3 @c Copyright (C) 1991-2021 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{} 1991-2021 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{} 1991-2021 Free Software Foundation, Inc.
99 Permission is granted to copy, distribute and/or modify this document
100 under the terms of the GNU Free Documentation License, Version 1.3
101 or any later version published by the Free Software Foundation;
102 with no Invariant Sections, with no Front-Cover Texts, and with no
103 Back-Cover Texts. A copy of the license is included in the
104 section entitled ``GNU Free Documentation License''.
110 @c FIXME: Talk about importance of *order* of args, cmds to linker!
115 This file documents the @sc{gnu} linker ld
116 @ifset VERSION_PACKAGE
117 @value{VERSION_PACKAGE}
119 version @value{VERSION}.
121 This document is distributed under the terms of the GNU Free
122 Documentation License version 1.3. A copy of the license is included
123 in the section entitled ``GNU Free Documentation License''.
126 * Overview:: Overview
127 * Invocation:: Invocation
128 * Scripts:: Linker Scripts
129 * Plugins:: Linker Plugins
131 * Machine Dependent:: Machine Dependent Features
135 * H8/300:: ld and the H8/300
138 * Renesas:: ld and other Renesas micros
141 * ARM:: ld and the ARM family
144 * M68HC11/68HC12:: ld and the Motorola 68HC11 and 68HC12 families
147 * HPPA ELF32:: ld and HPPA 32-bit ELF
150 * M68K:: ld and Motorola 68K family
153 * MIPS:: ld and MIPS family
156 * PowerPC ELF32:: ld and PowerPC 32-bit ELF Support
159 * PowerPC64 ELF64:: ld and PowerPC64 64-bit ELF Support
162 * S/390 ELF:: ld and S/390 ELF Support
165 * SPU ELF:: ld and SPU ELF Support
168 * TI COFF:: ld and the TI COFF
171 * Win32:: ld and WIN32 (cygwin/mingw)
174 * Xtensa:: ld and Xtensa Processors
177 @ifclear SingleFormat
180 @c Following blank line required for remaining bug in makeinfo conds/menus
182 * Reporting Bugs:: Reporting Bugs
183 * MRI:: MRI Compatible Script Files
184 * GNU Free Documentation License:: GNU Free Documentation License
185 * LD Index:: LD Index
192 @cindex @sc{gnu} linker
193 @cindex what is this?
196 @c man begin SYNOPSIS
197 ld [@b{options}] @var{objfile} @dots{}
201 ar(1), nm(1), objcopy(1), objdump(1), readelf(1) and
202 the Info entries for @file{binutils} and
207 @c man begin DESCRIPTION
209 @command{ld} combines a number of object and archive files, relocates
210 their data and ties up symbol references. Usually the last step in
211 compiling a program is to run @command{ld}.
213 @command{ld} accepts Linker Command Language files written in
214 a superset of AT&T's Link Editor Command Language syntax,
215 to provide explicit and total control over the linking process.
219 This man page does not describe the command language; see the
220 @command{ld} entry in @code{info} for full details on the command
221 language and on other aspects of the GNU linker.
224 @ifclear SingleFormat
225 This version of @command{ld} uses the general purpose BFD libraries
226 to operate on object files. This allows @command{ld} to read, combine, and
227 write object files in many different formats---for example, COFF or
228 @code{a.out}. Different formats may be linked together to produce any
229 available kind of object file. @xref{BFD}, for more information.
232 Aside from its flexibility, the @sc{gnu} linker is more helpful than other
233 linkers in providing diagnostic information. Many linkers abandon
234 execution immediately upon encountering an error; whenever possible,
235 @command{ld} continues executing, allowing you to identify other errors
236 (or, in some cases, to get an output file in spite of the error).
243 @c man begin DESCRIPTION
245 The @sc{gnu} linker @command{ld} is meant to cover a broad range of situations,
246 and to be as compatible as possible with other linkers. As a result,
247 you have many choices to control its behavior.
253 * Options:: Command-line Options
254 * Environment:: Environment Variables
258 @section Command-line Options
266 The linker supports a plethora of command-line options, but in actual
267 practice few of them are used in any particular context.
268 @cindex standard Unix system
269 For instance, a frequent use of @command{ld} is to link standard Unix
270 object files on a standard, supported Unix system. On such a system, to
271 link a file @code{hello.o}:
274 ld -o @var{output} /lib/crt0.o hello.o -lc
277 This tells @command{ld} to produce a file called @var{output} as the
278 result of linking the file @code{/lib/crt0.o} with @code{hello.o} and
279 the library @code{libc.a}, which will come from the standard search
280 directories. (See the discussion of the @samp{-l} option below.)
282 Some of the command-line options to @command{ld} may be specified at any
283 point in the command line. However, options which refer to files, such
284 as @samp{-l} or @samp{-T}, cause the file to be read at the point at
285 which the option appears in the command line, relative to the object
286 files and other file options. Repeating non-file options with a
287 different argument will either have no further effect, or override prior
288 occurrences (those further to the left on the command line) of that
289 option. Options which may be meaningfully specified more than once are
290 noted in the descriptions below.
293 Non-option arguments are object files or archives which are to be linked
294 together. They may follow, precede, or be mixed in with command-line
295 options, except that an object file argument may not be placed between
296 an option and its argument.
298 Usually the linker is invoked with at least one object file, but you can
299 specify other forms of binary input files using @samp{-l}, @samp{-R},
300 and the script command language. If @emph{no} binary input files at all
301 are specified, the linker does not produce any output, and issues the
302 message @samp{No input files}.
304 If the linker cannot recognize the format of an object file, it will
305 assume that it is a linker script. A script specified in this way
306 augments the main linker script used for the link (either the default
307 linker script or the one specified by using @samp{-T}). This feature
308 permits the linker to link against a file which appears to be an object
309 or an archive, but actually merely defines some symbol values, or uses
310 @code{INPUT} or @code{GROUP} to load other objects. Specifying a
311 script in this way merely augments the main linker script, with the
312 extra commands placed after the main script; use the @samp{-T} option
313 to replace the default linker script entirely, but note the effect of
314 the @code{INSERT} command. @xref{Scripts}.
316 For options whose names are a single letter,
317 option arguments must either follow the option letter without intervening
318 whitespace, or be given as separate arguments immediately following the
319 option that requires them.
321 For options whose names are multiple letters, either one dash or two can
322 precede the option name; for example, @samp{-trace-symbol} and
323 @samp{--trace-symbol} are equivalent. Note---there is one exception to
324 this rule. Multiple letter options that start with a lower case 'o' can
325 only be preceded by two dashes. This is to reduce confusion with the
326 @samp{-o} option. So for example @samp{-omagic} sets the output file
327 name to @samp{magic} whereas @samp{--omagic} sets the NMAGIC flag on the
330 Arguments to multiple-letter options must either be separated from the
331 option name by an equals sign, or be given as separate arguments
332 immediately following the option that requires them. For example,
333 @samp{--trace-symbol foo} and @samp{--trace-symbol=foo} are equivalent.
334 Unique abbreviations of the names of multiple-letter options are
337 Note---if the linker is being invoked indirectly, via a compiler driver
338 (e.g. @samp{gcc}) then all the linker command-line options should be
339 prefixed by @samp{-Wl,} (or whatever is appropriate for the particular
340 compiler driver) like this:
343 gcc -Wl,--start-group foo.o bar.o -Wl,--end-group
346 This is important, because otherwise the compiler driver program may
347 silently drop the linker options, resulting in a bad link. Confusion
348 may also arise when passing options that require values through a
349 driver, as the use of a space between option and argument acts as
350 a separator, and causes the driver to pass only the option to the linker
351 and the argument to the compiler. In this case, it is simplest to use
352 the joined forms of both single- and multiple-letter options, such as:
355 gcc foo.o bar.o -Wl,-eENTRY -Wl,-Map=a.map
358 Here is a table of the generic command-line switches accepted by the GNU
362 @include at-file.texi
364 @kindex -a @var{keyword}
365 @item -a @var{keyword}
366 This option is supported for HP/UX compatibility. The @var{keyword}
367 argument must be one of the strings @samp{archive}, @samp{shared}, or
368 @samp{default}. @samp{-aarchive} is functionally equivalent to
369 @samp{-Bstatic}, and the other two keywords are functionally equivalent
370 to @samp{-Bdynamic}. This option may be used any number of times.
372 @kindex --audit @var{AUDITLIB}
373 @item --audit @var{AUDITLIB}
374 Adds @var{AUDITLIB} to the @code{DT_AUDIT} entry of the dynamic section.
375 @var{AUDITLIB} is not checked for existence, nor will it use the DT_SONAME
376 specified in the library. If specified multiple times @code{DT_AUDIT}
377 will contain a colon separated list of audit interfaces to use. If the linker
378 finds an object with an audit entry while searching for shared libraries,
379 it will add a corresponding @code{DT_DEPAUDIT} entry in the output file.
380 This option is only meaningful on ELF platforms supporting the rtld-audit
383 @ifclear SingleFormat
384 @cindex binary input format
385 @kindex -b @var{format}
386 @kindex --format=@var{format}
389 @item -b @var{input-format}
390 @itemx --format=@var{input-format}
391 @command{ld} may be configured to support more than one kind of object
392 file. If your @command{ld} is configured this way, you can use the
393 @samp{-b} option to specify the binary format for input object files
394 that follow this option on the command line. Even when @command{ld} is
395 configured to support alternative object formats, you don't usually need
396 to specify this, as @command{ld} should be configured to expect as a
397 default input format the most usual format on each machine.
398 @var{input-format} is a text string, the name of a particular format
399 supported by the BFD libraries. (You can list the available binary
400 formats with @samp{objdump -i}.)
403 You may want to use this option if you are linking files with an unusual
404 binary format. You can also use @samp{-b} to switch formats explicitly (when
405 linking object files of different formats), by including
406 @samp{-b @var{input-format}} before each group of object files in a
409 The default format is taken from the environment variable
414 You can also define the input format from a script, using the command
417 see @ref{Format Commands}.
421 @kindex -c @var{MRI-cmdfile}
422 @kindex --mri-script=@var{MRI-cmdfile}
423 @cindex compatibility, MRI
424 @item -c @var{MRI-commandfile}
425 @itemx --mri-script=@var{MRI-commandfile}
426 For compatibility with linkers produced by MRI, @command{ld} accepts script
427 files written in an alternate, restricted command language, described in
429 @ref{MRI,,MRI Compatible Script Files}.
432 the MRI Compatible Script Files section of GNU ld documentation.
434 Introduce MRI script files with
435 the option @samp{-c}; use the @samp{-T} option to run linker
436 scripts written in the general-purpose @command{ld} scripting language.
437 If @var{MRI-cmdfile} does not exist, @command{ld} looks for it in the directories
438 specified by any @samp{-L} options.
440 @cindex common allocation
447 These three options are equivalent; multiple forms are supported for
448 compatibility with other linkers. They assign space to common symbols
449 even if a relocatable output file is specified (with @samp{-r}). The
450 script command @code{FORCE_COMMON_ALLOCATION} has the same effect.
451 @xref{Miscellaneous Commands}.
453 @kindex --depaudit @var{AUDITLIB}
454 @kindex -P @var{AUDITLIB}
455 @item --depaudit @var{AUDITLIB}
456 @itemx -P @var{AUDITLIB}
457 Adds @var{AUDITLIB} to the @code{DT_DEPAUDIT} entry of the dynamic section.
458 @var{AUDITLIB} is not checked for existence, nor will it use the DT_SONAME
459 specified in the library. If specified multiple times @code{DT_DEPAUDIT}
460 will contain a colon separated list of audit interfaces to use. This
461 option is only meaningful on ELF platforms supporting the rtld-audit interface.
462 The -P option is provided for Solaris compatibility.
464 @kindex --enable-non-contiguous-regions
465 @item --enable-non-contiguous-regions
466 This option avoids generating an error if an input section does not
467 fit a matching output section. The linker tries to allocate the input
468 section to subseque nt matching output sections, and generates an
469 error only if no output section is large enough. This is useful when
470 several non-contiguous memory regions are available and the input
471 section does not require a particular one. The order in which input
472 sections are evaluated does not change, for instance:
476 MEM1 (rwx) : ORIGIN : 0x1000, LENGTH = 0x14
477 MEM2 (rwx) : ORIGIN : 0x1000, LENGTH = 0x40
478 MEM3 (rwx) : ORIGIN : 0x2000, LENGTH = 0x40
481 mem1 : @{ *(.data.*); @} > MEM1
482 mem2 : @{ *(.data.*); @} > MEM2
483 mem3 : @{ *(.data.*); @} > MEM2
491 results in .data.1 affected to mem1, and .data.2 and .data.3
492 affected to mem2, even though .data.3 would fit in mem3.
495 This option is incompatible with INSERT statements because it changes
496 the way input sections are mapped to output sections.
498 @kindex --enable-non-contiguous-regions-warnings
499 @item --enable-non-contiguous-regions-warnings
500 This option enables warnings when
501 @code{--enable-non-contiguous-regions} allows possibly unexpected
502 matches in sections mapping, potentially leading to silently
503 discarding a section instead of failing because it does not fit any
506 @cindex entry point, from command line
507 @kindex -e @var{entry}
508 @kindex --entry=@var{entry}
510 @itemx --entry=@var{entry}
511 Use @var{entry} as the explicit symbol for beginning execution of your
512 program, rather than the default entry point. If there is no symbol
513 named @var{entry}, the linker will try to parse @var{entry} as a number,
514 and use that as the entry address (the number will be interpreted in
515 base 10; you may use a leading @samp{0x} for base 16, or a leading
516 @samp{0} for base 8). @xref{Entry Point}, for a discussion of defaults
517 and other ways of specifying the entry point.
519 @kindex --exclude-libs
520 @item --exclude-libs @var{lib},@var{lib},...
521 Specifies a list of archive libraries from which symbols should not be automatically
522 exported. The library names may be delimited by commas or colons. Specifying
523 @code{--exclude-libs ALL} excludes symbols in all archive libraries from
524 automatic export. This option is available only for the i386 PE targeted
525 port of the linker and for ELF targeted ports. For i386 PE, symbols
526 explicitly listed in a .def file are still exported, regardless of this
527 option. For ELF targeted ports, symbols affected by this option will
528 be treated as hidden.
530 @kindex --exclude-modules-for-implib
531 @item --exclude-modules-for-implib @var{module},@var{module},...
532 Specifies a list of object files or archive members, from which symbols
533 should not be automatically exported, but which should be copied wholesale
534 into the import library being generated during the link. The module names
535 may be delimited by commas or colons, and must match exactly the filenames
536 used by @command{ld} to open the files; for archive members, this is simply
537 the member name, but for object files the name listed must include and
538 match precisely any path used to specify the input file on the linker's
539 command-line. This option is available only for the i386 PE targeted port
540 of the linker. Symbols explicitly listed in a .def file are still exported,
541 regardless of this option.
543 @cindex dynamic symbol table
545 @kindex --export-dynamic
546 @kindex --no-export-dynamic
548 @itemx --export-dynamic
549 @itemx --no-export-dynamic
550 When creating a dynamically linked executable, using the @option{-E}
551 option or the @option{--export-dynamic} option causes the linker to add
552 all symbols to the dynamic symbol table. The dynamic symbol table is the
553 set of symbols which are visible from dynamic objects at run time.
555 If you do not use either of these options (or use the
556 @option{--no-export-dynamic} option to restore the default behavior), the
557 dynamic symbol table will normally contain only those symbols which are
558 referenced by some dynamic object mentioned in the link.
560 If you use @code{dlopen} to load a dynamic object which needs to refer
561 back to the symbols defined by the program, rather than some other
562 dynamic object, then you will probably need to use this option when
563 linking the program itself.
565 You can also use the dynamic list to control what symbols should
566 be added to the dynamic symbol table if the output format supports it.
567 See the description of @samp{--dynamic-list}.
569 Note that this option is specific to ELF targeted ports. PE targets
570 support a similar function to export all symbols from a DLL or EXE; see
571 the description of @samp{--export-all-symbols} below.
573 @kindex --export-dynamic-symbol=@var{glob}
574 @cindex export dynamic symbol
575 @item --export-dynamic-symbol=@var{glob}
576 When creating a dynamically linked executable, symbols matching
577 @var{glob} will be added to the dynamic symbol table. When creating a
578 shared library, references to symbols matching @var{glob} will not be
579 bound to the definitions within the shared library. This option is a
580 no-op when creating a shared library and @samp{-Bsymbolic} or
581 @samp{--dynamic-list} are not specified. This option is only meaningful
582 on ELF platforms which support shared libraries.
584 @kindex --export-dynamic-symbol-list=@var{file}
585 @cindex export dynamic symbol list
586 @item --export-dynamic-symbol-list=@var{file}
587 Specify a @samp{--export-dynamic-symbol} for each pattern in the file.
588 The format of the file is the same as the version node without
589 scope and node name. See @ref{VERSION} for more information.
591 @ifclear SingleFormat
592 @cindex big-endian objects
596 Link big-endian objects. This affects the default output format.
598 @cindex little-endian objects
601 Link little-endian objects. This affects the default output format.
604 @kindex -f @var{name}
605 @kindex --auxiliary=@var{name}
607 @itemx --auxiliary=@var{name}
608 When creating an ELF shared object, set the internal DT_AUXILIARY field
609 to the specified name. This tells the dynamic linker that the symbol
610 table of the shared object should be used as an auxiliary filter on the
611 symbol table of the shared object @var{name}.
613 If you later link a program against this filter object, then, when you
614 run the program, the dynamic linker will see the DT_AUXILIARY field. If
615 the dynamic linker resolves any symbols from the filter object, it will
616 first check whether there is a definition in the shared object
617 @var{name}. If there is one, it will be used instead of the definition
618 in the filter object. The shared object @var{name} need not exist.
619 Thus the shared object @var{name} may be used to provide an alternative
620 implementation of certain functions, perhaps for debugging or for
621 machine-specific performance.
623 This option may be specified more than once. The DT_AUXILIARY entries
624 will be created in the order in which they appear on the command line.
626 @kindex -F @var{name}
627 @kindex --filter=@var{name}
629 @itemx --filter=@var{name}
630 When creating an ELF shared object, set the internal DT_FILTER field to
631 the specified name. This tells the dynamic linker that the symbol table
632 of the shared object which is being created should be used as a filter
633 on the symbol table of the shared object @var{name}.
635 If you later link a program against this filter object, then, when you
636 run the program, the dynamic linker will see the DT_FILTER field. The
637 dynamic linker will resolve symbols according to the symbol table of the
638 filter object as usual, but it will actually link to the definitions
639 found in the shared object @var{name}. Thus the filter object can be
640 used to select a subset of the symbols provided by the object
643 Some older linkers used the @option{-F} option throughout a compilation
644 toolchain for specifying object-file format for both input and output
646 @ifclear SingleFormat
647 The @sc{gnu} linker uses other mechanisms for this purpose: the
648 @option{-b}, @option{--format}, @option{--oformat} options, the
649 @code{TARGET} command in linker scripts, and the @code{GNUTARGET}
650 environment variable.
652 The @sc{gnu} linker will ignore the @option{-F} option when not
653 creating an ELF shared object.
655 @cindex finalization function
656 @kindex -fini=@var{name}
657 @item -fini=@var{name}
658 When creating an ELF executable or shared object, call NAME when the
659 executable or shared object is unloaded, by setting DT_FINI to the
660 address of the function. By default, the linker uses @code{_fini} as
661 the function to call.
665 Ignored. Provided for compatibility with other tools.
667 @kindex -G @var{value}
668 @kindex --gpsize=@var{value}
671 @itemx --gpsize=@var{value}
672 Set the maximum size of objects to be optimized using the GP register to
673 @var{size}. This is only meaningful for object file formats such as
674 MIPS ELF that support putting large and small objects into different
675 sections. This is ignored for other object file formats.
677 @cindex runtime library name
678 @kindex -h @var{name}
679 @kindex -soname=@var{name}
681 @itemx -soname=@var{name}
682 When creating an ELF shared object, set the internal DT_SONAME field to
683 the specified name. When an executable is linked with a shared object
684 which has a DT_SONAME field, then when the executable is run the dynamic
685 linker will attempt to load the shared object specified by the DT_SONAME
686 field rather than the using the file name given to the linker.
689 @cindex incremental link
691 Perform an incremental link (same as option @samp{-r}).
693 @cindex initialization function
694 @kindex -init=@var{name}
695 @item -init=@var{name}
696 When creating an ELF executable or shared object, call NAME when the
697 executable or shared object is loaded, by setting DT_INIT to the address
698 of the function. By default, the linker uses @code{_init} as the
701 @cindex archive files, from cmd line
702 @kindex -l @var{namespec}
703 @kindex --library=@var{namespec}
704 @item -l @var{namespec}
705 @itemx --library=@var{namespec}
706 Add the archive or object file specified by @var{namespec} to the
707 list of files to link. This option may be used any number of times.
708 If @var{namespec} is of the form @file{:@var{filename}}, @command{ld}
709 will search the library path for a file called @var{filename}, otherwise it
710 will search the library path for a file called @file{lib@var{namespec}.a}.
712 On systems which support shared libraries, @command{ld} may also search for
713 files other than @file{lib@var{namespec}.a}. Specifically, on ELF
714 and SunOS systems, @command{ld} will search a directory for a library
715 called @file{lib@var{namespec}.so} before searching for one called
716 @file{lib@var{namespec}.a}. (By convention, a @code{.so} extension
717 indicates a shared library.) Note that this behavior does not apply
718 to @file{:@var{filename}}, which always specifies a file called
721 The linker will search an archive only once, at the location where it is
722 specified on the command line. If the archive defines a symbol which
723 was undefined in some object which appeared before the archive on the
724 command line, the linker will include the appropriate file(s) from the
725 archive. However, an undefined symbol in an object appearing later on
726 the command line will not cause the linker to search the archive again.
