex --no-copy-dt-needed-entries
1598 @item --copy-dt-needed-entries
1599 @itemx --no-copy-dt-needed-entries
1600 This option affects the treatment of dynamic libraries referred to
1601 by DT_NEEDED tags @emph{inside} ELF dynamic libraries mentioned on the
1602 command line. Normally the linker won't add a DT_NEEDED tag to the
1603 output binary for each library mentioned in a DT_NEEDED tag in an
1604 input dynamic library. With @option{--copy-dt-needed-entries}
1605 specified on the command line however any dynamic libraries that
1606 follow it will have their DT_NEEDED entries added. The default
1607 behaviour can be restored with @option{--no-copy-dt-needed-entries}.
1609 This option also has an effect on the resolution of symbols in dynamic
1610 libraries. With @option{--copy-dt-needed-entries} dynamic libraries
1611 mentioned on the command line will be recursively searched, following
1612 their DT_NEEDED tags to other libraries, in order to resolve symbols
1613 required by the output binary. With the default setting however
1614 the searching of dynamic libraries that follow it will stop with the
1615 dynamic library itself. No DT_NEEDED links will be traversed to resolve
1618 @cindex cross reference table
1621 Output a cross reference table. If a linker map file is being
1622 generated, the cross reference table is printed to the map file.
1623 Otherwise, it is printed on the standard output.
1625 The format of the table is intentionally simple, so that it may be
1626 easily processed by a script if necessary. The symbols are printed out,
1627 sorted by name. For each symbol, a list of file names is given. If the
1628 symbol is defined, the first file listed is the location of the
1629 definition. If the symbol is defined as a common value then any files
1630 where this happens appear next. Finally any files that reference the
1633 @cindex ctf variables
1634 @kindex --ctf-variables
1635 @kindex --no-ctf-variables
1636 @item --ctf-variables
1637 @item --no-ctf-variables
1638 The CTF debuginfo format supports a section which encodes the names and
1639 types of variables found in the program which do not appear in any symbol
1640 table. These variables clearly cannot be looked up by address by
1641 conventional debuggers, so the space used for their types and names is
1642 usually wasted: the types are usually small but the names are often not.
1643 @option{--ctf-variables} causes the generation of such a section.
1644 The default behaviour can be restored with @option{--no-ctf-variables}.
1646 @cindex ctf type sharing
1647 @kindex --ctf-share-types
1648 @item --ctf-share-types=@var{method}
1649 Adjust the method used to share types between translation units in CTF.
1652 @item share-unconflicted
1653 Put all types that do not have ambiguous definitions into the shared dictionary,
1654 where debuggers can easily access them, even if they only occur in one
1655 translation unit. This is the default.
1657 @item share-duplicated
1658 Put only types that occur in multiple translation units into the shared
1659 dictionary: types with only one definition go into per-translation-unit
1660 dictionaries. Types with ambiguous definitions in multiple translation units
1661 always go into per-translation-unit dictionaries. This tends to make the CTF
1662 larger, but may reduce the amount of CTF in the shared dictionary. For very
1663 large projects this may speed up opening the CTF and save memory in the CTF
1664 consumer at runtime.
1667 @cindex common allocation
1668 @kindex --no-define-common
1669 @item --no-define-common
1670 This option inhibits the assignment of addresses to common symbols.
1671 The script command @code{INHIBIT_COMMON_ALLOCATION} has the same effect.
1672 @xref{Miscellaneous Commands}.
1674 The @samp{--no-define-common} option allows decoupling
1675 the decision to assign addresses to Common symbols from the choice
1676 of the output file type; otherwise a non-Relocatable output type
1677 forces assigning addresses to Common symbols.
1678 Using @samp{--no-define-common} allows Common symbols that are referenced
1679 from a shared library to be assigned addresses only in the main program.
1680 This eliminates the unused duplicate space in the shared library,
1681 and also prevents any possible confusion over resolving to the wrong
1682 duplicate when there are many dynamic modules with specialized search
1683 paths for runtime symbol resolution.
1685 @cindex group allocation in linker script
1686 @cindex section groups
1688 @kindex --force-group-allocation
1689 @item --force-group-allocation
1690 This option causes the linker to place section group members like
1691 normal input sections, and to delete the section groups. This is the
1692 default behaviour for a final link but this option can be used to
1693 change the behaviour of a relocatable link (@samp{-r}). The script
1694 command @code{FORCE_GROUP_ALLOCATION} has the same
1695 effect. @xref{Miscellaneous Commands}.
1697 @cindex symbols, from command line
1698 @kindex --defsym=@var{symbol}=@var{exp}
1699 @item --defsym=@var{symbol}=@var{expression}
1700 Create a global symbol in the output file, containing the absolute
1701 address given by @var{expression}. You may use this option as many
1702 times as necessary to define multiple symbols in the command line. A
1703 limited form of arithmetic is supported for the @var{expression} in this
1704 context: you may give a hexadecimal constant or the name of an existing
1705 symbol, or use @code{+} and @code{-} to add or subtract hexadecimal
1706 constants or symbols. If you need more elaborate expressions, consider
1707 using the linker command language from a script (@pxref{Assignments}).
1708 @emph{Note:} there should be no white space between @var{symbol}, the
1709 equals sign (``@key{=}''), and @var{expression}.
1711 The linker processes @samp{--defsym} arguments and @samp{-T} arguments
1712 in order, placing @samp{--defsym} before @samp{-T} will define the
1713 symbol before the linker script from @samp{-T} is processed, while
1714 placing @samp{--defsym} after @samp{-T} will define the symbol after
1715 the linker script has been processed. This difference has
1716 consequences for expressions within the linker script that use the
1717 @samp{--defsym} symbols, which order is correct will depend on what
1718 you are trying to achieve.
1720 @cindex demangling, from command line
1721 @kindex --demangle[=@var{style}]
1722 @kindex --no-demangle
1723 @item --demangle[=@var{style}]
1724 @itemx --no-demangle
1725 These options control whether to demangle symbol names in error messages
1726 and other output. When the linker is told to demangle, it tries to
1727 present symbol names in a readable fashion: it strips leading
1728 underscores if they are used by the object file format, and converts C++
1729 mangled symbol names into user readable names. Different compilers have
1730 different mangling styles. The optional demangling style argument can be used
1731 to choose an appropriate demangling style for your compiler. The linker will
1732 demangle by default unless the environment variable @samp{COLLECT_NO_DEMANGLE}
1733 is set. These options may be used to override the default.
1735 @cindex dynamic linker, from command line
1736 @kindex -I@var{file}
1737 @kindex --dynamic-linker=@var{file}
1739 @itemx --dynamic-linker=@var{file}
1740 Set the name of the dynamic linker. This is only meaningful when
1741 generating dynamically linked ELF executables. The default dynamic
1742 linker is normally correct; don't use this unless you know what you are
1745 @kindex --no-dynamic-linker
1746 @item --no-dynamic-linker
1747 When producing an executable file, omit the request for a dynamic
1748 linker to be used at load-time. This is only meaningful for ELF
1749 executables that contain dynamic relocations, and usually requires
1750 entry point code that is capable of processing these relocations.
1752 @kindex --embedded-relocs
1753 @item --embedded-relocs
1754 This option is similar to the @option{--emit-relocs} option except
1755 that the relocs are stored in a target-specific section. This option
1756 is only supported by the @samp{BFIN}, @samp{CR16} and @emph{M68K}
1759 @kindex --disable-multiple-abs-defs
1760 @item --disable-multiple-abs-defs
1761 Do not allow multiple definitions with symbols included
1762 in filename invoked by -R or --just-symbols
1764 @kindex --fatal-warnings
1765 @kindex --no-fatal-warnings
1766 @item --fatal-warnings
1767 @itemx --no-fatal-warnings
1768 Treat all warnings as errors. The default behaviour can be restored
1769 with the option @option{--no-fatal-warnings}.
1771 @kindex --force-exe-suffix
1772 @item --force-exe-suffix
1773 Make sure that an output file has a .exe suffix.
1775 If a successfully built fully linked output file does not have a
1776 @code{.exe} or @code{.dll} suffix, this option forces the linker to copy
1777 the output file to one of the same name with a @code{.exe} suffix. This
1778 option is useful when using unmodified Unix makefiles on a Microsoft
1779 Windows host, since some versions of Windows won't run an image unless
1780 it ends in a @code{.exe} suffix.
1782 @kindex --gc-sections
1783 @kindex --no-gc-sections
1784 @cindex garbage collection
1786 @itemx --no-gc-sections
1787 Enable garbage collection of unused input sections. It is ignored on
1788 targets that do not support this option. The default behaviour (of not
1789 performing this garbage collection) can be restored by specifying
1790 @samp{--no-gc-sections} on the command line. Note that garbage
1791 collection for COFF and PE format targets is supported, but the
1792 implementation is currently considered to be experimental.
1794 @samp{--gc-sections} decides which input sections are used by
1795 examining symbols and relocations. The section containing the entry
1796 symbol and all sections containing symbols undefined on the
1797 command-line will be kept, as will sections containing symbols
1798 referenced by dynamic objects. Note that when building shared
1799 libraries, the linker must assume that any visible symbol is
1800 referenced. Once this initial set of sections has been determined,
1801 the linker recursively marks as used any section referenced by their
1802 relocations. See @samp{--entry}, @samp{--undefined}, and
1803 @samp{--gc-keep-exported}.
1805 This option can be set when doing a partial link (enabled with option
1806 @samp{-r}). In this case the root of symbols kept must be explicitly
1807 specified either by one of the options @samp{--entry},
1808 @samp{--undefined}, or @samp{--gc-keep-exported} or by a @code{ENTRY}
1809 command in the linker script.
1811 As a GNU extension, ELF input sections marked with the
1812 @code{SHF_GNU_RETAIN} flag will not be garbage collected.
1814 @kindex --print-gc-sections
1815 @kindex --no-print-gc-sections
1816 @cindex garbage collection
1817 @item --print-gc-sections
1818 @itemx --no-print-gc-sections
1819 List all sections removed by garbage collection. The listing is
1820 printed on stderr. This option is only effective if garbage
1821 collection has been enabled via the @samp{--gc-sections}) option. The
1822 default behaviour (of not listing the sections that are removed) can
1823 be restored by specifying @samp{--no-print-gc-sections} on the command
1826 @kindex --gc-keep-exported
1827 @cindex garbage collection
1828 @item --gc-keep-exported
1829 When @samp{--gc-sections} is enabled, this option prevents garbage
1830 collection of unused input sections that contain global symbols having
1831 default or protected visibility. This option is intended to be used for
1832 executables where unreferenced sections would otherwise be garbage
1833 collected regardless of the external visibility of contained symbols.
1834 Note that this option has no effect when linking shared objects since
1835 it is already the default behaviour. This option is only supported for
1838 @kindex --print-output-format
1839 @cindex output format
1840 @item --print-output-format
1841 Print the name of the default output format (perhaps influenced by
1842 other command-line options). This is the string that would appear
1843 in an @code{OUTPUT_FORMAT} linker script command (@pxref{File Commands}).
1845 @kindex --print-memory-usage
1846 @cindex memory usage
1847 @item --print-memory-usage
1848 Print used size, total size and used size of memory regions created with
1849 the @ref{MEMORY} command. This is useful on embedded targets to have a
1850 quick view of amount of free memory. The format of the output has one
1851 headline and one line per region. It is both human readable and easily
1852 parsable by tools. Here is an example of an output:
1855 Memory region Used Size Region Size %age Used
1856 ROM: 256 KB 1 MB 25.00%
1857 RAM: 32 B 2 GB 0.00%
1864 Print a summary of the command-line options on the standard output and exit.
1866 @kindex --target-help
1868 Print a summary of all target-specific options on the standard output and exit.
1870 @kindex -Map=@var{mapfile}
1871 @item -Map=@var{mapfile}
1872 Print a link map to the file @var{mapfile}. See the description of the
1873 @option{-M} option, above. If @var{mapfile} is just the character
1874 @code{-} then the map will be written to stdout.
1876 Specifying a directory as @var{mapfile} causes the linker map to be
1877 written as a file inside the directory. Normally name of the file
1878 inside the directory is computed as the basename of the @var{output}
1879 file with @code{.map} appended. If however the special character
1880 @code{%} is used then this will be replaced by the full path of the
1881 output file. Additionally if there are any characters after the
1882 @var{%} symbol then @code{.map} will no longer be appended.
1885 -o foo.exe -Map=bar [Creates ./bar]
1886 -o ../dir/foo.exe -Map=bar [Creates ./bar]
1887 -o foo.exe -Map=../dir [Creates ../dir/foo.exe.map]
1888 -o ../dir2/foo.exe -Map=../dir [Creates ../dir/foo.exe.map]
1889 -o foo.exe -Map=% [Creates ./foo.exe.map]
1890 -o ../dir/foo.exe -Map=% [Creates ../dir/foo.exe.map]
1891 -o foo.exe -Map=%.bar [Creates ./foo.exe.bar]
1892 -o ../dir/foo.exe -Map=%.bar [Creates ../dir/foo.exe.bar]
1893 -o ../dir2/foo.exe -Map=../dir/% [Creates ../dir/../dir2/foo.exe.map]
1894 -o ../dir2/foo.exe -Map=../dir/%.bar [Creates ../dir/../dir2/foo.exe.bar]
1897 It is an error to specify more than one @code{%} character.
1899 If the map file already exists then it will be overwritten by this
1902 @cindex memory usage
1903 @kindex --no-keep-memory
1904 @item --no-keep-memory
1905 @command{ld} normally optimizes for speed over memory usage by caching the
1906 symbol tables of input files in memory. This option tells @command{ld} to
1907 instead optimize for memory usage, by rereading the symbol tables as
1908 necessary. This may be required if @command{ld} runs out of memory space
1909 while linking a large executable.
1911 @kindex --no-undefined
1914 @item --no-undefined
1916 Report unresolved symbol references from regular object files. This
1917 is done even if the linker is creating a non-symbolic shared library.
1918 The switch @option{--[no-]allow-shlib-undefined} controls the
1919 behaviour for reporting unresolved references found in shared
1920 libraries being linked in.
1922 The effects of this option can be reverted by using @code{-z undefs}.
1924 @kindex --allow-multiple-definition
1926 @item --allow-multiple-definition
1928 Normally when a symbol is defined multiple times, the linker will
1929 report a fatal error. These options allow multiple definitions and the
1930 first definition will be used.
1932 @kindex --allow-shlib-undefined
1933 @kindex --no-allow-shlib-undefined
1934 @item --allow-shlib-undefined
1935 @itemx --no-allow-shlib-undefined
1936 Allows or disallows undefined symbols in shared libraries.
1937 This switch is similar to @option{--no-undefined} except that it
1938 determines the behaviour when the undefined symbols are in a
1939 shared library rather than a regular object file. It does not affect
1940 how undefined symbols in regular object files are handled.
1942 The default behaviour is to report errors for any undefined symbols
1943 referenced in shared libraries if the linker is being used to create
1944 an executable, but to allow them if the linker is being used to create
1947 The reasons for allowing undefined symbol references in shared
1948 libraries specified at link time are that:
1952 A shared library specified at link time may not be the same as the one
1953 that is available at load time, so the symbol might actually be
1954 resolvable at load time.
1956 There are some operating systems, eg BeOS and HPPA, where undefined
1957 symbols in shared libraries are normal.
1959 The BeOS kernel for example patches shared libraries at load time to
1960 select whichever function is most appropriate for the current
1961 architecture. This is used, for example, to dynamically select an
1962 appropriate memset function.
1965 @kindex --error-handling-script=@var{scriptname}
1966 @item --error-handling-script=@var{scriptname}
1967 If this option is provided then the linker will invoke
1968 @var{scriptname} whenever an error is encountered. Currently however
1969 only two kinds of error are supported: missing symbols and missing
1970 libraries. Two arguments will be passed to script: the keyword
1971 ``undefined-symbol'' or `missing-lib'' and the @var{name} of the
1972 undefined symbol or missing library. The intention is that the script
1973 will provide suggestions to the user as to where the symbol or library
1974 might be found. After the script has finished then the normal linker
1975 error message will be displayed.
1977 The availability of this option is controlled by a configure time
1978 switch, so it may not be present in specific implementations.
1980 @kindex --no-undefined-version
1981 @item --no-undefined-version
1982 Normally when a symbol has an undefined version, the linker will ignore
1983 it. This option disallows symbols with undefined version and a fatal error
1984 will be issued instead.
1986 @kindex --default-symver
1987 @item --default-symver
1988 Create and use a default symbol version (the soname) for unversioned
1991 @kindex --default-imported-symver
1992 @item --default-imported-symver
1993 Create and use a default symbol version (the soname) for unversioned
1996 @kindex --no-warn-mismatch
1997 @item --no-warn-mismatch
1998 Normally @command{ld} will give an error if you try to link together input
1999 files that are mismatched for some reason, perhaps because they have
2000 been compiled for different processors or for different endiannesses.
2001 This option tells @command{ld} that it should silently permit such possible
2002 errors. This option should only be used with care, in cases when you
2003 have taken some special action that ensures that the linker errors are
2006 @kindex --no-warn-search-mismatch
2007 @item --no-warn-search-mismatch
2008 Normally @command{ld} will give a warning if it finds an incompatible
2009 library during a library search. This option silences the warning.
2011 @kindex --no-whole-archive
2012 @item --no-whole-archive
2013 Turn off the effect of the @option{--whole-archive} option for subsequent
2016 @cindex output file after errors
2017 @kindex --noinhibit-exec
2018 @item --noinhibit-exec
2019 Retain the executable output file whenever it is still usable.
2020 Normally, the linker will not produce an output file if it encounters
2021 errors during the link process; it exits without writing an output file
2022 when it issues any error whatsoever.
2026 Only search library directories explicitly specified on the
2027 command line. Library directories specified in linker scripts
2028 (including linker scripts specified on the command line) are ignored.
2030 @ifclear SingleFormat
2031 @kindex --oformat=@var{output-format}
2032 @item --oformat=@var{output-format}
2033 @command{ld} may be configured to support more than one kind of object
2034 file. If your @command{ld} is configured this way, you can use the
2035 @samp{--oformat} option to specify the binary format for the output
2036 object file. Even when @command{ld} is configured to support alternative
2037 object formats, you don't usually need to specify this, as @command{ld}
2038 should be configured to produce as a default output format the most
2039 usual format on each machine. @var{output-format} is a text string, the
2040 name of a particular format supported by the BFD libraries. (You can
2041 list the available binary formats with @samp{objdump -i}.) The script
2042 command @code{OUTPUT_FORMAT} can also specify the output format, but
2043 this option overrides it. @xref{BFD}.
2046 @kindex --out-implib
2047 @item --out-implib @var{file}
2048 Create an import library in @var{file} corresponding to the executable
2049 the linker is generating (eg. a DLL or ELF program). This import
2050 library (which should be called @code{*.dll.a} or @code{*.a} for DLLs)
2051 may be used to link clients against the generated executable; this
2052 behaviour makes it possible to skip a separate import library creation
2053 step (eg. @code{dlltool} for DLLs). This option is only available for
2054 the i386 PE and ELF targetted ports of the linker.
2057 @kindex --pic-executable
2059 @itemx --pic-executable
2060 @cindex position independent executables
2061 Create a position independent executable. This is currently only supported on
2062 ELF platforms. Position independent executables are similar to shared
2063 libraries in that they are relocated by the dynamic linker to the virtual
2064 address the OS chooses for them (which can vary between invocations). Like
2065 normal dynamically linked executables they can be executed and symbols
2066 defined in the executable cannot be overridden by shared libraries.
2070 This option is ignored for Linux compatibility.
2074 This option is ignored for SVR4 compatibility.
2077 @cindex synthesizing linker
2078 @cindex relaxing addressing modes
2082 An option with machine dependent effects.
2084 This option is only supported on a few targets.
2087 @xref{H8/300,,@command{ld} and the H8/300}.
2090 @xref{Xtensa,, @command{ld} and Xtensa Processors}.
2093 @xref{M68HC11/68HC12,,@command{ld} and the 68HC11 and 68HC12}.
2096 @xref{Nios II,,@command{ld} and the Altera Nios II}.
2099 @xref{PowerPC ELF32,,@command{ld} and PowerPC 32-bit ELF Support}.
2102 On some platforms the @option{--relax} option performs target specific,
2103 global optimizations that become possible when the linker resolves
2104 addressing in the program, such as relaxing address modes,
2105 synthesizing new instructions, selecting shorter version of current
2106 instructions, and combining constant values.
2108 On some platforms these link time global optimizations may make symbolic
2109 debugging of the resulting executable impossible.
2111 This is known to be the case for the Matsushita MN10200 and MN10300
2112 family of processors.
2115 On platforms where the feature is supported, the option
2116 @option{--no-relax} will disable it.
2118 On platforms where the feature is not supported, both @option{--relax}
2119 and @option{--no-relax} are accepted, but ignored.
2121 @cindex retaining specified symbols
2122 @cindex stripping all but some symbols
2123 @cindex symbols, retaining selectively
2124 @kindex --retain-symbols-file=@var{filename}
2125 @item --retain-symbols-file=@var{filename}
2126 Retain @emph{only} the symbols listed in the file @var{filename},
2127 discarding all others. @var{filename} is simply a flat file, with one
2128 symbol name per line. This option is especially useful in environments
2132 where a large global symbol table is accumulated gradually, to conserve
2135 @samp{--retain-symbols-file} does @emph{not} discard undefined symbols,
2136 or symbols needed for relocations.
2138 You may only specify @samp{--retain-symbols-file} once in the command
2139 line. It overrides @samp{-s} and @samp{-S}.
2142 @item -rpath=@var{dir}
2143 @cindex runtime library search path
2144 @kindex -rpath=@var{dir}
2145 Add a directory to the runtime library search path. This is used when
2146 linking an ELF executable with shared objects. All @option{-rpath}
2147 arguments are concatenated and passed to the runtime linker, which uses
2148 them to locate shared objects at runtime.
2150 The @option{-rpath} option is also used when locating shared objects which
2151 are needed by shared objects explicitly included in the link; see the
2152 description of the @option{-rpath-link} option. Searching @option{-rpath}
2153 in this way is only supported by native linkers and cross linkers which
2154 have been configured with the @option{--with-sysroot} option.
2156 If @option{-rpath} is not used when linking an ELF executable, the
2157 contents of the environment variable @code{LD_RUN_PATH} will be used if it
2160 The @option{-rpath} option may also be used on SunOS. By default, on
2161 SunOS, the linker will form a runtime search path out of all the
2162 @option{-L} options it is given. If a @option{-rpath} option is used, the
2163 runtime search path will be formed exclusively using the @option{-rpath}
2164 options, ignoring the @option{-L} options. This can be useful when using
2165 gcc, which adds many @option{-L} options which may be on NFS mounted
2168 For compatibility with other ELF linkers, if the @option{-R} option is
2169 followed by a directory name, rather than a file name, it is treated as
2170 the @option{-rpath} option.
2174 @cindex link-time runtime library search path
2175 @kindex -rpath-link=@var{dir}
2176 @item -rpath-link=@var{dir}
2177 When using ELF or SunOS, one shared library may require another. This
2178 happens when an @code{ld -shared} link includes a shared library as one
2181 When the linker encounters such a dependency when doing a non-shared,
2182 non-relocatable link, it will automatically try to locate the required
2183 shared library and include it in the link, if it is not included
2184 explicitly. In such a case, the @option{-rpath-link} option
2185 specifies the first set of directories to search. The
2186 @option{-rpath-link} option may specify a sequence of directory names
2187 either by specifying a list of names separated by colons, or by
2188 appearing multiple times.
2190 The tokens @var{$ORIGIN} and @var{$LIB} can appear in these search
2191 directories. They will be replaced by the full path to the directory
2192 containing the program or shared object in the case of @var{$ORIGIN}
2193 and either @samp{lib} - for 32-bit binaries - or @samp{lib64} - for
2194 64-bit binaries - in the case of @var{$LIB}.
2196 The alternative form of these tokens - @var{$@{ORIGIN@}} and
2197 @var{$@{LIB@}} can also be used. The token @var{$PLATFORM} is not
2200 This option should be used with caution as it overrides the search path
2201 that may have been hard compiled into a shared library. In such a case it
2202 is possible to use unintentionally a different search path than the
2203 runtime linker would do.
2205 The linker uses the following search paths to locate required shared
2210 Any directories specified by @option{-rpath-link} options.
2212 Any directories specified by @option{-rpath} options. The difference
2213 between @option{-rpath} and @option{-rpath-link} is that directories
2214 specified by @option{-rpath} options are included in the executable and
2215 used at runtime, whereas the @option{-rpath-link} option is only effective
2216 at link time. Searching @option{-rpath} in this way is only supported
2217 by native linkers and cross linkers which have been configured with
2218 the @option{--with-sysroot} option.
2220 On an ELF system, for native linkers, if the @option{-rpath} and
2221 @option{-rpath-link} options were not used, search the contents of the
2222 environment variable @code{LD_RUN_PATH}.
2224 On SunOS, if the @option{-rpath} option was not used, search any
2225 directories specified using @option{-L} options.
2227 For a native linker, search the contents of the environment
2228 variable @code{LD_LIBRARY_PATH}.
2230 For a native ELF linker, the directories in @code{DT_RUNPATH} or
2231 @code{DT_RPATH} of a shared library are searched for shared
2232 libraries needed by it. The @code{DT_RPATH} entries are ignored if
2233 @code{DT_RUNPATH} entries exist.
2235 The default directories, normally @file{/lib} and @file{/usr/lib}.
2237 For a linker for a Linux system, if the file @file{/etc/ld.so.conf}
2238 exists, the list of directories found in that file. Note: the path
2239 to this file is prefixed with the @code{sysroot} value, if that is
2240 defined, and then any @code{prefix} string if the linker was
2241 configured with the @command{--prefix=<path>} option.
2243 For a native linker on a FreeBSD system, any directories specified by
2244 the @code{_PATH_ELF_HINTS} macro defined in the @file{elf-hints.h}
2247 Any directories specifed by a @code{SEARCH_DIR} command in the
2248 linker script being used.
2251 If the required shared library is not found, the linker will issue a
2252 warning and continue with the link.
2259 @cindex shared libraries
2260 Create a shared library. This is currently only supported on ELF, XCOFF
2261 and SunOS platforms. On SunOS, the linker will automatically create a
2262 shared library if the @option{-e} option is not used and there are
2263 undefined symbols in the link.
2265 @kindex --sort-common
2267 @itemx --sort-common=ascending
2268 @itemx --sort-common=descending
2269 This option tells @command{ld} to sort the common symbols by alignment in
2270 ascending or descending order when it places them in the appropriate output
2271 sections. The symbol alignments considered are sixteen-byte or larger,
2272 eight-byte, four-byte, two-byte, and one-byte. This is to prevent gaps
2273 between symbols due to alignment constraints. If no sorting order is
2274 specified, then descending order is assumed.
2276 @kindex --sort-section=name
2277 @item --sort-section=name
2278 This option will apply @code{SORT_BY_NAME} to all wildcard section
2279 patterns in the linker script.
2281 @kindex --sort-section=alignment
2282 @item --sort-section=alignment
2283 This option will apply @code{SORT_BY_ALIGNMENT} to all wildcard section
2284 patterns in the linker script.
2286 @kindex --spare-dynamic-tags
2287 @item --spare-dynamic-tags=@var{count}
2288 This option specifies the number of empty slots to leave in the
2289 .dynamic section of ELF shared objects. Empty slots may be needed by
2290 post processing tools, such as the prelinker. The default is 5.
2292 @kindex --split-by-file
2293 @item --split-by-file[=@var{size}]
2294 Similar to @option{--split-by-reloc} but creates a new output section for
2295 each input file when @var{size} is reached. @var{size} defaults to a
2296 size of 1 if not given.
2298 @kindex --split-by-reloc
2299 @item --split-by-reloc[=@var{count}]
2300 Tries to creates extra sections in the output file so that no single
2301 output section in the file contains more than @var{count} relocations.
2302 This is useful when generating huge relocatable files for downloading into
2303 certain real time kernels with the COFF object file format; since COFF
2304 cannot represent more than 65535 relocations in a single section. Note
2305 that this will fail to work with object file formats which do not
2306 support arbitrary sections. The linker will not split up individual
2307 input sections for redistribution, so if a single input section contains
2308 more than @var{count} relocations one output section will contain that
2309 many relocations. @var{count} defaults to a value of 32768.
2313 Compute and display statistics about the operation of the linker, such
2314 as execution time and memory usage.
2316 @kindex --sysroot=@var{directory}
2317 @item --sysroot=@var{directory}
2318 Use @var{directory} as the location of the sysroot, overriding the
2319 configure-time default. This option is only supported by linkers
2320 that were configured using @option{--with-sysroot}.
2324 This is used by COFF/PE based targets to create a task-linked object
2325 file where all of the global symbols have been converted to statics.
2327 @kindex --traditional-format
2328 @cindex traditional format
2329 @item --traditional-format
2330 For some targets, the output of @command{ld} is different in some ways from
2331 the output of some existing linker. This switch requests @command{ld} to
2332 use the traditional format instead.
2335 For example, on SunOS, @command{ld} combines duplicate entries in the
2336 symbol string table. This can reduce the size of an output file with
2337 full debugging information by over 30 percent. Unfortunately, the SunOS
2338 @code{dbx} program can not read the resulting program (@code{gdb} has no
2339 trouble). The @samp{--traditional-format} switch tells @command{ld} to not
2340 combine duplicate entries.
2342 @kindex --section-start=@var{sectionname}=@var{org}
2343 @item --section-start=@var{sectionname}=@var{org}
2344 Locate a section in the output file at the absolute
2345 address given by @var{org}. You may use this option as many
2346 times as necessary to locate multiple sections in the command
2348 @var{org} must be a single hexadecimal integer;
2349 for compatibility with other linkers, you may omit the leading
2350 @samp{0x} usually associated with hexadecimal values. @emph{Note:} there
2351 should be no white space between @var{sectionname}, the equals
2352 sign (``@key{=}''), and @var{org}.
2354 @kindex -Tbss=@var{org}
2355 @kindex -Tdata=@var{org}
2356 @kindex -Ttext=@var{org}
2357 @cindex segment origins, cmd line
2358 @item -Tbss=@var{org}
2359 @itemx -Tdata=@var{org}
2360 @itemx -Ttext=@var{org}
2361 Same as @option{--section-start}, with @code{.bss}, @code{.data} or
2362 @code{.text} as the @var{sectionname}.
2364 @kindex -Ttext-segment=@var{org}
2365 @item -Ttext-segment=@var{org}
2366 @cindex text segment origin, cmd line
2367 When creating an ELF executable, it will set the address of the first
2368 byte of the text segment.
2370 @kindex -Trodata-segment=@var{org}
2371 @item -Trodata-segment=@var{org}
2372 @cindex rodata segment origin, cmd line
2373 When creating an ELF executable or shared object for a target where
2374 the read-only data is in its own segment separate from the executable
2375 text, it will set the address of the first byte of the read-only data segment.
2377 @kindex -Tldata-segment=@var{org}
2378 @item -Tldata-segment=@var{org}
2379 @cindex ldata segment origin, cmd line
2380 When creating an ELF executable or shared object for x86-64 medium memory
2381 model, it will set the address of the first byte of the ldata segment.
2383 @kindex --unresolved-symbols
2384 @item --unresolved-symbols=@var{method}
2385 Determine how to handle unresolved symbols. There are four possible
2386 values for @samp{method}:
2390 Do not report any unresolved symbols.
2393 Report all unresolved symbols. This is the default.
2395 @item ignore-in-object-files
2396 Report unresolved symbols that are contained in shared libraries, but
2397 ignore them if they come from regular object files.
2399 @item ignore-in-shared-libs
2400 Report unresolved symbols that come from regular object files, but
2401 ignore them if they come from shared libraries. This can be useful
2402 when creating a dynamic binary and it is known that all the shared
2403 libraries that it should be referencing are included on the linker's
2407 The behaviour for shared libraries on their own can also be controlled
2408 by the @option{--[no-]allow-shlib-undefined} option.
2410 Normally the linker will generate an error message for each reported
2411 unresolved symbol but the option @option{--warn-unresolved-symbols}
2412 can change this to a warning.
2414 @kindex --verbose[=@var{NUMBER}]
2415 @cindex verbose[=@var{NUMBER}]
2417 @itemx --verbose[=@var{NUMBER}]
2418 Display the version number for @command{ld} and list the linker emulations
2419 supported. Display which input files can and cannot be opened. Display
2420 the linker script being used by the linker. If the optional @var{NUMBER}
2421 argument > 1, plugin symbol status will also be displayed.
2423 @kindex --version-script=@var{version-scriptfile}
2424 @cindex version script, symbol versions
2425 @item --version-script=@var{version-scriptfile}
2426 Specify the name of a version script to the linker. This is typically
2427 used when creating shared libraries to specify additional information
2428 about the version hierarchy for the library being created. This option
2429 is only fully supported on ELF platforms which support shared libraries;
2430 see @ref{VERSION}. It is partially supported on PE platforms, which can
2431 use version scripts to filter symbol visibility in auto-export mode: any
2432 symbols marked @samp{local} in the version script will not be exported.
2435 @kindex --warn-common
2436 @cindex warnings, on combining symbols
2437 @cindex combining symbols, warnings on
2439 Warn when a common symbol is combined with another common symbol or with
2440 a symbol definition. Unix linkers allow this somewhat sloppy practice,
2441 but linkers on some other operating systems do not. This option allows
2442 you to find potential problems from combining global symbols.
2443 Unfortunately, some C libraries use this practice, so you may get some
2444 warnings about symbols in the libraries as well as in your programs.
