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
130 * Machine Dependent:: Machine Dependent Features
134 * H8/300:: ld and the H8/300
137 * Renesas:: ld and other Renesas micros
140 * ARM:: ld and the ARM family
143 * M68HC11/68HC12:: ld and the Motorola 68HC11 and 68HC12 families
146 * HPPA ELF32:: ld and HPPA 32-bit ELF
149 * M68K:: ld and Motorola 68K family
152 * MIPS:: ld and MIPS family
155 * PowerPC ELF32:: ld and PowerPC 32-bit ELF Support
158 * PowerPC64 ELF64:: ld and PowerPC64 64-bit ELF Support
161 * S/390 ELF:: ld and S/390 ELF Support
164 * SPU ELF:: ld and SPU ELF Support
167 * TI COFF:: ld and the TI COFF
170 * Win32:: ld and WIN32 (cygwin/mingw)
173 * Xtensa:: ld and Xtensa Processors
176 @ifclear SingleFormat
179 @c Following blank line required for remaining bug in makeinfo conds/menus
181 * Reporting Bugs:: Reporting Bugs
182 * MRI:: MRI Compatible Script Files
183 * GNU Free Documentation License:: GNU Free Documentation License
184 * LD Index:: LD Index
191 @cindex @sc{gnu} linker
192 @cindex what is this?
195 @c man begin SYNOPSIS
196 ld [@b{options}] @var{objfile} @dots{}
200 ar(1), nm(1), objcopy(1), objdump(1), readelf(1) and
201 the Info entries for @file{binutils} and
206 @c man begin DESCRIPTION
208 @command{ld} combines a number of object and archive files, relocates
209 their data and ties up symbol references. Usually the last step in
210 compiling a program is to run @command{ld}.
212 @command{ld} accepts Linker Command Language files written in
213 a superset of AT&T's Link Editor Command Language syntax,
214 to provide explicit and total control over the linking process.
218 This man page does not describe the command language; see the
219 @command{ld} entry in @code{info} for full details on the command
220 language and on other aspects of the GNU linker.
223 @ifclear SingleFormat
224 This version of @command{ld} uses the general purpose BFD libraries
225 to operate on object files. This allows @command{ld} to read, combine, and
226 write object files in many different formats---for example, COFF or
227 @code{a.out}. Different formats may be linked together to produce any
228 available kind of object file. @xref{BFD}, for more information.
231 Aside from its flexibility, the @sc{gnu} linker is more helpful than other
232 linkers in providing diagnostic information. Many linkers abandon
233 execution immediately upon encountering an error; whenever possible,
234 @command{ld} continues executing, allowing you to identify other errors
235 (or, in some cases, to get an output file in spite of the error).
242 @c man begin DESCRIPTION
244 The @sc{gnu} linker @command{ld} is meant to cover a broad range of situations,
245 and to be as compatible as possible with other linkers. As a result,
246 you have many choices to control its behavior.
252 * Options:: Command-line Options
253 * Environment:: Environment Variables
257 @section Command-line Options
265 The linker supports a plethora of command-line options, but in actual
266 practice few of them are used in any particular context.
267 @cindex standard Unix system
268 For instance, a frequent use of @command{ld} is to link standard Unix
269 object files on a standard, supported Unix system. On such a system, to
270 link a file @code{hello.o}:
273 ld -o @var{output} /lib/crt0.o hello.o -lc
276 This tells @command{ld} to produce a file called @var{output} as the
277 result of linking the file @code{/lib/crt0.o} with @code{hello.o} and
278 the library @code{libc.a}, which will come from the standard search
279 directories. (See the discussion of the @samp{-l} option below.)
281 Some of the command-line options to @command{ld} may be specified at any
282 point in the command line. However, options which refer to files, such
283 as @samp{-l} or @samp{-T}, cause the file to be read at the point at
284 which the option appears in the command line, relative to the object
285 files and other file options. Repeating non-file options with a
286 different argument will either have no further effect, or override prior
287 occurrences (those further to the left on the command line) of that
288 option. Options which may be meaningfully specified more than once are
289 noted in the descriptions below.
292 Non-option arguments are object files or archives which are to be linked
293 together. They may follow, precede, or be mixed in with command-line
294 options, except that an object file argument may not be placed between
295 an option and its argument.
297 Usually the linker is invoked with at least one object file, but you can
298 specify other forms of binary input files using @samp{-l}, @samp{-R},
299 and the script command language. If @emph{no} binary input files at all
300 are specified, the linker does not produce any output, and issues the
301 message @samp{No input files}.
303 If the linker cannot recognize the format of an object file, it will
304 assume that it is a linker script. A script specified in this way
305 augments the main linker script used for the link (either the default
306 linker script or the one specified by using @samp{-T}). This feature
307 permits the linker to link against a file which appears to be an object
308 or an archive, but actually merely defines some symbol values, or uses
309 @code{INPUT} or @code{GROUP} to load other objects. Specifying a
310 script in this way merely augments the main linker script, with the
311 extra commands placed after the main script; use the @samp{-T} option
312 to replace the default linker script entirely, but note the effect of
313 the @code{INSERT} command. @xref{Scripts}.
315 For options whose names are a single letter,
316 option arguments must either follow the option letter without intervening
317 whitespace, or be given as separate arguments immediately following the
318 option that requires them.
320 For options whose names are multiple letters, either one dash or two can
321 precede the option name; for example, @samp{-trace-symbol} and
322 @samp{--trace-symbol} are equivalent. Note---there is one exception to
323 this rule. Multiple letter options that start with a lower case 'o' can
324 only be preceded by two dashes. This is to reduce confusion with the
325 @samp{-o} option. So for example @samp{-omagic} sets the output file
326 name to @samp{magic} whereas @samp{--omagic} sets the NMAGIC flag on the
329 Arguments to multiple-letter options must either be separated from the
330 option name by an equals sign, or be given as separate arguments
331 immediately following the option that requires them. For example,
332 @samp{--trace-symbol foo} and @samp{--trace-symbol=foo} are equivalent.
333 Unique abbreviations of the names of multiple-letter options are
336 Note---if the linker is being invoked indirectly, via a compiler driver
337 (e.g. @samp{gcc}) then all the linker command-line options should be
338 prefixed by @samp{-Wl,} (or whatever is appropriate for the particular
339 compiler driver) like this:
342 gcc -Wl,--start-group foo.o bar.o -Wl,--end-group
345 This is important, because otherwise the compiler driver program may
346 silently drop the linker options, resulting in a bad link. Confusion
347 may also arise when passing options that require values through a
348 driver, as the use of a space between option and argument acts as
349 a separator, and causes the driver to pass only the option to the linker
350 and the argument to the compiler. In this case, it is simplest to use
351 the joined forms of both single- and multiple-letter options, such as:
354 gcc foo.o bar.o -Wl,-eENTRY -Wl,-Map=a.map
357 Here is a table of the generic command-line switches accepted by the GNU
361 @include at-file.texi
363 @kindex -a @var{keyword}
364 @item -a @var{keyword}
365 This option is supported for HP/UX compatibility. The @var{keyword}
366 argument must be one of the strings @samp{archive}, @samp{shared}, or
367 @samp{default}. @samp{-aarchive} is functionally equivalent to
368 @samp{-Bstatic}, and the other two keywords are functionally equivalent
369 to @samp{-Bdynamic}. This option may be used any number of times.
371 @kindex --audit @var{AUDITLIB}
372 @item --audit @var{AUDITLIB}
373 Adds @var{AUDITLIB} to the @code{DT_AUDIT} entry of the dynamic section.
374 @var{AUDITLIB} is not checked for existence, nor will it use the DT_SONAME
375 specified in the library. If specified multiple times @code{DT_AUDIT}
376 will contain a colon separated list of audit interfaces to use. If the linker
377 finds an object with an audit entry while searching for shared libraries,
378 it will add a corresponding @code{DT_DEPAUDIT} entry in the output file.
379 This option is only meaningful on ELF platforms supporting the rtld-audit
382 @ifclear SingleFormat
383 @cindex binary input format
384 @kindex -b @var{format}
385 @kindex --format=@var{format}
388 @item -b @var{input-format}
389 @itemx --format=@var{input-format}
390 @command{ld} may be configured to support more than one kind of object
391 file. If your @command{ld} is configured this way, you can use the
392 @samp{-b} option to specify the binary format for input object files
393 that follow this option on the command line. Even when @command{ld} is
394 configured to support alternative object formats, you don't usually need
395 to specify this, as @command{ld} should be configured to expect as a
396 default input format the most usual format on each machine.
397 @var{input-format} is a text string, the name of a particular format
398 supported by the BFD libraries. (You can list the available binary
399 formats with @samp{objdump -i}.)
402 You may want to use this option if you are linking files with an unusual
403 binary format. You can also use @samp{-b} to switch formats explicitly (when
404 linking object files of different formats), by including
405 @samp{-b @var{input-format}} before each group of object files in a
408 The default format is taken from the environment variable
413 You can also define the input format from a script, using the command
416 see @ref{Format Commands}.
420 @kindex -c @var{MRI-cmdfile}
421 @kindex --mri-script=@var{MRI-cmdfile}
422 @cindex compatibility, MRI
423 @item -c @var{MRI-commandfile}
424 @itemx --mri-script=@var{MRI-commandfile}
425 For compatibility with linkers produced by MRI, @command{ld} accepts script
426 files written in an alternate, restricted command language, described in
428 @ref{MRI,,MRI Compatible Script Files}.
431 the MRI Compatible Script Files section of GNU ld documentation.
433 Introduce MRI script files with
434 the option @samp{-c}; use the @samp{-T} option to run linker
435 scripts written in the general-purpose @command{ld} scripting language.
436 If @var{MRI-cmdfile} does not exist, @command{ld} looks for it in the directories
437 specified by any @samp{-L} options.
439 @cindex common allocation
446 These three options are equivalent; multiple forms are supported for
447 compatibility with other linkers. They assign space to common symbols
448 even if a relocatable output file is specified (with @samp{-r}). The
449 script command @code{FORCE_COMMON_ALLOCATION} has the same effect.
450 @xref{Miscellaneous Commands}.
452 @kindex --depaudit @var{AUDITLIB}
453 @kindex -P @var{AUDITLIB}
454 @item --depaudit @var{AUDITLIB}
455 @itemx -P @var{AUDITLIB}
456 Adds @var{AUDITLIB} to the @code{DT_DEPAUDIT} entry of the dynamic section.
457 @var{AUDITLIB} is not checked for existence, nor will it use the DT_SONAME
458 specified in the library. If specified multiple times @code{DT_DEPAUDIT}
459 will contain a colon separated list of audit interfaces to use. This
460 option is only meaningful on ELF platforms supporting the rtld-audit interface.
461 The -P option is provided for Solaris compatibility.
463 @kindex --enable-non-contiguous-regions
464 @item --enable-non-contiguous-regions
465 This option avoids generating an error if an input section does not
466 fit a matching output section. The linker tries to allocate the input
467 section to subseque nt matching output sections, and generates an
468 error only if no output section is large enough. This is useful when
469 several non-contiguous memory regions are available and the input
470 section does not require a particular one. The order in which input
471 sections are evaluated does not change, for instance:
475 MEM1 (rwx) : ORIGIN : 0x1000, LENGTH = 0x14
476 MEM2 (rwx) : ORIGIN : 0x1000, LENGTH = 0x40
477 MEM3 (rwx) : ORIGIN : 0x2000, LENGTH = 0x40
480 mem1 : @{ *(.data.*); @} > MEM1
481 mem2 : @{ *(.data.*); @} > MEM2
482 mem3 : @{ *(.data.*); @} > MEM2
490 results in .data.1 affected to mem1, and .data.2 and .data.3
491 affected to mem2, even though .data.3 would fit in mem3.
494 This option is incompatible with INSERT statements because it changes
495 the way input sections are mapped to output sections.
497 @kindex --enable-non-contiguous-regions-warnings
498 @item --enable-non-contiguous-regions-warnings
499 This option enables warnings when
500 @code{--enable-non-contiguous-regions} allows possibly unexpected
501 matches in sections mapping, potentially leading to silently
502 discarding a section instead of failing because it does not fit any
505 @cindex entry point, from command line
506 @kindex -e @var{entry}
507 @kindex --entry=@var{entry}
509 @itemx --entry=@var{entry}
510 Use @var{entry} as the explicit symbol for beginning execution of your
511 program, rather than the default entry point. If there is no symbol
512 named @var{entry}, the linker will try to parse @var{entry} as a number,
513 and use that as the entry address (the number will be interpreted in
514 base 10; you may use a leading @samp{0x} for base 16, or a leading
515 @samp{0} for base 8). @xref{Entry Point}, for a discussion of defaults
516 and other ways of specifying the entry point.
518 @kindex --exclude-libs
519 @item --exclude-libs @var{lib},@var{lib},...
520 Specifies a list of archive libraries from which symbols should not be automatically
521 exported. The library names may be delimited by commas or colons. Specifying
522 @code{--exclude-libs ALL} excludes symbols in all archive libraries from
523 automatic export. This option is available only for the i386 PE targeted
524 port of the linker and for ELF targeted ports. For i386 PE, symbols
525 explicitly listed in a .def file are still exported, regardless of this
526 option. For ELF targeted ports, symbols affected by this option will
527 be treated as hidden.
529 @kindex --exclude-modules-for-implib
530 @item --exclude-modules-for-implib @var{module},@var{module},...
531 Specifies a list of object files or archive members, from which symbols
532 should not be automatically exported, but which should be copied wholesale
533 into the import library being generated during the link. The module names
534 may be delimited by commas or colons, and must match exactly the filenames
535 used by @command{ld} to open the files; for archive members, this is simply
536 the member name, but for object files the name listed must include and
537 match precisely any path used to specify the input file on the linker's
538 command-line. This option is available only for the i386 PE targeted port
539 of the linker. Symbols explicitly listed in a .def file are still exported,
540 regardless of this option.
542 @cindex dynamic symbol table
544 @kindex --export-dynamic
545 @kindex --no-export-dynamic
547 @itemx --export-dynamic
548 @itemx --no-export-dynamic
549 When creating a dynamically linked executable, using the @option{-E}
550 option or the @option{--export-dynamic} option causes the linker to add
551 all symbols to the dynamic symbol table. The dynamic symbol table is the
552 set of symbols which are visible from dynamic objects at run time.
554 If you do not use either of these options (or use the
555 @option{--no-export-dynamic} option to restore the default behavior), the
556 dynamic symbol table will normally contain only those symbols which are
557 referenced by some dynamic object mentioned in the link.
559 If you use @code{dlopen} to load a dynamic object which needs to refer
560 back to the symbols defined by the program, rather than some other
561 dynamic object, then you will probably need to use this option when
562 linking the program itself.
564 You can also use the dynamic list to control what symbols should
565 be added to the dynamic symbol table if the output format supports it.
566 See the description of @samp{--dynamic-list}.
568 Note that this option is specific to ELF targeted ports. PE targets
569 support a similar function to export all symbols from a DLL or EXE; see
570 the description of @samp{--export-all-symbols} below.
572 @kindex --export-dynamic-symbol=@var{glob}
573 @cindex export dynamic symbol
574 @item --export-dynamic-symbol=@var{glob}
575 When creating a dynamically linked executable, symbols matching
576 @var{glob} will be added to the dynamic symbol table. When creating a
577 shared library, references to symbols matching @var{glob} will not be
578 bound to the definitions within the shared library. This option is a
579 no-op when creating a shared library and @samp{-Bsymbolic} or
580 @samp{--dynamic-list} are not specified. This option is only meaningful
581 on ELF platforms which support shared libraries.
583 @kindex --export-dynamic-symbol-list=@var{file}
584 @cindex export dynamic symbol list
585 @item --export-dynamic-symbol-list=@var{file}
586 Specify a @samp{--export-dynamic-symbol} for each pattern in the file.
587 The format of the file is the same as the version node without
588 scope and node name. See @ref{VERSION} for more information.
590 @ifclear SingleFormat
591 @cindex big-endian objects
595 Link big-endian objects. This affects the default output format.
597 @cindex little-endian objects
600 Link little-endian objects. This affects the default output format.
603 @kindex -f @var{name}
604 @kindex --auxiliary=@var{name}
606 @itemx --auxiliary=@var{name}
607 When creating an ELF shared object, set the internal DT_AUXILIARY field
608 to the specified name. This tells the dynamic linker that the symbol
609 table of the shared object should be used as an auxiliary filter on the
610 symbol table of the shared object @var{name}.
612 If you later link a program against this filter object, then, when you
613 run the program, the dynamic linker will see the DT_AUXILIARY field. If
614 the dynamic linker resolves any symbols from the filter object, it will
615 first check whether there is a definition in the shared object
616 @var{name}. If there is one, it will be used instead of the definition
617 in the filter object. The shared object @var{name} need not exist.
618 Thus the shared object @var{name} may be used to provide an alternative
619 implementation of certain functions, perhaps for debugging or for
620 machine-specific performance.
622 This option may be specified more than once. The DT_AUXILIARY entries
623 will be created in the order in which they appear on the command line.
625 @kindex -F @var{name}
626 @kindex --filter=@var{name}
628 @itemx --filter=@var{name}
629 When creating an ELF shared object, set the internal DT_FILTER field to
630 the specified name. This tells the dynamic linker that the symbol table
631 of the shared object which is being created should be used as a filter
632 on the symbol table of the shared object @var{name}.
634 If you later link a program against this filter object, then, when you
635 run the program, the dynamic linker will see the DT_FILTER field. The
636 dynamic linker will resolve symbols according to the symbol table of the
637 filter object as usual, but it will actually link to the definitions
638 found in the shared object @var{name}. Thus the filter object can be
639 used to select a subset of the symbols provided by the object
642 Some older linkers used the @option{-F} option throughout a compilation
643 toolchain for specifying object-file format for both input and output
645 @ifclear SingleFormat
646 The @sc{gnu} linker uses other mechanisms for this purpose: the
647 @option{-b}, @option{--format}, @option{--oformat} options, the
648 @code{TARGET} command in linker scripts, and the @code{GNUTARGET}
649 environment variable.
651 The @sc{gnu} linker will ignore the @option{-F} option when not
652 creating an ELF shared object.
654 @cindex finalization function
655 @kindex -fini=@var{name}
656 @item -fini=@var{name}
657 When creating an ELF executable or shared object, call NAME when the
658 executable or shared object is unloaded, by setting DT_FINI to the
659 address of the function. By default, the linker uses @code{_fini} as
660 the function to call.
664 Ignored. Provided for compatibility with other tools.
666 @kindex -G @var{value}
667 @kindex --gpsize=@var{value}
670 @itemx --gpsize=@var{value}
671 Set the maximum size of objects to be optimized using the GP register to
672 @var{size}. This is only meaningful for object file formats such as
673 MIPS ELF that support putting large and small objects into different
674 sections. This is ignored for other object file formats.
676 @cindex runtime library name
677 @kindex -h @var{name}
678 @kindex -soname=@var{name}
680 @itemx -soname=@var{name}
681 When creating an ELF shared object, set the internal DT_SONAME field to
682 the specified name. When an executable is linked with a shared object
683 which has a DT_SONAME field, then when the executable is run the dynamic
684 linker will attempt to load the shared object specified by the DT_SONAME
685 field rather than the using the file name given to the linker.
688 @cindex incremental link
690 Perform an incremental link (same as option @samp{-r}).
692 @cindex initialization function
693 @kindex -init=@var{name}
694 @item -init=@var{name}
695 When creating an ELF executable or shared object, call NAME when the
696 executable or shared object is loaded, by setting DT_INIT to the address
697 of the function. By default, the linker uses @code{_init} as the
700 @cindex archive files, from cmd line
701 @kindex -l @var{namespec}
702 @kindex --library=@var{namespec}
703 @item -l @var{namespec}
704 @itemx --library=@var{namespec}
705 Add the archive or object file specified by @var{namespec} to the
706 list of files to link. This option may be used any number of times.
707 If @var{namespec} is of the form @file{:@var{filename}}, @command{ld}
708 will search the library path for a file called @var{filename}, otherwise it
709 will search the library path for a file called @file{lib@var{namespec}.a}.
711 On systems which support shared libraries, @command{ld} may also search for
712 files other than @file{lib@var{namespec}.a}. Specifically, on ELF
713 and SunOS systems, @command{ld} will search a directory for a library
714 called @file{lib@var{namespec}.so} before searching for one called
715 @file{lib@var{namespec}.a}. (By convention, a @code{.so} extension
716 indicates a shared library.) Note that this behavior does not apply
717 to @file{:@var{filename}}, which always specifies a file called
720 The linker will search an archive only once, at the location where it is
721 specified on the command line. If the archive defines a symbol which
722 was undefined in some object which appeared before the archive on the
723 command line, the linker will include the appropriate file(s) from the
724 archive. However, an undefined symbol in an object appearing later on
725 the command line will not cause the linker to search the archive again.
727 See the @option{-(} option for a way to force the linker to search
728 archives multiple times.
730 You may list the same archive multiple times on the command line.
733 This type of archive searching is standard for Unix linkers. However,
734 if you are using @command{ld} on AIX, note that it is different from the
735 behaviour of the AIX linker.
738 @cindex search directory, from cmd line
740 @kindex --library-path=@var{dir}
741 @item -L @var{searchdir}
742 @itemx --library-path=@var{searchdir}
743 Add path @var{searchdir} to the list of paths that @command{ld} will search
744 for archive libraries and @command{ld} control scripts. You may use this
745 option any number of times. The directories are searched in the order
746 in which they are specified on the command line. Directories specified
747 on the command line are searched before the default directories. All
748 @option{-L} options apply to all @option{-l} options, regardless of the
749 order in which the options appear. @option{-L} options do not affect
750 how @command{ld} searches for a linker script unless @option{-T}
753 If @var{searchdir} begins with @code{=} or @code{$SYSROOT}, then this
754 prefix will be replaced by the @dfn{sysroot prefix}, controlled by the
755 @samp{--sysroot} option, or specified when the linker is configured.
758 The default set of paths searched (without being specified with
759 @samp{-L}) depends on which emulation mode @command{ld} is using, and in
760 some cases also on how it was configured. @xref{Environment}.
763 The paths can also be specified in a link script with the
764 @code{SEARCH_DIR} command. Directories specified this way are searched
765 at the point in which the linker script appears in the command line.
768 @kindex -m @var{emulation}
769 @item -m @var{emulation}
770 Emulate the @var{emulation} linker. You can list the available
771 emulations with the @samp{--verbose} or @samp{-V} options.
773 If the @samp{-m} option is not used, the emulation is taken from the
774 @code{LDEMULATION} environment variable, if that is defined.
776 Otherwise, the default emulation depends upon how the linker was
784 Print a link map to the standard output. A link map provides
785 information about the link, including the following:
789 Where object files are mapped into memory.
791 How common symbols are allocated.
793 All archive members included in the link, with a mention of the symbol
794 which caused the archive member to be brought in.
796 The values assigned to symbols.
798 Note - symbols whose values are computed by an expression which
799 involves a reference to a previous value of the same symbol may not
800 have correct result displayed in the link map. This is because the
801 linker discards intermediate results and only retains the final value
802 of an expression. Under such circumstances the linker will display
803 the final value enclosed by square brackets. Thus for example a
804 linker script containing:
812 will produce the following output in the link map if the @option{-M}
817 [0x0000000c] foo = (foo * 0x4)
818 [0x0000000c] foo = (foo + 0x8)
821 See @ref{Expressions} for more information about expressions in linker
825 How GNU properties are merged.
827 When the linker merges input .note.gnu.property sections into one output
828 .note.gnu.property section, some properties are removed or updated.
829 These actions are reported in the link map. For example:
832 Removed property 0xc0000002 to merge foo.o (0x1) and bar.o (not found)
835 This indicates that property 0xc0000002 is removed from output when
836 merging properties in @file{foo.o}, whose property 0xc0000002 value
837 is 0x1, and @file{bar.o}, which doesn't have property 0xc0000002.
840 Updated property 0xc0010001 (0x1) to merge foo.o (0x1) and bar.o (0x1)
843 This indicates that property 0xc0010001 value is updated to 0x1 in output
844 when merging properties in @file{foo.o}, whose 0xc0010001 property value
845 is 0x1, and @file{bar.o}, whose 0xc0010001 property value is 0x1.
848 @cindex link map discarded
849 @kindex --print-map-discarded
850 @kindex --no-print-map-discarded
851 @item --print-map-discarded
852 @itemx --no-print-map-discarded
853 Print (or do not print) the list of discarded and garbage collected sections
854 in the link map. Enabled by default.
857 @cindex read-only text
862 Turn off page alignment of sections, and disable linking against shared
863 libraries. If the output format supports Unix style magic numbers,
864 mark the output as @code{NMAGIC}.
868 @cindex read/write from cmd line
872 Set the text and data sections to be readable and writable. Also, do
873 not page-align the data segment, and disable linking against shared
874 libraries. If the output format supports Unix style magic numbers,
875 mark the output as @code{OMAGIC}. Note: Although a writable text section
876 is allowed for PE-COFF targets, it does not conform to the format
877 specification published by Microsoft.
882 This option negates most of the effects of the @option{-N} option. It
883 sets the text section to be read-only, and forces the data segment to
884 be page-aligned. Note - this option does not enable linking against
885 shared libraries. Use @option{-Bdynamic} for this.
887 @kindex -o @var{output}
888 @kindex --output=@var{output}
889 @cindex naming the output file
890 @item -o @var{output}
891 @itemx --output=@var{output}
892 Use @var{output} as the name for the program produced by @command{ld}; if this
893 option is not specified, the name @file{a.out} is used by default. The
894 script command @code{OUTPUT} can also specify the output file name.
896 @kindex --dependency-file=@var{depfile}
897 @cindex dependency file
898 @item --dependency-file=@var{depfile}
899 Write a @dfn{dependency file} to @var{depfile}. This file contains a rule
900 suitable for @code{make} describing the output file and all the input files
901 that were read to produce it. The output is similar to the compiler's
902 output with @samp{-M -MP} (@pxref{Preprocessor Options,, Options
903 Controlling the Preprocessor, gcc.info, Using the GNU Compiler
904 Collection}). Note that there is no option like the compiler's @samp{-MM},
905 to exclude ``system files'' (which is not a well-specified concept in the
906 linker, unlike ``system headers'' in the compiler). So the output from
907 @samp{--dependency-file} is always specific to the exact state of the
908 installation where it was produced, and should not be copied into
909 distributed makefiles without careful editing.
911 @kindex -O @var{level}
912 @cindex generating optimized output
914 If @var{level} is a numeric values greater than zero @command{ld} optimizes
915 the output. This might take significantly longer and therefore probably
916 should only be enabled for the final binary. At the moment this
917 option only affects ELF shared library generation. Future releases of
918 the linker may make more use of this option. Also currently there is
919 no difference in the linker's behaviour for different non-zero values
920 of this option. Again this may change with future releases.
922 @kindex -plugin @var{name}
923 @item -plugin @var{name}
924 Involve a plugin in the linking process. The @var{name} parameter is
925 the absolute filename of the plugin. Usually this parameter is
926 automatically added by the complier, when using link time
927 optimization, but users can also add their own plugins if they so
930 Note that the location of the compiler originated plugins is different
931 from the place where the @command{ar}, @command{nm} and
932 @command{ranlib} programs search for their plugins. In order for
933 those commands to make use of a compiler based plugin it must first be
934 copied into the @file{$@{libdir@}/bfd-plugins} directory. All gcc
935 based linker plugins are backward compatible, so it is sufficient to
936 just copy in the newest one.
939 @cindex push state governing input file handling
941 The @option{--push-state} allows to preserve the current state of the
942 flags which govern the input file handling so that they can all be
943 restored with one corresponding @option{--pop-state} option.
945 The option which are covered are: @option{-Bdynamic}, @option{-Bstatic},
946 @option{-dn}, @option{-dy}, @option{-call_shared}, @option{-non_shared},
947 @option{-static}, @option{-N}, @option{-n}, @option{--whole-archive},
948 @option{--no-whole-archive}, @option{-r}, @option{-Ur},
949 @option{--copy-dt-needed-entries}, @option{--no-copy-dt-needed-entries},
950 @option{--as-needed}, @option{--no-as-needed}, and @option{-a}.
952 One target for this option are specifications for @file{pkg-config}. When
953 used with the @option{--libs} option all possibly needed libraries are
954 listed and then possibly linked with all the time. It is better to return
955 something as follows:
958 -Wl,--push-state,--as-needed -libone -libtwo -Wl,--pop-state
962 @cindex pop state governing input file handling
964 Undoes the effect of --push-state, restores the previous values of the
965 flags governing input file handling.
968 @kindex --emit-relocs
969 @cindex retain relocations in final executable
972 Leave relocation sections and contents in fully linked executables.
973 Post link analysis and optimization tools may need this information in
974 order to perform correct modifications of executables. This results
975 in larger executables.
977 This option is currently only supported on ELF platforms.
979 @kindex --force-dynamic
980 @cindex forcing the creation of dynamic sections
981 @item --force-dynamic
982 Force the output file to have dynamic sections. This option is specific
986 @cindex relocatable output
988 @kindex --relocatable
991 Generate relocatable output---i.e., generate an output file that can in
992 turn serve as input to @command{ld}. This is often called @dfn{partial
993 linking}. As a side effect, in environments that support standard Unix
994 magic numbers, this option also sets the output file's magic number to
996 @c ; see @option{-N}.
997 If this option is not specified, an absolute file is produced. When
998 linking C++ programs, this option @emph{will not} resolve references to
999 constructors; to do that, use @samp{-Ur}.
1001 When an input file does not have the same format as the output file,
1002 partial linking is only supported if that input file does not contain any
1003 relocations. Different output formats can have further restrictions; for
1004 example some @code{a.out}-based formats do not support partial linking
1005 with input files in other formats at all.
1007 This option does the same thing as @samp{-i}.
1009 @kindex -R @var{file}
1010 @kindex --just-symbols=@var{file}
1011 @cindex symbol-only input
1012 @item -R @var{filename}
1013 @itemx --just-symbols=@var{filename}
1014 Read symbol names and their addresses from @var{filename}, but do not
1015 relocate it or include it in the output. This allows your output file
1016 to refer symbolically to absolute locations of memory defined in other
1017 programs. You may use this option more than once.
1019 For compatibility with other ELF linkers, if the @option{-R} option is
1020 followed by a directory name, rather than a file name, it is treated as
1021 the @option{-rpath} option.
1025 @cindex strip all symbols
1028 Omit all symbol information from the output file.
1031 @kindex --strip-debug
1032 @cindex strip debugger symbols
1034 @itemx --strip-debug
1035 Omit debugger symbol information (but not all symbols) from the output file.
1037 @kindex --strip-discarded
1038 @kindex --no-strip-discarded
1039 @item --strip-discarded
1040 @itemx --no-strip-discarded
1041 Omit (or do not omit) global symbols defined in discarded sections.
1046 @cindex input files, displaying
1049 Print the names of the input files as @command{ld} processes them. If
1050 @samp{-t} is given twice then members within archives are also printed.
1051 @samp{-t} output is useful to generate a list of all the object files
1052 and scripts involved in linking, for example, when packaging files for
1053 a linker bug report.
1055 @kindex -T @var{script}
1056 @kindex --script=@var{script}
1057 @cindex script files
1058 @item -T @var{scriptfile}
1059 @itemx --script=@var{scriptfile}
1060 Use @var{scriptfile} as the linker script. This script replaces
1061 @command{ld}'s default linker script (rather than adding to it), so
1062 @var{commandfile} must specify everything necessary to describe the
1063 output file. @xref{Scripts}. If @var{scriptfile} does not exist in
1064 the current directory, @code{ld} looks for it in the directories
1065 specified by any preceding @samp{-L} options. Multiple @samp{-T}
1068 @kindex -dT @var{script}
1069 @kindex --default-script=@var{script}
1070 @cindex script files
1071 @item -dT @var{scriptfile}
1072 @itemx --default-script=@var{scriptfile}
1073 Use @var{scriptfile} as the default linker script. @xref{Scripts}.
1075 This option is similar to the @option{--script} option except that
1076 processing of the script is delayed until after the rest of the
1077 command line has been processed. This allows options placed after the
1078 @option{--default-script} option on the command line to affect the
1079 behaviour of the linker script, which can be important when the linker
1080 command line cannot be directly controlled by the user. (eg because
1081 the command line is being constructed by another tool, such as
1084 @kindex -u @var{symbol}
1085 @kindex --undefined=@var{symbol}
1086 @cindex undefined symbol
1087 @item -u @var{symbol}
1088 @itemx --undefined=@var{symbol}
1089 Force @var{symbol} to be entered in the output file as an undefined
1090 symbol. Doing this may, for example, trigger linking of additional
1091 modules from standard libraries. @samp{-u} may be repeated with
1092 different option arguments to enter additional undefined symbols. This
1093 option is equivalent to the @code{EXTERN} linker script command.
1095 If this option is being used to force additional modules to be pulled
1096 into the link, and if it is an error for the symbol to remain
1097 undefined, then the option @option{--require-defined} should be used
1100 @kindex --require-defined=@var{symbol}
1101 @cindex symbols, require defined
1102 @cindex defined symbol
1103 @item --require-defined=@var{symbol}
1104 Require that @var{symbol} is defined in the output file. This option
1105 is the same as option @option{--undefined} except that if @var{symbol}
1106 is not defined in the output file then the linker will issue an error
1107 and exit. The same effect can be achieved in a linker script by using
1108 @code{EXTERN}, @code{ASSERT} and @code{DEFINED} together. This option
1109 can be used multiple times to require additional symbols.
1112 @cindex constructors
1114 For anything other than C++ programs, this option is equivalent to
1115 @samp{-r}: it generates relocatable output---i.e., an output file that can in
1116 turn serve as input to @command{ld}. When linking C++ programs, @samp{-Ur}
1117 @emph{does} resolve references to constructors, unlike @samp{-r}.
1118 It does not work to use @samp{-Ur} on files that were themselves linked
1119 with @samp{-Ur}; once the constructor table has been built, it cannot
1120 be added to. Use @samp{-Ur} only for the last partial link, and
1121 @samp{-r} for the others.
1123 @kindex --orphan-handling=@var{MODE}
1124 @cindex orphan sections
1125 @cindex sections, orphan
1126 @item --orphan-handling=@var{MODE}
1127 Control how orphan sections are handled. An orphan section is one not
1128 specifically mentioned in a linker script. @xref{Orphan Sections}.
1130 @var{MODE} can have any of the following values:
1134 Orphan sections are placed into a suitable output section following
1135 the strategy described in @ref{Orphan Sections}. The option
1136 @samp{--unique} also affects how sections are placed.
1139 All orphan sections are discarded, by placing them in the
1140 @samp{/DISCARD/} section (@pxref{Output Section Discarding}).
1143 The linker will place the orphan section as for @code{place} and also
1147 The linker will exit with an error if any orphan section is found.
1150 The default if @samp{--orphan-handling} is not given is @code{place}.
1152 @kindex --unique[=@var{SECTION}]
1153 @item --unique[=@var{SECTION}]
1154 Creates a separate output section for every input section matching
1155 @var{SECTION}, or if the optional wildcard @var{SECTION} argument is
1156 missing, for every orphan input section. An orphan section is one not
1157 specifically mentioned in a linker script. You may use this option
1158 multiple times on the command line; It prevents the normal merging of
1159 input sections with the same name, overriding output section assignments
1169 Display the version number for @command{ld}. The @option{-V} option also
1170 lists the supported emulations.
1173 @kindex --discard-all
1174 @cindex deleting local symbols
1176 @itemx --discard-all
1177 Delete all local symbols.
1180 @kindex --discard-locals
1181 @cindex local symbols, deleting
1183 @itemx --discard-locals
1184 Delete all temporary local symbols. (These symbols start with
1185 system-specific local label prefixes, typically @samp{.L} for ELF systems
1186 or @samp{L} for traditional a.out systems.)
1188 @kindex -y @var{symbol}
1189 @kindex --trace-symbol=@var{symbol}
1190 @cindex symbol tracing
1191 @item -y @var{symbol}
1192 @itemx --trace-symbol=@var{symbol}
1193 Print the name of each linked file in which @var{symbol} appears. This
1194 option may be given any number of times. On many systems it is necessary
1195 to prepend an underscore.
1197 This option is useful when you have an undefined symbol in your link but
1198 don't know where the reference is coming from.
1200 @kindex -Y @var{path}
1202 Add @var{path} to the default library search path. This option exists
1203 for Solaris compatibility.
1205 @kindex -z @var{keyword}
1206 @item -z @var{keyword}
1207 The recognized keywords are:
1211 Always generate BND prefix in PLT entries. Supported for Linux/x86_64.
1213 @item call-nop=prefix-addr
1214 @itemx call-nop=suffix-nop
1215 @itemx call-nop=prefix-@var{byte}
1216 @itemx call-nop=suffix-@var{byte}
1217 Specify the 1-byte @code{NOP} padding when transforming indirect call
1218 to a locally defined function, foo, via its GOT slot.
1219 @option{call-nop=prefix-addr} generates @code{0x67 call foo}.
1220 @option{call-nop=suffix-nop} generates @code{call foo 0x90}.
1221 @option{call-nop=prefix-@var{byte}} generates @code{@var{byte} call foo}.
1222 @option{call-nop=suffix-@var{byte}} generates @code{call foo @var{byte}}.
1223 Supported for i386 and x86_64.
1225 @item cet-report=none
1226 @itemx cet-report=warning
1227 @itemx cet-report=error
1228 Specify how to report the missing GNU_PROPERTY_X86_FEATURE_1_IBT and
1229 GNU_PROPERTY_X86_FEATURE_1_SHSTK properties in input .note.gnu.property
1230 section. @option{cet-report=none}, which is the default, will make the
1231 linker not report missing properties in input files.
1232 @option{cet-report=warning} will make the linker issue a warning for
1233 missing properties in input files. @option{cet-report=error} will make
1234 the linker issue an error for missing properties in input files.
1235 Note that @option{ibt} will turn off the missing
1236 GNU_PROPERTY_X86_FEATURE_1_IBT property report and @option{shstk} will
1237 turn off the missing GNU_PROPERTY_X86_FEATURE_1_SHSTK property report.
1238 Supported for Linux/i386 and Linux/x86_64.
1242 Combine multiple dynamic relocation sections and sort to improve
1243 dynamic symbol lookup caching. Do not do this if @samp{nocombreloc}.
1247 Generate common symbols with STT_COMMON type during a relocatable
1248 link. Use STT_OBJECT type if @samp{nocommon}.
1250 @item common-page-size=@var{value}
1251 Set the page size most commonly used to @var{value}. Memory image
1252 layout will be optimized to minimize memory pages if the system is
1253 using pages of this size.
1256 Report unresolved symbol references from regular object files. This
1257 is done even if the linker is creating a non-symbolic shared library.
1258 This option is the inverse of @samp{-z undefs}.
1260 @item dynamic-undefined-weak
1261 @itemx nodynamic-undefined-weak
1262 Make undefined weak symbols dynamic when building a dynamic object,
1263 if they are referenced from a regular object file and not forced local
1264 by symbol visibility or versioning. Do not make them dynamic if
1265 @samp{nodynamic-undefined-weak}. If neither option is given, a target
1266 may default to either option being in force, or make some other
1267 selection of undefined weak symbols dynamic. Not all targets support
1271 Marks the object as requiring executable stack.
1274 This option is only meaningful when building a shared object. It makes
1275 the symbols defined by this shared object available for symbol resolution
1276 of subsequently loaded libraries.