728 See the @option{-(} option for a way to force the linker to search
729 archives multiple times.
731 You may list the same archive multiple times on the command line.
734 This type of archive searching is standard for Unix linkers. However,
735 if you are using @command{ld} on AIX, note that it is different from the
736 behaviour of the AIX linker.
739 @cindex search directory, from cmd line
741 @kindex --library-path=@var{dir}
742 @item -L @var{searchdir}
743 @itemx --library-path=@var{searchdir}
744 Add path @var{searchdir} to the list of paths that @command{ld} will search
745 for archive libraries and @command{ld} control scripts. You may use this
746 option any number of times. The directories are searched in the order
747 in which they are specified on the command line. Directories specified
748 on the command line are searched before the default directories. All
749 @option{-L} options apply to all @option{-l} options, regardless of the
750 order in which the options appear. @option{-L} options do not affect
751 how @command{ld} searches for a linker script unless @option{-T}
754 If @var{searchdir} begins with @code{=} or @code{$SYSROOT}, then this
755 prefix will be replaced by the @dfn{sysroot prefix}, controlled by the
756 @samp{--sysroot} option, or specified when the linker is configured.
759 The default set of paths searched (without being specified with
760 @samp{-L}) depends on which emulation mode @command{ld} is using, and in
761 some cases also on how it was configured. @xref{Environment}.
764 The paths can also be specified in a link script with the
765 @code{SEARCH_DIR} command. Directories specified this way are searched
766 at the point in which the linker script appears in the command line.
769 @kindex -m @var{emulation}
770 @item -m @var{emulation}
771 Emulate the @var{emulation} linker. You can list the available
772 emulations with the @samp{--verbose} or @samp{-V} options.
774 If the @samp{-m} option is not used, the emulation is taken from the
775 @code{LDEMULATION} environment variable, if that is defined.
777 Otherwise, the default emulation depends upon how the linker was
785 Print a link map to the standard output. A link map provides
786 information about the link, including the following:
790 Where object files are mapped into memory.
792 How common symbols are allocated.
794 All archive members included in the link, with a mention of the symbol
795 which caused the archive member to be brought in.
797 The values assigned to symbols.
799 Note - symbols whose values are computed by an expression which
800 involves a reference to a previous value of the same symbol may not
801 have correct result displayed in the link map. This is because the
802 linker discards intermediate results and only retains the final value
803 of an expression. Under such circumstances the linker will display
804 the final value enclosed by square brackets. Thus for example a
805 linker script containing:
813 will produce the following output in the link map if the @option{-M}
818 [0x0000000c] foo = (foo * 0x4)
819 [0x0000000c] foo = (foo + 0x8)
822 See @ref{Expressions} for more information about expressions in linker
826 How GNU properties are merged.
828 When the linker merges input .note.gnu.property sections into one output
829 .note.gnu.property section, some properties are removed or updated.
830 These actions are reported in the link map. For example:
833 Removed property 0xc0000002 to merge foo.o (0x1) and bar.o (not found)
836 This indicates that property 0xc0000002 is removed from output when
837 merging properties in @file{foo.o}, whose property 0xc0000002 value
838 is 0x1, and @file{bar.o}, which doesn't have property 0xc0000002.
841 Updated property 0xc0010001 (0x1) to merge foo.o (0x1) and bar.o (0x1)
844 This indicates that property 0xc0010001 value is updated to 0x1 in output
845 when merging properties in @file{foo.o}, whose 0xc0010001 property value
846 is 0x1, and @file{bar.o}, whose 0xc0010001 property value is 0x1.
849 @cindex link map discarded
850 @kindex --print-map-discarded
851 @kindex --no-print-map-discarded
852 @item --print-map-discarded
853 @itemx --no-print-map-discarded
854 Print (or do not print) the list of discarded and garbage collected sections
855 in the link map. Enabled by default.
858 @cindex read-only text
863 Turn off page alignment of sections, and disable linking against shared
864 libraries. If the output format supports Unix style magic numbers,
865 mark the output as @code{NMAGIC}.
869 @cindex read/write from cmd line
873 Set the text and data sections to be readable and writable. Also, do
874 not page-align the data segment, and disable linking against shared
875 libraries. If the output format supports Unix style magic numbers,
876 mark the output as @code{OMAGIC}. Note: Although a writable text section
877 is allowed for PE-COFF targets, it does not conform to the format
878 specification published by Microsoft.
883 This option negates most of the effects of the @option{-N} option. It
884 sets the text section to be read-only, and forces the data segment to
885 be page-aligned. Note - this option does not enable linking against
886 shared libraries. Use @option{-Bdynamic} for this.
888 @kindex -o @var{output}
889 @kindex --output=@var{output}
890 @cindex naming the output file
891 @item -o @var{output}
892 @itemx --output=@var{output}
893 Use @var{output} as the name for the program produced by @command{ld}; if this
894 option is not specified, the name @file{a.out} is used by default. The
895 script command @code{OUTPUT} can also specify the output file name.
897 @kindex --dependency-file=@var{depfile}
898 @cindex dependency file
899 @item --dependency-file=@var{depfile}
900 Write a @dfn{dependency file} to @var{depfile}. This file contains a rule
901 suitable for @code{make} describing the output file and all the input files
902 that were read to produce it. The output is similar to the compiler's
903 output with @samp{-M -MP} (@pxref{Preprocessor Options,, Options
904 Controlling the Preprocessor, gcc.info, Using the GNU Compiler
905 Collection}). Note that there is no option like the compiler's @samp{-MM},
906 to exclude ``system files'' (which is not a well-specified concept in the
907 linker, unlike ``system headers'' in the compiler). So the output from
908 @samp{--dependency-file} is always specific to the exact state of the
909 installation where it was produced, and should not be copied into
910 distributed makefiles without careful editing.
912 @kindex -O @var{level}
913 @cindex generating optimized output
915 If @var{level} is a numeric values greater than zero @command{ld} optimizes
916 the output. This might take significantly longer and therefore probably
917 should only be enabled for the final binary. At the moment this
918 option only affects ELF shared library generation. Future releases of
919 the linker may make more use of this option. Also currently there is
920 no difference in the linker's behaviour for different non-zero values
921 of this option. Again this may change with future releases.
923 @kindex -plugin @var{name}
924 @item -plugin @var{name}
925 Involve a plugin in the linking process. The @var{name} parameter is
926 the absolute filename of the plugin. Usually this parameter is
927 automatically added by the complier, when using link time
928 optimization, but users can also add their own plugins if they so
931 Note that the location of the compiler originated plugins is different
932 from the place where the @command{ar}, @command{nm} and
933 @command{ranlib} programs search for their plugins. In order for
934 those commands to make use of a compiler based plugin it must first be
935 copied into the @file{$@{libdir@}/bfd-plugins} directory. All gcc
936 based linker plugins are backward compatible, so it is sufficient to
937 just copy in the newest one.
940 @cindex push state governing input file handling
942 The @option{--push-state} allows one to preserve the current state of the
943 flags which govern the input file handling so that they can all be
944 restored with one corresponding @option{--pop-state} option.
946 The option which are covered are: @option{-Bdynamic}, @option{-Bstatic},
947 @option{-dn}, @option{-dy}, @option{-call_shared}, @option{-non_shared},
948 @option{-static}, @option{-N}, @option{-n}, @option{--whole-archive},
949 @option{--no-whole-archive}, @option{-r}, @option{-Ur},
950 @option{--copy-dt-needed-entries}, @option{--no-copy-dt-needed-entries},
951 @option{--as-needed}, @option{--no-as-needed}, and @option{-a}.
953 One target for this option are specifications for @file{pkg-config}. When
954 used with the @option{--libs} option all possibly needed libraries are
955 listed and then possibly linked with all the time. It is better to return
956 something as follows:
959 -Wl,--push-state,--as-needed -libone -libtwo -Wl,--pop-state
963 @cindex pop state governing input file handling
965 Undoes the effect of --push-state, restores the previous values of the
966 flags governing input file handling.
969 @kindex --emit-relocs
970 @cindex retain relocations in final executable
973 Leave relocation sections and contents in fully linked executables.
974 Post link analysis and optimization tools may need this information in
975 order to perform correct modifications of executables. This results
976 in larger executables.
978 This option is currently only supported on ELF platforms.
980 @kindex --force-dynamic
981 @cindex forcing the creation of dynamic sections
982 @item --force-dynamic
983 Force the output file to have dynamic sections. This option is specific
987 @cindex relocatable output
989 @kindex --relocatable
992 Generate relocatable output---i.e., generate an output file that can in
993 turn serve as input to @command{ld}. This is often called @dfn{partial
994 linking}. As a side effect, in environments that support standard Unix
995 magic numbers, this option also sets the output file's magic number to
997 @c ; see @option{-N}.
998 If this option is not specified, an absolute file is produced. When
999 linking C++ programs, this option @emph{will not} resolve references to
1000 constructors; to do that, use @samp{-Ur}.
1002 When an input file does not have the same format as the output file,
1003 partial linking is only supported if that input file does not contain any
1004 relocations. Different output formats can have further restrictions; for
1005 example some @code{a.out}-based formats do not support partial linking
1006 with input files in other formats at all.
1008 This option does the same thing as @samp{-i}.
1010 @kindex -R @var{file}
1011 @kindex --just-symbols=@var{file}
1012 @cindex symbol-only input
1013 @item -R @var{filename}
1014 @itemx --just-symbols=@var{filename}
1015 Read symbol names and their addresses from @var{filename}, but do not
1016 relocate it or include it in the output. This allows your output file
1017 to refer symbolically to absolute locations of memory defined in other
1018 programs. You may use this option more than once.
1020 For compatibility with other ELF linkers, if the @option{-R} option is
1021 followed by a directory name, rather than a file name, it is treated as
1022 the @option{-rpath} option.
1026 @cindex strip all symbols
1029 Omit all symbol information from the output file.
1032 @kindex --strip-debug
1033 @cindex strip debugger symbols
1035 @itemx --strip-debug
1036 Omit debugger symbol information (but not all symbols) from the output file.
1038 @kindex --strip-discarded
1039 @kindex --no-strip-discarded
1040 @item --strip-discarded
1041 @itemx --no-strip-discarded
1042 Omit (or do not omit) global symbols defined in discarded sections.
1047 @cindex input files, displaying
1050 Print the names of the input files as @command{ld} processes them. If
1051 @samp{-t} is given twice then members within archives are also printed.
1052 @samp{-t} output is useful to generate a list of all the object files
1053 and scripts involved in linking, for example, when packaging files for
1054 a linker bug report.
1056 @kindex -T @var{script}
1057 @kindex --script=@var{script}
1058 @cindex script files
1059 @item -T @var{scriptfile}
1060 @itemx --script=@var{scriptfile}
1061 Use @var{scriptfile} as the linker script. This script replaces
1062 @command{ld}'s default linker script (rather than adding to it), so
1063 @var{commandfile} must specify everything necessary to describe the
1064 output file. @xref{Scripts}. If @var{scriptfile} does not exist in
1065 the current directory, @code{ld} looks for it in the directories
1066 specified by any preceding @samp{-L} options. Multiple @samp{-T}
1069 @kindex -dT @var{script}
1070 @kindex --default-script=@var{script}
1071 @cindex script files
1072 @item -dT @var{scriptfile}
1073 @itemx --default-script=@var{scriptfile}
1074 Use @var{scriptfile} as the default linker script. @xref{Scripts}.
1076 This option is similar to the @option{--script} option except that
1077 processing of the script is delayed until after the rest of the
1078 command line has been processed. This allows options placed after the
1079 @option{--default-script} option on the command line to affect the
1080 behaviour of the linker script, which can be important when the linker
1081 command line cannot be directly controlled by the user. (eg because
1082 the command line is being constructed by another tool, such as
1085 @kindex -u @var{symbol}
1086 @kindex --undefined=@var{symbol}
1087 @cindex undefined symbol
1088 @item -u @var{symbol}
1089 @itemx --undefined=@var{symbol}
1090 Force @var{symbol} to be entered in the output file as an undefined
1091 symbol. Doing this may, for example, trigger linking of additional
1092 modules from standard libraries. @samp{-u} may be repeated with
1093 different option arguments to enter additional undefined symbols. This
1094 option is equivalent to the @code{EXTERN} linker script command.
1096 If this option is being used to force additional modules to be pulled
1097 into the link, and if it is an error for the symbol to remain
1098 undefined, then the option @option{--require-defined} should be used
1101 @kindex --require-defined=@var{symbol}
1102 @cindex symbols, require defined
1103 @cindex defined symbol
1104 @item --require-defined=@var{symbol}
1105 Require that @var{symbol} is defined in the output file. This option
1106 is the same as option @option{--undefined} except that if @var{symbol}
1107 is not defined in the output file then the linker will issue an error
1108 and exit. The same effect can be achieved in a linker script by using
1109 @code{EXTERN}, @code{ASSERT} and @code{DEFINED} together. This option
1110 can be used multiple times to require additional symbols.
1113 @cindex constructors
1115 For anything other than C++ programs, this option is equivalent to
1116 @samp{-r}: it generates relocatable output---i.e., an output file that can in
1117 turn serve as input to @command{ld}. When linking C++ programs, @samp{-Ur}
1118 @emph{does} resolve references to constructors, unlike @samp{-r}.
1119 It does not work to use @samp{-Ur} on files that were themselves linked
1120 with @samp{-Ur}; once the constructor table has been built, it cannot
1121 be added to. Use @samp{-Ur} only for the last partial link, and
1122 @samp{-r} for the others.
1124 @kindex --orphan-handling=@var{MODE}
1125 @cindex orphan sections
1126 @cindex sections, orphan
1127 @item --orphan-handling=@var{MODE}
1128 Control how orphan sections are handled. An orphan section is one not
1129 specifically mentioned in a linker script. @xref{Orphan Sections}.
1131 @var{MODE} can have any of the following values:
1135 Orphan sections are placed into a suitable output section following
1136 the strategy described in @ref{Orphan Sections}. The option
1137 @samp{--unique} also affects how sections are placed.
1140 All orphan sections are discarded, by placing them in the
1141 @samp{/DISCARD/} section (@pxref{Output Section Discarding}).
1144 The linker will place the orphan section as for @code{place} and also
1148 The linker will exit with an error if any orphan section is found.
1151 The default if @samp{--orphan-handling} is not given is @code{place}.
1153 @kindex --unique[=@var{SECTION}]
1154 @item --unique[=@var{SECTION}]
1155 Creates a separate output section for every input section matching
1156 @var{SECTION}, or if the optional wildcard @var{SECTION} argument is
1157 missing, for every orphan input section. An orphan section is one not
1158 specifically mentioned in a linker script. You may use this option
1159 multiple times on the command line; It prevents the normal merging of
1160 input sections with the same name, overriding output section assignments
1170 Display the version number for @command{ld}. The @option{-V} option also
1171 lists the supported emulations.
1174 @kindex --discard-all
1175 @cindex deleting local symbols
1177 @itemx --discard-all
1178 Delete all local symbols.
1181 @kindex --discard-locals
1182 @cindex local symbols, deleting
1184 @itemx --discard-locals
1185 Delete all temporary local symbols. (These symbols start with
1186 system-specific local label prefixes, typically @samp{.L} for ELF systems
1187 or @samp{L} for traditional a.out systems.)
1189 @kindex -y @var{symbol}
1190 @kindex --trace-symbol=@var{symbol}
1191 @cindex symbol tracing
1192 @item -y @var{symbol}
1193 @itemx --trace-symbol=@var{symbol}
1194 Print the name of each linked file in which @var{symbol} appears. This
1195 option may be given any number of times. On many systems it is necessary
1196 to prepend an underscore.
1198 This option is useful when you have an undefined symbol in your link but
1199 don't know where the reference is coming from.
1201 @kindex -Y @var{path}
1203 Add @var{path} to the default library search path. This option exists
1204 for Solaris compatibility.
1206 @kindex -z @var{keyword}
1207 @item -z @var{keyword}
1208 The recognized keywords are:
1212 Always generate BND prefix in PLT entries. Supported for Linux/x86_64.
1214 @item call-nop=prefix-addr
1215 @itemx call-nop=suffix-nop
1216 @itemx call-nop=prefix-@var{byte}
1217 @itemx call-nop=suffix-@var{byte}
1218 Specify the 1-byte @code{NOP} padding when transforming indirect call
1219 to a locally defined function, foo, via its GOT slot.
1220 @option{call-nop=prefix-addr} generates @code{0x67 call foo}.
1221 @option{call-nop=suffix-nop} generates @code{call foo 0x90}.
1222 @option{call-nop=prefix-@var{byte}} generates @code{@var{byte} call foo}.
1223 @option{call-nop=suffix-@var{byte}} generates @code{call foo @var{byte}}.
1224 Supported for i386 and x86_64.
1226 @item cet-report=none
1227 @itemx cet-report=warning
1228 @itemx cet-report=error
1229 Specify how to report the missing GNU_PROPERTY_X86_FEATURE_1_IBT and
1230 GNU_PROPERTY_X86_FEATURE_1_SHSTK properties in input .note.gnu.property
1231 section. @option{cet-report=none}, which is the default, will make the
1232 linker not report missing properties in input files.
1233 @option{cet-report=warning} will make the linker issue a warning for
1234 missing properties in input files. @option{cet-report=error} will make
1235 the linker issue an error for missing properties in input files.
1236 Note that @option{ibt} will turn off the missing
1237 GNU_PROPERTY_X86_FEATURE_1_IBT property report and @option{shstk} will
1238 turn off the missing GNU_PROPERTY_X86_FEATURE_1_SHSTK property report.
1239 Supported for Linux/i386 and Linux/x86_64.
1243 Combine multiple dynamic relocation sections and sort to improve
1244 dynamic symbol lookup caching. Do not do this if @samp{nocombreloc}.
1248 Generate common symbols with STT_COMMON type during a relocatable
1249 link. Use STT_OBJECT type if @samp{nocommon}.
1251 @item common-page-size=@var{value}
1252 Set the page size most commonly used to @var{value}. Memory image
1253 layout will be optimized to minimize memory pages if the system is
1254 using pages of this size.
1257 Report unresolved symbol references from regular object files. This
1258 is done even if the linker is creating a non-symbolic shared library.
1259 This option is the inverse of @samp{-z undefs}.
1261 @item dynamic-undefined-weak
1262 @itemx nodynamic-undefined-weak
1263 Make undefined weak symbols dynamic when building a dynamic object,
1264 if they are referenced from a regular object file and not forced local
1265 by symbol visibility or versioning. Do not make them dynamic if
1266 @samp{nodynamic-undefined-weak}. If neither option is given, a target
1267 may default to either option being in force, or make some other
1268 selection of undefined weak symbols dynamic. Not all targets support
1272 Marks the object as requiring executable stack.
1275 This option is only meaningful when building a shared object. It makes
1276 the symbols defined by this shared object available for symbol resolution
1277 of subsequently loaded libraries.
1280 This option is only meaningful when building a dynamic executable.
1281 This option marks the executable as requiring global auditing by
1282 setting the @code{DF_1_GLOBAUDIT} bit in the @code{DT_FLAGS_1} dynamic
1283 tag. Global auditing requires that any auditing library defined via
1284 the @option{--depaudit} or @option{-P} command-line options be run for
1285 all dynamic objects loaded by the application.
1288 Generate Intel Indirect Branch Tracking (IBT) enabled PLT entries.
1289 Supported for Linux/i386 and Linux/x86_64.
1292 Generate GNU_PROPERTY_X86_FEATURE_1_IBT in .note.gnu.property section
1293 to indicate compatibility with IBT. This also implies @option{ibtplt}.
1294 Supported for Linux/i386 and Linux/x86_64.
1297 This option is only meaningful when building a shared object.
1298 It marks the object so that its runtime initialization will occur
1299 before the runtime initialization of any other objects brought into
1300 the process at the same time. Similarly the runtime finalization of
1301 the object will occur after the runtime finalization of any other
1305 Specify that the dynamic loader should modify its symbol search order
1306 so that symbols in this shared library interpose all other shared
1307 libraries not so marked.
1311 When generating a shared library or other dynamically loadable ELF
1312 object mark it as one that should (by default) only ever be loaded once,
1313 and only in the main namespace (when using @code{dlmopen}). This is
1314 primarily used to mark fundamental libraries such as libc, libpthread et
1315 al which do not usually function correctly unless they are the sole instances
1316 of themselves. This behaviour can be overridden by the @code{dlmopen} caller
1317 and does not apply to certain loading mechanisms (such as audit libraries).
1320 Generate GNU_PROPERTY_X86_FEATURE_1_LAM_U48 in .note.gnu.property section
1321 to indicate compatibility with Intel LAM_U48. Supported for Linux/x86_64.
1324 Generate GNU_PROPERTY_X86_FEATURE_1_LAM_U57 in .note.gnu.property section
1325 to indicate compatibility with Intel LAM_U57. Supported for Linux/x86_64.
1327 @item lam-u48-report=none
1328 @itemx lam-u48-report=warning
1329 @itemx lam-u48-report=error
1330 Specify how to report the missing GNU_PROPERTY_X86_FEATURE_1_LAM_U48
1331 property in input .note.gnu.property section.