2446 There are three kinds of global symbols, illustrated here by C examples:
2450 A definition, which goes in the initialized data section of the output
2454 An undefined reference, which does not allocate space.
2455 There must be either a definition or a common symbol for the
2459 A common symbol. If there are only (one or more) common symbols for a
2460 variable, it goes in the uninitialized data area of the output file.
2461 The linker merges multiple common symbols for the same variable into a
2462 single symbol. If they are of different sizes, it picks the largest
2463 size. The linker turns a common symbol into a declaration, if there is
2464 a definition of the same variable.
2467 The @samp{--warn-common} option can produce five kinds of warnings.
2468 Each warning consists of a pair of lines: the first describes the symbol
2469 just encountered, and the second describes the previous symbol
2470 encountered with the same name. One or both of the two symbols will be
2475 Turning a common symbol into a reference, because there is already a
2476 definition for the symbol.
2478 @var{file}(@var{section}): warning: common of `@var{symbol}'
2479 overridden by definition
2480 @var{file}(@var{section}): warning: defined here
2484 Turning a common symbol into a reference, because a later definition for
2485 the symbol is encountered. This is the same as the previous case,
2486 except that the symbols are encountered in a different order.
2488 @var{file}(@var{section}): warning: definition of `@var{symbol}'
2490 @var{file}(@var{section}): warning: common is here
2494 Merging a common symbol with a previous same-sized common symbol.
2496 @var{file}(@var{section}): warning: multiple common
2498 @var{file}(@var{section}): warning: previous common is here
2502 Merging a common symbol with a previous larger common symbol.
2504 @var{file}(@var{section}): warning: common of `@var{symbol}'
2505 overridden by larger common
2506 @var{file}(@var{section}): warning: larger common is here
2510 Merging a common symbol with a previous smaller common symbol. This is
2511 the same as the previous case, except that the symbols are
2512 encountered in a different order.
2514 @var{file}(@var{section}): warning: common of `@var{symbol}'
2515 overriding smaller common
2516 @var{file}(@var{section}): warning: smaller common is here
2520 @kindex --warn-constructors
2521 @item --warn-constructors
2522 Warn if any global constructors are used. This is only useful for a few
2523 object file formats. For formats like COFF or ELF, the linker can not
2524 detect the use of global constructors.
2526 @kindex --warn-multiple-gp
2527 @item --warn-multiple-gp
2528 Warn if multiple global pointer values are required in the output file.
2529 This is only meaningful for certain processors, such as the Alpha.
2530 Specifically, some processors put large-valued constants in a special
2531 section. A special register (the global pointer) points into the middle
2532 of this section, so that constants can be loaded efficiently via a
2533 base-register relative addressing mode. Since the offset in
2534 base-register relative mode is fixed and relatively small (e.g., 16
2535 bits), this limits the maximum size of the constant pool. Thus, in
2536 large programs, it is often necessary to use multiple global pointer
2537 values in order to be able to address all possible constants. This
2538 option causes a warning to be issued whenever this case occurs.
2541 @cindex warnings, on undefined symbols
2542 @cindex undefined symbols, warnings on
2544 Only warn once for each undefined symbol, rather than once per module
2547 @kindex --warn-section-align
2548 @cindex warnings, on section alignment
2549 @cindex section alignment, warnings on
2550 @item --warn-section-align
2551 Warn if the address of an output section is changed because of
2552 alignment. Typically, the alignment will be set by an input section.
2553 The address will only be changed if it not explicitly specified; that
2554 is, if the @code{SECTIONS} command does not specify a start address for
2555 the section (@pxref{SECTIONS}).
2557 @kindex --warn-textrel
2558 @item --warn-textrel
2559 Warn if the linker adds DT_TEXTREL to a position-independent executable
2562 @kindex --warn-alternate-em
2563 @item --warn-alternate-em
2564 Warn if an object has alternate ELF machine code.
2566 @kindex --warn-unresolved-symbols
2567 @item --warn-unresolved-symbols
2568 If the linker is going to report an unresolved symbol (see the option
2569 @option{--unresolved-symbols}) it will normally generate an error.
2570 This option makes it generate a warning instead.
2572 @kindex --error-unresolved-symbols
2573 @item --error-unresolved-symbols
2574 This restores the linker's default behaviour of generating errors when
2575 it is reporting unresolved symbols.
2577 @kindex --whole-archive
2578 @cindex including an entire archive
2579 @item --whole-archive
2580 For each archive mentioned on the command line after the
2581 @option{--whole-archive} option, include every object file in the archive
2582 in the link, rather than searching the archive for the required object
2583 files. This is normally used to turn an archive file into a shared
2584 library, forcing every object to be included in the resulting shared
2585 library. This option may be used more than once.
2587 Two notes when using this option from gcc: First, gcc doesn't know
2588 about this option, so you have to use @option{-Wl,-whole-archive}.
2589 Second, don't forget to use @option{-Wl,-no-whole-archive} after your
2590 list of archives, because gcc will add its own list of archives to
2591 your link and you may not want this flag to affect those as well.
2593 @kindex --wrap=@var{symbol}
2594 @item --wrap=@var{symbol}
2595 Use a wrapper function for @var{symbol}. Any undefined reference to
2596 @var{symbol} will be resolved to @code{__wrap_@var{symbol}}. Any
2597 undefined reference to @code{__real_@var{symbol}} will be resolved to
2600 This can be used to provide a wrapper for a system function. The
2601 wrapper function should be called @code{__wrap_@var{symbol}}. If it
2602 wishes to call the system function, it should call
2603 @code{__real_@var{symbol}}.
2605 Here is a trivial example:
2609 __wrap_malloc (size_t c)
2611 printf ("malloc called with %zu\n", c);
2612 return __real_malloc (c);
2616 If you link other code with this file using @option{--wrap malloc}, then
2617 all calls to @code{malloc} will call the function @code{__wrap_malloc}
2618 instead. The call to @code{__real_malloc} in @code{__wrap_malloc} will
2619 call the real @code{malloc} function.
2621 You may wish to provide a @code{__real_malloc} function as well, so that
2622 links without the @option{--wrap} option will succeed. If you do this,
2623 you should not put the definition of @code{__real_malloc} in the same
2624 file as @code{__wrap_malloc}; if you do, the assembler may resolve the
2625 call before the linker has a chance to wrap it to @code{malloc}.
2627 Only undefined references are replaced by the linker. So, translation unit
2628 internal references to @var{symbol} are not resolved to
2629 @code{__wrap_@var{symbol}}. In the next example, the call to @code{f} in
2630 @code{g} is not resolved to @code{__wrap_f}.
2646 @kindex --eh-frame-hdr
2647 @kindex --no-eh-frame-hdr
2648 @item --eh-frame-hdr
2649 @itemx --no-eh-frame-hdr
2650 Request (@option{--eh-frame-hdr}) or suppress
2651 (@option{--no-eh-frame-hdr}) the creation of @code{.eh_frame_hdr}
2652 section and ELF @code{PT_GNU_EH_FRAME} segment header.
2654 @kindex --ld-generated-unwind-info
2655 @item --no-ld-generated-unwind-info
2656 Request creation of @code{.eh_frame} unwind info for linker
2657 generated code sections like PLT. This option is on by default
2658 if linker generated unwind info is supported.
2660 @kindex --enable-new-dtags
2661 @kindex --disable-new-dtags
2662 @item --enable-new-dtags
2663 @itemx --disable-new-dtags
2664 This linker can create the new dynamic tags in ELF. But the older ELF
2665 systems may not understand them. If you specify
2666 @option{--enable-new-dtags}, the new dynamic tags will be created as needed
2667 and older dynamic tags will be omitted.
2668 If you specify @option{--disable-new-dtags}, no new dynamic tags will be
2669 created. By default, the new dynamic tags are not created. Note that
2670 those options are only available for ELF systems.
2672 @kindex --hash-size=@var{number}
2673 @item --hash-size=@var{number}
2674 Set the default size of the linker's hash tables to a prime number
2675 close to @var{number}. Increasing this value can reduce the length of
2676 time it takes the linker to perform its tasks, at the expense of
2677 increasing the linker's memory requirements. Similarly reducing this
2678 value can reduce the memory requirements at the expense of speed.
2680 @kindex --hash-style=@var{style}
2681 @item --hash-style=@var{style}
2682 Set the type of linker's hash table(s). @var{style} can be either
2683 @code{sysv} for classic ELF @code{.hash} section, @code{gnu} for
2684 new style GNU @code{.gnu.hash} section or @code{both} for both
2685 the classic ELF @code{.hash} and new style GNU @code{.gnu.hash}
2686 hash tables. The default depends upon how the linker was configured,
2687 but for most Linux based systems it will be @code{both}.
2689 @kindex --compress-debug-sections=none
2690 @kindex --compress-debug-sections=zlib
2691 @kindex --compress-debug-sections=zlib-gnu
2692 @kindex --compress-debug-sections=zlib-gabi
2693 @item --compress-debug-sections=none
2694 @itemx --compress-debug-sections=zlib
2695 @itemx --compress-debug-sections=zlib-gnu
2696 @itemx --compress-debug-sections=zlib-gabi
2697 On ELF platforms, these options control how DWARF debug sections are
2698 compressed using zlib.
2700 @option{--compress-debug-sections=none} doesn't compress DWARF debug
2701 sections. @option{--compress-debug-sections=zlib-gnu} compresses
2702 DWARF debug sections and renames them to begin with @samp{.zdebug}
2703 instead of @samp{.debug}. @option{--compress-debug-sections=zlib-gabi}
2704 also compresses DWARF debug sections, but rather than renaming them it
2705 sets the SHF_COMPRESSED flag in the sections' headers.
2707 The @option{--compress-debug-sections=zlib} option is an alias for
2708 @option{--compress-debug-sections=zlib-gabi}.
2710 Note that this option overrides any compression in input debug
2711 sections, so if a binary is linked with @option{--compress-debug-sections=none}
2712 for example, then any compressed debug sections in input files will be
2713 uncompressed before they are copied into the output binary.
2715 The default compression behaviour varies depending upon the target
2716 involved and the configure options used to build the toolchain. The
2717 default can be determined by examining the output from the linker's
2718 @option{--help} option.
2720 @kindex --reduce-memory-overheads
2721 @item --reduce-memory-overheads
2722 This option reduces memory requirements at ld runtime, at the expense of
2723 linking speed. This was introduced to select the old O(n^2) algorithm
2724 for link map file generation, rather than the new O(n) algorithm which uses
2725 about 40% more memory for symbol storage.
2727 Another effect of the switch is to set the default hash table size to
2728 1021, which again saves memory at the cost of lengthening the linker's
2729 run time. This is not done however if the @option{--hash-size} switch
2732 The @option{--reduce-memory-overheads} switch may be also be used to
2733 enable other tradeoffs in future versions of the linker.
2736 @kindex --build-id=@var{style}
2738 @itemx --build-id=@var{style}
2739 Request the creation of a @code{.note.gnu.build-id} ELF note section
2740 or a @code{.buildid} COFF section. The contents of the note are
2741 unique bits identifying this linked file. @var{style} can be
2742 @code{uuid} to use 128 random bits, @code{sha1} to use a 160-bit
2743 @sc{SHA1} hash on the normative parts of the output contents,
2744 @code{md5} to use a 128-bit @sc{MD5} hash on the normative parts of
2745 the output contents, or @code{0x@var{hexstring}} to use a chosen bit
2746 string specified as an even number of hexadecimal digits (@code{-} and
2747 @code{:} characters between digit pairs are ignored). If @var{style}
2748 is omitted, @code{sha1} is used.
2750 The @code{md5} and @code{sha1} styles produces an identifier
2751 that is always the same in an identical output file, but will be
2752 unique among all nonidentical output files. It is not intended
2753 to be compared as a checksum for the file's contents. A linked
2754 file may be changed later by other tools, but the build ID bit
2755 string identifying the original linked file does not change.
2757 Passing @code{none} for @var{style} disables the setting from any
2758 @code{--build-id} options earlier on the command line.
2763 @subsection Options Specific to i386 PE Targets
2765 @c man begin OPTIONS
2767 The i386 PE linker supports the @option{-shared} option, which causes
2768 the output to be a dynamically linked library (DLL) instead of a
2769 normal executable. You should name the output @code{*.dll} when you
2770 use this option. In addition, the linker fully supports the standard
2771 @code{*.def} files, which may be specified on the linker command line
2772 like an object file (in fact, it should precede archives it exports
2773 symbols from, to ensure that they get linked in, just like a normal
2776 In addition to the options common to all targets, the i386 PE linker
2777 support additional command-line options that are specific to the i386
2778 PE target. Options that take values may be separated from their
2779 values by either a space or an equals sign.
2783 @kindex --add-stdcall-alias
2784 @item --add-stdcall-alias
2785 If given, symbols with a stdcall suffix (@@@var{nn}) will be exported
2786 as-is and also with the suffix stripped.
2787 [This option is specific to the i386 PE targeted port of the linker]
2790 @item --base-file @var{file}
2791 Use @var{file} as the name of a file in which to save the base
2792 addresses of all the relocations needed for generating DLLs with
2794 [This is an i386 PE specific option]
2798 Create a DLL instead of a regular executable. You may also use
2799 @option{-shared} or specify a @code{LIBRARY} in a given @code{.def}
2801 [This option is specific to the i386 PE targeted port of the linker]
2803 @kindex --enable-long-section-names
2804 @kindex --disable-long-section-names
2805 @item --enable-long-section-names
2806 @itemx --disable-long-section-names
2807 The PE variants of the COFF object format add an extension that permits
2808 the use of section names longer than eight characters, the normal limit
2809 for COFF. By default, these names are only allowed in object files, as
2810 fully-linked executable images do not carry the COFF string table required
2811 to support the longer names. As a GNU extension, it is possible to
2812 allow their use in executable images as well, or to (probably pointlessly!)
2813 disallow it in object files, by using these two options. Executable images
2814 generated with these long section names are slightly non-standard, carrying
2815 as they do a string table, and may generate confusing output when examined
2816 with non-GNU PE-aware tools, such as file viewers and dumpers. However,
2817 GDB relies on the use of PE long section names to find Dwarf-2 debug
2818 information sections in an executable image at runtime, and so if neither
2819 option is specified on the command-line, @command{ld} will enable long
2820 section names, overriding the default and technically correct behaviour,
2821 when it finds the presence of debug information while linking an executable
2822 image and not stripping symbols.
2823 [This option is valid for all PE targeted ports of the linker]
2825 @kindex --enable-stdcall-fixup
2826 @kindex --disable-stdcall-fixup
2827 @item --enable-stdcall-fixup
2828 @itemx --disable-stdcall-fixup
2829 If the link finds a symbol that it cannot resolve, it will attempt to
2830 do ``fuzzy linking'' by looking for another defined symbol that differs
2831 only in the format of the symbol name (cdecl vs stdcall) and will
2832 resolve that symbol by linking to the match. For example, the
2833 undefined symbol @code{_foo} might be linked to the function
2834 @code{_foo@@12}, or the undefined symbol @code{_bar@@16} might be linked
2835 to the function @code{_bar}. When the linker does this, it prints a
2836 warning, since it normally should have failed to link, but sometimes
2837 import libraries generated from third-party dlls may need this feature
2838 to be usable. If you specify @option{--enable-stdcall-fixup}, this
2839 feature is fully enabled and warnings are not printed. If you specify
2840 @option{--disable-stdcall-fixup}, this feature is disabled and such
2841 mismatches are considered to be errors.
2842 [This option is specific to the i386 PE targeted port of the linker]
2844 @kindex --leading-underscore
2845 @kindex --no-leading-underscore
2846 @item --leading-underscore
2847 @itemx --no-leading-underscore
2848 For most targets default symbol-prefix is an underscore and is defined
2849 in target's description. By this option it is possible to
2850 disable/enable the default underscore symbol-prefix.
2852 @cindex DLLs, creating
2853 @kindex --export-all-symbols
2854 @item --export-all-symbols
2855 If given, all global symbols in the objects used to build a DLL will
2856 be exported by the DLL. Note that this is the default if there
2857 otherwise wouldn't be any exported symbols. When symbols are
2858 explicitly exported via DEF files or implicitly exported via function
2859 attributes, the default is to not export anything else unless this
2860 option is given. Note that the symbols @code{DllMain@@12},
2861 @code{DllEntryPoint@@0}, @code{DllMainCRTStartup@@12}, and
2862 @code{impure_ptr} will not be automatically
2863 exported. Also, symbols imported from other DLLs will not be
2864 re-exported, nor will symbols specifying the DLL's internal layout
2865 such as those beginning with @code{_head_} or ending with
2866 @code{_iname}. In addition, no symbols from @code{libgcc},
2867 @code{libstd++}, @code{libmingw32}, or @code{crtX.o} will be exported.
2868 Symbols whose names begin with @code{__rtti_} or @code{__builtin_} will
2869 not be exported, to help with C++ DLLs. Finally, there is an
2870 extensive list of cygwin-private symbols that are not exported
2871 (obviously, this applies on when building DLLs for cygwin targets).
2872 These cygwin-excludes are: @code{_cygwin_dll_entry@@12},
2873 @code{_cygwin_crt0_common@@8}, @code{_cygwin_noncygwin_dll_entry@@12},
2874 @code{_fmode}, @code{_impure_ptr}, @code{cygwin_attach_dll},
2875 @code{cygwin_premain0}, @code{cygwin_premain1}, @code{cygwin_premain2},
2876 @code{cygwin_premain3}, and @code{environ}.
2877 [This option is specific to the i386 PE targeted port of the linker]
2879 @kindex --exclude-symbols
2880 @item --exclude-symbols @var{symbol},@var{symbol},...
2881 Specifies a list of symbols which should not be automatically
2882 exported. The symbol names may be delimited by commas or colons.
2883 [This option is specific to the i386 PE targeted port of the linker]
2885 @kindex --exclude-all-symbols
2886 @item --exclude-all-symbols
2887 Specifies no symbols should be automatically exported.
2888 [This option is specific to the i386 PE targeted port of the linker]
2890 @kindex --file-alignment
2891 @item --file-alignment
2892 Specify the file alignment. Sections in the file will always begin at
2893 file offsets which are multiples of this number. This defaults to
2895 [This option is specific to the i386 PE targeted port of the linker]
2899 @item --heap @var{reserve}
2900 @itemx --heap @var{reserve},@var{commit}
2901 Specify the number of bytes of memory to reserve (and optionally commit)
2902 to be used as heap for this program. The default is 1MB reserved, 4K
2904 [This option is specific to the i386 PE targeted port of the linker]
2907 @kindex --image-base
2908 @item --image-base @var{value}
2909 Use @var{value} as the base address of your program or dll. This is
2910 the lowest memory location that will be used when your program or dll
2911 is loaded. To reduce the need to relocate and improve performance of
2912 your dlls, each should have a unique base address and not overlap any
2913 other dlls. The default is 0x400000 for executables, and 0x10000000
2915 [This option is specific to the i386 PE targeted port of the linker]
2919 If given, the stdcall suffixes (@@@var{nn}) will be stripped from
2920 symbols before they are exported.
2921 [This option is specific to the i386 PE targeted port of the linker]
2923 @kindex --large-address-aware
2924 @item --large-address-aware
2925 If given, the appropriate bit in the ``Characteristics'' field of the COFF
2926 header is set to indicate that this executable supports virtual addresses
2927 greater than 2 gigabytes. This should be used in conjunction with the /3GB
2928 or /USERVA=@var{value} megabytes switch in the ``[operating systems]''
2929 section of the BOOT.INI. Otherwise, this bit has no effect.
2930 [This option is specific to PE targeted ports of the linker]
2932 @kindex --disable-large-address-aware
2933 @item --disable-large-address-aware
2934 Reverts the effect of a previous @samp{--large-address-aware} option.
2935 This is useful if @samp{--large-address-aware} is always set by the compiler
2936 driver (e.g. Cygwin gcc) and the executable does not support virtual
2937 addresses greater than 2 gigabytes.
2938 [This option is specific to PE targeted ports of the linker]
2940 @kindex --major-image-version
2941 @item --major-image-version @var{value}
2942 Sets the major number of the ``image version''. Defaults to 1.
2943 [This option is specific to the i386 PE targeted port of the linker]
2945 @kindex --major-os-version
2946 @item --major-os-version @var{value}
2947 Sets the major number of the ``os version''. Defaults to 4.
2948 [This option is specific to the i386 PE targeted port of the linker]
2950 @kindex --major-subsystem-version
2951 @item --major-subsystem-version @var{value}
2952 Sets the major number of the ``subsystem version''. Defaults to 4.
2953 [This option is specific to the i386 PE targeted port of the linker]
2955 @kindex --minor-image-version
2956 @item --minor-image-version @var{value}
2957 Sets the minor number of the ``image version''. Defaults to 0.
2958 [This option is specific to the i386 PE targeted port of the linker]
2960 @kindex --minor-os-version
2961 @item --minor-os-version @var{value}
2962 Sets the minor number of the ``os version''. Defaults to 0.
2963 [This option is specific to the i386 PE targeted port of the linker]
2965 @kindex --minor-subsystem-version
2966 @item --minor-subsystem-version @var{value}
2967 Sets the minor number of the ``subsystem version''. Defaults to 0.
2968 [This option is specific to the i386 PE targeted port of the linker]
2970 @cindex DEF files, creating
2971 @cindex DLLs, creating
2972 @kindex --output-def
2973 @item --output-def @var{file}
2974 The linker will create the file @var{file} which will contain a DEF
2975 file corresponding to the DLL the linker is generating. This DEF file
2976 (which should be called @code{*.def}) may be used to create an import
2977 library with @code{dlltool} or may be used as a reference to
2978 automatically or implicitly exported symbols.
2979 [This option is specific to the i386 PE targeted port of the linker]
2981 @cindex DLLs, creating
2982 @kindex --enable-auto-image-base
2983 @item --enable-auto-image-base
2984 @itemx --enable-auto-image-base=@var{value}
2985 Automatically choose the image base for DLLs, optionally starting with base
2986 @var{value}, unless one is specified using the @code{--image-base} argument.
2987 By using a hash generated from the dllname to create unique image bases
2988 for each DLL, in-memory collisions and relocations which can delay program
2989 execution are avoided.
2990 [This option is specific to the i386 PE targeted port of the linker]
2992 @kindex --disable-auto-image-base
2993 @item --disable-auto-image-base
2994 Do not automatically generate a unique image base. If there is no
2995 user-specified image base (@code{--image-base}) then use the platform
2997 [This option is specific to the i386 PE targeted port of the linker]
2999 @cindex DLLs, linking to
3000 @kindex --dll-search-prefix
3001 @item --dll-search-prefix @var{string}
3002 When linking dynamically to a dll without an import library,
3003 search for @code{<string><basename>.dll} in preference to
3004 @code{lib<basename>.dll}. This behaviour allows easy distinction
3005 between DLLs built for the various "subplatforms": native, cygwin,
3006 uwin, pw, etc. For instance, cygwin DLLs typically use
3007 @code{--dll-search-prefix=cyg}.
3008 [This option is specific to the i386 PE targeted port of the linker]
3010 @kindex --enable-auto-import
3011 @item --enable-auto-import
3012 Do sophisticated linking of @code{_symbol} to @code{__imp__symbol} for
3013 DATA imports from DLLs, thus making it possible to bypass the dllimport
3014 mechanism on the user side and to reference unmangled symbol names.
3015 [This option is specific to the i386 PE targeted port of the linker]
3017 The following remarks pertain to the original implementation of the
3018 feature and are obsolete nowadays for Cygwin and MinGW targets.
3020 Note: Use of the 'auto-import' extension will cause the text section
3021 of the image file to be made writable. This does not conform to the
3022 PE-COFF format specification published by Microsoft.
3024 Note - use of the 'auto-import' extension will also cause read only
3025 data which would normally be placed into the .rdata section to be
3026 placed into the .data section instead. This is in order to work
3027 around a problem with consts that is described here:
3028 http://www.cygwin.com/ml/cygwin/2004-09/msg01101.html
3030 Using 'auto-import' generally will 'just work' -- but sometimes you may
3033 "variable '<var>' can't be auto-imported. Please read the
3034 documentation for ld's @code{--enable-auto-import} for details."
3036 This message occurs when some (sub)expression accesses an address
3037 ultimately given by the sum of two constants (Win32 import tables only
3038 allow one). Instances where this may occur include accesses to member
3039 fields of struct variables imported from a DLL, as well as using a
3040 constant index into an array variable imported from a DLL. Any
3041 multiword variable (arrays, structs, long long, etc) may trigger
3042 this error condition. However, regardless of the exact data type
3043 of the offending exported variable, ld will always detect it, issue
3044 the warning, and exit.
3046 There are several ways to address this difficulty, regardless of the
3047 data type of the exported variable:
3049 One way is to use --enable-runtime-pseudo-reloc switch. This leaves the task
3050 of adjusting references in your client code for runtime environment, so
3051 this method works only when runtime environment supports this feature.
3053 A second solution is to force one of the 'constants' to be a variable --
3054 that is, unknown and un-optimizable at compile time. For arrays,
3055 there are two possibilities: a) make the indexee (the array's address)
3056 a variable, or b) make the 'constant' index a variable. Thus:
3059 extern type extern_array[];
3061 @{ volatile type *t=extern_array; t[1] @}
3067 extern type extern_array[];
3069 @{ volatile int t=1; extern_array[t] @}
3072 For structs (and most other multiword data types) the only option
3073 is to make the struct itself (or the long long, or the ...) variable:
3076 extern struct s extern_struct;
3077 extern_struct.field -->
3078 @{ volatile struct s *t=&extern_struct; t->field @}
3084 extern long long extern_ll;
3086 @{ volatile long long * local_ll=&extern_ll; *local_ll @}
3089 A third method of dealing with this difficulty is to abandon
3090 'auto-import' for the offending symbol and mark it with
3091 @code{__declspec(dllimport)}. However, in practice that
3092 requires using compile-time #defines to indicate whether you are
3093 building a DLL, building client code that will link to the DLL, or
3094 merely building/linking to a static library. In making the choice
3095 between the various methods of resolving the 'direct address with
3096 constant offset' problem, you should consider typical real-world usage:
3104 void main(int argc, char **argv)@{
3105 printf("%d\n",arr[1]);
3115 void main(int argc, char **argv)@{
3116 /* This workaround is for win32 and cygwin; do not "optimize" */
3117 volatile int *parr = arr;
3118 printf("%d\n",parr[1]);
3125 /* Note: auto-export is assumed (no __declspec(dllexport)) */
3126 #if (defined(_WIN32) || defined(__CYGWIN__)) && \
3127 !(defined(FOO_BUILD_DLL) || defined(FOO_STATIC))
3128 #define FOO_IMPORT __declspec(dllimport)
3132 extern FOO_IMPORT int arr[];
3135 void main(int argc, char **argv)@{
3136 printf("%d\n",arr[1]);
3140 A fourth way to avoid this problem is to re-code your
3141 library to use a functional interface rather than a data interface
3142 for the offending variables (e.g. set_foo() and get_foo() accessor
3145 @kindex --disable-auto-import
3146 @item --disable-auto-import
3147 Do not attempt to do sophisticated linking of @code{_symbol} to
3148 @code{__imp__symbol} for DATA imports from DLLs.
3149 [This option is specific to the i386 PE targeted port of the linker]
3151 @kindex --enable-runtime-pseudo-reloc
3152 @item --enable-runtime-pseudo-reloc
3153 If your code contains expressions described in --enable-auto-import section,
3154 that is, DATA imports from DLL with non-zero offset, this switch will create
3155 a vector of 'runtime pseudo relocations' which can be used by runtime
3156 environment to adjust references to such data in your client code.
3157 [This option is specific to the i386 PE targeted port of the linker]
3159 @kindex --disable-runtime-pseudo-reloc
3160 @item --disable-runtime-pseudo-reloc
3161 Do not create pseudo relocations for non-zero offset DATA imports from DLLs.
3162 [This option is specific to the i386 PE targeted port of the linker]
3164 @kindex --enable-extra-pe-debug
3165 @item --enable-extra-pe-debug
3166 Show additional debug info related to auto-import symbol thunking.
3167 [This option is specific to the i386 PE targeted port of the linker]
3169 @kindex --section-alignment
3170 @item --section-alignment
3171 Sets the section alignment. Sections in memory will always begin at
3172 addresses which are a multiple of this number. Defaults to 0x1000.
3173 [This option is specific to the i386 PE targeted port of the linker]
3177 @item --stack @var{reserve}
3178 @itemx --stack @var{reserve},@var{commit}
3179 Specify the number of bytes of memory to reserve (and optionally commit)
3180 to be used as stack for this program. The default is 2MB reserved, 4K
3182 [This option is specific to the i386 PE targeted port of the linker]
3185 @item --subsystem @var{which}
3186 @itemx --subsystem @var{which}:@var{major}
3187 @itemx --subsystem @var{which}:@var{major}.@var{minor}
3188 Specifies the subsystem under which your program will execute. The
3189 legal values for @var{which} are @code{native}, @code{windows},
3190 @code{console}, @code{posix}, and @code{xbox}. You may optionally set
3191 the subsystem version also. Numeric values are also accepted for
3193 [This option is specific to the i386 PE targeted port of the linker]
3195 The following options set flags in the @code{DllCharacteristics} field
3196 of the PE file header:
3197 [These options are specific to PE targeted ports of the linker]
3199 @kindex --high-entropy-va
3200 @item --high-entropy-va
3201 @itemx --disable-high-entropy-va
3202 Image is compatible with 64-bit address space layout randomization
3203 (ASLR). This option is enabled by default for 64-bit PE images.
3205 This option also implies @option{--dynamicbase} and
3206 @option{--enable-reloc-section}.
3208 @kindex --dynamicbase
3210 @itemx --disable-dynamicbase
3211 The image base address may be relocated using address space layout
3212 randomization (ASLR). This feature was introduced with MS Windows
3213 Vista for i386 PE targets. This option is enabled by default but
3214 can be disabled via the @option{--disable-dynamicbase} option.
3215 This option also implies @option{--enable-reloc-section}.
3217 @kindex --forceinteg
3219 @itemx --disable-forceinteg
3220 Code integrity checks are enforced. This option is disabled by
3225 @item --disable-nxcompat
3226 The image is compatible with the Data Execution Prevention.
3227 This feature was introduced with MS Windows XP SP2 for i386 PE
3228 targets. The option is enabled by default.
3230 @kindex --no-isolation
3231 @item --no-isolation
3232 @itemx --disable-no-isolation
3233 Although the image understands isolation, do not isolate the image.
3234 This option is disabled by default.
3238 @itemx --disable-no-seh
3239 The image does not use SEH. No SE handler may be called from
3240 this image. This option is disabled by default.
3244 @itemx --disable-no-bind
3245 Do not bind this image. This option is disabled by default.
3249 @itemx --disable-wdmdriver
3250 The driver uses the MS Windows Driver Model. This option is disabled
3255 @itemx --disable-tsaware
3256 The image is Terminal Server aware. This option is disabled by
3259 @kindex --insert-timestamp
3260 @item --insert-timestamp
3261 @itemx --no-insert-timestamp
3262 Insert a real timestamp into the image. This is the default behaviour
3263 as it matches legacy code and it means that the image will work with
3264 other, proprietary tools. The problem with this default is that it
3265 will result in slightly different images being produced each time the
3266 same sources are linked. The option @option{--no-insert-timestamp}
3267 can be used to insert a zero value for the timestamp, this ensuring
3268 that binaries produced from identical sources will compare
3271 @kindex --enable-reloc-section
3272 @item --enable-reloc-section
3273 @itemx --disable-reloc-section
3274 Create the base relocation table, which is necessary if the image
3275 is loaded at a different image base than specified in the PE header.
3276 This option is enabled by default.
3282 @subsection Options specific to C6X uClinux targets
3284 @c man begin OPTIONS
3286 The C6X uClinux target uses a binary format called DSBT to support shared
3287 libraries. Each shared library in the system needs to have a unique index;
3288 all executables use an index of 0.
3293 @item --dsbt-size @var{size}
3294 This option sets the number of entries in the DSBT of the current executable
3295 or shared library to @var{size}. The default is to create a table with 64
3298 @kindex --dsbt-index
3299 @item --dsbt-index @var{index}
3300 This option sets the DSBT index of the current executable or shared library
3301 to @var{index}. The default is 0, which is appropriate for generating
3302 executables. If a shared library is generated with a DSBT index of 0, the
3303 @code{R_C6000_DSBT_INDEX} relocs are copied into the output file.
3305 @kindex --no-merge-exidx-entries
3306 The @samp{--no-merge-exidx-entries} switch disables the merging of adjacent
3307 exidx entries in frame unwind info.
3315 @subsection Options specific to C-SKY targets
3317 @c man begin OPTIONS
3321 @kindex --branch-stub on C-SKY
3323 This option enables linker branch relaxation by inserting branch stub
3324 sections when needed to extend the range of branches. This option is
3325 usually not required since C-SKY supports branch and call instructions that
3326 can access the full memory range and branch relaxation is normally handled by
3327 the compiler or assembler.
3329 @kindex --stub-group-size on C-SKY
3330 @item --stub-group-size=@var{N}
3331 This option allows finer control of linker branch stub creation.
3332 It sets the maximum size of a group of input sections that can
3333 be handled by one stub section. A negative value of @var{N} locates
3334 stub sections after their branches, while a positive value allows stub
3335 sections to appear either before or after the branches. Values of
3336 @samp{1} or @samp{-1} indicate that the
3337 linker should choose suitable defaults.
3345 @subsection Options specific to Motorola 68HC11 and 68HC12 targets
3347 @c man begin OPTIONS
3349 The 68HC11 and 68HC12 linkers support specific options to control the
3350 memory bank switching mapping and trampoline code generation.
3354 @kindex --no-trampoline
3355 @item --no-trampoline
3356 This option disables the generation of trampoline. By default a trampoline
3357 is generated for each far function which is called using a @code{jsr}
3358 instruction (this happens when a pointer to a far function is taken).