1279 This option is only meaningful when building a dynamic executable.
1280 This option marks the executable as requiring global auditing by
1281 setting the @code{DF_1_GLOBAUDIT} bit in the @code{DT_FLAGS_1} dynamic
1282 tag. Global auditing requires that any auditing library defined via
1283 the @option{--depaudit} or @option{-P} command-line options be run for
1284 all dynamic objects loaded by the application.
1287 Generate Intel Indirect Branch Tracking (IBT) enabled PLT entries.
1288 Supported for Linux/i386 and Linux/x86_64.
1291 Generate GNU_PROPERTY_X86_FEATURE_1_IBT in .note.gnu.property section
1292 to indicate compatibility with IBT. This also implies @option{ibtplt}.
1293 Supported for Linux/i386 and Linux/x86_64.
1296 This option is only meaningful when building a shared object.
1297 It marks the object so that its runtime initialization will occur
1298 before the runtime initialization of any other objects brought into
1299 the process at the same time. Similarly the runtime finalization of
1300 the object will occur after the runtime finalization of any other
1304 Specify that the dynamic loader should modify its symbol search order
1305 so that symbols in this shared library interpose all other shared
1306 libraries not so marked.
1309 When generating an executable or shared library, mark it to tell the
1310 dynamic linker to defer function call resolution to the point when
1311 the function is called (lazy binding), rather than at load time.
1312 Lazy binding is the default.
1315 Specify that the object's filters be processed immediately at runtime.
1317 @item max-page-size=@var{value}
1318 Set the maximum memory page size supported to @var{value}.
1321 Allow multiple definitions.
1324 Disable linker generated .dynbss variables used in place of variables
1325 defined in shared libraries. May result in dynamic text relocations.
1328 Specify that the dynamic loader search for dependencies of this object
1329 should ignore any default library search paths.
1332 Specify that the object shouldn't be unloaded at runtime.
1335 Specify that the object is not available to @code{dlopen}.
1338 Specify that the object can not be dumped by @code{dldump}.
1341 Marks the object as not requiring executable stack.
1343 @item noextern-protected-data
1344 Don't treat protected data symbols as external when building a shared
1345 library. This option overrides the linker backend default. It can be
1346 used to work around incorrect relocations against protected data symbols
1347 generated by compiler. Updates on protected data symbols by another
1348 module aren't visible to the resulting shared library. Supported for
1351 @item noreloc-overflow
1352 Disable relocation overflow check. This can be used to disable
1353 relocation overflow check if there will be no dynamic relocation
1354 overflow at run-time. Supported for x86_64.
1357 When generating an executable or shared library, mark it to tell the
1358 dynamic linker to resolve all symbols when the program is started, or
1359 when the shared library is loaded by dlopen, instead of deferring
1360 function call resolution to the point when the function is first
1364 Specify that the object requires @samp{$ORIGIN} handling in paths.
1368 Create an ELF @code{PT_GNU_RELRO} segment header in the object. This
1369 specifies a memory segment that should be made read-only after
1370 relocation, if supported. Specifying @samp{common-page-size} smaller
1371 than the system page size will render this protection ineffective.
1372 Don't create an ELF @code{PT_GNU_RELRO} segment if @samp{norelro}.
1375 @itemx noseparate-code
1376 Create separate code @code{PT_LOAD} segment header in the object. This
1377 specifies a memory segment that should contain only instructions and must
1378 be in wholly disjoint pages from any other data. Don't create separate
1379 code @code{PT_LOAD} segment if @samp{noseparate-code} is used.
1382 @itemx nounique-symbol
1383 Avoid duplicated local symbol names in the symbol string table. Append
1384 ".@code{number}" to duplicated local symbol names if @samp{unique-symbol}
1385 is used. @option{nounique-symbol} is the default.
1388 Generate GNU_PROPERTY_X86_FEATURE_1_SHSTK in .note.gnu.property section
1389 to indicate compatibility with Intel Shadow Stack. Supported for
1390 Linux/i386 and Linux/x86_64.
1392 @item stack-size=@var{value}
1393 Specify a stack size for an ELF @code{PT_GNU_STACK} segment.
1394 Specifying zero will override any default non-zero sized
1395 @code{PT_GNU_STACK} segment creation.
1397 @item start-stop-visibility=@var{value}
1399 @cindex ELF symbol visibility
1400 Specify the ELF symbol visibility for synthesized
1401 @code{__start_SECNAME} and @code{__stop_SECNAME} symbols (@pxref{Input
1402 Section Example}). @var{value} must be exactly @samp{default},
1403 @samp{internal}, @samp{hidden}, or @samp{protected}. If no @samp{-z
1404 start-stop-visibility} option is given, @samp{protected} is used for
1405 compatibility with historical practice. However, it's highly
1406 recommended to use @samp{-z start-stop-visibility=hidden} in new
1407 programs and shared libraries so that these symbols are not exported
1408 between shared objects, which is not usually what's intended.
1413 Report an error if DT_TEXTREL is set, i.e., if the position-independent
1414 or shared object has dynamic relocations in read-only sections. Don't
1415 report an error if @samp{notext} or @samp{textoff}.
1418 Do not report unresolved symbol references from regular object files,
1419 either when creating an executable, or when creating a shared library.
1420 This option is the inverse of @samp{-z defs}.
1425 Specify the x86-64 ISA level needed in .note.gnu.property section.
1426 @option{x86-64-v2} generates @code{GNU_PROPERTY_X86_ISA_1_V2}.
1427 @option{x86-64-v3} generates @code{GNU_PROPERTY_X86_ISA_1_V3}.
1428 @option{x86-64-v4} generates @code{GNU_PROPERTY_X86_ISA_1_V4}.
1429 Supported for Linux/i386 and Linux/x86_64.
1433 Other keywords are ignored for Solaris compatibility.
1436 @cindex groups of archives
1437 @item -( @var{archives} -)
1438 @itemx --start-group @var{archives} --end-group
1439 The @var{archives} should be a list of archive files. They may be
1440 either explicit file names, or @samp{-l} options.
1442 The specified archives are searched repeatedly until no new undefined
1443 references are created. Normally, an archive is searched only once in
1444 the order that it is specified on the command line. If a symbol in that
1445 archive is needed to resolve an undefined symbol referred to by an
1446 object in an archive that appears later on the command line, the linker
1447 would not be able to resolve that reference. By grouping the archives,
1448 they will all be searched repeatedly until all possible references are
1451 Using this option has a significant performance cost. It is best to use
1452 it only when there are unavoidable circular references between two or
1455 @kindex --accept-unknown-input-arch
1456 @kindex --no-accept-unknown-input-arch
1457 @item --accept-unknown-input-arch
1458 @itemx --no-accept-unknown-input-arch
1459 Tells the linker to accept input files whose architecture cannot be
1460 recognised. The assumption is that the user knows what they are doing
1461 and deliberately wants to link in these unknown input files. This was
1462 the default behaviour of the linker, before release 2.14. The default
1463 behaviour from release 2.14 onwards is to reject such input files, and
1464 so the @samp{--accept-unknown-input-arch} option has been added to
1465 restore the old behaviour.
1468 @kindex --no-as-needed
1470 @itemx --no-as-needed
1471 This option affects ELF DT_NEEDED tags for dynamic libraries mentioned
1472 on the command line after the @option{--as-needed} option. Normally
1473 the linker will add a DT_NEEDED tag for each dynamic library mentioned
1474 on the command line, regardless of whether the library is actually
1475 needed or not. @option{--as-needed} causes a DT_NEEDED tag to only be
1476 emitted for a library that @emph{at that point in the link} satisfies a
1477 non-weak undefined symbol reference from a regular object file or, if
1478 the library is not found in the DT_NEEDED lists of other needed libraries, a
1479 non-weak undefined symbol reference from another needed dynamic library.
1480 Object files or libraries appearing on the command line @emph{after}
1481 the library in question do not affect whether the library is seen as
1482 needed. This is similar to the rules for extraction of object files
1483 from archives. @option{--no-as-needed} restores the default behaviour.
1485 @kindex --add-needed
1486 @kindex --no-add-needed
1488 @itemx --no-add-needed
1489 These two options have been deprecated because of the similarity of
1490 their names to the @option{--as-needed} and @option{--no-as-needed}
1491 options. They have been replaced by @option{--copy-dt-needed-entries}
1492 and @option{--no-copy-dt-needed-entries}.
1494 @kindex -assert @var{keyword}
1495 @item -assert @var{keyword}
1496 This option is ignored for SunOS compatibility.
1500 @kindex -call_shared
1504 Link against dynamic libraries. This is only meaningful on platforms
1505 for which shared libraries are supported. This option is normally the
1506 default on such platforms. The different variants of this option are
1507 for compatibility with various systems. You may use this option
1508 multiple times on the command line: it affects library searching for
1509 @option{-l} options which follow it.
1513 Set the @code{DF_1_GROUP} flag in the @code{DT_FLAGS_1} entry in the dynamic
1514 section. This causes the runtime linker to handle lookups in this
1515 object and its dependencies to be performed only inside the group.
1516 @option{--unresolved-symbols=report-all} is implied. This option is
1517 only meaningful on ELF platforms which support shared libraries.
1527 Do not link against shared libraries. This is only meaningful on
1528 platforms for which shared libraries are supported. The different
1529 variants of this option are for compatibility with various systems. You
1530 may use this option multiple times on the command line: it affects
1531 library searching for @option{-l} options which follow it. This
1532 option also implies @option{--unresolved-symbols=report-all}. This
1533 option can be used with @option{-shared}. Doing so means that a
1534 shared library is being created but that all of the library's external
1535 references must be resolved by pulling in entries from static
1540 When creating a shared library, bind references to global symbols to the
1541 definition within the shared library, if any. Normally, it is possible
1542 for a program linked against a shared library to override the definition
1543 within the shared library. This option is only meaningful on ELF
1544 platforms which support shared libraries.
1546 @kindex -Bsymbolic-functions
1547 @item -Bsymbolic-functions
1548 When creating a shared library, bind references to global function
1549 symbols to the definition within the shared library, if any.
1550 This option is only meaningful on ELF platforms which support shared
1553 @kindex --dynamic-list=@var{dynamic-list-file}
1554 @item --dynamic-list=@var{dynamic-list-file}
1555 Specify the name of a dynamic list file to the linker. This is
1556 typically used when creating shared libraries to specify a list of
1557 global symbols whose references shouldn't be bound to the definition
1558 within the shared library, or creating dynamically linked executables
1559 to specify a list of symbols which should be added to the symbol table
1560 in the executable. This option is only meaningful on ELF platforms
1561 which support shared libraries.
1563 The format of the dynamic list is the same as the version node without
1564 scope and node name. See @ref{VERSION} for more information.
1566 @kindex --dynamic-list-data
1567 @item --dynamic-list-data
1568 Include all global data symbols to the dynamic list.
1570 @kindex --dynamic-list-cpp-new
1571 @item --dynamic-list-cpp-new
1572 Provide the builtin dynamic list for C++ operator new and delete. It
1573 is mainly useful for building shared libstdc++.
1575 @kindex --dynamic-list-cpp-typeinfo
1576 @item --dynamic-list-cpp-typeinfo
1577 Provide the builtin dynamic list for C++ runtime type identification.
1579 @kindex --check-sections
1580 @kindex --no-check-sections
1581 @item --check-sections
1582 @itemx --no-check-sections
1583 Asks the linker @emph{not} to check section addresses after they have
1584 been assigned to see if there are any overlaps. Normally the linker will
1585 perform this check, and if it finds any overlaps it will produce
1586 suitable error messages. The linker does know about, and does make
1587 allowances for sections in overlays. The default behaviour can be
1588 restored by using the command-line switch @option{--check-sections}.
1589 Section overlap is not usually checked for relocatable links. You can
1590 force checking in that case by using the @option{--check-sections}
1593 @kindex --copy-dt-needed-entries
1594 @kindex --no-copy-dt-needed-entries
1595 @item --copy-dt-needed-entries
1596 @itemx --no-copy-dt-needed-entries
1597 This option affects the treatment of dynamic libraries referred to
1598 by DT_NEEDED tags @emph{inside} ELF dynamic libraries mentioned on the
1599 command line. Normally the linker won't add a DT_NEEDED tag to the
1600 output binary for each library mentioned in a DT_NEEDED tag in an
1601 input dynamic library. With @option{--copy-dt-needed-entries}
1602 specified on the command line however any dynamic libraries that
1603 follow it will have their DT_NEEDED entries added. The default
1604 behaviour can be restored with @option{--no-copy-dt-needed-entries}.
1606 This option also has an effect on the resolution of symbols in dynamic
1607 libraries. With @option{--copy-dt-needed-entries} dynamic libraries
1608 mentioned on the command line will be recursively searched, following
1609 their DT_NEEDED tags to other libraries, in order to resolve symbols
1610 required by the output binary. With the default setting however
1611 the searching of dynamic libraries that follow it will stop with the
1612 dynamic library itself. No DT_NEEDED links will be traversed to resolve
1615 @cindex cross reference table
1618 Output a cross reference table. If a linker map file is being
1619 generated, the cross reference table is printed to the map file.
1620 Otherwise, it is printed on the standard output.
1622 The format of the table is intentionally simple, so that it may be
1623 easily processed by a script if necessary. The symbols are printed out,
1624 sorted by name. For each symbol, a list of file names is given. If the
1625 symbol is defined, the first file listed is the location of the
1626 definition. If the symbol is defined as a common value then any files
1627 where this happens appear next. Finally any files that reference the
1630 @cindex ctf variables
1631 @kindex --ctf-variables
1632 @kindex --no-ctf-variables
1633 @item --ctf-variables
1634 @item --no-ctf-variables
1635 The CTF debuginfo format supports a section which encodes the names and
1636 types of variables found in the program which do not appear in any symbol
1637 table. These variables clearly cannot be looked up by address by
1638 conventional debuggers, so the space used for their types and names is
1639 usually wasted: the types are usually small but the names are often not.
1640 @option{--ctf-variables} causes the generation of such a section.
1641 The default behaviour can be restored with @option{--no-ctf-variables}.
1643 @cindex ctf type sharing
1644 @kindex --ctf-share-types
1645 @item --ctf-share-types=@var{method}
1646 Adjust the method used to share types between translation units in CTF.
1649 @item share-unconflicted
1650 Put all types that do not have ambiguous definitions into the shared dictionary,
1651 where debuggers can easily access them, even if they only occur in one
1652 translation unit. This is the default.
1654 @item share-duplicated
1655 Put only types that occur in multiple translation units into the shared
1656 dictionary: types with only one definition go into per-translation-unit
1657 dictionaries. Types with ambiguous definitions in multiple translation units
1658 always go into per-translation-unit dictionaries. This tends to make the CTF
1659 larger, but may reduce the amount of CTF in the shared dictionary. For very
1660 large projects this may speed up opening the CTF and save memory in the CTF
1661 consumer at runtime.
1664 @cindex common allocation
1665 @kindex --no-define-common
1666 @item --no-define-common
1667 This option inhibits the assignment of addresses to common symbols.
1668 The script command @code{INHIBIT_COMMON_ALLOCATION} has the same effect.
1669 @xref{Miscellaneous Commands}.
1671 The @samp{--no-define-common} option allows decoupling
1672 the decision to assign addresses to Common symbols from the choice
1673 of the output file type; otherwise a non-Relocatable output type
1674 forces assigning addresses to Common symbols.
1675 Using @samp{--no-define-common} allows Common symbols that are referenced
1676 from a shared library to be assigned addresses only in the main program.
1677 This eliminates the unused duplicate space in the shared library,
1678 and also prevents any possible confusion over resolving to the wrong
1679 duplicate when there are many dynamic modules with specialized search
1680 paths for runtime symbol resolution.
1682 @cindex group allocation in linker script
1683 @cindex section groups
1685 @kindex --force-group-allocation
1686 @item --force-group-allocation
1687 This option causes the linker to place section group members like
1688 normal input sections, and to delete the section groups. This is the
1689 default behaviour for a final link but this option can be used to
1690 change the behaviour of a relocatable link (@samp{-r}). The script
1691 command @code{FORCE_GROUP_ALLOCATION} has the same
1692 effect. @xref{Miscellaneous Commands}.
1694 @cindex symbols, from command line
1695 @kindex --defsym=@var{symbol}=@var{exp}
1696 @item --defsym=@var{symbol}=@var{expression}
1697 Create a global symbol in the output file, containing the absolute
1698 address given by @var{expression}. You may use this option as many
1699 times as necessary to define multiple symbols in the command line. A
1700 limited form of arithmetic is supported for the @var{expression} in this
1701 context: you may give a hexadecimal constant or the name of an existing
1702 symbol, or use @code{+} and @code{-} to add or subtract hexadecimal
1703 constants or symbols. If you need more elaborate expressions, consider
1704 using the linker command language from a script (@pxref{Assignments}).
1705 @emph{Note:} there should be no white space between @var{symbol}, the
1706 equals sign (``@key{=}''), and @var{expression}.
1708 The linker processes @samp{--defsym} arguments and @samp{-T} arguments
1709 in order, placing @samp{--defsym} before @samp{-T} will define the
1710 symbol before the linker script from @samp{-T} is processed, while
1711 placing @samp{--defsym} after @samp{-T} will define the symbol after
1712 the linker script has been processed. This difference has
1713 consequences for expressions within the linker script that use the
1714 @samp{--defsym} symbols, which order is correct will depend on what
1715 you are trying to achieve.
1717 @cindex demangling, from command line
1718 @kindex --demangle[=@var{style}]
1719 @kindex --no-demangle
1720 @item --demangle[=@var{style}]
1721 @itemx --no-demangle
1722 These options control whether to demangle symbol names in error messages
1723 and other output. When the linker is told to demangle, it tries to
1724 present symbol names in a readable fashion: it strips leading
1725 underscores if they are used by the object file format, and converts C++
1726 mangled symbol names into user readable names. Different compilers have
1727 different mangling styles. The optional demangling style argument can be used
1728 to choose an appropriate demangling style for your compiler. The linker will
1729 demangle by default unless the environment variable @samp{COLLECT_NO_DEMANGLE}
1730 is set. These options may be used to override the default.
1732 @cindex dynamic linker, from command line
1733 @kindex -I@var{file}
1734 @kindex --dynamic-linker=@var{file}
1736 @itemx --dynamic-linker=@var{file}
1737 Set the name of the dynamic linker. This is only meaningful when
1738 generating dynamically linked ELF executables. The default dynamic
1739 linker is normally correct; don't use this unless you know what you are
1742 @kindex --no-dynamic-linker
1743 @item --no-dynamic-linker
1744 When producing an executable file, omit the request for a dynamic
1745 linker to be used at load-time. This is only meaningful for ELF
1746 executables that contain dynamic relocations, and usually requires
1747 entry point code that is capable of processing these relocations.
1749 @kindex --embedded-relocs
1750 @item --embedded-relocs
1751 This option is similar to the @option{--emit-relocs} option except
1752 that the relocs are stored in a target-specific section. This option
1753 is only supported by the @samp{BFIN}, @samp{CR16} and @emph{M68K}
1756 @kindex --disable-multiple-abs-defs
1757 @item --disable-multiple-abs-defs
1758 Do not allow multiple definitions with symbols included
1759 in filename invoked by -R or --just-symbols
1761 @kindex --fatal-warnings
1762 @kindex --no-fatal-warnings
1763 @item --fatal-warnings
1764 @itemx --no-fatal-warnings
1765 Treat all warnings as errors. The default behaviour can be restored
1766 with the option @option{--no-fatal-warnings}.
1768 @kindex --force-exe-suffix
1769 @item --force-exe-suffix
1770 Make sure that an output file has a .exe suffix.
1772 If a successfully built fully linked output file does not have a
1773 @code{.exe} or @code{.dll} suffix, this option forces the linker to copy
1774 the output file to one of the same name with a @code{.exe} suffix. This
1775 option is useful when using unmodified Unix makefiles on a Microsoft
1776 Windows host, since some versions of Windows won't run an image unless
1777 it ends in a @code{.exe} suffix.
1779 @kindex --gc-sections
1780 @kindex --no-gc-sections
1781 @cindex garbage collection
1783 @itemx --no-gc-sections
1784 Enable garbage collection of unused input sections. It is ignored on
1785 targets that do not support this option. The default behaviour (of not
1786 performing this garbage collection) can be restored by specifying
1787 @samp{--no-gc-sections} on the command line. Note that garbage
1788 collection for COFF and PE format targets is supported, but the
1789 implementation is currently considered to be experimental.
1791 @samp{--gc-sections} decides which input sections are used by
1792 examining symbols and relocations. The section containing the entry
1793 symbol and all sections containing symbols undefined on the
1794 command-line will be kept, as will sections containing symbols
1795 referenced by dynamic objects. Note that when building shared
1796 libraries, the linker must assume that any visible symbol is
1797 referenced. Once this initial set of sections has been determined,
1798 the linker recursively marks as used any section referenced by their
1799 relocations. See @samp{--entry}, @samp{--undefined}, and
1800 @samp{--gc-keep-exported}.
1802 This option can be set when doing a partial link (enabled with option
1803 @samp{-r}). In this case the root of symbols kept must be explicitly
1804 specified either by one of the options @samp{--entry},
1805 @samp{--undefined}, or @samp{--gc-keep-exported} or by a @code{ENTRY}
1806 command in the linker script.
1808 @kindex --print-gc-sections
1809 @kindex --no-print-gc-sections
1810 @cindex garbage collection
1811 @item --print-gc-sections
1812 @itemx --no-print-gc-sections
1813 List all sections removed by garbage collection. The listing is
1814 printed on stderr. This option is only effective if garbage
1815 collection has been enabled via the @samp{--gc-sections}) option. The
1816 default behaviour (of not listing the sections that are removed) can
1817 be restored by specifying @samp{--no-print-gc-sections} on the command
1820 @kindex --gc-keep-exported
1821 @cindex garbage collection
1822 @item --gc-keep-exported
1823 When @samp{--gc-sections} is enabled, this option prevents garbage
1824 collection of unused input sections that contain global symbols having
1825 default or protected visibility. This option is intended to be used for
1826 executables where unreferenced sections would otherwise be garbage
1827 collected regardless of the external visibility of contained symbols.
1828 Note that this option has no effect when linking shared objects since
1829 it is already the default behaviour. This option is only supported for
1832 @kindex --print-output-format
1833 @cindex output format
1834 @item --print-output-format
1835 Print the name of the default output format (perhaps influenced by
1836 other command-line options). This is the string that would appear
1837 in an @code{OUTPUT_FORMAT} linker script command (@pxref{File Commands}).
1839 @kindex --print-memory-usage
1840 @cindex memory usage
1841 @item --print-memory-usage
1842 Print used size, total size and used size of memory regions created with
1843 the @ref{MEMORY} command. This is useful on embedded targets to have a
1844 quick view of amount of free memory. The format of the output has one
1845 headline and one line per region. It is both human readable and easily
1846 parsable by tools. Here is an example of an output:
1849 Memory region Used Size Region Size %age Used
1850 ROM: 256 KB 1 MB 25.00%
1851 RAM: 32 B 2 GB 0.00%
1858 Print a summary of the command-line options on the standard output and exit.
1860 @kindex --target-help
1862 Print a summary of all target-specific options on the standard output and exit.
1864 @kindex -Map=@var{mapfile}
1865 @item -Map=@var{mapfile}
1866 Print a link map to the file @var{mapfile}. See the description of the
1867 @option{-M} option, above. Specifying a directory as @var{mapfile}
1868 causes the linker map to be written into a file inside the directory.
1869 The name of the file is based upon the @var{output} filename with
1870 @code{.map} appended.
1872 @cindex memory usage
1873 @kindex --no-keep-memory
1874 @item --no-keep-memory
1875 @command{ld} normally optimizes for speed over memory usage by caching the
1876 symbol tables of input files in memory. This option tells @command{ld} to
1877 instead optimize for memory usage, by rereading the symbol tables as
1878 necessary. This may be required if @command{ld} runs out of memory space
1879 while linking a large executable.
1881 @kindex --no-undefined
1884 @item --no-undefined
1886 Report unresolved symbol references from regular object files. This
1887 is done even if the linker is creating a non-symbolic shared library.
1888 The switch @option{--[no-]allow-shlib-undefined} controls the
1889 behaviour for reporting unresolved references found in shared
1890 libraries being linked in.
1892 The effects of this option can be reverted by using @code{-z undefs}.
1894 @kindex --allow-multiple-definition
1896 @item --allow-multiple-definition
1898 Normally when a symbol is defined multiple times, the linker will
1899 report a fatal error. These options allow multiple definitions and the
1900 first definition will be used.
1902 @kindex --allow-shlib-undefined
1903 @kindex --no-allow-shlib-undefined
1904 @item --allow-shlib-undefined
1905 @itemx --no-allow-shlib-undefined
1906 Allows or disallows undefined symbols in shared libraries.
1907 This switch is similar to @option{--no-undefined} except that it
1908 determines the behaviour when the undefined symbols are in a
1909 shared library rather than a regular object file. It does not affect
1910 how undefined symbols in regular object files are handled.
1912 The default behaviour is to report errors for any undefined symbols
1913 referenced in shared libraries if the linker is being used to create
1914 an executable, but to allow them if the linker is being used to create
1917 The reasons for allowing undefined symbol references in shared
1918 libraries specified at link time are that:
1922 A shared library specified at link time may not be the same as the one
1923 that is available at load time, so the symbol might actually be
1924 resolvable at load time.
1926 There are some operating systems, eg BeOS and HPPA, where undefined
1927 symbols in shared libraries are normal.
1929 The BeOS kernel for example patches shared libraries at load time to
1930 select whichever function is most appropriate for the current
1931 architecture. This is used, for example, to dynamically select an
1932 appropriate memset function.
1935 @kindex --error-handling-script=@var{scriptname}
1936 @item --error-handling-script=@var{scriptname}
1937 If this option is provided then the linker will invoke
1938 @var{scriptname} whenever an error is encountered. Currently however
1939 only two kinds of error are supported: missing symbols and missing
1940 libraries. Two arguments will be passed to script: the keyword
1941 ``missing-symbol'' or `missing-lib'' and the @var{name} of the
1942 missing symbol or library. The intention is that the script will
1943 provide suggestions to the user as to where the symbol or library
1944 might be found. After the script has finished then the normal linker
1945 error message will be displayed.
1947 The availability of this option is controlled by a configure time
1948 switch, so it may not be present in specific implementations.
1950 @kindex --no-undefined-version
1951 @item --no-undefined-version
1952 Normally when a symbol has an undefined version, the linker will ignore
1953 it. This option disallows symbols with undefined version and a fatal error
1954 will be issued instead.
1956 @kindex --default-symver
1957 @item --default-symver
1958 Create and use a default symbol version (the soname) for unversioned
1961 @kindex --default-imported-symver
1962 @item --default-imported-symver
1963 Create and use a default symbol version (the soname) for unversioned
1966 @kindex --no-warn-mismatch
1967 @item --no-warn-mismatch
1968 Normally @command{ld} will give an error if you try to link together input
1969 files that are mismatched for some reason, perhaps because they have
1970 been compiled for different processors or for different endiannesses.
1971 This option tells @command{ld} that it should silently permit such possible
1972 errors. This option should only be used with care, in cases when you
1973 have taken some special action that ensures that the linker errors are
1976 @kindex --no-warn-search-mismatch
1977 @item --no-warn-search-mismatch
1978 Normally @command{ld} will give a warning if it finds an incompatible
1979 library during a library search. This option silences the warning.
1981 @kindex --no-whole-archive
1982 @item --no-whole-archive
1983 Turn off the effect of the @option{--whole-archive} option for subsequent
1986 @cindex output file after errors
1987 @kindex --noinhibit-exec
1988 @item --noinhibit-exec
1989 Retain the executable output file whenever it is still usable.
1990 Normally, the linker will not produce an output file if it encounters
1991 errors during the link process; it exits without writing an output file
1992 when it issues any error whatsoever.
1996 Only search library directories explicitly specified on the
1997 command line. Library directories specified in linker scripts
1998 (including linker scripts specified on the command line) are ignored.
2000 @ifclear SingleFormat
2001 @kindex --oformat=@var{output-format}
2002 @item --oformat=@var{output-format}
2003 @command{ld} may be configured to support more than one kind of object
2004 file. If your @command{ld} is configured this way, you can use the
2005 @samp{--oformat} option to specify the binary format for the output
2006 object file. Even when @command{ld} is configured to support alternative
2007 object formats, you don't usually need to specify this, as @command{ld}
2008 should be configured to produce as a default output format the most
2009 usual format on each machine. @var{output-format} is a text string, the
2010 name of a particular format supported by the BFD libraries. (You can
2011 list the available binary formats with @samp{objdump -i}.) The script
2012 command @code{OUTPUT_FORMAT} can also specify the output format, but
2013 this option overrides it. @xref{BFD}.
2016 @kindex --out-implib
2017 @item --out-implib @var{file}
2018 Create an import library in @var{file} corresponding to the executable
2019 the linker is generating (eg. a DLL or ELF program). This import
2020 library (which should be called @code{*.dll.a} or @code{*.a} for DLLs)
2021 may be used to link clients against the generated executable; this
2022 behaviour makes it possible to skip a separate import library creation
2023 step (eg. @code{dlltool} for DLLs). This option is only available for
2024 the i386 PE and ELF targetted ports of the linker.
2027 @kindex --pic-executable
2029 @itemx --pic-executable
2030 @cindex position independent executables
2031 Create a position independent executable. This is currently only supported on
2032 ELF platforms. Position independent executables are similar to shared
2033 libraries in that they are relocated by the dynamic linker to the virtual
2034 address the OS chooses for them (which can vary between invocations). Like
2035 normal dynamically linked executables they can be executed and symbols
2036 defined in the executable cannot be overridden by shared libraries.
2040 This option is ignored for Linux compatibility.
2044 This option is ignored for SVR4 compatibility.
2047 @cindex synthesizing linker
2048 @cindex relaxing addressing modes
2052 An option with machine dependent effects.
2054 This option is only supported on a few targets.
2057 @xref{H8/300,,@command{ld} and the H8/300}.
2060 @xref{Xtensa,, @command{ld} and Xtensa Processors}.
2063 @xref{M68HC11/68HC12,,@command{ld} and the 68HC11 and 68HC12}.
2066 @xref{Nios II,,@command{ld} and the Altera Nios II}.
2069 @xref{PowerPC ELF32,,@command{ld} and PowerPC 32-bit ELF Support}.
2072 On some platforms the @option{--relax} option performs target specific,
2073 global optimizations that become possible when the linker resolves
2074 addressing in the program, such as relaxing address modes,
2075 synthesizing new instructions, selecting shorter version of current
2076 instructions, and combining constant values.
2078 On some platforms these link time global optimizations may make symbolic
2079 debugging of the resulting executable impossible.
2081 This is known to be the case for the Matsushita MN10200 and MN10300
2082 family of processors.
2085 On platforms where the feature is supported, the option
2086 @option{--no-relax} will disable it.
2088 On platforms where the feature is not supported, both @option{--relax}
2089 and @option{--no-relax} are accepted, but ignored.
2091 @cindex retaining specified symbols
2092 @cindex stripping all but some symbols
2093 @cindex symbols, retaining selectively
2094 @kindex --retain-symbols-file=@var{filename}
2095 @item --retain-symbols-file=@var{filename}
2096 Retain @emph{only} the symbols listed in the file @var{filename},
2097 discarding all others. @var{filename} is simply a flat file, with one
2098 symbol name per line. This option is especially useful in environments
2102 where a large global symbol table is accumulated gradually, to conserve
2105 @samp{--retain-symbols-file} does @emph{not} discard undefined symbols,
2106 or symbols needed for relocations.
2108 You may only specify @samp{--retain-symbols-file} once in the command
2109 line. It overrides @samp{-s} and @samp{-S}.
2112 @item -rpath=@var{dir}
2113 @cindex runtime library search path
2114 @kindex -rpath=@var{dir}
2115 Add a directory to the runtime library search path. This is used when
2116 linking an ELF executable with shared objects. All @option{-rpath}
2117 arguments are concatenated and passed to the runtime linker, which uses
2118 them to locate shared objects at runtime.
2120 The @option{-rpath} option is also used when locating shared objects which
2121 are needed by shared objects explicitly included in the link; see the
2122 description of the @option{-rpath-link} option. Searching @option{-rpath}
2123 in this way is only supported by native linkers and cross linkers which
2124 have been configured with the @option{--with-sysroot} option.
2126 If @option{-rpath} is not used when linking an ELF executable, the
2127 contents of the environment variable @code{LD_RUN_PATH} will be used if it
2130 The @option{-rpath} option may also be used on SunOS. By default, on
2131 SunOS, the linker will form a runtime search path out of all the
2132 @option{-L} options it is given. If a @option{-rpath} option is used, the
2133 runtime search path will be formed exclusively using the @option{-rpath}
2134 options, ignoring the @option{-L} options. This can be useful when using
2135 gcc, which adds many @option{-L} options which may be on NFS mounted
2138 For compatibility with other ELF linkers, if the @option{-R} option is
2139 followed by a directory name, rather than a file name, it is treated as
2140 the @option{-rpath} option.
2144 @cindex link-time runtime library search path
2145 @kindex -rpath-link=@var{dir}
2146 @item -rpath-link=@var{dir}
2147 When using ELF or SunOS, one shared library may require another. This
2148 happens when an @code{ld -shared} link includes a shared library as one
2151 When the linker encounters such a dependency when doing a non-shared,
2152 non-relocatable link, it will automatically try to locate the required
2153 shared library and include it in the link, if it is not included
2154 explicitly. In such a case, the @option{-rpath-link} option
2155 specifies the first set of directories to search. The
2156 @option{-rpath-link} option may specify a sequence of directory names
2157 either by specifying a list of names separated by colons, or by
2158 appearing multiple times.
2160 The tokens @var{$ORIGIN} and @var{$LIB} can appear in these search
2161 directories. They will be replaced by the full path to the directory
2162 containing the program or shared object in the case of @var{$ORIGIN}
2163 and either @samp{lib} - for 32-bit binaries - or @samp{lib64} - for
2164 64-bit binaries - in the case of @var{$LIB}.
2166 The alternative form of these tokens - @var{$@{ORIGIN@}} and
2167 @var{$@{LIB@}} can also be used. The token @var{$PLATFORM} is not
2170 This option should be used with caution as it overrides the search path
2171 that may have been hard compiled into a shared library. In such a case it
2172 is possible to use unintentionally a different search path than the
2173 runtime linker would do.
2175 The linker uses the following search paths to locate required shared
2180 Any directories specified by @option{-rpath-link} options.
2182 Any directories specified by @option{-rpath} options. The difference
2183 between @option{-rpath} and @option{-rpath-link} is that directories
2184 specified by @option{-rpath} options are included in the executable and
2185 used at runtime, whereas the @option{-rpath-link} option is only effective
2186 at link time. Searching @option{-rpath} in this way is only supported
2187 by native linkers and cross linkers which have been configured with
2188 the @option{--with-sysroot} option.
2190 On an ELF system, for native linkers, if the @option{-rpath} and
2191 @option{-rpath-link} options were not used, search the contents of the
2192 environment variable @code{LD_RUN_PATH}.
2194 On SunOS, if the @option{-rpath} option was not used, search any
2195 directories specified using @option{-L} options.
2197 For a native linker, search the contents of the environment
2198 variable @code{LD_LIBRARY_PATH}.
2200 For a native ELF linker, the directories in @code{DT_RUNPATH} or
2201 @code{DT_RPATH} of a shared library are searched for shared
2202 libraries needed by it. The @code{DT_RPATH} entries are ignored if
2203 @code{DT_RUNPATH} entries exist.
2205 The default directories, normally @file{/lib} and @file{/usr/lib}.
2207 For a linker for a Linux system, if the file @file{/etc/ld.so.conf}
2208 exists, the list of directories found in that file. Note: the path
2209 to this file is prefixed with the @code{sysroot} value, if that is
2210 defined, and then any @code{prefix} string if the linker was
2211 configured with the @command{--prefix=<path>} option.
2213 For a native linker on a FreeBSD system, any directories specified by
2214 the @code{_PATH_ELF_HINTS} macro defined in the @file{elf-hints.h}
2217 Any directories specifed by a @code{SEARCH_DIR} command in the
2218 linker script being used.
2221 If the required shared library is not found, the linker will issue a
2222 warning and continue with the link.
2229 @cindex shared libraries
2230 Create a shared library. This is currently only supported on ELF, XCOFF
2231 and SunOS platforms. On SunOS, the linker will automatically create a
2232 shared library if the @option{-e} option is not used and there are
2233 undefined symbols in the link.
2235 @kindex --sort-common
2237 @itemx --sort-common=ascending
2238 @itemx --sort-common=descending
2239 This option tells @command{ld} to sort the common symbols by alignment in
2240 ascending or descending order when it places them in the appropriate output
2241 sections. The symbol alignments considered are sixteen-byte or larger,
2242 eight-byte, four-byte, two-byte, and one-byte. This is to prevent gaps
2243 between symbols due to alignment constraints. If no sorting order is
2244 specified, then descending order is assumed.
2246 @kindex --sort-section=name
2247 @item --sort-section=name
2248 This option will apply @code{SORT_BY_NAME} to all wildcard section
2249 patterns in the linker script.
2251 @kindex --sort-section=alignment
2252 @item --sort-section=alignment
2253 This option will apply @code{SORT_BY_ALIGNMENT} to all wildcard section
2254 patterns in the linker script.
2256 @kindex --spare-dynamic-tags
2257 @item --spare-dynamic-tags=@var{count}
2258 This option specifies the number of empty slots to leave in the
2259 .dynamic section of ELF shared objects. Empty slots may be needed by
2260 post processing tools, such as the prelinker. The default is 5.
2262 @kindex --split-by-file
2263 @item --split-by-file[=@var{size}]
2264 Similar to @option{--split-by-reloc} but creates a new output section for
2265 each input file when @var{size} is reached. @var{size} defaults to a
2266 size of 1 if not given.
2268 @kindex --split-by-reloc
2269 @item --split-by-reloc[=@var{count}]
2270 Tries to creates extra sections in the output file so that no single
2271 output section in the file contains more than @var{count} relocations.
2272 This is useful when generating huge relocatable files for downloading into
2273 certain real time kernels with the COFF object file format; since COFF
2274 cannot represent more than 65535 relocations in a single section. Note
2275 that this will fail to work with object file formats which do not
2276 support arbitrary sections. The linker will not split up individual
2277 input sections for redistribution, so if a single input section contains
2278 more than @var{count} relocations one output section will contain that
2279 many relocations. @var{count} defaults to a value of 32768.
2283 Compute and display statistics about the operation of the linker, such
2284 as execution time and memory usage.
2286 @kindex --sysroot=@var{directory}
2287 @item --sysroot=@var{directory}
2288 Use @var{directory} as the location of the sysroot, overriding the
2289 configure-time default. This option is only supported by linkers
2290 that were configured using @option{--with-sysroot}.
2294 This is used by COFF/PE based targets to create a task-linked object
2295 file where all of the global symbols have been converted to statics.
2297 @kindex --traditional-format
2298 @cindex traditional format
2299 @item --traditional-format
2300 For some targets, the output of @command{ld} is different in some ways from
2301 the output of some existing linker. This switch requests @command{ld} to
2302 use the traditional format instead.
2305 For example, on SunOS, @command{ld} combines duplicate entries in the
2306 symbol string table. This can reduce the size of an output file with
2307 full debugging information by over 30 percent. Unfortunately, the SunOS
2308 @code{dbx} program can not read the resulting program (@code{gdb} has no
2309 trouble). The @samp{--traditional-format} switch tells @command{ld} to not
2310 combine duplicate entries.