1332 @option{lam-u48-report=none}, which is the default, will make the
1333 linker not report missing properties in input files.
1334 @option{lam-u48-report=warning} will make the linker issue a warning for
1335 missing properties in input files. @option{lam-u48-report=error} will
1336 make the linker issue an error for missing properties in input files.
1337 Supported for Linux/x86_64.
1339 @item lam-u57-report=none
1340 @itemx lam-u57-report=warning
1341 @itemx lam-u57-report=error
1342 Specify how to report the missing GNU_PROPERTY_X86_FEATURE_1_LAM_U57
1343 property in input .note.gnu.property section.
1344 @option{lam-u57-report=none}, which is the default, will make the
1345 linker not report missing properties in input files.
1346 @option{lam-u57-report=warning} will make the linker issue a warning for
1347 missing properties in input files. @option{lam-u57-report=error} will
1348 make the linker issue an error for missing properties in input files.
1349 Supported for Linux/x86_64.
1351 @item lam-report=none
1352 @itemx lam-report=warning
1353 @itemx lam-report=error
1354 Specify how to report the missing GNU_PROPERTY_X86_FEATURE_1_LAM_U48 and
1355 GNU_PROPERTY_X86_FEATURE_1_LAM_U57 properties in input .note.gnu.property
1356 section. @option{lam-report=none}, which is the default, will make the
1357 linker not report missing properties in input files.
1358 @option{lam-report=warning} will make the linker issue a warning for
1359 missing properties in input files. @option{lam-report=error} will make
1360 the linker issue an error for missing properties in input files.
1361 Supported for Linux/x86_64.
1364 When generating an executable or shared library, mark it to tell the
1365 dynamic linker to defer function call resolution to the point when
1366 the function is called (lazy binding), rather than at load time.
1367 Lazy binding is the default.
1370 Specify that the object's filters be processed immediately at runtime.
1372 @item max-page-size=@var{value}
1373 Set the maximum memory page size supported to @var{value}.
1376 Allow multiple definitions.
1379 Disable linker generated .dynbss variables used in place of variables
1380 defined in shared libraries. May result in dynamic text relocations.
1383 Specify that the dynamic loader search for dependencies of this object
1384 should ignore any default library search paths.
1387 Specify that the object shouldn't be unloaded at runtime.
1390 Specify that the object is not available to @code{dlopen}.
1393 Specify that the object can not be dumped by @code{dldump}.
1396 Marks the object as not requiring executable stack.
1398 @item noextern-protected-data
1399 Don't treat protected data symbols as external when building a shared
1400 library. This option overrides the linker backend default. It can be
1401 used to work around incorrect relocations against protected data symbols
1402 generated by compiler. Updates on protected data symbols by another
1403 module aren't visible to the resulting shared library. Supported for
1406 @item noreloc-overflow
1407 Disable relocation overflow check. This can be used to disable
1408 relocation overflow check if there will be no dynamic relocation
1409 overflow at run-time. Supported for x86_64.
1412 When generating an executable or shared library, mark it to tell the
1413 dynamic linker to resolve all symbols when the program is started, or
1414 when the shared library is loaded by dlopen, instead of deferring
1415 function call resolution to the point when the function is first
1419 Specify that the object requires @samp{$ORIGIN} handling in paths.
1423 Create an ELF @code{PT_GNU_RELRO} segment header in the object. This
1424 specifies a memory segment that should be made read-only after
1425 relocation, if supported. Specifying @samp{common-page-size} smaller
1426 than the system page size will render this protection ineffective.
1427 Don't create an ELF @code{PT_GNU_RELRO} segment if @samp{norelro}.
1429 @item report-relative-reloc
1430 Report dynamic relative relocations generated by linker. Supported for
1431 Linux/i386 and Linux/x86_64.
1434 @itemx noseparate-code
1435 Create separate code @code{PT_LOAD} segment header in the object. This
1436 specifies a memory segment that should contain only instructions and must
1437 be in wholly disjoint pages from any other data. Don't create separate
1438 code @code{PT_LOAD} segment if @samp{noseparate-code} is used.
1441 Generate GNU_PROPERTY_X86_FEATURE_1_SHSTK in .note.gnu.property section
1442 to indicate compatibility with Intel Shadow Stack. Supported for
1443 Linux/i386 and Linux/x86_64.
1445 @item stack-size=@var{value}
1446 Specify a stack size for an ELF @code{PT_GNU_STACK} segment.
1447 Specifying zero will override any default non-zero sized
1448 @code{PT_GNU_STACK} segment creation.
1451 @itemx nostart-stop-gc
1452 @cindex start-stop-gc
1453 When @samp{--gc-sections} is in effect, a reference from a retained
1454 section to @code{__start_SECNAME} or @code{__stop_SECNAME} causes all
1455 input sections named @code{SECNAME} to also be retained, if
1456 @code{SECNAME} is representable as a C identifier and either
1457 @code{__start_SECNAME} or @code{__stop_SECNAME} is synthesized by the
1458 linker. @samp{-z start-stop-gc} disables this effect, allowing
1459 sections to be garbage collected as if the special synthesized symbols
1460 were not defined. @samp{-z start-stop-gc} has no effect on a
1461 definition of @code{__start_SECNAME} or @code{__stop_SECNAME} in an
1462 object file or linker script. Such a definition will prevent the
1463 linker providing a synthesized @code{__start_SECNAME} or
1464 @code{__stop_SECNAME} respectively, and therefore the special
1465 treatment by garbage collection for those references.
1467 @item start-stop-visibility=@var{value}
1469 @cindex ELF symbol visibility
1470 Specify the ELF symbol visibility for synthesized
1471 @code{__start_SECNAME} and @code{__stop_SECNAME} symbols (@pxref{Input
1472 Section Example}). @var{value} must be exactly @samp{default},
1473 @samp{internal}, @samp{hidden}, or @samp{protected}. If no @samp{-z
1474 start-stop-visibility} option is given, @samp{protected} is used for
1475 compatibility with historical practice. However, it's highly
1476 recommended to use @samp{-z start-stop-visibility=hidden} in new
1477 programs and shared libraries so that these symbols are not exported
1478 between shared objects, which is not usually what's intended.
1483 Report an error if DT_TEXTREL is set, i.e., if the position-independent
1484 or shared object has dynamic relocations in read-only sections. Don't
1485 report an error if @samp{notext} or @samp{textoff}.
1488 Do not report unresolved symbol references from regular object files,
1489 either when creating an executable, or when creating a shared library.
1490 This option is the inverse of @samp{-z defs}.
1493 @itemx nounique-symbol
1494 Avoid duplicated local symbol names in the symbol string table. Append
1495 ".@code{number}" to duplicated local symbol names if @samp{unique-symbol}
1496 is used. @option{nounique-symbol} is the default.
1498 @item x86-64-baseline
1502 Specify the x86-64 ISA level needed in .note.gnu.property section.
1503 @option{x86-64-baseline} generates @code{GNU_PROPERTY_X86_ISA_1_BASELINE}.
1504 @option{x86-64-v2} generates @code{GNU_PROPERTY_X86_ISA_1_V2}.
1505 @option{x86-64-v3} generates @code{GNU_PROPERTY_X86_ISA_1_V3}.
1506 @option{x86-64-v4} generates @code{GNU_PROPERTY_X86_ISA_1_V4}.
1507 Supported for Linux/i386 and Linux/x86_64.
1511 Other keywords are ignored for Solaris compatibility.
1514 @cindex groups of archives
1515 @item -( @var{archives} -)
1516 @itemx --start-group @var{archives} --end-group
1517 The @var{archives} should be a list of archive files. They may be
1518 either explicit file names, or @samp{-l} options.
1520 The specified archives are searched repeatedly until no new undefined
1521 references are created. Normally, an archive is searched only once in
1522 the order that it is specified on the command line. If a symbol in that
1523 archive is needed to resolve an undefined symbol referred to by an
1524 object in an archive that appears later on the command line, the linker
1525 would not be able to resolve that reference. By grouping the archives,
1526 they will all be searched repeatedly until all possible references are
1529 Using this option has a significant performance cost. It is best to use
1530 it only when there are unavoidable circular references between two or
1533 @kindex --accept-unknown-input-arch
1534 @kindex --no-accept-unknown-input-arch
1535 @item --accept-unknown-input-arch
1536 @itemx --no-accept-unknown-input-arch
1537 Tells the linker to accept input files whose architecture cannot be
1538 recognised. The assumption is that the user knows what they are doing
1539 and deliberately wants to link in these unknown input files. This was
1540 the default behaviour of the linker, before release 2.14. The default
1541 behaviour from release 2.14 onwards is to reject such input files, and
1542 so the @samp{--accept-unknown-input-arch} option has been added to
1543 restore the old behaviour.
1546 @kindex --no-as-needed
1548 @itemx --no-as-needed
1549 This option affects ELF DT_NEEDED tags for dynamic libraries mentioned
1550 on the command line after the @option{--as-needed} option. Normally
1551 the linker will add a DT_NEEDED tag for each dynamic library mentioned
1552 on the command line, regardless of whether the library is actually
1553 needed or not. @option{--as-needed} causes a DT_NEEDED tag to only be
1554 emitted for a library that @emph{at that point in the link} satisfies a
1555 non-weak undefined symbol reference from a regular object file or, if
1556 the library is not found in the DT_NEEDED lists of other needed libraries, a
1557 non-weak undefined symbol reference from another needed dynamic library.
1558 Object files or libraries appearing on the command line @emph{after}
1559 the library in question do not affect whether the library is seen as
1560 needed. This is similar to the rules for extraction of object files
1561 from archives. @option{--no-as-needed} restores the default behaviour.
1563 Note: On Linux based systems the @option{--as-needed} option also has
1564 an affect on the behaviour of the @option{--rpath} and
1565 @option{--rpath-link} options. See the description of
1566 @option{--rpath-link} for more details.
1568 @kindex --add-needed
1569 @kindex --no-add-needed
1571 @itemx --no-add-needed
1572 These two options have been deprecated because of the similarity of
1573 their names to the @option{--as-needed} and @option{--no-as-needed}
1574 options. They have been replaced by @option{--copy-dt-needed-entries}
1575 and @option{--no-copy-dt-needed-entries}.
1577 @kindex -assert @var{keyword}
1578 @item -assert @var{keyword}
1579 This option is ignored for SunOS compatibility.
1583 @kindex -call_shared
1587 Link against dynamic libraries. This is only meaningful on platforms
1588 for which shared libraries are supported. This option is normally the
1589 default on such platforms. The different variants of this option are
1590 for compatibility with various systems. You may use this option
1591 multiple times on the command line: it affects library searching for
1592 @option{-l} options which follow it.
1596 Set the @code{DF_1_GROUP} flag in the @code{DT_FLAGS_1} entry in the dynamic
1597 section. This causes the runtime linker to handle lookups in this
1598 object and its dependencies to be performed only inside the group.
1599 @option{--unresolved-symbols=report-all} is implied. This option is
1600 only meaningful on ELF platforms which support shared libraries.
1610 Do not link against shared libraries. This is only meaningful on
1611 platforms for which shared libraries are supported. The different
1612 variants of this option are for compatibility with various systems. You
1613 may use this option multiple times on the command line: it affects
1614 library searching for @option{-l} options which follow it. This
1615 option also implies @option{--unresolved-symbols=report-all}. This
1616 option can be used with @option{-shared}. Doing so means that a
1617 shared library is being created but that all of the library's external
1618 references must be resolved by pulling in entries from static
1623 When creating a shared library, bind references to global symbols to the
1624 definition within the shared library, if any. Normally, it is possible
1625 for a program linked against a shared library to override the definition
1626 within the shared library. This option is only meaningful on ELF
1627 platforms which support shared libraries.
1629 @kindex -Bsymbolic-functions
1630 @item -Bsymbolic-functions
1631 When creating a shared library, bind references to global function
1632 symbols to the definition within the shared library, if any.
1633 This option is only meaningful on ELF platforms which support shared
1636 @kindex -Bno-symbolic
1638 This option can cancel previously specified @samp{-Bsymbolic} and
1639 @samp{-Bsymbolic-functions}.
1641 @kindex --dynamic-list=@var{dynamic-list-file}
1642 @item --dynamic-list=@var{dynamic-list-file}
1643 Specify the name of a dynamic list file to the linker. This is
1644 typically used when creating shared libraries to specify a list of
1645 global symbols whose references shouldn't be bound to the definition
1646 within the shared library, or creating dynamically linked executables
1647 to specify a list of symbols which should be added to the symbol table
1648 in the executable. This option is only meaningful on ELF platforms
1649 which support shared libraries.
1651 The format of the dynamic list is the same as the version node without
1652 scope and node name. See @ref{VERSION} for more information.
1654 @kindex --dynamic-list-data
1655 @item --dynamic-list-data
1656 Include all global data symbols to the dynamic list.
1658 @kindex --dynamic-list-cpp-new
1659 @item --dynamic-list-cpp-new
1660 Provide the builtin dynamic list for C++ operator new and delete. It
1661 is mainly useful for building shared libstdc++.
1663 @kindex --dynamic-list-cpp-typeinfo
1664 @item --dynamic-list-cpp-typeinfo
1665 Provide the builtin dynamic list for C++ runtime type identification.
1667 @kindex --check-sections
1668 @kindex --no-check-sections
1669 @item --check-sections
1670 @itemx --no-check-sections
1671 Asks the linker @emph{not} to check section addresses after they have
1672 been assigned to see if there are any overlaps. Normally the linker will
1673 perform this check, and if it finds any overlaps it will produce
1674 suitable error messages. The linker does know about, and does make
1675 allowances for sections in overlays. The default behaviour can be
1676 restored by using the command-line switch @option{--check-sections}.
1677 Section overlap is not usually checked for relocatable links. You can
1678 force checking in that case by using the @option{--check-sections}
1681 @kindex --copy-dt-needed-entries
1682 @kindex --no-copy-dt-needed-entries
1683 @item --copy-dt-needed-entries
1684 @itemx --no-copy-dt-needed-entries
1685 This option affects the treatment of dynamic libraries referred to
1686 by DT_NEEDED tags @emph{inside} ELF dynamic libraries mentioned on the
1687 command line. Normally the linker won't add a DT_NEEDED tag to the
1688 output binary for each library mentioned in a DT_NEEDED tag in an
1689 input dynamic library. With @option{--copy-dt-needed-entries}
1690 specified on the command line however any dynamic libraries that
1691 follow it will have their DT_NEEDED entries added. The default
1692 behaviour can be restored with @option{--no-copy-dt-needed-entries}.
1694 This option also has an effect on the resolution of symbols in dynamic
1695 libraries. With @option{--copy-dt-needed-entries} dynamic libraries
1696 mentioned on the command line will be recursively searched, following
1697 their DT_NEEDED tags to other libraries, in order to resolve symbols
1698 required by the output binary. With the default setting however
1699 the searching of dynamic libraries that follow it will stop with the
1700 dynamic library itself. No DT_NEEDED links will be traversed to resolve
1703 @cindex cross reference table
1706 Output a cross reference table. If a linker map file is being
1707 generated, the cross reference table is printed to the map file.
1708 Otherwise, it is printed on the standard output.
1710 The format of the table is intentionally simple, so that it may be
1711 easily processed by a script if necessary. The symbols are printed out,
1712 sorted by name. For each symbol, a list of file names is given. If the
1713 symbol is defined, the first file listed is the location of the
1714 definition. If the symbol is defined as a common value then any files
1715 where this happens appear next. Finally any files that reference the
1718 @cindex ctf variables
1719 @kindex --ctf-variables
1720 @kindex --no-ctf-variables
1721 @item --ctf-variables
1722 @item --no-ctf-variables
1723 The CTF debuginfo format supports a section which encodes the names and
1724 types of variables found in the program which do not appear in any symbol
1725 table. These variables clearly cannot be looked up by address by
1726 conventional debuggers, so the space used for their types and names is
1727 usually wasted: the types are usually small but the names are often not.
1728 @option{--ctf-variables} causes the generation of such a section.
1729 The default behaviour can be restored with @option{--no-ctf-variables}.
1731 @cindex ctf type sharing
1732 @kindex --ctf-share-types
1733 @item --ctf-share-types=@var{method}
1734 Adjust the method used to share types between translation units in CTF.
1737 @item share-unconflicted
1738 Put all types that do not have ambiguous definitions into the shared dictionary,
1739 where debuggers can easily access them, even if they only occur in one
1740 translation unit. This is the default.
1742 @item share-duplicated
1743 Put only types that occur in multiple translation units into the shared
1744 dictionary: types with only one definition go into per-translation-unit
1745 dictionaries. Types with ambiguous definitions in multiple translation units
1746 always go into per-translation-unit dictionaries. This tends to make the CTF
1747 larger, but may reduce the amount of CTF in the shared dictionary. For very
1748 large projects this may speed up opening the CTF and save memory in the CTF
1749 consumer at runtime.
1752 @cindex common allocation
1753 @kindex --no-define-common
1754 @item --no-define-common
1755 This option inhibits the assignment of addresses to common symbols.
1756 The script command @code{INHIBIT_COMMON_ALLOCATION} has the same effect.
1757 @xref{Miscellaneous Commands}.
1759 The @samp{--no-define-common} option allows decoupling
1760 the decision to assign addresses to Common symbols from the choice
1761 of the output file type; otherwise a non-Relocatable output type
1762 forces assigning addresses to Common symbols.
1763 Using @samp{--no-define-common} allows Common symbols that are referenced
1764 from a shared library to be assigned addresses only in the main program.
1765 This eliminates the unused duplicate space in the shared library,
1766 and also prevents any possible confusion over resolving to the wrong
1767 duplicate when there are many dynamic modules with specialized search
1768 paths for runtime symbol resolution.
1770 @cindex group allocation in linker script
1771 @cindex section groups
1773 @kindex --force-group-allocation
1774 @item --force-group-allocation
1775 This option causes the linker to place section group members like
1776 normal input sections, and to delete the section groups. This is the
1777 default behaviour for a final link but this option can be used to
1778 change the behaviour of a relocatable link (@samp{-r}). The script
1779 command @code{FORCE_GROUP_ALLOCATION} has the same
1780 effect. @xref{Miscellaneous Commands}.
1782 @cindex symbols, from command line
1783 @kindex --defsym=@var{symbol}=@var{exp}
1784 @item --defsym=@var{symbol}=@var{expression}
1785 Create a global symbol in the output file, containing the absolute
1786 address given by @var{expression}. You may use this option as many
1787 times as necessary to define multiple symbols in the command line. A
1788 limited form of arithmetic is supported for the @var{expression} in this
1789 context: you may give a hexadecimal constant or the name of an existing
1790 symbol, or use @code{+} and @code{-} to add or subtract hexadecimal
1791 constants or symbols. If you need more elaborate expressions, consider
1792 using the linker command language from a script (@pxref{Assignments}).
1793 @emph{Note:} there should be no white space between @var{symbol}, the
1794 equals sign (``@key{=}''), and @var{expression}.
1796 The linker processes @samp{--defsym} arguments and @samp{-T} arguments
1797 in order, placing @samp{--defsym} before @samp{-T} will define the
1798 symbol before the linker script from @samp{-T} is processed, while
1799 placing @samp{--defsym} after @samp{-T} will define the symbol after
1800 the linker script has been processed. This difference has
1801 consequences for expressions within the linker script that use the
1802 @samp{--defsym} symbols, which order is correct will depend on what
1803 you are trying to achieve.
1805 @cindex demangling, from command line
1806 @kindex --demangle[=@var{style}]
1807 @kindex --no-demangle
1808 @item --demangle[=@var{style}]
1809 @itemx --no-demangle
1810 These options control whether to demangle symbol names in error messages
1811 and other output. When the linker is told to demangle, it tries to
1812 present symbol names in a readable fashion: it strips leading
1813 underscores if they are used by the object file format, and converts C++
1814 mangled symbol names into user readable names. Different compilers have
1815 different mangling styles. The optional demangling style argument can be used
1816 to choose an appropriate demangling style for your compiler. The linker will
1817 demangle by default unless the environment variable @samp{COLLECT_NO_DEMANGLE}
1818 is set. These options may be used to override the default.
1820 @cindex dynamic linker, from command line
1821 @kindex -I@var{file}
1822 @kindex --dynamic-linker=@var{file}
1824 @itemx --dynamic-linker=@var{file}
1825 Set the name of the dynamic linker. This is only meaningful when
1826 generating dynamically linked ELF executables. The default dynamic
1827 linker is normally correct; don't use this unless you know what you are
1830 @kindex --no-dynamic-linker
1831 @item --no-dynamic-linker
1832 When producing an executable file, omit the request for a dynamic
1833 linker to be used at load-time. This is only meaningful for ELF
1834 executables that contain dynamic relocations, and usually requires
1835 entry point code that is capable of processing these relocations.
1837 @kindex --embedded-relocs
1838 @item --embedded-relocs
1839 This option is similar to the @option{--emit-relocs} option except
1840 that the relocs are stored in a target-specific section. This option
1841 is only supported by the @samp{BFIN}, @samp{CR16} and @emph{M68K}
1844 @kindex --disable-multiple-abs-defs
1845 @item --disable-multiple-abs-defs
1846 Do not allow multiple definitions with symbols included
1847 in filename invoked by -R or --just-symbols
1849 @kindex --fatal-warnings
1850 @kindex --no-fatal-warnings
1851 @item --fatal-warnings
1852 @itemx --no-fatal-warnings
1853 Treat all warnings as errors. The default behaviour can be restored
1854 with the option @option{--no-fatal-warnings}.