3360 @kindex --bank-window
3361 @item --bank-window @var{name}
3362 This option indicates to the linker the name of the memory region in
3363 the @samp{MEMORY} specification that describes the memory bank window.
3364 The definition of such region is then used by the linker to compute
3365 paging and addresses within the memory window.
3373 @subsection Options specific to Motorola 68K target
3375 @c man begin OPTIONS
3377 The following options are supported to control handling of GOT generation
3378 when linking for 68K targets.
3383 @item --got=@var{type}
3384 This option tells the linker which GOT generation scheme to use.
3385 @var{type} should be one of @samp{single}, @samp{negative},
3386 @samp{multigot} or @samp{target}. For more information refer to the
3387 Info entry for @file{ld}.
3395 @subsection Options specific to MIPS targets
3397 @c man begin OPTIONS
3399 The following options are supported to control microMIPS instruction
3400 generation and branch relocation checks for ISA mode transitions when
3401 linking for MIPS targets.
3409 These options control the choice of microMIPS instructions used in code
3410 generated by the linker, such as that in the PLT or lazy binding stubs,
3411 or in relaxation. If @samp{--insn32} is used, then the linker only uses
3412 32-bit instruction encodings. By default or if @samp{--no-insn32} is
3413 used, all instruction encodings are used, including 16-bit ones where
3416 @kindex --ignore-branch-isa
3417 @item --ignore-branch-isa
3418 @kindex --no-ignore-branch-isa
3419 @itemx --no-ignore-branch-isa
3420 These options control branch relocation checks for invalid ISA mode
3421 transitions. If @samp{--ignore-branch-isa} is used, then the linker
3422 accepts any branch relocations and any ISA mode transition required
3423 is lost in relocation calculation, except for some cases of @code{BAL}
3424 instructions which meet relaxation conditions and are converted to
3425 equivalent @code{JALX} instructions as the associated relocation is
3426 calculated. By default or if @samp{--no-ignore-branch-isa} is used
3427 a check is made causing the loss of an ISA mode transition to produce
3430 @kindex --compact-branches
3431 @item --compact-branches
3432 @kindex --no-compact-branches
3433 @itemx --no-compact-branches
3434 These options control the generation of compact instructions by the linker
3435 in the PLT entries for MIPS R6.
3444 @subsection Options specific to PDP11 targets
3446 @c man begin OPTIONS
3448 For the pdp11-aout target, three variants of the output format can be
3449 produced as selected by the following options. The default variant
3450 for pdp11-aout is the @samp{--omagic} option, whereas for other
3451 targets @samp{--nmagic} is the default. The @samp{--imagic} option is
3452 defined only for the pdp11-aout target, while the others are described
3453 here as they apply to the pdp11-aout target.
3462 Mark the output as @code{OMAGIC} (0407) in the @file{a.out} header to
3463 indicate that the text segment is not to be write-protected and
3464 shared. Since the text and data sections are both readable and
3465 writable, the data section is allocated immediately contiguous after
3466 the text segment. This is the oldest format for PDP11 executable
3467 programs and is the default for @command{ld} on PDP11 Unix systems
3468 from the beginning through 2.11BSD.
3475 Mark the output as @code{NMAGIC} (0410) in the @file{a.out} header to
3476 indicate that when the output file is executed, the text portion will
3477 be read-only and shareable among all processes executing the same
3478 file. This involves moving the data areas up to the first possible 8K
3479 byte page boundary following the end of the text. This option creates
3480 a @emph{pure executable} format.
3487 Mark the output as @code{IMAGIC} (0411) in the @file{a.out} header to
3488 indicate that when the output file is executed, the program text and
3489 data areas will be loaded into separate address spaces using the split
3490 instruction and data space feature of the memory management unit in
3491 larger models of the PDP11. This doubles the address space available
3492 to the program. The text segment is again pure, write-protected, and
3493 shareable. The only difference in the output format between this
3494 option and the others, besides the magic number, is that both the text
3495 and data sections start at location 0. The @samp{-z} option selected
3496 this format in 2.11BSD. This option creates a @emph{separate
3502 Equivalent to @samp{--nmagic} for pdp11-aout.
3511 @section Environment Variables
3513 @c man begin ENVIRONMENT
3515 You can change the behaviour of @command{ld} with the environment variables
3516 @ifclear SingleFormat
3519 @code{LDEMULATION} and @code{COLLECT_NO_DEMANGLE}.
3521 @ifclear SingleFormat
3523 @cindex default input format
3524 @code{GNUTARGET} determines the input-file object format if you don't
3525 use @samp{-b} (or its synonym @samp{--format}). Its value should be one
3526 of the BFD names for an input format (@pxref{BFD}). If there is no
3527 @code{GNUTARGET} in the environment, @command{ld} uses the natural format
3528 of the target. If @code{GNUTARGET} is set to @code{default} then BFD
3529 attempts to discover the input format by examining binary input files;
3530 this method often succeeds, but there are potential ambiguities, since
3531 there is no method of ensuring that the magic number used to specify
3532 object-file formats is unique. However, the configuration procedure for
3533 BFD on each system places the conventional format for that system first
3534 in the search-list, so ambiguities are resolved in favor of convention.
3538 @cindex default emulation
3539 @cindex emulation, default
3540 @code{LDEMULATION} determines the default emulation if you don't use the
3541 @samp{-m} option. The emulation can affect various aspects of linker
3542 behaviour, particularly the default linker script. You can list the
3543 available emulations with the @samp{--verbose} or @samp{-V} options. If
3544 the @samp{-m} option is not used, and the @code{LDEMULATION} environment
3545 variable is not defined, the default emulation depends upon how the
3546 linker was configured.
3548 @kindex COLLECT_NO_DEMANGLE
3549 @cindex demangling, default
3550 Normally, the linker will default to demangling symbols. However, if
3551 @code{COLLECT_NO_DEMANGLE} is set in the environment, then it will
3552 default to not demangling symbols. This environment variable is used in
3553 a similar fashion by the @code{gcc} linker wrapper program. The default
3554 may be overridden by the @samp{--demangle} and @samp{--no-demangle}
3561 @chapter Linker Scripts
3564 @cindex linker scripts
3565 @cindex command files
3566 Every link is controlled by a @dfn{linker script}. This script is
3567 written in the linker command language.
3569 The main purpose of the linker script is to describe how the sections in
3570 the input files should be mapped into the output file, and to control
3571 the memory layout of the output file. Most linker scripts do nothing
3572 more than this. However, when necessary, the linker script can also
3573 direct the linker to perform many other operations, using the commands
3576 The linker always uses a linker script. If you do not supply one
3577 yourself, the linker will use a default script that is compiled into the
3578 linker executable. You can use the @samp{--verbose} command-line option
3579 to display the default linker script. Certain command-line options,
3580 such as @samp{-r} or @samp{-N}, will affect the default linker script.
3582 You may supply your own linker script by using the @samp{-T} command
3583 line option. When you do this, your linker script will replace the
3584 default linker script.
3586 You may also use linker scripts implicitly by naming them as input files
3587 to the linker, as though they were files to be linked. @xref{Implicit
3591 * Basic Script Concepts:: Basic Linker Script Concepts
3592 * Script Format:: Linker Script Format
3593 * Simple Example:: Simple Linker Script Example
3594 * Simple Commands:: Simple Linker Script Commands
3595 * Assignments:: Assigning Values to Symbols
3596 * SECTIONS:: SECTIONS Command
3597 * MEMORY:: MEMORY Command
3598 * PHDRS:: PHDRS Command
3599 * VERSION:: VERSION Command
3600 * Expressions:: Expressions in Linker Scripts
3601 * Implicit Linker Scripts:: Implicit Linker Scripts
3604 @node Basic Script Concepts
3605 @section Basic Linker Script Concepts
3606 @cindex linker script concepts
3607 We need to define some basic concepts and vocabulary in order to
3608 describe the linker script language.
3610 The linker combines input files into a single output file. The output
3611 file and each input file are in a special data format known as an
3612 @dfn{object file format}. Each file is called an @dfn{object file}.
3613 The output file is often called an @dfn{executable}, but for our
3614 purposes we will also call it an object file. Each object file has,
3615 among other things, a list of @dfn{sections}. We sometimes refer to a
3616 section in an input file as an @dfn{input section}; similarly, a section
3617 in the output file is an @dfn{output section}.
3619 Each section in an object file has a name and a size. Most sections
3620 also have an associated block of data, known as the @dfn{section
3621 contents}. A section may be marked as @dfn{loadable}, which means that
3622 the contents should be loaded into memory when the output file is run.
3623 A section with no contents may be @dfn{allocatable}, which means that an
3624 area in memory should be set aside, but nothing in particular should be
3625 loaded there (in some cases this memory must be zeroed out). A section
3626 which is neither loadable nor allocatable typically contains some sort
3627 of debugging information.
3629 Every loadable or allocatable output section has two addresses. The
3630 first is the @dfn{VMA}, or virtual memory address. This is the address
3631 the section will have when the output file is run. The second is the
3632 @dfn{LMA}, or load memory address. This is the address at which the
3633 section will be loaded. In most cases the two addresses will be the
3634 same. An example of when they might be different is when a data section
3635 is loaded into ROM, and then copied into RAM when the program starts up
3636 (this technique is often used to initialize global variables in a ROM
3637 based system). In this case the ROM address would be the LMA, and the
3638 RAM address would be the VMA.
3640 You can see the sections in an object file by using the @code{objdump}
3641 program with the @samp{-h} option.
3643 Every object file also has a list of @dfn{symbols}, known as the
3644 @dfn{symbol table}. A symbol may be defined or undefined. Each symbol
3645 has a name, and each defined symbol has an address, among other
3646 information. If you compile a C or C++ program into an object file, you
3647 will get a defined symbol for every defined function and global or
3648 static variable. Every undefined function or global variable which is
3649 referenced in the input file will become an undefined symbol.
3651 You can see the symbols in an object file by using the @code{nm}
3652 program, or by using the @code{objdump} program with the @samp{-t}
3656 @section Linker Script Format
3657 @cindex linker script format
3658 Linker scripts are text files.
3660 You write a linker script as a series of commands. Each command is
3661 either a keyword, possibly followed by arguments, or an assignment to a
3662 symbol. You may separate commands using semicolons. Whitespace is
3665 Strings such as file or format names can normally be entered directly.
3666 If the file name contains a character such as a comma which would
3667 otherwise serve to separate file names, you may put the file name in
3668 double quotes. There is no way to use a double quote character in a
3671 You may include comments in linker scripts just as in C, delimited by
3672 @samp{/*} and @samp{*/}. As in C, comments are syntactically equivalent
3675 @node Simple Example
3676 @section Simple Linker Script Example
3677 @cindex linker script example
3678 @cindex example of linker script
3679 Many linker scripts are fairly simple.
3681 The simplest possible linker script has just one command:
3682 @samp{SECTIONS}. You use the @samp{SECTIONS} command to describe the
3683 memory layout of the output file.
3685 The @samp{SECTIONS} command is a powerful command. Here we will
3686 describe a simple use of it. Let's assume your program consists only of
3687 code, initialized data, and uninitialized data. These will be in the
3688 @samp{.text}, @samp{.data}, and @samp{.bss} sections, respectively.
3689 Let's assume further that these are the only sections which appear in
3692 For this example, let's say that the code should be loaded at address
3693 0x10000, and that the data should start at address 0x8000000. Here is a
3694 linker script which will do that:
3699 .text : @{ *(.text) @}
3701 .data : @{ *(.data) @}
3702 .bss : @{ *(.bss) @}
3706 You write the @samp{SECTIONS} command as the keyword @samp{SECTIONS},
3707 followed by a series of symbol assignments and output section
3708 descriptions enclosed in curly braces.
3710 The first line inside the @samp{SECTIONS} command of the above example
3711 sets the value of the special symbol @samp{.}, which is the location
3712 counter. If you do not specify the address of an output section in some
3713 other way (other ways are described later), the address is set from the
3714 current value of the location counter. The location counter is then
3715 incremented by the size of the output section. At the start of the
3716 @samp{SECTIONS} command, the location counter has the value @samp{0}.
3718 The second line defines an output section, @samp{.text}. The colon is
3719 required syntax which may be ignored for now. Within the curly braces
3720 after the output section name, you list the names of the input sections
3721 which should be placed into this output section. The @samp{*} is a
3722 wildcard which matches any file name. The expression @samp{*(.text)}
3723 means all @samp{.text} input sections in all input files.
3725 Since the location counter is @samp{0x10000} when the output section
3726 @samp{.text} is defined, the linker will set the address of the
3727 @samp{.text} section in the output file to be @samp{0x10000}.
3729 The remaining lines define the @samp{.data} and @samp{.bss} sections in
3730 the output file. The linker will place the @samp{.data} output section
3731 at address @samp{0x8000000}. After the linker places the @samp{.data}
3732 output section, the value of the location counter will be
3733 @samp{0x8000000} plus the size of the @samp{.data} output section. The
3734 effect is that the linker will place the @samp{.bss} output section
3735 immediately after the @samp{.data} output section in memory.
3737 The linker will ensure that each output section has the required
3738 alignment, by increasing the location counter if necessary. In this
3739 example, the specified addresses for the @samp{.text} and @samp{.data}
3740 sections will probably satisfy any alignment constraints, but the linker
3741 may have to create a small gap between the @samp{.data} and @samp{.bss}
3744 That's it! That's a simple and complete linker script.
3746 @node Simple Commands
3747 @section Simple Linker Script Commands
3748 @cindex linker script simple commands
3749 In this section we describe the simple linker script commands.
3752 * Entry Point:: Setting the entry point
3753 * File Commands:: Commands dealing with files
3754 @ifclear SingleFormat
3755 * Format Commands:: Commands dealing with object file formats
3758 * REGION_ALIAS:: Assign alias names to memory regions
3759 * Miscellaneous Commands:: Other linker script commands
3763 @subsection Setting the Entry Point
3764 @kindex ENTRY(@var{symbol})
3765 @cindex start of execution
3766 @cindex first instruction
3768 The first instruction to execute in a program is called the @dfn{entry
3769 point}. You can use the @code{ENTRY} linker script command to set the
3770 entry point. The argument is a symbol name:
3775 There are several ways to set the entry point. The linker will set the
3776 entry point by trying each of the following methods in order, and
3777 stopping when one of them succeeds:
3780 the @samp{-e} @var{entry} command-line option;
3782 the @code{ENTRY(@var{symbol})} command in a linker script;
3784 the value of a target-specific symbol, if it is defined; For many
3785 targets this is @code{start}, but PE- and BeOS-based systems for example
3786 check a list of possible entry symbols, matching the first one found.
3788 the address of the first byte of the @samp{.text} section, if present;
3790 The address @code{0}.
3794 @subsection Commands Dealing with Files
3795 @cindex linker script file commands
3796 Several linker script commands deal with files.
3799 @item INCLUDE @var{filename}
3800 @kindex INCLUDE @var{filename}
3801 @cindex including a linker script
3802 Include the linker script @var{filename} at this point. The file will
3803 be searched for in the current directory, and in any directory specified
3804 with the @option{-L} option. You can nest calls to @code{INCLUDE} up to
3807 You can place @code{INCLUDE} directives at the top level, in @code{MEMORY} or
3808 @code{SECTIONS} commands, or in output section descriptions.
3810 @item INPUT(@var{file}, @var{file}, @dots{})
3811 @itemx INPUT(@var{file} @var{file} @dots{})
3812 @kindex INPUT(@var{files})
3813 @cindex input files in linker scripts
3814 @cindex input object files in linker scripts
3815 @cindex linker script input object files
3816 The @code{INPUT} command directs the linker to include the named files
3817 in the link, as though they were named on the command line.
3819 For example, if you always want to include @file{subr.o} any time you do
3820 a link, but you can't be bothered to put it on every link command line,
3821 then you can put @samp{INPUT (subr.o)} in your linker script.
3823 In fact, if you like, you can list all of your input files in the linker
3824 script, and then invoke the linker with nothing but a @samp{-T} option.
3826 In case a @dfn{sysroot prefix} is configured, and the filename starts
3827 with the @samp{/} character, and the script being processed was
3828 located inside the @dfn{sysroot prefix}, the filename will be looked
3829 for in the @dfn{sysroot prefix}. The @dfn{sysroot prefix} can also be forced by specifying
3830 @code{=} as the first character in the filename path, or prefixing the
3831 filename path with @code{$SYSROOT}. See also the description of
3832 @samp{-L} in @ref{Options,,Command-line Options}.
3834 If a @dfn{sysroot prefix} is not used then the linker will try to open
3835 the file in the directory containing the linker script. If it is not
3836 found the linker will then search the current directory. If it is still
3837 not found the linker will search through the archive library search
3840 If you use @samp{INPUT (-l@var{file})}, @command{ld} will transform the
3841 name to @code{lib@var{file}.a}, as with the command-line argument
3844 When you use the @code{INPUT} command in an implicit linker script, the
3845 files will be included in the link at the point at which the linker
3846 script file is included. This can affect archive searching.
3848 @item GROUP(@var{file}, @var{file}, @dots{})
3849 @itemx GROUP(@var{file} @var{file} @dots{})
3850 @kindex GROUP(@var{files})
3851 @cindex grouping input files
3852 The @code{GROUP} command is like @code{INPUT}, except that the named
3853 files should all be archives, and they are searched repeatedly until no
3854 new undefined references are created. See the description of @samp{-(}
3855 in @ref{Options,,Command-line Options}.
3857 @item AS_NEEDED(@var{file}, @var{file}, @dots{})
3858 @itemx AS_NEEDED(@var{file} @var{file} @dots{})
3859 @kindex AS_NEEDED(@var{files})
3860 This construct can appear only inside of the @code{INPUT} or @code{GROUP}
3861 commands, among other filenames. The files listed will be handled
3862 as if they appear directly in the @code{INPUT} or @code{GROUP} commands,
3863 with the exception of ELF shared libraries, that will be added only
3864 when they are actually needed. This construct essentially enables
3865 @option{--as-needed} option for all the files listed inside of it
3866 and restores previous @option{--as-needed} resp. @option{--no-as-needed}
3869 @item OUTPUT(@var{filename})
3870 @kindex OUTPUT(@var{filename})
3871 @cindex output file name in linker script
3872 The @code{OUTPUT} command names the output file. Using
3873 @code{OUTPUT(@var{filename})} in the linker script is exactly like using
3874 @samp{-o @var{filename}} on the command line (@pxref{Options,,Command
3875 Line Options}). If both are used, the command-line option takes
3878 You can use the @code{OUTPUT} command to define a default name for the
3879 output file other than the usual default of @file{a.out}.
3881 @item SEARCH_DIR(@var{path})
3882 @kindex SEARCH_DIR(@var{path})
3883 @cindex library search path in linker script
3884 @cindex archive search path in linker script
3885 @cindex search path in linker script
3886 The @code{SEARCH_DIR} command adds @var{path} to the list of paths where
3887 @command{ld} looks for archive libraries. Using
3888 @code{SEARCH_DIR(@var{path})} is exactly like using @samp{-L @var{path}}
3889 on the command line (@pxref{Options,,Command-line Options}). If both
3890 are used, then the linker will search both paths. Paths specified using
3891 the command-line option are searched first.
3893 @item STARTUP(@var{filename})
3894 @kindex STARTUP(@var{filename})
3895 @cindex first input file
3896 The @code{STARTUP} command is just like the @code{INPUT} command, except
3897 that @var{filename} will become the first input file to be linked, as
3898 though it were specified first on the command line. This may be useful
3899 when using a system in which the entry point is always the start of the
3903 @ifclear SingleFormat
3904 @node Format Commands
3905 @subsection Commands Dealing with Object File Formats
3906 A couple of linker script commands deal with object file formats.
3909 @item OUTPUT_FORMAT(@var{bfdname})
3910 @itemx OUTPUT_FORMAT(@var{default}, @var{big}, @var{little})
3911 @kindex OUTPUT_FORMAT(@var{bfdname})
3912 @cindex output file format in linker script
3913 The @code{OUTPUT_FORMAT} command names the BFD format to use for the
3914 output file (@pxref{BFD}). Using @code{OUTPUT_FORMAT(@var{bfdname})} is
3915 exactly like using @samp{--oformat @var{bfdname}} on the command line
3916 (@pxref{Options,,Command-line Options}). If both are used, the command
3917 line option takes precedence.
3919 You can use @code{OUTPUT_FORMAT} with three arguments to use different
3920 formats based on the @samp{-EB} and @samp{-EL} command-line options.
3921 This permits the linker script to set the output format based on the
3924 If neither @samp{-EB} nor @samp{-EL} are used, then the output format
3925 will be the first argument, @var{default}. If @samp{-EB} is used, the
3926 output format will be the second argument, @var{big}. If @samp{-EL} is
3927 used, the output format will be the third argument, @var{little}.
3929 For example, the default linker script for the MIPS ELF target uses this
3932 OUTPUT_FORMAT(elf32-bigmips, elf32-bigmips, elf32-littlemips)
3934 This says that the default format for the output file is
3935 @samp{elf32-bigmips}, but if the user uses the @samp{-EL} command-line
3936 option, the output file will be created in the @samp{elf32-littlemips}
3939 @item TARGET(@var{bfdname})
3940 @kindex TARGET(@var{bfdname})
3941 @cindex input file format in linker script
3942 The @code{TARGET} command names the BFD format to use when reading input
3943 files. It affects subsequent @code{INPUT} and @code{GROUP} commands.
3944 This command is like using @samp{-b @var{bfdname}} on the command line
3945 (@pxref{Options,,Command-line Options}). If the @code{TARGET} command
3946 is used but @code{OUTPUT_FORMAT} is not, then the last @code{TARGET}
3947 command is also used to set the format for the output file. @xref{BFD}.
3952 @subsection Assign alias names to memory regions
3953 @kindex REGION_ALIAS(@var{alias}, @var{region})
3954 @cindex region alias
3955 @cindex region names
3957 Alias names can be added to existing memory regions created with the
3958 @ref{MEMORY} command. Each name corresponds to at most one memory region.
3961 REGION_ALIAS(@var{alias}, @var{region})
3964 The @code{REGION_ALIAS} function creates an alias name @var{alias} for the
3965 memory region @var{region}. This allows a flexible mapping of output sections
3966 to memory regions. An example follows.
3968 Suppose we have an application for embedded systems which come with various
3969 memory storage devices. All have a general purpose, volatile memory @code{RAM}
3970 that allows code execution or data storage. Some may have a read-only,
3971 non-volatile memory @code{ROM} that allows code execution and read-only data
3972 access. The last variant is a read-only, non-volatile memory @code{ROM2} with
3973 read-only data access and no code execution capability. We have four output
3978 @code{.text} program code;
3980 @code{.rodata} read-only data;
3982 @code{.data} read-write initialized data;
3984 @code{.bss} read-write zero initialized data.
3987 The goal is to provide a linker command file that contains a system independent
3988 part defining the output sections and a system dependent part mapping the
3989 output sections to the memory regions available on the system. Our embedded
3990 systems come with three different memory setups @code{A}, @code{B} and
3992 @multitable @columnfractions .25 .25 .25 .25
3993 @item Section @tab Variant A @tab Variant B @tab Variant C
3994 @item .text @tab RAM @tab ROM @tab ROM
3995 @item .rodata @tab RAM @tab ROM @tab ROM2
3996 @item .data @tab RAM @tab RAM/ROM @tab RAM/ROM2
3997 @item .bss @tab RAM @tab RAM @tab RAM
3999 The notation @code{RAM/ROM} or @code{RAM/ROM2} means that this section is
4000 loaded into region @code{ROM} or @code{ROM2} respectively. Please note that
4001 the load address of the @code{.data} section starts in all three variants at
4002 the end of the @code{.rodata} section.
4004 The base linker script that deals with the output sections follows. It
4005 includes the system dependent @code{linkcmds.memory} file that describes the
4008 INCLUDE linkcmds.memory
4021 .data : AT (rodata_end)
4026 data_size = SIZEOF(.data);
4027 data_load_start = LOADADDR(.data);
4035 Now we need three different @code{linkcmds.memory} files to define memory
4036 regions and alias names. The content of @code{linkcmds.memory} for the three
4037 variants @code{A}, @code{B} and @code{C}:
4040 Here everything goes into the @code{RAM}.
4044 RAM : ORIGIN = 0, LENGTH = 4M
4047 REGION_ALIAS("REGION_TEXT", RAM);
4048 REGION_ALIAS("REGION_RODATA", RAM);
4049 REGION_ALIAS("REGION_DATA", RAM);
4050 REGION_ALIAS("REGION_BSS", RAM);
4053 Program code and read-only data go into the @code{ROM}. Read-write data goes
4054 into the @code{RAM}. An image of the initialized data is loaded into the
4055 @code{ROM} and will be copied during system start into the @code{RAM}.
4059 ROM : ORIGIN = 0, LENGTH = 3M
4060 RAM : ORIGIN = 0x10000000, LENGTH = 1M
4063 REGION_ALIAS("REGION_TEXT", ROM);
4064 REGION_ALIAS("REGION_RODATA", ROM);
4065 REGION_ALIAS("REGION_DATA", RAM);
4066 REGION_ALIAS("REGION_BSS", RAM);
4069 Program code goes into the @code{ROM}. Read-only data goes into the
4070 @code{ROM2}. Read-write data goes into the @code{RAM}. An image of the
4071 initialized data is loaded into the @code{ROM2} and will be copied during
4072 system start into the @code{RAM}.
4076 ROM : ORIGIN = 0, LENGTH = 2M
4077 ROM2 : ORIGIN = 0x10000000, LENGTH = 1M
4078 RAM : ORIGIN = 0x20000000, LENGTH = 1M
4081 REGION_ALIAS("REGION_TEXT", ROM);
4082 REGION_ALIAS("REGION_RODATA", ROM2);
4083 REGION_ALIAS("REGION_DATA", RAM);
4084 REGION_ALIAS("REGION_BSS", RAM);
4088 It is possible to write a common system initialization routine to copy the
4089 @code{.data} section from @code{ROM} or @code{ROM2} into the @code{RAM} if
4094 extern char data_start [];
4095 extern char data_size [];
4096 extern char data_load_start [];
4098 void copy_data(void)
4100 if (data_start != data_load_start)
4102 memcpy(data_start, data_load_start, (size_t) data_size);
4107 @node Miscellaneous Commands
4108 @subsection Other Linker Script Commands
4109 There are a few other linker scripts commands.
4112 @item ASSERT(@var{exp}, @var{message})
4114 @cindex assertion in linker script
4115 Ensure that @var{exp} is non-zero. If it is zero, then exit the linker
4116 with an error code, and print @var{message}.
4118 Note that assertions are checked before the final stages of linking
4119 take place. This means that expressions involving symbols PROVIDEd
4120 inside section definitions will fail if the user has not set values
4121 for those symbols. The only exception to this rule is PROVIDEd
4122 symbols that just reference dot. Thus an assertion like this:
4127 PROVIDE (__stack = .);
4128 PROVIDE (__stack_size = 0x100);
4129 ASSERT ((__stack > (_end + __stack_size)), "Error: No room left for the stack");
4133 will fail if @code{__stack_size} is not defined elsewhere. Symbols
4134 PROVIDEd outside of section definitions are evaluated earlier, so they
4135 can be used inside ASSERTions. Thus:
4138 PROVIDE (__stack_size = 0x100);
4141 PROVIDE (__stack = .);
4142 ASSERT ((__stack > (_end + __stack_size)), "Error: No room left for the stack");
4148 @item EXTERN(@var{symbol} @var{symbol} @dots{})
4150 @cindex undefined symbol in linker script
4151 Force @var{symbol} to be entered in the output file as an undefined
4152 symbol. Doing this may, for example, trigger linking of additional
4153 modules from standard libraries. You may list several @var{symbol}s for
4154 each @code{EXTERN}, and you may use @code{EXTERN} multiple times. This
4155 command has the same effect as the @samp{-u} command-line option.
4157 @item FORCE_COMMON_ALLOCATION
4158 @kindex FORCE_COMMON_ALLOCATION
4159 @cindex common allocation in linker script
4160 This command has the same effect as the @samp{-d} command-line option:
4161 to make @command{ld} assign space to common symbols even if a relocatable
4162 output file is specified (@samp{-r}).
4164 @item INHIBIT_COMMON_ALLOCATION
4165 @kindex INHIBIT_COMMON_ALLOCATION
4166 @cindex common allocation in linker script
4167 This command has the same effect as the @samp{--no-define-common}
4168 command-line option: to make @code{ld} omit the assignment of addresses
4169 to common symbols even for a non-relocatable output file.
4171 @item FORCE_GROUP_ALLOCATION
4172 @kindex FORCE_GROUP_ALLOCATION
4173 @cindex group allocation in linker script
4174 @cindex section groups
4176 This command has the same effect as the
4177 @samp{--force-group-allocation} command-line option: to make
4178 @command{ld} place section group members like normal input sections,
4179 and to delete the section groups even if a relocatable output file is
4180 specified (@samp{-r}).
4182 @item INSERT [ AFTER | BEFORE ] @var{output_section}
4184 @cindex insert user script into default script
4185 This command is typically used in a script specified by @samp{-T} to
4186 augment the default @code{SECTIONS} with, for example, overlays. It
4187 inserts all prior linker script statements after (or before)
4188 @var{output_section}, and also causes @samp{-T} to not override the
4189 default linker script. The exact insertion point is as for orphan
4190 sections. @xref{Location Counter}. The insertion happens after the
4191 linker has mapped input sections to output sections. Prior to the
4192 insertion, since @samp{-T} scripts are parsed before the default
4193 linker script, statements in the @samp{-T} script occur before the
4194 default linker script statements in the internal linker representation
4195 of the script. In particular, input section assignments will be made
4196 to @samp{-T} output sections before those in the default script. Here
4197 is an example of how a @samp{-T} script using @code{INSERT} might look:
4204 .ov1 @{ ov1*(.text) @}
4205 .ov2 @{ ov2*(.text) @}
4211 @item NOCROSSREFS(@var{section} @var{section} @dots{})
4212 @kindex NOCROSSREFS(@var{sections})
4213 @cindex cross references
4214 This command may be used to tell @command{ld} to issue an error about any
4215 references among certain output sections.
4217 In certain types of programs, particularly on embedded systems when
4218 using overlays, when one section is loaded into memory, another section
4219 will not be. Any direct references between the two sections would be
4220 errors. For example, it would be an error if code in one section called
4221 a function defined in the other section.
4223 The @code{NOCROSSREFS} command takes a list of output section names. If
4224 @command{ld} detects any cross references between the sections, it reports
4225 an error and returns a non-zero exit status. Note that the
4226 @code{NOCROSSREFS} command uses output section names, not input section
4229 @item NOCROSSREFS_TO(@var{tosection} @var{fromsection} @dots{})
4230 @kindex NOCROSSREFS_TO(@var{tosection} @var{fromsections})
4231 @cindex cross references
4232 This command may be used to tell @command{ld} to issue an error about any
4233 references to one section from a list of other sections.
4235 The @code{NOCROSSREFS} command is useful when ensuring that two or more
4236 output sections are entirely independent but there are situations where
4237 a one-way dependency is needed. For example, in a multi-core application
4238 there may be shared code that can be called from each core but for safety
4239 must never call back.
4241 The @code{NOCROSSREFS_TO} command takes a list of output section names.
4242 The first section can not be referenced from any of the other sections.
4243 If @command{ld} detects any references to the first section from any of
4244 the other sections, it reports an error and returns a non-zero exit
4245 status. Note that the @code{NOCROSSREFS_TO} command uses output section
4246 names, not input section names.
4248 @ifclear SingleFormat
4249 @item OUTPUT_ARCH(@var{bfdarch})
4250 @kindex OUTPUT_ARCH(@var{bfdarch})
4251 @cindex machine architecture
4252 @cindex architecture
4253 Specify a particular output machine architecture. The argument is one
4254 of the names used by the BFD library (@pxref{BFD}). You can see the
4255 architecture of an object file by using the @code{objdump} program with
4256 the @samp{-f} option.
4259 @item LD_FEATURE(@var{string})
4260 @kindex LD_FEATURE(@var{string})
4261 This command may be used to modify @command{ld} behavior. If
4262 @var{string} is @code{"SANE_EXPR"} then absolute symbols and numbers
4263 in a script are simply treated as numbers everywhere.
4264 @xref{Expression Section}.
4268 @section Assigning Values to Symbols
4269 @cindex assignment in scripts
4270 @cindex symbol definition, scripts
4271 @cindex variables, defining
4272 You may assign a value to a symbol in a linker script. This will define
4273 the symbol and place it into the symbol table with a global scope.
4276 * Simple Assignments:: Simple Assignments
4279 * PROVIDE_HIDDEN:: PROVIDE_HIDDEN
4280 * Source Code Reference:: How to use a linker script defined symbol in source code
4283 @node Simple Assignments
4284 @subsection Simple Assignments
4286 You may assign to a symbol using any of the C assignment operators:
4289 @item @var{symbol} = @var{expression} ;
4290 @itemx @var{symbol} += @var{expression} ;
4291 @itemx @var{symbol} -= @var{expression} ;
4292 @itemx @var{symbol} *= @var{expression} ;
4293 @itemx @var{symbol} /= @var{expression} ;
4294 @itemx @var{symbol} <<= @var{expression} ;
4295 @itemx @var{symbol} >>= @var{expression} ;
4296 @itemx @var{symbol} &= @var{expression} ;
4297 @itemx @var{symbol} |= @var{expression} ;
4300 The first case will define @var{symbol} to the value of
4301 @var{expression}. In the other cases, @var{symbol} must already be
4302 defined, and the value will be adjusted accordingly.
4304 The special symbol name @samp{.} indicates the location counter. You
4305 may only use this within a @code{SECTIONS} command. @xref{Location Counter}.