2312 @kindex --section-start=@var{sectionname}=@var{org}
2313 @item --section-start=@var{sectionname}=@var{org}
2314 Locate a section in the output file at the absolute
2315 address given by @var{org}. You may use this option as many
2316 times as necessary to locate multiple sections in the command
2318 @var{org} must be a single hexadecimal integer;
2319 for compatibility with other linkers, you may omit the leading
2320 @samp{0x} usually associated with hexadecimal values. @emph{Note:} there
2321 should be no white space between @var{sectionname}, the equals
2322 sign (``@key{=}''), and @var{org}.
2324 @kindex -Tbss=@var{org}
2325 @kindex -Tdata=@var{org}
2326 @kindex -Ttext=@var{org}
2327 @cindex segment origins, cmd line
2328 @item -Tbss=@var{org}
2329 @itemx -Tdata=@var{org}
2330 @itemx -Ttext=@var{org}
2331 Same as @option{--section-start}, with @code{.bss}, @code{.data} or
2332 @code{.text} as the @var{sectionname}.
2334 @kindex -Ttext-segment=@var{org}
2335 @item -Ttext-segment=@var{org}
2336 @cindex text segment origin, cmd line
2337 When creating an ELF executable, it will set the address of the first
2338 byte of the text segment.
2340 @kindex -Trodata-segment=@var{org}
2341 @item -Trodata-segment=@var{org}
2342 @cindex rodata segment origin, cmd line
2343 When creating an ELF executable or shared object for a target where
2344 the read-only data is in its own segment separate from the executable
2345 text, it will set the address of the first byte of the read-only data segment.
2347 @kindex -Tldata-segment=@var{org}
2348 @item -Tldata-segment=@var{org}
2349 @cindex ldata segment origin, cmd line
2350 When creating an ELF executable or shared object for x86-64 medium memory
2351 model, it will set the address of the first byte of the ldata segment.
2353 @kindex --unresolved-symbols
2354 @item --unresolved-symbols=@var{method}
2355 Determine how to handle unresolved symbols. There are four possible
2356 values for @samp{method}:
2360 Do not report any unresolved symbols.
2363 Report all unresolved symbols. This is the default.
2365 @item ignore-in-object-files
2366 Report unresolved symbols that are contained in shared libraries, but
2367 ignore them if they come from regular object files.
2369 @item ignore-in-shared-libs
2370 Report unresolved symbols that come from regular object files, but
2371 ignore them if they come from shared libraries. This can be useful
2372 when creating a dynamic binary and it is known that all the shared
2373 libraries that it should be referencing are included on the linker's
2377 The behaviour for shared libraries on their own can also be controlled
2378 by the @option{--[no-]allow-shlib-undefined} option.
2380 Normally the linker will generate an error message for each reported
2381 unresolved symbol but the option @option{--warn-unresolved-symbols}
2382 can change this to a warning.
2384 @kindex --verbose[=@var{NUMBER}]
2385 @cindex verbose[=@var{NUMBER}]
2387 @itemx --verbose[=@var{NUMBER}]
2388 Display the version number for @command{ld} and list the linker emulations
2389 supported. Display which input files can and cannot be opened. Display
2390 the linker script being used by the linker. If the optional @var{NUMBER}
2391 argument > 1, plugin symbol status will also be displayed.
2393 @kindex --version-script=@var{version-scriptfile}
2394 @cindex version script, symbol versions
2395 @item --version-script=@var{version-scriptfile}
2396 Specify the name of a version script to the linker. This is typically
2397 used when creating shared libraries to specify additional information
2398 about the version hierarchy for the library being created. This option
2399 is only fully supported on ELF platforms which support shared libraries;
2400 see @ref{VERSION}. It is partially supported on PE platforms, which can
2401 use version scripts to filter symbol visibility in auto-export mode: any
2402 symbols marked @samp{local} in the version script will not be exported.
2405 @kindex --warn-common
2406 @cindex warnings, on combining symbols
2407 @cindex combining symbols, warnings on
2409 Warn when a common symbol is combined with another common symbol or with
2410 a symbol definition. Unix linkers allow this somewhat sloppy practice,
2411 but linkers on some other operating systems do not. This option allows
2412 you to find potential problems from combining global symbols.
2413 Unfortunately, some C libraries use this practice, so you may get some
2414 warnings about symbols in the libraries as well as in your programs.
2416 There are three kinds of global symbols, illustrated here by C examples:
2420 A definition, which goes in the initialized data section of the output
2424 An undefined reference, which does not allocate space.
2425 There must be either a definition or a common symbol for the
2429 A common symbol. If there are only (one or more) common symbols for a
2430 variable, it goes in the uninitialized data area of the output file.
2431 The linker merges multiple common symbols for the same variable into a
2432 single symbol. If they are of different sizes, it picks the largest
2433 size. The linker turns a common symbol into a declaration, if there is
2434 a definition of the same variable.
2437 The @samp{--warn-common} option can produce five kinds of warnings.
2438 Each warning consists of a pair of lines: the first describes the symbol
2439 just encountered, and the second describes the previous symbol
2440 encountered with the same name. One or both of the two symbols will be
2445 Turning a common symbol into a reference, because there is already a
2446 definition for the symbol.
2448 @var{file}(@var{section}): warning: common of `@var{symbol}'
2449 overridden by definition
2450 @var{file}(@var{section}): warning: defined here
2454 Turning a common symbol into a reference, because a later definition for
2455 the symbol is encountered. This is the same as the previous case,
2456 except that the symbols are encountered in a different order.
2458 @var{file}(@var{section}): warning: definition of `@var{symbol}'
2460 @var{file}(@var{section}): warning: common is here
2464 Merging a common symbol with a previous same-sized common symbol.
2466 @var{file}(@var{section}): warning: multiple common
2468 @var{file}(@var{section}): warning: previous common is here
2472 Merging a common symbol with a previous larger common symbol.
2474 @var{file}(@var{section}): warning: common of `@var{symbol}'
2475 overridden by larger common
2476 @var{file}(@var{section}): warning: larger common is here
2480 Merging a common symbol with a previous smaller common symbol. This is
2481 the same as the previous case, except that the symbols are
2482 encountered in a different order.
2484 @var{file}(@var{section}): warning: common of `@var{symbol}'
2485 overriding smaller common
2486 @var{file}(@var{section}): warning: smaller common is here
2490 @kindex --warn-constructors
2491 @item --warn-constructors
2492 Warn if any global constructors are used. This is only useful for a few
2493 object file formats. For formats like COFF or ELF, the linker can not
2494 detect the use of global constructors.
2496 @kindex --warn-multiple-gp
2497 @item --warn-multiple-gp
2498 Warn if multiple global pointer values are required in the output file.
2499 This is only meaningful for certain processors, such as the Alpha.
2500 Specifically, some processors put large-valued constants in a special
2501 section. A special register (the global pointer) points into the middle
2502 of this section, so that constants can be loaded efficiently via a
2503 base-register relative addressing mode. Since the offset in
2504 base-register relative mode is fixed and relatively small (e.g., 16
2505 bits), this limits the maximum size of the constant pool. Thus, in
2506 large programs, it is often necessary to use multiple global pointer
2507 values in order to be able to address all possible constants. This
2508 option causes a warning to be issued whenever this case occurs.
2511 @cindex warnings, on undefined symbols
2512 @cindex undefined symbols, warnings on
2514 Only warn once for each undefined symbol, rather than once per module
2517 @kindex --warn-section-align
2518 @cindex warnings, on section alignment
2519 @cindex section alignment, warnings on
2520 @item --warn-section-align
2521 Warn if the address of an output section is changed because of
2522 alignment. Typically, the alignment will be set by an input section.
2523 The address will only be changed if it not explicitly specified; that
2524 is, if the @code{SECTIONS} command does not specify a start address for
2525 the section (@pxref{SECTIONS}).
2527 @kindex --warn-textrel
2528 @item --warn-textrel
2529 Warn if the linker adds DT_TEXTREL to a position-independent executable
2532 @kindex --warn-alternate-em
2533 @item --warn-alternate-em
2534 Warn if an object has alternate ELF machine code.
2536 @kindex --warn-unresolved-symbols
2537 @item --warn-unresolved-symbols
2538 If the linker is going to report an unresolved symbol (see the option
2539 @option{--unresolved-symbols}) it will normally generate an error.
2540 This option makes it generate a warning instead.
2542 @kindex --error-unresolved-symbols
2543 @item --error-unresolved-symbols
2544 This restores the linker's default behaviour of generating errors when
2545 it is reporting unresolved symbols.
2547 @kindex --whole-archive
2548 @cindex including an entire archive
2549 @item --whole-archive
2550 For each archive mentioned on the command line after the
2551 @option{--whole-archive} option, include every object file in the archive
2552 in the link, rather than searching the archive for the required object
2553 files. This is normally used to turn an archive file into a shared
2554 library, forcing every object to be included in the resulting shared
2555 library. This option may be used more than once.
2557 Two notes when using this option from gcc: First, gcc doesn't know
2558 about this option, so you have to use @option{-Wl,-whole-archive}.
2559 Second, don't forget to use @option{-Wl,-no-whole-archive} after your
2560 list of archives, because gcc will add its own list of archives to
2561 your link and you may not want this flag to affect those as well.
2563 @kindex --wrap=@var{symbol}
2564 @item --wrap=@var{symbol}
2565 Use a wrapper function for @var{symbol}. Any undefined reference to
2566 @var{symbol} will be resolved to @code{__wrap_@var{symbol}}. Any
2567 undefined reference to @code{__real_@var{symbol}} will be resolved to
2570 This can be used to provide a wrapper for a system function. The
2571 wrapper function should be called @code{__wrap_@var{symbol}}. If it
2572 wishes to call the system function, it should call
2573 @code{__real_@var{symbol}}.
2575 Here is a trivial example:
2579 __wrap_malloc (size_t c)
2581 printf ("malloc called with %zu\n", c);
2582 return __real_malloc (c);
2586 If you link other code with this file using @option{--wrap malloc}, then
2587 all calls to @code{malloc} will call the function @code{__wrap_malloc}
2588 instead. The call to @code{__real_malloc} in @code{__wrap_malloc} will
2589 call the real @code{malloc} function.
2591 You may wish to provide a @code{__real_malloc} function as well, so that
2592 links without the @option{--wrap} option will succeed. If you do this,
2593 you should not put the definition of @code{__real_malloc} in the same
2594 file as @code{__wrap_malloc}; if you do, the assembler may resolve the
2595 call before the linker has a chance to wrap it to @code{malloc}.
2597 Only undefined references are replaced by the linker. So, translation unit
2598 internal references to @var{symbol} are not resolved to
2599 @code{__wrap_@var{symbol}}. In the next example, the call to @code{f} in
2600 @code{g} is not resolved to @code{__wrap_f}.
2616 @kindex --eh-frame-hdr
2617 @kindex --no-eh-frame-hdr
2618 @item --eh-frame-hdr
2619 @itemx --no-eh-frame-hdr
2620 Request (@option{--eh-frame-hdr}) or suppress
2621 (@option{--no-eh-frame-hdr}) the creation of @code{.eh_frame_hdr}
2622 section and ELF @code{PT_GNU_EH_FRAME} segment header.
2624 @kindex --ld-generated-unwind-info
2625 @item --no-ld-generated-unwind-info
2626 Request creation of @code{.eh_frame} unwind info for linker
2627 generated code sections like PLT. This option is on by default
2628 if linker generated unwind info is supported.
2630 @kindex --enable-new-dtags
2631 @kindex --disable-new-dtags
2632 @item --enable-new-dtags
2633 @itemx --disable-new-dtags
2634 This linker can create the new dynamic tags in ELF. But the older ELF
2635 systems may not understand them. If you specify
2636 @option{--enable-new-dtags}, the new dynamic tags will be created as needed
2637 and older dynamic tags will be omitted.
2638 If you specify @option{--disable-new-dtags}, no new dynamic tags will be
2639 created. By default, the new dynamic tags are not created. Note that
2640 those options are only available for ELF systems.
2642 @kindex --hash-size=@var{number}
2643 @item --hash-size=@var{number}
2644 Set the default size of the linker's hash tables to a prime number
2645 close to @var{number}. Increasing this value can reduce the length of
2646 time it takes the linker to perform its tasks, at the expense of
2647 increasing the linker's memory requirements. Similarly reducing this
2648 value can reduce the memory requirements at the expense of speed.
2650 @kindex --hash-style=@var{style}
2651 @item --hash-style=@var{style}
2652 Set the type of linker's hash table(s). @var{style} can be either
2653 @code{sysv} for classic ELF @code{.hash} section, @code{gnu} for
2654 new style GNU @code{.gnu.hash} section or @code{both} for both
2655 the classic ELF @code{.hash} and new style GNU @code{.gnu.hash}
2656 hash tables. The default depends upon how the linker was configured,
2657 but for most Linux based systems it will be @code{both}.
2659 @kindex --compress-debug-sections=none
2660 @kindex --compress-debug-sections=zlib
2661 @kindex --compress-debug-sections=zlib-gnu
2662 @kindex --compress-debug-sections=zlib-gabi
2663 @item --compress-debug-sections=none
2664 @itemx --compress-debug-sections=zlib
2665 @itemx --compress-debug-sections=zlib-gnu
2666 @itemx --compress-debug-sections=zlib-gabi
2667 On ELF platforms, these options control how DWARF debug sections are
2668 compressed using zlib.
2670 @option{--compress-debug-sections=none} doesn't compress DWARF debug
2671 sections. @option{--compress-debug-sections=zlib-gnu} compresses
2672 DWARF debug sections and renames them to begin with @samp{.zdebug}
2673 instead of @samp{.debug}. @option{--compress-debug-sections=zlib-gabi}
2674 also compresses DWARF debug sections, but rather than renaming them it
2675 sets the SHF_COMPRESSED flag in the sections' headers.
2677 The @option{--compress-debug-sections=zlib} option is an alias for
2678 @option{--compress-debug-sections=zlib-gabi}.
2680 Note that this option overrides any compression in input debug
2681 sections, so if a binary is linked with @option{--compress-debug-sections=none}
2682 for example, then any compressed debug sections in input files will be
2683 uncompressed before they are copied into the output binary.
2685 The default compression behaviour varies depending upon the target
2686 involved and the configure options used to build the toolchain. The
2687 default can be determined by examining the output from the linker's
2688 @option{--help} option.
2690 @kindex --reduce-memory-overheads
2691 @item --reduce-memory-overheads
2692 This option reduces memory requirements at ld runtime, at the expense of
2693 linking speed. This was introduced to select the old O(n^2) algorithm
2694 for link map file generation, rather than the new O(n) algorithm which uses
2695 about 40% more memory for symbol storage.
2697 Another effect of the switch is to set the default hash table size to
2698 1021, which again saves memory at the cost of lengthening the linker's
2699 run time. This is not done however if the @option{--hash-size} switch
2702 The @option{--reduce-memory-overheads} switch may be also be used to
2703 enable other tradeoffs in future versions of the linker.
2706 @kindex --build-id=@var{style}
2708 @itemx --build-id=@var{style}
2709 Request the creation of a @code{.note.gnu.build-id} ELF note section
2710 or a @code{.buildid} COFF section. The contents of the note are
2711 unique bits identifying this linked file. @var{style} can be
2712 @code{uuid} to use 128 random bits, @code{sha1} to use a 160-bit
2713 @sc{SHA1} hash on the normative parts of the output contents,
2714 @code{md5} to use a 128-bit @sc{MD5} hash on the normative parts of
2715 the output contents, or @code{0x@var{hexstring}} to use a chosen bit
2716 string specified as an even number of hexadecimal digits (@code{-} and
2717 @code{:} characters between digit pairs are ignored). If @var{style}
2718 is omitted, @code{sha1} is used.
2720 The @code{md5} and @code{sha1} styles produces an identifier
2721 that is always the same in an identical output file, but will be
2722 unique among all nonidentical output files. It is not intended
2723 to be compared as a checksum for the file's contents. A linked
2724 file may be changed later by other tools, but the build ID bit
2725 string identifying the original linked file does not change.
2727 Passing @code{none} for @var{style} disables the setting from any
2728 @code{--build-id} options earlier on the command line.
2733 @subsection Options Specific to i386 PE Targets
2735 @c man begin OPTIONS
2737 The i386 PE linker supports the @option{-shared} option, which causes
2738 the output to be a dynamically linked library (DLL) instead of a
2739 normal executable. You should name the output @code{*.dll} when you
2740 use this option. In addition, the linker fully supports the standard
2741 @code{*.def} files, which may be specified on the linker command line
2742 like an object file (in fact, it should precede archives it exports
2743 symbols from, to ensure that they get linked in, just like a normal
2746 In addition to the options common to all targets, the i386 PE linker
2747 support additional command-line options that are specific to the i386
2748 PE target. Options that take values may be separated from their
2749 values by either a space or an equals sign.
2753 @kindex --add-stdcall-alias
2754 @item --add-stdcall-alias
2755 If given, symbols with a stdcall suffix (@@@var{nn}) will be exported
2756 as-is and also with the suffix stripped.
2757 [This option is specific to the i386 PE targeted port of the linker]
2760 @item --base-file @var{file}
2761 Use @var{file} as the name of a file in which to save the base
2762 addresses of all the relocations needed for generating DLLs with
2764 [This is an i386 PE specific option]
2768 Create a DLL instead of a regular executable. You may also use
2769 @option{-shared} or specify a @code{LIBRARY} in a given @code{.def}
2771 [This option is specific to the i386 PE targeted port of the linker]
2773 @kindex --enable-long-section-names
2774 @kindex --disable-long-section-names
2775 @item --enable-long-section-names
2776 @itemx --disable-long-section-names
2777 The PE variants of the COFF object format add an extension that permits
2778 the use of section names longer than eight characters, the normal limit
2779 for COFF. By default, these names are only allowed in object files, as
2780 fully-linked executable images do not carry the COFF string table required
2781 to support the longer names. As a GNU extension, it is possible to
2782 allow their use in executable images as well, or to (probably pointlessly!)
2783 disallow it in object files, by using these two options. Executable images
2784 generated with these long section names are slightly non-standard, carrying
2785 as they do a string table, and may generate confusing output when examined
2786 with non-GNU PE-aware tools, such as file viewers and dumpers. However,
2787 GDB relies on the use of PE long section names to find Dwarf-2 debug
2788 information sections in an executable image at runtime, and so if neither
2789 option is specified on the command-line, @command{ld} will enable long
2790 section names, overriding the default and technically correct behaviour,
2791 when it finds the presence of debug information while linking an executable
2792 image and not stripping symbols.
2793 [This option is valid for all PE targeted ports of the linker]
2795 @kindex --enable-stdcall-fixup
2796 @kindex --disable-stdcall-fixup
2797 @item --enable-stdcall-fixup
2798 @itemx --disable-stdcall-fixup
2799 If the link finds a symbol that it cannot resolve, it will attempt to
2800 do ``fuzzy linking'' by looking for another defined symbol that differs
2801 only in the format of the symbol name (cdecl vs stdcall) and will
2802 resolve that symbol by linking to the match. For example, the
2803 undefined symbol @code{_foo} might be linked to the function
2804 @code{_foo@@12}, or the undefined symbol @code{_bar@@16} might be linked
2805 to the function @code{_bar}. When the linker does this, it prints a
2806 warning, since it normally should have failed to link, but sometimes
2807 import libraries generated from third-party dlls may need this feature
2808 to be usable. If you specify @option{--enable-stdcall-fixup}, this
2809 feature is fully enabled and warnings are not printed. If you specify
2810 @option{--disable-stdcall-fixup}, this feature is disabled and such
2811 mismatches are considered to be errors.
2812 [This option is specific to the i386 PE targeted port of the linker]
2814 @kindex --leading-underscore
2815 @kindex --no-leading-underscore
2816 @item --leading-underscore
2817 @itemx --no-leading-underscore
2818 For most targets default symbol-prefix is an underscore and is defined
2819 in target's description. By this option it is possible to
2820 disable/enable the default underscore symbol-prefix.
2822 @cindex DLLs, creating
2823 @kindex --export-all-symbols
2824 @item --export-all-symbols
2825 If given, all global symbols in the objects used to build a DLL will
2826 be exported by the DLL. Note that this is the default if there
2827 otherwise wouldn't be any exported symbols. When symbols are
2828 explicitly exported via DEF files or implicitly exported via function
2829 attributes, the default is to not export anything else unless this
2830 option is given. Note that the symbols @code{DllMain@@12},
2831 @code{DllEntryPoint@@0}, @code{DllMainCRTStartup@@12}, and
2832 @code{impure_ptr} will not be automatically
2833 exported. Also, symbols imported from other DLLs will not be
2834 re-exported, nor will symbols specifying the DLL's internal layout
2835 such as those beginning with @code{_head_} or ending with
2836 @code{_iname}. In addition, no symbols from @code{libgcc},
2837 @code{libstd++}, @code{libmingw32}, or @code{crtX.o} will be exported.
2838 Symbols whose names begin with @code{__rtti_} or @code{__builtin_} will
2839 not be exported, to help with C++ DLLs. Finally, there is an
2840 extensive list of cygwin-private symbols that are not exported
2841 (obviously, this applies on when building DLLs for cygwin targets).
2842 These cygwin-excludes are: @code{_cygwin_dll_entry@@12},
2843 @code{_cygwin_crt0_common@@8}, @code{_cygwin_noncygwin_dll_entry@@12},
2844 @code{_fmode}, @code{_impure_ptr}, @code{cygwin_attach_dll},
2845 @code{cygwin_premain0}, @code{cygwin_premain1}, @code{cygwin_premain2},
2846 @code{cygwin_premain3}, and @code{environ}.
2847 [This option is specific to the i386 PE targeted port of the linker]
2849 @kindex --exclude-symbols
2850 @item --exclude-symbols @var{symbol},@var{symbol},...
2851 Specifies a list of symbols which should not be automatically
2852 exported. The symbol names may be delimited by commas or colons.
2853 [This option is specific to the i386 PE targeted port of the linker]
2855 @kindex --exclude-all-symbols
2856 @item --exclude-all-symbols
2857 Specifies no symbols should be automatically exported.
2858 [This option is specific to the i386 PE targeted port of the linker]
2860 @kindex --file-alignment
2861 @item --file-alignment
2862 Specify the file alignment. Sections in the file will always begin at
2863 file offsets which are multiples of this number. This defaults to
2865 [This option is specific to the i386 PE targeted port of the linker]
2869 @item --heap @var{reserve}
2870 @itemx --heap @var{reserve},@var{commit}
2871 Specify the number of bytes of memory to reserve (and optionally commit)
2872 to be used as heap for this program. The default is 1MB reserved, 4K
2874 [This option is specific to the i386 PE targeted port of the linker]
2877 @kindex --image-base
2878 @item --image-base @var{value}
2879 Use @var{value} as the base address of your program or dll. This is
2880 the lowest memory location that will be used when your program or dll
2881 is loaded. To reduce the need to relocate and improve performance of
2882 your dlls, each should have a unique base address and not overlap any
2883 other dlls. The default is 0x400000 for executables, and 0x10000000
2885 [This option is specific to the i386 PE targeted port of the linker]
2889 If given, the stdcall suffixes (@@@var{nn}) will be stripped from
2890 symbols before they are exported.
2891 [This option is specific to the i386 PE targeted port of the linker]
2893 @kindex --large-address-aware
2894 @item --large-address-aware
2895 If given, the appropriate bit in the ``Characteristics'' field of the COFF
2896 header is set to indicate that this executable supports virtual addresses
2897 greater than 2 gigabytes. This should be used in conjunction with the /3GB
2898 or /USERVA=@var{value} megabytes switch in the ``[operating systems]''
2899 section of the BOOT.INI. Otherwise, this bit has no effect.
2900 [This option is specific to PE targeted ports of the linker]
2902 @kindex --disable-large-address-aware
2903 @item --disable-large-address-aware
2904 Reverts the effect of a previous @samp{--large-address-aware} option.
2905 This is useful if @samp{--large-address-aware} is always set by the compiler
2906 driver (e.g. Cygwin gcc) and the executable does not support virtual
2907 addresses greater than 2 gigabytes.
2908 [This option is specific to PE targeted ports of the linker]
2910 @kindex --major-image-version
2911 @item --major-image-version @var{value}
2912 Sets the major number of the ``image version''. Defaults to 1.
2913 [This option is specific to the i386 PE targeted port of the linker]
2915 @kindex --major-os-version
2916 @item --major-os-version @var{value}
2917 Sets the major number of the ``os version''. Defaults to 4.
2918 [This option is specific to the i386 PE targeted port of the linker]
2920 @kindex --major-subsystem-version
2921 @item --major-subsystem-version @var{value}
2922 Sets the major number of the ``subsystem version''. Defaults to 4.
2923 [This option is specific to the i386 PE targeted port of the linker]
2925 @kindex --minor-image-version
2926 @item --minor-image-version @var{value}
2927 Sets the minor number of the ``image version''. Defaults to 0.
2928 [This option is specific to the i386 PE targeted port of the linker]
2930 @kindex --minor-os-version
2931 @item --minor-os-version @var{value}
2932 Sets the minor number of the ``os version''. Defaults to 0.
2933 [This option is specific to the i386 PE targeted port of the linker]
2935 @kindex --minor-subsystem-version
2936 @item --minor-subsystem-version @var{value}
2937 Sets the minor number of the ``subsystem version''. Defaults to 0.
2938 [This option is specific to the i386 PE targeted port of the linker]
2940 @cindex DEF files, creating
2941 @cindex DLLs, creating
2942 @kindex --output-def
2943 @item --output-def @var{file}
2944 The linker will create the file @var{file} which will contain a DEF
2945 file corresponding to the DLL the linker is generating. This DEF file
2946 (which should be called @code{*.def}) may be used to create an import
2947 library with @code{dlltool} or may be used as a reference to
2948 automatically or implicitly exported symbols.
2949 [This option is specific to the i386 PE targeted port of the linker]
2951 @cindex DLLs, creating
2952 @kindex --enable-auto-image-base
2953 @item --enable-auto-image-base
2954 @itemx --enable-auto-image-base=@var{value}
2955 Automatically choose the image base for DLLs, optionally starting with base
2956 @var{value}, unless one is specified using the @code{--image-base} argument.
2957 By using a hash generated from the dllname to create unique image bases
2958 for each DLL, in-memory collisions and relocations which can delay program
2959 execution are avoided.
2960 [This option is specific to the i386 PE targeted port of the linker]
2962 @kindex --disable-auto-image-base
2963 @item --disable-auto-image-base
2964 Do not automatically generate a unique image base. If there is no
2965 user-specified image base (@code{--image-base}) then use the platform
2967 [This option is specific to the i386 PE targeted port of the linker]
2969 @cindex DLLs, linking to
2970 @kindex --dll-search-prefix
2971 @item --dll-search-prefix @var{string}
2972 When linking dynamically to a dll without an import library,
2973 search for @code{<string><basename>.dll} in preference to
2974 @code{lib<basename>.dll}. This behaviour allows easy distinction
2975 between DLLs built for the various "subplatforms": native, cygwin,
2976 uwin, pw, etc. For instance, cygwin DLLs typically use
2977 @code{--dll-search-prefix=cyg}.
2978 [This option is specific to the i386 PE targeted port of the linker]
2980 @kindex --enable-auto-import
2981 @item --enable-auto-import
2982 Do sophisticated linking of @code{_symbol} to @code{__imp__symbol} for
2983 DATA imports from DLLs, thus making it possible to bypass the dllimport
2984 mechanism on the user side and to reference unmangled symbol names.
2985 [This option is specific to the i386 PE targeted port of the linker]
2987 The following remarks pertain to the original implementation of the
2988 feature and are obsolete nowadays for Cygwin and MinGW targets.
2990 Note: Use of the 'auto-import' extension will cause the text section
2991 of the image file to be made writable. This does not conform to the
2992 PE-COFF format specification published by Microsoft.
2994 Note - use of the 'auto-import' extension will also cause read only
2995 data which would normally be placed into the .rdata section to be
2996 placed into the .data section instead. This is in order to work
2997 around a problem with consts that is described here:
2998 http://www.cygwin.com/ml/cygwin/2004-09/msg01101.html
3000 Using 'auto-import' generally will 'just work' -- but sometimes you may
3003 "variable '<var>' can't be auto-imported. Please read the
3004 documentation for ld's @code{--enable-auto-import} for details."
3006 This message occurs when some (sub)expression accesses an address
3007 ultimately given by the sum of two constants (Win32 import tables only
3008 allow one). Instances where this may occur include accesses to member
3009 fields of struct variables imported from a DLL, as well as using a
3010 constant index into an array variable imported from a DLL. Any
3011 multiword variable (arrays, structs, long long, etc) may trigger
3012 this error condition. However, regardless of the exact data type
3013 of the offending exported variable, ld will always detect it, issue
3014 the warning, and exit.
3016 There are several ways to address this difficulty, regardless of the
3017 data type of the exported variable:
3019 One way is to use --enable-runtime-pseudo-reloc switch. This leaves the task
3020 of adjusting references in your client code for runtime environment, so
3021 this method works only when runtime environment supports this feature.
3023 A second solution is to force one of the 'constants' to be a variable --
3024 that is, unknown and un-optimizable at compile time. For arrays,
3025 there are two possibilities: a) make the indexee (the array's address)
3026 a variable, or b) make the 'constant' index a variable. Thus:
3029 extern type extern_array[];
3031 @{ volatile type *t=extern_array; t[1] @}
3037 extern type extern_array[];
3039 @{ volatile int t=1; extern_array[t] @}
3042 For structs (and most other multiword data types) the only option
3043 is to make the struct itself (or the long long, or the ...) variable:
3046 extern struct s extern_struct;
3047 extern_struct.field -->
3048 @{ volatile struct s *t=&extern_struct; t->field @}
3054 extern long long extern_ll;
3056 @{ volatile long long * local_ll=&extern_ll; *local_ll @}
3059 A third method of dealing with this difficulty is to abandon
3060 'auto-import' for the offending symbol and mark it with
3061 @code{__declspec(dllimport)}. However, in practice that
3062 requires using compile-time #defines to indicate whether you are
3063 building a DLL, building client code that will link to the DLL, or
3064 merely building/linking to a static library. In making the choice
3065 between the various methods of resolving the 'direct address with
3066 constant offset' problem, you should consider typical real-world usage:
3074 void main(int argc, char **argv)@{
3075 printf("%d\n",arr[1]);
3085 void main(int argc, char **argv)@{
3086 /* This workaround is for win32 and cygwin; do not "optimize" */
3087 volatile int *parr = arr;
3088 printf("%d\n",parr[1]);
3095 /* Note: auto-export is assumed (no __declspec(dllexport)) */
3096 #if (defined(_WIN32) || defined(__CYGWIN__)) && \
3097 !(defined(FOO_BUILD_DLL) || defined(FOO_STATIC))
3098 #define FOO_IMPORT __declspec(dllimport)
3102 extern FOO_IMPORT int arr[];
3105 void main(int argc, char **argv)@{
3106 printf("%d\n",arr[1]);
3110 A fourth way to avoid this problem is to re-code your
3111 library to use a functional interface rather than a data interface
3112 for the offending variables (e.g. set_foo() and get_foo() accessor
3115 @kindex --disable-auto-import
3116 @item --disable-auto-import
3117 Do not attempt to do sophisticated linking of @code{_symbol} to
3118 @code{__imp__symbol} for DATA imports from DLLs.
3119 [This option is specific to the i386 PE targeted port of the linker]
3121 @kindex --enable-runtime-pseudo-reloc
3122 @item --enable-runtime-pseudo-reloc
3123 If your code contains expressions described in --enable-auto-import section,
3124 that is, DATA imports from DLL with non-zero offset, this switch will create
3125 a vector of 'runtime pseudo relocations' which can be used by runtime
3126 environment to adjust references to such data in your client code.
3127 [This option is specific to the i386 PE targeted port of the linker]
3129 @kindex --disable-runtime-pseudo-reloc
3130 @item --disable-runtime-pseudo-reloc
3131 Do not create pseudo relocations for non-zero offset DATA imports from DLLs.
3132 [This option is specific to the i386 PE targeted port of the linker]
3134 @kindex --enable-extra-pe-debug
3135 @item --enable-extra-pe-debug
3136 Show additional debug info related to auto-import symbol thunking.
3137 [This option is specific to the i386 PE targeted port of the linker]
3139 @kindex --section-alignment
3140 @item --section-alignment
3141 Sets the section alignment. Sections in memory will always begin at
3142 addresses which are a multiple of this number. Defaults to 0x1000.
3143 [This option is specific to the i386 PE targeted port of the linker]
3147 @item --stack @var{reserve}
3148 @itemx --stack @var{reserve},@var{commit}
3149 Specify the number of bytes of memory to reserve (and optionally commit)
3150 to be used as stack for this program. The default is 2MB reserved, 4K
3152 [This option is specific to the i386 PE targeted port of the linker]
3155 @item --subsystem @var{which}
3156 @itemx --subsystem @var{which}:@var{major}
3157 @itemx --subsystem @var{which}:@var{major}.@var{minor}
3158 Specifies the subsystem under which your program will execute. The
3159 legal values for @var{which} are @code{native}, @code{windows},
3160 @code{console}, @code{posix}, and @code{xbox}. You may optionally set
3161 the subsystem version also. Numeric values are also accepted for
3163 [This option is specific to the i386 PE targeted port of the linker]
3165 The following options set flags in the @code{DllCharacteristics} field
3166 of the PE file header:
3167 [These options are specific to PE targeted ports of the linker]
3169 @kindex --high-entropy-va
3170 @item --high-entropy-va
3171 @itemx --disable-high-entropy-va
3172 Image is compatible with 64-bit address space layout randomization
3173 (ASLR). This option is enabled by default for 64-bit PE images.
3175 This option also implies @option{--dynamicbase} and
3176 @option{--enable-reloc-section}.
3178 @kindex --dynamicbase
3180 @itemx --disable-dynamicbase
3181 The image base address may be relocated using address space layout
3182 randomization (ASLR). This feature was introduced with MS Windows
3183 Vista for i386 PE targets. This option is enabled by default but
3184 can be disabled via the @option{--disable-dynamicbase} option.
3185 This option also implies @option{--enable-reloc-section}.
3187 @kindex --forceinteg
3189 @itemx --disable-forceinteg
3190 Code integrity checks are enforced. This option is disabled by
3195 @item --disable-nxcompat
3196 The image is compatible with the Data Execution Prevention.
3197 This feature was introduced with MS Windows XP SP2 for i386 PE
3198 targets. The option is enabled by default.
3200 @kindex --no-isolation
3201 @item --no-isolation
3202 @itemx --disable-no-isolation
3203 Although the image understands isolation, do not isolate the image.
3204 This option is disabled by default.
3208 @itemx --disable-no-seh
3209 The image does not use SEH. No SE handler may be called from
3210 this image. This option is disabled by default.
3214 @itemx --disable-no-bind
3215 Do not bind this image. This option is disabled by default.
3219 @itemx --disable-wdmdriver
3220 The driver uses the MS Windows Driver Model. This option is disabled
3225 @itemx --disable-tsaware
3226 The image is Terminal Server aware. This option is disabled by
3229 @kindex --insert-timestamp
3230 @item --insert-timestamp
3231 @itemx --no-insert-timestamp
3232 Insert a real timestamp into the image. This is the default behaviour
3233 as it matches legacy code and it means that the image will work with
3234 other, proprietary tools. The problem with this default is that it
3235 will result in slightly different images being produced each time the
3236 same sources are linked. The option @option{--no-insert-timestamp}
3237 can be used to insert a zero value for the timestamp, this ensuring
3238 that binaries produced from identical sources will compare
3241 @kindex --enable-reloc-section
3242 @item --enable-reloc-section
3243 @itemx --disable-reloc-section
3244 Create the base relocation table, which is necessary if the image
3245 is loaded at a different image base than specified in the PE header.
3246 This option is enabled by default.
3252 @subsection Options specific to C6X uClinux targets
3254 @c man begin OPTIONS
3256 The C6X uClinux target uses a binary format called DSBT to support shared
3257 libraries. Each shared library in the system needs to have a unique index;
3258 all executables use an index of 0.
3263 @item --dsbt-size @var{size}
3264 This option sets the number of entries in the DSBT of the current executable
3265 or shared library to @var{size}. The default is to create a table with 64
3268 @kindex --dsbt-index
3269 @item --dsbt-index @var{index}
3270 This option sets the DSBT index of the current executable or shared library
3271 to @var{index}. The default is 0, which is appropriate for generating
3272 executables. If a shared library is generated with a DSBT index of 0, the
3273 @code{R_C6000_DSBT_INDEX} relocs are copied into the output file.
3275 @kindex --no-merge-exidx-entries
3276 The @samp{--no-merge-exidx-entries} switch disables the merging of adjacent
3277 exidx entries in frame unwind info.
3285 @subsection Options specific to C-SKY targets
3287 @c man begin OPTIONS
3291 @kindex --branch-stub on C-SKY
3293 This option enables linker branch relaxation by inserting branch stub
3294 sections when needed to extend the range of branches. This option is
3295 usually not required since C-SKY supports branch and call instructions that
3296 can access the full memory range and branch relaxation is normally handled by
3297 the compiler or assembler.
3299 @kindex --stub-group-size on C-SKY
3300 @item --stub-group-size=@var{N}
3301 This option allows finer control of linker branch stub creation.
3302 It sets the maximum size of a group of input sections that can
3303 be handled by one stub section. A negative value of @var{N} locates
3304 stub sections after their branches, while a positive value allows stub
3305 sections to appear either before or after the branches. Values of
3306 @samp{1} or @samp{-1} indicate that the
3307 linker should choose suitable defaults.
3315 @subsection Options specific to Motorola 68HC11 and 68HC12 targets
3317 @c man begin OPTIONS
3319 The 68HC11 and 68HC12 linkers support specific options to control the
3320 memory bank switching mapping and trampoline code generation.
3324 @kindex --no-trampoline
3325 @item --no-trampoline
3326 This option disables the generation of trampoline. By default a trampoline
3327 is generated for each far function which is called using a @code{jsr}
3328 instruction (this happens when a pointer to a far function is taken).
3330 @kindex --bank-window
3331 @item --bank-window @var{name}
3332 This option indicates to the linker the name of the memory region in
3333 the @samp{MEMORY} specification that describes the memory bank window.
3334 The definition of such region is then used by the linker to compute
3335 paging and addresses within the memory window.
3343 @subsection Options specific to Motorola 68K target
3345 @c man begin OPTIONS
3347 The following options are supported to control handling of GOT generation
3348 when linking for 68K targets.
3353 @item --got=@var{type}
3354 This option tells the linker which GOT generation scheme to use.
3355 @var{type} should be one of @samp{single}, @samp{negative},
3356 @samp{multigot} or @samp{target}. For more information refer to the
3357 Info entry for @file{ld}.
3365 @subsection Options specific to MIPS targets
3367 @c man begin OPTIONS
3369 The following options are supported to control microMIPS instruction
3370 generation and branch relocation checks for ISA mode transitions when
3371 linking for MIPS targets.
3379 These options control the choice of microMIPS instructions used in code
3380 generated by the linker, such as that in the PLT or lazy binding stubs,
3381 or in relaxation. If @samp{--insn32} is used, then the linker only uses
3382 32-bit instruction encodings. By default or if @samp{--no-insn32} is
3383 used, all instruction encodings are used, including 16-bit ones where
3386 @kindex --ignore-branch-isa
3387 @item --ignore-branch-isa
3388 @kindex --no-ignore-branch-isa
3389 @itemx --no-ignore-branch-isa
3390 These options control branch relocation checks for invalid ISA mode
3391 transitions. If @samp{--ignore-branch-isa} is used, then the linker
3392 accepts any branch relocations and any ISA mode transition required
3393 is lost in relocation calculation, except for some cases of @code{BAL}
3394 instructions which meet relaxation conditions and are converted to
3395 equivalent @code{JALX} instructions as the associated relocation is
3396 calculated. By default or if @samp{--no-ignore-branch-isa} is used
3397 a check is made causing the loss of an ISA mode transition to produce
3400 @kindex --compact-branches
3401 @item --compact-branches
3402 @kindex --no-compact-branches
3403 @itemx --no-compact-branches
3404 These options control the generation of compact instructions by the linker
3405 in the PLT entries for MIPS R6.