1856 @kindex --force-exe-suffix
1857 @item --force-exe-suffix
1858 Make sure that an output file has a .exe suffix.
1860 If a successfully built fully linked output file does not have a
1861 @code{.exe} or @code{.dll} suffix, this option forces the linker to copy
1862 the output file to one of the same name with a @code{.exe} suffix. This
1863 option is useful when using unmodified Unix makefiles on a Microsoft
1864 Windows host, since some versions of Windows won't run an image unless
1865 it ends in a @code{.exe} suffix.
1867 @kindex --gc-sections
1868 @kindex --no-gc-sections
1869 @cindex garbage collection
1871 @itemx --no-gc-sections
1872 Enable garbage collection of unused input sections. It is ignored on
1873 targets that do not support this option. The default behaviour (of not
1874 performing this garbage collection) can be restored by specifying
1875 @samp{--no-gc-sections} on the command line. Note that garbage
1876 collection for COFF and PE format targets is supported, but the
1877 implementation is currently considered to be experimental.
1879 @samp{--gc-sections} decides which input sections are used by
1880 examining symbols and relocations. The section containing the entry
1881 symbol and all sections containing symbols undefined on the
1882 command-line will be kept, as will sections containing symbols
1883 referenced by dynamic objects. Note that when building shared
1884 libraries, the linker must assume that any visible symbol is
1885 referenced. Once this initial set of sections has been determined,
1886 the linker recursively marks as used any section referenced by their
1887 relocations. See @samp{--entry}, @samp{--undefined}, and
1888 @samp{--gc-keep-exported}.
1890 This option can be set when doing a partial link (enabled with option
1891 @samp{-r}). In this case the root of symbols kept must be explicitly
1892 specified either by one of the options @samp{--entry},
1893 @samp{--undefined}, or @samp{--gc-keep-exported} or by a @code{ENTRY}
1894 command in the linker script.
1896 As a GNU extension, ELF input sections marked with the
1897 @code{SHF_GNU_RETAIN} flag will not be garbage collected.
1899 @kindex --print-gc-sections
1900 @kindex --no-print-gc-sections
1901 @cindex garbage collection
1902 @item --print-gc-sections
1903 @itemx --no-print-gc-sections
1904 List all sections removed by garbage collection. The listing is
1905 printed on stderr. This option is only effective if garbage
1906 collection has been enabled via the @samp{--gc-sections}) option. The
1907 default behaviour (of not listing the sections that are removed) can
1908 be restored by specifying @samp{--no-print-gc-sections} on the command
1911 @kindex --gc-keep-exported
1912 @cindex garbage collection
1913 @item --gc-keep-exported
1914 When @samp{--gc-sections} is enabled, this option prevents garbage
1915 collection of unused input sections that contain global symbols having
1916 default or protected visibility. This option is intended to be used for
1917 executables where unreferenced sections would otherwise be garbage
1918 collected regardless of the external visibility of contained symbols.
1919 Note that this option has no effect when linking shared objects since
1920 it is already the default behaviour. This option is only supported for
1923 @kindex --print-output-format
1924 @cindex output format
1925 @item --print-output-format
1926 Print the name of the default output format (perhaps influenced by
1927 other command-line options). This is the string that would appear
1928 in an @code{OUTPUT_FORMAT} linker script command (@pxref{File Commands}).
1930 @kindex --print-memory-usage
1931 @cindex memory usage
1932 @item --print-memory-usage
1933 Print used size, total size and used size of memory regions created with
1934 the @ref{MEMORY} command. This is useful on embedded targets to have a
1935 quick view of amount of free memory. The format of the output has one
1936 headline and one line per region. It is both human readable and easily
1937 parsable by tools. Here is an example of an output:
1940 Memory region Used Size Region Size %age Used
1941 ROM: 256 KB 1 MB 25.00%
1942 RAM: 32 B 2 GB 0.00%
1949 Print a summary of the command-line options on the standard output and exit.
1951 @kindex --target-help
1953 Print a summary of all target-specific options on the standard output and exit.
1955 @kindex -Map=@var{mapfile}
1956 @item -Map=@var{mapfile}
1957 Print a link map to the file @var{mapfile}. See the description of the
1958 @option{-M} option, above. If @var{mapfile} is just the character
1959 @code{-} then the map will be written to stdout.
1961 Specifying a directory as @var{mapfile} causes the linker map to be
1962 written as a file inside the directory. Normally name of the file
1963 inside the directory is computed as the basename of the @var{output}
1964 file with @code{.map} appended. If however the special character
1965 @code{%} is used then this will be replaced by the full path of the
1966 output file. Additionally if there are any characters after the
1967 @var{%} symbol then @code{.map} will no longer be appended.
1970 -o foo.exe -Map=bar [Creates ./bar]
1971 -o ../dir/foo.exe -Map=bar [Creates ./bar]
1972 -o foo.exe -Map=../dir [Creates ../dir/foo.exe.map]
1973 -o ../dir2/foo.exe -Map=../dir [Creates ../dir/foo.exe.map]
1974 -o foo.exe -Map=% [Creates ./foo.exe.map]
1975 -o ../dir/foo.exe -Map=% [Creates ../dir/foo.exe.map]
1976 -o foo.exe -Map=%.bar [Creates ./foo.exe.bar]
1977 -o ../dir/foo.exe -Map=%.bar [Creates ../dir/foo.exe.bar]
1978 -o ../dir2/foo.exe -Map=../dir/% [Creates ../dir/../dir2/foo.exe.map]
1979 -o ../dir2/foo.exe -Map=../dir/%.bar [Creates ../dir/../dir2/foo.exe.bar]
1982 It is an error to specify more than one @code{%} character.
1984 If the map file already exists then it will be overwritten by this
1987 @cindex memory usage
1988 @kindex --no-keep-memory
1989 @item --no-keep-memory
1990 @command{ld} normally optimizes for speed over memory usage by caching the
1991 symbol tables of input files in memory. This option tells @command{ld} to
1992 instead optimize for memory usage, by rereading the symbol tables as
1993 necessary. This may be required if @command{ld} runs out of memory space
1994 while linking a large executable.
1996 @kindex --no-undefined
1999 @item --no-undefined
2001 Report unresolved symbol references from regular object files. This
2002 is done even if the linker is creating a non-symbolic shared library.
2003 The switch @option{--[no-]allow-shlib-undefined} controls the
2004 behaviour for reporting unresolved references found in shared
2005 libraries being linked in.
2007 The effects of this option can be reverted by using @code{-z undefs}.
2009 @kindex --allow-multiple-definition
2011 @item --allow-multiple-definition
2013 Normally when a symbol is defined multiple times, the linker will
2014 report a fatal error. These options allow multiple definitions and the
2015 first definition will be used.
2017 @kindex --allow-shlib-undefined
2018 @kindex --no-allow-shlib-undefined
2019 @item --allow-shlib-undefined
2020 @itemx --no-allow-shlib-undefined
2021 Allows or disallows undefined symbols in shared libraries.
2022 This switch is similar to @option{--no-undefined} except that it
2023 determines the behaviour when the undefined symbols are in a
2024 shared library rather than a regular object file. It does not affect
2025 how undefined symbols in regular object files are handled.
2027 The default behaviour is to report errors for any undefined symbols
2028 referenced in shared libraries if the linker is being used to create
2029 an executable, but to allow them if the linker is being used to create
2032 The reasons for allowing undefined symbol references in shared
2033 libraries specified at link time are that:
2037 A shared library specified at link time may not be the same as the one
2038 that is available at load time, so the symbol might actually be
2039 resolvable at load time.
2041 There are some operating systems, eg BeOS and HPPA, where undefined
2042 symbols in shared libraries are normal.
2044 The BeOS kernel for example patches shared libraries at load time to
2045 select whichever function is most appropriate for the current
2046 architecture. This is used, for example, to dynamically select an
2047 appropriate memset function.
2050 @kindex --error-handling-script=@var{scriptname}
2051 @item --error-handling-script=@var{scriptname}
2052 If this option is provided then the linker will invoke
2053 @var{scriptname} whenever an error is encountered. Currently however
2054 only two kinds of error are supported: missing symbols and missing
2055 libraries. Two arguments will be passed to script: the keyword
2056 ``undefined-symbol'' or `missing-lib'' and the @var{name} of the
2057 undefined symbol or missing library. The intention is that the script
2058 will provide suggestions to the user as to where the symbol or library
2059 might be found. After the script has finished then the normal linker
2060 error message will be displayed.
2062 The availability of this option is controlled by a configure time
2063 switch, so it may not be present in specific implementations.
2065 @kindex --no-undefined-version
2066 @item --no-undefined-version
2067 Normally when a symbol has an undefined version, the linker will ignore
2068 it. This option disallows symbols with undefined version and a fatal error
2069 will be issued instead.
2071 @kindex --default-symver
2072 @item --default-symver
2073 Create and use a default symbol version (the soname) for unversioned
2076 @kindex --default-imported-symver
2077 @item --default-imported-symver
2078 Create and use a default symbol version (the soname) for unversioned
2081 @kindex --no-warn-mismatch
2082 @item --no-warn-mismatch
2083 Normally @command{ld} will give an error if you try to link together input
2084 files that are mismatched for some reason, perhaps because they have
2085 been compiled for different processors or for different endiannesses.
2086 This option tells @command{ld} that it should silently permit such possible
2087 errors. This option should only be used with care, in cases when you
2088 have taken some special action that ensures that the linker errors are
2091 @kindex --no-warn-search-mismatch
2092 @item --no-warn-search-mismatch
2093 Normally @command{ld} will give a warning if it finds an incompatible
2094 library during a library search. This option silences the warning.
2096 @kindex --no-whole-archive
2097 @item --no-whole-archive
2098 Turn off the effect of the @option{--whole-archive} option for subsequent
2101 @cindex output file after errors
2102 @kindex --noinhibit-exec
2103 @item --noinhibit-exec
2104 Retain the executable output file whenever it is still usable.
2105 Normally, the linker will not produce an output file if it encounters
2106 errors during the link process; it exits without writing an output file
2107 when it issues any error whatsoever.
2111 Only search library directories explicitly specified on the
2112 command line. Library directories specified in linker scripts
2113 (including linker scripts specified on the command line) are ignored.
2115 @ifclear SingleFormat
2116 @kindex --oformat=@var{output-format}
2117 @item --oformat=@var{output-format}
2118 @command{ld} may be configured to support more than one kind of object
2119 file. If your @command{ld} is configured this way, you can use the
2120 @samp{--oformat} option to specify the binary format for the output
2121 object file. Even when @command{ld} is configured to support alternative
2122 object formats, you don't usually need to specify this, as @command{ld}
2123 should be configured to produce as a default output format the most
2124 usual format on each machine. @var{output-format} is a text string, the
2125 name of a particular format supported by the BFD libraries. (You can
2126 list the available binary formats with @samp{objdump -i}.) The script
2127 command @code{OUTPUT_FORMAT} can also specify the output format, but
2128 this option overrides it. @xref{BFD}.
2131 @kindex --out-implib
2132 @item --out-implib @var{file}
2133 Create an import library in @var{file} corresponding to the executable
2134 the linker is generating (eg. a DLL or ELF program). This import
2135 library (which should be called @code{*.dll.a} or @code{*.a} for DLLs)
2136 may be used to link clients against the generated executable; this
2137 behaviour makes it possible to skip a separate import library creation
2138 step (eg. @code{dlltool} for DLLs). This option is only available for
2139 the i386 PE and ELF targetted ports of the linker.
2142 @kindex --pic-executable
2144 @itemx --pic-executable
2145 @cindex position independent executables
2146 Create a position independent executable. This is currently only supported on
2147 ELF platforms. Position independent executables are similar to shared
2148 libraries in that they are relocated by the dynamic linker to the virtual
2149 address the OS chooses for them (which can vary between invocations). Like
2150 normal dynamically linked executables they can be executed and symbols
2151 defined in the executable cannot be overridden by shared libraries.
2155 @cindex position dependent executables
2156 Create a position dependent executable. This is the default.
2160 This option is ignored for Linux compatibility.
2164 This option is ignored for SVR4 compatibility.
2167 @cindex synthesizing linker
2168 @cindex relaxing addressing modes
2172 An option with machine dependent effects.
2174 This option is only supported on a few targets.
2177 @xref{H8/300,,@command{ld} and the H8/300}.
2180 @xref{Xtensa,, @command{ld} and Xtensa Processors}.
2183 @xref{M68HC11/68HC12,,@command{ld} and the 68HC11 and 68HC12}.
2186 @xref{Nios II,,@command{ld} and the Altera Nios II}.
2189 @xref{PowerPC ELF32,,@command{ld} and PowerPC 32-bit ELF Support}.
2192 On some platforms the @option{--relax} option performs target specific,
2193 global optimizations that become possible when the linker resolves
2194 addressing in the program, such as relaxing address modes,
2195 synthesizing new instructions, selecting shorter version of current
2196 instructions, and combining constant values.
2198 On some platforms these link time global optimizations may make symbolic
2199 debugging of the resulting executable impossible.
2201 This is known to be the case for the Matsushita MN10200 and MN10300
2202 family of processors.
2205 On platforms where the feature is supported, the option
2206 @option{--no-relax} will disable it.
2208 On platforms where the feature is not supported, both @option{--relax}
2209 and @option{--no-relax} are accepted, but ignored.
2211 @cindex retaining specified symbols
2212 @cindex stripping all but some symbols
2213 @cindex symbols, retaining selectively
2214 @kindex --retain-symbols-file=@var{filename}
2215 @item --retain-symbols-file=@var{filename}
2216 Retain @emph{only} the symbols listed in the file @var{filename},
2217 discarding all others. @var{filename} is simply a flat file, with one
2218 symbol name per line. This option is especially useful in environments
2222 where a large global symbol table is accumulated gradually, to conserve
2225 @samp{--retain-symbols-file} does @emph{not} discard undefined symbols,
2226 or symbols needed for relocations.
2228 You may only specify @samp{--retain-symbols-file} once in the command
2229 line. It overrides @samp{-s} and @samp{-S}.
2232 @item -rpath=@var{dir}
2233 @cindex runtime library search path
2234 @kindex -rpath=@var{dir}
2235 Add a directory to the runtime library search path. This is used when
2236 linking an ELF executable with shared objects. All @option{-rpath}
2237 arguments are concatenated and passed to the runtime linker, which uses
2238 them to locate shared objects at runtime.
2240 The @option{-rpath} option is also used when locating shared objects which
2241 are needed by shared objects explicitly included in the link; see the
2242 description of the @option{-rpath-link} option. Searching @option{-rpath}
2243 in this way is only supported by native linkers and cross linkers which
2244 have been configured with the @option{--with-sysroot} option.
2246 If @option{-rpath} is not used when linking an ELF executable, the
2247 contents of the environment variable @code{LD_RUN_PATH} will be used if it
2250 The @option{-rpath} option may also be used on SunOS. By default, on
2251 SunOS, the linker will form a runtime search path out of all the
2252 @option{-L} options it is given. If a @option{-rpath} option is used, the
2253 runtime search path will be formed exclusively using the @option{-rpath}
2254 options, ignoring the @option{-L} options. This can be useful when using
2255 gcc, which adds many @option{-L} options which may be on NFS mounted
2258 For compatibility with other ELF linkers, if the @option{-R} option is
2259 followed by a directory name, rather than a file name, it is treated as
2260 the @option{-rpath} option.
2264 @cindex link-time runtime library search path
2265 @kindex -rpath-link=@var{dir}
2266 @item -rpath-link=@var{dir}
2267 When using ELF or SunOS, one shared library may require another. This
2268 happens when an @code{ld -shared} link includes a shared library as one
2271 When the linker encounters such a dependency when doing a non-shared,
2272 non-relocatable link, it will automatically try to locate the required
2273 shared library and include it in the link, if it is not included
2274 explicitly. In such a case, the @option{-rpath-link} option
2275 specifies the first set of directories to search. The
2276 @option{-rpath-link} option may specify a sequence of directory names
2277 either by specifying a list of names separated by colons, or by
2278 appearing multiple times.
2280 The tokens @var{$ORIGIN} and @var{$LIB} can appear in these search
2281 directories. They will be replaced by the full path to the directory
2282 containing the program or shared object in the case of @var{$ORIGIN}
2283 and either @samp{lib} - for 32-bit binaries - or @samp{lib64} - for
2284 64-bit binaries - in the case of @var{$LIB}.
2286 The alternative form of these tokens - @var{$@{ORIGIN@}} and
2287 @var{$@{LIB@}} can also be used. The token @var{$PLATFORM} is not
2290 This option should be used with caution as it overrides the search path
2291 that may have been hard compiled into a shared library. In such a case it
2292 is possible to use unintentionally a different search path than the
2293 runtime linker would do.
2295 The linker uses the following search paths to locate required shared
2300 Any directories specified by @option{-rpath-link} options.
2302 Any directories specified by @option{-rpath} options. The difference
2303 between @option{-rpath} and @option{-rpath-link} is that directories
2304 specified by @option{-rpath} options are included in the executable and
2305 used at runtime, whereas the @option{-rpath-link} option is only effective
2306 at link time. Searching @option{-rpath} in this way is only supported
2307 by native linkers and cross linkers which have been configured with
2308 the @option{--with-sysroot} option.
2310 On an ELF system, for native linkers, if the @option{-rpath} and
2311 @option{-rpath-link} options were not used, search the contents of the
2312 environment variable @code{LD_RUN_PATH}.
2314 On SunOS, if the @option{-rpath} option was not used, search any
2315 directories specified using @option{-L} options.
2317 For a native linker, search the contents of the environment
2318 variable @code{LD_LIBRARY_PATH}.
2320 For a native ELF linker, the directories in @code{DT_RUNPATH} or
2321 @code{DT_RPATH} of a shared library are searched for shared
2322 libraries needed by it. The @code{DT_RPATH} entries are ignored if
2323 @code{DT_RUNPATH} entries exist.
2325 For a linker for a Linux system, if the file @file{/etc/ld.so.conf}
2326 exists, the list of directories found in that file. Note: the path
2327 to this file is prefixed with the @code{sysroot} value, if that is
2328 defined, and then any @code{prefix} string if the linker was
2329 configured with the @command{--prefix=<path>} option.
2331 For a native linker on a FreeBSD system, any directories specified by
2332 the @code{_PATH_ELF_HINTS} macro defined in the @file{elf-hints.h}
2335 Any directories specified by a @code{SEARCH_DIR} command in a
2336 linker script given on the command line, including scripts specified
2337 by @option{-T} (but not @option{-dT}).
2339 The default directories, normally @file{/lib} and @file{/usr/lib}.
2341 Any directories specified by a plugin LDPT_SET_EXTRA_LIBRARY_PATH.
2343 Any directories specified by a @code{SEARCH_DIR} command in a default
2347 Note however on Linux based systems there is an additional caveat: If
2348 the @option{--as-needed} option is active @emph{and} a shared library
2349 is located which would normally satisfy the search @emph{and} this
2350 library does not have DT_NEEDED tag for @file{libc.so}
2351 @emph{and} there is a shared library later on in the set of search
2352 directories which also satisfies the search @emph{and}
2353 this second shared library does have a DT_NEEDED tag for
2354 @file{libc.so} @emph{then} the second library will be selected instead
2357 If the required shared library is not found, the linker will issue a
2358 warning and continue with the link.
2366 @cindex shared libraries
2367 Create a shared library. This is currently only supported on ELF, XCOFF
2368 and SunOS platforms. On SunOS, the linker will automatically create a
2369 shared library if the @option{-e} option is not used and there are
2370 undefined symbols in the link.
2372 @kindex --sort-common
2374 @itemx --sort-common=ascending
2375 @itemx --sort-common=descending
2376 This option tells @command{ld} to sort the common symbols by alignment in
2377 ascending or descending order when it places them in the appropriate output
2378 sections. The symbol alignments considered are sixteen-byte or larger,
2379 eight-byte, four-byte, two-byte, and one-byte. This is to prevent gaps
2380 between symbols due to alignment constraints. If no sorting order is
2381 specified, then descending order is assumed.
2383 @kindex --sort-section=name
2384 @item --sort-section=name
2385 This option will apply @code{SORT_BY_NAME} to all wildcard section
2386 patterns in the linker script.
2388 @kindex --sort-section=alignment
2389 @item --sort-section=alignment
2390 This option will apply @code{SORT_BY_ALIGNMENT} to all wildcard section
2391 patterns in the linker script.
2393 @kindex --spare-dynamic-tags
2394 @item --spare-dynamic-tags=@var{count}
2395 This option specifies the number of empty slots to leave in the
2396 .dynamic section of ELF shared objects. Empty slots may be needed by
2397 post processing tools, such as the prelinker. The default is 5.
2399 @kindex --split-by-file
2400 @item --split-by-file[=@var{size}]
2401 Similar to @option{--split-by-reloc} but creates a new output section for
2402 each input file when @var{size} is reached. @var{size} defaults to a
2403 size of 1 if not given.
2405 @kindex --split-by-reloc
2406 @item --split-by-reloc[=@var{count}]
2407 Tries to creates extra sections in the output file so that no single
2408 output section in the file contains more than @var{count} relocations.
2409 This is useful when generating huge relocatable files for downloading into
2410 certain real time kernels with the COFF object file format; since COFF
2411 cannot represent more than 65535 relocations in a single section. Note
2412 that this will fail to work with object file formats which do not
2413 support arbitrary sections. The linker will not split up individual
2414 input sections for redistribution, so if a single input section contains
2415 more than @var{count} relocations one output section will contain that
2416 many relocations. @var{count} defaults to a value of 32768.