4307 The semicolon after @var{expression} is required.
4309 Expressions are defined below; see @ref{Expressions}.
4311 You may write symbol assignments as commands in their own right, or as
4312 statements within a @code{SECTIONS} command, or as part of an output
4313 section description in a @code{SECTIONS} command.
4315 The section of the symbol will be set from the section of the
4316 expression; for more information, see @ref{Expression Section}.
4318 Here is an example showing the three different places that symbol
4319 assignments may be used:
4330 _bdata = (. + 3) & ~ 3;
4331 .data : @{ *(.data) @}
4335 In this example, the symbol @samp{floating_point} will be defined as
4336 zero. The symbol @samp{_etext} will be defined as the address following
4337 the last @samp{.text} input section. The symbol @samp{_bdata} will be
4338 defined as the address following the @samp{.text} output section aligned
4339 upward to a 4 byte boundary.
4344 For ELF targeted ports, define a symbol that will be hidden and won't be
4345 exported. The syntax is @code{HIDDEN(@var{symbol} = @var{expression})}.
4347 Here is the example from @ref{Simple Assignments}, rewritten to use
4351 HIDDEN(floating_point = 0);
4359 HIDDEN(_bdata = (. + 3) & ~ 3);
4360 .data : @{ *(.data) @}
4364 In this case none of the three symbols will be visible outside this module.
4369 In some cases, it is desirable for a linker script to define a symbol
4370 only if it is referenced and is not defined by any object included in
4371 the link. For example, traditional linkers defined the symbol
4372 @samp{etext}. However, ANSI C requires that the user be able to use
4373 @samp{etext} as a function name without encountering an error. The
4374 @code{PROVIDE} keyword may be used to define a symbol, such as
4375 @samp{etext}, only if it is referenced but not defined. The syntax is
4376 @code{PROVIDE(@var{symbol} = @var{expression})}.
4378 Here is an example of using @code{PROVIDE} to define @samp{etext}:
4391 In this example, if the program defines @samp{_etext} (with a leading
4392 underscore), the linker will give a multiple definition error. If, on
4393 the other hand, the program defines @samp{etext} (with no leading
4394 underscore), the linker will silently use the definition in the program.
4395 If the program references @samp{etext} but does not define it, the
4396 linker will use the definition in the linker script.
4398 Note - the @code{PROVIDE} directive considers a common symbol to be
4399 defined, even though such a symbol could be combined with the symbol
4400 that the @code{PROVIDE} would create. This is particularly important
4401 when considering constructor and destructor list symbols such as
4402 @samp{__CTOR_LIST__} as these are often defined as common symbols.
4404 @node PROVIDE_HIDDEN
4405 @subsection PROVIDE_HIDDEN
4406 @cindex PROVIDE_HIDDEN
4407 Similar to @code{PROVIDE}. For ELF targeted ports, the symbol will be
4408 hidden and won't be exported.
4410 @node Source Code Reference
4411 @subsection Source Code Reference
4413 Accessing a linker script defined variable from source code is not
4414 intuitive. In particular a linker script symbol is not equivalent to
4415 a variable declaration in a high level language, it is instead a
4416 symbol that does not have a value.
4418 Before going further, it is important to note that compilers often
4419 transform names in the source code into different names when they are
4420 stored in the symbol table. For example, Fortran compilers commonly
4421 prepend or append an underscore, and C++ performs extensive @samp{name
4422 mangling}. Therefore there might be a discrepancy between the name
4423 of a variable as it is used in source code and the name of the same
4424 variable as it is defined in a linker script. For example in C a
4425 linker script variable might be referred to as:
4431 But in the linker script it might be defined as:
4437 In the remaining examples however it is assumed that no name
4438 transformation has taken place.
4440 When a symbol is declared in a high level language such as C, two
4441 things happen. The first is that the compiler reserves enough space
4442 in the program's memory to hold the @emph{value} of the symbol. The
4443 second is that the compiler creates an entry in the program's symbol
4444 table which holds the symbol's @emph{address}. ie the symbol table
4445 contains the address of the block of memory holding the symbol's
4446 value. So for example the following C declaration, at file scope:
4452 creates an entry called @samp{foo} in the symbol table. This entry
4453 holds the address of an @samp{int} sized block of memory where the
4454 number 1000 is initially stored.
4456 When a program references a symbol the compiler generates code that
4457 first accesses the symbol table to find the address of the symbol's
4458 memory block and then code to read the value from that memory block.
4465 looks up the symbol @samp{foo} in the symbol table, gets the address
4466 associated with this symbol and then writes the value 1 into that
4473 looks up the symbol @samp{foo} in the symbol table, gets its address
4474 and then copies this address into the block of memory associated with
4475 the variable @samp{a}.
4477 Linker scripts symbol declarations, by contrast, create an entry in
4478 the symbol table but do not assign any memory to them. Thus they are
4479 an address without a value. So for example the linker script definition:
4485 creates an entry in the symbol table called @samp{foo} which holds
4486 the address of memory location 1000, but nothing special is stored at
4487 address 1000. This means that you cannot access the @emph{value} of a
4488 linker script defined symbol - it has no value - all you can do is
4489 access the @emph{address} of a linker script defined symbol.
4491 Hence when you are using a linker script defined symbol in source code
4492 you should always take the address of the symbol, and never attempt to
4493 use its value. For example suppose you want to copy the contents of a
4494 section of memory called .ROM into a section called .FLASH and the
4495 linker script contains these declarations:
4499 start_of_ROM = .ROM;
4500 end_of_ROM = .ROM + sizeof (.ROM);
4501 start_of_FLASH = .FLASH;
4505 Then the C source code to perform the copy would be:
4509 extern char start_of_ROM, end_of_ROM, start_of_FLASH;
4511 memcpy (& start_of_FLASH, & start_of_ROM, & end_of_ROM - & start_of_ROM);
4515 Note the use of the @samp{&} operators. These are correct.
4516 Alternatively the symbols can be treated as the names of vectors or
4517 arrays and then the code will again work as expected:
4521 extern char start_of_ROM[], end_of_ROM[], start_of_FLASH[];
4523 memcpy (start_of_FLASH, start_of_ROM, end_of_ROM - start_of_ROM);
4527 Note how using this method does not require the use of @samp{&}
4531 @section SECTIONS Command
4533 The @code{SECTIONS} command tells the linker how to map input sections
4534 into output sections, and how to place the output sections in memory.
4536 The format of the @code{SECTIONS} command is:
4540 @var{sections-command}
4541 @var{sections-command}
4546 Each @var{sections-command} may of be one of the following:
4550 an @code{ENTRY} command (@pxref{Entry Point,,Entry command})
4552 a symbol assignment (@pxref{Assignments})
4554 an output section description
4556 an overlay description
4559 The @code{ENTRY} command and symbol assignments are permitted inside the
4560 @code{SECTIONS} command for convenience in using the location counter in
4561 those commands. This can also make the linker script easier to
4562 understand because you can use those commands at meaningful points in
4563 the layout of the output file.
4565 Output section descriptions and overlay descriptions are described
4568 If you do not use a @code{SECTIONS} command in your linker script, the
4569 linker will place each input section into an identically named output
4570 section in the order that the sections are first encountered in the
4571 input files. If all input sections are present in the first file, for
4572 example, the order of sections in the output file will match the order
4573 in the first input file. The first section will be at address zero.
4576 * Output Section Description:: Output section description
4577 * Output Section Name:: Output section name
4578 * Output Section Address:: Output section address
4579 * Input Section:: Input section description
4580 * Output Section Data:: Output section data
4581 * Output Section Keywords:: Output section keywords
4582 * Output Section Discarding:: Output section discarding
4583 * Output Section Attributes:: Output section attributes
4584 * Overlay Description:: Overlay description
4587 @node Output Section Description
4588 @subsection Output Section Description
4589 The full description of an output section looks like this:
4592 @var{section} [@var{address}] [(@var{type})] :
4594 [ALIGN(@var{section_align}) | ALIGN_WITH_INPUT]
4595 [SUBALIGN(@var{subsection_align})]
4598 @var{output-section-command}
4599 @var{output-section-command}
4601 @} [>@var{region}] [AT>@var{lma_region}] [:@var{phdr} :@var{phdr} @dots{}] [=@var{fillexp}] [,]
4605 Most output sections do not use most of the optional section attributes.
4607 The whitespace around @var{section} is required, so that the section
4608 name is unambiguous. The colon and the curly braces are also required.
4609 The comma at the end may be required if a @var{fillexp} is used and
4610 the next @var{sections-command} looks like a continuation of the expression.
4611 The line breaks and other white space are optional.
4613 Each @var{output-section-command} may be one of the following:
4617 a symbol assignment (@pxref{Assignments})
4619 an input section description (@pxref{Input Section})
4621 data values to include directly (@pxref{Output Section Data})
4623 a special output section keyword (@pxref{Output Section Keywords})
4626 @node Output Section Name
4627 @subsection Output Section Name
4628 @cindex name, section
4629 @cindex section name
4630 The name of the output section is @var{section}. @var{section} must
4631 meet the constraints of your output format. In formats which only
4632 support a limited number of sections, such as @code{a.out}, the name
4633 must be one of the names supported by the format (@code{a.out}, for
4634 example, allows only @samp{.text}, @samp{.data} or @samp{.bss}). If the
4635 output format supports any number of sections, but with numbers and not
4636 names (as is the case for Oasys), the name should be supplied as a
4637 quoted numeric string. A section name may consist of any sequence of
4638 characters, but a name which contains any unusual characters such as
4639 commas must be quoted.
4641 The output section name @samp{/DISCARD/} is special; @ref{Output Section
4644 @node Output Section Address
4645 @subsection Output Section Address
4646 @cindex address, section
4647 @cindex section address
4648 The @var{address} is an expression for the VMA (the virtual memory
4649 address) of the output section. This address is optional, but if it
4650 is provided then the output address will be set exactly as specified.
4652 If the output address is not specified then one will be chosen for the
4653 section, based on the heuristic below. This address will be adjusted
4654 to fit the alignment requirement of the output section. The
4655 alignment requirement is the strictest alignment of any input section
4656 contained within the output section.
4658 The output section address heuristic is as follows:
4662 If an output memory @var{region} is set for the section then it
4663 is added to this region and its address will be the next free address
4667 If the MEMORY command has been used to create a list of memory
4668 regions then the first region which has attributes compatible with the
4669 section is selected to contain it. The section's output address will
4670 be the next free address in that region; @ref{MEMORY}.
4673 If no memory regions were specified, or none match the section then
4674 the output address will be based on the current value of the location
4682 .text . : @{ *(.text) @}
4689 .text : @{ *(.text) @}
4693 are subtly different. The first will set the address of the
4694 @samp{.text} output section to the current value of the location
4695 counter. The second will set it to the current value of the location
4696 counter aligned to the strictest alignment of any of the @samp{.text}
4699 The @var{address} may be an arbitrary expression; @ref{Expressions}.
4700 For example, if you want to align the section on a 0x10 byte boundary,
4701 so that the lowest four bits of the section address are zero, you could
4702 do something like this:
4704 .text ALIGN(0x10) : @{ *(.text) @}
4707 This works because @code{ALIGN} returns the current location counter
4708 aligned upward to the specified value.
4710 Specifying @var{address} for a section will change the value of the
4711 location counter, provided that the section is non-empty. (Empty
4712 sections are ignored).
4715 @subsection Input Section Description
4716 @cindex input sections
4717 @cindex mapping input sections to output sections
4718 The most common output section command is an input section description.
4720 The input section description is the most basic linker script operation.
4721 You use output sections to tell the linker how to lay out your program
4722 in memory. You use input section descriptions to tell the linker how to
4723 map the input files into your memory layout.
4726 * Input Section Basics:: Input section basics
4727 * Input Section Wildcards:: Input section wildcard patterns
4728 * Input Section Common:: Input section for common symbols
4729 * Input Section Keep:: Input section and garbage collection
4730 * Input Section Example:: Input section example
4733 @node Input Section Basics
4734 @subsubsection Input Section Basics
4735 @cindex input section basics
4736 An input section description consists of a file name optionally followed
4737 by a list of section names in parentheses.
4739 The file name and the section name may be wildcard patterns, which we
4740 describe further below (@pxref{Input Section Wildcards}).
4742 The most common input section description is to include all input
4743 sections with a particular name in the output section. For example, to
4744 include all input @samp{.text} sections, you would write:
4749 Here the @samp{*} is a wildcard which matches any file name. To exclude a list
4750 @cindex EXCLUDE_FILE
4751 of files from matching the file name wildcard, EXCLUDE_FILE may be used to
4752 match all files except the ones specified in the EXCLUDE_FILE list. For
4755 EXCLUDE_FILE (*crtend.o *otherfile.o) *(.ctors)
4758 will cause all .ctors sections from all files except @file{crtend.o}
4759 and @file{otherfile.o} to be included. The EXCLUDE_FILE can also be
4760 placed inside the section list, for example:
4762 *(EXCLUDE_FILE (*crtend.o *otherfile.o) .ctors)
4765 The result of this is identically to the previous example. Supporting
4766 two syntaxes for EXCLUDE_FILE is useful if the section list contains
4767 more than one section, as described below.
4769 There are two ways to include more than one section:
4775 The difference between these is the order in which the @samp{.text} and
4776 @samp{.rdata} input sections will appear in the output section. In the
4777 first example, they will be intermingled, appearing in the same order as
4778 they are found in the linker input. In the second example, all
4779 @samp{.text} input sections will appear first, followed by all
4780 @samp{.rdata} input sections.
4782 When using EXCLUDE_FILE with more than one section, if the exclusion
4783 is within the section list then the exclusion only applies to the
4784 immediately following section, for example:
4786 *(EXCLUDE_FILE (*somefile.o) .text .rdata)
4789 will cause all @samp{.text} sections from all files except
4790 @file{somefile.o} to be included, while all @samp{.rdata} sections
4791 from all files, including @file{somefile.o}, will be included. To
4792 exclude the @samp{.rdata} sections from @file{somefile.o} the example
4793 could be modified to:
4795 *(EXCLUDE_FILE (*somefile.o) .text EXCLUDE_FILE (*somefile.o) .rdata)
4798 Alternatively, placing the EXCLUDE_FILE outside of the section list,
4799 before the input file selection, will cause the exclusion to apply for
4800 all sections. Thus the previous example can be rewritten as:
4802 EXCLUDE_FILE (*somefile.o) *(.text .rdata)
4805 You can specify a file name to include sections from a particular file.
4806 You would do this if one or more of your files contain special data that
4807 needs to be at a particular location in memory. For example:
4812 To refine the sections that are included based on the section flags
4813 of an input section, INPUT_SECTION_FLAGS may be used.
4815 Here is a simple example for using Section header flags for ELF sections:
4820 .text : @{ INPUT_SECTION_FLAGS (SHF_MERGE & SHF_STRINGS) *(.text) @}
4821 .text2 : @{ INPUT_SECTION_FLAGS (!SHF_WRITE) *(.text) @}
4826 In this example, the output section @samp{.text} will be comprised of any
4827 input section matching the name *(.text) whose section header flags
4828 @code{SHF_MERGE} and @code{SHF_STRINGS} are set. The output section
4829 @samp{.text2} will be comprised of any input section matching the name *(.text)
4830 whose section header flag @code{SHF_WRITE} is clear.
4832 You can also specify files within archives by writing a pattern
4833 matching the archive, a colon, then the pattern matching the file,
4834 with no whitespace around the colon.
4838 matches file within archive
4840 matches the whole archive
4842 matches file but not one in an archive
4845 Either one or both of @samp{archive} and @samp{file} can contain shell
4846 wildcards. On DOS based file systems, the linker will assume that a
4847 single letter followed by a colon is a drive specifier, so
4848 @samp{c:myfile.o} is a simple file specification, not @samp{myfile.o}
4849 within an archive called @samp{c}. @samp{archive:file} filespecs may
4850 also be used within an @code{EXCLUDE_FILE} list, but may not appear in
4851 other linker script contexts. For instance, you cannot extract a file
4852 from an archive by using @samp{archive:file} in an @code{INPUT}
4855 If you use a file name without a list of sections, then all sections in
4856 the input file will be included in the output section. This is not
4857 commonly done, but it may by useful on occasion. For example:
4862 When you use a file name which is not an @samp{archive:file} specifier
4863 and does not contain any wild card
4864 characters, the linker will first see if you also specified the file
4865 name on the linker command line or in an @code{INPUT} command. If you
4866 did not, the linker will attempt to open the file as an input file, as
4867 though it appeared on the command line. Note that this differs from an
4868 @code{INPUT} command, because the linker will not search for the file in
4869 the archive search path.
4871 @node Input Section Wildcards
4872 @subsubsection Input Section Wildcard Patterns
4873 @cindex input section wildcards
4874 @cindex wildcard file name patterns
4875 @cindex file name wildcard patterns
4876 @cindex section name wildcard patterns
4877 In an input section description, either the file name or the section
4878 name or both may be wildcard patterns.
4880 The file name of @samp{*} seen in many examples is a simple wildcard
4881 pattern for the file name.
4883 The wildcard patterns are like those used by the Unix shell.
4887 matches any number of characters
4889 matches any single character
4891 matches a single instance of any of the @var{chars}; the @samp{-}
4892 character may be used to specify a range of characters, as in
4893 @samp{[a-z]} to match any lower case letter
4895 quotes the following character
4898 When a file name is matched with a wildcard, the wildcard characters
4899 will not match a @samp{/} character (used to separate directory names on
4900 Unix). A pattern consisting of a single @samp{*} character is an
4901 exception; it will always match any file name, whether it contains a
4902 @samp{/} or not. In a section name, the wildcard characters will match
4903 a @samp{/} character.
4905 File name wildcard patterns only match files which are explicitly
4906 specified on the command line or in an @code{INPUT} command. The linker
4907 does not search directories to expand wildcards.
4909 If a file name matches more than one wildcard pattern, or if a file name
4910 appears explicitly and is also matched by a wildcard pattern, the linker
4911 will use the first match in the linker script. For example, this
4912 sequence of input section descriptions is probably in error, because the
4913 @file{data.o} rule will not be used:
4915 .data : @{ *(.data) @}
4916 .data1 : @{ data.o(.data) @}
4919 @cindex SORT_BY_NAME
4920 Normally, the linker will place files and sections matched by wildcards
4921 in the order in which they are seen during the link. You can change
4922 this by using the @code{SORT_BY_NAME} keyword, which appears before a wildcard
4923 pattern in parentheses (e.g., @code{SORT_BY_NAME(.text*)}). When the
4924 @code{SORT_BY_NAME} keyword is used, the linker will sort the files or sections
4925 into ascending order by name before placing them in the output file.
4927 @cindex SORT_BY_ALIGNMENT
4928 @code{SORT_BY_ALIGNMENT} is similar to @code{SORT_BY_NAME}.
4929 @code{SORT_BY_ALIGNMENT} will sort sections into descending order of
4930 alignment before placing them in the output file. Placing larger
4931 alignments before smaller alignments can reduce the amount of padding
4934 @cindex SORT_BY_INIT_PRIORITY
4935 @code{SORT_BY_INIT_PRIORITY} is also similar to @code{SORT_BY_NAME}.
4936 @code{SORT_BY_INIT_PRIORITY} will sort sections into ascending
4937 numerical order of the GCC init_priority attribute encoded in the
4938 section name before placing them in the output file. In
4939 @code{.init_array.NNNNN} and @code{.fini_array.NNNNN}, @code{NNNNN} is
4940 the init_priority. In @code{.ctors.NNNNN} and @code{.dtors.NNNNN},
4941 @code{NNNNN} is 65535 minus the init_priority.
4944 @code{SORT} is an alias for @code{SORT_BY_NAME}.
4946 When there are nested section sorting commands in linker script, there
4947 can be at most 1 level of nesting for section sorting commands.
4951 @code{SORT_BY_NAME} (@code{SORT_BY_ALIGNMENT} (wildcard section pattern)).
4952 It will sort the input sections by name first, then by alignment if two
4953 sections have the same name.
4955 @code{SORT_BY_ALIGNMENT} (@code{SORT_BY_NAME} (wildcard section pattern)).
4956 It will sort the input sections by alignment first, then by name if two
4957 sections have the same alignment.
4959 @code{SORT_BY_NAME} (@code{SORT_BY_NAME} (wildcard section pattern)) is
4960 treated the same as @code{SORT_BY_NAME} (wildcard section pattern).
4962 @code{SORT_BY_ALIGNMENT} (@code{SORT_BY_ALIGNMENT} (wildcard section pattern))
4963 is treated the same as @code{SORT_BY_ALIGNMENT} (wildcard section pattern).
4965 All other nested section sorting commands are invalid.
4968 When both command-line section sorting option and linker script
4969 section sorting command are used, section sorting command always
4970 takes precedence over the command-line option.
4972 If the section sorting command in linker script isn't nested, the
4973 command-line option will make the section sorting command to be
4974 treated as nested sorting command.
4978 @code{SORT_BY_NAME} (wildcard section pattern ) with
4979 @option{--sort-sections alignment} is equivalent to
4980 @code{SORT_BY_NAME} (@code{SORT_BY_ALIGNMENT} (wildcard section pattern)).
4982 @code{SORT_BY_ALIGNMENT} (wildcard section pattern) with
4983 @option{--sort-section name} is equivalent to
4984 @code{SORT_BY_ALIGNMENT} (@code{SORT_BY_NAME} (wildcard section pattern)).
4987 If the section sorting command in linker script is nested, the
4988 command-line option will be ignored.
4991 @code{SORT_NONE} disables section sorting by ignoring the command-line
4992 section sorting option.
4994 If you ever get confused about where input sections are going, use the
4995 @samp{-M} linker option to generate a map file. The map file shows
4996 precisely how input sections are mapped to output sections.
4998 This example shows how wildcard patterns might be used to partition
4999 files. This linker script directs the linker to place all @samp{.text}
5000 sections in @samp{.text} and all @samp{.bss} sections in @samp{.bss}.
5001 The linker will place the @samp{.data} section from all files beginning
5002 with an upper case character in @samp{.DATA}; for all other files, the
5003 linker will place the @samp{.data} section in @samp{.data}.
5007 .text : @{ *(.text) @}
5008 .DATA : @{ [A-Z]*(.data) @}
5009 .data : @{ *(.data) @}
5010 .bss : @{ *(.bss) @}
5015 @node Input Section Common
5016 @subsubsection Input Section for Common Symbols
5017 @cindex common symbol placement
5018 @cindex uninitialized data placement
5019 A special notation is needed for common symbols, because in many object
5020 file formats common symbols do not have a particular input section. The
5021 linker treats common symbols as though they are in an input section
5022 named @samp{COMMON}.
5024 You may use file names with the @samp{COMMON} section just as with any
5025 other input sections. You can use this to place common symbols from a
5026 particular input file in one section while common symbols from other
5027 input files are placed in another section.
5029 In most cases, common symbols in input files will be placed in the
5030 @samp{.bss} section in the output file. For example:
5032 .bss @{ *(.bss) *(COMMON) @}
5035 @cindex scommon section
5036 @cindex small common symbols
5037 Some object file formats have more than one type of common symbol. For
5038 example, the MIPS ELF object file format distinguishes standard common
5039 symbols and small common symbols. In this case, the linker will use a
5040 different special section name for other types of common symbols. In
5041 the case of MIPS ELF, the linker uses @samp{COMMON} for standard common
5042 symbols and @samp{.scommon} for small common symbols. This permits you
5043 to map the different types of common symbols into memory at different
5047 You will sometimes see @samp{[COMMON]} in old linker scripts. This
5048 notation is now considered obsolete. It is equivalent to
5051 @node Input Section Keep
5052 @subsubsection Input Section and Garbage Collection
5054 @cindex garbage collection
5055 When link-time garbage collection is in use (@samp{--gc-sections}),
5056 it is often useful to mark sections that should not be eliminated.
5057 This is accomplished by surrounding an input section's wildcard entry
5058 with @code{KEEP()}, as in @code{KEEP(*(.init))} or
5059 @code{KEEP(SORT_BY_NAME(*)(.ctors))}.
5061 @node Input Section Example
5062 @subsubsection Input Section Example
5063 The following example is a complete linker script. It tells the linker
5064 to read all of the sections from file @file{all.o} and place them at the
5065 start of output section @samp{outputa} which starts at location
5066 @samp{0x10000}. All of section @samp{.input1} from file @file{foo.o}
5067 follows immediately, in the same output section. All of section
5068 @samp{.input2} from @file{foo.o} goes into output section
5069 @samp{outputb}, followed by section @samp{.input1} from @file{foo1.o}.
5070 All of the remaining @samp{.input1} and @samp{.input2} sections from any
5071 files are written to output section @samp{outputc}.
5099 If an output section's name is the same as the input section's name
5100 and is representable as a C identifier, then the linker will
5101 automatically @pxref{PROVIDE} two symbols: __start_SECNAME and
5102 __stop_SECNAME, where SECNAME is the name of the section. These
5103 indicate the start address and end address of the output section
5104 respectively. Note: most section names are not representable as
5105 C identifiers because they contain a @samp{.} character.
5107 @node Output Section Data
5108 @subsection Output Section Data
5110 @cindex section data
5111 @cindex output section data
5112 @kindex BYTE(@var{expression})
5113 @kindex SHORT(@var{expression})
5114 @kindex LONG(@var{expression})
5115 @kindex QUAD(@var{expression})
5116 @kindex SQUAD(@var{expression})
5117 You can include explicit bytes of data in an output section by using
5118 @code{BYTE}, @code{SHORT}, @code{LONG}, @code{QUAD}, or @code{SQUAD} as
5119 an output section command. Each keyword is followed by an expression in
5120 parentheses providing the value to store (@pxref{Expressions}). The
5121 value of the expression is stored at the current value of the location
5124 The @code{BYTE}, @code{SHORT}, @code{LONG}, and @code{QUAD} commands
5125 store one, two, four, and eight bytes (respectively). After storing the
5126 bytes, the location counter is incremented by the number of bytes
5129 For example, this will store the byte 1 followed by the four byte value
5130 of the symbol @samp{addr}:
5136 When using a 64 bit host or target, @code{QUAD} and @code{SQUAD} are the
5137 same; they both store an 8 byte, or 64 bit, value. When both host and
5138 target are 32 bits, an expression is computed as 32 bits. In this case
5139 @code{QUAD} stores a 32 bit value zero extended to 64 bits, and
5140 @code{SQUAD} stores a 32 bit value sign extended to 64 bits.
5142 If the object file format of the output file has an explicit endianness,
5143 which is the normal case, the value will be stored in that endianness.
5144 When the object file format does not have an explicit endianness, as is
5145 true of, for example, S-records, the value will be stored in the
5146 endianness of the first input object file.
5148 Note---these commands only work inside a section description and not
5149 between them, so the following will produce an error from the linker:
5151 SECTIONS @{@ .text : @{@ *(.text) @}@ LONG(1) .data : @{@ *(.data) @}@ @}@
5153 whereas this will work:
5155 SECTIONS @{@ .text : @{@ *(.text) ; LONG(1) @}@ .data : @{@ *(.data) @}@ @}@
5158 @kindex FILL(@var{expression})
5159 @cindex holes, filling
5160 @cindex unspecified memory
5161 You may use the @code{FILL} command to set the fill pattern for the
5162 current section. It is followed by an expression in parentheses. Any
5163 otherwise unspecified regions of memory within the section (for example,
5164 gaps left due to the required alignment of input sections) are filled
5165 with the value of the expression, repeated as
5166 necessary. A @code{FILL} statement covers memory locations after the
5167 point at which it occurs in the section definition; by including more
5168 than one @code{FILL} statement, you can have different fill patterns in
5169 different parts of an output section.
5171 This example shows how to fill unspecified regions of memory with the
5177 The @code{FILL} command is similar to the @samp{=@var{fillexp}} output
5178 section attribute, but it only affects the
5179 part of the section following the @code{FILL} command, rather than the
5180 entire section. If both are used, the @code{FILL} command takes
5181 precedence. @xref{Output Section Fill}, for details on the fill
5184 @node Output Section Keywords
5185 @subsection Output Section Keywords
5186 There are a couple of keywords which can appear as output section
5190 @kindex CREATE_OBJECT_SYMBOLS
5191 @cindex input filename symbols
5192 @cindex filename symbols
5193 @item CREATE_OBJECT_SYMBOLS
5194 The command tells the linker to create a symbol for each input file.
5195 The name of each symbol will be the name of the corresponding input
5196 file. The section of each symbol will be the output section in which
5197 the @code{CREATE_OBJECT_SYMBOLS} command appears.
5199 This is conventional for the a.out object file format. It is not
5200 normally used for any other object file format.
5202 @kindex CONSTRUCTORS
5203 @cindex C++ constructors, arranging in link
5204 @cindex constructors, arranging in link
5206 When linking using the a.out object file format, the linker uses an
5207 unusual set construct to support C++ global constructors and
5208 destructors. When linking object file formats which do not support
5209 arbitrary sections, such as ECOFF and XCOFF, the linker will
5210 automatically recognize C++ global constructors and destructors by name.
5211 For these object file formats, the @code{CONSTRUCTORS} command tells the
5212 linker to place constructor information in the output section where the
5213 @code{CONSTRUCTORS} command appears. The @code{CONSTRUCTORS} command is
5214 ignored for other object file formats.
5216 The symbol @w{@code{__CTOR_LIST__}} marks the start of the global
5217 constructors, and the symbol @w{@code{__CTOR_END__}} marks the end.
5218 Similarly, @w{@code{__DTOR_LIST__}} and @w{@code{__DTOR_END__}} mark
5219 the start and end of the global destructors. The
5220 first word in the list is the number of entries, followed by the address
5221 of each constructor or destructor, followed by a zero word. The
5222 compiler must arrange to actually run the code. For these object file
5223 formats @sc{gnu} C++ normally calls constructors from a subroutine
5224 @code{__main}; a call to @code{__main} is automatically inserted into
5225 the startup code for @code{main}. @sc{gnu} C++ normally runs
5226 destructors either by using @code{atexit}, or directly from the function
5229 For object file formats such as @code{COFF} or @code{ELF} which support
5230 arbitrary section names, @sc{gnu} C++ will normally arrange to put the
5231 addresses of global constructors and destructors into the @code{.ctors}
5232 and @code{.dtors} sections. Placing the following sequence into your
5233 linker script will build the sort of table which the @sc{gnu} C++
5234 runtime code expects to see.
5238 LONG((__CTOR_END__ - __CTOR_LIST__) / 4 - 2)
5243 LONG((__DTOR_END__ - __DTOR_LIST__) / 4 - 2)
5249 If you are using the @sc{gnu} C++ support for initialization priority,
5250 which provides some control over the order in which global constructors
5251 are run, you must sort the constructors at link time to ensure that they
5252 are executed in the correct order. When using the @code{CONSTRUCTORS}
5253 command, use @samp{SORT_BY_NAME(CONSTRUCTORS)} instead. When using the
5254 @code{.ctors} and @code{.dtors} sections, use @samp{*(SORT_BY_NAME(.ctors))} and
5255 @samp{*(SORT_BY_NAME(.dtors))} instead of just @samp{*(.ctors)} and
5258 Normally the compiler and linker will handle these issues automatically,
5259 and you will not need to concern yourself with them. However, you may
5260 need to consider this if you are using C++ and writing your own linker
5265 @node Output Section Discarding
5266 @subsection Output Section Discarding
5267 @cindex discarding sections
5268 @cindex sections, discarding
5269 @cindex removing sections
5270 The linker will not normally create output sections with no contents.
5271 This is for convenience when referring to input sections that may or
5272 may not be present in any of the input files. For example:
5274 .foo : @{ *(.foo) @}
5277 will only create a @samp{.foo} section in the output file if there is a
5278 @samp{.foo} section in at least one input file, and if the input
5279 sections are not all empty. Other link script directives that allocate
5280 space in an output section will also create the output section. So
5281 too will assignments to dot even if the assignment does not create
5282 space, except for @samp{. = 0}, @samp{. = . + 0}, @samp{. = sym},
5283 @samp{. = . + sym} and @samp{. = ALIGN (. != 0, expr, 1)} when
5284 @samp{sym} is an absolute symbol of value 0 defined in the script.
5285 This allows you to force output of an empty section with @samp{. = .}.
5287 The linker will ignore address assignments (@pxref{Output Section Address})
5288 on discarded output sections, except when the linker script defines
5289 symbols in the output section. In that case the linker will obey
5290 the address assignments, possibly advancing dot even though the
5291 section is discarded.
5294 The special output section name @samp{/DISCARD/} may be used to discard
5295 input sections. Any input sections which are assigned to an output
5296 section named @samp{/DISCARD/} are not included in the output file.
5298 This can be used to discard input sections marked with the ELF flag
5299 @code{SHF_GNU_RETAIN}, which would otherwise have been saved from linker
5302 Note, sections that match the @samp{/DISCARD/} output section will be
5303 discarded even if they are in an ELF section group which has other
5304 members which are not being discarded. This is deliberate.
5305 Discarding takes precedence over grouping.
5307 @node Output Section Attributes
5308 @subsection Output Section Attributes
5309 @cindex output section attributes
5310 We showed above that the full description of an output section looked
5315 @var{section} [@var{address}] [(@var{type})] :
5317 [ALIGN(@var{section_align}) | ALIGN_WITH_INPUT]
5318 [SUBALIGN(@var{subsection_align})]
5321 @var{output-section-command}
5322 @var{output-section-command}
5324 @} [>@var{region}] [AT>@var{lma_region}] [:@var{phdr} :@var{phdr} @dots{}] [=@var{fillexp}]
5328 We've already described @var{section}, @var{address}, and
5329 @var{output-section-command}. In this section we will describe the
5330 remaining section attributes.