3414 @subsection Options specific to PDP11 targets
3416 @c man begin OPTIONS
3418 For the pdp11-aout target, three variants of the output format can be
3419 produced as selected by the following options. The default variant
3420 for pdp11-aout is the @samp{--omagic} option, whereas for other
3421 targets @samp{--nmagic} is the default. The @samp{--imagic} option is
3422 defined only for the pdp11-aout target, while the others are described
3423 here as they apply to the pdp11-aout target.
3432 Mark the output as @code{OMAGIC} (0407) in the @file{a.out} header to
3433 indicate that the text segment is not to be write-protected and
3434 shared. Since the text and data sections are both readable and
3435 writable, the data section is allocated immediately contiguous after
3436 the text segment. This is the oldest format for PDP11 executable
3437 programs and is the default for @command{ld} on PDP11 Unix systems
3438 from the beginning through 2.11BSD.
3445 Mark the output as @code{NMAGIC} (0410) in the @file{a.out} header to
3446 indicate that when the output file is executed, the text portion will
3447 be read-only and shareable among all processes executing the same
3448 file. This involves moving the data areas up to the first possible 8K
3449 byte page boundary following the end of the text. This option creates
3450 a @emph{pure executable} format.
3457 Mark the output as @code{IMAGIC} (0411) in the @file{a.out} header to
3458 indicate that when the output file is executed, the program text and
3459 data areas will be loaded into separate address spaces using the split
3460 instruction and data space feature of the memory management unit in
3461 larger models of the PDP11. This doubles the address space available
3462 to the program. The text segment is again pure, write-protected, and
3463 shareable. The only difference in the output format between this
3464 option and the others, besides the magic number, is that both the text
3465 and data sections start at location 0. The @samp{-z} option selected
3466 this format in 2.11BSD. This option creates a @emph{separate
3472 Equivalent to @samp{--nmagic} for pdp11-aout.
3481 @section Environment Variables
3483 @c man begin ENVIRONMENT
3485 You can change the behaviour of @command{ld} with the environment variables
3486 @ifclear SingleFormat
3489 @code{LDEMULATION} and @code{COLLECT_NO_DEMANGLE}.
3491 @ifclear SingleFormat
3493 @cindex default input format
3494 @code{GNUTARGET} determines the input-file object format if you don't
3495 use @samp{-b} (or its synonym @samp{--format}). Its value should be one
3496 of the BFD names for an input format (@pxref{BFD}). If there is no
3497 @code{GNUTARGET} in the environment, @command{ld} uses the natural format
3498 of the target. If @code{GNUTARGET} is set to @code{default} then BFD
3499 attempts to discover the input format by examining binary input files;
3500 this method often succeeds, but there are potential ambiguities, since
3501 there is no method of ensuring that the magic number used to specify
3502 object-file formats is unique. However, the configuration procedure for
3503 BFD on each system places the conventional format for that system first
3504 in the search-list, so ambiguities are resolved in favor of convention.
3508 @cindex default emulation
3509 @cindex emulation, default
3510 @code{LDEMULATION} determines the default emulation if you don't use the
3511 @samp{-m} option. The emulation can affect various aspects of linker
3512 behaviour, particularly the default linker script. You can list the
3513 available emulations with the @samp{--verbose} or @samp{-V} options. If
3514 the @samp{-m} option is not used, and the @code{LDEMULATION} environment
3515 variable is not defined, the default emulation depends upon how the
3516 linker was configured.
3518 @kindex COLLECT_NO_DEMANGLE
3519 @cindex demangling, default
3520 Normally, the linker will default to demangling symbols. However, if
3521 @code{COLLECT_NO_DEMANGLE} is set in the environment, then it will
3522 default to not demangling symbols. This environment variable is used in
3523 a similar fashion by the @code{gcc} linker wrapper program. The default
3524 may be overridden by the @samp{--demangle} and @samp{--no-demangle}
3531 @chapter Linker Scripts
3534 @cindex linker scripts
3535 @cindex command files
3536 Every link is controlled by a @dfn{linker script}. This script is
3537 written in the linker command language.
3539 The main purpose of the linker script is to describe how the sections in
3540 the input files should be mapped into the output file, and to control
3541 the memory layout of the output file. Most linker scripts do nothing
3542 more than this. However, when necessary, the linker script can also
3543 direct the linker to perform many other operations, using the commands
3546 The linker always uses a linker script. If you do not supply one
3547 yourself, the linker will use a default script that is compiled into the
3548 linker executable. You can use the @samp{--verbose} command-line option
3549 to display the default linker script. Certain command-line options,
3550 such as @samp{-r} or @samp{-N}, will affect the default linker script.
3552 You may supply your own linker script by using the @samp{-T} command
3553 line option. When you do this, your linker script will replace the
3554 default linker script.
3556 You may also use linker scripts implicitly by naming them as input files
3557 to the linker, as though they were files to be linked. @xref{Implicit
3561 * Basic Script Concepts:: Basic Linker Script Concepts
3562 * Script Format:: Linker Script Format
3563 * Simple Example:: Simple Linker Script Example
3564 * Simple Commands:: Simple Linker Script Commands
3565 * Assignments:: Assigning Values to Symbols
3566 * SECTIONS:: SECTIONS Command
3567 * MEMORY:: MEMORY Command
3568 * PHDRS:: PHDRS Command
3569 * VERSION:: VERSION Command
3570 * Expressions:: Expressions in Linker Scripts
3571 * Implicit Linker Scripts:: Implicit Linker Scripts
3574 @node Basic Script Concepts
3575 @section Basic Linker Script Concepts
3576 @cindex linker script concepts
3577 We need to define some basic concepts and vocabulary in order to
3578 describe the linker script language.
3580 The linker combines input files into a single output file. The output
3581 file and each input file are in a special data format known as an
3582 @dfn{object file format}. Each file is called an @dfn{object file}.
3583 The output file is often called an @dfn{executable}, but for our
3584 purposes we will also call it an object file. Each object file has,
3585 among other things, a list of @dfn{sections}. We sometimes refer to a
3586 section in an input file as an @dfn{input section}; similarly, a section
3587 in the output file is an @dfn{output section}.
3589 Each section in an object file has a name and a size. Most sections
3590 also have an associated block of data, known as the @dfn{section
3591 contents}. A section may be marked as @dfn{loadable}, which means that
3592 the contents should be loaded into memory when the output file is run.
3593 A section with no contents may be @dfn{allocatable}, which means that an
3594 area in memory should be set aside, but nothing in particular should be
3595 loaded there (in some cases this memory must be zeroed out). A section
3596 which is neither loadable nor allocatable typically contains some sort
3597 of debugging information.
3599 Every loadable or allocatable output section has two addresses. The
3600 first is the @dfn{VMA}, or virtual memory address. This is the address
3601 the section will have when the output file is run. The second is the
3602 @dfn{LMA}, or load memory address. This is the address at which the
3603 section will be loaded. In most cases the two addresses will be the
3604 same. An example of when they might be different is when a data section
3605 is loaded into ROM, and then copied into RAM when the program starts up
3606 (this technique is often used to initialize global variables in a ROM
3607 based system). In this case the ROM address would be the LMA, and the
3608 RAM address would be the VMA.
3610 You can see the sections in an object file by using the @code{objdump}
3611 program with the @samp{-h} option.
3613 Every object file also has a list of @dfn{symbols}, known as the
3614 @dfn{symbol table}. A symbol may be defined or undefined. Each symbol
3615 has a name, and each defined symbol has an address, among other
3616 information. If you compile a C or C++ program into an object file, you
3617 will get a defined symbol for every defined function and global or
3618 static variable. Every undefined function or global variable which is
3619 referenced in the input file will become an undefined symbol.
3621 You can see the symbols in an object file by using the @code{nm}
3622 program, or by using the @code{objdump} program with the @samp{-t}
3626 @section Linker Script Format
3627 @cindex linker script format
3628 Linker scripts are text files.
3630 You write a linker script as a series of commands. Each command is
3631 either a keyword, possibly followed by arguments, or an assignment to a
3632 symbol. You may separate commands using semicolons. Whitespace is
3635 Strings such as file or format names can normally be entered directly.
3636 If the file name contains a character such as a comma which would
3637 otherwise serve to separate file names, you may put the file name in
3638 double quotes. There is no way to use a double quote character in a
3641 You may include comments in linker scripts just as in C, delimited by
3642 @samp{/*} and @samp{*/}. As in C, comments are syntactically equivalent
3645 @node Simple Example
3646 @section Simple Linker Script Example
3647 @cindex linker script example
3648 @cindex example of linker script
3649 Many linker scripts are fairly simple.
3651 The simplest possible linker script has just one command:
3652 @samp{SECTIONS}. You use the @samp{SECTIONS} command to describe the
3653 memory layout of the output file.
3655 The @samp{SECTIONS} command is a powerful command. Here we will
3656 describe a simple use of it. Let's assume your program consists only of
3657 code, initialized data, and uninitialized data. These will be in the
3658 @samp{.text}, @samp{.data}, and @samp{.bss} sections, respectively.
3659 Let's assume further that these are the only sections which appear in
3662 For this example, let's say that the code should be loaded at address
3663 0x10000, and that the data should start at address 0x8000000. Here is a
3664 linker script which will do that:
3669 .text : @{ *(.text) @}
3671 .data : @{ *(.data) @}
3672 .bss : @{ *(.bss) @}
3676 You write the @samp{SECTIONS} command as the keyword @samp{SECTIONS},
3677 followed by a series of symbol assignments and output section
3678 descriptions enclosed in curly braces.
3680 The first line inside the @samp{SECTIONS} command of the above example
3681 sets the value of the special symbol @samp{.}, which is the location
3682 counter. If you do not specify the address of an output section in some
3683 other way (other ways are described later), the address is set from the
3684 current value of the location counter. The location counter is then
3685 incremented by the size of the output section. At the start of the
3686 @samp{SECTIONS} command, the location counter has the value @samp{0}.
3688 The second line defines an output section, @samp{.text}. The colon is
3689 required syntax which may be ignored for now. Within the curly braces
3690 after the output section name, you list the names of the input sections
3691 which should be placed into this output section. The @samp{*} is a
3692 wildcard which matches any file name. The expression @samp{*(.text)}
3693 means all @samp{.text} input sections in all input files.
3695 Since the location counter is @samp{0x10000} when the output section
3696 @samp{.text} is defined, the linker will set the address of the
3697 @samp{.text} section in the output file to be @samp{0x10000}.
3699 The remaining lines define the @samp{.data} and @samp{.bss} sections in
3700 the output file. The linker will place the @samp{.data} output section
3701 at address @samp{0x8000000}. After the linker places the @samp{.data}
3702 output section, the value of the location counter will be
3703 @samp{0x8000000} plus the size of the @samp{.data} output section. The
3704 effect is that the linker will place the @samp{.bss} output section
3705 immediately after the @samp{.data} output section in memory.
3707 The linker will ensure that each output section has the required
3708 alignment, by increasing the location counter if necessary. In this
3709 example, the specified addresses for the @samp{.text} and @samp{.data}
3710 sections will probably satisfy any alignment constraints, but the linker
3711 may have to create a small gap between the @samp{.data} and @samp{.bss}
3714 That's it! That's a simple and complete linker script.
3716 @node Simple Commands
3717 @section Simple Linker Script Commands
3718 @cindex linker script simple commands
3719 In this section we describe the simple linker script commands.
3722 * Entry Point:: Setting the entry point
3723 * File Commands:: Commands dealing with files
3724 @ifclear SingleFormat
3725 * Format Commands:: Commands dealing with object file formats
3728 * REGION_ALIAS:: Assign alias names to memory regions
3729 * Miscellaneous Commands:: Other linker script commands
3733 @subsection Setting the Entry Point
3734 @kindex ENTRY(@var{symbol})
3735 @cindex start of execution
3736 @cindex first instruction
3738 The first instruction to execute in a program is called the @dfn{entry
3739 point}. You can use the @code{ENTRY} linker script command to set the
3740 entry point. The argument is a symbol name:
3745 There are several ways to set the entry point. The linker will set the
3746 entry point by trying each of the following methods in order, and
3747 stopping when one of them succeeds:
3750 the @samp{-e} @var{entry} command-line option;
3752 the @code{ENTRY(@var{symbol})} command in a linker script;
3754 the value of a target-specific symbol, if it is defined; For many
3755 targets this is @code{start}, but PE- and BeOS-based systems for example
3756 check a list of possible entry symbols, matching the first one found.
3758 the address of the first byte of the @samp{.text} section, if present;
3760 The address @code{0}.
3764 @subsection Commands Dealing with Files
3765 @cindex linker script file commands
3766 Several linker script commands deal with files.
3769 @item INCLUDE @var{filename}
3770 @kindex INCLUDE @var{filename}
3771 @cindex including a linker script
3772 Include the linker script @var{filename} at this point. The file will
3773 be searched for in the current directory, and in any directory specified
3774 with the @option{-L} option. You can nest calls to @code{INCLUDE} up to
3777 You can place @code{INCLUDE} directives at the top level, in @code{MEMORY} or
3778 @code{SECTIONS} commands, or in output section descriptions.
3780 @item INPUT(@var{file}, @var{file}, @dots{})
3781 @itemx INPUT(@var{file} @var{file} @dots{})
3782 @kindex INPUT(@var{files})
3783 @cindex input files in linker scripts
3784 @cindex input object files in linker scripts
3785 @cindex linker script input object files
3786 The @code{INPUT} command directs the linker to include the named files
3787 in the link, as though they were named on the command line.
3789 For example, if you always want to include @file{subr.o} any time you do
3790 a link, but you can't be bothered to put it on every link command line,
3791 then you can put @samp{INPUT (subr.o)} in your linker script.
3793 In fact, if you like, you can list all of your input files in the linker
3794 script, and then invoke the linker with nothing but a @samp{-T} option.
3796 In case a @dfn{sysroot prefix} is configured, and the filename starts
3797 with the @samp{/} character, and the script being processed was
3798 located inside the @dfn{sysroot prefix}, the filename will be looked
3799 for in the @dfn{sysroot prefix}. The @dfn{sysroot prefix} can also be forced by specifying
3800 @code{=} as the first character in the filename path, or prefixing the
3801 filename path with @code{$SYSROOT}. See also the description of
3802 @samp{-L} in @ref{Options,,Command-line Options}.
3804 If a @dfn{sysroot prefix} is not used then the linker will try to open
3805 the file in the directory containing the linker script. If it is not
3806 found the linker will then search the current directory. If it is still
3807 not found the linker will search through the archive library search
3810 If you use @samp{INPUT (-l@var{file})}, @command{ld} will transform the
3811 name to @code{lib@var{file}.a}, as with the command-line argument
3814 When you use the @code{INPUT} command in an implicit linker script, the
3815 files will be included in the link at the point at which the linker
3816 script file is included. This can affect archive searching.
3818 @item GROUP(@var{file}, @var{file}, @dots{})
3819 @itemx GROUP(@var{file} @var{file} @dots{})
3820 @kindex GROUP(@var{files})
3821 @cindex grouping input files
3822 The @code{GROUP} command is like @code{INPUT}, except that the named
3823 files should all be archives, and they are searched repeatedly until no
3824 new undefined references are created. See the description of @samp{-(}
3825 in @ref{Options,,Command-line Options}.
3827 @item AS_NEEDED(@var{file}, @var{file}, @dots{})
3828 @itemx AS_NEEDED(@var{file} @var{file} @dots{})
3829 @kindex AS_NEEDED(@var{files})
3830 This construct can appear only inside of the @code{INPUT} or @code{GROUP}
3831 commands, among other filenames. The files listed will be handled
3832 as if they appear directly in the @code{INPUT} or @code{GROUP} commands,
3833 with the exception of ELF shared libraries, that will be added only
3834 when they are actually needed. This construct essentially enables
3835 @option{--as-needed} option for all the files listed inside of it
3836 and restores previous @option{--as-needed} resp. @option{--no-as-needed}
3839 @item OUTPUT(@var{filename})
3840 @kindex OUTPUT(@var{filename})
3841 @cindex output file name in linker script
3842 The @code{OUTPUT} command names the output file. Using
3843 @code{OUTPUT(@var{filename})} in the linker script is exactly like using
3844 @samp{-o @var{filename}} on the command line (@pxref{Options,,Command
3845 Line Options}). If both are used, the command-line option takes
3848 You can use the @code{OUTPUT} command to define a default name for the
3849 output file other than the usual default of @file{a.out}.
3851 @item SEARCH_DIR(@var{path})
3852 @kindex SEARCH_DIR(@var{path})
3853 @cindex library search path in linker script
3854 @cindex archive search path in linker script
3855 @cindex search path in linker script
3856 The @code{SEARCH_DIR} command adds @var{path} to the list of paths where
3857 @command{ld} looks for archive libraries. Using
3858 @code{SEARCH_DIR(@var{path})} is exactly like using @samp{-L @var{path}}
3859 on the command line (@pxref{Options,,Command-line Options}). If both
3860 are used, then the linker will search both paths. Paths specified using
3861 the command-line option are searched first.
3863 @item STARTUP(@var{filename})
3864 @kindex STARTUP(@var{filename})
3865 @cindex first input file
3866 The @code{STARTUP} command is just like the @code{INPUT} command, except
3867 that @var{filename} will become the first input file to be linked, as
3868 though it were specified first on the command line. This may be useful
3869 when using a system in which the entry point is always the start of the
3873 @ifclear SingleFormat
3874 @node Format Commands
3875 @subsection Commands Dealing with Object File Formats
3876 A couple of linker script commands deal with object file formats.
3879 @item OUTPUT_FORMAT(@var{bfdname})
3880 @itemx OUTPUT_FORMAT(@var{default}, @var{big}, @var{little})
3881 @kindex OUTPUT_FORMAT(@var{bfdname})
3882 @cindex output file format in linker script
3883 The @code{OUTPUT_FORMAT} command names the BFD format to use for the
3884 output file (@pxref{BFD}). Using @code{OUTPUT_FORMAT(@var{bfdname})} is
3885 exactly like using @samp{--oformat @var{bfdname}} on the command line
3886 (@pxref{Options,,Command-line Options}). If both are used, the command
3887 line option takes precedence.
3889 You can use @code{OUTPUT_FORMAT} with three arguments to use different
3890 formats based on the @samp{-EB} and @samp{-EL} command-line options.
3891 This permits the linker script to set the output format based on the
3894 If neither @samp{-EB} nor @samp{-EL} are used, then the output format
3895 will be the first argument, @var{default}. If @samp{-EB} is used, the
3896 output format will be the second argument, @var{big}. If @samp{-EL} is
3897 used, the output format will be the third argument, @var{little}.
3899 For example, the default linker script for the MIPS ELF target uses this
3902 OUTPUT_FORMAT(elf32-bigmips, elf32-bigmips, elf32-littlemips)
3904 This says that the default format for the output file is
3905 @samp{elf32-bigmips}, but if the user uses the @samp{-EL} command-line
3906 option, the output file will be created in the @samp{elf32-littlemips}
3909 @item TARGET(@var{bfdname})
3910 @kindex TARGET(@var{bfdname})
3911 @cindex input file format in linker script
3912 The @code{TARGET} command names the BFD format to use when reading input
3913 files. It affects subsequent @code{INPUT} and @code{GROUP} commands.
3914 This command is like using @samp{-b @var{bfdname}} on the command line
3915 (@pxref{Options,,Command-line Options}). If the @code{TARGET} command
3916 is used but @code{OUTPUT_FORMAT} is not, then the last @code{TARGET}
3917 command is also used to set the format for the output file. @xref{BFD}.
3922 @subsection Assign alias names to memory regions
3923 @kindex REGION_ALIAS(@var{alias}, @var{region})
3924 @cindex region alias
3925 @cindex region names
3927 Alias names can be added to existing memory regions created with the
3928 @ref{MEMORY} command. Each name corresponds to at most one memory region.
3931 REGION_ALIAS(@var{alias}, @var{region})
3934 The @code{REGION_ALIAS} function creates an alias name @var{alias} for the
3935 memory region @var{region}. This allows a flexible mapping of output sections
3936 to memory regions. An example follows.
3938 Suppose we have an application for embedded systems which come with various
3939 memory storage devices. All have a general purpose, volatile memory @code{RAM}
3940 that allows code execution or data storage. Some may have a read-only,
3941 non-volatile memory @code{ROM} that allows code execution and read-only data
3942 access. The last variant is a read-only, non-volatile memory @code{ROM2} with
3943 read-only data access and no code execution capability. We have four output
3948 @code{.text} program code;
3950 @code{.rodata} read-only data;
3952 @code{.data} read-write initialized data;
3954 @code{.bss} read-write zero initialized data.
3957 The goal is to provide a linker command file that contains a system independent
3958 part defining the output sections and a system dependent part mapping the
3959 output sections to the memory regions available on the system. Our embedded
3960 systems come with three different memory setups @code{A}, @code{B} and
3962 @multitable @columnfractions .25 .25 .25 .25
3963 @item Section @tab Variant A @tab Variant B @tab Variant C
3964 @item .text @tab RAM @tab ROM @tab ROM
3965 @item .rodata @tab RAM @tab ROM @tab ROM2
3966 @item .data @tab RAM @tab RAM/ROM @tab RAM/ROM2
3967 @item .bss @tab RAM @tab RAM @tab RAM
3969 The notation @code{RAM/ROM} or @code{RAM/ROM2} means that this section is
3970 loaded into region @code{ROM} or @code{ROM2} respectively. Please note that
3971 the load address of the @code{.data} section starts in all three variants at
3972 the end of the @code{.rodata} section.
3974 The base linker script that deals with the output sections follows. It
3975 includes the system dependent @code{linkcmds.memory} file that describes the
3978 INCLUDE linkcmds.memory
3991 .data : AT (rodata_end)
3996 data_size = SIZEOF(.data);
3997 data_load_start = LOADADDR(.data);
4005 Now we need three different @code{linkcmds.memory} files to define memory
4006 regions and alias names. The content of @code{linkcmds.memory} for the three
4007 variants @code{A}, @code{B} and @code{C}:
4010 Here everything goes into the @code{RAM}.
4014 RAM : ORIGIN = 0, LENGTH = 4M
4017 REGION_ALIAS("REGION_TEXT", RAM);
4018 REGION_ALIAS("REGION_RODATA", RAM);
4019 REGION_ALIAS("REGION_DATA", RAM);
4020 REGION_ALIAS("REGION_BSS", RAM);
4023 Program code and read-only data go into the @code{ROM}. Read-write data goes
4024 into the @code{RAM}. An image of the initialized data is loaded into the
4025 @code{ROM} and will be copied during system start into the @code{RAM}.
4029 ROM : ORIGIN = 0, LENGTH = 3M
4030 RAM : ORIGIN = 0x10000000, LENGTH = 1M
4033 REGION_ALIAS("REGION_TEXT", ROM);
4034 REGION_ALIAS("REGION_RODATA", ROM);
4035 REGION_ALIAS("REGION_DATA", RAM);
4036 REGION_ALIAS("REGION_BSS", RAM);
4039 Program code goes into the @code{ROM}. Read-only data goes into the
4040 @code{ROM2}. Read-write data goes into the @code{RAM}. An image of the
4041 initialized data is loaded into the @code{ROM2} and will be copied during
4042 system start into the @code{RAM}.
4046 ROM : ORIGIN = 0, LENGTH = 2M
4047 ROM2 : ORIGIN = 0x10000000, LENGTH = 1M
4048 RAM : ORIGIN = 0x20000000, LENGTH = 1M
4051 REGION_ALIAS("REGION_TEXT", ROM);
4052 REGION_ALIAS("REGION_RODATA", ROM2);
4053 REGION_ALIAS("REGION_DATA", RAM);
4054 REGION_ALIAS("REGION_BSS", RAM);
4058 It is possible to write a common system initialization routine to copy the
4059 @code{.data} section from @code{ROM} or @code{ROM2} into the @code{RAM} if
4064 extern char data_start [];
4065 extern char data_size [];
4066 extern char data_load_start [];
4068 void copy_data(void)
4070 if (data_start != data_load_start)
4072 memcpy(data_start, data_load_start, (size_t) data_size);
4077 @node Miscellaneous Commands
4078 @subsection Other Linker Script Commands
4079 There are a few other linker scripts commands.
4082 @item ASSERT(@var{exp}, @var{message})
4084 @cindex assertion in linker script
4085 Ensure that @var{exp} is non-zero. If it is zero, then exit the linker
4086 with an error code, and print @var{message}.
4088 Note that assertions are checked before the final stages of linking
4089 take place. This means that expressions involving symbols PROVIDEd
4090 inside section definitions will fail if the user has not set values
4091 for those symbols. The only exception to this rule is PROVIDEd
4092 symbols that just reference dot. Thus an assertion like this:
4097 PROVIDE (__stack = .);
4098 PROVIDE (__stack_size = 0x100);
4099 ASSERT ((__stack > (_end + __stack_size)), "Error: No room left for the stack");
4103 will fail if @code{__stack_size} is not defined elsewhere. Symbols
4104 PROVIDEd outside of section definitions are evaluated earlier, so they
4105 can be used inside ASSERTions. Thus:
4108 PROVIDE (__stack_size = 0x100);
4111 PROVIDE (__stack = .);
4112 ASSERT ((__stack > (_end + __stack_size)), "Error: No room left for the stack");
4118 @item EXTERN(@var{symbol} @var{symbol} @dots{})
4120 @cindex undefined symbol in linker script
4121 Force @var{symbol} to be entered in the output file as an undefined
4122 symbol. Doing this may, for example, trigger linking of additional
4123 modules from standard libraries. You may list several @var{symbol}s for
4124 each @code{EXTERN}, and you may use @code{EXTERN} multiple times. This
4125 command has the same effect as the @samp{-u} command-line option.
4127 @item FORCE_COMMON_ALLOCATION
4128 @kindex FORCE_COMMON_ALLOCATION
4129 @cindex common allocation in linker script
4130 This command has the same effect as the @samp{-d} command-line option:
4131 to make @command{ld} assign space to common symbols even if a relocatable
4132 output file is specified (@samp{-r}).
4134 @item INHIBIT_COMMON_ALLOCATION
4135 @kindex INHIBIT_COMMON_ALLOCATION
4136 @cindex common allocation in linker script
4137 This command has the same effect as the @samp{--no-define-common}
4138 command-line option: to make @code{ld} omit the assignment of addresses
4139 to common symbols even for a non-relocatable output file.
4141 @item FORCE_GROUP_ALLOCATION
4142 @kindex FORCE_GROUP_ALLOCATION
4143 @cindex group allocation in linker script
4144 @cindex section groups
4146 This command has the same effect as the
4147 @samp{--force-group-allocation} command-line option: to make
4148 @command{ld} place section group members like normal input sections,
4149 and to delete the section groups even if a relocatable output file is
4150 specified (@samp{-r}).
4152 @item INSERT [ AFTER | BEFORE ] @var{output_section}
4154 @cindex insert user script into default script
4155 This command is typically used in a script specified by @samp{-T} to
4156 augment the default @code{SECTIONS} with, for example, overlays. It
4157 inserts all prior linker script statements after (or before)
4158 @var{output_section}, and also causes @samp{-T} to not override the
4159 default linker script. The exact insertion point is as for orphan
4160 sections. @xref{Location Counter}. The insertion happens after the
4161 linker has mapped input sections to output sections. Prior to the
4162 insertion, since @samp{-T} scripts are parsed before the default
4163 linker script, statements in the @samp{-T} script occur before the
4164 default linker script statements in the internal linker representation
4165 of the script. In particular, input section assignments will be made
4166 to @samp{-T} output sections before those in the default script. Here
4167 is an example of how a @samp{-T} script using @code{INSERT} might look:
4174 .ov1 @{ ov1*(.text) @}
4175 .ov2 @{ ov2*(.text) @}
4181 @item NOCROSSREFS(@var{section} @var{section} @dots{})
4182 @kindex NOCROSSREFS(@var{sections})
4183 @cindex cross references
4184 This command may be used to tell @command{ld} to issue an error about any
4185 references among certain output sections.
4187 In certain types of programs, particularly on embedded systems when
4188 using overlays, when one section is loaded into memory, another section
4189 will not be. Any direct references between the two sections would be
4190 errors. For example, it would be an error if code in one section called
4191 a function defined in the other section.
4193 The @code{NOCROSSREFS} command takes a list of output section names. If
4194 @command{ld} detects any cross references between the sections, it reports
4195 an error and returns a non-zero exit status. Note that the
4196 @code{NOCROSSREFS} command uses output section names, not input section
4199 @item NOCROSSREFS_TO(@var{tosection} @var{fromsection} @dots{})
4200 @kindex NOCROSSREFS_TO(@var{tosection} @var{fromsections})
4201 @cindex cross references
4202 This command may be used to tell @command{ld} to issue an error about any
4203 references to one section from a list of other sections.
4205 The @code{NOCROSSREFS} command is useful when ensuring that two or more
4206 output sections are entirely independent but there are situations where
4207 a one-way dependency is needed. For example, in a multi-core application
4208 there may be shared code that can be called from each core but for safety
4209 must never call back.
4211 The @code{NOCROSSREFS_TO} command takes a list of output section names.
4212 The first section can not be referenced from any of the other sections.
4213 If @command{ld} detects any references to the first section from any of
4214 the other sections, it reports an error and returns a non-zero exit
4215 status. Note that the @code{NOCROSSREFS_TO} command uses output section
4216 names, not input section names.
4218 @ifclear SingleFormat
4219 @item OUTPUT_ARCH(@var{bfdarch})
4220 @kindex OUTPUT_ARCH(@var{bfdarch})
4221 @cindex machine architecture
4222 @cindex architecture
4223 Specify a particular output machine architecture. The argument is one
4224 of the names used by the BFD library (@pxref{BFD}). You can see the
4225 architecture of an object file by using the @code{objdump} program with
4226 the @samp{-f} option.
4229 @item LD_FEATURE(@var{string})
4230 @kindex LD_FEATURE(@var{string})
4231 This command may be used to modify @command{ld} behavior. If
4232 @var{string} is @code{"SANE_EXPR"} then absolute symbols and numbers
4233 in a script are simply treated as numbers everywhere.
4234 @xref{Expression Section}.
4238 @section Assigning Values to Symbols
4239 @cindex assignment in scripts
4240 @cindex symbol definition, scripts
4241 @cindex variables, defining
4242 You may assign a value to a symbol in a linker script. This will define
4243 the symbol and place it into the symbol table with a global scope.
4246 * Simple Assignments:: Simple Assignments
4249 * PROVIDE_HIDDEN:: PROVIDE_HIDDEN
4250 * Source Code Reference:: How to use a linker script defined symbol in source code
4253 @node Simple Assignments
4254 @subsection Simple Assignments
4256 You may assign to a symbol using any of the C assignment operators:
4259 @item @var{symbol} = @var{expression} ;
4260 @itemx @var{symbol} += @var{expression} ;
4261 @itemx @var{symbol} -= @var{expression} ;
4262 @itemx @var{symbol} *= @var{expression} ;
4263 @itemx @var{symbol} /= @var{expression} ;
4264 @itemx @var{symbol} <<= @var{expression} ;
4265 @itemx @var{symbol} >>= @var{expression} ;
4266 @itemx @var{symbol} &= @var{expression} ;
4267 @itemx @var{symbol} |= @var{expression} ;
4270 The first case will define @var{symbol} to the value of
4271 @var{expression}. In the other cases, @var{symbol} must already be
4272 defined, and the value will be adjusted accordingly.
4274 The special symbol name @samp{.} indicates the location counter. You
4275 may only use this within a @code{SECTIONS} command. @xref{Location Counter}.
4277 The semicolon after @var{expression} is required.
4279 Expressions are defined below; see @ref{Expressions}.
4281 You may write symbol assignments as commands in their own right, or as
4282 statements within a @code{SECTIONS} command, or as part of an output
4283 section description in a @code{SECTIONS} command.
4285 The section of the symbol will be set from the section of the
4286 expression; for more information, see @ref{Expression Section}.
4288 Here is an example showing the three different places that symbol
4289 assignments may be used:
4300 _bdata = (. + 3) & ~ 3;
4301 .data : @{ *(.data) @}
4305 In this example, the symbol @samp{floating_point} will be defined as
4306 zero. The symbol @samp{_etext} will be defined as the address following
4307 the last @samp{.text} input section. The symbol @samp{_bdata} will be
4308 defined as the address following the @samp{.text} output section aligned
4309 upward to a 4 byte boundary.
4314 For ELF targeted ports, define a symbol that will be hidden and won't be
4315 exported. The syntax is @code{HIDDEN(@var{symbol} = @var{expression})}.
4317 Here is the example from @ref{Simple Assignments}, rewritten to use
4321 HIDDEN(floating_point = 0);
4329 HIDDEN(_bdata = (. + 3) & ~ 3);
4330 .data : @{ *(.data) @}
4334 In this case none of the three symbols will be visible outside this module.
4339 In some cases, it is desirable for a linker script to define a symbol
4340 only if it is referenced and is not defined by any object included in
4341 the link. For example, traditional linkers defined the symbol
4342 @samp{etext}. However, ANSI C requires that the user be able to use
4343 @samp{etext} as a function name without encountering an error. The
4344 @code{PROVIDE} keyword may be used to define a symbol, such as
4345 @samp{etext}, only if it is referenced but not defined. The syntax is
4346 @code{PROVIDE(@var{symbol} = @var{expression})}.
4348 Here is an example of using @code{PROVIDE} to define @samp{etext}:
4361 In this example, if the program defines @samp{_etext} (with a leading
4362 underscore), the linker will give a multiple definition error. If, on
4363 the other hand, the program defines @samp{etext} (with no leading
4364 underscore), the linker will silently use the definition in the program.
4365 If the program references @samp{etext} but does not define it, the
4366 linker will use the definition in the linker script.
4368 Note - the @code{PROVIDE} directive considers a common symbol to be
4369 defined, even though such a symbol could be combined with the symbol
4370 that the @code{PROVIDE} would create. This is particularly important
4371 when considering constructor and destructor list symbols such as
4372 @samp{__CTOR_LIST__} as these are often defined as common symbols.
4374 @node PROVIDE_HIDDEN
4375 @subsection PROVIDE_HIDDEN
4376 @cindex PROVIDE_HIDDEN
4377 Similar to @code{PROVIDE}. For ELF targeted ports, the symbol will be
4378 hidden and won't be exported.
4380 @node Source Code Reference
4381 @subsection Source Code Reference
4383 Accessing a linker script defined variable from source code is not
4384 intuitive. In particular a linker script symbol is not equivalent to
4385 a variable declaration in a high level language, it is instead a
4386 symbol that does not have a value.
4388 Before going further, it is important to note that compilers often
4389 transform names in the source code into different names when they are
4390 stored in the symbol table. For example, Fortran compilers commonly
4391 prepend or append an underscore, and C++ performs extensive @samp{name
4392 mangling}. Therefore there might be a discrepancy between the name
4393 of a variable as it is used in source code and the name of the same
4394 variable as it is defined in a linker script. For example in C a
4395 linker script variable might be referred to as:
4401 But in the linker script it might be defined as:
4407 In the remaining examples however it is assumed that no name
4408 transformation has taken place.
4410 When a symbol is declared in a high level language such as C, two
4411 things happen. The first is that the compiler reserves enough space
4412 in the program's memory to hold the @emph{value} of the symbol. The
4413 second is that the compiler creates an entry in the program's symbol
4414 table which holds the symbol's @emph{address}. ie the symbol table
4415 contains the address of the block of memory holding the symbol's
4416 value. So for example the following C declaration, at file scope:
4422 creates an entry called @samp{foo} in the symbol table. This entry
4423 holds the address of an @samp{int} sized block of memory where the
4424 number 1000 is initially stored.
4426 When a program references a symbol the compiler generates code that
4427 first accesses the symbol table to find the address of the symbol's
4428 memory block and then code to read the value from that memory block.
4435 looks up the symbol @samp{foo} in the symbol table, gets the address
4436 associated with this symbol and then writes the value 1 into that
4443 looks up the symbol @samp{foo} in the symbol table, gets its address
4444 and then copies this address into the block of memory associated with
4445 the variable @samp{a}.
4447 Linker scripts symbol declarations, by contrast, create an entry in
4448 the symbol table but do not assign any memory to them. Thus they are
4449 an address without a value. So for example the linker script definition:
4455 creates an entry in the symbol table called @samp{foo} which holds
4456 the address of memory location 1000, but nothing special is stored at
4457 address 1000. This means that you cannot access the @emph{value} of a
4458 linker script defined symbol - it has no value - all you can do is
4459 access the @emph{address} of a linker script defined symbol.
4461 Hence when you are using a linker script defined symbol in source code
4462 you should always take the address of the symbol, and never attempt to
4463 use its value. For example suppose you want to copy the contents of a
4464 section of memory called .ROM into a section called .FLASH and the
4465 linker script contains these declarations:
4469 start_of_ROM = .ROM;
4470 end_of_ROM = .ROM + sizeof (.ROM);
4471 start_of_FLASH = .FLASH;
4475 Then the C source code to perform the copy would be:
4479 extern char start_of_ROM, end_of_ROM, start_of_FLASH;
4481 memcpy (& start_of_FLASH, & start_of_ROM, & end_of_ROM - & start_of_ROM);
4485 Note the use of the @samp{&} operators. These are correct.
4486 Alternatively the symbols can be treated as the names of vectors or
4487 arrays and then the code will again work as expected:
4491 extern char start_of_ROM[], end_of_ROM[], start_of_FLASH[];
4493 memcpy (start_of_FLASH, start_of_ROM, end_of_ROM - start_of_ROM);
4497 Note how using this method does not require the use of @samp{&}
4501 @section SECTIONS Command
4503 The @code{SECTIONS} command tells the linker how to map input sections
4504 into output sections, and how to place the output sections in memory.
4506 The format of the @code{SECTIONS} command is:
4510 @var{sections-command}
4511 @var{sections-command}
4516 Each @var{sections-command} may of be one of the following:
4520 an @code{ENTRY} command (@pxref{Entry Point,,Entry command})
4522 a symbol assignment (@pxref{Assignments})
4524 an output section description
4526 an overlay description
4529 The @code{ENTRY} command and symbol assignments are permitted inside the
4530 @code{SECTIONS} command for convenience in using the location counter in
4531 those commands. This can also make the linker script easier to
4532 understand because you can use those commands at meaningful points in
4533 the layout of the output file.
4535 Output section descriptions and overlay descriptions are described
4538 If you do not use a @code{SECTIONS} command in your linker script, the
4539 linker will place each input section into an identically named output
4540 section in the order that the sections are first encountered in the
4541 input files. If all input sections are present in the first file, for
4542 example, the order of sections in the output file will match the order
4543 in the first input file. The first section will be at address zero.