2420 Compute and display statistics about the operation of the linker, such
2421 as execution time and memory usage.
2423 @kindex --sysroot=@var{directory}
2424 @item --sysroot=@var{directory}
2425 Use @var{directory} as the location of the sysroot, overriding the
2426 configure-time default. This option is only supported by linkers
2427 that were configured using @option{--with-sysroot}.
2431 This is used by COFF/PE based targets to create a task-linked object
2432 file where all of the global symbols have been converted to statics.
2434 @kindex --traditional-format
2435 @cindex traditional format
2436 @item --traditional-format
2437 For some targets, the output of @command{ld} is different in some ways from
2438 the output of some existing linker. This switch requests @command{ld} to
2439 use the traditional format instead.
2442 For example, on SunOS, @command{ld} combines duplicate entries in the
2443 symbol string table. This can reduce the size of an output file with
2444 full debugging information by over 30 percent. Unfortunately, the SunOS
2445 @code{dbx} program can not read the resulting program (@code{gdb} has no
2446 trouble). The @samp{--traditional-format} switch tells @command{ld} to not
2447 combine duplicate entries.
2449 @kindex --section-start=@var{sectionname}=@var{org}
2450 @item --section-start=@var{sectionname}=@var{org}
2451 Locate a section in the output file at the absolute
2452 address given by @var{org}. You may use this option as many
2453 times as necessary to locate multiple sections in the command
2455 @var{org} must be a single hexadecimal integer;
2456 for compatibility with other linkers, you may omit the leading
2457 @samp{0x} usually associated with hexadecimal values. @emph{Note:} there
2458 should be no white space between @var{sectionname}, the equals
2459 sign (``@key{=}''), and @var{org}.
2461 @kindex -Tbss=@var{org}
2462 @kindex -Tdata=@var{org}
2463 @kindex -Ttext=@var{org}
2464 @cindex segment origins, cmd line
2465 @item -Tbss=@var{org}
2466 @itemx -Tdata=@var{org}
2467 @itemx -Ttext=@var{org}
2468 Same as @option{--section-start}, with @code{.bss}, @code{.data} or
2469 @code{.text} as the @var{sectionname}.
2471 @kindex -Ttext-segment=@var{org}
2472 @item -Ttext-segment=@var{org}
2473 @cindex text segment origin, cmd line
2474 When creating an ELF executable, it will set the address of the first
2475 byte of the text segment.
2477 @kindex -Trodata-segment=@var{org}
2478 @item -Trodata-segment=@var{org}
2479 @cindex rodata segment origin, cmd line
2480 When creating an ELF executable or shared object for a target where
2481 the read-only data is in its own segment separate from the executable
2482 text, it will set the address of the first byte of the read-only data segment.
2484 @kindex -Tldata-segment=@var{org}
2485 @item -Tldata-segment=@var{org}
2486 @cindex ldata segment origin, cmd line
2487 When creating an ELF executable or shared object for x86-64 medium memory
2488 model, it will set the address of the first byte of the ldata segment.
2490 @kindex --unresolved-symbols
2491 @item --unresolved-symbols=@var{method}
2492 Determine how to handle unresolved symbols. There are four possible
2493 values for @samp{method}:
2497 Do not report any unresolved symbols.
2500 Report all unresolved symbols. This is the default.
2502 @item ignore-in-object-files
2503 Report unresolved symbols that are contained in shared libraries, but
2504 ignore them if they come from regular object files.
2506 @item ignore-in-shared-libs
2507 Report unresolved symbols that come from regular object files, but
2508 ignore them if they come from shared libraries. This can be useful
2509 when creating a dynamic binary and it is known that all the shared
2510 libraries that it should be referencing are included on the linker's
2514 The behaviour for shared libraries on their own can also be controlled
2515 by the @option{--[no-]allow-shlib-undefined} option.
2517 Normally the linker will generate an error message for each reported
2518 unresolved symbol but the option @option{--warn-unresolved-symbols}
2519 can change this to a warning.
2521 @kindex --verbose[=@var{NUMBER}]
2522 @cindex verbose[=@var{NUMBER}]
2524 @itemx --verbose[=@var{NUMBER}]
2525 Display the version number for @command{ld} and list the linker emulations
2526 supported. Display which input files can and cannot be opened. Display
2527 the linker script being used by the linker. If the optional @var{NUMBER}
2528 argument > 1, plugin symbol status will also be displayed.
2530 @kindex --version-script=@var{version-scriptfile}
2531 @cindex version script, symbol versions
2532 @item --version-script=@var{version-scriptfile}
2533 Specify the name of a version script to the linker. This is typically
2534 used when creating shared libraries to specify additional information
2535 about the version hierarchy for the library being created. This option
2536 is only fully supported on ELF platforms which support shared libraries;
2537 see @ref{VERSION}. It is partially supported on PE platforms, which can
2538 use version scripts to filter symbol visibility in auto-export mode: any
2539 symbols marked @samp{local} in the version script will not be exported.
2542 @kindex --warn-common
2543 @cindex warnings, on combining symbols
2544 @cindex combining symbols, warnings on
2546 Warn when a common symbol is combined with another common symbol or with
2547 a symbol definition. Unix linkers allow this somewhat sloppy practice,
2548 but linkers on some other operating systems do not. This option allows
2549 you to find potential problems from combining global symbols.
2550 Unfortunately, some C libraries use this practice, so you may get some
2551 warnings about symbols in the libraries as well as in your programs.
2553 There are three kinds of global symbols, illustrated here by C examples:
2557 A definition, which goes in the initialized data section of the output
2561 An undefined reference, which does not allocate space.
2562 There must be either a definition or a common symbol for the
2566 A common symbol. If there are only (one or more) common symbols for a
2567 variable, it goes in the uninitialized data area of the output file.
2568 The linker merges multiple common symbols for the same variable into a
2569 single symbol. If they are of different sizes, it picks the largest
2570 size. The linker turns a common symbol into a declaration, if there is
2571 a definition of the same variable.
2574 The @samp{--warn-common} option can produce five kinds of warnings.
2575 Each warning consists of a pair of lines: the first describes the symbol
2576 just encountered, and the second describes the previous symbol
2577 encountered with the same name. One or both of the two symbols will be
2582 Turning a common symbol into a reference, because there is already a
2583 definition for the symbol.
2585 @var{file}(@var{section}): warning: common of `@var{symbol}'
2586 overridden by definition
2587 @var{file}(@var{section}): warning: defined here
2591 Turning a common symbol into a reference, because a later definition for
2592 the symbol is encountered. This is the same as the previous case,
2593 except that the symbols are encountered in a different order.
2595 @var{file}(@var{section}): warning: definition of `@var{symbol}'
2597 @var{file}(@var{section}): warning: common is here
2601 Merging a common symbol with a previous same-sized common symbol.
2603 @var{file}(@var{section}): warning: multiple common
2605 @var{file}(@var{section}): warning: previous common is here
2609 Merging a common symbol with a previous larger common symbol.
2611 @var{file}(@var{section}): warning: common of `@var{symbol}'
2612 overridden by larger common
2613 @var{file}(@var{section}): warning: larger common is here
2617 Merging a common symbol with a previous smaller common symbol. This is
2618 the same as the previous case, except that the symbols are
2619 encountered in a different order.
2621 @var{file}(@var{section}): warning: common of `@var{symbol}'
2622 overriding smaller common
2623 @var{file}(@var{section}): warning: smaller common is here
2627 @kindex --warn-constructors
2628 @item --warn-constructors
2629 Warn if any global constructors are used. This is only useful for a few
2630 object file formats. For formats like COFF or ELF, the linker can not
2631 detect the use of global constructors.
2633 @kindex --warn-multiple-gp
2634 @item --warn-multiple-gp
2635 Warn if multiple global pointer values are required in the output file.
2636 This is only meaningful for certain processors, such as the Alpha.
2637 Specifically, some processors put large-valued constants in a special
2638 section. A special register (the global pointer) points into the middle
2639 of this section, so that constants can be loaded efficiently via a
2640 base-register relative addressing mode. Since the offset in
2641 base-register relative mode is fixed and relatively small (e.g., 16
2642 bits), this limits the maximum size of the constant pool. Thus, in
2643 large programs, it is often necessary to use multiple global pointer
2644 values in order to be able to address all possible constants. This
2645 option causes a warning to be issued whenever this case occurs.
2648 @cindex warnings, on undefined symbols
2649 @cindex undefined symbols, warnings on
2651 Only warn once for each undefined symbol, rather than once per module
2654 @kindex --warn-section-align
2655 @cindex warnings, on section alignment
2656 @cindex section alignment, warnings on
2657 @item --warn-section-align
2658 Warn if the address of an output section is changed because of
2659 alignment. Typically, the alignment will be set by an input section.
2660 The address will only be changed if it not explicitly specified; that
2661 is, if the @code{SECTIONS} command does not specify a start address for
2662 the section (@pxref{SECTIONS}).
2664 @kindex --warn-textrel
2665 @item --warn-textrel
2666 Warn if the linker adds DT_TEXTREL to a position-independent executable
2669 @kindex --warn-alternate-em
2670 @item --warn-alternate-em
2671 Warn if an object has alternate ELF machine code.
2673 @kindex --warn-unresolved-symbols
2674 @item --warn-unresolved-symbols
2675 If the linker is going to report an unresolved symbol (see the option
2676 @option{--unresolved-symbols}) it will normally generate an error.
2677 This option makes it generate a warning instead.
2679 @kindex --error-unresolved-symbols
2680 @item --error-unresolved-symbols
2681 This restores the linker's default behaviour of generating errors when
2682 it is reporting unresolved symbols.
2684 @kindex --whole-archive
2685 @cindex including an entire archive
2686 @item --whole-archive
2687 For each archive mentioned on the command line after the
2688 @option{--whole-archive} option, include every object file in the archive
2689 in the link, rather than searching the archive for the required object
2690 files. This is normally used to turn an archive file into a shared
2691 library, forcing every object to be included in the resulting shared
2692 library. This option may be used more than once.
2694 Two notes when using this option from gcc: First, gcc doesn't know
2695 about this option, so you have to use @option{-Wl,-whole-archive}.
2696 Second, don't forget to use @option{-Wl,-no-whole-archive} after your
2697 list of archives, because gcc will add its own list of archives to
2698 your link and you may not want this flag to affect those as well.
2700 @kindex --wrap=@var{symbol}
2701 @item --wrap=@var{symbol}
2702 Use a wrapper function for @var{symbol}. Any undefined reference to
2703 @var{symbol} will be resolved to @code{__wrap_@var{symbol}}. Any
2704 undefined reference to @code{__real_@var{symbol}} will be resolved to
2707 This can be used to provide a wrapper for a system function. The
2708 wrapper function should be called @code{__wrap_@var{symbol}}. If it
2709 wishes to call the system function, it should call
2710 @code{__real_@var{symbol}}.
2712 Here is a trivial example:
2716 __wrap_malloc (size_t c)
2718 printf ("malloc called with %zu\n", c);
2719 return __real_malloc (c);
2723 If you link other code with this file using @option{--wrap malloc}, then
2724 all calls to @code{malloc} will call the function @code{__wrap_malloc}
2725 instead. The call to @code{__real_malloc} in @code{__wrap_malloc} will
2726 call the real @code{malloc} function.
2728 You may wish to provide a @code{__real_malloc} function as well, so that
2729 links without the @option{--wrap} option will succeed. If you do this,
2730 you should not put the definition of @code{__real_malloc} in the same
2731 file as @code{__wrap_malloc}; if you do, the assembler may resolve the
2732 call before the linker has a chance to wrap it to @code{malloc}.
2734 Only undefined references are replaced by the linker. So, translation unit
2735 internal references to @var{symbol} are not resolved to
2736 @code{__wrap_@var{symbol}}. In the next example, the call to @code{f} in
2737 @code{g} is not resolved to @code{__wrap_f}.
2753 @kindex --eh-frame-hdr
2754 @kindex --no-eh-frame-hdr
2755 @item --eh-frame-hdr
2756 @itemx --no-eh-frame-hdr
2757 Request (@option{--eh-frame-hdr}) or suppress
2758 (@option{--no-eh-frame-hdr}) the creation of @code{.eh_frame_hdr}
2759 section and ELF @code{PT_GNU_EH_FRAME} segment header.
2761 @kindex --ld-generated-unwind-info
2762 @item --no-ld-generated-unwind-info
2763 Request creation of @code{.eh_frame} unwind info for linker
2764 generated code sections like PLT. This option is on by default
2765 if linker generated unwind info is supported.
2767 @kindex --enable-new-dtags
2768 @kindex --disable-new-dtags
2769 @item --enable-new-dtags
2770 @itemx --disable-new-dtags
2771 This linker can create the new dynamic tags in ELF. But the older ELF
2772 systems may not understand them. If you specify
2773 @option{--enable-new-dtags}, the new dynamic tags will be created as needed
2774 and older dynamic tags will be omitted.
2775 If you specify @option{--disable-new-dtags}, no new dynamic tags will be
2776 created. By default, the new dynamic tags are not created. Note that
2777 those options are only available for ELF systems.
2779 @kindex --hash-size=@var{number}
2780 @item --hash-size=@var{number}
2781 Set the default size of the linker's hash tables to a prime number
2782 close to @var{number}. Increasing this value can reduce the length of
2783 time it takes the linker to perform its tasks, at the expense of
2784 increasing the linker's memory requirements. Similarly reducing this
2785 value can reduce the memory requirements at the expense of speed.
2787 @kindex --hash-style=@var{style}
2788 @item --hash-style=@var{style}
2789 Set the type of linker's hash table(s). @var{style} can be either
2790 @code{sysv} for classic ELF @code{.hash} section, @code{gnu} for
2791 new style GNU @code{.gnu.hash} section or @code{both} for both
2792 the classic ELF @code{.hash} and new style GNU @code{.gnu.hash}
2793 hash tables. The default depends upon how the linker was configured,
2794 but for most Linux based systems it will be @code{both}.
2796 @kindex --compress-debug-sections=none
2797 @kindex --compress-debug-sections=zlib
2798 @kindex --compress-debug-sections=zlib-gnu
2799 @kindex --compress-debug-sections=zlib-gabi
2800 @item --compress-debug-sections=none
2801 @itemx --compress-debug-sections=zlib
2802 @itemx --compress-debug-sections=zlib-gnu
2803 @itemx --compress-debug-sections=zlib-gabi
2804 On ELF platforms, these options control how DWARF debug sections are
2805 compressed using zlib.
2807 @option{--compress-debug-sections=none} doesn't compress DWARF debug
2808 sections. @option{--compress-debug-sections=zlib-gnu} compresses
2809 DWARF debug sections and renames them to begin with @samp{.zdebug}
2810 instead of @samp{.debug}. @option{--compress-debug-sections=zlib-gabi}
2811 also compresses DWARF debug sections, but rather than renaming them it
2812 sets the SHF_COMPRESSED flag in the sections' headers.
2814 The @option{--compress-debug-sections=zlib} option is an alias for
2815 @option{--compress-debug-sections=zlib-gabi}.
2817 Note that this option overrides any compression in input debug
2818 sections, so if a binary is linked with @option{--compress-debug-sections=none}
2819 for example, then any compressed debug sections in input files will be
2820 uncompressed before they are copied into the output binary.
2822 The default compression behaviour varies depending upon the target
2823 involved and the configure options used to build the toolchain. The
2824 default can be determined by examining the output from the linker's
2825 @option{--help} option.
2827 @kindex --reduce-memory-overheads
2828 @item --reduce-memory-overheads
2829 This option reduces memory requirements at ld runtime, at the expense of
2830 linking speed. This was introduced to select the old O(n^2) algorithm
2831 for link map file generation, rather than the new O(n) algorithm which uses
2832 about 40% more memory for symbol storage.
2834 Another effect of the switch is to set the default hash table size to
2835 1021, which again saves memory at the cost of lengthening the linker's
2836 run time. This is not done however if the @option{--hash-size} switch
2839 The @option{--reduce-memory-overheads} switch may be also be used to
2840 enable other tradeoffs in future versions of the linker.
2842 @kindex --max-cache-size=@var{size}
2843 @item --max-cache-size=@var{size}
2844 @command{ld} normally caches the relocation information and symbol tables
2845 of input files in memory with the unlimited size. This option sets the
2846 maximum cache size to @var{size}.
2849 @kindex --build-id=@var{style}
2851 @itemx --build-id=@var{style}
2852 Request the creation of a @code{.note.gnu.build-id} ELF note section
2853 or a @code{.buildid} COFF section. The contents of the note are
2854 unique bits identifying this linked file. @var{style} can be
2855 @code{uuid} to use 128 random bits, @code{sha1} to use a 160-bit
2856 @sc{SHA1} hash on the normative parts of the output contents,
2857 @code{md5} to use a 128-bit @sc{MD5} hash on the normative parts of
2858 the output contents, or @code{0x@var{hexstring}} to use a chosen bit
2859 string specified as an even number of hexadecimal digits (@code{-} and
2860 @code{:} characters between digit pairs are ignored). If @var{style}
2861 is omitted, @code{sha1} is used.
2863 The @code{md5} and @code{sha1} styles produces an identifier
2864 that is always the same in an identical output file, but will be
2865 unique among all nonidentical output files. It is not intended
2866 to be compared as a checksum for the file's contents. A linked
2867 file may be changed later by other tools, but the build ID bit
2868 string identifying the original linked file does not change.
2870 Passing @code{none} for @var{style} disables the setting from any
2871 @code{--build-id} options earlier on the command line.
2876 @subsection Options Specific to i386 PE Targets
2878 @c man begin OPTIONS
2880 The i386 PE linker supports the @option{-shared} option, which causes
2881 the output to be a dynamically linked library (DLL) instead of a
2882 normal executable. You should name the output @code{*.dll} when you
2883 use this option. In addition, the linker fully supports the standard
2884 @code{*.def} files, which may be specified on the linker command line
2885 like an object file (in fact, it should precede archives it exports
2886 symbols from, to ensure that they get linked in, just like a normal
2889 In addition to the options common to all targets, the i386 PE linker
2890 support additional command-line options that are specific to the i386
2891 PE target. Options that take values may be separated from their
2892 values by either a space or an equals sign.
2896 @kindex --add-stdcall-alias
2897 @item --add-stdcall-alias
2898 If given, symbols with a stdcall suffix (@@@var{nn}) will be exported
2899 as-is and also with the suffix stripped.
2900 [This option is specific to the i386 PE targeted port of the linker]
2903 @item --base-file @var{file}
2904 Use @var{file} as the name of a file in which to save the base
2905 addresses of all the relocations needed for generating DLLs with
2907 [This is an i386 PE specific option]
2911 Create a DLL instead of a regular executable. You may also use
2912 @option{-shared} or specify a @code{LIBRARY} in a given @code{.def}
2914 [This option is specific to the i386 PE targeted port of the linker]
2916 @kindex --enable-long-section-names
2917 @kindex --disable-long-section-names
2918 @item --enable-long-section-names
2919 @itemx --disable-long-section-names
2920 The PE variants of the COFF object format add an extension that permits
2921 the use of section names longer than eight characters, the normal limit
2922 for COFF. By default, these names are only allowed in object files, as
2923 fully-linked executable images do not carry the COFF string table required
2924 to support the longer names. As a GNU extension, it is possible to
2925 allow their use in executable images as well, or to (probably pointlessly!)
2926 disallow it in object files, by using these two options. Executable images
2927 generated with these long section names are slightly non-standard, carrying
2928 as they do a string table, and may generate confusing output when examined
2929 with non-GNU PE-aware tools, such as file viewers and dumpers. However,
2930 GDB relies on the use of PE long section names to find Dwarf-2 debug
2931 information sections in an executable image at runtime, and so if neither
2932 option is specified on the command-line, @command{ld} will enable long
2933 section names, overriding the default and technically correct behaviour,
2934 when it finds the presence of debug information while linking an executable
2935 image and not stripping symbols.
2936 [This option is valid for all PE targeted ports of the linker]
2938 @kindex --enable-stdcall-fixup
2939 @kindex --disable-stdcall-fixup
2940 @item --enable-stdcall-fixup
2941 @itemx --disable-stdcall-fixup
2942 If the link finds a symbol that it cannot resolve, it will attempt to
2943 do ``fuzzy linking'' by looking for another defined symbol that differs
2944 only in the format of the symbol name (cdecl vs stdcall) and will
2945 resolve that symbol by linking to the match. For example, the
2946 undefined symbol @code{_foo} might be linked to the function
2947 @code{_foo@@12}, or the undefined symbol @code{_bar@@16} might be linked
2948 to the function @code{_bar}. When the linker does this, it prints a
2949 warning, since it normally should have failed to link, but sometimes
2950 import libraries generated from third-party dlls may need this feature
2951 to be usable. If you specify @option{--enable-stdcall-fixup}, this
2952 feature is fully enabled and warnings are not printed. If you specify
2953 @option{--disable-stdcall-fixup}, this feature is disabled and such
2954 mismatches are considered to be errors.
2955 [This option is specific to the i386 PE targeted port of the linker]
2957 @kindex --leading-underscore
2958 @kindex --no-leading-underscore
2959 @item --leading-underscore
2960 @itemx --no-leading-underscore
2961 For most targets default symbol-prefix is an underscore and is defined
2962 in target's description. By this option it is possible to
2963 disable/enable the default underscore symbol-prefix.