5333 * Output Section Type:: Output section type
5334 * Output Section LMA:: Output section LMA
5335 * Forced Output Alignment:: Forced Output Alignment
5336 * Forced Input Alignment:: Forced Input Alignment
5337 * Output Section Constraint:: Output section constraint
5338 * Output Section Region:: Output section region
5339 * Output Section Phdr:: Output section phdr
5340 * Output Section Fill:: Output section fill
5343 @node Output Section Type
5344 @subsubsection Output Section Type
5345 Each output section may have a type. The type is a keyword in
5346 parentheses. The following types are defined:
5350 The section should be marked as not loadable, so that it will not be
5351 loaded into memory when the program is run.
5356 These type names are supported for backward compatibility, and are
5357 rarely used. They all have the same effect: the section should be
5358 marked as not allocatable, so that no memory is allocated for the
5359 section when the program is run.
5363 @cindex prevent unnecessary loading
5364 @cindex loading, preventing
5365 The linker normally sets the attributes of an output section based on
5366 the input sections which map into it. You can override this by using
5367 the section type. For example, in the script sample below, the
5368 @samp{ROM} section is addressed at memory location @samp{0} and does not
5369 need to be loaded when the program is run.
5373 ROM 0 (NOLOAD) : @{ @dots{} @}
5379 @node Output Section LMA
5380 @subsubsection Output Section LMA
5381 @kindex AT>@var{lma_region}
5382 @kindex AT(@var{lma})
5383 @cindex load address
5384 @cindex section load address
5385 Every section has a virtual address (VMA) and a load address (LMA); see
5386 @ref{Basic Script Concepts}. The virtual address is specified by the
5387 @pxref{Output Section Address} described earlier. The load address is
5388 specified by the @code{AT} or @code{AT>} keywords. Specifying a load
5389 address is optional.
5391 The @code{AT} keyword takes an expression as an argument. This
5392 specifies the exact load address of the section. The @code{AT>} keyword
5393 takes the name of a memory region as an argument. @xref{MEMORY}. The
5394 load address of the section is set to the next free address in the
5395 region, aligned to the section's alignment requirements.
5397 If neither @code{AT} nor @code{AT>} is specified for an allocatable
5398 section, the linker will use the following heuristic to determine the
5403 If the section has a specific VMA address, then this is used as
5404 the LMA address as well.
5407 If the section is not allocatable then its LMA is set to its VMA.
5410 Otherwise if a memory region can be found that is compatible
5411 with the current section, and this region contains at least one
5412 section, then the LMA is set so the difference between the
5413 VMA and LMA is the same as the difference between the VMA and LMA of
5414 the last section in the located region.
5417 If no memory regions have been declared then a default region
5418 that covers the entire address space is used in the previous step.
5421 If no suitable region could be found, or there was no previous
5422 section then the LMA is set equal to the VMA.
5425 @cindex ROM initialized data
5426 @cindex initialized data in ROM
5427 This feature is designed to make it easy to build a ROM image. For
5428 example, the following linker script creates three output sections: one
5429 called @samp{.text}, which starts at @code{0x1000}, one called
5430 @samp{.mdata}, which is loaded at the end of the @samp{.text} section
5431 even though its VMA is @code{0x2000}, and one called @samp{.bss} to hold
5432 uninitialized data at address @code{0x3000}. The symbol @code{_data} is
5433 defined with the value @code{0x2000}, which shows that the location
5434 counter holds the VMA value, not the LMA value.
5440 .text 0x1000 : @{ *(.text) _etext = . ; @}
5442 AT ( ADDR (.text) + SIZEOF (.text) )
5443 @{ _data = . ; *(.data); _edata = . ; @}
5445 @{ _bstart = . ; *(.bss) *(COMMON) ; _bend = . ;@}
5450 The run-time initialization code for use with a program generated with
5451 this linker script would include something like the following, to copy
5452 the initialized data from the ROM image to its runtime address. Notice
5453 how this code takes advantage of the symbols defined by the linker
5458 extern char _etext, _data, _edata, _bstart, _bend;
5459 char *src = &_etext;
5462 /* ROM has data at end of text; copy it. */
5463 while (dst < &_edata)
5467 for (dst = &_bstart; dst< &_bend; dst++)
5472 @node Forced Output Alignment
5473 @subsubsection Forced Output Alignment
5474 @kindex ALIGN(@var{section_align})
5475 @cindex forcing output section alignment
5476 @cindex output section alignment
5477 You can increase an output section's alignment by using ALIGN. As an
5478 alternative you can enforce that the difference between the VMA and LMA remains
5479 intact throughout this output section with the ALIGN_WITH_INPUT attribute.
5481 @node Forced Input Alignment
5482 @subsubsection Forced Input Alignment
5483 @kindex SUBALIGN(@var{subsection_align})
5484 @cindex forcing input section alignment
5485 @cindex input section alignment
5486 You can force input section alignment within an output section by using
5487 SUBALIGN. The value specified overrides any alignment given by input
5488 sections, whether larger or smaller.
5490 @node Output Section Constraint
5491 @subsubsection Output Section Constraint
5494 @cindex constraints on output sections
5495 You can specify that an output section should only be created if all
5496 of its input sections are read-only or all of its input sections are
5497 read-write by using the keyword @code{ONLY_IF_RO} and
5498 @code{ONLY_IF_RW} respectively.
5500 @node Output Section Region
5501 @subsubsection Output Section Region
5502 @kindex >@var{region}
5503 @cindex section, assigning to memory region
5504 @cindex memory regions and sections
5505 You can assign a section to a previously defined region of memory by
5506 using @samp{>@var{region}}. @xref{MEMORY}.
5508 Here is a simple example:
5511 MEMORY @{ rom : ORIGIN = 0x1000, LENGTH = 0x1000 @}
5512 SECTIONS @{ ROM : @{ *(.text) @} >rom @}
5516 @node Output Section Phdr
5517 @subsubsection Output Section Phdr
5519 @cindex section, assigning to program header
5520 @cindex program headers and sections
5521 You can assign a section to a previously defined program segment by
5522 using @samp{:@var{phdr}}. @xref{PHDRS}. If a section is assigned to
5523 one or more segments, then all subsequent allocated sections will be
5524 assigned to those segments as well, unless they use an explicitly
5525 @code{:@var{phdr}} modifier. You can use @code{:NONE} to tell the
5526 linker to not put the section in any segment at all.
5528 Here is a simple example:
5531 PHDRS @{ text PT_LOAD ; @}
5532 SECTIONS @{ .text : @{ *(.text) @} :text @}
5536 @node Output Section Fill
5537 @subsubsection Output Section Fill
5538 @kindex =@var{fillexp}
5539 @cindex section fill pattern
5540 @cindex fill pattern, entire section
5541 You can set the fill pattern for an entire section by using
5542 @samp{=@var{fillexp}}. @var{fillexp} is an expression
5543 (@pxref{Expressions}). Any otherwise unspecified regions of memory
5544 within the output section (for example, gaps left due to the required
5545 alignment of input sections) will be filled with the value, repeated as
5546 necessary. If the fill expression is a simple hex number, ie. a string
5547 of hex digit starting with @samp{0x} and without a trailing @samp{k} or @samp{M}, then
5548 an arbitrarily long sequence of hex digits can be used to specify the
5549 fill pattern; Leading zeros become part of the pattern too. For all
5550 other cases, including extra parentheses or a unary @code{+}, the fill
5551 pattern is the four least significant bytes of the value of the
5552 expression. In all cases, the number is big-endian.
5554 You can also change the fill value with a @code{FILL} command in the
5555 output section commands; (@pxref{Output Section Data}).
5557 Here is a simple example:
5560 SECTIONS @{ .text : @{ *(.text) @} =0x90909090 @}
5564 @node Overlay Description
5565 @subsection Overlay Description
5568 An overlay description provides an easy way to describe sections which
5569 are to be loaded as part of a single memory image but are to be run at
5570 the same memory address. At run time, some sort of overlay manager will
5571 copy the overlaid sections in and out of the runtime memory address as
5572 required, perhaps by simply manipulating addressing bits. This approach
5573 can be useful, for example, when a certain region of memory is faster
5576 Overlays are described using the @code{OVERLAY} command. The
5577 @code{OVERLAY} command is used within a @code{SECTIONS} command, like an
5578 output section description. The full syntax of the @code{OVERLAY}
5579 command is as follows:
5582 OVERLAY [@var{start}] : [NOCROSSREFS] [AT ( @var{ldaddr} )]
5586 @var{output-section-command}
5587 @var{output-section-command}
5589 @} [:@var{phdr}@dots{}] [=@var{fill}]
5592 @var{output-section-command}
5593 @var{output-section-command}
5595 @} [:@var{phdr}@dots{}] [=@var{fill}]
5597 @} [>@var{region}] [:@var{phdr}@dots{}] [=@var{fill}] [,]
5601 Everything is optional except @code{OVERLAY} (a keyword), and each
5602 section must have a name (@var{secname1} and @var{secname2} above). The
5603 section definitions within the @code{OVERLAY} construct are identical to
5604 those within the general @code{SECTIONS} construct (@pxref{SECTIONS}),
5605 except that no addresses and no memory regions may be defined for
5606 sections within an @code{OVERLAY}.
5608 The comma at the end may be required if a @var{fill} is used and
5609 the next @var{sections-command} looks like a continuation of the expression.
5611 The sections are all defined with the same starting address. The load
5612 addresses of the sections are arranged such that they are consecutive in
5613 memory starting at the load address used for the @code{OVERLAY} as a
5614 whole (as with normal section definitions, the load address is optional,
5615 and defaults to the start address; the start address is also optional,
5616 and defaults to the current value of the location counter).
5618 If the @code{NOCROSSREFS} keyword is used, and there are any
5619 references among the sections, the linker will report an error. Since
5620 the sections all run at the same address, it normally does not make
5621 sense for one section to refer directly to another.
5622 @xref{Miscellaneous Commands, NOCROSSREFS}.
5624 For each section within the @code{OVERLAY}, the linker automatically
5625 provides two symbols. The symbol @code{__load_start_@var{secname}} is
5626 defined as the starting load address of the section. The symbol
5627 @code{__load_stop_@var{secname}} is defined as the final load address of
5628 the section. Any characters within @var{secname} which are not legal
5629 within C identifiers are removed. C (or assembler) code may use these
5630 symbols to move the overlaid sections around as necessary.
5632 At the end of the overlay, the value of the location counter is set to
5633 the start address of the overlay plus the size of the largest section.
5635 Here is an example. Remember that this would appear inside a
5636 @code{SECTIONS} construct.
5639 OVERLAY 0x1000 : AT (0x4000)
5641 .text0 @{ o1/*.o(.text) @}
5642 .text1 @{ o2/*.o(.text) @}
5647 This will define both @samp{.text0} and @samp{.text1} to start at
5648 address 0x1000. @samp{.text0} will be loaded at address 0x4000, and
5649 @samp{.text1} will be loaded immediately after @samp{.text0}. The
5650 following symbols will be defined if referenced: @code{__load_start_text0},
5651 @code{__load_stop_text0}, @code{__load_start_text1},
5652 @code{__load_stop_text1}.
5654 C code to copy overlay @code{.text1} into the overlay area might look
5659 extern char __load_start_text1, __load_stop_text1;
5660 memcpy ((char *) 0x1000, &__load_start_text1,
5661 &__load_stop_text1 - &__load_start_text1);
5665 Note that the @code{OVERLAY} command is just syntactic sugar, since
5666 everything it does can be done using the more basic commands. The above
5667 example could have been written identically as follows.
5671 .text0 0x1000 : AT (0x4000) @{ o1/*.o(.text) @}
5672 PROVIDE (__load_start_text0 = LOADADDR (.text0));
5673 PROVIDE (__load_stop_text0 = LOADADDR (.text0) + SIZEOF (.text0));
5674 .text1 0x1000 : AT (0x4000 + SIZEOF (.text0)) @{ o2/*.o(.text) @}
5675 PROVIDE (__load_start_text1 = LOADADDR (.text1));
5676 PROVIDE (__load_stop_text1 = LOADADDR (.text1) + SIZEOF (.text1));
5677 . = 0x1000 + MAX (SIZEOF (.text0), SIZEOF (.text1));
5682 @section MEMORY Command
5684 @cindex memory regions
5685 @cindex regions of memory
5686 @cindex allocating memory
5687 @cindex discontinuous memory
5688 The linker's default configuration permits allocation of all available
5689 memory. You can override this by using the @code{MEMORY} command.
5691 The @code{MEMORY} command describes the location and size of blocks of
5692 memory in the target. You can use it to describe which memory regions
5693 may be used by the linker, and which memory regions it must avoid. You
5694 can then assign sections to particular memory regions. The linker will
5695 set section addresses based on the memory regions, and will warn about
5696 regions that become too full. The linker will not shuffle sections
5697 around to fit into the available regions.
5699 A linker script may contain many uses of the @code{MEMORY} command,
5700 however, all memory blocks defined are treated as if they were
5701 specified inside a single @code{MEMORY} command. The syntax for
5707 @var{name} [(@var{attr})] : ORIGIN = @var{origin}, LENGTH = @var{len}
5713 The @var{name} is a name used in the linker script to refer to the
5714 region. The region name has no meaning outside of the linker script.
5715 Region names are stored in a separate name space, and will not conflict
5716 with symbol names, file names, or section names. Each memory region
5717 must have a distinct name within the @code{MEMORY} command. However you can
5718 add later alias names to existing memory regions with the @ref{REGION_ALIAS}
5721 @cindex memory region attributes
5722 The @var{attr} string is an optional list of attributes that specify
5723 whether to use a particular memory region for an input section which is
5724 not explicitly mapped in the linker script. As described in
5725 @ref{SECTIONS}, if you do not specify an output section for some input
5726 section, the linker will create an output section with the same name as
5727 the input section. If you define region attributes, the linker will use
5728 them to select the memory region for the output section that it creates.
5730 The @var{attr} string must consist only of the following characters:
5745 Invert the sense of any of the attributes that follow
5748 If an unmapped section matches any of the listed attributes other than
5749 @samp{!}, it will be placed in the memory region. The @samp{!}
5750 attribute reverses the test for the characters that follow, so that an
5751 unmapped section will be placed in the memory region only if it does
5752 not match any of the attributes listed afterwards. Thus an attribute
5753 string of @samp{RW!X} will match any unmapped section that has either
5754 or both of the @samp{R} and @samp{W} attributes, but only as long as
5755 the section does not also have the @samp{X} attribute.
5760 The @var{origin} is an numerical expression for the start address of
5761 the memory region. The expression must evaluate to a constant and it
5762 cannot involve any symbols. The keyword @code{ORIGIN} may be
5763 abbreviated to @code{org} or @code{o} (but not, for example,
5769 The @var{len} is an expression for the size in bytes of the memory
5770 region. As with the @var{origin} expression, the expression must
5771 be numerical only and must evaluate to a constant. The keyword
5772 @code{LENGTH} may be abbreviated to @code{len} or @code{l}.
5774 In the following example, we specify that there are two memory regions
5775 available for allocation: one starting at @samp{0} for 256 kilobytes,
5776 and the other starting at @samp{0x40000000} for four megabytes. The
5777 linker will place into the @samp{rom} memory region every section which
5778 is not explicitly mapped into a memory region, and is either read-only
5779 or executable. The linker will place other sections which are not
5780 explicitly mapped into a memory region into the @samp{ram} memory
5787 rom (rx) : ORIGIN = 0, LENGTH = 256K
5788 ram (!rx) : org = 0x40000000, l = 4M
5793 Once you define a memory region, you can direct the linker to place
5794 specific output sections into that memory region by using the
5795 @samp{>@var{region}} output section attribute. For example, if you have
5796 a memory region named @samp{mem}, you would use @samp{>mem} in the
5797 output section definition. @xref{Output Section Region}. If no address
5798 was specified for the output section, the linker will set the address to
5799 the next available address within the memory region. If the combined
5800 output sections directed to a memory region are too large for the
5801 region, the linker will issue an error message.
5803 It is possible to access the origin and length of a memory in an
5804 expression via the @code{ORIGIN(@var{memory})} and
5805 @code{LENGTH(@var{memory})} functions:
5809 _fstack = ORIGIN(ram) + LENGTH(ram) - 4;
5814 @section PHDRS Command
5816 @cindex program headers
5817 @cindex ELF program headers
5818 @cindex program segments
5819 @cindex segments, ELF
5820 The ELF object file format uses @dfn{program headers}, also knows as
5821 @dfn{segments}. The program headers describe how the program should be
5822 loaded into memory. You can print them out by using the @code{objdump}
5823 program with the @samp{-p} option.
5825 When you run an ELF program on a native ELF system, the system loader
5826 reads the program headers in order to figure out how to load the
5827 program. This will only work if the program headers are set correctly.
5828 This manual does not describe the details of how the system loader
5829 interprets program headers; for more information, see the ELF ABI.
5831 The linker will create reasonable program headers by default. However,
5832 in some cases, you may need to specify the program headers more
5833 precisely. You may use the @code{PHDRS} command for this purpose. When
5834 the linker sees the @code{PHDRS} command in the linker script, it will
5835 not create any program headers other than the ones specified.
5837 The linker only pays attention to the @code{PHDRS} command when
5838 generating an ELF output file. In other cases, the linker will simply
5839 ignore @code{PHDRS}.
5841 This is the syntax of the @code{PHDRS} command. The words @code{PHDRS},
5842 @code{FILEHDR}, @code{AT}, and @code{FLAGS} are keywords.
5848 @var{name} @var{type} [ FILEHDR ] [ PHDRS ] [ AT ( @var{address} ) ]
5849 [ FLAGS ( @var{flags} ) ] ;
5854 The @var{name} is used only for reference in the @code{SECTIONS} command
5855 of the linker script. It is not put into the output file. Program
5856 header names are stored in a separate name space, and will not conflict
5857 with symbol names, file names, or section names. Each program header
5858 must have a distinct name. The headers are processed in order and it
5859 is usual for them to map to sections in ascending load address order.
5861 Certain program header types describe segments of memory which the
5862 system loader will load from the file. In the linker script, you
5863 specify the contents of these segments by placing allocatable output
5864 sections in the segments. You use the @samp{:@var{phdr}} output section
5865 attribute to place a section in a particular segment. @xref{Output
5868 It is normal to put certain sections in more than one segment. This
5869 merely implies that one segment of memory contains another. You may
5870 repeat @samp{:@var{phdr}}, using it once for each segment which should
5871 contain the section.
5873 If you place a section in one or more segments using @samp{:@var{phdr}},
5874 then the linker will place all subsequent allocatable sections which do
5875 not specify @samp{:@var{phdr}} in the same segments. This is for
5876 convenience, since generally a whole set of contiguous sections will be
5877 placed in a single segment. You can use @code{:NONE} to override the
5878 default segment and tell the linker to not put the section in any
5883 You may use the @code{FILEHDR} and @code{PHDRS} keywords after
5884 the program header type to further describe the contents of the segment.
5885 The @code{FILEHDR} keyword means that the segment should include the ELF
5886 file header. The @code{PHDRS} keyword means that the segment should
5887 include the ELF program headers themselves. If applied to a loadable
5888 segment (@code{PT_LOAD}), all prior loadable segments must have one of
5891 The @var{type} may be one of the following. The numbers indicate the
5892 value of the keyword.
5895 @item @code{PT_NULL} (0)
5896 Indicates an unused program header.
5898 @item @code{PT_LOAD} (1)
5899 Indicates that this program header describes a segment to be loaded from
5902 @item @code{PT_DYNAMIC} (2)
5903 Indicates a segment where dynamic linking information can be found.
5905 @item @code{PT_INTERP} (3)
5906 Indicates a segment where the name of the program interpreter may be
5909 @item @code{PT_NOTE} (4)
5910 Indicates a segment holding note information.
5912 @item @code{PT_SHLIB} (5)
5913 A reserved program header type, defined but not specified by the ELF
5916 @item @code{PT_PHDR} (6)
5917 Indicates a segment where the program headers may be found.
5919 @item @code{PT_TLS} (7)
5920 Indicates a segment containing thread local storage.
5922 @item @var{expression}
5923 An expression giving the numeric type of the program header. This may
5924 be used for types not defined above.
5927 You can specify that a segment should be loaded at a particular address
5928 in memory by using an @code{AT} expression. This is identical to the
5929 @code{AT} command used as an output section attribute (@pxref{Output
5930 Section LMA}). The @code{AT} command for a program header overrides the
5931 output section attribute.
5933 The linker will normally set the segment flags based on the sections
5934 which comprise the segment. You may use the @code{FLAGS} keyword to
5935 explicitly specify the segment flags. The value of @var{flags} must be
5936 an integer. It is used to set the @code{p_flags} field of the program
5939 Here is an example of @code{PHDRS}. This shows a typical set of program
5940 headers used on a native ELF system.
5946 headers PT_PHDR PHDRS ;
5948 text PT_LOAD FILEHDR PHDRS ;
5950 dynamic PT_DYNAMIC ;
5956 .interp : @{ *(.interp) @} :text :interp
5957 .text : @{ *(.text) @} :text
5958 .rodata : @{ *(.rodata) @} /* defaults to :text */
5960 . = . + 0x1000; /* move to a new page in memory */
5961 .data : @{ *(.data) @} :data
5962 .dynamic : @{ *(.dynamic) @} :data :dynamic
5969 @section VERSION Command
5970 @kindex VERSION @{script text@}
5971 @cindex symbol versions
5972 @cindex version script
5973 @cindex versions of symbols
5974 The linker supports symbol versions when using ELF. Symbol versions are
5975 only useful when using shared libraries. The dynamic linker can use
5976 symbol versions to select a specific version of a function when it runs
5977 a program that may have been linked against an earlier version of the
5980 You can include a version script directly in the main linker script, or
5981 you can supply the version script as an implicit linker script. You can
5982 also use the @samp{--version-script} linker option.
5984 The syntax of the @code{VERSION} command is simply
5986 VERSION @{ version-script-commands @}
5989 The format of the version script commands is identical to that used by
5990 Sun's linker in Solaris 2.5. The version script defines a tree of
5991 version nodes. You specify the node names and interdependencies in the
5992 version script. You can specify which symbols are bound to which
5993 version nodes, and you can reduce a specified set of symbols to local
5994 scope so that they are not globally visible outside of the shared
5997 The easiest way to demonstrate the version script language is with a few
6023 This example version script defines three version nodes. The first
6024 version node defined is @samp{VERS_1.1}; it has no other dependencies.
6025 The script binds the symbol @samp{foo1} to @samp{VERS_1.1}. It reduces
6026 a number of symbols to local scope so that they are not visible outside
6027 of the shared library; this is done using wildcard patterns, so that any
6028 symbol whose name begins with @samp{old}, @samp{original}, or @samp{new}
6029 is matched. The wildcard patterns available are the same as those used
6030 in the shell when matching filenames (also known as ``globbing'').
6031 However, if you specify the symbol name inside double quotes, then the
6032 name is treated as literal, rather than as a glob pattern.
6034 Next, the version script defines node @samp{VERS_1.2}. This node
6035 depends upon @samp{VERS_1.1}. The script binds the symbol @samp{foo2}
6036 to the version node @samp{VERS_1.2}.
6038 Finally, the version script defines node @samp{VERS_2.0}. This node
6039 depends upon @samp{VERS_1.2}. The scripts binds the symbols @samp{bar1}
6040 and @samp{bar2} are bound to the version node @samp{VERS_2.0}.
6042 When the linker finds a symbol defined in a library which is not
6043 specifically bound to a version node, it will effectively bind it to an
6044 unspecified base version of the library. You can bind all otherwise
6045 unspecified symbols to a given version node by using @samp{global: *;}
6046 somewhere in the version script. Note that it's slightly crazy to use
6047 wildcards in a global spec except on the last version node. Global
6048 wildcards elsewhere run the risk of accidentally adding symbols to the
6049 set exported for an old version. That's wrong since older versions
6050 ought to have a fixed set of symbols.
6052 The names of the version nodes have no specific meaning other than what
6053 they might suggest to the person reading them. The @samp{2.0} version
6054 could just as well have appeared in between @samp{1.1} and @samp{1.2}.
6055 However, this would be a confusing way to write a version script.
6057 Node name can be omitted, provided it is the only version node
6058 in the version script. Such version script doesn't assign any versions to
6059 symbols, only selects which symbols will be globally visible out and which
6063 @{ global: foo; bar; local: *; @};
6066 When you link an application against a shared library that has versioned
6067 symbols, the application itself knows which version of each symbol it
6068 requires, and it also knows which version nodes it needs from each
6069 shared library it is linked against. Thus at runtime, the dynamic
6070 loader can make a quick check to make sure that the libraries you have
6071 linked against do in fact supply all of the version nodes that the
6072 application will need to resolve all of the dynamic symbols. In this
6073 way it is possible for the dynamic linker to know with certainty that
6074 all external symbols that it needs will be resolvable without having to
6075 search for each symbol reference.
6077 The symbol versioning is in effect a much more sophisticated way of
6078 doing minor version checking that SunOS does. The fundamental problem
6079 that is being addressed here is that typically references to external
6080 functions are bound on an as-needed basis, and are not all bound when
6081 the application starts up. If a shared library is out of date, a
6082 required interface may be missing; when the application tries to use
6083 that interface, it may suddenly and unexpectedly fail. With symbol
6084 versioning, the user will get a warning when they start their program if
6085 the libraries being used with the application are too old.
6087 There are several GNU extensions to Sun's versioning approach. The
6088 first of these is the ability to bind a symbol to a version node in the
6089 source file where the symbol is defined instead of in the versioning
6090 script. This was done mainly to reduce the burden on the library
6091 maintainer. You can do this by putting something like:
6093 __asm__(".symver original_foo,foo@@VERS_1.1");
6096 in the C source file. This renames the function @samp{original_foo} to
6097 be an alias for @samp{foo} bound to the version node @samp{VERS_1.1}.
6098 The @samp{local:} directive can be used to prevent the symbol
6099 @samp{original_foo} from being exported. A @samp{.symver} directive
6100 takes precedence over a version script.
6102 The second GNU extension is to allow multiple versions of the same
6103 function to appear in a given shared library. In this way you can make
6104 an incompatible change to an interface without increasing the major
6105 version number of the shared library, while still allowing applications
6106 linked against the old interface to continue to function.
6108 To do this, you must use multiple @samp{.symver} directives in the
6109 source file. Here is an example:
6112 __asm__(".symver original_foo,foo@@");
6113 __asm__(".symver old_foo,foo@@VERS_1.1");
6114 __asm__(".symver old_foo1,foo@@VERS_1.2");
6115 __asm__(".symver new_foo,foo@@@@VERS_2.0");
6118 In this example, @samp{foo@@} represents the symbol @samp{foo} bound to the
6119 unspecified base version of the symbol. The source file that contains this
6120 example would define 4 C functions: @samp{original_foo}, @samp{old_foo},
6121 @samp{old_foo1}, and @samp{new_foo}.
6123 When you have multiple definitions of a given symbol, there needs to be
6124 some way to specify a default version to which external references to
6125 this symbol will be bound. You can do this with the
6126 @samp{foo@@@@VERS_2.0} type of @samp{.symver} directive. You can only
6127 declare one version of a symbol as the default in this manner; otherwise
6128 you would effectively have multiple definitions of the same symbol.
6130 If you wish to bind a reference to a specific version of the symbol
6131 within the shared library, you can use the aliases of convenience
6132 (i.e., @samp{old_foo}), or you can use the @samp{.symver} directive to
6133 specifically bind to an external version of the function in question.
6135 You can also specify the language in the version script:
6138 VERSION extern "lang" @{ version-script-commands @}
6141 The supported @samp{lang}s are @samp{C}, @samp{C++}, and @samp{Java}.
6142 The linker will iterate over the list of symbols at the link time and
6143 demangle them according to @samp{lang} before matching them to the
6144 patterns specified in @samp{version-script-commands}. The default
6145 @samp{lang} is @samp{C}.
6147 Demangled names may contains spaces and other special characters. As
6148 described above, you can use a glob pattern to match demangled names,
6149 or you can use a double-quoted string to match the string exactly. In
6150 the latter case, be aware that minor differences (such as differing
6151 whitespace) between the version script and the demangler output will
6152 cause a mismatch. As the exact string generated by the demangler
6153 might change in the future, even if the mangled name does not, you
6154 should check that all of your version directives are behaving as you
6155 expect when you upgrade.
6158 @section Expressions in Linker Scripts
6161 The syntax for expressions in the linker script language is identical to
6162 that of C expressions. All expressions are evaluated as integers. All
6163 expressions are evaluated in the same size, which is 32 bits if both the
6164 host and target are 32 bits, and is otherwise 64 bits.
6166 You can use and set symbol values in expressions.
6168 The linker defines several special purpose builtin functions for use in
6172 * Constants:: Constants
6173 * Symbolic Constants:: Symbolic constants
6174 * Symbols:: Symbol Names
6175 * Orphan Sections:: Orphan Sections
6176 * Location Counter:: The Location Counter
6177 * Operators:: Operators
6178 * Evaluation:: Evaluation
6179 * Expression Section:: The Section of an Expression
6180 * Builtin Functions:: Builtin Functions
6184 @subsection Constants
6185 @cindex integer notation
6186 @cindex constants in linker scripts
6187 All constants are integers.
6189 As in C, the linker considers an integer beginning with @samp{0} to be
6190 octal, and an integer beginning with @samp{0x} or @samp{0X} to be
6191 hexadecimal. Alternatively the linker accepts suffixes of @samp{h} or
6192 @samp{H} for hexadecimal, @samp{o} or @samp{O} for octal, @samp{b} or
6193 @samp{B} for binary and @samp{d} or @samp{D} for decimal. Any integer
6194 value without a prefix or a suffix is considered to be decimal.
6196 @cindex scaled integers
6197 @cindex K and M integer suffixes
6198 @cindex M and K integer suffixes
6199 @cindex suffixes for integers
6200 @cindex integer suffixes
6201 In addition, you can use the suffixes @code{K} and @code{M} to scale a
6205 @c END TEXI2ROFF-KILL
6206 @code{1024} or @code{1024*1024}
6210 ${\rm 1024}$ or ${\rm 1024}^2$
6212 @c END TEXI2ROFF-KILL
6213 respectively. For example, the following
6214 all refer to the same quantity:
6223 Note - the @code{K} and @code{M} suffixes cannot be used in
6224 conjunction with the base suffixes mentioned above.
6226 @node Symbolic Constants
6227 @subsection Symbolic Constants
6228 @cindex symbolic constants
6230 It is possible to refer to target-specific constants via the use of
6231 the @code{CONSTANT(@var{name})} operator, where @var{name} is one of:
6236 The target's maximum page size.
6238 @item COMMONPAGESIZE
6239 @kindex COMMONPAGESIZE
6240 The target's default page size.
6246 .text ALIGN (CONSTANT (MAXPAGESIZE)) : @{ *(.text) @}
6249 will create a text section aligned to the largest page boundary
6250 supported by the target.
6253 @subsection Symbol Names
6254 @cindex symbol names
6256 @cindex quoted symbol names
6258 Unless quoted, symbol names start with a letter, underscore, or period
6259 and may include letters, digits, underscores, periods, and hyphens.
6260 Unquoted symbol names must not conflict with any keywords. You can
6261 specify a symbol which contains odd characters or has the same name as a
6262 keyword by surrounding the symbol name in double quotes:
6265 "with a space" = "also with a space" + 10;
6268 Since symbols can contain many non-alphabetic characters, it is safest
6269 to delimit symbols with spaces. For example, @samp{A-B} is one symbol,
6270 whereas @samp{A - B} is an expression involving subtraction.
6272 @node Orphan Sections
6273 @subsection Orphan Sections
6275 Orphan sections are sections present in the input files which
6276 are not explicitly placed into the output file by the linker
6277 script. The linker will still copy these sections into the
6278 output file by either finding, or creating a suitable output section
6279 in which to place the orphaned input section.
6281 If the name of an orphaned input section exactly matches the name of
6282 an existing output section, then the orphaned input section will be
6283 placed at the end of that output section.
6285 If there is no output section with a matching name then new output
6286 sections will be created. Each new output section will have the same
6287 name as the orphan section placed within it. If there are multiple
6288 orphan sections with the same name, these will all be combined into
6289 one new output section.
6291 If new output sections are created to hold orphaned input sections,
6292 then the linker must decide where to place these new output sections
6293 in relation to existing output sections. On most modern targets, the
6294 linker attempts to place orphan sections after sections of the same
6295 attribute, such as code vs data, loadable vs non-loadable, etc. If no
6296 sections with matching attributes are found, or your target lacks this
6297 support, the orphan section is placed at the end of the file.
6299 The command-line options @samp{--orphan-handling} and @samp{--unique}
6300 (@pxref{Options,,Command-line Options}) can be used to control which
6301 output sections an orphan is placed in.
6303 @node Location Counter
6304 @subsection The Location Counter
6307 @cindex location counter
6308 @cindex current output location
6309 The special linker variable @dfn{dot} @samp{.} always contains the
6310 current output location counter. Since the @code{.} always refers to a
6311 location in an output section, it may only appear in an expression
6312 within a @code{SECTIONS} command. The @code{.} symbol may appear
6313 anywhere that an ordinary symbol is allowed in an expression.
6316 Assigning a value to @code{.} will cause the location counter to be
6317 moved. This may be used to create holes in the output section. The
6318 location counter may not be moved backwards inside an output section,
6319 and may not be moved backwards outside of an output section if so
6320 doing creates areas with overlapping LMAs.
6336 In the previous example, the @samp{.text} section from @file{file1} is
6337 located at the beginning of the output section @samp{output}. It is
6338 followed by a 1000 byte gap. Then the @samp{.text} section from
6339 @file{file2} appears, also with a 1000 byte gap following before the
6340 @samp{.text} section from @file{file3}. The notation @samp{= 0x12345678}
6341 specifies what data to write in the gaps (@pxref{Output Section Fill}).