4546 * Output Section Description:: Output section description
4547 * Output Section Name:: Output section name
4548 * Output Section Address:: Output section address
4549 * Input Section:: Input section description
4550 * Output Section Data:: Output section data
4551 * Output Section Keywords:: Output section keywords
4552 * Output Section Discarding:: Output section discarding
4553 * Output Section Attributes:: Output section attributes
4554 * Overlay Description:: Overlay description
4557 @node Output Section Description
4558 @subsection Output Section Description
4559 The full description of an output section looks like this:
4562 @var{section} [@var{address}] [(@var{type})] :
4564 [ALIGN(@var{section_align}) | ALIGN_WITH_INPUT]
4565 [SUBALIGN(@var{subsection_align})]
4568 @var{output-section-command}
4569 @var{output-section-command}
4571 @} [>@var{region}] [AT>@var{lma_region}] [:@var{phdr} :@var{phdr} @dots{}] [=@var{fillexp}] [,]
4575 Most output sections do not use most of the optional section attributes.
4577 The whitespace around @var{section} is required, so that the section
4578 name is unambiguous. The colon and the curly braces are also required.
4579 The comma at the end may be required if a @var{fillexp} is used and
4580 the next @var{sections-command} looks like a continuation of the expression.
4581 The line breaks and other white space are optional.
4583 Each @var{output-section-command} may be one of the following:
4587 a symbol assignment (@pxref{Assignments})
4589 an input section description (@pxref{Input Section})
4591 data values to include directly (@pxref{Output Section Data})
4593 a special output section keyword (@pxref{Output Section Keywords})
4596 @node Output Section Name
4597 @subsection Output Section Name
4598 @cindex name, section
4599 @cindex section name
4600 The name of the output section is @var{section}. @var{section} must
4601 meet the constraints of your output format. In formats which only
4602 support a limited number of sections, such as @code{a.out}, the name
4603 must be one of the names supported by the format (@code{a.out}, for
4604 example, allows only @samp{.text}, @samp{.data} or @samp{.bss}). If the
4605 output format supports any number of sections, but with numbers and not
4606 names (as is the case for Oasys), the name should be supplied as a
4607 quoted numeric string. A section name may consist of any sequence of
4608 characters, but a name which contains any unusual characters such as
4609 commas must be quoted.
4611 The output section name @samp{/DISCARD/} is special; @ref{Output Section
4614 @node Output Section Address
4615 @subsection Output Section Address
4616 @cindex address, section
4617 @cindex section address
4618 The @var{address} is an expression for the VMA (the virtual memory
4619 address) of the output section. This address is optional, but if it
4620 is provided then the output address will be set exactly as specified.
4622 If the output address is not specified then one will be chosen for the
4623 section, based on the heuristic below. This address will be adjusted
4624 to fit the alignment requirement of the output section. The
4625 alignment requirement is the strictest alignment of any input section
4626 contained within the output section.
4628 The output section address heuristic is as follows:
4632 If an output memory @var{region} is set for the section then it
4633 is added to this region and its address will be the next free address
4637 If the MEMORY command has been used to create a list of memory
4638 regions then the first region which has attributes compatible with the
4639 section is selected to contain it. The section's output address will
4640 be the next free address in that region; @ref{MEMORY}.
4643 If no memory regions were specified, or none match the section then
4644 the output address will be based on the current value of the location
4652 .text . : @{ *(.text) @}
4659 .text : @{ *(.text) @}
4663 are subtly different. The first will set the address of the
4664 @samp{.text} output section to the current value of the location
4665 counter. The second will set it to the current value of the location
4666 counter aligned to the strictest alignment of any of the @samp{.text}
4669 The @var{address} may be an arbitrary expression; @ref{Expressions}.
4670 For example, if you want to align the section on a 0x10 byte boundary,
4671 so that the lowest four bits of the section address are zero, you could
4672 do something like this:
4674 .text ALIGN(0x10) : @{ *(.text) @}
4677 This works because @code{ALIGN} returns the current location counter
4678 aligned upward to the specified value.
4680 Specifying @var{address} for a section will change the value of the
4681 location counter, provided that the section is non-empty. (Empty
4682 sections are ignored).
4685 @subsection Input Section Description
4686 @cindex input sections
4687 @cindex mapping input sections to output sections
4688 The most common output section command is an input section description.
4690 The input section description is the most basic linker script operation.
4691 You use output sections to tell the linker how to lay out your program
4692 in memory. You use input section descriptions to tell the linker how to
4693 map the input files into your memory layout.
4696 * Input Section Basics:: Input section basics
4697 * Input Section Wildcards:: Input section wildcard patterns
4698 * Input Section Common:: Input section for common symbols
4699 * Input Section Keep:: Input section and garbage collection
4700 * Input Section Example:: Input section example
4703 @node Input Section Basics
4704 @subsubsection Input Section Basics
4705 @cindex input section basics
4706 An input section description consists of a file name optionally followed
4707 by a list of section names in parentheses.
4709 The file name and the section name may be wildcard patterns, which we
4710 describe further below (@pxref{Input Section Wildcards}).
4712 The most common input section description is to include all input
4713 sections with a particular name in the output section. For example, to
4714 include all input @samp{.text} sections, you would write:
4719 Here the @samp{*} is a wildcard which matches any file name. To exclude a list
4720 @cindex EXCLUDE_FILE
4721 of files from matching the file name wildcard, EXCLUDE_FILE may be used to
4722 match all files except the ones specified in the EXCLUDE_FILE list. For
4725 EXCLUDE_FILE (*crtend.o *otherfile.o) *(.ctors)
4728 will cause all .ctors sections from all files except @file{crtend.o}
4729 and @file{otherfile.o} to be included. The EXCLUDE_FILE can also be
4730 placed inside the section list, for example:
4732 *(EXCLUDE_FILE (*crtend.o *otherfile.o) .ctors)
4735 The result of this is identically to the previous example. Supporting
4736 two syntaxes for EXCLUDE_FILE is useful if the section list contains
4737 more than one section, as described below.
4739 There are two ways to include more than one section:
4745 The difference between these is the order in which the @samp{.text} and
4746 @samp{.rdata} input sections will appear in the output section. In the
4747 first example, they will be intermingled, appearing in the same order as
4748 they are found in the linker input. In the second example, all
4749 @samp{.text} input sections will appear first, followed by all
4750 @samp{.rdata} input sections.
4752 When using EXCLUDE_FILE with more than one section, if the exclusion
4753 is within the section list then the exclusion only applies to the
4754 immediately following section, for example:
4756 *(EXCLUDE_FILE (*somefile.o) .text .rdata)
4759 will cause all @samp{.text} sections from all files except
4760 @file{somefile.o} to be included, while all @samp{.rdata} sections
4761 from all files, including @file{somefile.o}, will be included. To
4762 exclude the @samp{.rdata} sections from @file{somefile.o} the example
4763 could be modified to:
4765 *(EXCLUDE_FILE (*somefile.o) .text EXCLUDE_FILE (*somefile.o) .rdata)
4768 Alternatively, placing the EXCLUDE_FILE outside of the section list,
4769 before the input file selection, will cause the exclusion to apply for
4770 all sections. Thus the previous example can be rewritten as:
4772 EXCLUDE_FILE (*somefile.o) *(.text .rdata)
4775 You can specify a file name to include sections from a particular file.
4776 You would do this if one or more of your files contain special data that
4777 needs to be at a particular location in memory. For example:
4782 To refine the sections that are included based on the section flags
4783 of an input section, INPUT_SECTION_FLAGS may be used.
4785 Here is a simple example for using Section header flags for ELF sections:
4790 .text : @{ INPUT_SECTION_FLAGS (SHF_MERGE & SHF_STRINGS) *(.text) @}
4791 .text2 : @{ INPUT_SECTION_FLAGS (!SHF_WRITE) *(.text) @}
4796 In this example, the output section @samp{.text} will be comprised of any
4797 input section matching the name *(.text) whose section header flags
4798 @code{SHF_MERGE} and @code{SHF_STRINGS} are set. The output section
4799 @samp{.text2} will be comprised of any input section matching the name *(.text)
4800 whose section header flag @code{SHF_WRITE} is clear.
4802 You can also specify files within archives by writing a pattern
4803 matching the archive, a colon, then the pattern matching the file,
4804 with no whitespace around the colon.
4808 matches file within archive
4810 matches the whole archive
4812 matches file but not one in an archive
4815 Either one or both of @samp{archive} and @samp{file} can contain shell
4816 wildcards. On DOS based file systems, the linker will assume that a
4817 single letter followed by a colon is a drive specifier, so
4818 @samp{c:myfile.o} is a simple file specification, not @samp{myfile.o}
4819 within an archive called @samp{c}. @samp{archive:file} filespecs may
4820 also be used within an @code{EXCLUDE_FILE} list, but may not appear in
4821 other linker script contexts. For instance, you cannot extract a file
4822 from an archive by using @samp{archive:file} in an @code{INPUT}
4825 If you use a file name without a list of sections, then all sections in
4826 the input file will be included in the output section. This is not
4827 commonly done, but it may by useful on occasion. For example:
4832 When you use a file name which is not an @samp{archive:file} specifier
4833 and does not contain any wild card
4834 characters, the linker will first see if you also specified the file
4835 name on the linker command line or in an @code{INPUT} command. If you
4836 did not, the linker will attempt to open the file as an input file, as
4837 though it appeared on the command line. Note that this differs from an
4838 @code{INPUT} command, because the linker will not search for the file in
4839 the archive search path.
4841 @node Input Section Wildcards
4842 @subsubsection Input Section Wildcard Patterns
4843 @cindex input section wildcards
4844 @cindex wildcard file name patterns
4845 @cindex file name wildcard patterns
4846 @cindex section name wildcard patterns
4847 In an input section description, either the file name or the section
4848 name or both may be wildcard patterns.
4850 The file name of @samp{*} seen in many examples is a simple wildcard
4851 pattern for the file name.
4853 The wildcard patterns are like those used by the Unix shell.
4857 matches any number of characters
4859 matches any single character
4861 matches a single instance of any of the @var{chars}; the @samp{-}
4862 character may be used to specify a range of characters, as in
4863 @samp{[a-z]} to match any lower case letter
4865 quotes the following character
4868 When a file name is matched with a wildcard, the wildcard characters
4869 will not match a @samp{/} character (used to separate directory names on
4870 Unix). A pattern consisting of a single @samp{*} character is an
4871 exception; it will always match any file name, whether it contains a
4872 @samp{/} or not. In a section name, the wildcard characters will match
4873 a @samp{/} character.
4875 File name wildcard patterns only match files which are explicitly
4876 specified on the command line or in an @code{INPUT} command. The linker
4877 does not search directories to expand wildcards.
4879 If a file name matches more than one wildcard pattern, or if a file name
4880 appears explicitly and is also matched by a wildcard pattern, the linker
4881 will use the first match in the linker script. For example, this
4882 sequence of input section descriptions is probably in error, because the
4883 @file{data.o} rule will not be used:
4885 .data : @{ *(.data) @}
4886 .data1 : @{ data.o(.data) @}
4889 @cindex SORT_BY_NAME
4890 Normally, the linker will place files and sections matched by wildcards
4891 in the order in which they are seen during the link. You can change
4892 this by using the @code{SORT_BY_NAME} keyword, which appears before a wildcard
4893 pattern in parentheses (e.g., @code{SORT_BY_NAME(.text*)}). When the
4894 @code{SORT_BY_NAME} keyword is used, the linker will sort the files or sections
4895 into ascending order by name before placing them in the output file.
4897 @cindex SORT_BY_ALIGNMENT
4898 @code{SORT_BY_ALIGNMENT} is similar to @code{SORT_BY_NAME}.
4899 @code{SORT_BY_ALIGNMENT} will sort sections into descending order of
4900 alignment before placing them in the output file. Placing larger
4901 alignments before smaller alignments can reduce the amount of padding
4904 @cindex SORT_BY_INIT_PRIORITY
4905 @code{SORT_BY_INIT_PRIORITY} is also similar to @code{SORT_BY_NAME}.
4906 @code{SORT_BY_INIT_PRIORITY} will sort sections into ascending
4907 numerical order of the GCC init_priority attribute encoded in the
4908 section name before placing them in the output file. In
4909 @code{.init_array.NNNNN} and @code{.fini_array.NNNNN}, @code{NNNNN} is
4910 the init_priority. In @code{.ctors.NNNNN} and @code{.dtors.NNNNN},
4911 @code{NNNNN} is 65535 minus the init_priority.
4914 @code{SORT} is an alias for @code{SORT_BY_NAME}.
4916 When there are nested section sorting commands in linker script, there
4917 can be at most 1 level of nesting for section sorting commands.
4921 @code{SORT_BY_NAME} (@code{SORT_BY_ALIGNMENT} (wildcard section pattern)).
4922 It will sort the input sections by name first, then by alignment if two
4923 sections have the same name.
4925 @code{SORT_BY_ALIGNMENT} (@code{SORT_BY_NAME} (wildcard section pattern)).
4926 It will sort the input sections by alignment first, then by name if two
4927 sections have the same alignment.
4929 @code{SORT_BY_NAME} (@code{SORT_BY_NAME} (wildcard section pattern)) is
4930 treated the same as @code{SORT_BY_NAME} (wildcard section pattern).
4932 @code{SORT_BY_ALIGNMENT} (@code{SORT_BY_ALIGNMENT} (wildcard section pattern))
4933 is treated the same as @code{SORT_BY_ALIGNMENT} (wildcard section pattern).
4935 All other nested section sorting commands are invalid.
4938 When both command-line section sorting option and linker script
4939 section sorting command are used, section sorting command always
4940 takes precedence over the command-line option.
4942 If the section sorting command in linker script isn't nested, the
4943 command-line option will make the section sorting command to be
4944 treated as nested sorting command.
4948 @code{SORT_BY_NAME} (wildcard section pattern ) with
4949 @option{--sort-sections alignment} is equivalent to
4950 @code{SORT_BY_NAME} (@code{SORT_BY_ALIGNMENT} (wildcard section pattern)).
4952 @code{SORT_BY_ALIGNMENT} (wildcard section pattern) with
4953 @option{--sort-section name} is equivalent to
4954 @code{SORT_BY_ALIGNMENT} (@code{SORT_BY_NAME} (wildcard section pattern)).
4957 If the section sorting command in linker script is nested, the
4958 command-line option will be ignored.
4961 @code{SORT_NONE} disables section sorting by ignoring the command-line
4962 section sorting option.
4964 If you ever get confused about where input sections are going, use the
4965 @samp{-M} linker option to generate a map file. The map file shows
4966 precisely how input sections are mapped to output sections.
4968 This example shows how wildcard patterns might be used to partition
4969 files. This linker script directs the linker to place all @samp{.text}
4970 sections in @samp{.text} and all @samp{.bss} sections in @samp{.bss}.
4971 The linker will place the @samp{.data} section from all files beginning
4972 with an upper case character in @samp{.DATA}; for all other files, the
4973 linker will place the @samp{.data} section in @samp{.data}.
4977 .text : @{ *(.text) @}
4978 .DATA : @{ [A-Z]*(.data) @}
4979 .data : @{ *(.data) @}
4980 .bss : @{ *(.bss) @}
4985 @node Input Section Common
4986 @subsubsection Input Section for Common Symbols
4987 @cindex common symbol placement
4988 @cindex uninitialized data placement
4989 A special notation is needed for common symbols, because in many object
4990 file formats common symbols do not have a particular input section. The
4991 linker treats common symbols as though they are in an input section
4992 named @samp{COMMON}.
4994 You may use file names with the @samp{COMMON} section just as with any
4995 other input sections. You can use this to place common symbols from a
4996 particular input file in one section while common symbols from other
4997 input files are placed in another section.
4999 In most cases, common symbols in input files will be placed in the
5000 @samp{.bss} section in the output file. For example:
5002 .bss @{ *(.bss) *(COMMON) @}
5005 @cindex scommon section
5006 @cindex small common symbols
5007 Some object file formats have more than one type of common symbol. For
5008 example, the MIPS ELF object file format distinguishes standard common
5009 symbols and small common symbols. In this case, the linker will use a
5010 different special section name for other types of common symbols. In
5011 the case of MIPS ELF, the linker uses @samp{COMMON} for standard common
5012 symbols and @samp{.scommon} for small common symbols. This permits you
5013 to map the different types of common symbols into memory at different
5017 You will sometimes see @samp{[COMMON]} in old linker scripts. This
5018 notation is now considered obsolete. It is equivalent to
5021 @node Input Section Keep
5022 @subsubsection Input Section and Garbage Collection
5024 @cindex garbage collection
5025 When link-time garbage collection is in use (@samp{--gc-sections}),
5026 it is often useful to mark sections that should not be eliminated.
5027 This is accomplished by surrounding an input section's wildcard entry
5028 with @code{KEEP()}, as in @code{KEEP(*(.init))} or
5029 @code{KEEP(SORT_BY_NAME(*)(.ctors))}.
5031 @node Input Section Example
5032 @subsubsection Input Section Example
5033 The following example is a complete linker script. It tells the linker
5034 to read all of the sections from file @file{all.o} and place them at the
5035 start of output section @samp{outputa} which starts at location
5036 @samp{0x10000}. All of section @samp{.input1} from file @file{foo.o}
5037 follows immediately, in the same output section. All of section
5038 @samp{.input2} from @file{foo.o} goes into output section
5039 @samp{outputb}, followed by section @samp{.input1} from @file{foo1.o}.
5040 All of the remaining @samp{.input1} and @samp{.input2} sections from any
5041 files are written to output section @samp{outputc}.
5069 If an output section's name is the same as the input section's name
5070 and is representable as a C identifier, then the linker will
5071 automatically @pxref{PROVIDE} two symbols: __start_SECNAME and
5072 __stop_SECNAME, where SECNAME is the name of the section. These
5073 indicate the start address and end address of the output section
5074 respectively. Note: most section names are not representable as
5075 C identifiers because they contain a @samp{.} character.
5077 @node Output Section Data
5078 @subsection Output Section Data
5080 @cindex section data
5081 @cindex output section data
5082 @kindex BYTE(@var{expression})
5083 @kindex SHORT(@var{expression})
5084 @kindex LONG(@var{expression})
5085 @kindex QUAD(@var{expression})
5086 @kindex SQUAD(@var{expression})
5087 You can include explicit bytes of data in an output section by using
5088 @code{BYTE}, @code{SHORT}, @code{LONG}, @code{QUAD}, or @code{SQUAD} as
5089 an output section command. Each keyword is followed by an expression in
5090 parentheses providing the value to store (@pxref{Expressions}). The
5091 value of the expression is stored at the current value of the location
5094 The @code{BYTE}, @code{SHORT}, @code{LONG}, and @code{QUAD} commands
5095 store one, two, four, and eight bytes (respectively). After storing the
5096 bytes, the location counter is incremented by the number of bytes
5099 For example, this will store the byte 1 followed by the four byte value
5100 of the symbol @samp{addr}:
5106 When using a 64 bit host or target, @code{QUAD} and @code{SQUAD} are the
5107 same; they both store an 8 byte, or 64 bit, value. When both host and
5108 target are 32 bits, an expression is computed as 32 bits. In this case
5109 @code{QUAD} stores a 32 bit value zero extended to 64 bits, and
5110 @code{SQUAD} stores a 32 bit value sign extended to 64 bits.
5112 If the object file format of the output file has an explicit endianness,
5113 which is the normal case, the value will be stored in that endianness.
5114 When the object file format does not have an explicit endianness, as is
5115 true of, for example, S-records, the value will be stored in the
5116 endianness of the first input object file.
5118 Note---these commands only work inside a section description and not
5119 between them, so the following will produce an error from the linker:
5121 SECTIONS @{@ .text : @{@ *(.text) @}@ LONG(1) .data : @{@ *(.data) @}@ @}@
5123 whereas this will work:
5125 SECTIONS @{@ .text : @{@ *(.text) ; LONG(1) @}@ .data : @{@ *(.data) @}@ @}@
5128 @kindex FILL(@var{expression})
5129 @cindex holes, filling
5130 @cindex unspecified memory
5131 You may use the @code{FILL} command to set the fill pattern for the
5132 current section. It is followed by an expression in parentheses. Any
5133 otherwise unspecified regions of memory within the section (for example,
5134 gaps left due to the required alignment of input sections) are filled
5135 with the value of the expression, repeated as
5136 necessary. A @code{FILL} statement covers memory locations after the
5137 point at which it occurs in the section definition; by including more
5138 than one @code{FILL} statement, you can have different fill patterns in
5139 different parts of an output section.
5141 This example shows how to fill unspecified regions of memory with the
5147 The @code{FILL} command is similar to the @samp{=@var{fillexp}} output
5148 section attribute, but it only affects the
5149 part of the section following the @code{FILL} command, rather than the
5150 entire section. If both are used, the @code{FILL} command takes
5151 precedence. @xref{Output Section Fill}, for details on the fill
5154 @node Output Section Keywords
5155 @subsection Output Section Keywords
5156 There are a couple of keywords which can appear as output section
5160 @kindex CREATE_OBJECT_SYMBOLS
5161 @cindex input filename symbols
5162 @cindex filename symbols
5163 @item CREATE_OBJECT_SYMBOLS
5164 The command tells the linker to create a symbol for each input file.
5165 The name of each symbol will be the name of the corresponding input
5166 file. The section of each symbol will be the output section in which
5167 the @code{CREATE_OBJECT_SYMBOLS} command appears.
5169 This is conventional for the a.out object file format. It is not
5170 normally used for any other object file format.
5172 @kindex CONSTRUCTORS
5173 @cindex C++ constructors, arranging in link
5174 @cindex constructors, arranging in link
5176 When linking using the a.out object file format, the linker uses an
5177 unusual set construct to support C++ global constructors and
5178 destructors. When linking object file formats which do not support
5179 arbitrary sections, such as ECOFF and XCOFF, the linker will
5180 automatically recognize C++ global constructors and destructors by name.
5181 For these object file formats, the @code{CONSTRUCTORS} command tells the
5182 linker to place constructor information in the output section where the
5183 @code{CONSTRUCTORS} command appears. The @code{CONSTRUCTORS} command is
5184 ignored for other object file formats.
5186 The symbol @w{@code{__CTOR_LIST__}} marks the start of the global
5187 constructors, and the symbol @w{@code{__CTOR_END__}} marks the end.
5188 Similarly, @w{@code{__DTOR_LIST__}} and @w{@code{__DTOR_END__}} mark
5189 the start and end of the global destructors. The
5190 first word in the list is the number of entries, followed by the address
5191 of each constructor or destructor, followed by a zero word. The
5192 compiler must arrange to actually run the code. For these object file
5193 formats @sc{gnu} C++ normally calls constructors from a subroutine
5194 @code{__main}; a call to @code{__main} is automatically inserted into
5195 the startup code for @code{main}. @sc{gnu} C++ normally runs
5196 destructors either by using @code{atexit}, or directly from the function
5199 For object file formats such as @code{COFF} or @code{ELF} which support
5200 arbitrary section names, @sc{gnu} C++ will normally arrange to put the
5201 addresses of global constructors and destructors into the @code{.ctors}
5202 and @code{.dtors} sections. Placing the following sequence into your
5203 linker script will build the sort of table which the @sc{gnu} C++
5204 runtime code expects to see.
5208 LONG((__CTOR_END__ - __CTOR_LIST__) / 4 - 2)
5213 LONG((__DTOR_END__ - __DTOR_LIST__) / 4 - 2)
5219 If you are using the @sc{gnu} C++ support for initialization priority,
5220 which provides some control over the order in which global constructors
5221 are run, you must sort the constructors at link time to ensure that they
5222 are executed in the correct order. When using the @code{CONSTRUCTORS}
5223 command, use @samp{SORT_BY_NAME(CONSTRUCTORS)} instead. When using the
5224 @code{.ctors} and @code{.dtors} sections, use @samp{*(SORT_BY_NAME(.ctors))} and
5225 @samp{*(SORT_BY_NAME(.dtors))} instead of just @samp{*(.ctors)} and
5228 Normally the compiler and linker will handle these issues automatically,
5229 and you will not need to concern yourself with them. However, you may
5230 need to consider this if you are using C++ and writing your own linker
5235 @node Output Section Discarding
5236 @subsection Output Section Discarding
5237 @cindex discarding sections
5238 @cindex sections, discarding
5239 @cindex removing sections
5240 The linker will not normally create output sections with no contents.
5241 This is for convenience when referring to input sections that may or
5242 may not be present in any of the input files. For example:
5244 .foo : @{ *(.foo) @}
5247 will only create a @samp{.foo} section in the output file if there is a
5248 @samp{.foo} section in at least one input file, and if the input
5249 sections are not all empty. Other link script directives that allocate
5250 space in an output section will also create the output section. So
5251 too will assignments to dot even if the assignment does not create
5252 space, except for @samp{. = 0}, @samp{. = . + 0}, @samp{. = sym},
5253 @samp{. = . + sym} and @samp{. = ALIGN (. != 0, expr, 1)} when
5254 @samp{sym} is an absolute symbol of value 0 defined in the script.
5255 This allows you to force output of an empty section with @samp{. = .}.
5257 The linker will ignore address assignments (@pxref{Output Section Address})
5258 on discarded output sections, except when the linker script defines
5259 symbols in the output section. In that case the linker will obey
5260 the address assignments, possibly advancing dot even though the
5261 section is discarded.
5264 The special output section name @samp{/DISCARD/} may be used to discard
5265 input sections. Any input sections which are assigned to an output
5266 section named @samp{/DISCARD/} are not included in the output file.
5268 Note, sections that match the @samp{/DISCARD/} output section will be
5269 discarded even if they are in an ELF section group which has other
5270 members which are not being discarded. This is deliberate.
5271 Discarding takes precedence over grouping.
5273 @node Output Section Attributes
5274 @subsection Output Section Attributes
5275 @cindex output section attributes
5276 We showed above that the full description of an output section looked
5281 @var{section} [@var{address}] [(@var{type})] :
5283 [ALIGN(@var{section_align}) | ALIGN_WITH_INPUT]
5284 [SUBALIGN(@var{subsection_align})]
5287 @var{output-section-command}
5288 @var{output-section-command}
5290 @} [>@var{region}] [AT>@var{lma_region}] [:@var{phdr} :@var{phdr} @dots{}] [=@var{fillexp}]
5294 We've already described @var{section}, @var{address}, and
5295 @var{output-section-command}. In this section we will describe the
5296 remaining section attributes.
5299 * Output Section Type:: Output section type
5300 * Output Section LMA:: Output section LMA
5301 * Forced Output Alignment:: Forced Output Alignment
5302 * Forced Input Alignment:: Forced Input Alignment
5303 * Output Section Constraint:: Output section constraint
5304 * Output Section Region:: Output section region
5305 * Output Section Phdr:: Output section phdr
5306 * Output Section Fill:: Output section fill
5309 @node Output Section Type
5310 @subsubsection Output Section Type
5311 Each output section may have a type. The type is a keyword in
5312 parentheses. The following types are defined:
5316 The section should be marked as not loadable, so that it will not be
5317 loaded into memory when the program is run.
5322 These type names are supported for backward compatibility, and are
5323 rarely used. They all have the same effect: the section should be
5324 marked as not allocatable, so that no memory is allocated for the
5325 section when the program is run.
5329 @cindex prevent unnecessary loading
5330 @cindex loading, preventing
5331 The linker normally sets the attributes of an output section based on
5332 the input sections which map into it. You can override this by using
5333 the section type. For example, in the script sample below, the
5334 @samp{ROM} section is addressed at memory location @samp{0} and does not
5335 need to be loaded when the program is run.
5339 ROM 0 (NOLOAD) : @{ @dots{} @}
5345 @node Output Section LMA
5346 @subsubsection Output Section LMA
5347 @kindex AT>@var{lma_region}
5348 @kindex AT(@var{lma})
5349 @cindex load address
5350 @cindex section load address
5351 Every section has a virtual address (VMA) and a load address (LMA); see
5352 @ref{Basic Script Concepts}. The virtual address is specified by the
5353 @pxref{Output Section Address} described earlier. The load address is
5354 specified by the @code{AT} or @code{AT>} keywords. Specifying a load
5355 address is optional.
5357 The @code{AT} keyword takes an expression as an argument. This
5358 specifies the exact load address of the section. The @code{AT>} keyword
5359 takes the name of a memory region as an argument. @xref{MEMORY}. The
5360 load address of the section is set to the next free address in the
5361 region, aligned to the section's alignment requirements.
5363 If neither @code{AT} nor @code{AT>} is specified for an allocatable
5364 section, the linker will use the following heuristic to determine the
5369 If the section has a specific VMA address, then this is used as
5370 the LMA address as well.
5373 If the section is not allocatable then its LMA is set to its VMA.
5376 Otherwise if a memory region can be found that is compatible
5377 with the current section, and this region contains at least one
5378 section, then the LMA is set so the difference between the
5379 VMA and LMA is the same as the difference between the VMA and LMA of
5380 the last section in the located region.
5383 If no memory regions have been declared then a default region
5384 that covers the entire address space is used in the previous step.
5387 If no suitable region could be found, or there was no previous
5388 section then the LMA is set equal to the VMA.
5391 @cindex ROM initialized data
5392 @cindex initialized data in ROM
5393 This feature is designed to make it easy to build a ROM image. For
5394 example, the following linker script creates three output sections: one
5395 called @samp{.text}, which starts at @code{0x1000}, one called
5396 @samp{.mdata}, which is loaded at the end of the @samp{.text} section
5397 even though its VMA is @code{0x2000}, and one called @samp{.bss} to hold
5398 uninitialized data at address @code{0x3000}. The symbol @code{_data} is
5399 defined with the value @code{0x2000}, which shows that the location
5400 counter holds the VMA value, not the LMA value.
5406 .text 0x1000 : @{ *(.text) _etext = . ; @}
5408 AT ( ADDR (.text) + SIZEOF (.text) )
5409 @{ _data = . ; *(.data); _edata = . ; @}
5411 @{ _bstart = . ; *(.bss) *(COMMON) ; _bend = . ;@}
5416 The run-time initialization code for use with a program generated with
5417 this linker script would include something like the following, to copy
5418 the initialized data from the ROM image to its runtime address. Notice
5419 how this code takes advantage of the symbols defined by the linker
5424 extern char _etext, _data, _edata, _bstart, _bend;
5425 char *src = &_etext;
5428 /* ROM has data at end of text; copy it. */
5429 while (dst < &_edata)
5433 for (dst = &_bstart; dst< &_bend; dst++)
5438 @node Forced Output Alignment
5439 @subsubsection Forced Output Alignment
5440 @kindex ALIGN(@var{section_align})
5441 @cindex forcing output section alignment
5442 @cindex output section alignment
5443 You can increase an output section's alignment by using ALIGN. As an
5444 alternative you can enforce that the difference between the VMA and LMA remains
5445 intact throughout this output section with the ALIGN_WITH_INPUT attribute.
5447 @node Forced Input Alignment
5448 @subsubsection Forced Input Alignment
5449 @kindex SUBALIGN(@var{subsection_align})
5450 @cindex forcing input section alignment
5451 @cindex input section alignment
5452 You can force input section alignment within an output section by using
5453 SUBALIGN. The value specified overrides any alignment given by input
5454 sections, whether larger or smaller.
5456 @node Output Section Constraint
5457 @subsubsection Output Section Constraint
5460 @cindex constraints on output sections
5461 You can specify that an output section should only be created if all
5462 of its input sections are read-only or all of its input sections are
5463 read-write by using the keyword @code{ONLY_IF_RO} and
5464 @code{ONLY_IF_RW} respectively.
5466 @node Output Section Region
5467 @subsubsection Output Section Region
5468 @kindex >@var{region}
5469 @cindex section, assigning to memory region
5470 @cindex memory regions and sections
5471 You can assign a section to a previously defined region of memory by
5472 using @samp{>@var{region}}. @xref{MEMORY}.
5474 Here is a simple example:
5477 MEMORY @{ rom : ORIGIN = 0x1000, LENGTH = 0x1000 @}
5478 SECTIONS @{ ROM : @{ *(.text) @} >rom @}
5482 @node Output Section Phdr
5483 @subsubsection Output Section Phdr
5485 @cindex section, assigning to program header
5486 @cindex program headers and sections
5487 You can assign a section to a previously defined program segment by
5488 using @samp{:@var{phdr}}. @xref{PHDRS}. If a section is assigned to
5489 one or more segments, then all subsequent allocated sections will be
5490 assigned to those segments as well, unless they use an explicitly
5491 @code{:@var{phdr}} modifier. You can use @code{:NONE} to tell the
5492 linker to not put the section in any segment at all.
5494 Here is a simple example:
5497 PHDRS @{ text PT_LOAD ; @}
5498 SECTIONS @{ .text : @{ *(.text) @} :text @}
5502 @node Output Section Fill
5503 @subsubsection Output Section Fill
5504 @kindex =@var{fillexp}
5505 @cindex section fill pattern
5506 @cindex fill pattern, entire section
5507 You can set the fill pattern for an entire section by using
5508 @samp{=@var{fillexp}}. @var{fillexp} is an expression
5509 (@pxref{Expressions}). Any otherwise unspecified regions of memory
5510 within the output section (for example, gaps left due to the required
5511 alignment of input sections) will be filled with the value, repeated as
5512 necessary. If the fill expression is a simple hex number, ie. a string
5513 of hex digit starting with @samp{0x} and without a trailing @samp{k} or @samp{M}, then
5514 an arbitrarily long sequence of hex digits can be used to specify the
5515 fill pattern; Leading zeros become part of the pattern too. For all
5516 other cases, including extra parentheses or a unary @code{+}, the fill
5517 pattern is the four least significant bytes of the value of the
5518 expression. In all cases, the number is big-endian.
5520 You can also change the fill value with a @code{FILL} command in the
5521 output section commands; (@pxref{Output Section Data}).
5523 Here is a simple example:
5526 SECTIONS @{ .text : @{ *(.text) @} =0x90909090 @}
5530 @node Overlay Description
5531 @subsection Overlay Description
5534 An overlay description provides an easy way to describe sections which
5535 are to be loaded as part of a single memory image but are to be run at
5536 the same memory address. At run time, some sort of overlay manager will
5537 copy the overlaid sections in and out of the runtime memory address as
5538 required, perhaps by simply manipulating addressing bits. This approach
5539 can be useful, for example, when a certain region of memory is faster
5542 Overlays are described using the @code{OVERLAY} command. The
5543 @code{OVERLAY} command is used within a @code{SECTIONS} command, like an
5544 output section description. The full syntax of the @code{OVERLAY}
5545 command is as follows:
5548 OVERLAY [@var{start}] : [NOCROSSREFS] [AT ( @var{ldaddr} )]
5552 @var{output-section-command}
5553 @var{output-section-command}
5555 @} [:@var{phdr}@dots{}] [=@var{fill}]
5558 @var{output-section-command}
5559 @var{output-section-command}
5561 @} [:@var{phdr}@dots{}] [=@var{fill}]
5563 @} [>@var{region}] [:@var{phdr}@dots{}] [=@var{fill}] [,]
5567 Everything is optional except @code{OVERLAY} (a keyword), and each
5568 section must have a name (@var{secname1} and @var{secname2} above). The
5569 section definitions within the @code{OVERLAY} construct are identical to
5570 those within the general @code{SECTIONS} construct (@pxref{SECTIONS}),
5571 except that no addresses and no memory regions may be defined for
5572 sections within an @code{OVERLAY}.
5574 The comma at the end may be required if a @var{fill} is used and
5575 the next @var{sections-command} looks like a continuation of the expression.
5577 The sections are all defined with the same starting address. The load
5578 addresses of the sections are arranged such that they are consecutive in
5579 memory starting at the load address used for the @code{OVERLAY} as a
5580 whole (as with normal section definitions, the load address is optional,
5581 and defaults to the start address; the start address is also optional,
5582 and defaults to the current value of the location counter).
5584 If the @code{NOCROSSREFS} keyword is used, and there are any
5585 references among the sections, the linker will report an error. Since
5586 the sections all run at the same address, it normally does not make
5587 sense for one section to refer directly to another.
5588 @xref{Miscellaneous Commands, NOCROSSREFS}.
5590 For each section within the @code{OVERLAY}, the linker automatically
5591 provides two symbols. The symbol @code{__load_start_@var{secname}} is
5592 defined as the starting load address of the section. The symbol
5593 @code{__load_stop_@var{secname}} is defined as the final load address of
5594 the section. Any characters within @var{secname} which are not legal
5595 within C identifiers are removed. C (or assembler) code may use these
5596 symbols to move the overlaid sections around as necessary.
5598 At the end of the overlay, the value of the location counter is set to
5599 the start address of the overlay plus the size of the largest section.
5601 Here is an example. Remember that this would appear inside a
5602 @code{SECTIONS} construct.
5605 OVERLAY 0x1000 : AT (0x4000)
5607 .text0 @{ o1/*.o(.text) @}
5608 .text1 @{ o2/*.o(.text) @}
5613 This will define both @samp{.text0} and @samp{.text1} to start at
5614 address 0x1000. @samp{.text0} will be loaded at address 0x4000, and
5615 @samp{.text1} will be loaded immediately after @samp{.text0}. The
5616 following symbols will be defined if referenced: @code{__load_start_text0},
5617 @code{__load_stop_text0}, @code{__load_start_text1},
5618 @code{__load_stop_text1}.
5620 C code to copy overlay @code{.text1} into the overlay area might look
5625 extern char __load_start_text1, __load_stop_text1;
5626 memcpy ((char *) 0x1000, &__load_start_text1,
5627 &__load_stop_text1 - &__load_start_text1);
5631 Note that the @code{OVERLAY} command is just syntactic sugar, since
5632 everything it does can be done using the more basic commands. The above
5633 example could have been written identically as follows.
5637 .text0 0x1000 : AT (0x4000) @{ o1/*.o(.text) @}
5638 PROVIDE (__load_start_text0 = LOADADDR (.text0));
5639 PROVIDE (__load_stop_text0 = LOADADDR (.text0) + SIZEOF (.text0));
5640 .text1 0x1000 : AT (0x4000 + SIZEOF (.text0)) @{ o2/*.o(.text) @}
5641 PROVIDE (__load_start_text1 = LOADADDR (.text1));
5642 PROVIDE (__load_stop_text1 = LOADADDR (.text1) + SIZEOF (.text1));
5643 . = 0x1000 + MAX (SIZEOF (.text0), SIZEOF (.text1));
5648 @section MEMORY Command
5650 @cindex memory regions
5651 @cindex regions of memory
5652 @cindex allocating memory
5653 @cindex discontinuous memory
5654 The linker's default configuration permits allocation of all available
5655 memory. You can override this by using the @code{MEMORY} command.
5657 The @code{MEMORY} command describes the location and size of blocks of
5658 memory in the target. You can use it to describe which memory regions
5659 may be used by the linker, and which memory regions it must avoid. You
5660 can then assign sections to particular memory regions. The linker will
5661 set section addresses based on the memory regions, and will warn about
5662 regions that become too full. The linker will not shuffle sections
5663 around to fit into the available regions.
5665 A linker script may contain many uses of the @code{MEMORY} command,
5666 however, all memory blocks defined are treated as if they were
5667 specified inside a single @code{MEMORY} command. The syntax for
5673 @var{name} [(@var{attr})] : ORIGIN = @var{origin}, LENGTH = @var{len}
5679 The @var{name} is a name used in the linker script to refer to the
5680 region. The region name has no meaning outside of the linker script.
5681 Region names are stored in a separate name space, and will not conflict
5682 with symbol names, file names, or section names. Each memory region
5683 must have a distinct name within the @code{MEMORY} command. However you can
5684 add later alias names to existing memory regions with the @ref{REGION_ALIAS}
5687 @cindex memory region attributes
5688 The @var{attr} string is an optional list of attributes that specify
5689 whether to use a particular memory region for an input section which is
5690 not explicitly mapped in the linker script. As described in
5691 @ref{SECTIONS}, if you do not specify an output section for some input
5692 section, the linker will create an output section with the same name as
5693 the input section. If you define region attributes, the linker will use
5694 them to select the memory region for the output section that it creates.