2965 @cindex DLLs, creating
2966 @kindex --export-all-symbols
2967 @item --export-all-symbols
2968 If given, all global symbols in the objects used to build a DLL will
2969 be exported by the DLL. Note that this is the default if there
2970 otherwise wouldn't be any exported symbols. When symbols are
2971 explicitly exported via DEF files or implicitly exported via function
2972 attributes, the default is to not export anything else unless this
2973 option is given. Note that the symbols @code{DllMain@@12},
2974 @code{DllEntryPoint@@0}, @code{DllMainCRTStartup@@12}, and
2975 @code{impure_ptr} will not be automatically
2976 exported. Also, symbols imported from other DLLs will not be
2977 re-exported, nor will symbols specifying the DLL's internal layout
2978 such as those beginning with @code{_head_} or ending with
2979 @code{_iname}. In addition, no symbols from @code{libgcc},
2980 @code{libstd++}, @code{libmingw32}, or @code{crtX.o} will be exported.
2981 Symbols whose names begin with @code{__rtti_} or @code{__builtin_} will
2982 not be exported, to help with C++ DLLs. Finally, there is an
2983 extensive list of cygwin-private symbols that are not exported
2984 (obviously, this applies on when building DLLs for cygwin targets).
2985 These cygwin-excludes are: @code{_cygwin_dll_entry@@12},
2986 @code{_cygwin_crt0_common@@8}, @code{_cygwin_noncygwin_dll_entry@@12},
2987 @code{_fmode}, @code{_impure_ptr}, @code{cygwin_attach_dll},
2988 @code{cygwin_premain0}, @code{cygwin_premain1}, @code{cygwin_premain2},
2989 @code{cygwin_premain3}, and @code{environ}.
2990 [This option is specific to the i386 PE targeted port of the linker]
2992 @kindex --exclude-symbols
2993 @item --exclude-symbols @var{symbol},@var{symbol},...
2994 Specifies a list of symbols which should not be automatically
2995 exported. The symbol names may be delimited by commas or colons.
2996 [This option is specific to the i386 PE targeted port of the linker]
2998 @kindex --exclude-all-symbols
2999 @item --exclude-all-symbols
3000 Specifies no symbols should be automatically exported.
3001 [This option is specific to the i386 PE targeted port of the linker]
3003 @kindex --file-alignment
3004 @item --file-alignment
3005 Specify the file alignment. Sections in the file will always begin at
3006 file offsets which are multiples of this number. This defaults to
3008 [This option is specific to the i386 PE targeted port of the linker]
3012 @item --heap @var{reserve}
3013 @itemx --heap @var{reserve},@var{commit}
3014 Specify the number of bytes of memory to reserve (and optionally commit)
3015 to be used as heap for this program. The default is 1MB reserved, 4K
3017 [This option is specific to the i386 PE targeted port of the linker]
3020 @kindex --image-base
3021 @item --image-base @var{value}
3022 Use @var{value} as the base address of your program or dll. This is
3023 the lowest memory location that will be used when your program or dll
3024 is loaded. To reduce the need to relocate and improve performance of
3025 your dlls, each should have a unique base address and not overlap any
3026 other dlls. The default is 0x400000 for executables, and 0x10000000
3028 [This option is specific to the i386 PE targeted port of the linker]
3032 If given, the stdcall suffixes (@@@var{nn}) will be stripped from
3033 symbols before they are exported.
3034 [This option is specific to the i386 PE targeted port of the linker]
3036 @kindex --large-address-aware
3037 @item --large-address-aware
3038 If given, the appropriate bit in the ``Characteristics'' field of the COFF
3039 header is set to indicate that this executable supports virtual addresses
3040 greater than 2 gigabytes. This should be used in conjunction with the /3GB
3041 or /USERVA=@var{value} megabytes switch in the ``[operating systems]''
3042 section of the BOOT.INI. Otherwise, this bit has no effect.
3043 [This option is specific to PE targeted ports of the linker]
3045 @kindex --disable-large-address-aware
3046 @item --disable-large-address-aware
3047 Reverts the effect of a previous @samp{--large-address-aware} option.
3048 This is useful if @samp{--large-address-aware} is always set by the compiler
3049 driver (e.g. Cygwin gcc) and the executable does not support virtual
3050 addresses greater than 2 gigabytes.
3051 [This option is specific to PE targeted ports of the linker]
3053 @kindex --major-image-version
3054 @item --major-image-version @var{value}
3055 Sets the major number of the ``image version''. Defaults to 1.
3056 [This option is specific to the i386 PE targeted port of the linker]
3058 @kindex --major-os-version
3059 @item --major-os-version @var{value}
3060 Sets the major number of the ``os version''. Defaults to 4.
3061 [This option is specific to the i386 PE targeted port of the linker]
3063 @kindex --major-subsystem-version
3064 @item --major-subsystem-version @var{value}
3065 Sets the major number of the ``subsystem version''. Defaults to 4.
3066 [This option is specific to the i386 PE targeted port of the linker]
3068 @kindex --minor-image-version
3069 @item --minor-image-version @var{value}
3070 Sets the minor number of the ``image version''. Defaults to 0.
3071 [This option is specific to the i386 PE targeted port of the linker]
3073 @kindex --minor-os-version
3074 @item --minor-os-version @var{value}
3075 Sets the minor number of the ``os version''. Defaults to 0.
3076 [This option is specific to the i386 PE targeted port of the linker]
3078 @kindex --minor-subsystem-version
3079 @item --minor-subsystem-version @var{value}
3080 Sets the minor number of the ``subsystem version''. Defaults to 0.
3081 [This option is specific to the i386 PE targeted port of the linker]
3083 @cindex DEF files, creating
3084 @cindex DLLs, creating
3085 @kindex --output-def
3086 @item --output-def @var{file}
3087 The linker will create the file @var{file} which will contain a DEF
3088 file corresponding to the DLL the linker is generating. This DEF file
3089 (which should be called @code{*.def}) may be used to create an import
3090 library with @code{dlltool} or may be used as a reference to
3091 automatically or implicitly exported symbols.
3092 [This option is specific to the i386 PE targeted port of the linker]
3094 @cindex DLLs, creating
3095 @kindex --enable-auto-image-base
3096 @item --enable-auto-image-base
3097 @itemx --enable-auto-image-base=@var{value}
3098 Automatically choose the image base for DLLs, optionally starting with base
3099 @var{value}, unless one is specified using the @code{--image-base} argument.
3100 By using a hash generated from the dllname to create unique image bases
3101 for each DLL, in-memory collisions and relocations which can delay program
3102 execution are avoided.
3103 [This option is specific to the i386 PE targeted port of the linker]
3105 @kindex --disable-auto-image-base
3106 @item --disable-auto-image-base
3107 Do not automatically generate a unique image base. If there is no
3108 user-specified image base (@code{--image-base}) then use the platform
3110 [This option is specific to the i386 PE targeted port of the linker]
3112 @cindex DLLs, linking to
3113 @kindex --dll-search-prefix
3114 @item --dll-search-prefix @var{string}
3115 When linking dynamically to a dll without an import library,
3116 search for @code{<string><basename>.dll} in preference to
3117 @code{lib<basename>.dll}. This behaviour allows easy distinction
3118 between DLLs built for the various "subplatforms": native, cygwin,
3119 uwin, pw, etc. For instance, cygwin DLLs typically use
3120 @code{--dll-search-prefix=cyg}.
3121 [This option is specific to the i386 PE targeted port of the linker]
3123 @kindex --enable-auto-import
3124 @item --enable-auto-import
3125 Do sophisticated linking of @code{_symbol} to @code{__imp__symbol} for
3126 DATA imports from DLLs, thus making it possible to bypass the dllimport
3127 mechanism on the user side and to reference unmangled symbol names.
3128 [This option is specific to the i386 PE targeted port of the linker]
3130 The following remarks pertain to the original implementation of the
3131 feature and are obsolete nowadays for Cygwin and MinGW targets.
3133 Note: Use of the 'auto-import' extension will cause the text section
3134 of the image file to be made writable. This does not conform to the
3135 PE-COFF format specification published by Microsoft.
3137 Note - use of the 'auto-import' extension will also cause read only
3138 data which would normally be placed into the .rdata section to be
3139 placed into the .data section instead. This is in order to work
3140 around a problem with consts that is described here:
3141 http://www.cygwin.com/ml/cygwin/2004-09/msg01101.html
3143 Using 'auto-import' generally will 'just work' -- but sometimes you may
3146 "variable '<var>' can't be auto-imported. Please read the
3147 documentation for ld's @code{--enable-auto-import} for details."
3149 This message occurs when some (sub)expression accesses an address
3150 ultimately given by the sum of two constants (Win32 import tables only
3151 allow one). Instances where this may occur include accesses to member
3152 fields of struct variables imported from a DLL, as well as using a
3153 constant index into an array variable imported from a DLL. Any
3154 multiword variable (arrays, structs, long long, etc) may trigger
3155 this error condition. However, regardless of the exact data type
3156 of the offending exported variable, ld will always detect it, issue
3157 the warning, and exit.
3159 There are several ways to address this difficulty, regardless of the
3160 data type of the exported variable:
3162 One way is to use --enable-runtime-pseudo-reloc switch. This leaves the task
3163 of adjusting references in your client code for runtime environment, so
3164 this method works only when runtime environment supports this feature.
3166 A second solution is to force one of the 'constants' to be a variable --
3167 that is, unknown and un-optimizable at compile time. For arrays,
3168 there are two possibilities: a) make the indexee (the array's address)
3169 a variable, or b) make the 'constant' index a variable. Thus:
3172 extern type extern_array[];
3174 @{ volatile type *t=extern_array; t[1] @}
3180 extern type extern_array[];
3182 @{ volatile int t=1; extern_array[t] @}
3185 For structs (and most other multiword data types) the only option
3186 is to make the struct itself (or the long long, or the ...) variable:
3189 extern struct s extern_struct;
3190 extern_struct.field -->
3191 @{ volatile struct s *t=&extern_struct; t->field @}
3197 extern long long extern_ll;
3199 @{ volatile long long * local_ll=&extern_ll; *local_ll @}
3202 A third method of dealing with this difficulty is to abandon
3203 'auto-import' for the offending symbol and mark it with
3204 @code{__declspec(dllimport)}. However, in practice that
3205 requires using compile-time #defines to indicate whether you are
3206 building a DLL, building client code that will link to the DLL, or
3207 merely building/linking to a static library. In making the choice
3208 between the various methods of resolving the 'direct address with
3209 constant offset' problem, you should consider typical real-world usage:
3217 void main(int argc, char **argv)@{
3218 printf("%d\n",arr[1]);
3228 void main(int argc, char **argv)@{
3229 /* This workaround is for win32 and cygwin; do not "optimize" */
3230 volatile int *parr = arr;
3231 printf("%d\n",parr[1]);
3238 /* Note: auto-export is assumed (no __declspec(dllexport)) */
3239 #if (defined(_WIN32) || defined(__CYGWIN__)) && \
3240 !(defined(FOO_BUILD_DLL) || defined(FOO_STATIC))
3241 #define FOO_IMPORT __declspec(dllimport)
3245 extern FOO_IMPORT int arr[];
3248 void main(int argc, char **argv)@{
3249 printf("%d\n",arr[1]);
3253 A fourth way to avoid this problem is to re-code your
3254 library to use a functional interface rather than a data interface
3255 for the offending variables (e.g. set_foo() and get_foo() accessor
3258 @kindex --disable-auto-import
3259 @item --disable-auto-import
3260 Do not attempt to do sophisticated linking of @code{_symbol} to
3261 @code{__imp__symbol} for DATA imports from DLLs.
3262 [This option is specific to the i386 PE targeted port of the linker]
3264 @kindex --enable-runtime-pseudo-reloc
3265 @item --enable-runtime-pseudo-reloc
3266 If your code contains expressions described in --enable-auto-import section,
3267 that is, DATA imports from DLL with non-zero offset, this switch will create
3268 a vector of 'runtime pseudo relocations' which can be used by runtime
3269 environment to adjust references to such data in your client code.
3270 [This option is specific to the i386 PE targeted port of the linker]
3272 @kindex --disable-runtime-pseudo-reloc
3273 @item --disable-runtime-pseudo-reloc
3274 Do not create pseudo relocations for non-zero offset DATA imports from DLLs.
3275 [This option is specific to the i386 PE targeted port of the linker]
3277 @kindex --enable-extra-pe-debug
3278 @item --enable-extra-pe-debug
3279 Show additional debug info related to auto-import symbol thunking.
3280 [This option is specific to the i386 PE targeted port of the linker]
3282 @kindex --section-alignment
3283 @item --section-alignment
3284 Sets the section alignment. Sections in memory will always begin at
3285 addresses which are a multiple of this number. Defaults to 0x1000.
3286 [This option is specific to the i386 PE targeted port of the linker]
3290 @item --stack @var{reserve}
3291 @itemx --stack @var{reserve},@var{commit}
3292 Specify the number of bytes of memory to reserve (and optionally commit)
3293 to be used as stack for this program. The default is 2MB reserved, 4K
3295 [This option is specific to the i386 PE targeted port of the linker]
3298 @item --subsystem @var{which}
3299 @itemx --subsystem @var{which}:@var{major}
3300 @itemx --subsystem @var{which}:@var{major}.@var{minor}
3301 Specifies the subsystem under which your program will execute. The
3302 legal values for @var{which} are @code{native}, @code{windows},
3303 @code{console}, @code{posix}, and @code{xbox}. You may optionally set
3304 the subsystem version also. Numeric values are also accepted for
3306 [This option is specific to the i386 PE targeted port of the linker]
3308 The following options set flags in the @code{DllCharacteristics} field
3309 of the PE file header:
3310 [These options are specific to PE targeted ports of the linker]
3312 @kindex --high-entropy-va
3313 @item --high-entropy-va
3314 @itemx --disable-high-entropy-va
3315 Image is compatible with 64-bit address space layout randomization
3316 (ASLR). This option is enabled by default for 64-bit PE images.
3318 This option also implies @option{--dynamicbase} and
3319 @option{--enable-reloc-section}.
3321 @kindex --dynamicbase
3323 @itemx --disable-dynamicbase
3324 The image base address may be relocated using address space layout
3325 randomization (ASLR). This feature was introduced with MS Windows
3326 Vista for i386 PE targets. This option is enabled by default but
3327 can be disabled via the @option{--disable-dynamicbase} option.
3328 This option also implies @option{--enable-reloc-section}.
3330 @kindex --forceinteg
3332 @itemx --disable-forceinteg
3333 Code integrity checks are enforced. This option is disabled by
3338 @item --disable-nxcompat
3339 The image is compatible with the Data Execution Prevention.
3340 This feature was introduced with MS Windows XP SP2 for i386 PE
3341 targets. The option is enabled by default.
3343 @kindex --no-isolation
3344 @item --no-isolation
3345 @itemx --disable-no-isolation
3346 Although the image understands isolation, do not isolate the image.
3347 This option is disabled by default.
3351 @itemx --disable-no-seh
3352 The image does not use SEH. No SE handler may be called from
3353 this image. This option is disabled by default.
3357 @itemx --disable-no-bind
3358 Do not bind this image. This option is disabled by default.
3362 @itemx --disable-wdmdriver
3363 The driver uses the MS Windows Driver Model. This option is disabled
3368 @itemx --disable-tsaware
3369 The image is Terminal Server aware. This option is disabled by
3372 @kindex --insert-timestamp
3373 @item --insert-timestamp
3374 @itemx --no-insert-timestamp
3375 Insert a real timestamp into the image. This is the default behaviour
3376 as it matches legacy code and it means that the image will work with
3377 other, proprietary tools. The problem with this default is that it
3378 will result in slightly different images being produced each time the
3379 same sources are linked. The option @option{--no-insert-timestamp}
3380 can be used to insert a zero value for the timestamp, this ensuring
3381 that binaries produced from identical sources will compare
3384 @kindex --enable-reloc-section
3385 @item --enable-reloc-section
3386 @itemx --disable-reloc-section
3387 Create the base relocation table, which is necessary if the image
3388 is loaded at a different image base than specified in the PE header.
3389 This option is enabled by default.
3395 @subsection Options specific to C6X uClinux targets
3397 @c man begin OPTIONS
3399 The C6X uClinux target uses a binary format called DSBT to support shared
3400 libraries. Each shared library in the system needs to have a unique index;
3401 all executables use an index of 0.
3406 @item --dsbt-size @var{size}
3407 This option sets the number of entries in the DSBT of the current executable
3408 or shared library to @var{size}. The default is to create a table with 64
3411 @kindex --dsbt-index
3412 @item --dsbt-index @var{index}
3413 This option sets the DSBT index of the current executable or shared library
3414 to @var{index}. The default is 0, which is appropriate for generating
3415 executables. If a shared library is generated with a DSBT index of 0, the
3416 @code{R_C6000_DSBT_INDEX} relocs are copied into the output file.
3418 @kindex --no-merge-exidx-entries
3419 The @samp{--no-merge-exidx-entries} switch disables the merging of adjacent
3420 exidx entries in frame unwind info.
3428 @subsection Options specific to C-SKY targets
3430 @c man begin OPTIONS
3434 @kindex --branch-stub on C-SKY
3436 This option enables linker branch relaxation by inserting branch stub
3437 sections when needed to extend the range of branches. This option is
3438 usually not required since C-SKY supports branch and call instructions that
3439 can access the full memory range and branch relaxation is normally handled by
3440 the compiler or assembler.
3442 @kindex --stub-group-size on C-SKY
3443 @item --stub-group-size=@var{N}
3444 This option allows finer control of linker branch stub creation.
3445 It sets the maximum size of a group of input sections that can
3446 be handled by one stub section. A negative value of @var{N} locates
3447 stub sections after their branches, while a positive value allows stub
3448 sections to appear either before or after the branches. Values of
3449 @samp{1} or @samp{-1} indicate that the
3450 linker should choose suitable defaults.
3458 @subsection Options specific to Motorola 68HC11 and 68HC12 targets
3460 @c man begin OPTIONS
3462 The 68HC11 and 68HC12 linkers support specific options to control the
3463 memory bank switching mapping and trampoline code generation.
3467 @kindex --no-trampoline
3468 @item --no-trampoline
3469 This option disables the generation of trampoline. By default a trampoline
3470 is generated for each far function which is called using a @code{jsr}
3471 instruction (this happens when a pointer to a far function is taken).
3473 @kindex --bank-window
3474 @item --bank-window @var{name}
3475 This option indicates to the linker the name of the memory region in
3476 the @samp{MEMORY} specification that describes the memory bank window.
3477 The definition of such region is then used by the linker to compute
3478 paging and addresses within the memory window.
3486 @subsection Options specific to Motorola 68K target
3488 @c man begin OPTIONS
3490 The following options are supported to control handling of GOT generation
3491 when linking for 68K targets.
3496 @item --got=@var{type}
3497 This option tells the linker which GOT generation scheme to use.
3498 @var{type} should be one of @samp{single}, @samp{negative},
3499 @samp{multigot} or @samp{target}. For more information refer to the
3500 Info entry for @file{ld}.
3508 @subsection Options specific to MIPS targets
3510 @c man begin OPTIONS
3512 The following options are supported to control microMIPS instruction
3513 generation and branch relocation checks for ISA mode transitions when
3514 linking for MIPS targets.
3522 These options control the choice of microMIPS instructions used in code
3523 generated by the linker, such as that in the PLT or lazy binding stubs,
3524 or in relaxation. If @samp{--insn32} is used, then the linker only uses
3525 32-bit instruction encodings. By default or if @samp{--no-insn32} is
3526 used, all instruction encodings are used, including 16-bit ones where
3529 @kindex --ignore-branch-isa
3530 @item --ignore-branch-isa
3531 @kindex --no-ignore-branch-isa
3532 @itemx --no-ignore-branch-isa
3533 These options control branch relocation checks for invalid ISA mode
3534 transitions. If @samp{--ignore-branch-isa} is used, then the linker
3535 accepts any branch relocations and any ISA mode transition required
3536 is lost in relocation calculation, except for some cases of @code{BAL}
3537 instructions which meet relaxation conditions and are converted to
3538 equivalent @code{JALX} instructions as the associated relocation is
3539 calculated. By default or if @samp{--no-ignore-branch-isa} is used
3540 a check is made causing the loss of an ISA mode transition to produce
3543 @kindex --compact-branches
3544 @item --compact-branches
3545 @kindex --no-compact-branches
3546 @itemx --no-compact-branches
3547 These options control the generation of compact instructions by the linker
3548 in the PLT entries for MIPS R6.
3557 @subsection Options specific to PDP11 targets
3559 @c man begin OPTIONS
3561 For the pdp11-aout target, three variants of the output format can be
3562 produced as selected by the following options. The default variant
3563 for pdp11-aout is the @samp{--omagic} option, whereas for other
3564 targets @samp{--nmagic} is the default. The @samp{--imagic} option is
3565 defined only for the pdp11-aout target, while the others are described
3566 here as they apply to the pdp11-aout target.
3575 Mark the output as @code{OMAGIC} (0407) in the @file{a.out} header to
3576 indicate that the text segment is not to be write-protected and
3577 shared. Since the text and data sections are both readable and
3578 writable, the data section is allocated immediately contiguous after
3579 the text segment. This is the oldest format for PDP11 executable
3580 programs and is the default for @command{ld} on PDP11 Unix systems
3581 from the beginning through 2.11BSD.
3588 Mark the output as @code{NMAGIC} (0410) in the @file{a.out} header to
3589 indicate that when the output file is executed, the text portion will
3590 be read-only and shareable among all processes executing the same
3591 file. This involves moving the data areas up to the first possible 8K
3592 byte page boundary following the end of the text. This option creates
3593 a @emph{pure executable} format.