6343 @cindex dot inside sections
6344 Note: @code{.} actually refers to the byte offset from the start of the
6345 current containing object. Normally this is the @code{SECTIONS}
6346 statement, whose start address is 0, hence @code{.} can be used as an
6347 absolute address. If @code{.} is used inside a section description
6348 however, it refers to the byte offset from the start of that section,
6349 not an absolute address. Thus in a script like this:
6367 The @samp{.text} section will be assigned a starting address of 0x100
6368 and a size of exactly 0x200 bytes, even if there is not enough data in
6369 the @samp{.text} input sections to fill this area. (If there is too
6370 much data, an error will be produced because this would be an attempt to
6371 move @code{.} backwards). The @samp{.data} section will start at 0x500
6372 and it will have an extra 0x600 bytes worth of space after the end of
6373 the values from the @samp{.data} input sections and before the end of
6374 the @samp{.data} output section itself.
6376 @cindex dot outside sections
6377 Setting symbols to the value of the location counter outside of an
6378 output section statement can result in unexpected values if the linker
6379 needs to place orphan sections. For example, given the following:
6385 .text: @{ *(.text) @}
6389 .data: @{ *(.data) @}
6394 If the linker needs to place some input section, e.g. @code{.rodata},
6395 not mentioned in the script, it might choose to place that section
6396 between @code{.text} and @code{.data}. You might think the linker
6397 should place @code{.rodata} on the blank line in the above script, but
6398 blank lines are of no particular significance to the linker. As well,
6399 the linker doesn't associate the above symbol names with their
6400 sections. Instead, it assumes that all assignments or other
6401 statements belong to the previous output section, except for the
6402 special case of an assignment to @code{.}. I.e., the linker will
6403 place the orphan @code{.rodata} section as if the script was written
6410 .text: @{ *(.text) @}
6414 .rodata: @{ *(.rodata) @}
6415 .data: @{ *(.data) @}
6420 This may or may not be the script author's intention for the value of
6421 @code{start_of_data}. One way to influence the orphan section
6422 placement is to assign the location counter to itself, as the linker
6423 assumes that an assignment to @code{.} is setting the start address of
6424 a following output section and thus should be grouped with that
6425 section. So you could write:
6431 .text: @{ *(.text) @}
6436 .data: @{ *(.data) @}
6441 Now, the orphan @code{.rodata} section will be placed between
6442 @code{end_of_text} and @code{start_of_data}.
6446 @subsection Operators
6447 @cindex operators for arithmetic
6448 @cindex arithmetic operators
6449 @cindex precedence in expressions
6450 The linker recognizes the standard C set of arithmetic operators, with
6451 the standard bindings and precedence levels:
6454 @c END TEXI2ROFF-KILL
6456 precedence associativity Operators Notes
6462 5 left == != > < <= >=
6468 11 right &= += -= *= /= (2)
6472 (1) Prefix operators
6473 (2) @xref{Assignments}.
6477 \vskip \baselineskip
6478 %"lispnarrowing" is the extra indent used generally for smallexample
6479 \hskip\lispnarrowing\vbox{\offinterlineskip
6482 {\vrule#&\strut\hfil\ #\ \hfil&\vrule#&\strut\hfil\ #\ \hfil&\vrule#&\strut\hfil\ {\tt #}\ \hfil&\vrule#\cr
6483 height2pt&\omit&&\omit&&\omit&\cr
6484 &Precedence&& Associativity &&{\rm Operators}&\cr
6485 height2pt&\omit&&\omit&&\omit&\cr
6487 height2pt&\omit&&\omit&&\omit&\cr
6489 % '176 is tilde, '~' in tt font
6490 &1&&left&&\qquad- \char'176\ !\qquad\dag&\cr
6491 &2&&left&&* / \%&\cr
6494 &5&&left&&== != > < <= >=&\cr
6497 &8&&left&&{\&\&}&\cr
6500 &11&&right&&\qquad\&= += -= *= /=\qquad\ddag&\cr
6502 height2pt&\omit&&\omit&&\omit&\cr}
6507 @obeylines@parskip=0pt@parindent=0pt
6508 @dag@quad Prefix operators.
6509 @ddag@quad @xref{Assignments}.
6512 @c END TEXI2ROFF-KILL
6515 @subsection Evaluation
6516 @cindex lazy evaluation
6517 @cindex expression evaluation order
6518 The linker evaluates expressions lazily. It only computes the value of
6519 an expression when absolutely necessary.
6521 The linker needs some information, such as the value of the start
6522 address of the first section, and the origins and lengths of memory
6523 regions, in order to do any linking at all. These values are computed
6524 as soon as possible when the linker reads in the linker script.
6526 However, other values (such as symbol values) are not known or needed
6527 until after storage allocation. Such values are evaluated later, when
6528 other information (such as the sizes of output sections) is available
6529 for use in the symbol assignment expression.
6531 The sizes of sections cannot be known until after allocation, so
6532 assignments dependent upon these are not performed until after
6535 Some expressions, such as those depending upon the location counter
6536 @samp{.}, must be evaluated during section allocation.
6538 If the result of an expression is required, but the value is not
6539 available, then an error results. For example, a script like the
6545 .text 9+this_isnt_constant :
6551 will cause the error message @samp{non constant expression for initial
6554 @node Expression Section
6555 @subsection The Section of an Expression
6556 @cindex expression sections
6557 @cindex absolute expressions
6558 @cindex relative expressions
6559 @cindex absolute and relocatable symbols
6560 @cindex relocatable and absolute symbols
6561 @cindex symbols, relocatable and absolute
6562 Addresses and symbols may be section relative, or absolute. A section
6563 relative symbol is relocatable. If you request relocatable output
6564 using the @samp{-r} option, a further link operation may change the
6565 value of a section relative symbol. On the other hand, an absolute
6566 symbol will retain the same value throughout any further link
6569 Some terms in linker expressions are addresses. This is true of
6570 section relative symbols and for builtin functions that return an
6571 address, such as @code{ADDR}, @code{LOADADDR}, @code{ORIGIN} and
6572 @code{SEGMENT_START}. Other terms are simply numbers, or are builtin
6573 functions that return a non-address value, such as @code{LENGTH}.
6574 One complication is that unless you set @code{LD_FEATURE ("SANE_EXPR")}
6575 (@pxref{Miscellaneous Commands}), numbers and absolute symbols are treated
6576 differently depending on their location, for compatibility with older
6577 versions of @code{ld}. Expressions appearing outside an output
6578 section definition treat all numbers as absolute addresses.
6579 Expressions appearing inside an output section definition treat
6580 absolute symbols as numbers. If @code{LD_FEATURE ("SANE_EXPR")} is
6581 given, then absolute symbols and numbers are simply treated as numbers
6584 In the following simple example,
6591 __executable_start = 0x100;
6595 __data_start = 0x10;
6603 both @code{.} and @code{__executable_start} are set to the absolute
6604 address 0x100 in the first two assignments, then both @code{.} and
6605 @code{__data_start} are set to 0x10 relative to the @code{.data}
6606 section in the second two assignments.
6608 For expressions involving numbers, relative addresses and absolute
6609 addresses, ld follows these rules to evaluate terms:
6613 Unary operations on an absolute address or number, and binary
6614 operations on two absolute addresses or two numbers, or between one
6615 absolute address and a number, apply the operator to the value(s).
6617 Unary operations on a relative address, and binary operations on two
6618 relative addresses in the same section or between one relative address
6619 and a number, apply the operator to the offset part of the address(es).
6621 Other binary operations, that is, between two relative addresses not
6622 in the same section, or between a relative address and an absolute
6623 address, first convert any non-absolute term to an absolute address
6624 before applying the operator.
6627 The result section of each sub-expression is as follows:
6631 An operation involving only numbers results in a number.
6633 The result of comparisons, @samp{&&} and @samp{||} is also a number.
6635 The result of other binary arithmetic and logical operations on two
6636 relative addresses in the same section or two absolute addresses
6637 (after above conversions) is also a number when
6638 @code{LD_FEATURE ("SANE_EXPR")} or inside an output section definition
6639 but an absolute address otherwise.
6641 The result of other operations on relative addresses or one
6642 relative address and a number, is a relative address in the same
6643 section as the relative operand(s).
6645 The result of other operations on absolute addresses (after above
6646 conversions) is an absolute address.
6649 You can use the builtin function @code{ABSOLUTE} to force an expression
6650 to be absolute when it would otherwise be relative. For example, to
6651 create an absolute symbol set to the address of the end of the output
6652 section @samp{.data}:
6656 .data : @{ *(.data) _edata = ABSOLUTE(.); @}
6660 If @samp{ABSOLUTE} were not used, @samp{_edata} would be relative to the
6661 @samp{.data} section.
6663 Using @code{LOADADDR} also forces an expression absolute, since this
6664 particular builtin function returns an absolute address.
6666 @node Builtin Functions
6667 @subsection Builtin Functions
6668 @cindex functions in expressions
6669 The linker script language includes a number of builtin functions for
6670 use in linker script expressions.
6673 @item ABSOLUTE(@var{exp})
6674 @kindex ABSOLUTE(@var{exp})
6675 @cindex expression, absolute
6676 Return the absolute (non-relocatable, as opposed to non-negative) value
6677 of the expression @var{exp}. Primarily useful to assign an absolute
6678 value to a symbol within a section definition, where symbol values are
6679 normally section relative. @xref{Expression Section}.
6681 @item ADDR(@var{section})
6682 @kindex ADDR(@var{section})
6683 @cindex section address in expression
6684 Return the address (VMA) of the named @var{section}. Your
6685 script must previously have defined the location of that section. In
6686 the following example, @code{start_of_output_1}, @code{symbol_1} and
6687 @code{symbol_2} are assigned equivalent values, except that
6688 @code{symbol_1} will be relative to the @code{.output1} section while
6689 the other two will be absolute:
6695 start_of_output_1 = ABSOLUTE(.);
6700 symbol_1 = ADDR(.output1);
6701 symbol_2 = start_of_output_1;
6707 @item ALIGN(@var{align})
6708 @itemx ALIGN(@var{exp},@var{align})
6709 @kindex ALIGN(@var{align})
6710 @kindex ALIGN(@var{exp},@var{align})
6711 @cindex round up location counter
6712 @cindex align location counter
6713 @cindex round up expression
6714 @cindex align expression
6715 Return the location counter (@code{.}) or arbitrary expression aligned
6716 to the next @var{align} boundary. The single operand @code{ALIGN}
6717 doesn't change the value of the location counter---it just does
6718 arithmetic on it. The two operand @code{ALIGN} allows an arbitrary
6719 expression to be aligned upwards (@code{ALIGN(@var{align})} is
6720 equivalent to @code{ALIGN(ABSOLUTE(.), @var{align})}).
6722 Here is an example which aligns the output @code{.data} section to the
6723 next @code{0x2000} byte boundary after the preceding section and sets a
6724 variable within the section to the next @code{0x8000} boundary after the
6729 .data ALIGN(0x2000): @{
6731 variable = ALIGN(0x8000);
6737 The first use of @code{ALIGN} in this example specifies the location of
6738 a section because it is used as the optional @var{address} attribute of
6739 a section definition (@pxref{Output Section Address}). The second use
6740 of @code{ALIGN} is used to defines the value of a symbol.
6742 The builtin function @code{NEXT} is closely related to @code{ALIGN}.
6744 @item ALIGNOF(@var{section})
6745 @kindex ALIGNOF(@var{section})
6746 @cindex section alignment
6747 Return the alignment in bytes of the named @var{section}, if that section has
6748 been allocated. If the section has not been allocated when this is
6749 evaluated, the linker will report an error. In the following example,
6750 the alignment of the @code{.output} section is stored as the first
6751 value in that section.
6756 LONG (ALIGNOF (.output))
6763 @item BLOCK(@var{exp})
6764 @kindex BLOCK(@var{exp})
6765 This is a synonym for @code{ALIGN}, for compatibility with older linker
6766 scripts. It is most often seen when setting the address of an output
6769 @item DATA_SEGMENT_ALIGN(@var{maxpagesize}, @var{commonpagesize})
6770 @kindex DATA_SEGMENT_ALIGN(@var{maxpagesize}, @var{commonpagesize})
6771 This is equivalent to either
6773 (ALIGN(@var{maxpagesize}) + (. & (@var{maxpagesize} - 1)))
6777 (ALIGN(@var{maxpagesize})
6778 + ((. + @var{commonpagesize} - 1) & (@var{maxpagesize} - @var{commonpagesize})))
6781 depending on whether the latter uses fewer @var{commonpagesize} sized pages
6782 for the data segment (area between the result of this expression and
6783 @code{DATA_SEGMENT_END}) than the former or not.
6784 If the latter form is used, it means @var{commonpagesize} bytes of runtime
6785 memory will be saved at the expense of up to @var{commonpagesize} wasted
6786 bytes in the on-disk file.
6788 This expression can only be used directly in @code{SECTIONS} commands, not in
6789 any output section descriptions and only once in the linker script.
6790 @var{commonpagesize} should be less or equal to @var{maxpagesize} and should
6791 be the system page size the object wants to be optimized for while still
6792 running on system page sizes up to @var{maxpagesize}. Note however
6793 that @samp{-z relro} protection will not be effective if the system
6794 page size is larger than @var{commonpagesize}.
6799 . = DATA_SEGMENT_ALIGN(0x10000, 0x2000);
6802 @item DATA_SEGMENT_END(@var{exp})
6803 @kindex DATA_SEGMENT_END(@var{exp})
6804 This defines the end of data segment for @code{DATA_SEGMENT_ALIGN}
6805 evaluation purposes.
6808 . = DATA_SEGMENT_END(.);
6811 @item DATA_SEGMENT_RELRO_END(@var{offset}, @var{exp})
6812 @kindex DATA_SEGMENT_RELRO_END(@var{offset}, @var{exp})
6813 This defines the end of the @code{PT_GNU_RELRO} segment when
6814 @samp{-z relro} option is used.
6815 When @samp{-z relro} option is not present, @code{DATA_SEGMENT_RELRO_END}
6816 does nothing, otherwise @code{DATA_SEGMENT_ALIGN} is padded so that
6817 @var{exp} + @var{offset} is aligned to the @var{commonpagesize}
6818 argument given to @code{DATA_SEGMENT_ALIGN}. If present in the linker
6819 script, it must be placed between @code{DATA_SEGMENT_ALIGN} and
6820 @code{DATA_SEGMENT_END}. Evaluates to the second argument plus any
6821 padding needed at the end of the @code{PT_GNU_RELRO} segment due to
6825 . = DATA_SEGMENT_RELRO_END(24, .);
6828 @item DEFINED(@var{symbol})
6829 @kindex DEFINED(@var{symbol})
6830 @cindex symbol defaults
6831 Return 1 if @var{symbol} is in the linker global symbol table and is
6832 defined before the statement using DEFINED in the script, otherwise
6833 return 0. You can use this function to provide
6834 default values for symbols. For example, the following script fragment
6835 shows how to set a global symbol @samp{begin} to the first location in
6836 the @samp{.text} section---but if a symbol called @samp{begin} already
6837 existed, its value is preserved:
6843 begin = DEFINED(begin) ? begin : . ;
6851 @item LENGTH(@var{memory})
6852 @kindex LENGTH(@var{memory})
6853 Return the length of the memory region named @var{memory}.
6855 @item LOADADDR(@var{section})
6856 @kindex LOADADDR(@var{section})
6857 @cindex section load address in expression
6858 Return the absolute LMA of the named @var{section}. (@pxref{Output
6861 @item LOG2CEIL(@var{exp})
6862 @kindex LOG2CEIL(@var{exp})
6863 Return the binary logarithm of @var{exp} rounded towards infinity.
6864 @code{LOG2CEIL(0)} returns 0.
6867 @item MAX(@var{exp1}, @var{exp2})
6868 Returns the maximum of @var{exp1} and @var{exp2}.
6871 @item MIN(@var{exp1}, @var{exp2})
6872 Returns the minimum of @var{exp1} and @var{exp2}.
6874 @item NEXT(@var{exp})
6875 @kindex NEXT(@var{exp})
6876 @cindex unallocated address, next
6877 Return the next unallocated address that is a multiple of @var{exp}.
6878 This function is closely related to @code{ALIGN(@var{exp})}; unless you
6879 use the @code{MEMORY} command to define discontinuous memory for the
6880 output file, the two functions are equivalent.
6882 @item ORIGIN(@var{memory})
6883 @kindex ORIGIN(@var{memory})
6884 Return the origin of the memory region named @var{memory}.
6886 @item SEGMENT_START(@var{segment}, @var{default})
6887 @kindex SEGMENT_START(@var{segment}, @var{default})
6888 Return the base address of the named @var{segment}. If an explicit
6889 value has already been given for this segment (with a command-line
6890 @samp{-T} option) then that value will be returned otherwise the value
6891 will be @var{default}. At present, the @samp{-T} command-line option
6892 can only be used to set the base address for the ``text'', ``data'', and
6893 ``bss'' sections, but you can use @code{SEGMENT_START} with any segment
6896 @item SIZEOF(@var{section})
6897 @kindex SIZEOF(@var{section})
6898 @cindex section size
6899 Return the size in bytes of the named @var{section}, if that section has
6900 been allocated. If the section has not been allocated when this is
6901 evaluated, the linker will report an error. In the following example,
6902 @code{symbol_1} and @code{symbol_2} are assigned identical values:
6911 symbol_1 = .end - .start ;
6912 symbol_2 = SIZEOF(.output);
6917 @item SIZEOF_HEADERS
6918 @itemx sizeof_headers
6919 @kindex SIZEOF_HEADERS
6921 Return the size in bytes of the output file's headers. This is
6922 information which appears at the start of the output file. You can use
6923 this number when setting the start address of the first section, if you
6924 choose, to facilitate paging.
6926 @cindex not enough room for program headers
6927 @cindex program headers, not enough room
6928 When producing an ELF output file, if the linker script uses the
6929 @code{SIZEOF_HEADERS} builtin function, the linker must compute the
6930 number of program headers before it has determined all the section
6931 addresses and sizes. If the linker later discovers that it needs
6932 additional program headers, it will report an error @samp{not enough
6933 room for program headers}. To avoid this error, you must avoid using
6934 the @code{SIZEOF_HEADERS} function, or you must rework your linker
6935 script to avoid forcing the linker to use additional program headers, or
6936 you must define the program headers yourself using the @code{PHDRS}
6937 command (@pxref{PHDRS}).
6940 @node Implicit Linker Scripts
6941 @section Implicit Linker Scripts
6942 @cindex implicit linker scripts
6943 If you specify a linker input file which the linker can not recognize as
6944 an object file or an archive file, it will try to read the file as a
6945 linker script. If the file can not be parsed as a linker script, the
6946 linker will report an error.
6948 An implicit linker script will not replace the default linker script.
6950 Typically an implicit linker script would contain only symbol
6951 assignments, or the @code{INPUT}, @code{GROUP}, or @code{VERSION}
6954 Any input files read because of an implicit linker script will be read
6955 at the position in the command line where the implicit linker script was
6956 read. This can affect archive searching.
6959 @chapter Linker Plugins
6962 @cindex linker plugins
6963 The linker can use dynamically loaded plugins to modify its behavior.
6964 For example, the link-time optimization feature that some compilers
6965 support is implemented with a linker plugin.
6967 Currently there is only one plugin shipped by default, but more may
6968 be added here later.
6971 * libdep Plugin:: Static Library Dependencies Plugin
6975 @section Static Library Dependencies Plugin
6976 @cindex static library dependencies
6977 Originally, static libraries were contained in an archive file consisting
6978 just of a collection of relocatable object files. Later they evolved to
6979 optionally include a symbol table, to assist in finding the needed objects
6980 within a library. There their evolution ended, and dynamic libraries
6983 One useful feature of dynamic libraries was that, more than just collecting
6984 multiple objects into a single file, they also included a list of their
6985 dependencies, such that one could specify just the name of a single dynamic
6986 library at link time, and all of its dependencies would be implicitly
6987 referenced as well. But static libraries lacked this feature, so if a
6988 link invocation was switched from using dynamic libraries to static
6989 libraries, the link command would usually fail unless it was rewritten to
6990 explicitly list the dependencies of the static library.
6992 The GNU @command{ar} utility now supports a @option{--record-libdeps} option
6993 to embed dependency lists into static libraries as well, and the @file{libdep}
6994 plugin may be used to read this dependency information at link time. The
6995 dependency information is stored as a single string, carrying @option{-l}
6996 and @option{-L} arguments as they would normally appear in a linker
6997 command line. As such, the information can be written with any text
6998 utility and stored into any archive, even if GNU @command{ar} is not
6999 being used to create the archive. The information is stored in an
7000 archive member named @samp{__.LIBDEP}.
7002 For example, given a library @file{libssl.a} that depends on another
7003 library @file{libcrypto.a} which may be found in @file{/usr/local/lib},
7004 the @samp{__.LIBDEP} member of @file{libssl.a} would contain
7007 -L/usr/local/lib -lcrypto
7011 @node Machine Dependent
7012 @chapter Machine Dependent Features
7014 @cindex machine dependencies
7015 @command{ld} has additional features on some platforms; the following
7016 sections describe them. Machines where @command{ld} has no additional
7017 functionality are not listed.
7021 * H8/300:: @command{ld} and the H8/300
7024 * M68HC11/68HC12:: @code{ld} and the Motorola 68HC11 and 68HC12 families
7027 * ARM:: @command{ld} and the ARM family
7030 * HPPA ELF32:: @command{ld} and HPPA 32-bit ELF
7033 * M68K:: @command{ld} and the Motorola 68K family
7036 * MIPS:: @command{ld} and the MIPS family
7039 * MMIX:: @command{ld} and MMIX
7042 * MSP430:: @command{ld} and MSP430
7045 * NDS32:: @command{ld} and NDS32
7048 * Nios II:: @command{ld} and the Altera Nios II
7051 * PowerPC ELF32:: @command{ld} and PowerPC 32-bit ELF Support
7054 * PowerPC64 ELF64:: @command{ld} and PowerPC64 64-bit ELF Support
7057 * S/390 ELF:: @command{ld} and S/390 ELF Support
7060 * SPU ELF:: @command{ld} and SPU ELF Support
7063 * TI COFF:: @command{ld} and TI COFF
7066 * WIN32:: @command{ld} and WIN32 (cygwin/mingw)
7069 * Xtensa:: @command{ld} and Xtensa Processors
7080 @section @command{ld} and the H8/300
7082 @cindex H8/300 support
7083 For the H8/300, @command{ld} can perform these global optimizations when
7084 you specify the @samp{--relax} command-line option.
7087 @cindex relaxing on H8/300
7088 @item relaxing address modes
7089 @command{ld} finds all @code{jsr} and @code{jmp} instructions whose
7090 targets are within eight bits, and turns them into eight-bit
7091 program-counter relative @code{bsr} and @code{bra} instructions,
7094 @cindex synthesizing on H8/300
7095 @item synthesizing instructions
7096 @c FIXME: specifically mov.b, or any mov instructions really? -> mov.b only, at least on H8, H8H, H8S
7097 @command{ld} finds all @code{mov.b} instructions which use the
7098 sixteen-bit absolute address form, but refer to the top
7099 page of memory, and changes them to use the eight-bit address form.
7100 (That is: the linker turns @samp{mov.b @code{@@}@var{aa}:16} into
7101 @samp{mov.b @code{@@}@var{aa}:8} whenever the address @var{aa} is in the
7102 top page of memory).
7104 @command{ld} finds all @code{mov} instructions which use the register
7105 indirect with 32-bit displacement addressing mode, but use a small
7106 displacement inside 16-bit displacement range, and changes them to use
7107 the 16-bit displacement form. (That is: the linker turns @samp{mov.b
7108 @code{@@}@var{d}:32,ERx} into @samp{mov.b @code{@@}@var{d}:16,ERx}
7109 whenever the displacement @var{d} is in the 16 bit signed integer
7110 range. Only implemented in ELF-format ld).
7112 @item bit manipulation instructions
7113 @command{ld} finds all bit manipulation instructions like @code{band, bclr,
7114 biand, bild, bior, bist, bixor, bld, bnot, bor, bset, bst, btst, bxor}
7115 which use 32 bit and 16 bit absolute address form, but refer to the top
7116 page of memory, and changes them to use the 8 bit address form.
7117 (That is: the linker turns @samp{bset #xx:3,@code{@@}@var{aa}:32} into
7118 @samp{bset #xx:3,@code{@@}@var{aa}:8} whenever the address @var{aa} is in
7119 the top page of memory).
7121 @item system control instructions
7122 @command{ld} finds all @code{ldc.w, stc.w} instructions which use the
7123 32 bit absolute address form, but refer to the top page of memory, and
7124 changes them to use 16 bit address form.
7125 (That is: the linker turns @samp{ldc.w @code{@@}@var{aa}:32,ccr} into
7126 @samp{ldc.w @code{@@}@var{aa}:16,ccr} whenever the address @var{aa} is in
7127 the top page of memory).
7137 @c This stuff is pointless to say unless you're especially concerned
7138 @c with Renesas chips; don't enable it for generic case, please.
7140 @chapter @command{ld} and Other Renesas Chips
7142 @command{ld} also supports the Renesas (formerly Hitachi) H8/300H,
7143 H8/500, and SH chips. No special features, commands, or command-line
7144 options are required for these chips.
7158 @node M68HC11/68HC12
7159 @section @command{ld} and the Motorola 68HC11 and 68HC12 families
7161 @cindex M68HC11 and 68HC12 support
7163 @subsection Linker Relaxation
7165 For the Motorola 68HC11, @command{ld} can perform these global
7166 optimizations when you specify the @samp{--relax} command-line option.
7169 @cindex relaxing on M68HC11
7170 @item relaxing address modes
7171 @command{ld} finds all @code{jsr} and @code{jmp} instructions whose
7172 targets are within eight bits, and turns them into eight-bit
7173 program-counter relative @code{bsr} and @code{bra} instructions,
7176 @command{ld} also looks at all 16-bit extended addressing modes and
7177 transforms them in a direct addressing mode when the address is in
7178 page 0 (between 0 and 0x0ff).
7180 @item relaxing gcc instruction group
7181 When @command{gcc} is called with @option{-mrelax}, it can emit group
7182 of instructions that the linker can optimize to use a 68HC11 direct
7183 addressing mode. These instructions consists of @code{bclr} or
7184 @code{bset} instructions.
7188 @subsection Trampoline Generation
7190 @cindex trampoline generation on M68HC11
7191 @cindex trampoline generation on M68HC12
7192 For 68HC11 and 68HC12, @command{ld} can generate trampoline code to
7193 call a far function using a normal @code{jsr} instruction. The linker
7194 will also change the relocation to some far function to use the
7195 trampoline address instead of the function address. This is typically the
7196 case when a pointer to a function is taken. The pointer will in fact
7197 point to the function trampoline.
7205 @section @command{ld} and the ARM family
7207 @cindex ARM interworking support
7208 @kindex --support-old-code
7209 For the ARM, @command{ld} will generate code stubs to allow functions calls
7210 between ARM and Thumb code. These stubs only work with code that has
7211 been compiled and assembled with the @samp{-mthumb-interwork} command
7212 line option. If it is necessary to link with old ARM object files or
7213 libraries, which have not been compiled with the -mthumb-interwork
7214 option then the @samp{--support-old-code} command-line switch should be
7215 given to the linker. This will make it generate larger stub functions
7216 which will work with non-interworking aware ARM code. Note, however,
7217 the linker does not support generating stubs for function calls to
7218 non-interworking aware Thumb code.
7220 @cindex thumb entry point
7221 @cindex entry point, thumb
7222 @kindex --thumb-entry=@var{entry}
7223 The @samp{--thumb-entry} switch is a duplicate of the generic
7224 @samp{--entry} switch, in that it sets the program's starting address.
7225 But it also sets the bottom bit of the address, so that it can be
7226 branched to using a BX instruction, and the program will start
7227 executing in Thumb mode straight away.
7229 @cindex PE import table prefixing
7230 @kindex --use-nul-prefixed-import-tables
7231 The @samp{--use-nul-prefixed-import-tables} switch is specifying, that
7232 the import tables idata4 and idata5 have to be generated with a zero
7233 element prefix for import libraries. This is the old style to generate
7234 import tables. By default this option is turned off.
7238 The @samp{--be8} switch instructs @command{ld} to generate BE8 format
7239 executables. This option is only valid when linking big-endian
7240 objects - ie ones which have been assembled with the @option{-EB}
7241 option. The resulting image will contain big-endian data and
7245 @kindex --target1-rel
7246 @kindex --target1-abs
7247 The @samp{R_ARM_TARGET1} relocation is typically used for entries in the
7248 @samp{.init_array} section. It is interpreted as either @samp{R_ARM_REL32}
7249 or @samp{R_ARM_ABS32}, depending on the target. The @samp{--target1-rel}
7250 and @samp{--target1-abs} switches override the default.
7253 @kindex --target2=@var{type}
7254 The @samp{--target2=type} switch overrides the default definition of the
7255 @samp{R_ARM_TARGET2} relocation. Valid values for @samp{type}, their
7256 meanings, and target defaults are as follows:
7259 @samp{R_ARM_REL32} (arm*-*-elf, arm*-*-eabi)
7261 @samp{R_ARM_ABS32} (arm*-*-symbianelf)
7263 @samp{R_ARM_GOT_PREL} (arm*-*-linux, arm*-*-*bsd)
7268 The @samp{R_ARM_V4BX} relocation (defined by the ARM AAELF
7269 specification) enables objects compiled for the ARMv4 architecture to be
7270 interworking-safe when linked with other objects compiled for ARMv4t, but
7271 also allows pure ARMv4 binaries to be built from the same ARMv4 objects.
7273 In the latter case, the switch @option{--fix-v4bx} must be passed to the
7274 linker, which causes v4t @code{BX rM} instructions to be rewritten as
7275 @code{MOV PC,rM}, since v4 processors do not have a @code{BX} instruction.
7277 In the former case, the switch should not be used, and @samp{R_ARM_V4BX}
7278 relocations are ignored.
7280 @cindex FIX_V4BX_INTERWORKING
7281 @kindex --fix-v4bx-interworking
7282 Replace @code{BX rM} instructions identified by @samp{R_ARM_V4BX}
7283 relocations with a branch to the following veneer:
7291 This allows generation of libraries/applications that work on ARMv4 cores
7292 and are still interworking safe. Note that the above veneer clobbers the
7293 condition flags, so may cause incorrect program behavior in rare cases.
7297 The @samp{--use-blx} switch enables the linker to use ARM/Thumb
7298 BLX instructions (available on ARMv5t and above) in various
7299 situations. Currently it is used to perform calls via the PLT from Thumb
7300 code using BLX rather than using BX and a mode-switching stub before
7301 each PLT entry. This should lead to such calls executing slightly faster.
7303 This option is enabled implicitly for SymbianOS, so there is no need to
7304 specify it if you are using that target.
7306 @cindex VFP11_DENORM_FIX
7307 @kindex --vfp11-denorm-fix
7308 The @samp{--vfp11-denorm-fix} switch enables a link-time workaround for a
7309 bug in certain VFP11 coprocessor hardware, which sometimes allows
7310 instructions with denorm operands (which must be handled by support code)
7311 to have those operands overwritten by subsequent instructions before
7312 the support code can read the intended values.
7314 The bug may be avoided in scalar mode if you allow at least one
7315 intervening instruction between a VFP11 instruction which uses a register
7316 and another instruction which writes to the same register, or at least two
7317 intervening instructions if vector mode is in use. The bug only affects
7318 full-compliance floating-point mode: you do not need this workaround if
7319 you are using "runfast" mode. Please contact ARM for further details.
7321 If you know you are using buggy VFP11 hardware, you can
7322 enable this workaround by specifying the linker option
7323 @samp{--vfp-denorm-fix=scalar} if you are using the VFP11 scalar
7324 mode only, or @samp{--vfp-denorm-fix=vector} if you are using
7325 vector mode (the latter also works for scalar code). The default is
7326 @samp{--vfp-denorm-fix=none}.
7328 If the workaround is enabled, instructions are scanned for
7329 potentially-troublesome sequences, and a veneer is created for each
7330 such sequence which may trigger the erratum. The veneer consists of the
7331 first instruction of the sequence and a branch back to the subsequent
7332 instruction. The original instruction is then replaced with a branch to
7333 the veneer. The extra cycles required to call and return from the veneer
7334 are sufficient to avoid the erratum in both the scalar and vector cases.
7336 @cindex ARM1176 erratum workaround
7337 @kindex --fix-arm1176
7338 @kindex --no-fix-arm1176
7339 The @samp{--fix-arm1176} switch enables a link-time workaround for an erratum
7340 in certain ARM1176 processors. The workaround is enabled by default if you
7341 are targeting ARM v6 (excluding ARM v6T2) or earlier. It can be disabled
7342 unconditionally by specifying @samp{--no-fix-arm1176}.
7344 Further information is available in the ``ARM1176JZ-S and ARM1176JZF-S
7345 Programmer Advice Notice'' available on the ARM documentation website at:
7346 http://infocenter.arm.com/.
7348 @cindex STM32L4xx erratum workaround
7349 @kindex --fix-stm32l4xx-629360
7351 The @samp{--fix-stm32l4xx-629360} switch enables a link-time
7352 workaround for a bug in the bus matrix / memory controller for some of
7353 the STM32 Cortex-M4 based products (STM32L4xx). When accessing
7354 off-chip memory via the affected bus for bus reads of 9 words or more,
7355 the bus can generate corrupt data and/or abort. These are only
7356 core-initiated accesses (not DMA), and might affect any access:
7357 integer loads such as LDM, POP and floating-point loads such as VLDM,
7358 VPOP. Stores are not affected.
7360 The bug can be avoided by splitting memory accesses into the
7361 necessary chunks to keep bus reads below 8 words.