5696 The @var{attr} string must consist only of the following characters:
5711 Invert the sense of any of the attributes that follow
5714 If an unmapped section matches any of the listed attributes other than
5715 @samp{!}, it will be placed in the memory region. The @samp{!}
5716 attribute reverses the test for the characters that follow, so that an
5717 unmapped section will be placed in the memory region only if it does
5718 not match any of the attributes listed afterwards. Thus an attribute
5719 string of @samp{RW!X} will match any unmapped section that has either
5720 or both of the @samp{R} and @samp{W} attributes, but only as long as
5721 the section does not also have the @samp{X} attribute.
5726 The @var{origin} is an numerical expression for the start address of
5727 the memory region. The expression must evaluate to a constant and it
5728 cannot involve any symbols. The keyword @code{ORIGIN} may be
5729 abbreviated to @code{org} or @code{o} (but not, for example,
5735 The @var{len} is an expression for the size in bytes of the memory
5736 region. As with the @var{origin} expression, the expression must
5737 be numerical only and must evaluate to a constant. The keyword
5738 @code{LENGTH} may be abbreviated to @code{len} or @code{l}.
5740 In the following example, we specify that there are two memory regions
5741 available for allocation: one starting at @samp{0} for 256 kilobytes,
5742 and the other starting at @samp{0x40000000} for four megabytes. The
5743 linker will place into the @samp{rom} memory region every section which
5744 is not explicitly mapped into a memory region, and is either read-only
5745 or executable. The linker will place other sections which are not
5746 explicitly mapped into a memory region into the @samp{ram} memory
5753 rom (rx) : ORIGIN = 0, LENGTH = 256K
5754 ram (!rx) : org = 0x40000000, l = 4M
5759 Once you define a memory region, you can direct the linker to place
5760 specific output sections into that memory region by using the
5761 @samp{>@var{region}} output section attribute. For example, if you have
5762 a memory region named @samp{mem}, you would use @samp{>mem} in the
5763 output section definition. @xref{Output Section Region}. If no address
5764 was specified for the output section, the linker will set the address to
5765 the next available address within the memory region. If the combined
5766 output sections directed to a memory region are too large for the
5767 region, the linker will issue an error message.
5769 It is possible to access the origin and length of a memory in an
5770 expression via the @code{ORIGIN(@var{memory})} and
5771 @code{LENGTH(@var{memory})} functions:
5775 _fstack = ORIGIN(ram) + LENGTH(ram) - 4;
5780 @section PHDRS Command
5782 @cindex program headers
5783 @cindex ELF program headers
5784 @cindex program segments
5785 @cindex segments, ELF
5786 The ELF object file format uses @dfn{program headers}, also knows as
5787 @dfn{segments}. The program headers describe how the program should be
5788 loaded into memory. You can print them out by using the @code{objdump}
5789 program with the @samp{-p} option.
5791 When you run an ELF program on a native ELF system, the system loader
5792 reads the program headers in order to figure out how to load the
5793 program. This will only work if the program headers are set correctly.
5794 This manual does not describe the details of how the system loader
5795 interprets program headers; for more information, see the ELF ABI.
5797 The linker will create reasonable program headers by default. However,
5798 in some cases, you may need to specify the program headers more
5799 precisely. You may use the @code{PHDRS} command for this purpose. When
5800 the linker sees the @code{PHDRS} command in the linker script, it will
5801 not create any program headers other than the ones specified.
5803 The linker only pays attention to the @code{PHDRS} command when
5804 generating an ELF output file. In other cases, the linker will simply
5805 ignore @code{PHDRS}.
5807 This is the syntax of the @code{PHDRS} command. The words @code{PHDRS},
5808 @code{FILEHDR}, @code{AT}, and @code{FLAGS} are keywords.
5814 @var{name} @var{type} [ FILEHDR ] [ PHDRS ] [ AT ( @var{address} ) ]
5815 [ FLAGS ( @var{flags} ) ] ;
5820 The @var{name} is used only for reference in the @code{SECTIONS} command
5821 of the linker script. It is not put into the output file. Program
5822 header names are stored in a separate name space, and will not conflict
5823 with symbol names, file names, or section names. Each program header
5824 must have a distinct name. The headers are processed in order and it
5825 is usual for them to map to sections in ascending load address order.
5827 Certain program header types describe segments of memory which the
5828 system loader will load from the file. In the linker script, you
5829 specify the contents of these segments by placing allocatable output
5830 sections in the segments. You use the @samp{:@var{phdr}} output section
5831 attribute to place a section in a particular segment. @xref{Output
5834 It is normal to put certain sections in more than one segment. This
5835 merely implies that one segment of memory contains another. You may
5836 repeat @samp{:@var{phdr}}, using it once for each segment which should
5837 contain the section.
5839 If you place a section in one or more segments using @samp{:@var{phdr}},
5840 then the linker will place all subsequent allocatable sections which do
5841 not specify @samp{:@var{phdr}} in the same segments. This is for
5842 convenience, since generally a whole set of contiguous sections will be
5843 placed in a single segment. You can use @code{:NONE} to override the
5844 default segment and tell the linker to not put the section in any
5849 You may use the @code{FILEHDR} and @code{PHDRS} keywords after
5850 the program header type to further describe the contents of the segment.
5851 The @code{FILEHDR} keyword means that the segment should include the ELF
5852 file header. The @code{PHDRS} keyword means that the segment should
5853 include the ELF program headers themselves. If applied to a loadable
5854 segment (@code{PT_LOAD}), all prior loadable segments must have one of
5857 The @var{type} may be one of the following. The numbers indicate the
5858 value of the keyword.
5861 @item @code{PT_NULL} (0)
5862 Indicates an unused program header.
5864 @item @code{PT_LOAD} (1)
5865 Indicates that this program header describes a segment to be loaded from
5868 @item @code{PT_DYNAMIC} (2)
5869 Indicates a segment where dynamic linking information can be found.
5871 @item @code{PT_INTERP} (3)
5872 Indicates a segment where the name of the program interpreter may be
5875 @item @code{PT_NOTE} (4)
5876 Indicates a segment holding note information.
5878 @item @code{PT_SHLIB} (5)
5879 A reserved program header type, defined but not specified by the ELF
5882 @item @code{PT_PHDR} (6)
5883 Indicates a segment where the program headers may be found.
5885 @item @code{PT_TLS} (7)
5886 Indicates a segment containing thread local storage.
5888 @item @var{expression}
5889 An expression giving the numeric type of the program header. This may
5890 be used for types not defined above.
5893 You can specify that a segment should be loaded at a particular address
5894 in memory by using an @code{AT} expression. This is identical to the
5895 @code{AT} command used as an output section attribute (@pxref{Output
5896 Section LMA}). The @code{AT} command for a program header overrides the
5897 output section attribute.
5899 The linker will normally set the segment flags based on the sections
5900 which comprise the segment. You may use the @code{FLAGS} keyword to
5901 explicitly specify the segment flags. The value of @var{flags} must be
5902 an integer. It is used to set the @code{p_flags} field of the program
5905 Here is an example of @code{PHDRS}. This shows a typical set of program
5906 headers used on a native ELF system.
5912 headers PT_PHDR PHDRS ;
5914 text PT_LOAD FILEHDR PHDRS ;
5916 dynamic PT_DYNAMIC ;
5922 .interp : @{ *(.interp) @} :text :interp
5923 .text : @{ *(.text) @} :text
5924 .rodata : @{ *(.rodata) @} /* defaults to :text */
5926 . = . + 0x1000; /* move to a new page in memory */
5927 .data : @{ *(.data) @} :data
5928 .dynamic : @{ *(.dynamic) @} :data :dynamic
5935 @section VERSION Command
5936 @kindex VERSION @{script text@}
5937 @cindex symbol versions
5938 @cindex version script
5939 @cindex versions of symbols
5940 The linker supports symbol versions when using ELF. Symbol versions are
5941 only useful when using shared libraries. The dynamic linker can use
5942 symbol versions to select a specific version of a function when it runs
5943 a program that may have been linked against an earlier version of the
5946 You can include a version script directly in the main linker script, or
5947 you can supply the version script as an implicit linker script. You can
5948 also use the @samp{--version-script} linker option.
5950 The syntax of the @code{VERSION} command is simply
5952 VERSION @{ version-script-commands @}
5955 The format of the version script commands is identical to that used by
5956 Sun's linker in Solaris 2.5. The version script defines a tree of
5957 version nodes. You specify the node names and interdependencies in the
5958 version script. You can specify which symbols are bound to which
5959 version nodes, and you can reduce a specified set of symbols to local
5960 scope so that they are not globally visible outside of the shared
5963 The easiest way to demonstrate the version script language is with a few
5989 This example version script defines three version nodes. The first
5990 version node defined is @samp{VERS_1.1}; it has no other dependencies.
5991 The script binds the symbol @samp{foo1} to @samp{VERS_1.1}. It reduces
5992 a number of symbols to local scope so that they are not visible outside
5993 of the shared library; this is done using wildcard patterns, so that any
5994 symbol whose name begins with @samp{old}, @samp{original}, or @samp{new}
5995 is matched. The wildcard patterns available are the same as those used
5996 in the shell when matching filenames (also known as ``globbing'').
5997 However, if you specify the symbol name inside double quotes, then the
5998 name is treated as literal, rather than as a glob pattern.
6000 Next, the version script defines node @samp{VERS_1.2}. This node
6001 depends upon @samp{VERS_1.1}. The script binds the symbol @samp{foo2}
6002 to the version node @samp{VERS_1.2}.
6004 Finally, the version script defines node @samp{VERS_2.0}. This node
6005 depends upon @samp{VERS_1.2}. The scripts binds the symbols @samp{bar1}
6006 and @samp{bar2} are bound to the version node @samp{VERS_2.0}.
6008 When the linker finds a symbol defined in a library which is not
6009 specifically bound to a version node, it will effectively bind it to an
6010 unspecified base version of the library. You can bind all otherwise
6011 unspecified symbols to a given version node by using @samp{global: *;}
6012 somewhere in the version script. Note that it's slightly crazy to use
6013 wildcards in a global spec except on the last version node. Global
6014 wildcards elsewhere run the risk of accidentally adding symbols to the
6015 set exported for an old version. That's wrong since older versions
6016 ought to have a fixed set of symbols.
6018 The names of the version nodes have no specific meaning other than what
6019 they might suggest to the person reading them. The @samp{2.0} version
6020 could just as well have appeared in between @samp{1.1} and @samp{1.2}.
6021 However, this would be a confusing way to write a version script.
6023 Node name can be omitted, provided it is the only version node
6024 in the version script. Such version script doesn't assign any versions to
6025 symbols, only selects which symbols will be globally visible out and which
6029 @{ global: foo; bar; local: *; @};
6032 When you link an application against a shared library that has versioned
6033 symbols, the application itself knows which version of each symbol it
6034 requires, and it also knows which version nodes it needs from each
6035 shared library it is linked against. Thus at runtime, the dynamic
6036 loader can make a quick check to make sure that the libraries you have
6037 linked against do in fact supply all of the version nodes that the
6038 application will need to resolve all of the dynamic symbols. In this
6039 way it is possible for the dynamic linker to know with certainty that
6040 all external symbols that it needs will be resolvable without having to
6041 search for each symbol reference.
6043 The symbol versioning is in effect a much more sophisticated way of
6044 doing minor version checking that SunOS does. The fundamental problem
6045 that is being addressed here is that typically references to external
6046 functions are bound on an as-needed basis, and are not all bound when
6047 the application starts up. If a shared library is out of date, a
6048 required interface may be missing; when the application tries to use
6049 that interface, it may suddenly and unexpectedly fail. With symbol
6050 versioning, the user will get a warning when they start their program if
6051 the libraries being used with the application are too old.
6053 There are several GNU extensions to Sun's versioning approach. The
6054 first of these is the ability to bind a symbol to a version node in the
6055 source file where the symbol is defined instead of in the versioning
6056 script. This was done mainly to reduce the burden on the library
6057 maintainer. You can do this by putting something like:
6059 __asm__(".symver original_foo,foo@@VERS_1.1");
6062 in the C source file. This renames the function @samp{original_foo} to
6063 be an alias for @samp{foo} bound to the version node @samp{VERS_1.1}.
6064 The @samp{local:} directive can be used to prevent the symbol
6065 @samp{original_foo} from being exported. A @samp{.symver} directive
6066 takes precedence over a version script.
6068 The second GNU extension is to allow multiple versions of the same
6069 function to appear in a given shared library. In this way you can make
6070 an incompatible change to an interface without increasing the major
6071 version number of the shared library, while still allowing applications
6072 linked against the old interface to continue to function.
6074 To do this, you must use multiple @samp{.symver} directives in the
6075 source file. Here is an example:
6078 __asm__(".symver original_foo,foo@@");
6079 __asm__(".symver old_foo,foo@@VERS_1.1");
6080 __asm__(".symver old_foo1,foo@@VERS_1.2");
6081 __asm__(".symver new_foo,foo@@@@VERS_2.0");
6084 In this example, @samp{foo@@} represents the symbol @samp{foo} bound to the
6085 unspecified base version of the symbol. The source file that contains this
6086 example would define 4 C functions: @samp{original_foo}, @samp{old_foo},
6087 @samp{old_foo1}, and @samp{new_foo}.
6089 When you have multiple definitions of a given symbol, there needs to be
6090 some way to specify a default version to which external references to
6091 this symbol will be bound. You can do this with the
6092 @samp{foo@@@@VERS_2.0} type of @samp{.symver} directive. You can only
6093 declare one version of a symbol as the default in this manner; otherwise
6094 you would effectively have multiple definitions of the same symbol.
6096 If you wish to bind a reference to a specific version of the symbol
6097 within the shared library, you can use the aliases of convenience
6098 (i.e., @samp{old_foo}), or you can use the @samp{.symver} directive to
6099 specifically bind to an external version of the function in question.
6101 You can also specify the language in the version script:
6104 VERSION extern "lang" @{ version-script-commands @}
6107 The supported @samp{lang}s are @samp{C}, @samp{C++}, and @samp{Java}.
6108 The linker will iterate over the list of symbols at the link time and
6109 demangle them according to @samp{lang} before matching them to the
6110 patterns specified in @samp{version-script-commands}. The default
6111 @samp{lang} is @samp{C}.
6113 Demangled names may contains spaces and other special characters. As
6114 described above, you can use a glob pattern to match demangled names,
6115 or you can use a double-quoted string to match the string exactly. In
6116 the latter case, be aware that minor differences (such as differing
6117 whitespace) between the version script and the demangler output will
6118 cause a mismatch. As the exact string generated by the demangler
6119 might change in the future, even if the mangled name does not, you
6120 should check that all of your version directives are behaving as you
6121 expect when you upgrade.
6124 @section Expressions in Linker Scripts
6127 The syntax for expressions in the linker script language is identical to
6128 that of C expressions. All expressions are evaluated as integers. All
6129 expressions are evaluated in the same size, which is 32 bits if both the
6130 host and target are 32 bits, and is otherwise 64 bits.
6132 You can use and set symbol values in expressions.
6134 The linker defines several special purpose builtin functions for use in
6138 * Constants:: Constants
6139 * Symbolic Constants:: Symbolic constants
6140 * Symbols:: Symbol Names
6141 * Orphan Sections:: Orphan Sections
6142 * Location Counter:: The Location Counter
6143 * Operators:: Operators
6144 * Evaluation:: Evaluation
6145 * Expression Section:: The Section of an Expression
6146 * Builtin Functions:: Builtin Functions
6150 @subsection Constants
6151 @cindex integer notation
6152 @cindex constants in linker scripts
6153 All constants are integers.
6155 As in C, the linker considers an integer beginning with @samp{0} to be
6156 octal, and an integer beginning with @samp{0x} or @samp{0X} to be
6157 hexadecimal. Alternatively the linker accepts suffixes of @samp{h} or
6158 @samp{H} for hexadecimal, @samp{o} or @samp{O} for octal, @samp{b} or
6159 @samp{B} for binary and @samp{d} or @samp{D} for decimal. Any integer
6160 value without a prefix or a suffix is considered to be decimal.
6162 @cindex scaled integers
6163 @cindex K and M integer suffixes
6164 @cindex M and K integer suffixes
6165 @cindex suffixes for integers
6166 @cindex integer suffixes
6167 In addition, you can use the suffixes @code{K} and @code{M} to scale a
6171 @c END TEXI2ROFF-KILL
6172 @code{1024} or @code{1024*1024}
6176 ${\rm 1024}$ or ${\rm 1024}^2$
6178 @c END TEXI2ROFF-KILL
6179 respectively. For example, the following
6180 all refer to the same quantity:
6189 Note - the @code{K} and @code{M} suffixes cannot be used in
6190 conjunction with the base suffixes mentioned above.
6192 @node Symbolic Constants
6193 @subsection Symbolic Constants
6194 @cindex symbolic constants
6196 It is possible to refer to target-specific constants via the use of
6197 the @code{CONSTANT(@var{name})} operator, where @var{name} is one of:
6202 The target's maximum page size.
6204 @item COMMONPAGESIZE
6205 @kindex COMMONPAGESIZE
6206 The target's default page size.
6212 .text ALIGN (CONSTANT (MAXPAGESIZE)) : @{ *(.text) @}
6215 will create a text section aligned to the largest page boundary
6216 supported by the target.
6219 @subsection Symbol Names
6220 @cindex symbol names
6222 @cindex quoted symbol names
6224 Unless quoted, symbol names start with a letter, underscore, or period
6225 and may include letters, digits, underscores, periods, and hyphens.
6226 Unquoted symbol names must not conflict with any keywords. You can
6227 specify a symbol which contains odd characters or has the same name as a
6228 keyword by surrounding the symbol name in double quotes:
6231 "with a space" = "also with a space" + 10;
6234 Since symbols can contain many non-alphabetic characters, it is safest
6235 to delimit symbols with spaces. For example, @samp{A-B} is one symbol,
6236 whereas @samp{A - B} is an expression involving subtraction.
6238 @node Orphan Sections
6239 @subsection Orphan Sections
6241 Orphan sections are sections present in the input files which
6242 are not explicitly placed into the output file by the linker
6243 script. The linker will still copy these sections into the
6244 output file by either finding, or creating a suitable output section
6245 in which to place the orphaned input section.
6247 If the name of an orphaned input section exactly matches the name of
6248 an existing output section, then the orphaned input section will be
6249 placed at the end of that output section.
6251 If there is no output section with a matching name then new output
6252 sections will be created. Each new output section will have the same
6253 name as the orphan section placed within it. If there are multiple
6254 orphan sections with the same name, these will all be combined into
6255 one new output section.
6257 If new output sections are created to hold orphaned input sections,
6258 then the linker must decide where to place these new output sections
6259 in relation to existing output sections. On most modern targets, the
6260 linker attempts to place orphan sections after sections of the same
6261 attribute, such as code vs data, loadable vs non-loadable, etc. If no
6262 sections with matching attributes are found, or your target lacks this
6263 support, the orphan section is placed at the end of the file.
6265 The command-line options @samp{--orphan-handling} and @samp{--unique}
6266 (@pxref{Options,,Command-line Options}) can be used to control which
6267 output sections an orphan is placed in.
6269 @node Location Counter
6270 @subsection The Location Counter
6273 @cindex location counter
6274 @cindex current output location
6275 The special linker variable @dfn{dot} @samp{.} always contains the
6276 current output location counter. Since the @code{.} always refers to a
6277 location in an output section, it may only appear in an expression
6278 within a @code{SECTIONS} command. The @code{.} symbol may appear
6279 anywhere that an ordinary symbol is allowed in an expression.
6282 Assigning a value to @code{.} will cause the location counter to be
6283 moved. This may be used to create holes in the output section. The
6284 location counter may not be moved backwards inside an output section,
6285 and may not be moved backwards outside of an output section if so
6286 doing creates areas with overlapping LMAs.
6302 In the previous example, the @samp{.text} section from @file{file1} is
6303 located at the beginning of the output section @samp{output}. It is
6304 followed by a 1000 byte gap. Then the @samp{.text} section from
6305 @file{file2} appears, also with a 1000 byte gap following before the
6306 @samp{.text} section from @file{file3}. The notation @samp{= 0x12345678}
6307 specifies what data to write in the gaps (@pxref{Output Section Fill}).
6309 @cindex dot inside sections
6310 Note: @code{.} actually refers to the byte offset from the start of the
6311 current containing object. Normally this is the @code{SECTIONS}
6312 statement, whose start address is 0, hence @code{.} can be used as an
6313 absolute address. If @code{.} is used inside a section description
6314 however, it refers to the byte offset from the start of that section,
6315 not an absolute address. Thus in a script like this:
6333 The @samp{.text} section will be assigned a starting address of 0x100
6334 and a size of exactly 0x200 bytes, even if there is not enough data in
6335 the @samp{.text} input sections to fill this area. (If there is too
6336 much data, an error will be produced because this would be an attempt to
6337 move @code{.} backwards). The @samp{.data} section will start at 0x500
6338 and it will have an extra 0x600 bytes worth of space after the end of
6339 the values from the @samp{.data} input sections and before the end of
6340 the @samp{.data} output section itself.
6342 @cindex dot outside sections
6343 Setting symbols to the value of the location counter outside of an
6344 output section statement can result in unexpected values if the linker
6345 needs to place orphan sections. For example, given the following:
6351 .text: @{ *(.text) @}
6355 .data: @{ *(.data) @}
6360 If the linker needs to place some input section, e.g. @code{.rodata},
6361 not mentioned in the script, it might choose to place that section
6362 between @code{.text} and @code{.data}. You might think the linker
6363 should place @code{.rodata} on the blank line in the above script, but
6364 blank lines are of no particular significance to the linker. As well,
6365 the linker doesn't associate the above symbol names with their
6366 sections. Instead, it assumes that all assignments or other
6367 statements belong to the previous output section, except for the
6368 special case of an assignment to @code{.}. I.e., the linker will
6369 place the orphan @code{.rodata} section as if the script was written
6376 .text: @{ *(.text) @}
6380 .rodata: @{ *(.rodata) @}
6381 .data: @{ *(.data) @}
6386 This may or may not be the script author's intention for the value of
6387 @code{start_of_data}. One way to influence the orphan section
6388 placement is to assign the location counter to itself, as the linker
6389 assumes that an assignment to @code{.} is setting the start address of
6390 a following output section and thus should be grouped with that
6391 section. So you could write:
6397 .text: @{ *(.text) @}
6402 .data: @{ *(.data) @}
6407 Now, the orphan @code{.rodata} section will be placed between
6408 @code{end_of_text} and @code{start_of_data}.
6412 @subsection Operators
6413 @cindex operators for arithmetic
6414 @cindex arithmetic operators
6415 @cindex precedence in expressions
6416 The linker recognizes the standard C set of arithmetic operators, with
6417 the standard bindings and precedence levels:
6420 @c END TEXI2ROFF-KILL
6422 precedence associativity Operators Notes
6428 5 left == != > < <= >=
6434 11 right &= += -= *= /= (2)
6438 (1) Prefix operators
6439 (2) @xref{Assignments}.
6443 \vskip \baselineskip
6444 %"lispnarrowing" is the extra indent used generally for smallexample
6445 \hskip\lispnarrowing\vbox{\offinterlineskip
6448 {\vrule#&\strut\hfil\ #\ \hfil&\vrule#&\strut\hfil\ #\ \hfil&\vrule#&\strut\hfil\ {\tt #}\ \hfil&\vrule#\cr
6449 height2pt&\omit&&\omit&&\omit&\cr
6450 &Precedence&& Associativity &&{\rm Operators}&\cr
6451 height2pt&\omit&&\omit&&\omit&\cr
6453 height2pt&\omit&&\omit&&\omit&\cr
6455 % '176 is tilde, '~' in tt font
6456 &1&&left&&\qquad- \char'176\ !\qquad\dag&\cr
6457 &2&&left&&* / \%&\cr
6460 &5&&left&&== != > < <= >=&\cr
6463 &8&&left&&{\&\&}&\cr
6466 &11&&right&&\qquad\&= += -= *= /=\qquad\ddag&\cr
6468 height2pt&\omit&&\omit&&\omit&\cr}
6473 @obeylines@parskip=0pt@parindent=0pt
6474 @dag@quad Prefix operators.
6475 @ddag@quad @xref{Assignments}.
6478 @c END TEXI2ROFF-KILL
6481 @subsection Evaluation
6482 @cindex lazy evaluation
6483 @cindex expression evaluation order
6484 The linker evaluates expressions lazily. It only computes the value of
6485 an expression when absolutely necessary.
6487 The linker needs some information, such as the value of the start
6488 address of the first section, and the origins and lengths of memory
6489 regions, in order to do any linking at all. These values are computed
6490 as soon as possible when the linker reads in the linker script.
6492 However, other values (such as symbol values) are not known or needed
6493 until after storage allocation. Such values are evaluated later, when
6494 other information (such as the sizes of output sections) is available
6495 for use in the symbol assignment expression.
6497 The sizes of sections cannot be known until after allocation, so
6498 assignments dependent upon these are not performed until after
6501 Some expressions, such as those depending upon the location counter
6502 @samp{.}, must be evaluated during section allocation.
6504 If the result of an expression is required, but the value is not
6505 available, then an error results. For example, a script like the
6511 .text 9+this_isnt_constant :
6517 will cause the error message @samp{non constant expression for initial
6520 @node Expression Section
6521 @subsection The Section of an Expression
6522 @cindex expression sections
6523 @cindex absolute expressions
6524 @cindex relative expressions
6525 @cindex absolute and relocatable symbols
6526 @cindex relocatable and absolute symbols
6527 @cindex symbols, relocatable and absolute
6528 Addresses and symbols may be section relative, or absolute. A section
6529 relative symbol is relocatable. If you request relocatable output
6530 using the @samp{-r} option, a further link operation may change the
6531 value of a section relative symbol. On the other hand, an absolute
6532 symbol will retain the same value throughout any further link
6535 Some terms in linker expressions are addresses. This is true of
6536 section relative symbols and for builtin functions that return an
6537 address, such as @code{ADDR}, @code{LOADADDR}, @code{ORIGIN} and
6538 @code{SEGMENT_START}. Other terms are simply numbers, or are builtin
6539 functions that return a non-address value, such as @code{LENGTH}.
6540 One complication is that unless you set @code{LD_FEATURE ("SANE_EXPR")}
6541 (@pxref{Miscellaneous Commands}), numbers and absolute symbols are treated
6542 differently depending on their location, for compatibility with older
6543 versions of @code{ld}. Expressions appearing outside an output
6544 section definition treat all numbers as absolute addresses.
6545 Expressions appearing inside an output section definition treat
6546 absolute symbols as numbers. If @code{LD_FEATURE ("SANE_EXPR")} is
6547 given, then absolute symbols and numbers are simply treated as numbers
6550 In the following simple example,
6557 __executable_start = 0x100;
6561 __data_start = 0x10;
6569 both @code{.} and @code{__executable_start} are set to the absolute
6570 address 0x100 in the first two assignments, then both @code{.} and
6571 @code{__data_start} are set to 0x10 relative to the @code{.data}
6572 section in the second two assignments.
6574 For expressions involving numbers, relative addresses and absolute
6575 addresses, ld follows these rules to evaluate terms:
6579 Unary operations on an absolute address or number, and binary
6580 operations on two absolute addresses or two numbers, or between one
6581 absolute address and a number, apply the operator to the value(s).
6583 Unary operations on a relative address, and binary operations on two
6584 relative addresses in the same section or between one relative address
6585 and a number, apply the operator to the offset part of the address(es).
6587 Other binary operations, that is, between two relative addresses not
6588 in the same section, or between a relative address and an absolute
6589 address, first convert any non-absolute term to an absolute address
6590 before applying the operator.
6593 The result section of each sub-expression is as follows:
6597 An operation involving only numbers results in a number.
6599 The result of comparisons, @samp{&&} and @samp{||} is also a number.
6601 The result of other binary arithmetic and logical operations on two
6602 relative addresses in the same section or two absolute addresses
6603 (after above conversions) is also a number when
6604 @code{LD_FEATURE ("SANE_EXPR")} or inside an output section definition
6605 but an absolute address otherwise.
6607 The result of other operations on relative addresses or one
6608 relative address and a number, is a relative address in the same
6609 section as the relative operand(s).
6611 The result of other operations on absolute addresses (after above
6612 conversions) is an absolute address.
6615 You can use the builtin function @code{ABSOLUTE} to force an expression
6616 to be absolute when it would otherwise be relative. For example, to
6617 create an absolute symbol set to the address of the end of the output
6618 section @samp{.data}:
6622 .data : @{ *(.data) _edata = ABSOLUTE(.); @}
6626 If @samp{ABSOLUTE} were not used, @samp{_edata} would be relative to the
6627 @samp{.data} section.
6629 Using @code{LOADADDR} also forces an expression absolute, since this
6630 particular builtin function returns an absolute address.
6632 @node Builtin Functions
6633 @subsection Builtin Functions
6634 @cindex functions in expressions
6635 The linker script language includes a number of builtin functions for
6636 use in linker script expressions.
6639 @item ABSOLUTE(@var{exp})
6640 @kindex ABSOLUTE(@var{exp})
6641 @cindex expression, absolute
6642 Return the absolute (non-relocatable, as opposed to non-negative) value
6643 of the expression @var{exp}. Primarily useful to assign an absolute
6644 value to a symbol within a section definition, where symbol values are
6645 normally section relative. @xref{Expression Section}.
6647 @item ADDR(@var{section})
6648 @kindex ADDR(@var{section})
6649 @cindex section address in expression
6650 Return the address (VMA) of the named @var{section}. Your
6651 script must previously have defined the location of that section. In
6652 the following example, @code{start_of_output_1}, @code{symbol_1} and
6653 @code{symbol_2} are assigned equivalent values, except that
6654 @code{symbol_1} will be relative to the @code{.output1} section while
6655 the other two will be absolute:
6661 start_of_output_1 = ABSOLUTE(.);
6666 symbol_1 = ADDR(.output1);
6667 symbol_2 = start_of_output_1;
6673 @item ALIGN(@var{align})
6674 @itemx ALIGN(@var{exp},@var{align})
6675 @kindex ALIGN(@var{align})
6676 @kindex ALIGN(@var{exp},@var{align})
6677 @cindex round up location counter
6678 @cindex align location counter
6679 @cindex round up expression
6680 @cindex align expression
6681 Return the location counter (@code{.}) or arbitrary expression aligned
6682 to the next @var{align} boundary. The single operand @code{ALIGN}
6683 doesn't change the value of the location counter---it just does
6684 arithmetic on it. The two operand @code{ALIGN} allows an arbitrary
6685 expression to be aligned upwards (@code{ALIGN(@var{align})} is
6686 equivalent to @code{ALIGN(ABSOLUTE(.), @var{align})}).
6688 Here is an example which aligns the output @code{.data} section to the
6689 next @code{0x2000} byte boundary after the preceding section and sets a
6690 variable within the section to the next @code{0x8000} boundary after the
6695 .data ALIGN(0x2000): @{
6697 variable = ALIGN(0x8000);
6703 The first use of @code{ALIGN} in this example specifies the location of
6704 a section because it is used as the optional @var{address} attribute of
6705 a section definition (@pxref{Output Section Address}). The second use
6706 of @code{ALIGN} is used to defines the value of a symbol.
6708 The builtin function @code{NEXT} is closely related to @code{ALIGN}.
6710 @item ALIGNOF(@var{section})
6711 @kindex ALIGNOF(@var{section})
6712 @cindex section alignment
6713 Return the alignment in bytes of the named @var{section}, if that section has
6714 been allocated. If the section has not been allocated when this is
6715 evaluated, the linker will report an error. In the following example,
6716 the alignment of the @code{.output} section is stored as the first
6717 value in that section.
6722 LONG (ALIGNOF (.output))
6729 @item BLOCK(@var{exp})
6730 @kindex BLOCK(@var{exp})
6731 This is a synonym for @code{ALIGN}, for compatibility with older linker
6732 scripts. It is most often seen when setting the address of an output
6735 @item DATA_SEGMENT_ALIGN(@var{maxpagesize}, @var{commonpagesize})
6736 @kindex DATA_SEGMENT_ALIGN(@var{maxpagesize}, @var{commonpagesize})
6737 This is equivalent to either
6739 (ALIGN(@var{maxpagesize}) + (. & (@var{maxpagesize} - 1)))
6743 (ALIGN(@var{maxpagesize})
6744 + ((. + @var{commonpagesize} - 1) & (@var{maxpagesize} - @var{commonpagesize})))
6747 depending on whether the latter uses fewer @var{commonpagesize} sized pages
6748 for the data segment (area between the result of this expression and
6749 @code{DATA_SEGMENT_END}) than the former or not.
6750 If the latter form is used, it means @var{commonpagesize} bytes of runtime
6751 memory will be saved at the expense of up to @var{commonpagesize} wasted
6752 bytes in the on-disk file.
6754 This expression can only be used directly in @code{SECTIONS} commands, not in
6755 any output section descriptions and only once in the linker script.
6756 @var{commonpagesize} should be less or equal to @var{maxpagesize} and should
6757 be the system page size the object wants to be optimized for while still
6758 running on system page sizes up to @var{maxpagesize}. Note however
6759 that @samp{-z relro} protection will not be effective if the system
6760 page size is larger than @var{commonpagesize}.
6765 . = DATA_SEGMENT_ALIGN(0x10000, 0x2000);
6768 @item DATA_SEGMENT_END(@var{exp})
6769 @kindex DATA_SEGMENT_END(@var{exp})
6770 This defines the end of data segment for @code{DATA_SEGMENT_ALIGN}
6771 evaluation purposes.
6774 . = DATA_SEGMENT_END(.);
6777 @item DATA_SEGMENT_RELRO_END(@var{offset}, @var{exp})
6778 @kindex DATA_SEGMENT_RELRO_END(@var{offset}, @var{exp})
6779 This defines the end of the @code{PT_GNU_RELRO} segment when
6780 @samp{-z relro} option is used.
6781 When @samp{-z relro} option is not present, @code{DATA_SEGMENT_RELRO_END}
6782 does nothing, otherwise @code{DATA_SEGMENT_ALIGN} is padded so that
6783 @var{exp} + @var{offset} is aligned to the @var{commonpagesize}
6784 argument given to @code{DATA_SEGMENT_ALIGN}. If present in the linker
6785 script, it must be placed between @code{DATA_SEGMENT_ALIGN} and
6786 @code{DATA_SEGMENT_END}. Evaluates to the second argument plus any
6787 padding needed at the end of the @code{PT_GNU_RELRO} segment due to
6791 . = DATA_SEGMENT_RELRO_END(24, .);
6794 @item DEFINED(@var{symbol})
6795 @kindex DEFINED(@var{symbol})
6796 @cindex symbol defaults
6797 Return 1 if @var{symbol} is in the linker global symbol table and is
6798 defined before the statement using DEFINED in the script, otherwise
6799 return 0. You can use this function to provide
6800 default values for symbols. For example, the following script fragment
6801 shows how to set a global symbol @samp{begin} to the first location in
6802 the @samp{.text} section---but if a symbol called @samp{begin} already
6803 existed, its value is preserved:
6809 begin = DEFINED(begin) ? begin : . ;
6817 @item LENGTH(@var{memory})
6818 @kindex LENGTH(@var{memory})
6819 Return the length of the memory region named @var{memory}.
6821 @item LOADADDR(@var{section})
6822 @kindex LOADADDR(@var{section})
6823 @cindex section load address in expression
6824 Return the absolute LMA of the named @var{section}. (@pxref{Output
6827 @item LOG2CEIL(@var{exp})
6828 @kindex LOG2CEIL(@var{exp})
6829 Return the binary logarithm of @var{exp} rounded towards infinity.
6830 @code{LOG2CEIL(0)} returns 0.
6833 @item MAX(@var{exp1}, @var{exp2})
6834 Returns the maximum of @var{exp1} and @var{exp2}.
6837 @item MIN(@var{exp1}, @var{exp2})
6838 Returns the minimum of @var{exp1} and @var{exp2}.
6840 @item NEXT(@var{exp})
6841 @kindex NEXT(@var{exp})
6842 @cindex unallocated address, next
6843 Return the next unallocated address that is a multiple of @var{exp}.
6844 This function is closely related to @code{ALIGN(@var{exp})}; unless you
6845 use the @code{MEMORY} command to define discontinuous memory for the
6846 output file, the two functions are equivalent.
6848 @item ORIGIN(@var{memory})
6849 @kindex ORIGIN(@var{memory})
6850 Return the origin of the memory region named @var{memory}.
6852 @item SEGMENT_START(@var{segment}, @var{default})
6853 @kindex SEGMENT_START(@var{segment}, @var{default})
6854 Return the base address of the named @var{segment}. If an explicit
6855 value has already been given for this segment (with a command-line
6856 @samp{-T} option) then that value will be returned otherwise the value
6857 will be @var{default}. At present, the @samp{-T} command-line option
6858 can only be used to set the base address for the ``text'', ``data'', and
6859 ``bss'' sections, but you can use @code{SEGMENT_START} with any segment
6862 @item SIZEOF(@var{section})
6863 @kindex SIZEOF(@var{section})
6864 @cindex section size
6865 Return the size in bytes of the named @var{section}, if that section has
6866 been allocated. If the section has not been allocated when this is
6867 evaluated, the linker will report an error. In the following example,
6868 @code{symbol_1} and @code{symbol_2} are assigned identical values:
6877 symbol_1 = .end - .start ;
6878 symbol_2 = SIZEOF(.output);
6883 @item SIZEOF_HEADERS
6884 @itemx sizeof_headers
6885 @kindex SIZEOF_HEADERS
6887 Return the size in bytes of the output file's headers. This is
6888 information which appears at the start of the output file. You can use
6889 this number when setting the start address of the first section, if you
6890 choose, to facilitate paging.
6892 @cindex not enough room for program headers
6893 @cindex program headers, not enough room
6894 When producing an ELF output file, if the linker script uses the
6895 @code{SIZEOF_HEADERS} builtin function, the linker must compute the
6896 number of program headers before it has determined all the section
6897 addresses and sizes. If the linker later discovers that it needs
6898 additional program headers, it will report an error @samp{not enough
6899 room for program headers}. To avoid this error, you must avoid using
6900 the @code{SIZEOF_HEADERS} function, or you must rework your linker
6901 script to avoid forcing the linker to use additional program headers, or
6902 you must define the program headers yourself using the @code{PHDRS}
6903 command (@pxref{PHDRS}).
6906 @node Implicit Linker Scripts
6907 @section Implicit Linker Scripts
6908 @cindex implicit linker scripts
6909 If you specify a linker input file which the linker can not recognize as
6910 an object file or an archive file, it will try to read the file as a
6911 linker script. If the file can not be parsed as a linker script, the
6912 linker will report an error.
6914 An implicit linker script will not replace the default linker script.
6916 Typically an implicit linker script would contain only symbol
6917 assignments, or the @code{INPUT}, @code{GROUP}, or @code{VERSION}
6920 Any input files read because of an implicit linker script will be read
6921 at the position in the command line where the implicit linker script was
6922 read. This can affect archive searching.
6925 @node Machine Dependent
6926 @chapter Machine Dependent Features
6928 @cindex machine dependencies
6929 @command{ld} has additional features on some platforms; the following
6930 sections describe them. Machines where @command{ld} has no additional
6931 functionality are not listed.