3600 Mark the output as @code{IMAGIC} (0411) in the @file{a.out} header to
3601 indicate that when the output file is executed, the program text and
3602 data areas will be loaded into separate address spaces using the split
3603 instruction and data space feature of the memory management unit in
3604 larger models of the PDP11. This doubles the address space available
3605 to the program. The text segment is again pure, write-protected, and
3606 shareable. The only difference in the output format between this
3607 option and the others, besides the magic number, is that both the text
3608 and data sections start at location 0. The @samp{-z} option selected
3609 this format in 2.11BSD. This option creates a @emph{separate
3615 Equivalent to @samp{--nmagic} for pdp11-aout.
3624 @section Environment Variables
3626 @c man begin ENVIRONMENT
3628 You can change the behaviour of @command{ld} with the environment variables
3629 @ifclear SingleFormat
3632 @code{LDEMULATION} and @code{COLLECT_NO_DEMANGLE}.
3634 @ifclear SingleFormat
3636 @cindex default input format
3637 @code{GNUTARGET} determines the input-file object format if you don't
3638 use @samp{-b} (or its synonym @samp{--format}). Its value should be one
3639 of the BFD names for an input format (@pxref{BFD}). If there is no
3640 @code{GNUTARGET} in the environment, @command{ld} uses the natural format
3641 of the target. If @code{GNUTARGET} is set to @code{default} then BFD
3642 attempts to discover the input format by examining binary input files;
3643 this method often succeeds, but there are potential ambiguities, since
3644 there is no method of ensuring that the magic number used to specify
3645 object-file formats is unique. However, the configuration procedure for
3646 BFD on each system places the conventional format for that system first
3647 in the search-list, so ambiguities are resolved in favor of convention.
3651 @cindex default emulation
3652 @cindex emulation, default
3653 @code{LDEMULATION} determines the default emulation if you don't use the
3654 @samp{-m} option. The emulation can affect various aspects of linker
3655 behaviour, particularly the default linker script. You can list the
3656 available emulations with the @samp{--verbose} or @samp{-V} options. If
3657 the @samp{-m} option is not used, and the @code{LDEMULATION} environment
3658 variable is not defined, the default emulation depends upon how the
3659 linker was configured.
3661 @kindex COLLECT_NO_DEMANGLE
3662 @cindex demangling, default
3663 Normally, the linker will default to demangling symbols. However, if
3664 @code{COLLECT_NO_DEMANGLE} is set in the environment, then it will
3665 default to not demangling symbols. This environment variable is used in
3666 a similar fashion by the @code{gcc} linker wrapper program. The default
3667 may be overridden by the @samp{--demangle} and @samp{--no-demangle}
3674 @chapter Linker Scripts
3677 @cindex linker scripts
3678 @cindex command files
3679 Every link is controlled by a @dfn{linker script}. This script is
3680 written in the linker command language.
3682 The main purpose of the linker script is to describe how the sections in
3683 the input files should be mapped into the output file, and to control
3684 the memory layout of the output file. Most linker scripts do nothing
3685 more than this. However, when necessary, the linker script can also
3686 direct the linker to perform many other operations, using the commands
3689 The linker always uses a linker script. If you do not supply one
3690 yourself, the linker will use a default script that is compiled into the
3691 linker executable. You can use the @samp{--verbose} command-line option
3692 to display the default linker script. Certain command-line options,
3693 such as @samp{-r} or @samp{-N}, will affect the default linker script.
3695 You may supply your own linker script by using the @samp{-T} command
3696 line option. When you do this, your linker script will replace the
3697 default linker script.
3699 You may also use linker scripts implicitly by naming them as input files
3700 to the linker, as though they were files to be linked. @xref{Implicit
3704 * Basic Script Concepts:: Basic Linker Script Concepts
3705 * Script Format:: Linker Script Format
3706 * Simple Example:: Simple Linker Script Example
3707 * Simple Commands:: Simple Linker Script Commands
3708 * Assignments:: Assigning Values to Symbols
3709 * SECTIONS:: SECTIONS Command
3710 * MEMORY:: MEMORY Command
3711 * PHDRS:: PHDRS Command
3712 * VERSION:: VERSION Command
3713 * Expressions:: Expressions in Linker Scripts
3714 * Implicit Linker Scripts:: Implicit Linker Scripts
3717 @node Basic Script Concepts
3718 @section Basic Linker Script Concepts
3719 @cindex linker script concepts
3720 We need to define some basic concepts and vocabulary in order to
3721 describe the linker script language.
3723 The linker combines input files into a single output file. The output
3724 file and each input file are in a special data format known as an
3725 @dfn{object file format}. Each file is called an @dfn{object file}.
3726 The output file is often called an @dfn{executable}, but for our
3727 purposes we will also call it an object file. Each object file has,
3728 among other things, a list of @dfn{sections}. We sometimes refer to a
3729 section in an input file as an @dfn{input section}; similarly, a section
3730 in the output file is an @dfn{output section}.
3732 Each section in an object file has a name and a size. Most sections
3733 also have an associated block of data, known as the @dfn{section
3734 contents}. A section may be marked as @dfn{loadable}, which means that
3735 the contents should be loaded into memory when the output file is run.
3736 A section with no contents may be @dfn{allocatable}, which means that an
3737 area in memory should be set aside, but nothing in particular should be
3738 loaded there (in some cases this memory must be zeroed out). A section
3739 which is neither loadable nor allocatable typically contains some sort
3740 of debugging information.
3742 Every loadable or allocatable output section has two addresses. The
3743 first is the @dfn{VMA}, or virtual memory address. This is the address
3744 the section will have when the output file is run. The second is the
3745 @dfn{LMA}, or load memory address. This is the address at which the
3746 section will be loaded. In most cases the two addresses will be the
3747 same. An example of when they might be different is when a data section
3748 is loaded into ROM, and then copied into RAM when the program starts up
3749 (this technique is often used to initialize global variables in a ROM
3750 based system). In this case the ROM address would be the LMA, and the
3751 RAM address would be the VMA.
3753 You can see the sections in an object file by using the @code{objdump}
3754 program with the @samp{-h} option.
3756 Every object file also has a list of @dfn{symbols}, known as the
3757 @dfn{symbol table}. A symbol may be defined or undefined. Each symbol
3758 has a name, and each defined symbol has an address, among other
3759 information. If you compile a C or C++ program into an object file, you
3760 will get a defined symbol for every defined function and global or
3761 static variable. Every undefined function or global variable which is
3762 referenced in the input file will become an undefined symbol.
3764 You can see the symbols in an object file by using the @code{nm}
3765 program, or by using the @code{objdump} program with the @samp{-t}
3769 @section Linker Script Format
3770 @cindex linker script format
3771 Linker scripts are text files.
3773 You write a linker script as a series of commands. Each command is
3774 either a keyword, possibly followed by arguments, or an assignment to a
3775 symbol. You may separate commands using semicolons. Whitespace is
3778 Strings such as file or format names can normally be entered directly.
3779 If the file name contains a character such as a comma which would
3780 otherwise serve to separate file names, you may put the file name in
3781 double quotes. There is no way to use a double quote character in a
3784 You may include comments in linker scripts just as in C, delimited by
3785 @samp{/*} and @samp{*/}. As in C, comments are syntactically equivalent
3788 @node Simple Example
3789 @section Simple Linker Script Example
3790 @cindex linker script example
3791 @cindex example of linker script
3792 Many linker scripts are fairly simple.
3794 The simplest possible linker script has just one command:
3795 @samp{SECTIONS}. You use the @samp{SECTIONS} command to describe the
3796 memory layout of the output file.
3798 The @samp{SECTIONS} command is a powerful command. Here we will
3799 describe a simple use of it. Let's assume your program consists only of
3800 code, initialized data, and uninitialized data. These will be in the
3801 @samp{.text}, @samp{.data}, and @samp{.bss} sections, respectively.
3802 Let's assume further that these are the only sections which appear in
3805 For this example, let's say that the code should be loaded at address
3806 0x10000, and that the data should start at address 0x8000000. Here is a
3807 linker script which will do that:
3812 .text : @{ *(.text) @}
3814 .data : @{ *(.data) @}
3815 .bss : @{ *(.bss) @}
3819 You write the @samp{SECTIONS} command as the keyword @samp{SECTIONS},
3820 followed by a series of symbol assignments and output section
3821 descriptions enclosed in curly braces.
3823 The first line inside the @samp{SECTIONS} command of the above example
3824 sets the value of the special symbol @samp{.}, which is the location
3825 counter. If you do not specify the address of an output section in some
3826 other way (other ways are described later), the address is set from the
3827 current value of the location counter. The location counter is then
3828 incremented by the size of the output section. At the start of the
3829 @samp{SECTIONS} command, the location counter has the value @samp{0}.
3831 The second line defines an output section, @samp{.text}. The colon is
3832 required syntax which may be ignored for now. Within the curly braces
3833 after the output section name, you list the names of the input sections
3834 which should be placed into this output section. The @samp{*} is a
3835 wildcard which matches any file name. The expression @samp{*(.text)}
3836 means all @samp{.text} input sections in all input files.
3838 Since the location counter is @samp{0x10000} when the output section
3839 @samp{.text} is defined, the linker will set the address of the
3840 @samp{.text} section in the output file to be @samp{0x10000}.
3842 The remaining lines define the @samp{.data} and @samp{.bss} sections in
3843 the output file. The linker will place the @samp{.data} output section
3844 at address @samp{0x8000000}. After the linker places the @samp{.data}
3845 output section, the value of the location counter will be
3846 @samp{0x8000000} plus the size of the @samp{.data} output section. The
3847 effect is that the linker will place the @samp{.bss} output section
3848 immediately after the @samp{.data} output section in memory.
3850 The linker will ensure that each output section has the required
3851 alignment, by increasing the location counter if necessary. In this
3852 example, the specified addresses for the @samp{.text} and @samp{.data}
3853 sections will probably satisfy any alignment constraints, but the linker
3854 may have to create a small gap between the @samp{.data} and @samp{.bss}
3857 That's it! That's a simple and complete linker script.
3859 @node Simple Commands
3860 @section Simple Linker Script Commands
3861 @cindex linker script simple commands
3862 In this section we describe the simple linker script commands.
3865 * Entry Point:: Setting the entry point
3866 * File Commands:: Commands dealing with files
3867 @ifclear SingleFormat
3868 * Format Commands:: Commands dealing with object file formats
3871 * REGION_ALIAS:: Assign alias names to memory regions
3872 * Miscellaneous Commands:: Other linker script commands
3876 @subsection Setting the Entry Point
3877 @kindex ENTRY(@var{symbol})
3878 @cindex start of execution
3879 @cindex first instruction
3881 The first instruction to execute in a program is called the @dfn{entry
3882 point}. You can use the @code{ENTRY} linker script command to set the
3883 entry point. The argument is a symbol name:
3888 There are several ways to set the entry point. The linker will set the
3889 entry point by trying each of the following methods in order, and
3890 stopping when one of them succeeds:
3893 the @samp{-e} @var{entry} command-line option;
3895 the @code{ENTRY(@var{symbol})} command in a linker script;
3897 the value of a target-specific symbol, if it is defined; For many
3898 targets this is @code{start}, but PE- and BeOS-based systems for example
3899 check a list of possible entry symbols, matching the first one found.
3901 the address of the first byte of the @samp{.text} section, if present;
3903 The address @code{0}.
3907 @subsection Commands Dealing with Files
3908 @cindex linker script file commands
3909 Several linker script commands deal with files.
3912 @item INCLUDE @var{filename}
3913 @kindex INCLUDE @var{filename}
3914 @cindex including a linker script
3915 Include the linker script @var{filename} at this point. The file will
3916 be searched for in the current directory, and in any directory specified
3917 with the @option{-L} option. You can nest calls to @code{INCLUDE} up to
3920 You can place @code{INCLUDE} directives at the top level, in @code{MEMORY} or
3921 @code{SECTIONS} commands, or in output section descriptions.
3923 @item INPUT(@var{file}, @var{file}, @dots{})
3924 @itemx INPUT(@var{file} @var{file} @dots{})
3925 @kindex INPUT(@var{files})
3926 @cindex input files in linker scripts
3927 @cindex input object files in linker scripts
3928 @cindex linker script input object files
3929 The @code{INPUT} command directs the linker to include the named files
3930 in the link, as though they were named on the command line.
3932 For example, if you always want to include @file{subr.o} any time you do
3933 a link, but you can't be bothered to put it on every link command line,
3934 then you can put @samp{INPUT (subr.o)} in your linker script.
3936 In fact, if you like, you can list all of your input files in the linker
3937 script, and then invoke the linker with nothing but a @samp{-T} option.
3939 In case a @dfn{sysroot prefix} is configured, and the filename starts
3940 with the @samp{/} character, and the script being processed was
3941 located inside the @dfn{sysroot prefix}, the filename will be looked
3942 for in the @dfn{sysroot prefix}. The @dfn{sysroot prefix} can also be forced by specifying
3943 @code{=} as the first character in the filename path, or prefixing the
3944 filename path with @code{$SYSROOT}. See also the description of
3945 @samp{-L} in @ref{Options,,Command-line Options}.
3947 If a @dfn{sysroot prefix} is not used then the linker will try to open
3948 the file in the directory containing the linker script. If it is not
3949 found the linker will then search the current directory. If it is still
3950 not found the linker will search through the archive library search
3953 If you use @samp{INPUT (-l@var{file})}, @command{ld} will transform the
3954 name to @code{lib@var{file}.a}, as with the command-line argument
3957 When you use the @code{INPUT} command in an implicit linker script, the
3958 files will be included in the link at the point at which the linker
3959 script file is included. This can affect archive searching.
3961 @item GROUP(@var{file}, @var{file}, @dots{})
3962 @itemx GROUP(@var{file} @var{file} @dots{})
3963 @kindex GROUP(@var{files})
3964 @cindex grouping input files
3965 The @code{GROUP} command is like @code{INPUT}, except that the named
3966 files should all be archives, and they are searched repeatedly until no
3967 new undefined references are created. See the description of @samp{-(}
3968 in @ref{Options,,Command-line Options}.
3970 @item AS_NEEDED(@var{file}, @var{file}, @dots{})
3971 @itemx AS_NEEDED(@var{file} @var{file} @dots{})
3972 @kindex AS_NEEDED(@var{files})
3973 This construct can appear only inside of the @code{INPUT} or @code{GROUP}
3974 commands, among other filenames. The files listed will be handled
3975 as if they appear directly in the @code{INPUT} or @code{GROUP} commands,
3976 with the exception of ELF shared libraries, that will be added only
3977 when they are actually needed. This construct essentially enables
3978 @option{--as-needed} option for all the files listed inside of it
3979 and restores previous @option{--as-needed} resp. @option{--no-as-needed}
3982 @item OUTPUT(@var{filename})
3983 @kindex OUTPUT(@var{filename})
3984 @cindex output file name in linker script
3985 The @code{OUTPUT} command names the output file. Using
3986 @code{OUTPUT(@var{filename})} in the linker script is exactly like using
3987 @samp{-o @var{filename}} on the command line (@pxref{Options,,Command
3988 Line Options}). If both are used, the command-line option takes
3991 You can use the @code{OUTPUT} command to define a default name for the
3992 output file other than the usual default of @file{a.out}.
3994 @item SEARCH_DIR(@var{path})
3995 @kindex SEARCH_DIR(@var{path})
3996 @cindex library search path in linker script
3997 @cindex archive search path in linker script
3998 @cindex search path in linker script
3999 The @code{SEARCH_DIR} command adds @var{path} to the list of paths where
4000 @command{ld} looks for archive libraries. Using
4001 @code{SEARCH_DIR(@var{path})} is exactly like using @samp{-L @var{path}}
4002 on the command line (@pxref{Options,,Command-line Options}). If both
4003 are used, then the linker will search both paths. Paths specified using
4004 the command-line option are searched first.
4006 @item STARTUP(@var{filename})
4007 @kindex STARTUP(@var{filename})
4008 @cindex first input file
4009 The @code{STARTUP} command is just like the @code{INPUT} command, except
4010 that @var{filename} will become the first input file to be linked, as
4011 though it were specified first on the command line. This may be useful
4012 when using a system in which the entry point is always the start of the
4016 @ifclear SingleFormat
4017 @node Format Commands
4018 @subsection Commands Dealing with Object File Formats
4019 A couple of linker script commands deal with object file formats.
4022 @item OUTPUT_FORMAT(@var{bfdname})
4023 @itemx OUTPUT_FORMAT(@var{default}, @var{big}, @var{little})
4024 @kindex OUTPUT_FORMAT(@var{bfdname})
4025 @cindex output file format in linker script
4026 The @code{OUTPUT_FORMAT} command names the BFD format to use for the
4027 output file (@pxref{BFD}). Using @code{OUTPUT_FORMAT(@var{bfdname})} is
4028 exactly like using @samp{--oformat @var{bfdname}} on the command line
4029 (@pxref{Options,,Command-line Options}). If both are used, the command
4030 line option takes precedence.
4032 You can use @code{OUTPUT_FORMAT} with three arguments to use different
4033 formats based on the @samp{-EB} and @samp{-EL} command-line options.
4034 This permits the linker script to set the output format based on the
4037 If neither @samp{-EB} nor @samp{-EL} are used, then the output format
4038 will be the first argument, @var{default}. If @samp{-EB} is used, the
4039 output format will be the second argument, @var{big}. If @samp{-EL} is
4040 used, the output format will be the third argument, @var{little}.
4042 For example, the default linker script for the MIPS ELF target uses this
4045 OUTPUT_FORMAT(elf32-bigmips, elf32-bigmips, elf32-littlemips)
4047 This says that the default format for the output file is
4048 @samp{elf32-bigmips}, but if the user uses the @samp{-EL} command-line
4049 option, the output file will be created in the @samp{elf32-littlemips}
4052 @item TARGET(@var{bfdname})
4053 @kindex TARGET(@var{bfdname})
4054 @cindex input file format in linker script
4055 The @code{TARGET} command names the BFD format to use when reading input
4056 files. It affects subsequent @code{INPUT} and @code{GROUP} commands.
4057 This command is like using @samp{-b @var{bfdname}} on the command line
4058 (@pxref{Options,,Command-line Options}). If the @code{TARGET} command
4059 is used but @code{OUTPUT_FORMAT} is not, then the last @code{TARGET}
4060 command is also used to set the format for the output file. @xref{BFD}.
4065 @subsection Assign alias names to memory regions
4066 @kindex REGION_ALIAS(@var{alias}, @var{region})
4067 @cindex region alias
4068 @cindex region names
4070 Alias names can be added to existing memory regions created with the
4071 @ref{MEMORY} command. Each name corresponds to at most one memory region.
4074 REGION_ALIAS(@var{alias}, @var{region})
4077 The @code{REGION_ALIAS} function creates an alias name @var{alias} for the
4078 memory region @var{region}. This allows a flexible mapping of output sections
4079 to memory regions. An example follows.
4081 Suppose we have an application for embedded systems which come with various
4082 memory storage devices. All have a general purpose, volatile memory @code{RAM}
4083 that allows code execution or data storage. Some may have a read-only,
4084 non-volatile memory @code{ROM} that allows code execution and read-only data
4085 access. The last variant is a read-only, non-volatile memory @code{ROM2} with
4086 read-only data access and no code execution capability. We have four output
4091 @code{.text} program code;
4093 @code{.rodata} read-only data;
4095 @code{.data} read-write initialized data;
4097 @code{.bss} read-write zero initialized data.
4100 The goal is to provide a linker command file that contains a system independent
4101 part defining the output sections and a system dependent part mapping the
4102 output sections to the memory regions available on the system. Our embedded
4103 systems come with three different memory setups @code{A}, @code{B} and
4105 @multitable @columnfractions .25 .25 .25 .25
4106 @item Section @tab Variant A @tab Variant B @tab Variant C
4107 @item .text @tab RAM @tab ROM @tab ROM
4108 @item .rodata @tab RAM @tab ROM @tab ROM2
4109 @item .data @tab RAM @tab RAM/ROM @tab RAM/ROM2
4110 @item .bss @tab RAM @tab RAM @tab RAM
4112 The notation @code{RAM/ROM} or @code{RAM/ROM2} means that this section is
4113 loaded into region @code{ROM} or @code{ROM2} respectively. Please note that
4114 the load address of the @code{.data} section starts in all three variants at
4115 the end of the @code{.rodata} section.
4117 The base linker script that deals with the output sections follows. It
4118 includes the system dependent @code{linkcmds.memory} file that describes the
4121 INCLUDE linkcmds.memory
4134 .data : AT (rodata_end)
4139 data_size = SIZEOF(.data);
4140 data_load_start = LOADADDR(.data);
4148 Now we need three different @code{linkcmds.memory} files to define memory
4149 regions and alias names. The content of @code{linkcmds.memory} for the three
4150 variants @code{A}, @code{B} and @code{C}:
4153 Here everything goes into the @code{RAM}.
4157 RAM : ORIGIN = 0, LENGTH = 4M
4160 REGION_ALIAS("REGION_TEXT", RAM);
4161 REGION_ALIAS("REGION_RODATA", RAM);
4162 REGION_ALIAS("REGION_DATA", RAM);
4163 REGION_ALIAS("REGION_BSS", RAM);
4166 Program code and read-only data go into the @code{ROM}. Read-write data goes
4167 into the @code{RAM}. An image of the initialized data is loaded into the
4168 @code{ROM} and will be copied during system start into the @code{RAM}.