7363 The workaround is not enabled by default, this is equivalent to use
7364 @samp{--fix-stm32l4xx-629360=none}. If you know you are using buggy
7365 STM32L4xx hardware, you can enable the workaround by specifying the
7366 linker option @samp{--fix-stm32l4xx-629360}, or the equivalent
7367 @samp{--fix-stm32l4xx-629360=default}.
7369 If the workaround is enabled, instructions are scanned for
7370 potentially-troublesome sequences, and a veneer is created for each
7371 such sequence which may trigger the erratum. The veneer consists in a
7372 replacement sequence emulating the behaviour of the original one and a
7373 branch back to the subsequent instruction. The original instruction is
7374 then replaced with a branch to the veneer.
7376 The workaround does not always preserve the memory access order for
7377 the LDMDB instruction, when the instruction loads the PC.
7379 The workaround is not able to handle problematic instructions when
7380 they are in the middle of an IT block, since a branch is not allowed
7381 there. In that case, the linker reports a warning and no replacement
7384 The workaround is not able to replace problematic instructions with a
7385 PC-relative branch instruction if the @samp{.text} section is too
7386 large. In that case, when the branch that replaces the original code
7387 cannot be encoded, the linker reports a warning and no replacement
7390 @cindex NO_ENUM_SIZE_WARNING
7391 @kindex --no-enum-size-warning
7392 The @option{--no-enum-size-warning} switch prevents the linker from
7393 warning when linking object files that specify incompatible EABI
7394 enumeration size attributes. For example, with this switch enabled,
7395 linking of an object file using 32-bit enumeration values with another
7396 using enumeration values fitted into the smallest possible space will
7399 @cindex NO_WCHAR_SIZE_WARNING
7400 @kindex --no-wchar-size-warning
7401 The @option{--no-wchar-size-warning} switch prevents the linker from
7402 warning when linking object files that specify incompatible EABI
7403 @code{wchar_t} size attributes. For example, with this switch enabled,
7404 linking of an object file using 32-bit @code{wchar_t} values with another
7405 using 16-bit @code{wchar_t} values will not be diagnosed.
7408 @kindex --pic-veneer
7409 The @samp{--pic-veneer} switch makes the linker use PIC sequences for
7410 ARM/Thumb interworking veneers, even if the rest of the binary
7411 is not PIC. This avoids problems on uClinux targets where
7412 @samp{--emit-relocs} is used to generate relocatable binaries.
7414 @cindex STUB_GROUP_SIZE
7415 @kindex --stub-group-size=@var{N}
7416 The linker will automatically generate and insert small sequences of
7417 code into a linked ARM ELF executable whenever an attempt is made to
7418 perform a function call to a symbol that is too far away. The
7419 placement of these sequences of instructions - called stubs - is
7420 controlled by the command-line option @option{--stub-group-size=N}.
7421 The placement is important because a poor choice can create a need for
7422 duplicate stubs, increasing the code size. The linker will try to
7423 group stubs together in order to reduce interruptions to the flow of
7424 code, but it needs guidance as to how big these groups should be and
7425 where they should be placed.
7427 The value of @samp{N}, the parameter to the
7428 @option{--stub-group-size=} option controls where the stub groups are
7429 placed. If it is negative then all stubs are placed after the first
7430 branch that needs them. If it is positive then the stubs can be
7431 placed either before or after the branches that need them. If the
7432 value of @samp{N} is 1 (either +1 or -1) then the linker will choose
7433 exactly where to place groups of stubs, using its built in heuristics.
7434 A value of @samp{N} greater than 1 (or smaller than -1) tells the
7435 linker that a single group of stubs can service at most @samp{N} bytes
7436 from the input sections.
7438 The default, if @option{--stub-group-size=} is not specified, is
7441 Farcalls stubs insertion is fully supported for the ARM-EABI target
7442 only, because it relies on object files properties not present
7445 @cindex Cortex-A8 erratum workaround
7446 @kindex --fix-cortex-a8
7447 @kindex --no-fix-cortex-a8
7448 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}.
7450 The erratum only affects Thumb-2 code. Please contact ARM for further details.
7452 @cindex Cortex-A53 erratum 835769 workaround
7453 @kindex --fix-cortex-a53-835769
7454 @kindex --no-fix-cortex-a53-835769
7455 The @samp{--fix-cortex-a53-835769} switch enables a link-time workaround for erratum 835769 present on certain early revisions of Cortex-A53 processors. The workaround is disabled by default. It can be enabled by specifying @samp{--fix-cortex-a53-835769}, or disabled unconditionally by specifying @samp{--no-fix-cortex-a53-835769}.
7457 Please contact ARM for further details.
7459 @kindex --merge-exidx-entries
7460 @kindex --no-merge-exidx-entries
7461 @cindex Merging exidx entries
7462 The @samp{--no-merge-exidx-entries} switch disables the merging of adjacent exidx entries in debuginfo.
7465 @cindex 32-bit PLT entries
7466 The @samp{--long-plt} option enables the use of 16 byte PLT entries
7467 which support up to 4Gb of code. The default is to use 12 byte PLT
7468 entries which only support 512Mb of code.
7470 @kindex --no-apply-dynamic-relocs
7471 @cindex AArch64 rela addend
7472 The @samp{--no-apply-dynamic-relocs} option makes AArch64 linker do not apply
7473 link-time values for dynamic relocations.
7475 @cindex Placement of SG veneers
7476 All SG veneers are placed in the special output section @code{.gnu.sgstubs}.
7477 Its start address must be set, either with the command-line option
7478 @samp{--section-start} or in a linker script, to indicate where to place these
7481 @kindex --cmse-implib
7482 @cindex Secure gateway import library
7483 The @samp{--cmse-implib} option requests that the import libraries
7484 specified by the @samp{--out-implib} and @samp{--in-implib} options are
7485 secure gateway import libraries, suitable for linking a non-secure
7486 executable against secure code as per ARMv8-M Security Extensions.
7488 @kindex --in-implib=@var{file}
7489 @cindex Input import library
7490 The @samp{--in-implib=file} specifies an input import library whose symbols
7491 must keep the same address in the executable being produced. A warning is
7492 given if no @samp{--out-implib} is given but new symbols have been introduced
7493 in the executable that should be listed in its import library. Otherwise, if
7494 @samp{--out-implib} is specified, the symbols are added to the output import
7495 library. A warning is also given if some symbols present in the input import
7496 library have disappeared from the executable. This option is only effective
7497 for Secure Gateway import libraries, ie. when @samp{--cmse-implib} is
7511 @section @command{ld} and HPPA 32-bit ELF Support
7512 @cindex HPPA multiple sub-space stubs
7513 @kindex --multi-subspace
7514 When generating a shared library, @command{ld} will by default generate
7515 import stubs suitable for use with a single sub-space application.
7516 The @samp{--multi-subspace} switch causes @command{ld} to generate export
7517 stubs, and different (larger) import stubs suitable for use with
7518 multiple sub-spaces.
7520 @cindex HPPA stub grouping
7521 @kindex --stub-group-size=@var{N}
7522 Long branch stubs and import/export stubs are placed by @command{ld} in
7523 stub sections located between groups of input sections.
7524 @samp{--stub-group-size} specifies the maximum size of a group of input
7525 sections handled by one stub section. Since branch offsets are signed,
7526 a stub section may serve two groups of input sections, one group before
7527 the stub section, and one group after it. However, when using
7528 conditional branches that require stubs, it may be better (for branch
7529 prediction) that stub sections only serve one group of input sections.
7530 A negative value for @samp{N} chooses this scheme, ensuring that
7531 branches to stubs always use a negative offset. Two special values of
7532 @samp{N} are recognized, @samp{1} and @samp{-1}. These both instruct
7533 @command{ld} to automatically size input section groups for the branch types
7534 detected, with the same behaviour regarding stub placement as other
7535 positive or negative values of @samp{N} respectively.
7537 Note that @samp{--stub-group-size} does not split input sections. A
7538 single input section larger than the group size specified will of course
7539 create a larger group (of one section). If input sections are too
7540 large, it may not be possible for a branch to reach its stub.
7553 @section @command{ld} and the Motorola 68K family
7555 @cindex Motorola 68K GOT generation
7556 @kindex --got=@var{type}
7557 The @samp{--got=@var{type}} option lets you choose the GOT generation scheme.
7558 The choices are @samp{single}, @samp{negative}, @samp{multigot} and
7559 @samp{target}. When @samp{target} is selected the linker chooses
7560 the default GOT generation scheme for the current target.
7561 @samp{single} tells the linker to generate a single GOT with
7562 entries only at non-negative offsets.
7563 @samp{negative} instructs the linker to generate a single GOT with
7564 entries at both negative and positive offsets. Not all environments
7566 @samp{multigot} allows the linker to generate several GOTs in the
7567 output file. All GOT references from a single input object
7568 file access the same GOT, but references from different input object
7569 files might access different GOTs. Not all environments support such GOTs.
7582 @section @command{ld} and the MIPS family
7584 @cindex MIPS microMIPS instruction choice selection
7587 The @samp{--insn32} and @samp{--no-insn32} options control the choice of
7588 microMIPS instructions used in code generated by the linker, such as that
7589 in the PLT or lazy binding stubs, or in relaxation. If @samp{--insn32} is
7590 used, then the linker only uses 32-bit instruction encodings. By default
7591 or if @samp{--no-insn32} is used, all instruction encodings are used,
7592 including 16-bit ones where possible.
7594 @cindex MIPS branch relocation check control
7595 @kindex --ignore-branch-isa
7596 @kindex --no-ignore-branch-isa
7597 The @samp{--ignore-branch-isa} and @samp{--no-ignore-branch-isa} options
7598 control branch relocation checks for invalid ISA mode transitions. If
7599 @samp{--ignore-branch-isa} is used, then the linker accepts any branch
7600 relocations and any ISA mode transition required is lost in relocation
7601 calculation, except for some cases of @code{BAL} instructions which meet
7602 relaxation conditions and are converted to equivalent @code{JALX}
7603 instructions as the associated relocation is calculated. By default
7604 or if @samp{--no-ignore-branch-isa} is used a check is made causing
7605 the loss of an ISA mode transition to produce an error.
7618 @section @code{ld} and MMIX
7619 For MMIX, there is a choice of generating @code{ELF} object files or
7620 @code{mmo} object files when linking. The simulator @code{mmix}
7621 understands the @code{mmo} format. The binutils @code{objcopy} utility
7622 can translate between the two formats.
7624 There is one special section, the @samp{.MMIX.reg_contents} section.
7625 Contents in this section is assumed to correspond to that of global
7626 registers, and symbols referring to it are translated to special symbols,
7627 equal to registers. In a final link, the start address of the
7628 @samp{.MMIX.reg_contents} section corresponds to the first allocated
7629 global register multiplied by 8. Register @code{$255} is not included in
7630 this section; it is always set to the program entry, which is at the
7631 symbol @code{Main} for @code{mmo} files.
7633 Global symbols with the prefix @code{__.MMIX.start.}, for example
7634 @code{__.MMIX.start..text} and @code{__.MMIX.start..data} are special.
7635 The default linker script uses these to set the default start address
7638 Initial and trailing multiples of zero-valued 32-bit words in a section,
7639 are left out from an mmo file.
7652 @section @code{ld} and MSP430
7653 For the MSP430 it is possible to select the MPU architecture. The flag @samp{-m [mpu type]}
7654 will select an appropriate linker script for selected MPU type. (To get a list of known MPUs
7655 just pass @samp{-m help} option to the linker).
7657 @cindex MSP430 extra sections
7658 The linker will recognize some extra sections which are MSP430 specific:
7661 @item @samp{.vectors}
7662 Defines a portion of ROM where interrupt vectors located.
7664 @item @samp{.bootloader}
7665 Defines the bootloader portion of the ROM (if applicable). Any code
7666 in this section will be uploaded to the MPU.
7668 @item @samp{.infomem}
7669 Defines an information memory section (if applicable). Any code in
7670 this section will be uploaded to the MPU.
7672 @item @samp{.infomemnobits}
7673 This is the same as the @samp{.infomem} section except that any code
7674 in this section will not be uploaded to the MPU.
7676 @item @samp{.noinit}
7677 Denotes a portion of RAM located above @samp{.bss} section.
7679 The last two sections are used by gcc.
7683 @cindex MSP430 Options
7684 @kindex --code-region
7685 @item --code-region=[either,lower,upper,none]
7686 This will transform .text* sections to [either,lower,upper].text* sections. The
7687 argument passed to GCC for -mcode-region is propagated to the linker
7690 @kindex --data-region
7691 @item --data-region=[either,lower,upper,none]
7692 This will transform .data*, .bss* and .rodata* sections to
7693 [either,lower,upper].[data,bss,rodata]* sections. The argument passed to GCC
7694 for -mdata-region is propagated to the linker using this option.
7696 @kindex --disable-sec-transformation
7697 @item --disable-sec-transformation
7698 Prevent the transformation of sections as specified by the @code{--code-region}
7699 and @code{--data-region} options.
7700 This is useful if you are compiling and linking using a single call to the GCC
7701 wrapper, and want to compile the source files using -m[code,data]-region but
7702 not transform the sections for prebuilt libraries and objects.
7716 @section @code{ld} and NDS32
7717 @kindex relaxing on NDS32
7718 For NDS32, there are some options to select relaxation behavior. The linker
7719 relaxes objects according to these options.
7722 @item @samp{--m[no-]fp-as-gp}
7723 Disable/enable fp-as-gp relaxation.
7725 @item @samp{--mexport-symbols=FILE}
7726 Exporting symbols and their address into FILE as linker script.
7728 @item @samp{--m[no-]ex9}
7729 Disable/enable link-time EX9 relaxation.
7731 @item @samp{--mexport-ex9=FILE}
7732 Export the EX9 table after linking.
7734 @item @samp{--mimport-ex9=FILE}
7735 Import the Ex9 table for EX9 relaxation.
7737 @item @samp{--mupdate-ex9}
7738 Update the existing EX9 table.
7740 @item @samp{--mex9-limit=NUM}
7741 Maximum number of entries in the ex9 table.
7743 @item @samp{--mex9-loop-aware}
7744 Avoid generating the EX9 instruction inside the loop.
7746 @item @samp{--m[no-]ifc}
7747 Disable/enable the link-time IFC optimization.
7749 @item @samp{--mifc-loop-aware}
7750 Avoid generating the IFC instruction inside the loop.
7764 @section @command{ld} and the Altera Nios II
7765 @cindex Nios II call relaxation
7766 @kindex --relax on Nios II
7768 Call and immediate jump instructions on Nios II processors are limited to
7769 transferring control to addresses in the same 256MB memory segment,
7770 which may result in @command{ld} giving
7771 @samp{relocation truncated to fit} errors with very large programs.
7772 The command-line option @option{--relax} enables the generation of
7773 trampolines that can access the entire 32-bit address space for calls
7774 outside the normal @code{call} and @code{jmpi} address range. These
7775 trampolines are inserted at section boundaries, so may not themselves
7776 be reachable if an input section and its associated call trampolines are
7779 The @option{--relax} option is enabled by default unless @option{-r}
7780 is also specified. You can disable trampoline generation by using the
7781 @option{--no-relax} linker option. You can also disable this optimization
7782 locally by using the @samp{set .noat} directive in assembly-language
7783 source files, as the linker-inserted trampolines use the @code{at}
7784 register as a temporary.
7786 Note that the linker @option{--relax} option is independent of assembler
7787 relaxation options, and that using the GNU assembler's @option{-relax-all}
7788 option interferes with the linker's more selective call instruction relaxation.
7801 @section @command{ld} and PowerPC 32-bit ELF Support
7802 @cindex PowerPC long branches
7803 @kindex --relax on PowerPC
7804 Branches on PowerPC processors are limited to a signed 26-bit
7805 displacement, which may result in @command{ld} giving
7806 @samp{relocation truncated to fit} errors with very large programs.
7807 @samp{--relax} enables the generation of trampolines that can access
7808 the entire 32-bit address space. These trampolines are inserted at
7809 section boundaries, so may not themselves be reachable if an input
7810 section exceeds 33M in size. You may combine @samp{-r} and
7811 @samp{--relax} to add trampolines in a partial link. In that case
7812 both branches to undefined symbols and inter-section branches are also
7813 considered potentially out of range, and trampolines inserted.
7815 @cindex PowerPC ELF32 options
7820 Current PowerPC GCC accepts a @samp{-msecure-plt} option that
7821 generates code capable of using a newer PLT and GOT layout that has
7822 the security advantage of no executable section ever needing to be
7823 writable and no writable section ever being executable. PowerPC
7824 @command{ld} will generate this layout, including stubs to access the
7825 PLT, if all input files (including startup and static libraries) were
7826 compiled with @samp{-msecure-plt}. @samp{--bss-plt} forces the old
7827 BSS PLT (and GOT layout) which can give slightly better performance.
7829 @kindex --secure-plt
7831 @command{ld} will use the new PLT and GOT layout if it is linking new
7832 @samp{-fpic} or @samp{-fPIC} code, but does not do so automatically
7833 when linking non-PIC code. This option requests the new PLT and GOT
7834 layout. A warning will be given if some object file requires the old
7840 The new secure PLT and GOT are placed differently relative to other
7841 sections compared to older BSS PLT and GOT placement. The location of
7842 @code{.plt} must change because the new secure PLT is an initialized
7843 section while the old PLT is uninitialized. The reason for the
7844 @code{.got} change is more subtle: The new placement allows
7845 @code{.got} to be read-only in applications linked with
7846 @samp{-z relro -z now}. However, this placement means that
7847 @code{.sdata} cannot always be used in shared libraries, because the
7848 PowerPC ABI accesses @code{.sdata} in shared libraries from the GOT
7849 pointer. @samp{--sdata-got} forces the old GOT placement. PowerPC
7850 GCC doesn't use @code{.sdata} in shared libraries, so this option is
7851 really only useful for other compilers that may do so.
7853 @cindex PowerPC stub symbols
7854 @kindex --emit-stub-syms
7855 @item --emit-stub-syms
7856 This option causes @command{ld} to label linker stubs with a local
7857 symbol that encodes the stub type and destination.
7859 @cindex PowerPC TLS optimization
7860 @kindex --no-tls-optimize
7861 @item --no-tls-optimize
7862 PowerPC @command{ld} normally performs some optimization of code
7863 sequences used to access Thread-Local Storage. Use this option to
7864 disable the optimization.
7877 @node PowerPC64 ELF64
7878 @section @command{ld} and PowerPC64 64-bit ELF Support
7880 @cindex PowerPC64 ELF64 options
7882 @cindex PowerPC64 stub grouping
7883 @kindex --stub-group-size
7884 @item --stub-group-size
7885 Long branch stubs, PLT call stubs and TOC adjusting stubs are placed
7886 by @command{ld} in stub sections located between groups of input sections.
7887 @samp{--stub-group-size} specifies the maximum size of a group of input
7888 sections handled by one stub section. Since branch offsets are signed,
7889 a stub section may serve two groups of input sections, one group before
7890 the stub section, and one group after it. However, when using
7891 conditional branches that require stubs, it may be better (for branch
7892 prediction) that stub sections only serve one group of input sections.
7893 A negative value for @samp{N} chooses this scheme, ensuring that
7894 branches to stubs always use a negative offset. Two special values of
7895 @samp{N} are recognized, @samp{1} and @samp{-1}. These both instruct
7896 @command{ld} to automatically size input section groups for the branch types
7897 detected, with the same behaviour regarding stub placement as other
7898 positive or negative values of @samp{N} respectively.
7900 Note that @samp{--stub-group-size} does not split input sections. A
7901 single input section larger than the group size specified will of course
7902 create a larger group (of one section). If input sections are too
7903 large, it may not be possible for a branch to reach its stub.
7905 @cindex PowerPC64 stub symbols
7906 @kindex --emit-stub-syms
7907 @item --emit-stub-syms
7908 This option causes @command{ld} to label linker stubs with a local
7909 symbol that encodes the stub type and destination.
7911 @cindex PowerPC64 dot symbols
7913 @kindex --no-dotsyms
7916 These two options control how @command{ld} interprets version patterns
7917 in a version script. Older PowerPC64 compilers emitted both a
7918 function descriptor symbol with the same name as the function, and a
7919 code entry symbol with the name prefixed by a dot (@samp{.}). To
7920 properly version a function @samp{foo}, the version script thus needs
7921 to control both @samp{foo} and @samp{.foo}. The option
7922 @samp{--dotsyms}, on by default, automatically adds the required
7923 dot-prefixed patterns. Use @samp{--no-dotsyms} to disable this
7926 @cindex PowerPC64 register save/restore functions
7927 @kindex --save-restore-funcs
7928 @kindex --no-save-restore-funcs
7929 @item --save-restore-funcs
7930 @itemx --no-save-restore-funcs
7931 These two options control whether PowerPC64 @command{ld} automatically
7932 provides out-of-line register save and restore functions used by
7933 @samp{-Os} code. The default is to provide any such referenced
7934 function for a normal final link, and to not do so for a relocatable
7937 @cindex PowerPC64 TLS optimization
7938 @kindex --no-tls-optimize
7939 @item --no-tls-optimize
7940 PowerPC64 @command{ld} normally performs some optimization of code
7941 sequences used to access Thread-Local Storage. Use this option to
7942 disable the optimization.
7944 @cindex PowerPC64 __tls_get_addr optimization
7945 @kindex --tls-get-addr-optimize
7946 @kindex --no-tls-get-addr-optimize
7947 @kindex --tls-get-addr-regsave
7948 @kindex --no-tls-get-addr-regsave
7949 @item --tls-get-addr-optimize
7950 @itemx --no-tls-get-addr-optimize
7951 These options control how PowerPC64 @command{ld} uses a special
7952 stub to call __tls_get_addr. PowerPC64 glibc 2.22 and later support
7953 an optimization that allows the second and subsequent calls to
7954 @code{__tls_get_addr} for a given symbol to be resolved by the special
7955 stub without calling in to glibc. By default the linker enables
7956 generation of the stub when glibc advertises the availability of
7958 Using @option{--tls-get-addr-optimize} with an older glibc won't do
7959 much besides slow down your applications, but may be useful if linking
7960 an application against an older glibc with the expectation that it
7961 will normally be used on systems having a newer glibc.
7962 @option{--tls-get-addr-regsave} forces generation of a stub that saves
7963 and restores volatile registers around the call into glibc. Normally,
7964 this is done when the linker detects a call to __tls_get_addr_desc.
7965 Such calls then go via the register saving stub to __tls_get_addr_opt.
7966 @option{--no-tls-get-addr-regsave} disables generation of the
7969 @cindex PowerPC64 OPD optimization
7970 @kindex --no-opd-optimize
7971 @item --no-opd-optimize
7972 PowerPC64 @command{ld} normally removes @code{.opd} section entries
7973 corresponding to deleted link-once functions, or functions removed by
7974 the action of @samp{--gc-sections} or linker script @code{/DISCARD/}.
7975 Use this option to disable @code{.opd} optimization.
7977 @cindex PowerPC64 OPD spacing
7978 @kindex --non-overlapping-opd
7979 @item --non-overlapping-opd
7980 Some PowerPC64 compilers have an option to generate compressed
7981 @code{.opd} entries spaced 16 bytes apart, overlapping the third word,
7982 the static chain pointer (unused in C) with the first word of the next
7983 entry. This option expands such entries to the full 24 bytes.
7985 @cindex PowerPC64 TOC optimization
7986 @kindex --no-toc-optimize
7987 @item --no-toc-optimize
7988 PowerPC64 @command{ld} normally removes unused @code{.toc} section
7989 entries. Such entries are detected by examining relocations that
7990 reference the TOC in code sections. A reloc in a deleted code section
7991 marks a TOC word as unneeded, while a reloc in a kept code section
7992 marks a TOC word as needed. Since the TOC may reference itself, TOC
7993 relocs are also examined. TOC words marked as both needed and
7994 unneeded will of course be kept. TOC words without any referencing
7995 reloc are assumed to be part of a multi-word entry, and are kept or
7996 discarded as per the nearest marked preceding word. This works
7997 reliably for compiler generated code, but may be incorrect if assembly
7998 code is used to insert TOC entries. Use this option to disable the
8001 @cindex PowerPC64 inline PLT call optimization
8002 @kindex --no-inline-optimize
8003 @item --no-inline-optimize
8004 PowerPC64 @command{ld} normally replaces inline PLT call sequences
8005 marked with @code{R_PPC64_PLTSEQ}, @code{R_PPC64_PLTCALL},
8006 @code{R_PPC64_PLT16_HA} and @code{R_PPC64_PLT16_LO_DS} relocations by
8007 a number of @code{nop}s and a direct call when the function is defined
8008 locally and can't be overridden by some other definition. This option
8009 disables that optimization.
8011 @cindex PowerPC64 multi-TOC
8012 @kindex --no-multi-toc
8013 @item --no-multi-toc
8014 If given any toc option besides @code{-mcmodel=medium} or
8015 @code{-mcmodel=large}, PowerPC64 GCC generates code for a TOC model
8017 entries are accessed with a 16-bit offset from r2. This limits the
8018 total TOC size to 64K. PowerPC64 @command{ld} extends this limit by
8019 grouping code sections such that each group uses less than 64K for its
8020 TOC entries, then inserts r2 adjusting stubs between inter-group
8021 calls. @command{ld} does not split apart input sections, so cannot
8022 help if a single input file has a @code{.toc} section that exceeds
8023 64K, most likely from linking multiple files with @command{ld -r}.
8024 Use this option to turn off this feature.
8026 @cindex PowerPC64 TOC sorting
8027 @kindex --no-toc-sort
8029 By default, @command{ld} sorts TOC sections so that those whose file
8030 happens to have a section called @code{.init} or @code{.fini} are
8031 placed first, followed by TOC sections referenced by code generated
8032 with PowerPC64 gcc's @code{-mcmodel=small}, and lastly TOC sections
8033 referenced only by code generated with PowerPC64 gcc's
8034 @code{-mcmodel=medium} or @code{-mcmodel=large} options. Doing this
8035 results in better TOC grouping for multi-TOC. Use this option to turn
8038 @cindex PowerPC64 PLT stub alignment
8040 @kindex --no-plt-align
8042 @itemx --no-plt-align
8043 Use these options to control whether individual PLT call stubs are
8044 aligned to a 32-byte boundary, or to the specified power of two
8045 boundary when using @code{--plt-align=}. A negative value may be
8046 specified to pad PLT call stubs so that they do not cross the
8047 specified power of two boundary (or the minimum number of boundaries
8048 if a PLT stub is so large that it must cross a boundary). By default
8049 PLT call stubs are aligned to 32-byte boundaries.
8051 @cindex PowerPC64 PLT call stub static chain
8052 @kindex --plt-static-chain
8053 @kindex --no-plt-static-chain
8054 @item --plt-static-chain
8055 @itemx --no-plt-static-chain
8056 Use these options to control whether PLT call stubs load the static
8057 chain pointer (r11). @code{ld} defaults to not loading the static
8058 chain since there is never any need to do so on a PLT call.
8060 @cindex PowerPC64 PLT call stub thread safety
8061 @kindex --plt-thread-safe
8062 @kindex --no-plt-thread-safe
8063 @item --plt-thread-safe
8064 @itemx --no-plt-thread-safe
8065 With power7's weakly ordered memory model, it is possible when using
8066 lazy binding for ld.so to update a plt entry in one thread and have
8067 another thread see the individual plt entry words update in the wrong
8068 order, despite ld.so carefully writing in the correct order and using
8069 memory write barriers. To avoid this we need some sort of read
8070 barrier in the call stub, or use LD_BIND_NOW=1. By default, @code{ld}
8071 looks for calls to commonly used functions that create threads, and if
8072 seen, adds the necessary barriers. Use these options to change the
8075 @cindex PowerPC64 ELFv2 PLT localentry optimization
8076 @kindex --plt-localentry
8077 @kindex --no-plt-localentry
8078 @item --plt-localentry
8079 @itemx --no-localentry
8080 ELFv2 functions with localentry:0 are those with a single entry point,
8081 ie. global entry == local entry, and that have no requirement on r2
8082 (the TOC/GOT pointer) or r12, and guarantee r2 is unchanged on return.
8083 Such an external function can be called via the PLT without saving r2
8084 or restoring it on return, avoiding a common load-hit-store for small
8085 functions. The optimization is attractive, with up to 40% reduction
8086 in execution time for a small function, but can result in symbol
8087 interposition failures. Also, minor changes in a shared library,
8088 including system libraries, can cause a function that was localentry:0
8089 to become localentry:8. This will result in a dynamic loader
8090 complaint and failure to run. The option is experimental, use with
8091 care. @option{--no-plt-localentry} is the default.
8093 @cindex PowerPC64 Power10 stubs
8094 @kindex --power10-stubs
8095 @kindex --no-power10-stubs
8096 @item --power10-stubs
8097 @itemx --no-power10-stubs
8098 When PowerPC64 @command{ld} links input object files containing
8099 relocations used on power10 prefixed instructions it normally creates
8100 linkage stubs (PLT call and long branch) using power10 instructions
8101 for @code{@@notoc} PLT calls where @code{r2} is not known. The
8102 power10 notoc stubs are smaller and faster, so are preferred for
8103 power10. @option{--power10-stubs} and @option{--no-power10-stubs}
8104 allow you to override the linker's selection of stub instructions.
8105 @option{--power10-stubs=auto} allows the user to select the default
8120 @section @command{ld} and S/390 ELF Support
8122 @cindex S/390 ELF options
8126 @kindex --s390-pgste
8128 This option marks the result file with a @code{PT_S390_PGSTE}
8129 segment. The Linux kernel is supposed to allocate 4k page tables for
8130 binaries marked that way.
8144 @section @command{ld} and SPU ELF Support
8146 @cindex SPU ELF options
8152 This option marks an executable as a PIC plugin module.
8154 @cindex SPU overlays
8155 @kindex --no-overlays
8157 Normally, @command{ld} recognizes calls to functions within overlay
8158 regions, and redirects such calls to an overlay manager via a stub.
8159 @command{ld} also provides a built-in overlay manager. This option
8160 turns off all this special overlay handling.
8162 @cindex SPU overlay stub symbols
8163 @kindex --emit-stub-syms
8164 @item --emit-stub-syms
8165 This option causes @command{ld} to label overlay stubs with a local
8166 symbol that encodes the stub type and destination.
8168 @cindex SPU extra overlay stubs
8169 @kindex --extra-overlay-stubs
8170 @item --extra-overlay-stubs
8171 This option causes @command{ld} to add overlay call stubs on all
8172 function calls out of overlay regions. Normally stubs are not added
8173 on calls to non-overlay regions.
8175 @cindex SPU local store size
8176 @kindex --local-store=lo:hi
8177 @item --local-store=lo:hi
8178 @command{ld} usually checks that a final executable for SPU fits in
8179 the address range 0 to 256k. This option may be used to change the
8180 range. Disable the check entirely with @option{--local-store=0:0}.
8183 @kindex --stack-analysis
8184 @item --stack-analysis
8185 SPU local store space is limited. Over-allocation of stack space
8186 unnecessarily limits space available for code and data, while
8187 under-allocation results in runtime failures. If given this option,
8188 @command{ld} will provide an estimate of maximum stack usage.
8189 @command{ld} does this by examining symbols in code sections to
8190 determine the extents of functions, and looking at function prologues
8191 for stack adjusting instructions. A call-graph is created by looking
8192 for relocations on branch instructions. The graph is then searched
8193 for the maximum stack usage path. Note that this analysis does not
8194 find calls made via function pointers, and does not handle recursion
8195 and other cycles in the call graph. Stack usage may be
8196 under-estimated if your code makes such calls. Also, stack usage for
8197 dynamic allocation, e.g. alloca, will not be detected. If a link map
8198 is requested, detailed information about each function's stack usage
8199 and calls will be given.
8202 @kindex --emit-stack-syms
8203 @item --emit-stack-syms
8204 This option, if given along with @option{--stack-analysis} will result
8205 in @command{ld} emitting stack sizing symbols for each function.
8206 These take the form @code{__stack_<function_name>} for global
8207 functions, and @code{__stack_<number>_<function_name>} for static
8208 functions. @code{<number>} is the section id in hex. The value of
8209 such symbols is the stack requirement for the corresponding function.
8210 The symbol size will be zero, type @code{STT_NOTYPE}, binding
8211 @code{STB_LOCAL}, and section @code{SHN_ABS}.
8225 @section @command{ld}'s Support for Various TI COFF Versions
8226 @cindex TI COFF versions
8227 @kindex --format=@var{version}
8228 The @samp{--format} switch allows selection of one of the various
8229 TI COFF versions. The latest of this writing is 2; versions 0 and 1 are
8230 also supported. The TI COFF versions also vary in header byte-order
8231 format; @command{ld} will read any version or byte order, but the output
8232 header format depends on the default specified by the specific target.
8245 @section @command{ld} and WIN32 (cygwin/mingw)
8247 This section describes some of the win32 specific @command{ld} issues.
8248 See @ref{Options,,Command-line Options} for detailed description of the
8249 command-line options mentioned here.
8252 @cindex import libraries
8253 @item import libraries
8254 The standard Windows linker creates and uses so-called import
8255 libraries, which contains information for linking to dll's. They are
8256 regular static archives and are handled as any other static
8257 archive. The cygwin and mingw ports of @command{ld} have specific
8258 support for creating such libraries provided with the
8259 @samp{--out-implib} command-line option.