6935 * H8/300:: @command{ld} and the H8/300
6938 * M68HC11/68HC12:: @code{ld} and the Motorola 68HC11 and 68HC12 families
6941 * ARM:: @command{ld} and the ARM family
6944 * HPPA ELF32:: @command{ld} and HPPA 32-bit ELF
6947 * M68K:: @command{ld} and the Motorola 68K family
6950 * MIPS:: @command{ld} and the MIPS family
6953 * MMIX:: @command{ld} and MMIX
6956 * MSP430:: @command{ld} and MSP430
6959 * NDS32:: @command{ld} and NDS32
6962 * Nios II:: @command{ld} and the Altera Nios II
6965 * PowerPC ELF32:: @command{ld} and PowerPC 32-bit ELF Support
6968 * PowerPC64 ELF64:: @command{ld} and PowerPC64 64-bit ELF Support
6971 * S/390 ELF:: @command{ld} and S/390 ELF Support
6974 * SPU ELF:: @command{ld} and SPU ELF Support
6977 * TI COFF:: @command{ld} and TI COFF
6980 * WIN32:: @command{ld} and WIN32 (cygwin/mingw)
6983 * Xtensa:: @command{ld} and Xtensa Processors
6994 @section @command{ld} and the H8/300
6996 @cindex H8/300 support
6997 For the H8/300, @command{ld} can perform these global optimizations when
6998 you specify the @samp{--relax} command-line option.
7001 @cindex relaxing on H8/300
7002 @item relaxing address modes
7003 @command{ld} finds all @code{jsr} and @code{jmp} instructions whose
7004 targets are within eight bits, and turns them into eight-bit
7005 program-counter relative @code{bsr} and @code{bra} instructions,
7008 @cindex synthesizing on H8/300
7009 @item synthesizing instructions
7010 @c FIXME: specifically mov.b, or any mov instructions really? -> mov.b only, at least on H8, H8H, H8S
7011 @command{ld} finds all @code{mov.b} instructions which use the
7012 sixteen-bit absolute address form, but refer to the top
7013 page of memory, and changes them to use the eight-bit address form.
7014 (That is: the linker turns @samp{mov.b @code{@@}@var{aa}:16} into
7015 @samp{mov.b @code{@@}@var{aa}:8} whenever the address @var{aa} is in the
7016 top page of memory).
7018 @command{ld} finds all @code{mov} instructions which use the register
7019 indirect with 32-bit displacement addressing mode, but use a small
7020 displacement inside 16-bit displacement range, and changes them to use
7021 the 16-bit displacement form. (That is: the linker turns @samp{mov.b
7022 @code{@@}@var{d}:32,ERx} into @samp{mov.b @code{@@}@var{d}:16,ERx}
7023 whenever the displacement @var{d} is in the 16 bit signed integer
7024 range. Only implemented in ELF-format ld).
7026 @item bit manipulation instructions
7027 @command{ld} finds all bit manipulation instructions like @code{band, bclr,
7028 biand, bild, bior, bist, bixor, bld, bnot, bor, bset, bst, btst, bxor}
7029 which use 32 bit and 16 bit absolute address form, but refer to the top
7030 page of memory, and changes them to use the 8 bit address form.
7031 (That is: the linker turns @samp{bset #xx:3,@code{@@}@var{aa}:32} into
7032 @samp{bset #xx:3,@code{@@}@var{aa}:8} whenever the address @var{aa} is in
7033 the top page of memory).
7035 @item system control instructions
7036 @command{ld} finds all @code{ldc.w, stc.w} instructions which use the
7037 32 bit absolute address form, but refer to the top page of memory, and
7038 changes them to use 16 bit address form.
7039 (That is: the linker turns @samp{ldc.w @code{@@}@var{aa}:32,ccr} into
7040 @samp{ldc.w @code{@@}@var{aa}:16,ccr} whenever the address @var{aa} is in
7041 the top page of memory).
7051 @c This stuff is pointless to say unless you're especially concerned
7052 @c with Renesas chips; don't enable it for generic case, please.
7054 @chapter @command{ld} and Other Renesas Chips
7056 @command{ld} also supports the Renesas (formerly Hitachi) H8/300H,
7057 H8/500, and SH chips. No special features, commands, or command-line
7058 options are required for these chips.
7072 @node M68HC11/68HC12
7073 @section @command{ld} and the Motorola 68HC11 and 68HC12 families
7075 @cindex M68HC11 and 68HC12 support
7077 @subsection Linker Relaxation
7079 For the Motorola 68HC11, @command{ld} can perform these global
7080 optimizations when you specify the @samp{--relax} command-line option.
7083 @cindex relaxing on M68HC11
7084 @item relaxing address modes
7085 @command{ld} finds all @code{jsr} and @code{jmp} instructions whose
7086 targets are within eight bits, and turns them into eight-bit
7087 program-counter relative @code{bsr} and @code{bra} instructions,
7090 @command{ld} also looks at all 16-bit extended addressing modes and
7091 transforms them in a direct addressing mode when the address is in
7092 page 0 (between 0 and 0x0ff).
7094 @item relaxing gcc instruction group
7095 When @command{gcc} is called with @option{-mrelax}, it can emit group
7096 of instructions that the linker can optimize to use a 68HC11 direct
7097 addressing mode. These instructions consists of @code{bclr} or
7098 @code{bset} instructions.
7102 @subsection Trampoline Generation
7104 @cindex trampoline generation on M68HC11
7105 @cindex trampoline generation on M68HC12
7106 For 68HC11 and 68HC12, @command{ld} can generate trampoline code to
7107 call a far function using a normal @code{jsr} instruction. The linker
7108 will also change the relocation to some far function to use the
7109 trampoline address instead of the function address. This is typically the
7110 case when a pointer to a function is taken. The pointer will in fact
7111 point to the function trampoline.
7119 @section @command{ld} and the ARM family
7121 @cindex ARM interworking support
7122 @kindex --support-old-code
7123 For the ARM, @command{ld} will generate code stubs to allow functions calls
7124 between ARM and Thumb code. These stubs only work with code that has
7125 been compiled and assembled with the @samp{-mthumb-interwork} command
7126 line option. If it is necessary to link with old ARM object files or
7127 libraries, which have not been compiled with the -mthumb-interwork
7128 option then the @samp{--support-old-code} command-line switch should be
7129 given to the linker. This will make it generate larger stub functions
7130 which will work with non-interworking aware ARM code. Note, however,
7131 the linker does not support generating stubs for function calls to
7132 non-interworking aware Thumb code.
7134 @cindex thumb entry point
7135 @cindex entry point, thumb
7136 @kindex --thumb-entry=@var{entry}
7137 The @samp{--thumb-entry} switch is a duplicate of the generic
7138 @samp{--entry} switch, in that it sets the program's starting address.
7139 But it also sets the bottom bit of the address, so that it can be
7140 branched to using a BX instruction, and the program will start
7141 executing in Thumb mode straight away.
7143 @cindex PE import table prefixing
7144 @kindex --use-nul-prefixed-import-tables
7145 The @samp{--use-nul-prefixed-import-tables} switch is specifying, that
7146 the import tables idata4 and idata5 have to be generated with a zero
7147 element prefix for import libraries. This is the old style to generate
7148 import tables. By default this option is turned off.
7152 The @samp{--be8} switch instructs @command{ld} to generate BE8 format
7153 executables. This option is only valid when linking big-endian
7154 objects - ie ones which have been assembled with the @option{-EB}
7155 option. The resulting image will contain big-endian data and
7159 @kindex --target1-rel
7160 @kindex --target1-abs
7161 The @samp{R_ARM_TARGET1} relocation is typically used for entries in the
7162 @samp{.init_array} section. It is interpreted as either @samp{R_ARM_REL32}
7163 or @samp{R_ARM_ABS32}, depending on the target. The @samp{--target1-rel}
7164 and @samp{--target1-abs} switches override the default.
7167 @kindex --target2=@var{type}
7168 The @samp{--target2=type} switch overrides the default definition of the
7169 @samp{R_ARM_TARGET2} relocation. Valid values for @samp{type}, their
7170 meanings, and target defaults are as follows:
7173 @samp{R_ARM_REL32} (arm*-*-elf, arm*-*-eabi)
7175 @samp{R_ARM_ABS32} (arm*-*-symbianelf)
7177 @samp{R_ARM_GOT_PREL} (arm*-*-linux, arm*-*-*bsd)
7182 The @samp{R_ARM_V4BX} relocation (defined by the ARM AAELF
7183 specification) enables objects compiled for the ARMv4 architecture to be
7184 interworking-safe when linked with other objects compiled for ARMv4t, but
7185 also allows pure ARMv4 binaries to be built from the same ARMv4 objects.
7187 In the latter case, the switch @option{--fix-v4bx} must be passed to the
7188 linker, which causes v4t @code{BX rM} instructions to be rewritten as
7189 @code{MOV PC,rM}, since v4 processors do not have a @code{BX} instruction.
7191 In the former case, the switch should not be used, and @samp{R_ARM_V4BX}
7192 relocations are ignored.
7194 @cindex FIX_V4BX_INTERWORKING
7195 @kindex --fix-v4bx-interworking
7196 Replace @code{BX rM} instructions identified by @samp{R_ARM_V4BX}
7197 relocations with a branch to the following veneer:
7205 This allows generation of libraries/applications that work on ARMv4 cores
7206 and are still interworking safe. Note that the above veneer clobbers the
7207 condition flags, so may cause incorrect program behavior in rare cases.
7211 The @samp{--use-blx} switch enables the linker to use ARM/Thumb
7212 BLX instructions (available on ARMv5t and above) in various
7213 situations. Currently it is used to perform calls via the PLT from Thumb
7214 code using BLX rather than using BX and a mode-switching stub before
7215 each PLT entry. This should lead to such calls executing slightly faster.
7217 This option is enabled implicitly for SymbianOS, so there is no need to
7218 specify it if you are using that target.
7220 @cindex VFP11_DENORM_FIX
7221 @kindex --vfp11-denorm-fix
7222 The @samp{--vfp11-denorm-fix} switch enables a link-time workaround for a
7223 bug in certain VFP11 coprocessor hardware, which sometimes allows
7224 instructions with denorm operands (which must be handled by support code)
7225 to have those operands overwritten by subsequent instructions before
7226 the support code can read the intended values.
7228 The bug may be avoided in scalar mode if you allow at least one
7229 intervening instruction between a VFP11 instruction which uses a register
7230 and another instruction which writes to the same register, or at least two
7231 intervening instructions if vector mode is in use. The bug only affects
7232 full-compliance floating-point mode: you do not need this workaround if
7233 you are using "runfast" mode. Please contact ARM for further details.
7235 If you know you are using buggy VFP11 hardware, you can
7236 enable this workaround by specifying the linker option
7237 @samp{--vfp-denorm-fix=scalar} if you are using the VFP11 scalar
7238 mode only, or @samp{--vfp-denorm-fix=vector} if you are using
7239 vector mode (the latter also works for scalar code). The default is
7240 @samp{--vfp-denorm-fix=none}.
7242 If the workaround is enabled, instructions are scanned for
7243 potentially-troublesome sequences, and a veneer is created for each
7244 such sequence which may trigger the erratum. The veneer consists of the
7245 first instruction of the sequence and a branch back to the subsequent
7246 instruction. The original instruction is then replaced with a branch to
7247 the veneer. The extra cycles required to call and return from the veneer
7248 are sufficient to avoid the erratum in both the scalar and vector cases.
7250 @cindex ARM1176 erratum workaround
7251 @kindex --fix-arm1176
7252 @kindex --no-fix-arm1176
7253 The @samp{--fix-arm1176} switch enables a link-time workaround for an erratum
7254 in certain ARM1176 processors. The workaround is enabled by default if you
7255 are targeting ARM v6 (excluding ARM v6T2) or earlier. It can be disabled
7256 unconditionally by specifying @samp{--no-fix-arm1176}.
7258 Further information is available in the ``ARM1176JZ-S and ARM1176JZF-S
7259 Programmer Advice Notice'' available on the ARM documentation website at:
7260 http://infocenter.arm.com/.
7262 @cindex STM32L4xx erratum workaround
7263 @kindex --fix-stm32l4xx-629360
7265 The @samp{--fix-stm32l4xx-629360} switch enables a link-time
7266 workaround for a bug in the bus matrix / memory controller for some of
7267 the STM32 Cortex-M4 based products (STM32L4xx). When accessing
7268 off-chip memory via the affected bus for bus reads of 9 words or more,
7269 the bus can generate corrupt data and/or abort. These are only
7270 core-initiated accesses (not DMA), and might affect any access:
7271 integer loads such as LDM, POP and floating-point loads such as VLDM,
7272 VPOP. Stores are not affected.
7274 The bug can be avoided by splitting memory accesses into the
7275 necessary chunks to keep bus reads below 8 words.
7277 The workaround is not enabled by default, this is equivalent to use
7278 @samp{--fix-stm32l4xx-629360=none}. If you know you are using buggy
7279 STM32L4xx hardware, you can enable the workaround by specifying the
7280 linker option @samp{--fix-stm32l4xx-629360}, or the equivalent
7281 @samp{--fix-stm32l4xx-629360=default}.
7283 If the workaround is enabled, instructions are scanned for
7284 potentially-troublesome sequences, and a veneer is created for each
7285 such sequence which may trigger the erratum. The veneer consists in a
7286 replacement sequence emulating the behaviour of the original one and a
7287 branch back to the subsequent instruction. The original instruction is
7288 then replaced with a branch to the veneer.
7290 The workaround does not always preserve the memory access order for
7291 the LDMDB instruction, when the instruction loads the PC.
7293 The workaround is not able to handle problematic instructions when
7294 they are in the middle of an IT block, since a branch is not allowed
7295 there. In that case, the linker reports a warning and no replacement
7298 The workaround is not able to replace problematic instructions with a
7299 PC-relative branch instruction if the @samp{.text} section is too
7300 large. In that case, when the branch that replaces the original code
7301 cannot be encoded, the linker reports a warning and no replacement
7304 @cindex NO_ENUM_SIZE_WARNING
7305 @kindex --no-enum-size-warning
7306 The @option{--no-enum-size-warning} switch prevents the linker from
7307 warning when linking object files that specify incompatible EABI
7308 enumeration size attributes. For example, with this switch enabled,
7309 linking of an object file using 32-bit enumeration values with another
7310 using enumeration values fitted into the smallest possible space will
7313 @cindex NO_WCHAR_SIZE_WARNING
7314 @kindex --no-wchar-size-warning
7315 The @option{--no-wchar-size-warning} switch prevents the linker from
7316 warning when linking object files that specify incompatible EABI
7317 @code{wchar_t} size attributes. For example, with this switch enabled,
7318 linking of an object file using 32-bit @code{wchar_t} values with another
7319 using 16-bit @code{wchar_t} values will not be diagnosed.
7322 @kindex --pic-veneer
7323 The @samp{--pic-veneer} switch makes the linker use PIC sequences for
7324 ARM/Thumb interworking veneers, even if the rest of the binary
7325 is not PIC. This avoids problems on uClinux targets where
7326 @samp{--emit-relocs} is used to generate relocatable binaries.
7328 @cindex STUB_GROUP_SIZE
7329 @kindex --stub-group-size=@var{N}
7330 The linker will automatically generate and insert small sequences of
7331 code into a linked ARM ELF executable whenever an attempt is made to
7332 perform a function call to a symbol that is too far away. The
7333 placement of these sequences of instructions - called stubs - is
7334 controlled by the command-line option @option{--stub-group-size=N}.
7335 The placement is important because a poor choice can create a need for
7336 duplicate stubs, increasing the code size. The linker will try to
7337 group stubs together in order to reduce interruptions to the flow of
7338 code, but it needs guidance as to how big these groups should be and
7339 where they should be placed.
7341 The value of @samp{N}, the parameter to the
7342 @option{--stub-group-size=} option controls where the stub groups are
7343 placed. If it is negative then all stubs are placed after the first
7344 branch that needs them. If it is positive then the stubs can be
7345 placed either before or after the branches that need them. If the
7346 value of @samp{N} is 1 (either +1 or -1) then the linker will choose
7347 exactly where to place groups of stubs, using its built in heuristics.
7348 A value of @samp{N} greater than 1 (or smaller than -1) tells the
7349 linker that a single group of stubs can service at most @samp{N} bytes
7350 from the input sections.
7352 The default, if @option{--stub-group-size=} is not specified, is
7355 Farcalls stubs insertion is fully supported for the ARM-EABI target
7356 only, because it relies on object files properties not present
7359 @cindex Cortex-A8 erratum workaround
7360 @kindex --fix-cortex-a8
7361 @kindex --no-fix-cortex-a8
7362 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}.
7364 The erratum only affects Thumb-2 code. Please contact ARM for further details.
7366 @cindex Cortex-A53 erratum 835769 workaround
7367 @kindex --fix-cortex-a53-835769
7368 @kindex --no-fix-cortex-a53-835769
7369 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}.
7371 Please contact ARM for further details.
7373 @kindex --merge-exidx-entries
7374 @kindex --no-merge-exidx-entries
7375 @cindex Merging exidx entries
7376 The @samp{--no-merge-exidx-entries} switch disables the merging of adjacent exidx entries in debuginfo.
7379 @cindex 32-bit PLT entries
7380 The @samp{--long-plt} option enables the use of 16 byte PLT entries
7381 which support up to 4Gb of code. The default is to use 12 byte PLT
7382 entries which only support 512Mb of code.
7384 @kindex --no-apply-dynamic-relocs
7385 @cindex AArch64 rela addend
7386 The @samp{--no-apply-dynamic-relocs} option makes AArch64 linker do not apply
7387 link-time values for dynamic relocations.
7389 @cindex Placement of SG veneers
7390 All SG veneers are placed in the special output section @code{.gnu.sgstubs}.
7391 Its start address must be set, either with the command-line option
7392 @samp{--section-start} or in a linker script, to indicate where to place these
7395 @kindex --cmse-implib
7396 @cindex Secure gateway import library
7397 The @samp{--cmse-implib} option requests that the import libraries
7398 specified by the @samp{--out-implib} and @samp{--in-implib} options are
7399 secure gateway import libraries, suitable for linking a non-secure
7400 executable against secure code as per ARMv8-M Security Extensions.
7402 @kindex --in-implib=@var{file}
7403 @cindex Input import library
7404 The @samp{--in-implib=file} specifies an input import library whose symbols
7405 must keep the same address in the executable being produced. A warning is
7406 given if no @samp{--out-implib} is given but new symbols have been introduced
7407 in the executable that should be listed in its import library. Otherwise, if
7408 @samp{--out-implib} is specified, the symbols are added to the output import
7409 library. A warning is also given if some symbols present in the input import
7410 library have disappeared from the executable. This option is only effective
7411 for Secure Gateway import libraries, ie. when @samp{--cmse-implib} is
7425 @section @command{ld} and HPPA 32-bit ELF Support
7426 @cindex HPPA multiple sub-space stubs
7427 @kindex --multi-subspace
7428 When generating a shared library, @command{ld} will by default generate
7429 import stubs suitable for use with a single sub-space application.
7430 The @samp{--multi-subspace} switch causes @command{ld} to generate export
7431 stubs, and different (larger) import stubs suitable for use with
7432 multiple sub-spaces.
7434 @cindex HPPA stub grouping
7435 @kindex --stub-group-size=@var{N}
7436 Long branch stubs and import/export stubs are placed by @command{ld} in
7437 stub sections located between groups of input sections.
7438 @samp{--stub-group-size} specifies the maximum size of a group of input
7439 sections handled by one stub section. Since branch offsets are signed,
7440 a stub section may serve two groups of input sections, one group before
7441 the stub section, and one group after it. However, when using
7442 conditional branches that require stubs, it may be better (for branch
7443 prediction) that stub sections only serve one group of input sections.
7444 A negative value for @samp{N} chooses this scheme, ensuring that
7445 branches to stubs always use a negative offset. Two special values of
7446 @samp{N} are recognized, @samp{1} and @samp{-1}. These both instruct
7447 @command{ld} to automatically size input section groups for the branch types
7448 detected, with the same behaviour regarding stub placement as other
7449 positive or negative values of @samp{N} respectively.
7451 Note that @samp{--stub-group-size} does not split input sections. A
7452 single input section larger than the group size specified will of course
7453 create a larger group (of one section). If input sections are too
7454 large, it may not be possible for a branch to reach its stub.
7467 @section @command{ld} and the Motorola 68K family
7469 @cindex Motorola 68K GOT generation
7470 @kindex --got=@var{type}
7471 The @samp{--got=@var{type}} option lets you choose the GOT generation scheme.
7472 The choices are @samp{single}, @samp{negative}, @samp{multigot} and
7473 @samp{target}. When @samp{target} is selected the linker chooses
7474 the default GOT generation scheme for the current target.
7475 @samp{single} tells the linker to generate a single GOT with
7476 entries only at non-negative offsets.
7477 @samp{negative} instructs the linker to generate a single GOT with
7478 entries at both negative and positive offsets. Not all environments
7480 @samp{multigot} allows the linker to generate several GOTs in the
7481 output file. All GOT references from a single input object
7482 file access the same GOT, but references from different input object
7483 files might access different GOTs. Not all environments support such GOTs.
7496 @section @command{ld} and the MIPS family
7498 @cindex MIPS microMIPS instruction choice selection
7501 The @samp{--insn32} and @samp{--no-insn32} options control the choice of
7502 microMIPS instructions used in code generated by the linker, such as that
7503 in the PLT or lazy binding stubs, or in relaxation. If @samp{--insn32} is
7504 used, then the linker only uses 32-bit instruction encodings. By default
7505 or if @samp{--no-insn32} is used, all instruction encodings are used,
7506 including 16-bit ones where possible.
7508 @cindex MIPS branch relocation check control
7509 @kindex --ignore-branch-isa
7510 @kindex --no-ignore-branch-isa
7511 The @samp{--ignore-branch-isa} and @samp{--no-ignore-branch-isa} options
7512 control branch relocation checks for invalid ISA mode transitions. If
7513 @samp{--ignore-branch-isa} is used, then the linker accepts any branch
7514 relocations and any ISA mode transition required is lost in relocation
7515 calculation, except for some cases of @code{BAL} instructions which meet
7516 relaxation conditions and are converted to equivalent @code{JALX}
7517 instructions as the associated relocation is calculated. By default
7518 or if @samp{--no-ignore-branch-isa} is used a check is made causing
7519 the loss of an ISA mode transition to produce an error.
7532 @section @code{ld} and MMIX
7533 For MMIX, there is a choice of generating @code{ELF} object files or
7534 @code{mmo} object files when linking. The simulator @code{mmix}
7535 understands the @code{mmo} format. The binutils @code{objcopy} utility
7536 can translate between the two formats.
7538 There is one special section, the @samp{.MMIX.reg_contents} section.
7539 Contents in this section is assumed to correspond to that of global
7540 registers, and symbols referring to it are translated to special symbols,
7541 equal to registers. In a final link, the start address of the
7542 @samp{.MMIX.reg_contents} section corresponds to the first allocated
7543 global register multiplied by 8. Register @code{$255} is not included in
7544 this section; it is always set to the program entry, which is at the
7545 symbol @code{Main} for @code{mmo} files.
7547 Global symbols with the prefix @code{__.MMIX.start.}, for example
7548 @code{__.MMIX.start..text} and @code{__.MMIX.start..data} are special.
7549 The default linker script uses these to set the default start address
7552 Initial and trailing multiples of zero-valued 32-bit words in a section,
7553 are left out from an mmo file.
7566 @section @code{ld} and MSP430
7567 For the MSP430 it is possible to select the MPU architecture. The flag @samp{-m [mpu type]}
7568 will select an appropriate linker script for selected MPU type. (To get a list of known MPUs
7569 just pass @samp{-m help} option to the linker).
7571 @cindex MSP430 extra sections
7572 The linker will recognize some extra sections which are MSP430 specific:
7575 @item @samp{.vectors}
7576 Defines a portion of ROM where interrupt vectors located.
7578 @item @samp{.bootloader}
7579 Defines the bootloader portion of the ROM (if applicable). Any code
7580 in this section will be uploaded to the MPU.
7582 @item @samp{.infomem}
7583 Defines an information memory section (if applicable). Any code in
7584 this section will be uploaded to the MPU.
7586 @item @samp{.infomemnobits}
7587 This is the same as the @samp{.infomem} section except that any code
7588 in this section will not be uploaded to the MPU.
7590 @item @samp{.noinit}
7591 Denotes a portion of RAM located above @samp{.bss} section.
7593 The last two sections are used by gcc.
7597 @cindex MSP430 Options
7598 @kindex --code-region
7599 @item --code-region=[either,lower,upper,none]
7600 This will transform .text* sections to [either,lower,upper].text* sections. The
7601 argument passed to GCC for -mcode-region is propagated to the linker
7604 @kindex --data-region
7605 @item --data-region=[either,lower,upper,none]
7606 This will transform .data*, .bss* and .rodata* sections to
7607 [either,lower,upper].[data,bss,rodata]* sections. The argument passed to GCC
7608 for -mdata-region is propagated to the linker using this option.
7610 @kindex --disable-sec-transformation
7611 @item --disable-sec-transformation
7612 Prevent the transformation of sections as specified by the @code{--code-region}
7613 and @code{--data-region} options.
7614 This is useful if you are compiling and linking using a single call to the GCC
7615 wrapper, and want to compile the source files using -m[code,data]-region but
7616 not transform the sections for prebuilt libraries and objects.
7630 @section @code{ld} and NDS32
7631 @kindex relaxing on NDS32
7632 For NDS32, there are some options to select relaxation behavior. The linker
7633 relaxes objects according to these options.
7636 @item @samp{--m[no-]fp-as-gp}
7637 Disable/enable fp-as-gp relaxation.
7639 @item @samp{--mexport-symbols=FILE}
7640 Exporting symbols and their address into FILE as linker script.
7642 @item @samp{--m[no-]ex9}
7643 Disable/enable link-time EX9 relaxation.
7645 @item @samp{--mexport-ex9=FILE}
7646 Export the EX9 table after linking.
7648 @item @samp{--mimport-ex9=FILE}
7649 Import the Ex9 table for EX9 relaxation.
7651 @item @samp{--mupdate-ex9}
7652 Update the existing EX9 table.
7654 @item @samp{--mex9-limit=NUM}
7655 Maximum number of entries in the ex9 table.
7657 @item @samp{--mex9-loop-aware}
7658 Avoid generating the EX9 instruction inside the loop.
7660 @item @samp{--m[no-]ifc}
7661 Disable/enable the link-time IFC optimization.
7663 @item @samp{--mifc-loop-aware}
7664 Avoid generating the IFC instruction inside the loop.
7678 @section @command{ld} and the Altera Nios II
7679 @cindex Nios II call relaxation
7680 @kindex --relax on Nios II
7682 Call and immediate jump instructions on Nios II processors are limited to
7683 transferring control to addresses in the same 256MB memory segment,
7684 which may result in @command{ld} giving
7685 @samp{relocation truncated to fit} errors with very large programs.
7686 The command-line option @option{--relax} enables the generation of
7687 trampolines that can access the entire 32-bit address space for calls
7688 outside the normal @code{call} and @code{jmpi} address range. These
7689 trampolines are inserted at section boundaries, so may not themselves
7690 be reachable if an input section and its associated call trampolines are
7693 The @option{--relax} option is enabled by default unless @option{-r}
7694 is also specified. You can disable trampoline generation by using the
7695 @option{--no-relax} linker option. You can also disable this optimization
7696 locally by using the @samp{set .noat} directive in assembly-language
7697 source files, as the linker-inserted trampolines use the @code{at}
7698 register as a temporary.
7700 Note that the linker @option{--relax} option is independent of assembler
7701 relaxation options, and that using the GNU assembler's @option{-relax-all}
7702 option interferes with the linker's more selective call instruction relaxation.
7715 @section @command{ld} and PowerPC 32-bit ELF Support
7716 @cindex PowerPC long branches
7717 @kindex --relax on PowerPC
7718 Branches on PowerPC processors are limited to a signed 26-bit
7719 displacement, which may result in @command{ld} giving
7720 @samp{relocation truncated to fit} errors with very large programs.
7721 @samp{--relax} enables the generation of trampolines that can access
7722 the entire 32-bit address space. These trampolines are inserted at
7723 section boundaries, so may not themselves be reachable if an input
7724 section exceeds 33M in size. You may combine @samp{-r} and
7725 @samp{--relax} to add trampolines in a partial link. In that case
7726 both branches to undefined symbols and inter-section branches are also
7727 considered potentially out of range, and trampolines inserted.
7729 @cindex PowerPC ELF32 options
7734 Current PowerPC GCC accepts a @samp{-msecure-plt} option that
7735 generates code capable of using a newer PLT and GOT layout that has
7736 the security advantage of no executable section ever needing to be
7737 writable and no writable section ever being executable. PowerPC
7738 @command{ld} will generate this layout, including stubs to access the
7739 PLT, if all input files (including startup and static libraries) were
7740 compiled with @samp{-msecure-plt}. @samp{--bss-plt} forces the old
7741 BSS PLT (and GOT layout) which can give slightly better performance.
7743 @kindex --secure-plt
7745 @command{ld} will use the new PLT and GOT layout if it is linking new
7746 @samp{-fpic} or @samp{-fPIC} code, but does not do so automatically
7747 when linking non-PIC code. This option requests the new PLT and GOT
7748 layout. A warning will be given if some object file requires the old
7754 The new secure PLT and GOT are placed differently relative to other
7755 sections compared to older BSS PLT and GOT placement. The location of
7756 @code{.plt} must change because the new secure PLT is an initialized
7757 section while the old PLT is uninitialized. The reason for the
7758 @code{.got} change is more subtle: The new placement allows
7759 @code{.got} to be read-only in applications linked with
7760 @samp{-z relro -z now}. However, this placement means that
7761 @code{.sdata} cannot always be used in shared libraries, because the
7762 PowerPC ABI accesses @code{.sdata} in shared libraries from the GOT
7763 pointer. @samp{--sdata-got} forces the old GOT placement. PowerPC
7764 GCC doesn't use @code{.sdata} in shared libraries, so this option is
7765 really only useful for other compilers that may do so.
7767 @cindex PowerPC stub symbols
7768 @kindex --emit-stub-syms
7769 @item --emit-stub-syms
7770 This option causes @command{ld} to label linker stubs with a local
7771 symbol that encodes the stub type and destination.
7773 @cindex PowerPC TLS optimization
7774 @kindex --no-tls-optimize
7775 @item --no-tls-optimize
7776 PowerPC @command{ld} normally performs some optimization of code
7777 sequences used to access Thread-Local Storage. Use this option to
7778 disable the optimization.
7791 @node PowerPC64 ELF64
7792 @section @command{ld} and PowerPC64 64-bit ELF Support
7794 @cindex PowerPC64 ELF64 options
7796 @cindex PowerPC64 stub grouping
7797 @kindex --stub-group-size
7798 @item --stub-group-size
7799 Long branch stubs, PLT call stubs and TOC adjusting stubs are placed
7800 by @command{ld} in stub sections located between groups of input sections.
7801 @samp{--stub-group-size} specifies the maximum size of a group of input
7802 sections handled by one stub section. Since branch offsets are signed,
7803 a stub section may serve two groups of input sections, one group before
7804 the stub section, and one group after it. However, when using
7805 conditional branches that require stubs, it may be better (for branch
7806 prediction) that stub sections only serve one group of input sections.
7807 A negative value for @samp{N} chooses this scheme, ensuring that
7808 branches to stubs always use a negative offset. Two special values of
7809 @samp{N} are recognized, @samp{1} and @samp{-1}. These both instruct
7810 @command{ld} to automatically size input section groups for the branch types
7811 detected, with the same behaviour regarding stub placement as other
7812 positive or negative values of @samp{N} respectively.
7814 Note that @samp{--stub-group-size} does not split input sections. A
7815 single input section larger than the group size specified will of course
7816 create a larger group (of one section). If input sections are too
7817 large, it may not be possible for a branch to reach its stub.
7819 @cindex PowerPC64 stub symbols
7820 @kindex --emit-stub-syms
7821 @item --emit-stub-syms
7822 This option causes @command{ld} to label linker stubs with a local
7823 symbol that encodes the stub type and destination.
7825 @cindex PowerPC64 dot symbols
7827 @kindex --no-dotsyms
7830 These two options control how @command{ld} interprets version patterns
7831 in a version script. Older PowerPC64 compilers emitted both a
7832 function descriptor symbol with the same name as the function, and a
7833 code entry symbol with the name prefixed by a dot (@samp{.}). To
7834 properly version a function @samp{foo}, the version script thus needs
7835 to control both @samp{foo} and @samp{.foo}. The option
7836 @samp{--dotsyms}, on by default, automatically adds the required
7837 dot-prefixed patterns. Use @samp{--no-dotsyms} to disable this
7840 @cindex PowerPC64 register save/restore functions
7841 @kindex --save-restore-funcs
7842 @kindex --no-save-restore-funcs
7843 @item --save-restore-funcs
7844 @itemx --no-save-restore-funcs
7845 These two options control whether PowerPC64 @command{ld} automatically
7846 provides out-of-line register save and restore functions used by
7847 @samp{-Os} code. The default is to provide any such referenced
7848 function for a normal final link, and to not do so for a relocatable
7851 @cindex PowerPC64 TLS optimization
7852 @kindex --no-tls-optimize
7853 @item --no-tls-optimize
7854 PowerPC64 @command{ld} normally performs some optimization of code
7855 sequences used to access Thread-Local Storage. Use this option to
7856 disable the optimization.
7858 @cindex PowerPC64 __tls_get_addr optimization
7859 @kindex --tls-get-addr-optimize
7860 @kindex --no-tls-get-addr-optimize
7861 @kindex --tls-get-addr-regsave
7862 @kindex --no-tls-get-addr-regsave
7863 @item --tls-get-addr-optimize
7864 @itemx --no-tls-get-addr-optimize
7865 These options control how PowerPC64 @command{ld} uses a special
7866 stub to call __tls_get_addr. PowerPC64 glibc 2.22 and later support
7867 an optimization that allows the second and subsequent calls to
7868 @code{__tls_get_addr} for a given symbol to be resolved by the special
7869 stub without calling in to glibc. By default the linker enables
7870 generation of the stub when glibc advertises the availability of
7872 Using @option{--tls-get-addr-optimize} with an older glibc won't do
7873 much besides slow down your applications, but may be useful if linking
7874 an application against an older glibc with the expectation that it
7875 will normally be used on systems having a newer glibc.
7876 @option{--tls-get-addr-regsave} forces generation of a stub that saves
7877 and restores volatile registers around the call into glibc. Normally,
7878 this is done when the linker detects a call to __tls_get_addr_desc.
7879 Such calls then go via the register saving stub to __tls_get_addr_opt.
7880 @option{--no-tls-get-addr-regsave} disables generation of the
7883 @cindex PowerPC64 OPD optimization
7884 @kindex --no-opd-optimize
7885 @item --no-opd-optimize
7886 PowerPC64 @command{ld} normally removes @code{.opd} section entries
7887 corresponding to deleted link-once functions, or functions removed by
7888 the action of @samp{--gc-sections} or linker script @code{/DISCARD/}.
7889 Use this option to disable @code{.opd} optimization.
7891 @cindex PowerPC64 OPD spacing
7892 @kindex --non-overlapping-opd
7893 @item --non-overlapping-opd
7894 Some PowerPC64 compilers have an option to generate compressed
7895 @code{.opd} entries spaced 16 bytes apart, overlapping the third word,
7896 the static chain pointer (unused in C) with the first word of the next
7897 entry. This option expands such entries to the full 24 bytes.
7899 @cindex PowerPC64 TOC optimization
7900 @kindex --no-toc-optimize
7901 @item --no-toc-optimize
7902 PowerPC64 @command{ld} normally removes unused @code{.toc} section
7903 entries. Such entries are detected by examining relocations that
7904 reference the TOC in code sections. A reloc in a deleted code section
7905 marks a TOC word as unneeded, while a reloc in a kept code section
7906 marks a TOC word as needed. Since the TOC may reference itself, TOC
7907 relocs are also examined. TOC words marked as both needed and
7908 unneeded will of course be kept. TOC words without any referencing
7909 reloc are assumed to be part of a multi-word entry, and are kept or
7910 discarded as per the nearest marked preceding word. This works
7911 reliably for compiler generated code, but may be incorrect if assembly
7912 code is used to insert TOC entries. Use this option to disable the
7915 @cindex PowerPC64 inline PLT call optimization
7916 @kindex --no-inline-optimize
7917 @item --no-inline-optimize
7918 PowerPC64 @command{ld} normally replaces inline PLT call sequences
7919 marked with @code{R_PPC64_PLTSEQ}, @code{R_PPC64_PLTCALL},
7920 @code{R_PPC64_PLT16_HA} and @code{R_PPC64_PLT16_LO_DS} relocations by
7921 a number of @code{nop}s and a direct call when the function is defined
7922 locally and can't be overridden by some other definition. This option
7923 disables that optimization.
7925 @cindex PowerPC64 multi-TOC
7926 @kindex --no-multi-toc
7927 @item --no-multi-toc
7928 If given any toc option besides @code{-mcmodel=medium} or
7929 @code{-mcmodel=large}, PowerPC64 GCC generates code for a TOC model
7931 entries are accessed with a 16-bit offset from r2. This limits the
7932 total TOC size to 64K. PowerPC64 @command{ld} extends this limit by
7933 grouping code sections such that each group uses less than 64K for its
7934 TOC entries, then inserts r2 adjusting stubs between inter-group
7935 calls. @command{ld} does not split apart input sections, so cannot
7936 help if a single input file has a @code{.toc} section that exceeds
7937 64K, most likely from linking multiple files with @command{ld -r}.
7938 Use this option to turn off this feature.
7940 @cindex PowerPC64 TOC sorting
7941 @kindex --no-toc-sort
7943 By default, @command{ld} sorts TOC sections so that those whose file
7944 happens to have a section called @code{.init} or @code{.fini} are
7945 placed first, followed by TOC sections referenced by code generated
7946 with PowerPC64 gcc's @code{-mcmodel=small}, and lastly TOC sections
7947 referenced only by code generated with PowerPC64 gcc's
7948 @code{-mcmodel=medium} or @code{-mcmodel=large} options. Doing this
7949 results in better TOC grouping for multi-TOC. Use this option to turn
7952 @cindex PowerPC64 PLT stub alignment
7954 @kindex --no-plt-align
7956 @itemx --no-plt-align
7957 Use these options to control whether individual PLT call stubs are
7958 aligned to a 32-byte boundary, or to the specified power of two
7959 boundary when using @code{--plt-align=}. A negative value may be
7960 specified to pad PLT call stubs so that they do not cross the
7961 specified power of two boundary (or the minimum number of boundaries
7962 if a PLT stub is so large that it must cross a boundary). By default
7963 PLT call stubs are aligned to 32-byte boundaries.
7965 @cindex PowerPC64 PLT call stub static chain
7966 @kindex --plt-static-chain
7967 @kindex --no-plt-static-chain
7968 @item --plt-static-chain
7969 @itemx --no-plt-static-chain
7970 Use these options to control whether PLT call stubs load the static
7971 chain pointer (r11). @code{ld} defaults to not loading the static
7972 chain since there is never any need to do so on a PLT call.
7974 @cindex PowerPC64 PLT call stub thread safety
7975 @kindex --plt-thread-safe
7976 @kindex --no-plt-thread-safe
7977 @item --plt-thread-safe
7978 @itemx --no-plt-thread-safe
7979 With power7's weakly ordered memory model, it is possible when using
7980 lazy binding for ld.so to update a plt entry in one thread and have
7981 another thread see the individual plt entry words update in the wrong
7982 order, despite ld.so carefully writing in the correct order and using
7983 memory write barriers. To avoid this we need some sort of read
7984 barrier in the call stub, or use LD_BIND_NOW=1. By default, @code{ld}
7985 looks for calls to commonly used functions that create threads, and if
7986 seen, adds the necessary barriers. Use these options to change the
7989 @cindex PowerPC64 ELFv2 PLT localentry optimization
7990 @kindex --plt-localentry
7991 @kindex --no-plt-localentry
7992 @item --plt-localentry
7993 @itemx --no-localentry
7994 ELFv2 functions with localentry:0 are those with a single entry point,
7995 ie. global entry == local entry, and that have no requirement on r2
7996 (the TOC/GOT pointer) or r12, and guarantee r2 is unchanged on return.
7997 Such an external function can be called via the PLT without saving r2
7998 or restoring it on return, avoiding a common load-hit-store for small
7999 functions. The optimization is attractive, with up to 40% reduction
8000 in execution time for a small function, but can result in symbol
8001 interposition failures. Also, minor changes in a shared library,
8002 including system libraries, can cause a function that was localentry:0
8003 to become localentry:8. This will result in a dynamic loader
8004 complaint and failure to run. The option is experimental, use with
8005 care. @option{--no-plt-localentry} is the default.
8007 @cindex PowerPC64 Power10 stubs
8008 @kindex --power10-stubs
8009 @kindex --no-power10-stubs
8010 @item --power10-stubs
8011 @itemx --no-power10-stubs
8012 When PowerPC64 @command{ld} links input object files containing
8013 relocations used on power10 prefixed instructions it normally creates
8014 linkage stubs (PLT call and long branch) using power10 instructions
8015 for @code{@@notoc} PLT calls where @code{r2} is not known. The
8016 power10 notoc stubs are smaller and faster, so are preferred for
8017 power10. @option{--power10-stubs} and @option{--no-power10-stubs}
8018 allow you to override the linker's selection of stub instructions.
8019 @option{--power10-stubs=auto} allows the user to select the default
8034 @section @command{ld} and S/390 ELF Support
8036 @cindex S/390 ELF options
8040 @kindex --s390-pgste
8042 This option marks the result file with a @code{PT_S390_PGSTE}
8043 segment. The Linux kernel is supposed to allocate 4k page tables for
8044 binaries marked that way.
8058 @section @command{ld} and SPU ELF Support
8060 @cindex SPU ELF options
8066 This option marks an executable as a PIC plugin module.
8068 @cindex SPU overlays
8069 @kindex --no-overlays
8071 Normally, @command{ld} recognizes calls to functions within overlay
8072 regions, and redirects such calls to an overlay manager via a stub.
8073 @command{ld} also provides a built-in overlay manager. This option
8074 turns off all this special overlay handling.
8076 @cindex SPU overlay stub symbols
8077 @kindex --emit-stub-syms
8078 @item --emit-stub-syms
8079 This option causes @command{ld} to label overlay stubs with a local
8080 symbol that encodes the stub type and destination.
8082 @cindex SPU extra overlay stubs
8083 @kindex --extra-overlay-stubs
8084 @item --extra-overlay-stubs
8085 This option causes @command{ld} to add overlay call stubs on all
8086 function calls out of overlay regions. Normally stubs are not added
8087 on calls to non-overlay regions.
8089 @cindex SPU local store size
8090 @kindex --local-store=lo:hi
8091 @item --local-store=lo:hi
8092 @command{ld} usually checks that a final executable for SPU fits in
8093 the address range 0 to 256k. This option may be used to change the
8094 range. Disable the check entirely with @option{--local-store=0:0}.
8097 @kindex --stack-analysis
8098 @item --stack-analysis
8099 SPU local store space is limited. Over-allocation of stack space
8100 unnecessarily limits space available for code and data, while
8101 under-allocation results in runtime failures. If given this option,
8102 @command{ld} will provide an estimate of maximum stack usage.
8103 @command{ld} does this by examining symbols in code sections to
8104 determine the extents of functions, and looking at function prologues
8105 for stack adjusting instructions. A call-graph is created by looking
8106 for relocations on branch instructions. The graph is then searched
8107 for the maximum stack usage path. Note that this analysis does not
8108 find calls made via function pointers, and does not handle recursion
8109 and other cycles in the call graph. Stack usage may be
8110 under-estimated if your code makes such calls. Also, stack usage for
8111 dynamic allocation, e.g. alloca, will not be detected. If a link map
8112 is requested, detailed information about each function's stack usage
8113 and calls will be given.
8116 @kindex --emit-stack-syms
8117 @item --emit-stack-syms
8118 This option, if given along with @option{--stack-analysis} will result
8119 in @command{ld} emitting stack sizing symbols for each function.
8120 These take the form @code{__stack_<function_name>} for global
8121 functions, and @code{__stack_<number>_<function_name>} for static
8122 functions. @code{<number>} is the section id in hex. The value of
8123 such symbols is the stack requirement for the corresponding function.
8124 The symbol size will be zero, type @code{STT_NOTYPE}, binding
8125 @code{STB_LOCAL}, and section @code{SHN_ABS}.
8139 @section @command{ld}'s Support for Various TI COFF Versions
8140 @cindex TI COFF versions
8141 @kindex --format=@var{version}
8142 The @samp{--format} switch allows selection of one of the various
8143 TI COFF versions. The latest of this writing is 2; versions 0 and 1 are
8144 also supported. The TI COFF versions also vary in header byte-order
8145 format; @command{ld} will read any version or byte order, but the output
8146 header format depends on the default specified by the specific target.
8159 @section @command{ld} and WIN32 (cygwin/mingw)
8161 This section describes some of the win32 specific @command{ld} issues.
8162 See @ref{Options,,Command-line Options} for detailed description of the
8163 command-line options mentioned here.
8166 @cindex import libraries
8167 @item import libraries
8168 The standard Windows linker creates and uses so-called import
8169 libraries, which contains information for linking to dll's. They are
8170 regular static archives and are handled as any other static
8171 archive. The cygwin and mingw ports of @command{ld} have specific
8172 support for creating such libraries provided with the
8173 @samp{--out-implib} command-line option.
8175 @item exporting DLL symbols
8176 @cindex exporting DLL symbols
8177 The cygwin/mingw @command{ld} has several ways to export symbols for dll's.
8180 @item using auto-export functionality
8181 @cindex using auto-export functionality
8182 By default @command{ld} exports symbols with the auto-export functionality,
8183 which is controlled by the following command-line options:
8186 @item --export-all-symbols [This is the default]
8187 @item --exclude-symbols
8188 @item --exclude-libs
8189 @item --exclude-modules-for-implib
8190 @item --version-script
8193 When auto-export is in operation, @command{ld} will export all the non-local
8194 (global and common) symbols it finds in a DLL, with the exception of a few
8195 symbols known to belong to the system's runtime and libraries. As it will
8196 often not be desirable to export all of a DLL's symbols, which may include
8197 private functions that are not part of any public interface, the command-line
8198 options listed above may be used to filter symbols out from the list for
8199 exporting. The @samp{--output-def} option can be used in order to see the
8200 final list of exported symbols with all exclusions taken into effect.
8202 If @samp{--export-all-symbols} is not given explicitly on the
8203 command line, then the default auto-export behavior will be @emph{disabled}
8204 if either of the following are true:
8207 @item A DEF file is used.
8208 @item Any symbol in any object file was marked with the __declspec(dllexport) attribute.
8211 @item using a DEF file
8212 @cindex using a DEF file
8213 Another way of exporting symbols is using a DEF file. A DEF file is
8214 an ASCII file containing definitions of symbols which should be
8215 exported when a dll is created. Usually it is named @samp{<dll
8216 name>.def} and is added as any other object file to the linker's
8217 command line. The file's name must end in @samp{.def} or @samp{.DEF}.
8220 gcc -o <output> <objectfiles> <dll name>.def
8223 Using a DEF file turns off the normal auto-export behavior, unless the
8224 @samp{--export-all-symbols} option is also used.
8226 Here is an example of a DEF file for a shared library called @samp{xyz.dll}:
8229 LIBRARY "xyz.dll" BASE=0x20000000
8235 another_foo = abc.dll.afoo
8241 This example defines a DLL with a non-default base address and seven
8242 symbols in the export table. The third exported symbol @code{_bar} is an
8243 alias for the second. The fourth symbol, @code{another_foo} is resolved
8244 by "forwarding" to another module and treating it as an alias for
8245 @code{afoo} exported from the DLL @samp{abc.dll}. The final symbol
8246 @code{var1} is declared to be a data object. The @samp{doo} symbol in
8247 export library is an alias of @samp{foo}, which gets the string name
8248 in export table @samp{foo2}. The @samp{eoo} symbol is an data export
8249 symbol, which gets in export table the name @samp{var1}.
8251 The optional @code{LIBRARY <name>} command indicates the @emph{internal}
8252 name of the output DLL. If @samp{<name>} does not include a suffix,
8253 the default library suffix, @samp{.DLL} is appended.
8255 When the .DEF file is used to build an application, rather than a
8256 library, the @code{NAME <name>} command should be used instead of
8257 @code{LIBRARY}. If @samp{<name>} does not include a suffix, the default
8258 executable suffix, @samp{.EXE} is appended.
8260 With either @code{LIBRARY <name>} or @code{NAME <name>} the optional
8261 specification @code{BASE = <number>} may be used to specify a
8262 non-default base address for the image.
8264 If neither @code{LIBRARY <name>} nor @code{NAME <name>} is specified,
8265 or they specify an empty string, the internal name is the same as the
8266 filename specified on the command line.
8268 The complete specification of an export symbol is:
8272 ( ( ( <name1> [ = <name2> ] )
8273 | ( <name1> = <module-name> . <external-name>))
8274 [ @@ <integer> ] [NONAME] [DATA] [CONSTANT] [PRIVATE] [== <name3>] ) *
8277 Declares @samp{<name1>} as an exported symbol from the DLL, or declares
8278 @samp{<name1>} as an exported alias for @samp{<name2>}; or declares
8279 @samp{<name1>} as a "forward" alias for the symbol
8280 @samp{<external-name>} in the DLL @samp{<module-name>}.
8281 Optionally, the symbol may be exported by the specified ordinal
8282 @samp{<integer>} alias. The optional @samp{<name3>} is the to be used
8283 string in import/export table for the symbol.
8285 The optional keywords that follow the declaration indicate:
8287 @code{NONAME}: Do not put the symbol name in the DLL's export table. It
8288 will still be exported by its ordinal alias (either the value specified
8289 by the .def specification or, otherwise, the value assigned by the
8290 linker). The symbol name, however, does remain visible in the import
8291 library (if any), unless @code{PRIVATE} is also specified.
8293 @code{DATA}: The symbol is a variable or object, rather than a function.
8294 The import lib will export only an indirect reference to @code{foo} as
8295 the symbol @code{_imp__foo} (ie, @code{foo} must be resolved as
8298 @code{CONSTANT}: Like @code{DATA}, but put the undecorated @code{foo} as
8299 well as @code{_imp__foo} into the import library. Both refer to the
8300 read-only import address table's pointer to the variable, not to the
8301 variable itself. This can be dangerous. If the user code fails to add
8302 the @code{dllimport} attribute and also fails to explicitly add the
8303 extra indirection that the use of the attribute enforces, the
8304 application will behave unexpectedly.
8306 @code{PRIVATE}: Put the symbol in the DLL's export table, but do not put
8307 it into the static import library used to resolve imports at link time. The
8308 symbol can still be imported using the @code{LoadLibrary/GetProcAddress}
8309 API at runtime or by using the GNU ld extension of linking directly to
8310 the DLL without an import library.
8312 See ld/deffilep.y in the binutils sources for the full specification of
8313 other DEF file statements
8315 @cindex creating a DEF file
8316 While linking a shared dll, @command{ld} is able to create a DEF file
8317 with the @samp{--output-def <file>} command-line option.
8319 @item Using decorations
8320 @cindex Using decorations
8321 Another way of marking symbols for export is to modify the source code
8322 itself, so that when building the DLL each symbol to be exported is
8326 __declspec(dllexport) int a_variable
8327 __declspec(dllexport) void a_function(int with_args)
8330 All such symbols will be exported from the DLL. If, however,
8331 any of the object files in the DLL contain symbols decorated in
8332 this way, then the normal auto-export behavior is disabled, unless
8333 the @samp{--export-all-symbols} option is also used.
8335 Note that object files that wish to access these symbols must @emph{not}
8336 decorate them with dllexport. Instead, they should use dllimport,
8340 __declspec(dllimport) int a_variable
8341 __declspec(dllimport) void a_function(int with_args)
8344 This complicates the structure of library header files, because
8345 when included by the library itself the header must declare the
8346 variables and functions as dllexport, but when included by client
8347 code the header must declare them as dllimport. There are a number
8348 of idioms that are typically used to do this; often client code can
8349 omit the __declspec() declaration completely. See
8350 @samp{--enable-auto-import} and @samp{automatic data imports} for more
8354 @cindex automatic data imports
8355 @item automatic data imports
8356 The standard Windows dll format supports data imports from dlls only
8357 by adding special decorations (dllimport/dllexport), which let the
8358 compiler produce specific assembler instructions to deal with this
8359 issue. This increases the effort necessary to port existing Un*x
8360 code to these platforms, especially for large
8361 c++ libraries and applications. The auto-import feature, which was
8362 initially provided by Paul Sokolovsky, allows one to omit the
8363 decorations to achieve a behavior that conforms to that on POSIX/Un*x
8364 platforms. This feature is enabled with the @samp{--enable-auto-import}
8365 command-line option, although it is enabled by default on cygwin/mingw.
8366 The @samp{--enable-auto-import} option itself now serves mainly to
8367 suppress any warnings that are ordinarily emitted when linked objects
8368 trigger the feature's use.
8370 auto-import of variables does not always work flawlessly without
8371 additional assistance. Sometimes, you will see this message
8373 "variable '<var>' can't be auto-imported. Please read the
8374 documentation for ld's @code{--enable-auto-import} for details."
8376 The @samp{--enable-auto-import} documentation explains why this error
8377 occurs, and several methods that can be used to overcome this difficulty.
8378 One of these methods is the @emph{runtime pseudo-relocs} feature, described
8381 @cindex runtime pseudo-relocation
8382 For complex variables imported from DLLs (such as structs or classes),
8383 object files typically contain a base address for the variable and an
8384 offset (@emph{addend}) within the variable--to specify a particular
8385 field or public member, for instance. Unfortunately, the runtime loader used
8386 in win32 environments is incapable of fixing these references at runtime
8387 without the additional information supplied by dllimport/dllexport decorations.
8388 The standard auto-import feature described above is unable to resolve these
8391 The @samp{--enable-runtime-pseudo-relocs} switch allows these references to
8392 be resolved without error, while leaving the task of adjusting the references
8393 themselves (with their non-zero addends) to specialized code provided by the
8394 runtime environment. Recent versions of the cygwin and mingw environments and
8395 compilers provide this runtime support; older versions do not. However, the
8396 support is only necessary on the developer's platform; the compiled result will
8397 run without error on an older system.
8399 @samp{--enable-runtime-pseudo-relocs} is not the default; it must be explicitly
8402 @cindex direct linking to a dll
8403 @item direct linking to a dll
8404 The cygwin/mingw ports of @command{ld} support the direct linking,
8405 including data symbols, to a dll without the usage of any import
8406 libraries. This is much faster and uses much less memory than does the
8407 traditional import library method, especially when linking large
8408 libraries or applications. When @command{ld} creates an import lib, each
8409 function or variable exported from the dll is stored in its own bfd, even
8410 though a single bfd could contain many exports. The overhead involved in
8411 storing, loading, and processing so many bfd's is quite large, and explains the
8412 tremendous time, memory, and storage needed to link against particularly
8413 large or complex libraries when using import libs.
8415 Linking directly to a dll uses no extra command-line switches other than
8416 @samp{-L} and @samp{-l}, because @command{ld} already searches for a number
8417 of names to match each library. All that is needed from the developer's
8418 perspective is an understanding of this search, in order to force ld to
8419 select the dll instead of an import library.
8422 For instance, when ld is called with the argument @samp{-lxxx} it will attempt
8423 to find, in the first directory of its search path,
8436 before moving on to the next directory in the search path.
8438 (*) Actually, this is not @samp{cygxxx.dll} but in fact is @samp{<prefix>xxx.dll},
8439 where @samp{<prefix>} is set by the @command{ld} option
8440 @samp{--dll-search-prefix=<prefix>}. In the case of cygwin, the standard gcc spec
8441 file includes @samp{--dll-search-prefix=cyg}, so in effect we actually search for
8444 Other win32-based unix environments, such as mingw or pw32, may use other
8445 @samp{<prefix>}es, although at present only cygwin makes use of this feature. It
8446 was originally intended to help avoid name conflicts among dll's built for the
8447 various win32/un*x environments, so that (for example) two versions of a zlib dll
8448 could coexist on the same machine.
8450 The generic cygwin/mingw path layout uses a @samp{bin} directory for
8451 applications and dll's and a @samp{lib} directory for the import
8452 libraries (using cygwin nomenclature):
8458 libxxx.dll.a (in case of dll's)
8459 libxxx.a (in case of static archive)
8462 Linking directly to a dll without using the import library can be
8465 1. Use the dll directly by adding the @samp{bin} path to the link line
8467 gcc -Wl,-verbose -o a.exe -L../bin/ -lxxx
8470 However, as the dll's often have version numbers appended to their names
8471 (@samp{cygncurses-5.dll}) this will often fail, unless one specifies
8472 @samp{-L../bin -lncurses-5} to include the version. Import libs are generally
8473 not versioned, and do not have this difficulty.
8475 2. Create a symbolic link from the dll to a file in the @samp{lib}
8476 directory according to the above mentioned search pattern. This
8477 should be used to avoid unwanted changes in the tools needed for
8481 ln -s bin/cygxxx.dll lib/[cyg|lib|]xxx.dll[.a]
8484 Then you can link without any make environment changes.
8487 gcc -Wl,-verbose -o a.exe -L../lib/ -lxxx
8490 This technique also avoids the version number problems, because the following is
8497 libxxx.dll.a -> ../bin/cygxxx-5.dll
8500 Linking directly to a dll without using an import lib will work
8501 even when auto-import features are exercised, and even when
8502 @samp{--enable-runtime-pseudo-relocs} is used.
8504 Given the improvements in speed and memory usage, one might justifiably
8505 wonder why import libraries are used at all. There are three reasons:
8507 1. Until recently, the link-directly-to-dll functionality did @emph{not}
8508 work with auto-imported data.
8510 2. Sometimes it is necessary to include pure static objects within the
8511 import library (which otherwise contains only bfd's for indirection
8512 symbols that point to the exports of a dll). Again, the import lib
8513 for the cygwin kernel makes use of this ability, and it is not
8514 possible to do this without an import lib.
8516 3. Symbol aliases can only be resolved using an import lib. This is
8517 critical when linking against OS-supplied dll's (eg, the win32 API)
8518 in which symbols are usually exported as undecorated aliases of their
8519 stdcall-decorated assembly names.
8521 So, import libs are not going away. But the ability to replace
8522 true import libs with a simple symbolic link to (or a copy of)
8523 a dll, in many cases, is a useful addition to the suite of tools
8524 binutils makes available to the win32 developer. Given the
8525 massive improvements in memory requirements during linking, storage
8526 requirements, and linking speed, we expect that many developers
8527 will soon begin to use this feature whenever possible.
8529 @item symbol aliasing
8531 @item adding additional names
8532 Sometimes, it is useful to export symbols with additional names.
8533 A symbol @samp{foo} will be exported as @samp{foo}, but it can also be
8534 exported as @samp{_foo} by using special directives in the DEF file
8535 when creating the dll. This will affect also the optional created
8536 import library. Consider the following DEF file:
8539 LIBRARY "xyz.dll" BASE=0x61000000
8546 The line @samp{_foo = foo} maps the symbol @samp{foo} to @samp{_foo}.
8548 Another method for creating a symbol alias is to create it in the
8549 source code using the "weak" attribute:
8552 void foo () @{ /* Do something. */; @}
8553 void _foo () __attribute__ ((weak, alias ("foo")));
8556 See the gcc manual for more information about attributes and weak
8559 @item renaming symbols
8560 Sometimes it is useful to rename exports. For instance, the cygwin
8561 kernel does this regularly. A symbol @samp{_foo} can be exported as
8562 @samp{foo} but not as @samp{_foo} by using special directives in the
8563 DEF file. (This will also affect the import library, if it is
8564 created). In the following example:
8567 LIBRARY "xyz.dll" BASE=0x61000000
8573 The line @samp{_foo = foo} maps the exported symbol @samp{foo} to
8577 Note: using a DEF file disables the default auto-export behavior,
8578 unless the @samp{--export-all-symbols} command-line option is used.
8579 If, however, you are trying to rename symbols, then you should list
8580 @emph{all} desired exports in the DEF file, including the symbols
8581 that are not being renamed, and do @emph{not} use the
8582 @samp{--export-all-symbols} option. If you list only the
8583 renamed symbols in the DEF file, and use @samp{--export-all-symbols}
8584 to handle the other symbols, then the both the new names @emph{and}
8585 the original names for the renamed symbols will be exported.
8586 In effect, you'd be aliasing those symbols, not renaming them,
8587 which is probably not what you wanted.
8589 @cindex weak externals
8590 @item weak externals
8591 The Windows object format, PE, specifies a form of weak symbols called
8592 weak externals. When a weak symbol is linked and the symbol is not
8593 defined, the weak symbol becomes an alias for some other symbol. There
8594 are three variants of weak externals:
8596 @item Definition is searched for in objects and libraries, historically
8597 called lazy externals.
8598 @item Definition is searched for only in other objects, not in libraries.
8599 This form is not presently implemented.
8600 @item No search; the symbol is an alias. This form is not presently
8603 As a GNU extension, weak symbols that do not specify an alternate symbol
8604 are supported. If the symbol is undefined when linking, the symbol
8605 uses a default value.
8607 @cindex aligned common symbols
8608 @item aligned common symbols
8609 As a GNU extension to the PE file format, it is possible to specify the
8610 desired alignment for a common symbol. This information is conveyed from
8611 the assembler or compiler to the linker by means of GNU-specific commands
8612 carried in the object file's @samp{.drectve} section, which are recognized
8613 by @command{ld} and respected when laying out the common symbols. Native
8614 tools will be able to process object files employing this GNU extension,
8615 but will fail to respect the alignment instructions, and may issue noisy
8616 warnings about unknown linker directives.
8631 @section @code{ld} and Xtensa Processors
8633 @cindex Xtensa processors
8634 The default @command{ld} behavior for Xtensa processors is to interpret
8635 @code{SECTIONS} commands so that lists of explicitly named sections in a
8636 specification with a wildcard file will be interleaved when necessary to
8637 keep literal pools within the range of PC-relative load offsets. For
8638 example, with the command:
8650 @command{ld} may interleave some of the @code{.literal}
8651 and @code{.text} sections from different object files to ensure that the
8652 literal pools are within the range of PC-relative load offsets. A valid
8653 interleaving might place the @code{.literal} sections from an initial
8654 group of files followed by the @code{.text} sections of that group of
8655 files. Then, the @code{.literal} sections from the rest of the files
8656 and the @code{.text} sections from the rest of the files would follow.
8658 @cindex @option{--relax} on Xtensa
8659 @cindex relaxing on Xtensa
8660 Relaxation is enabled by default for the Xtensa version of @command{ld} and
8661 provides two important link-time optimizations. The first optimization
8662 is to combine identical literal values to reduce code size. A redundant
8663 literal will be removed and all the @code{L32R} instructions that use it
8664 will be changed to reference an identical literal, as long as the
8665 location of the replacement literal is within the offset range of all
8666 the @code{L32R} instructions. The second optimization is to remove
8667 unnecessary overhead from assembler-generated ``longcall'' sequences of
8668 @code{L32R}/@code{CALLX@var{n}} when the target functions are within
8669 range of direct @code{CALL@var{n}} instructions.
8671 For each of these cases where an indirect call sequence can be optimized
8672 to a direct call, the linker will change the @code{CALLX@var{n}}
8673 instruction to a @code{CALL@var{n}} instruction, remove the @code{L32R}
8674 instruction, and remove the literal referenced by the @code{L32R}
8675 instruction if it is not used for anything else. Removing the
8676 @code{L32R} instruction always reduces code size but can potentially
8677 hurt performance by changing the alignment of subsequent branch targets.
8678 By default, the linker will always preserve alignments, either by
8679 switching some instructions between 24-bit encodings and the equivalent
8680 density instructions or by inserting a no-op in place of the @code{L32R}
8681 instruction that was removed. If code size is more important than
8682 performance, the @option{--size-opt} option can be used to prevent the
8683 linker from widening density instructions or inserting no-ops, except in
8684 a few cases where no-ops are required for correctness.
8686 The following Xtensa-specific command-line options can be used to
8689 @cindex Xtensa options
8692 When optimizing indirect calls to direct calls, optimize for code size
8693 more than performance. With this option, the linker will not insert
8694 no-ops or widen density instructions to preserve branch target
8695 alignment. There may still be some cases where no-ops are required to
8696 preserve the correctness of the code.
8698 @item --abi-windowed
8700 Choose ABI for the output object and for the generated PLT code.
8701 PLT code inserted by the linker must match ABI of the output object
8702 because windowed and call0 ABI use incompatible function call
8704 Default ABI is chosen by the ABI tag in the @code{.xtensa.info} section
8705 of the first input object.
8706 A warning is issued if ABI tags of input objects do not match each other
8707 or the chosen output object ABI.
8715 @ifclear SingleFormat
8720 @cindex object file management
8721 @cindex object formats available
8723 The linker accesses object and archive files using the BFD libraries.
8724 These libraries allow the linker to use the same routines to operate on
8725 object files whatever the object file format. A different object file
8726 format can be supported simply by creating a new BFD back end and adding
8727 it to the library. To conserve runtime memory, however, the linker and
8728 associated tools are usually configured to support only a subset of the
8729 object file formats available. You can use @code{objdump -i}
8730 (@pxref{objdump,,objdump,binutils.info,The GNU Binary Utilities}) to
8731 list all the formats available for your configuration.
8733 @cindex BFD requirements
8734 @cindex requirements for BFD
8735 As with most implementations, BFD is a compromise between
8736 several conflicting requirements. The major factor influencing
8737 BFD design was efficiency: any time used converting between
8738 formats is time which would not have been spent had BFD not
8739 been involved. This is partly offset by abstraction payback; since
8740 BFD simplifies applications and back ends, more time and care
8741 may be spent optimizing algorithms for a greater speed.
8743 One minor artifact of the BFD solution which you should bear in
8744 mind is the potential for information loss. There are two places where
8745 useful information can be lost using the BFD mechanism: during
8746 conversion and during output. @xref{BFD information loss}.
8749 * BFD outline:: How it works: an outline of BFD
8753 @section How It Works: An Outline of BFD
8754 @cindex opening object files
8755 @include bfdsumm.texi
8758 @node Reporting Bugs
8759 @chapter Reporting Bugs
8760 @cindex bugs in @command{ld}
8761 @cindex reporting bugs in @command{ld}
8763 Your bug reports play an essential role in making @command{ld} reliable.
8765 Reporting a bug may help you by bringing a solution to your problem, or
8766 it may not. But in any case the principal function of a bug report is
8767 to help the entire community by making the next version of @command{ld}
8768 work better. Bug reports are your contribution to the maintenance of
8771 In order for a bug report to serve its purpose, you must include the
8772 information that enables us to fix the bug.
8775 * Bug Criteria:: Have you found a bug?
8776 * Bug Reporting:: How to report bugs
8780 @section Have You Found a Bug?
8781 @cindex bug criteria
8783 If you are not sure whether you have found a bug, here are some guidelines:
8786 @cindex fatal signal
8787 @cindex linker crash
8788 @cindex crash of linker
8790 If the linker gets a fatal signal, for any input whatever, that is a
8791 @command{ld} bug. Reliable linkers never crash.
8793 @cindex error on valid input
8795 If @command{ld} produces an error message for valid input, that is a bug.
8797 @cindex invalid input
8799 If @command{ld} does not produce an error message for invalid input, that
8800 may be a bug. In the general case, the linker can not verify that
8801 object files are correct.
8804 If you are an experienced user of linkers, your suggestions for
8805 improvement of @command{ld} are welcome in any case.
8809 @section How to Report Bugs
8811 @cindex @command{ld} bugs, reporting
8813 A number of companies and individuals offer support for @sc{gnu}
8814 products. If you obtained @command{ld} from a support organization, we
8815 recommend you contact that organization first.
8817 You can find contact information for many support companies and
8818 individuals in the file @file{etc/SERVICE} in the @sc{gnu} Emacs
8822 Otherwise, send bug reports for @command{ld} to
8826 The fundamental principle of reporting bugs usefully is this:
8827 @strong{report all the facts}. If you are not sure whether to state a
8828 fact or leave it out, state it!
8830 Often people omit facts because they think they know what causes the
8831 problem and assume that some details do not matter. Thus, you might
8832 assume that the name of a symbol you use in an example does not
8833 matter. Well, probably it does not, but one cannot be sure. Perhaps
8834 the bug is a stray memory reference which happens to fetch from the
8835 location where that name is stored in memory; perhaps, if the name
8836 were different, the contents of that location would fool the linker
8837 into doing the right thing despite the bug. Play it safe and give a
8838 specific, complete example. That is the easiest thing for you to do,
8839 and the most helpful.
8841 Keep in mind that the purpose of a bug report is to enable us to fix
8842 the bug if it is new to us. Therefore, always write your bug reports
8843 on the assumption that the bug has not been reported previously.
8845 Sometimes people give a few sketchy facts and ask, ``Does this ring a
8846 bell?'' This cannot help us fix a bug, so it is basically useless. We
8847 respond by asking for enough details to enable us to investigate.
8848 You might as well expedite matters by sending them to begin with.
8850 To enable us to fix the bug, you should include all these things:
8854 The version of @command{ld}. @command{ld} announces it if you start it with
8855 the @samp{--version} argument.
8857 Without this, we will not know whether there is any point in looking for
8858 the bug in the current version of @command{ld}.
8861 Any patches you may have applied to the @command{ld} source, including any
8862 patches made to the @code{BFD} library.
8865 The type of machine you are using, and the operating system name and
8869 What compiler (and its version) was used to compile @command{ld}---e.g.
8873 The command arguments you gave the linker to link your example and
8874 observe the bug. To guarantee you will not omit something important,
8875 list them all. A copy of the Makefile (or the output from make) is
8878 If we were to try to guess the arguments, we would probably guess wrong
8879 and then we might not encounter the bug.
8882 A complete input file, or set of input files, that will reproduce the
8883 bug. It is generally most helpful to send the actual object files
8884 provided that they are reasonably small. Say no more than 10K. For
8885 bigger files you can either make them available by FTP or HTTP or else
8886 state that you are willing to send the object file(s) to whomever
8887 requests them. (Note - your email will be going to a mailing list, so
8888 we do not want to clog it up with large attachments). But small
8889 attachments are best.
8891 If the source files were assembled using @code{gas} or compiled using
8892 @code{gcc}, then it may be OK to send the source files rather than the
8893 object files. In this case, be sure to say exactly what version of
8894 @code{gas} or @code{gcc} was used to produce the object files. Also say
8895 how @code{gas} or @code{gcc} were configured.
8898 A description of what behavior you observe that you believe is
8899 incorrect. For example, ``It gets a fatal signal.''
8901 Of course, if the bug is that @command{ld} gets a fatal signal, then we
8902 will certainly notice it. But if the bug is incorrect output, we might
8903 not notice unless it is glaringly wrong. You might as well not give us
8904 a chance to make a mistake.
8906 Even if the problem you experience is a fatal signal, you should still
8907 say so explicitly. Suppose something strange is going on, such as, your
8908 copy of @command{ld} is out of sync, or you have encountered a bug in the
8909 C library on your system. (This has happened!) Your copy might crash
8910 and ours would not. If you told us to expect a crash, then when ours
8911 fails to crash, we would know that the bug was not happening for us. If
8912 you had not told us to expect a crash, then we would not be able to draw
8913 any conclusion from our observations.
8916 If you wish to suggest changes to the @command{ld} source, send us context
8917 diffs, as generated by @code{diff} with the @samp{-u}, @samp{-c}, or
8918 @samp{-p} option. Always send diffs from the old file to the new file.
8919 If you even discuss something in the @command{ld} source, refer to it by
8920 context, not by line number.
8922 The line numbers in our development sources will not match those in your
8923 sources. Your line numbers would convey no useful information to us.
8926 Here are some things that are not necessary:
8930 A description of the envelope of the bug.
8932 Often people who encounter a bug spend a lot of time investigating
8933 which changes to the input file will make the bug go away and which
8934 changes will not affect it.
8936 This is often time consuming and not very useful, because the way we
8937 will find the bug is by running a single example under the debugger
8938 with breakpoints, not by pure deduction from a series of examples.
8939 We recommend that you save your time for something else.
8941 Of course, if you can find a simpler example to report @emph{instead}
8942 of the original one, that is a convenience for us. Errors in the
8943 output will be easier to spot, running under the debugger will take
8944 less time, and so on.
8946 However, simplification is not vital; if you do not want to do this,
8947 report the bug anyway and send us the entire test case you used.
8950 A patch for the bug.
8952 A patch for the bug does help us if it is a good one. But do not omit
8953 the necessary information, such as the test case, on the assumption that
8954 a patch is all we need. We might see problems with your patch and decide
8955 to fix the problem another way, or we might not understand it at all.
8957 Sometimes with a program as complicated as @command{ld} it is very hard to
8958 construct an example that will make the program follow a certain path
8959 through the code. If you do not send us the example, we will not be
8960 able to construct one, so we will not be able to verify that the bug is
8963 And if we cannot understand what bug you are trying to fix, or why your
8964 patch should be an improvement, we will not install it. A test case will
8965 help us to understand.
8968 A guess about what the bug is or what it depends on.
8970 Such guesses are usually wrong. Even we cannot guess right about such
8971 things without first using the debugger to find the facts.
8975 @appendix MRI Compatible Script Files
8976 @cindex MRI compatibility
8977 To aid users making the transition to @sc{gnu} @command{ld} from the MRI
8978 linker, @command{ld} can use MRI compatible linker scripts as an
8979 alternative to the more general-purpose linker scripting language
8980 described in @ref{Scripts}. MRI compatible linker scripts have a much
8981 simpler command set than the scripting language otherwise used with
8982 @command{ld}. @sc{gnu} @command{ld} supports the most commonly used MRI
8983 linker commands; these commands are described here.
8985 In general, MRI scripts aren't of much use with the @code{a.out} object
8986 file format, since it only has three sections and MRI scripts lack some
8987 features to make use of them.
8989 You can specify a file containing an MRI-compatible script using the
8990 @samp{-c} command-line option.
8992 Each command in an MRI-compatible script occupies its own line; each
8993 command line starts with the keyword that identifies the command (though
8994 blank lines are also allowed for punctuation). If a line of an
8995 MRI-compatible script begins with an unrecognized keyword, @command{ld}
8996 issues a warning message, but continues processing the script.
8998 Lines beginning with @samp{*} are comments.
9000 You can write these commands using all upper-case letters, or all
9001 lower case; for example, @samp{chip} is the same as @samp{CHIP}.
9002 The following list shows only the upper-case form of each command.
9005 @cindex @code{ABSOLUTE} (MRI)
9006 @item ABSOLUTE @var{secname}
9007 @itemx ABSOLUTE @var{secname}, @var{secname}, @dots{} @var{secname}
9008 Normally, @command{ld} includes in the output file all sections from all
9009 the input files. However, in an MRI-compatible script, you can use the
9010 @code{ABSOLUTE} command to restrict the sections that will be present in
9011 your output program. If the @code{ABSOLUTE} command is used at all in a
9012 script, then only the sections named explicitly in @code{ABSOLUTE}
9013 commands will appear in the linker output. You can still use other
9014 input sections (whatever you select on the command line, or using
9015 @code{LOAD}) to resolve addresses in the output file.
9017 @cindex @code{ALIAS} (MRI)
9018 @item ALIAS @var{out-secname}, @var{in-secname}
9019 Use this command to place the data from input section @var{in-secname}
9020 in a section called @var{out-secname} in the linker output file.
9022 @var{in-secname} may be an integer.
9024 @cindex @code{ALIGN} (MRI)
9025 @item ALIGN @var{secname} = @var{expression}
9026 Align the section called @var{secname} to @var{expression}. The
9027 @var{expression} should be a power of two.
9029 @cindex @code{BASE} (MRI)
9030 @item BASE @var{expression}
9031 Use the value of @var{expression} as the lowest address (other than
9032 absolute addresses) in the output file.
9034 @cindex @code{CHIP} (MRI)
9035 @item CHIP @var{expression}
9036 @itemx CHIP @var{expression}, @var{expression}
9037 This command does nothing; it is accepted only for compatibility.
9039 @cindex @code{END} (MRI)
9041 This command does nothing whatever; it's only accepted for compatibility.
9043 @cindex @code{FORMAT} (MRI)
9044 @item FORMAT @var{output-format}
9045 Similar to the @code{OUTPUT_FORMAT} command in the more general linker
9046 language, but restricted to S-records, if @var{output-format} is @samp{S}
9048 @cindex @code{LIST} (MRI)
9049 @item LIST @var{anything}@dots{}
9050 Print (to the standard output file) a link map, as produced by the
9051 @command{ld} command-line option @samp{-M}.
9053 The keyword @code{LIST} may be followed by anything on the
9054 same line, with no change in its effect.
9056 @cindex @code{LOAD} (MRI)
9057 @item LOAD @var{filename}
9058 @itemx LOAD @var{filename}, @var{filename}, @dots{} @var{filename}
9059 Include one or more object file @var{filename} in the link; this has the
9060 same effect as specifying @var{filename} directly on the @command{ld}
9063 @cindex @code{NAME} (MRI)
9064 @item NAME @var{output-name}
9065 @var{output-name} is the name for the program produced by @command{ld}; the
9066 MRI-compatible command @code{NAME} is equivalent to the command-line
9067 option @samp{-o} or the general script language command @code{OUTPUT}.
9069 @cindex @code{ORDER} (MRI)
9070 @item ORDER @var{secname}, @var{secname}, @dots{} @var{secname}
9071 @itemx ORDER @var{secname} @var{secname} @var{secname}
9072 Normally, @command{ld} orders the sections in its output file in the
9073 order in which they first appear in the input files. In an MRI-compatible
9074 script, you can override this ordering with the @code{ORDER} command. The
9075 sections you list with @code{ORDER} will appear first in your output
9076 file, in the order specified.
9078 @cindex @code{PUBLIC} (MRI)
9079 @item PUBLIC @var{name}=@var{expression}
9080 @itemx PUBLIC @var{name},@var{expression}
9081 @itemx PUBLIC @var{name} @var{expression}
9082 Supply a value (@var{expression}) for external symbol
9083 @var{name} used in the linker input files.
9085 @cindex @code{SECT} (MRI)
9086 @item SECT @var{secname}, @var{expression}
9087 @itemx SECT @var{secname}=@var{expression}
9088 @itemx SECT @var{secname} @var{expression}
9089 You can use any of these three forms of the @code{SECT} command to
9090 specify the start address (@var{expression}) for section @var{secname}.
9091 If you have more than one @code{SECT} statement for the same
9092 @var{secname}, only the @emph{first} sets the start address.
9095 @node GNU Free Documentation License
9096 @appendix GNU Free Documentation License
9100 @unnumbered LD Index
9105 % I think something like @@colophon should be in texinfo. In the
9107 \long\def\colophon{\hbox to0pt{}\vfill
9108 \centerline{The body of this manual is set in}
9109 \centerline{\fontname\tenrm,}
9110 \centerline{with headings in {\bf\fontname\tenbf}}
9111 \centerline{and examples in {\tt\fontname\tentt}.}
9112 \centerline{{\it\fontname\tenit\/} and}
9113 \centerline{{\sl\fontname\tensl\/}}
9114 \centerline{are used for emphasis.}\vfill}
9116 % Blame: doc@@cygnus.com, 28mar91.