8261 @item exporting DLL symbols
8262 @cindex exporting DLL symbols
8263 The cygwin/mingw @command{ld} has several ways to export symbols for dll's.
8266 @item using auto-export functionality
8267 @cindex using auto-export functionality
8268 By default @command{ld} exports symbols with the auto-export functionality,
8269 which is controlled by the following command-line options:
8272 @item --export-all-symbols [This is the default]
8273 @item --exclude-symbols
8274 @item --exclude-libs
8275 @item --exclude-modules-for-implib
8276 @item --version-script
8279 When auto-export is in operation, @command{ld} will export all the non-local
8280 (global and common) symbols it finds in a DLL, with the exception of a few
8281 symbols known to belong to the system's runtime and libraries. As it will
8282 often not be desirable to export all of a DLL's symbols, which may include
8283 private functions that are not part of any public interface, the command-line
8284 options listed above may be used to filter symbols out from the list for
8285 exporting. The @samp{--output-def} option can be used in order to see the
8286 final list of exported symbols with all exclusions taken into effect.
8288 If @samp{--export-all-symbols} is not given explicitly on the
8289 command line, then the default auto-export behavior will be @emph{disabled}
8290 if either of the following are true:
8293 @item A DEF file is used.
8294 @item Any symbol in any object file was marked with the __declspec(dllexport) attribute.
8297 @item using a DEF file
8298 @cindex using a DEF file
8299 Another way of exporting symbols is using a DEF file. A DEF file is
8300 an ASCII file containing definitions of symbols which should be
8301 exported when a dll is created. Usually it is named @samp{<dll
8302 name>.def} and is added as any other object file to the linker's
8303 command line. The file's name must end in @samp{.def} or @samp{.DEF}.
8306 gcc -o <output> <objectfiles> <dll name>.def
8309 Using a DEF file turns off the normal auto-export behavior, unless the
8310 @samp{--export-all-symbols} option is also used.
8312 Here is an example of a DEF file for a shared library called @samp{xyz.dll}:
8315 LIBRARY "xyz.dll" BASE=0x20000000
8321 another_foo = abc.dll.afoo
8327 This example defines a DLL with a non-default base address and seven
8328 symbols in the export table. The third exported symbol @code{_bar} is an
8329 alias for the second. The fourth symbol, @code{another_foo} is resolved
8330 by "forwarding" to another module and treating it as an alias for
8331 @code{afoo} exported from the DLL @samp{abc.dll}. The final symbol
8332 @code{var1} is declared to be a data object. The @samp{doo} symbol in
8333 export library is an alias of @samp{foo}, which gets the string name
8334 in export table @samp{foo2}. The @samp{eoo} symbol is an data export
8335 symbol, which gets in export table the name @samp{var1}.
8337 The optional @code{LIBRARY <name>} command indicates the @emph{internal}
8338 name of the output DLL. If @samp{<name>} does not include a suffix,
8339 the default library suffix, @samp{.DLL} is appended.
8341 When the .DEF file is used to build an application, rather than a
8342 library, the @code{NAME <name>} command should be used instead of
8343 @code{LIBRARY}. If @samp{<name>} does not include a suffix, the default
8344 executable suffix, @samp{.EXE} is appended.
8346 With either @code{LIBRARY <name>} or @code{NAME <name>} the optional
8347 specification @code{BASE = <number>} may be used to specify a
8348 non-default base address for the image.
8350 If neither @code{LIBRARY <name>} nor @code{NAME <name>} is specified,
8351 or they specify an empty string, the internal name is the same as the
8352 filename specified on the command line.
8354 The complete specification of an export symbol is:
8358 ( ( ( <name1> [ = <name2> ] )
8359 | ( <name1> = <module-name> . <external-name>))
8360 [ @@ <integer> ] [NONAME] [DATA] [CONSTANT] [PRIVATE] [== <name3>] ) *
8363 Declares @samp{<name1>} as an exported symbol from the DLL, or declares
8364 @samp{<name1>} as an exported alias for @samp{<name2>}; or declares
8365 @samp{<name1>} as a "forward" alias for the symbol
8366 @samp{<external-name>} in the DLL @samp{<module-name>}.
8367 Optionally, the symbol may be exported by the specified ordinal
8368 @samp{<integer>} alias. The optional @samp{<name3>} is the to be used
8369 string in import/export table for the symbol.
8371 The optional keywords that follow the declaration indicate:
8373 @code{NONAME}: Do not put the symbol name in the DLL's export table. It
8374 will still be exported by its ordinal alias (either the value specified
8375 by the .def specification or, otherwise, the value assigned by the
8376 linker). The symbol name, however, does remain visible in the import
8377 library (if any), unless @code{PRIVATE} is also specified.
8379 @code{DATA}: The symbol is a variable or object, rather than a function.
8380 The import lib will export only an indirect reference to @code{foo} as
8381 the symbol @code{_imp__foo} (ie, @code{foo} must be resolved as
8384 @code{CONSTANT}: Like @code{DATA}, but put the undecorated @code{foo} as
8385 well as @code{_imp__foo} into the import library. Both refer to the
8386 read-only import address table's pointer to the variable, not to the
8387 variable itself. This can be dangerous. If the user code fails to add
8388 the @code{dllimport} attribute and also fails to explicitly add the
8389 extra indirection that the use of the attribute enforces, the
8390 application will behave unexpectedly.
8392 @code{PRIVATE}: Put the symbol in the DLL's export table, but do not put
8393 it into the static import library used to resolve imports at link time. The
8394 symbol can still be imported using the @code{LoadLibrary/GetProcAddress}
8395 API at runtime or by using the GNU ld extension of linking directly to
8396 the DLL without an import library.
8398 See ld/deffilep.y in the binutils sources for the full specification of
8399 other DEF file statements
8401 @cindex creating a DEF file
8402 While linking a shared dll, @command{ld} is able to create a DEF file
8403 with the @samp{--output-def <file>} command-line option.
8405 @item Using decorations
8406 @cindex Using decorations
8407 Another way of marking symbols for export is to modify the source code
8408 itself, so that when building the DLL each symbol to be exported is
8412 __declspec(dllexport) int a_variable
8413 __declspec(dllexport) void a_function(int with_args)
8416 All such symbols will be exported from the DLL. If, however,
8417 any of the object files in the DLL contain symbols decorated in
8418 this way, then the normal auto-export behavior is disabled, unless
8419 the @samp{--export-all-symbols} option is also used.
8421 Note that object files that wish to access these symbols must @emph{not}
8422 decorate them with dllexport. Instead, they should use dllimport,
8426 __declspec(dllimport) int a_variable
8427 __declspec(dllimport) void a_function(int with_args)
8430 This complicates the structure of library header files, because
8431 when included by the library itself the header must declare the
8432 variables and functions as dllexport, but when included by client
8433 code the header must declare them as dllimport. There are a number
8434 of idioms that are typically used to do this; often client code can
8435 omit the __declspec() declaration completely. See
8436 @samp{--enable-auto-import} and @samp{automatic data imports} for more
8440 @cindex automatic data imports
8441 @item automatic data imports
8442 The standard Windows dll format supports data imports from dlls only
8443 by adding special decorations (dllimport/dllexport), which let the
8444 compiler produce specific assembler instructions to deal with this
8445 issue. This increases the effort necessary to port existing Un*x
8446 code to these platforms, especially for large
8447 c++ libraries and applications. The auto-import feature, which was
8448 initially provided by Paul Sokolovsky, allows one to omit the
8449 decorations to achieve a behavior that conforms to that on POSIX/Un*x
8450 platforms. This feature is enabled with the @samp{--enable-auto-import}
8451 command-line option, although it is enabled by default on cygwin/mingw.
8452 The @samp{--enable-auto-import} option itself now serves mainly to
8453 suppress any warnings that are ordinarily emitted when linked objects
8454 trigger the feature's use.
8456 auto-import of variables does not always work flawlessly without
8457 additional assistance. Sometimes, you will see this message
8459 "variable '<var>' can't be auto-imported. Please read the
8460 documentation for ld's @code{--enable-auto-import} for details."
8462 The @samp{--enable-auto-import} documentation explains why this error
8463 occurs, and several methods that can be used to overcome this difficulty.
8464 One of these methods is the @emph{runtime pseudo-relocs} feature, described
8467 @cindex runtime pseudo-relocation
8468 For complex variables imported from DLLs (such as structs or classes),
8469 object files typically contain a base address for the variable and an
8470 offset (@emph{addend}) within the variable--to specify a particular
8471 field or public member, for instance. Unfortunately, the runtime loader used
8472 in win32 environments is incapable of fixing these references at runtime
8473 without the additional information supplied by dllimport/dllexport decorations.
8474 The standard auto-import feature described above is unable to resolve these
8477 The @samp{--enable-runtime-pseudo-relocs} switch allows these references to
8478 be resolved without error, while leaving the task of adjusting the references
8479 themselves (with their non-zero addends) to specialized code provided by the
8480 runtime environment. Recent versions of the cygwin and mingw environments and
8481 compilers provide this runtime support; older versions do not. However, the
8482 support is only necessary on the developer's platform; the compiled result will
8483 run without error on an older system.
8485 @samp{--enable-runtime-pseudo-relocs} is not the default; it must be explicitly
8488 @cindex direct linking to a dll
8489 @item direct linking to a dll
8490 The cygwin/mingw ports of @command{ld} support the direct linking,
8491 including data symbols, to a dll without the usage of any import
8492 libraries. This is much faster and uses much less memory than does the
8493 traditional import library method, especially when linking large
8494 libraries or applications. When @command{ld} creates an import lib, each
8495 function or variable exported from the dll is stored in its own bfd, even
8496 though a single bfd could contain many exports. The overhead involved in
8497 storing, loading, and processing so many bfd's is quite large, and explains the
8498 tremendous time, memory, and storage needed to link against particularly
8499 large or complex libraries when using import libs.
8501 Linking directly to a dll uses no extra command-line switches other than
8502 @samp{-L} and @samp{-l}, because @command{ld} already searches for a number
8503 of names to match each library. All that is needed from the developer's
8504 perspective is an understanding of this search, in order to force ld to
8505 select the dll instead of an import library.
8508 For instance, when ld is called with the argument @samp{-lxxx} it will attempt
8509 to find, in the first directory of its search path,
8522 before moving on to the next directory in the search path.
8524 (*) Actually, this is not @samp{cygxxx.dll} but in fact is @samp{<prefix>xxx.dll},
8525 where @samp{<prefix>} is set by the @command{ld} option
8526 @samp{--dll-search-prefix=<prefix>}. In the case of cygwin, the standard gcc spec
8527 file includes @samp{--dll-search-prefix=cyg}, so in effect we actually search for
8530 Other win32-based unix environments, such as mingw or pw32, may use other
8531 @samp{<prefix>}es, although at present only cygwin makes use of this feature. It
8532 was originally intended to help avoid name conflicts among dll's built for the
8533 various win32/un*x environments, so that (for example) two versions of a zlib dll
8534 could coexist on the same machine.
8536 The generic cygwin/mingw path layout uses a @samp{bin} directory for
8537 applications and dll's and a @samp{lib} directory for the import
8538 libraries (using cygwin nomenclature):
8544 libxxx.dll.a (in case of dll's)
8545 libxxx.a (in case of static archive)
8548 Linking directly to a dll without using the import library can be
8551 1. Use the dll directly by adding the @samp{bin} path to the link line
8553 gcc -Wl,-verbose -o a.exe -L../bin/ -lxxx
8556 However, as the dll's often have version numbers appended to their names
8557 (@samp{cygncurses-5.dll}) this will often fail, unless one specifies
8558 @samp{-L../bin -lncurses-5} to include the version. Import libs are generally
8559 not versioned, and do not have this difficulty.
8561 2. Create a symbolic link from the dll to a file in the @samp{lib}
8562 directory according to the above mentioned search pattern. This
8563 should be used to avoid unwanted changes in the tools needed for
8567 ln -s bin/cygxxx.dll lib/[cyg|lib|]xxx.dll[.a]
8570 Then you can link without any make environment changes.
8573 gcc -Wl,-verbose -o a.exe -L../lib/ -lxxx
8576 This technique also avoids the version number problems, because the following is
8583 libxxx.dll.a -> ../bin/cygxxx-5.dll
8586 Linking directly to a dll without using an import lib will work
8587 even when auto-import features are exercised, and even when
8588 @samp{--enable-runtime-pseudo-relocs} is used.
8590 Given the improvements in speed and memory usage, one might justifiably
8591 wonder why import libraries are used at all. There are three reasons:
8593 1. Until recently, the link-directly-to-dll functionality did @emph{not}
8594 work with auto-imported data.
8596 2. Sometimes it is necessary to include pure static objects within the
8597 import library (which otherwise contains only bfd's for indirection
8598 symbols that point to the exports of a dll). Again, the import lib
8599 for the cygwin kernel makes use of this ability, and it is not
8600 possible to do this without an import lib.
8602 3. Symbol aliases can only be resolved using an import lib. This is
8603 critical when linking against OS-supplied dll's (eg, the win32 API)
8604 in which symbols are usually exported as undecorated aliases of their
8605 stdcall-decorated assembly names.
8607 So, import libs are not going away. But the ability to replace
8608 true import libs with a simple symbolic link to (or a copy of)
8609 a dll, in many cases, is a useful addition to the suite of tools
8610 binutils makes available to the win32 developer. Given the
8611 massive improvements in memory requirements during linking, storage
8612 requirements, and linking speed, we expect that many developers
8613 will soon begin to use this feature whenever possible.
8615 @item symbol aliasing
8617 @item adding additional names
8618 Sometimes, it is useful to export symbols with additional names.
8619 A symbol @samp{foo} will be exported as @samp{foo}, but it can also be
8620 exported as @samp{_foo} by using special directives in the DEF file
8621 when creating the dll. This will affect also the optional created
8622 import library. Consider the following DEF file:
8625 LIBRARY "xyz.dll" BASE=0x61000000
8632 The line @samp{_foo = foo} maps the symbol @samp{foo} to @samp{_foo}.
8634 Another method for creating a symbol alias is to create it in the
8635 source code using the "weak" attribute:
8638 void foo () @{ /* Do something. */; @}
8639 void _foo () __attribute__ ((weak, alias ("foo")));
8642 See the gcc manual for more information about attributes and weak
8645 @item renaming symbols
8646 Sometimes it is useful to rename exports. For instance, the cygwin
8647 kernel does this regularly. A symbol @samp{_foo} can be exported as
8648 @samp{foo} but not as @samp{_foo} by using special directives in the
8649 DEF file. (This will also affect the import library, if it is
8650 created). In the following example:
8653 LIBRARY "xyz.dll" BASE=0x61000000
8659 The line @samp{_foo = foo} maps the exported symbol @samp{foo} to
8663 Note: using a DEF file disables the default auto-export behavior,
8664 unless the @samp{--export-all-symbols} command-line option is used.
8665 If, however, you are trying to rename symbols, then you should list
8666 @emph{all} desired exports in the DEF file, including the symbols
8667 that are not being renamed, and do @emph{not} use the
8668 @samp{--export-all-symbols} option. If you list only the
8669 renamed symbols in the DEF file, and use @samp{--export-all-symbols}
8670 to handle the other symbols, then the both the new names @emph{and}
8671 the original names for the renamed symbols will be exported.
8672 In effect, you'd be aliasing those symbols, not renaming them,
8673 which is probably not what you wanted.
8675 @cindex weak externals
8676 @item weak externals
8677 The Windows object format, PE, specifies a form of weak symbols called
8678 weak externals. When a weak symbol is linked and the symbol is not
8679 defined, the weak symbol becomes an alias for some other symbol. There
8680 are three variants of weak externals:
8682 @item Definition is searched for in objects and libraries, historically
8683 called lazy externals.
8684 @item Definition is searched for only in other objects, not in libraries.
8685 This form is not presently implemented.
8686 @item No search; the symbol is an alias. This form is not presently
8689 As a GNU extension, weak symbols that do not specify an alternate symbol
8690 are supported. If the symbol is undefined when linking, the symbol
8691 uses a default value.
8693 @cindex aligned common symbols
8694 @item aligned common symbols
8695 As a GNU extension to the PE file format, it is possible to specify the
8696 desired alignment for a common symbol. This information is conveyed from
8697 the assembler or compiler to the linker by means of GNU-specific commands
8698 carried in the object file's @samp{.drectve} section, which are recognized
8699 by @command{ld} and respected when laying out the common symbols. Native
8700 tools will be able to process object files employing this GNU extension,
8701 but will fail to respect the alignment instructions, and may issue noisy
8702 warnings about unknown linker directives.
8717 @section @code{ld} and Xtensa Processors
8719 @cindex Xtensa processors
8720 The default @command{ld} behavior for Xtensa processors is to interpret
8721 @code{SECTIONS} commands so that lists of explicitly named sections in a
8722 specification with a wildcard file will be interleaved when necessary to
8723 keep literal pools within the range of PC-relative load offsets. For
8724 example, with the command:
8736 @command{ld} may interleave some of the @code{.literal}
8737 and @code{.text} sections from different object files to ensure that the
8738 literal pools are within the range of PC-relative load offsets. A valid
8739 interleaving might place the @code{.literal} sections from an initial
8740 group of files followed by the @code{.text} sections of that group of
8741 files. Then, the @code{.literal} sections from the rest of the files
8742 and the @code{.text} sections from the rest of the files would follow.
8744 @cindex @option{--relax} on Xtensa
8745 @cindex relaxing on Xtensa
8746 Relaxation is enabled by default for the Xtensa version of @command{ld} and
8747 provides two important link-time optimizations. The first optimization
8748 is to combine identical literal values to reduce code size. A redundant
8749 literal will be removed and all the @code{L32R} instructions that use it
8750 will be changed to reference an identical literal, as long as the
8751 location of the replacement literal is within the offset range of all
8752 the @code{L32R} instructions. The second optimization is to remove
8753 unnecessary overhead from assembler-generated ``longcall'' sequences of
8754 @code{L32R}/@code{CALLX@var{n}} when the target functions are within
8755 range of direct @code{CALL@var{n}} instructions.
8757 For each of these cases where an indirect call sequence can be optimized
8758 to a direct call, the linker will change the @code{CALLX@var{n}}
8759 instruction to a @code{CALL@var{n}} instruction, remove the @code{L32R}
8760 instruction, and remove the literal referenced by the @code{L32R}
8761 instruction if it is not used for anything else. Removing the
8762 @code{L32R} instruction always reduces code size but can potentially
8763 hurt performance by changing the alignment of subsequent branch targets.
8764 By default, the linker will always preserve alignments, either by
8765 switching some instructions between 24-bit encodings and the equivalent
8766 density instructions or by inserting a no-op in place of the @code{L32R}
8767 instruction that was removed. If code size is more important than
8768 performance, the @option{--size-opt} option can be used to prevent the
8769 linker from widening density instructions or inserting no-ops, except in
8770 a few cases where no-ops are required for correctness.
8772 The following Xtensa-specific command-line options can be used to
8775 @cindex Xtensa options
8778 When optimizing indirect calls to direct calls, optimize for code size
8779 more than performance. With this option, the linker will not insert
8780 no-ops or widen density instructions to preserve branch target
8781 alignment. There may still be some cases where no-ops are required to
8782 preserve the correctness of the code.
8784 @item --abi-windowed
8786 Choose ABI for the output object and for the generated PLT code.
8787 PLT code inserted by the linker must match ABI of the output object
8788 because windowed and call0 ABI use incompatible function call
8790 Default ABI is chosen by the ABI tag in the @code{.xtensa.info} section
8791 of the first input object.
8792 A warning is issued if ABI tags of input objects do not match each other
8793 or the chosen output object ABI.
8801 @ifclear SingleFormat
8806 @cindex object file management
8807 @cindex object formats available
8809 The linker accesses object and archive files using the BFD libraries.
8810 These libraries allow the linker to use the same routines to operate on
8811 object files whatever the object file format. A different object file
8812 format can be supported simply by creating a new BFD back end and adding
8813 it to the library. To conserve runtime memory, however, the linker and
8814 associated tools are usually configured to support only a subset of the
8815 object file formats available. You can use @code{objdump -i}
8816 (@pxref{objdump,,objdump,binutils.info,The GNU Binary Utilities}) to
8817 list all the formats available for your configuration.
8819 @cindex BFD requirements
8820 @cindex requirements for BFD
8821 As with most implementations, BFD is a compromise between
8822 several conflicting requirements. The major factor influencing
8823 BFD design was efficiency: any time used converting between
8824 formats is time which would not have been spent had BFD not
8825 been involved. This is partly offset by abstraction payback; since
8826 BFD simplifies applications and back ends, more time and care
8827 may be spent optimizing algorithms for a greater speed.
8829 One minor artifact of the BFD solution which you should bear in
8830 mind is the potential for information loss. There are two places where
8831 useful information can be lost using the BFD mechanism: during
8832 conversion and during output. @xref{BFD information loss}.
8835 * BFD outline:: How it works: an outline of BFD
8839 @section How It Works: An Outline of BFD
8840 @cindex opening object files
8841 @include bfdsumm.texi
8844 @node Reporting Bugs
8845 @chapter Reporting Bugs
8846 @cindex bugs in @command{ld}
8847 @cindex reporting bugs in @command{ld}
8849 Your bug reports play an essential role in making @command{ld} reliable.
8851 Reporting a bug may help you by bringing a solution to your problem, or
8852 it may not. But in any case the principal function of a bug report is
8853 to help the entire community by making the next version of @command{ld}
8854 work better. Bug reports are your contribution to the maintenance of
8857 In order for a bug report to serve its purpose, you must include the
8858 information that enables us to fix the bug.
8861 * Bug Criteria:: Have you found a bug?
8862 * Bug Reporting:: How to report bugs
8866 @section Have You Found a Bug?
8867 @cindex bug criteria
8869 If you are not sure whether you have found a bug, here are some guidelines:
8872 @cindex fatal signal
8873 @cindex linker crash
8874 @cindex crash of linker
8876 If the linker gets a fatal signal, for any input whatever, that is a
8877 @command{ld} bug. Reliable linkers never crash.
8879 @cindex error on valid input
8881 If @command{ld} produces an error message for valid input, that is a bug.
8883 @cindex invalid input
8885 If @command{ld} does not produce an error message for invalid input, that
8886 may be a bug. In the general case, the linker can not verify that
8887 object files are correct.
8890 If you are an experienced user of linkers, your suggestions for
8891 improvement of @command{ld} are welcome in any case.
8895 @section How to Report Bugs
8897 @cindex @command{ld} bugs, reporting
8899 A number of companies and individuals offer support for @sc{gnu}
8900 products. If you obtained @command{ld} from a support organization, we
8901 recommend you contact that organization first.
8903 You can find contact information for many support companies and
8904 individuals in the file @file{etc/SERVICE} in the @sc{gnu} Emacs
8908 Otherwise, send bug reports for @command{ld} to
8912 The fundamental principle of reporting bugs usefully is this:
8913 @strong{report all the facts}. If you are not sure whether to state a
8914 fact or leave it out, state it!
8916 Often people omit facts because they think they know what causes the
8917 problem and assume that some details do not matter. Thus, you might
8918 assume that the name of a symbol you use in an example does not
8919 matter. Well, probably it does not, but one cannot be sure. Perhaps
8920 the bug is a stray memory reference which happens to fetch from the
8921 location where that name is stored in memory; perhaps, if the name
8922 were different, the contents of that location would fool the linker
8923 into doing the right thing despite the bug. Play it safe and give a
8924 specific, complete example. That is the easiest thing for you to do,
8925 and the most helpful.
8927 Keep in mind that the purpose of a bug report is to enable us to fix
8928 the bug if it is new to us. Therefore, always write your bug reports
8929 on the assumption that the bug has not been reported previously.
8931 Sometimes people give a few sketchy facts and ask, ``Does this ring a
8932 bell?'' This cannot help us fix a bug, so it is basically useless. We
8933 respond by asking for enough details to enable us to investigate.
8934 You might as well expedite matters by sending them to begin with.
8936 To enable us to fix the bug, you should include all these things:
8940 The version of @command{ld}. @command{ld} announces it if you start it with
8941 the @samp{--version} argument.
8943 Without this, we will not know whether there is any point in looking for
8944 the bug in the current version of @command{ld}.
8947 Any patches you may have applied to the @command{ld} source, including any
8948 patches made to the @code{BFD} library.
8951 The type of machine you are using, and the operating system name and
8955 What compiler (and its version) was used to compile @command{ld}---e.g.
8959 The command arguments you gave the linker to link your example and
8960 observe the bug. To guarantee you will not omit something important,
8961 list them all. A copy of the Makefile (or the output from make) is
8964 If we were to try to guess the arguments, we would probably guess wrong
8965 and then we might not encounter the bug.
8968 A complete input file, or set of input files, that will reproduce the
8969 bug. It is generally most helpful to send the actual object files
8970 provided that they are reasonably small. Say no more than 10K. For
8971 bigger files you can either make them available by FTP or HTTP or else
8972 state that you are willing to send the object file(s) to whomever
8973 requests them. (Note - your email will be going to a mailing list, so
8974 we do not want to clog it up with large attachments). But small
8975 attachments are best.
8977 If the source files were assembled using @code{gas} or compiled using
8978 @code{gcc}, then it may be OK to send the source files rather than the
8979 object files. In this case, be sure to say exactly what version of
8980 @code{gas} or @code{gcc} was used to produce the object files. Also say
8981 how @code{gas} or @code{gcc} were configured.
8984 A description of what behavior you observe that you believe is
8985 incorrect. For example, ``It gets a fatal signal.''
8987 Of course, if the bug is that @command{ld} gets a fatal signal, then we
8988 will certainly notice it. But if the bug is incorrect output, we might
8989 not notice unless it is glaringly wrong. You might as well not give us
8990 a chance to make a mistake.
8992 Even if the problem you experience is a fatal signal, you should still
8993 say so explicitly. Suppose something strange is going on, such as, your
8994 copy of @command{ld} is out of sync, or you have encountered a bug in the
8995 C library on your system. (This has happened!) Your copy might crash
8996 and ours would not. If you told us to expect a crash, then when ours
8997 fails to crash, we would know that the bug was not happening for us. If
8998 you had not told us to expect a crash, then we would not be able to draw
8999 any conclusion from our observations.
9002 If you wish to suggest changes to the @command{ld} source, send us context
9003 diffs, as generated by @code{diff} with the @samp{-u}, @samp{-c}, or
9004 @samp{-p} option. Always send diffs from the old file to the new file.
9005 If you even discuss something in the @command{ld} source, refer to it by
9006 context, not by line number.
9008 The line numbers in our development sources will not match those in your
9009 sources. Your line numbers would convey no useful information to us.
9012 Here are some things that are not necessary:
9016 A description of the envelope of the bug.
9018 Often people who encounter a bug spend a lot of time investigating
9019 which changes to the input file will make the bug go away and which
9020 changes will not affect it.
9022 This is often time consuming and not very useful, because the way we
9023 will find the bug is by running a single example under the debugger
9024 with breakpoints, not by pure deduction from a series of examples.
9025 We recommend that you save your time for something else.
9027 Of course, if you can find a simpler example to report @emph{instead}
9028 of the original one, that is a convenience for us. Errors in the
9029 output will be easier to spot, running under the debugger will take
9030 less time, and so on.
9032 However, simplification is not vital; if you do not want to do this,
9033 report the bug anyway and send us the entire test case you used.
9036 A patch for the bug.
9038 A patch for the bug does help us if it is a good one. But do not omit
9039 the necessary information, such as the test case, on the assumption that
9040 a patch is all we need. We might see problems with your patch and decide
9041 to fix the problem another way, or we might not understand it at all.
9043 Sometimes with a program as complicated as @command{ld} it is very hard to
9044 construct an example that will make the program follow a certain path
9045 through the code. If you do not send us the example, we will not be
9046 able to construct one, so we will not be able to verify that the bug is
9049 And if we cannot understand what bug you are trying to fix, or why your
9050 patch should be an improvement, we will not install it. A test case will
9051 help us to understand.
9054 A guess about what the bug is or what it depends on.
9056 Such guesses are usually wrong. Even we cannot guess right about such
9057 things without first using the debugger to find the facts.
9061 @appendix MRI Compatible Script Files
9062 @cindex MRI compatibility
9063 To aid users making the transition to @sc{gnu} @command{ld} from the MRI
9064 linker, @command{ld} can use MRI compatible linker scripts as an
9065 alternative to the more general-purpose linker scripting language
9066 described in @ref{Scripts}. MRI compatible linker scripts have a much
9067 simpler command set than the scripting language otherwise used with
9068 @command{ld}. @sc{gnu} @command{ld} supports the most commonly used MRI
9069 linker commands; these commands are described here.
9071 In general, MRI scripts aren't of much use with the @code{a.out} object
9072 file format, since it only has three sections and MRI scripts lack some
9073 features to make use of them.
9075 You can specify a file containing an MRI-compatible script using the
9076 @samp{-c} command-line option.
9078 Each command in an MRI-compatible script occupies its own line; each
9079 command line starts with the keyword that identifies the command (though
9080 blank lines are also allowed for punctuation). If a line of an
9081 MRI-compatible script begins with an unrecognized keyword, @command{ld}
9082 issues a warning message, but continues processing the script.
9084 Lines beginning with @samp{*} are comments.
9086 You can write these commands using all upper-case letters, or all
9087 lower case; for example, @samp{chip} is the same as @samp{CHIP}.
9088 The following list shows only the upper-case form of each command.
9091 @cindex @code{ABSOLUTE} (MRI)
9092 @item ABSOLUTE @var{secname}
9093 @itemx ABSOLUTE @var{secname}, @var{secname}, @dots{} @var{secname}
9094 Normally, @command{ld} includes in the output file all sections from all
9095 the input files. However, in an MRI-compatible script, you can use the
9096 @code{ABSOLUTE} command to restrict the sections that will be present in
9097 your output program. If the @code{ABSOLUTE} command is used at all in a
9098 script, then only the sections named explicitly in @code{ABSOLUTE}
9099 commands will appear in the linker output. You can still use other
9100 input sections (whatever you select on the command line, or using
9101 @code{LOAD}) to resolve addresses in the output file.
9103 @cindex @code{ALIAS} (MRI)
9104 @item ALIAS @var{out-secname}, @var{in-secname}
9105 Use this command to place the data from input section @var{in-secname}
9106 in a section called @var{out-secname} in the linker output file.
9108 @var{in-secname} may be an integer.
9110 @cindex @code{ALIGN} (MRI)
9111 @item ALIGN @var{secname} = @var{expression}
9112 Align the section called @var{secname} to @var{expression}. The
9113 @var{expression} should be a power of two.
9115 @cindex @code{BASE} (MRI)
9116 @item BASE @var{expression}
9117 Use the value of @var{expression} as the lowest address (other than
9118 absolute addresses) in the output file.
9120 @cindex @code{CHIP} (MRI)
9121 @item CHIP @var{expression}
9122 @itemx CHIP @var{expression}, @var{expression}
9123 This command does nothing; it is accepted only for compatibility.
9125 @cindex @code{END} (MRI)
9127 This command does nothing whatever; it's only accepted for compatibility.
9129 @cindex @code{FORMAT} (MRI)
9130 @item FORMAT @var{output-format}
9131 Similar to the @code{OUTPUT_FORMAT} command in the more general linker
9132 language, but restricted to S-records, if @var{output-format} is @samp{S}
9134 @cindex @code{LIST} (MRI)
9135 @item LIST @var{anything}@dots{}
9136 Print (to the standard output file) a link map, as produced by the
9137 @command{ld} command-line option @samp{-M}.
9139 The keyword @code{LIST} may be followed by anything on the
9140 same line, with no change in its effect.
9142 @cindex @code{LOAD} (MRI)
9143 @item LOAD @var{filename}
9144 @itemx LOAD @var{filename}, @var{filename}, @dots{} @var{filename}
9145 Include one or more object file @var{filename} in the link; this has the
9146 same effect as specifying @var{filename} directly on the @command{ld}
9149 @cindex @code{NAME} (MRI)
9150 @item NAME @var{output-name}
9151 @var{output-name} is the name for the program produced by @command{ld}; the
9152 MRI-compatible command @code{NAME} is equivalent to the command-line
9153 option @samp{-o} or the general script language command @code{OUTPUT}.
9155 @cindex @code{ORDER} (MRI)
9156 @item ORDER @var{secname}, @var{secname}, @dots{} @var{secname}
9157 @itemx ORDER @var{secname} @var{secname} @var{secname}
9158 Normally, @command{ld} orders the sections in its output file in the
9159 order in which they first appear in the input files. In an MRI-compatible
9160 script, you can override this ordering with the @code{ORDER} command. The
9161 sections you list with @code{ORDER} will appear first in your output
9162 file, in the order specified.
9164 @cindex @code{PUBLIC} (MRI)
9165 @item PUBLIC @var{name}=@var{expression}
9166 @itemx PUBLIC @var{name},@var{expression}
9167 @itemx PUBLIC @var{name} @var{expression}
9168 Supply a value (@var{expression}) for external symbol
9169 @var{name} used in the linker input files.
9171 @cindex @code{SECT} (MRI)
9172 @item SECT @var{secname}, @var{expression}
9173 @itemx SECT @var{secname}=@var{expression}
9174 @itemx SECT @var{secname} @var{expression}
9175 You can use any of these three forms of the @code{SECT} command to
9176 specify the start address (@var{expression}) for section @var{secname}.
9177 If you have more than one @code{SECT} statement for the same
9178 @var{secname}, only the @emph{first} sets the start address.
9181 @node GNU Free Documentation License
9182 @appendix GNU Free Documentation License
9186 @unnumbered LD Index
9191 % I think something like @@colophon should be in texinfo. In the
9193 \long\def\colophon{\hbox to0pt{}\vfill
9194 \centerline{The body of this manual is set in}
9195 \centerline{\fontname\tenrm,}
9196 \centerline{with headings in {\bf\fontname\tenbf}}
9197 \centerline{and examples in {\tt\fontname\tentt}.}
9198 \centerline{{\it\fontname\tenit\/} and}
9199 \centerline{{\sl\fontname\tensl\/}}
9200 \centerline{are used for emphasis.}\vfill}
9202 % Blame: doc@@cygnus.com, 28mar91.