4172 ROM : ORIGIN = 0, LENGTH = 3M
4173 RAM : ORIGIN = 0x10000000, LENGTH = 1M
4176 REGION_ALIAS("REGION_TEXT", ROM);
4177 REGION_ALIAS("REGION_RODATA", ROM);
4178 REGION_ALIAS("REGION_DATA", RAM);
4179 REGION_ALIAS("REGION_BSS", RAM);
4182 Program code goes into the @code{ROM}. Read-only data goes into the
4183 @code{ROM2}. Read-write data goes into the @code{RAM}. An image of the
4184 initialized data is loaded into the @code{ROM2} and will be copied during
4185 system start into the @code{RAM}.
4189 ROM : ORIGIN = 0, LENGTH = 2M
4190 ROM2 : ORIGIN = 0x10000000, LENGTH = 1M
4191 RAM : ORIGIN = 0x20000000, LENGTH = 1M
4194 REGION_ALIAS("REGION_TEXT", ROM);
4195 REGION_ALIAS("REGION_RODATA", ROM2);
4196 REGION_ALIAS("REGION_DATA", RAM);
4197 REGION_ALIAS("REGION_BSS", RAM);
4201 It is possible to write a common system initialization routine to copy the
4202 @code{.data} section from @code{ROM} or @code{ROM2} into the @code{RAM} if
4207 extern char data_start [];
4208 extern char data_size [];
4209 extern char data_load_start [];
4211 void copy_data(void)
4213 if (data_start != data_load_start)
4215 memcpy(data_start, data_load_start, (size_t) data_size);
4220 @node Miscellaneous Commands
4221 @subsection Other Linker Script Commands
4222 There are a few other linker scripts commands.
4225 @item ASSERT(@var{exp}, @var{message})
4227 @cindex assertion in linker script
4228 Ensure that @var{exp} is non-zero. If it is zero, then exit the linker
4229 with an error code, and print @var{message}.
4231 Note that assertions are checked before the final stages of linking
4232 take place. This means that expressions involving symbols PROVIDEd
4233 inside section definitions will fail if the user has not set values
4234 for those symbols. The only exception to this rule is PROVIDEd
4235 symbols that just reference dot. Thus an assertion like this:
4240 PROVIDE (__stack = .);
4241 PROVIDE (__stack_size = 0x100);
4242 ASSERT ((__stack > (_end + __stack_size)), "Error: No room left for the stack");
4246 will fail if @code{__stack_size} is not defined elsewhere. Symbols
4247 PROVIDEd outside of section definitions are evaluated earlier, so they
4248 can be used inside ASSERTions. Thus:
4251 PROVIDE (__stack_size = 0x100);
4254 PROVIDE (__stack = .);
4255 ASSERT ((__stack > (_end + __stack_size)), "Error: No room left for the stack");
4261 @item EXTERN(@var{symbol} @var{symbol} @dots{})
4263 @cindex undefined symbol in linker script
4264 Force @var{symbol} to be entered in the output file as an undefined
4265 symbol. Doing this may, for example, trigger linking of additional
4266 modules from standard libraries. You may list several @var{symbol}s for
4267 each @code{EXTERN}, and you may use @code{EXTERN} multiple times. This
4268 command has the same effect as the @samp{-u} command-line option.
4270 @item FORCE_COMMON_ALLOCATION
4271 @kindex FORCE_COMMON_ALLOCATION
4272 @cindex common allocation in linker script
4273 This command has the same effect as the @samp{-d} command-line option:
4274 to make @command{ld} assign space to common symbols even if a relocatable
4275 output file is specified (@samp{-r}).
4277 @item INHIBIT_COMMON_ALLOCATION
4278 @kindex INHIBIT_COMMON_ALLOCATION
4279 @cindex common allocation in linker script
4280 This command has the same effect as the @samp{--no-define-common}
4281 command-line option: to make @code{ld} omit the assignment of addresses
4282 to common symbols even for a non-relocatable output file.
4284 @item FORCE_GROUP_ALLOCATION
4285 @kindex FORCE_GROUP_ALLOCATION
4286 @cindex group allocation in linker script
4287 @cindex section groups
4289 This command has the same effect as the
4290 @samp{--force-group-allocation} command-line option: to make
4291 @command{ld} place section group members like normal input sections,
4292 and to delete the section groups even if a relocatable output file is
4293 specified (@samp{-r}).
4295 @item INSERT [ AFTER | BEFORE ] @var{output_section}
4297 @cindex insert user script into default script
4298 This command is typically used in a script specified by @samp{-T} to
4299 augment the default @code{SECTIONS} with, for example, overlays. It
4300 inserts all prior linker script statements after (or before)
4301 @var{output_section}, and also causes @samp{-T} to not override the
4302 default linker script. The exact insertion point is as for orphan
4303 sections. @xref{Location Counter}. The insertion happens after the
4304 linker has mapped input sections to output sections. Prior to the
4305 insertion, since @samp{-T} scripts are parsed before the default
4306 linker script, statements in the @samp{-T} script occur before the
4307 default linker script statements in the internal linker representation
4308 of the script. In particular, input section assignments will be made
4309 to @samp{-T} output sections before those in the default script. Here
4310 is an example of how a @samp{-T} script using @code{INSERT} might look:
4317 .ov1 @{ ov1*(.text) @}
4318 .ov2 @{ ov2*(.text) @}
4324 @item NOCROSSREFS(@var{section} @var{section} @dots{})
4325 @kindex NOCROSSREFS(@var{sections})
4326 @cindex cross references
4327 This command may be used to tell @command{ld} to issue an error about any
4328 references among certain output sections.
4330 In certain types of programs, particularly on embedded systems when
4331 using overlays, when one section is loaded into memory, another section
4332 will not be. Any direct references between the two sections would be
4333 errors. For example, it would be an error if code in one section called
4334 a function defined in the other section.
4336 The @code{NOCROSSREFS} command takes a list of output section names. If
4337 @command{ld} detects any cross references between the sections, it reports
4338 an error and returns a non-zero exit status. Note that the
4339 @code{NOCROSSREFS} command uses output section names, not input section
4342 @item NOCROSSREFS_TO(@var{tosection} @var{fromsection} @dots{})
4343 @kindex NOCROSSREFS_TO(@var{tosection} @var{fromsections})
4344 @cindex cross references
4345 This command may be used to tell @command{ld} to issue an error about any
4346 references to one section from a list of other sections.
4348 The @code{NOCROSSREFS} command is useful when ensuring that two or more
4349 output sections are entirely independent but there are situations where
4350 a one-way dependency is needed. For example, in a multi-core application
4351 there may be shared code that can be called from each core but for safety
4352 must never call back.
4354 The @code{NOCROSSREFS_TO} command takes a list of output section names.
4355 The first section can not be referenced from any of the other sections.
4356 If @command{ld} detects any references to the first section from any of
4357 the other sections, it reports an error and returns a non-zero exit
4358 status. Note that the @code{NOCROSSREFS_TO} command uses output section
4359 names, not input section names.
4361 @ifclear SingleFormat
4362 @item OUTPUT_ARCH(@var{bfdarch})
4363 @kindex OUTPUT_ARCH(@var{bfdarch})
4364 @cindex machine architecture
4365 @cindex architecture
4366 Specify a particular output machine architecture. The argument is one
4367 of the names used by the BFD library (@pxref{BFD}). You can see the
4368 architecture of an object file by using the @code{objdump} program with
4369 the @samp{-f} option.
4372 @item LD_FEATURE(@var{string})
4373 @kindex LD_FEATURE(@var{string})
4374 This command may be used to modify @command{ld} behavior. If
4375 @var{string} is @code{"SANE_EXPR"} then absolute symbols and numbers
4376 in a script are simply treated as numbers everywhere.
4377 @xref{Expression Section}.
4381 @section Assigning Values to Symbols
4382 @cindex assignment in scripts
4383 @cindex symbol definition, scripts
4384 @cindex variables, defining
4385 You may assign a value to a symbol in a linker script. This will define
4386 the symbol and place it into the symbol table with a global scope.
4389 * Simple Assignments:: Simple Assignments
4392 * PROVIDE_HIDDEN:: PROVIDE_HIDDEN
4393 * Source Code Reference:: How to use a linker script defined symbol in source code
4396 @node Simple Assignments
4397 @subsection Simple Assignments
4399 You may assign to a symbol using any of the C assignment operators:
4402 @item @var{symbol} = @var{expression} ;
4403 @itemx @var{symbol} += @var{expression} ;
4404 @itemx @var{symbol} -= @var{expression} ;
4405 @itemx @var{symbol} *= @var{expression} ;
4406 @itemx @var{symbol} /= @var{expression} ;
4407 @itemx @var{symbol} <<= @var{expression} ;
4408 @itemx @var{symbol} >>= @var{expression} ;
4409 @itemx @var{symbol} &= @var{expression} ;
4410 @itemx @var{symbol} |= @var{expression} ;
4413 The first case will define @var{symbol} to the value of
4414 @var{expression}. In the other cases, @var{symbol} must already be
4415 defined, and the value will be adjusted accordingly.
4417 The special symbol name @samp{.} indicates the location counter. You
4418 may only use this within a @code{SECTIONS} command. @xref{Location Counter}.
4420 The semicolon after @var{expression} is required.
4422 Expressions are defined below; see @ref{Expressions}.
4424 You may write symbol assignments as commands in their own right, or as
4425 statements within a @code{SECTIONS} command, or as part of an output
4426 section description in a @code{SECTIONS} command.
4428 The section of the symbol will be set from the section of the
4429 expression; for more information, see @ref{Expression Section}.
4431 Here is an example showing the three different places that symbol
4432 assignments may be used:
4443 _bdata = (. + 3) & ~ 3;
4444 .data : @{ *(.data) @}
4448 In this example, the symbol @samp{floating_point} will be defined as
4449 zero. The symbol @samp{_etext} will be defined as the address following
4450 the last @samp{.text} input section. The symbol @samp{_bdata} will be
4451 defined as the address following the @samp{.text} output section aligned
4452 upward to a 4 byte boundary.
4457 For ELF targeted ports, define a symbol that will be hidden and won't be
4458 exported. The syntax is @code{HIDDEN(@var{symbol} = @var{expression})}.
4460 Here is the example from @ref{Simple Assignments}, rewritten to use
4464 HIDDEN(floating_point = 0);
4472 HIDDEN(_bdata = (. + 3) & ~ 3);
4473 .data : @{ *(.data) @}
4477 In this case none of the three symbols will be visible outside this module.
4482 In some cases, it is desirable for a linker script to define a symbol
4483 only if it is referenced and is not defined by any object included in
4484 the link. For example, traditional linkers defined the symbol
4485 @samp{etext}. However, ANSI C requires that the user be able to use
4486 @samp{etext} as a function name without encountering an error. The
4487 @code{PROVIDE} keyword may be used to define a symbol, such as
4488 @samp{etext}, only if it is referenced but not defined. The syntax is
4489 @code{PROVIDE(@var{symbol} = @var{expression})}.
4491 Here is an example of using @code{PROVIDE} to define @samp{etext}:
4504 In this example, if the program defines @samp{_etext} (with a leading
4505 underscore), the linker will give a multiple definition diagnostic. If,
4506 on the other hand, the program defines @samp{etext} (with no leading
4507 underscore), the linker will silently use the definition in the program.
4508 If the program references @samp{etext} but does not define it, the
4509 linker will use the definition in the linker script.
4511 Note - the @code{PROVIDE} directive considers a common symbol to be
4512 defined, even though such a symbol could be combined with the symbol
4513 that the @code{PROVIDE} would create. This is particularly important
4514 when considering constructor and destructor list symbols such as
4515 @samp{__CTOR_LIST__} as these are often defined as common symbols.
4517 @node PROVIDE_HIDDEN
4518 @subsection PROVIDE_HIDDEN
4519 @cindex PROVIDE_HIDDEN
4520 Similar to @code{PROVIDE}. For ELF targeted ports, the symbol will be
4521 hidden and won't be exported.
4523 @node Source Code Reference
4524 @subsection Source Code Reference
4526 Accessing a linker script defined variable from source code is not
4527 intuitive. In particular a linker script symbol is not equivalent to
4528 a variable declaration in a high level language, it is instead a
4529 symbol that does not have a value.
4531 Before going further, it is important to note that compilers often
4532 transform names in the source code into different names when they are
4533 stored in the symbol table. For example, Fortran compilers commonly
4534 prepend or append an underscore, and C++ performs extensive @samp{name
4535 mangling}. Therefore there might be a discrepancy between the name
4536 of a variable as it is used in source code and the name of the same
4537 variable as it is defined in a linker script. For example in C a
4538 linker script variable might be referred to as:
4544 But in the linker script it might be defined as:
4550 In the remaining examples however it is assumed that no name
4551 transformation has taken place.
4553 When a symbol is declared in a high level language such as C, two
4554 things happen. The first is that the compiler reserves enough space
4555 in the program's memory to hold the @emph{value} of the symbol. The
4556 second is that the compiler creates an entry in the program's symbol
4557 table which holds the symbol's @emph{address}. ie the symbol table
4558 contains the address of the block of memory holding the symbol's
4559 value. So for example the following C declaration, at file scope:
4565 creates an entry called @samp{foo} in the symbol table. This entry
4566 holds the address of an @samp{int} sized block of memory where the
4567 number 1000 is initially stored.
4569 When a program references a symbol the compiler generates code that
4570 first accesses the symbol table to find the address of the symbol's
4571 memory block and then code to read the value from that memory block.
4578 looks up the symbol @samp{foo} in the symbol table, gets the address
4579 associated with this symbol and then writes the value 1 into that
4586 looks up the symbol @samp{foo} in the symbol table, gets its address
4587 and then copies this address into the block of memory associated with
4588 the variable @samp{a}.
4590 Linker scripts symbol declarations, by contrast, create an entry in
4591 the symbol table but do not assign any memory to them. Thus they are
4592 an address without a value. So for example the linker script definition:
4598 creates an entry in the symbol table called @samp{foo} which holds
4599 the address of memory location 1000, but nothing special is stored at
4600 address 1000. This means that you cannot access the @emph{value} of a
4601 linker script defined symbol - it has no value - all you can do is
4602 access the @emph{address} of a linker script defined symbol.
4604 Hence when you are using a linker script defined symbol in source code
4605 you should always take the address of the symbol, and never attempt to
4606 use its value. For example suppose you want to copy the contents of a
4607 section of memory called .ROM into a section called .FLASH and the
4608 linker script contains these declarations:
4612 start_of_ROM = .ROM;
4613 end_of_ROM = .ROM + sizeof (.ROM);
4614 start_of_FLASH = .FLASH;
4618 Then the C source code to perform the copy would be:
4622 extern char start_of_ROM, end_of_ROM, start_of_FLASH;
4624 memcpy (& start_of_FLASH, & start_of_ROM, & end_of_ROM - & start_of_ROM);
4628 Note the use of the @samp{&} operators. These are correct.
4629 Alternatively the symbols can be treated as the names of vectors or
4630 arrays and then the code will again work as expected:
4634 extern char start_of_ROM[], end_of_ROM[], start_of_FLASH[];
4636 memcpy (start_of_FLASH, start_of_ROM, end_of_ROM - start_of_ROM);
4640 Note how using this method does not require the use of @samp{&}
4644 @section SECTIONS Command
4646 The @code{SECTIONS} command tells the linker how to map input sections
4647 into output sections, and how to place the output sections in memory.
4649 The format of the @code{SECTIONS} command is:
4653 @var{sections-command}
4654 @var{sections-command}
4659 Each @var{sections-command} may of be one of the following:
4663 an @code{ENTRY} command (@pxref{Entry Point,,Entry command})
4665 a symbol assignment (@pxref{Assignments})
4667 an output section description
4669 an overlay description
4672 The @code{ENTRY} command and symbol assignments are permitted inside the
4673 @code{SECTIONS} command for convenience in using the location counter in
4674 those commands. This can also make the linker script easier to
4675 understand because you can use those commands at meaningful points in
4676 the layout of the output file.
4678 Output section descriptions and overlay descriptions are described
4681 If you do not use a @code{SECTIONS} command in your linker script, the
4682 linker will place each input section into an identically named output
4683 section in the order that the sections are first encountered in the
4684 input files. If all input sections are present in the first file, for
4685 example, the order of sections in the output file will match the order
4686 in the first input file. The first section will be at address zero.
4689 * Output Section Description:: Output section description
4690 * Output Section Name:: Output section name
4691 * Output Section Address:: Output section address
4692 * Input Section:: Input section description
4693 * Output Section Data:: Output section data
4694 * Output Section Keywords:: Output section keywords
4695 * Output Section Discarding:: Output section discarding
4696 * Output Section Attributes:: Output section attributes
4697 * Overlay Description:: Overlay description
4700 @node Output Section Description
4701 @subsection Output Section Description
4702 The full description of an output section looks like this:
4705 @var{section} [@var{address}] [(@var{type})] :
4707 [ALIGN(@var{section_align}) | ALIGN_WITH_INPUT]
4708 [SUBALIGN(@var{subsection_align})]
4711 @var{output-section-command}
4712 @var{output-section-command}
4714 @} [>@var{region}] [AT>@var{lma_region}] [:@var{phdr} :@var{phdr} @dots{}] [=@var{fillexp}] [,]
4718 Most output sections do not use most of the optional section attributes.
4720 The whitespace around @var{section} is required, so that the section
4721 name is unambiguous. The colon and the curly braces are also required.
4722 The comma at the end may be required if a @var{fillexp} is used and
4723 the next @var{sections-command} looks like a continuation of the expression.
4724 The line breaks and other white space are optional.
4726 Each @var{output-section-command} may be one of the following:
4730 a symbol assignment (@pxref{Assignments})
4732 an input section description (@pxref{Input Section})
4734 data values to include directly (@pxref{Output Section Data})
4736 a special output section keyword (@pxref{Output Section Keywords})
4739 @node Output Section Name
4740 @subsection Output Section Name
4741 @cindex name, section
4742 @cindex section name
4743 The name of the output section is @var{section}. @var{section} must
4744 meet the constraints of your output format. In formats which only
4745 support a limited number of sections, such as @code{a.out}, the name
4746 must be one of the names supported by the format (@code{a.out}, for
4747 example, allows only @samp{.text}, @samp{.data} or @samp{.bss}). If the
4748 output format supports any number of sections, but with numbers and not
4749 names (as is the case for Oasys), the name should be supplied as a
4750 quoted numeric string. A section name may consist of any sequence of
4751 characters, but a name which contains any unusual characters such as
4752 commas must be quoted.
4754 The output section name @samp{/DISCARD/} is special; @ref{Output Section
4757 @node Output Section Address
4758 @subsection Output Section Address
4759 @cindex address, section
4760 @cindex section address
4761 The @var{address} is an expression for the VMA (the virtual memory
4762 address) of the output section. This address is optional, but if it
4763 is provided then the output address will be set exactly as specified.
4765 If the output address is not specified then one will be chosen for the
4766 section, based on the heuristic below. This address will be adjusted
4767 to fit the alignment requirement of the output section. The
4768 alignment requirement is the strictest alignment of any input section
4769 contained within the output section.
4771 The output section address heuristic is as follows:
4775 If an output memory @var{region} is set for the section then it
4776 is added to this region and its address will be the next free address
4780 If the MEMORY command has been used to create a list of memory
4781 regions then the first region which has attributes compatible with the
4782 section is selected to contain it. The section's output address will
4783 be the next free address in that region; @ref{MEMORY}.
4786 If no memory regions were specified, or none match the section then
4787 the output address will be based on the current value of the location
4795 .text . : @{ *(.text) @}
4802 .text : @{ *(.text) @}
4806 are subtly different. The first will set the address of the
4807 @samp{.text} output section to the current value of the location
4808 counter. The second will set it to the current value of the location
4809 counter aligned to the strictest alignment of any of the @samp{.text}
4812 The @var{address} may be an arbitrary expression; @ref{Expressions}.
4813 For example, if you want to align the section on a 0x10 byte boundary,
4814 so that the lowest four bits of the section address are zero, you could
4815 do something like this:
4817 .text ALIGN(0x10) : @{ *(.text) @}
4820 This works because @code{ALIGN} returns the current location counter
4821 aligned upward to the specified value.
4823 Specifying @var{address} for a section will change the value of the
4824 location counter, provided that the section is non-empty. (Empty
4825 sections are ignored).
4828 @subsection Input Section Description
4829 @cindex input sections
4830 @cindex mapping input sections to output sections
4831 The most common output section command is an input section description.
4833 The input section description is the most basic linker script operation.
4834 You use output sections to tell the linker how to lay out your program
4835 in memory. You use input section descriptions to tell the linker how to
4836 map the input files into your memory layout.
4839 * Input Section Basics:: Input section basics
4840 * Input Section Wildcards:: Input section wildcard patterns
4841 * Input Section Common:: Input section for common symbols
4842 * Input Section Keep:: Input section and garbage collection
4843 * Input Section Example:: Input section example
4846 @node Input Section Basics
4847 @subsubsection Input Section Basics
4848 @cindex input section basics
4849 An input section description consists of a file name optionally followed
4850 by a list of section names in parentheses.
4852 The file name and the section name may be wildcard patterns, which we
4853 describe further below (@pxref{Input Section Wildcards}).
4855 The most common input section description is to include all input
4856 sections with a particular name in the output section. For example, to
4857 include all input @samp{.text} sections, you would write:
4862 Here the @samp{*} is a wildcard which matches any file name. To exclude a list
4863 @cindex EXCLUDE_FILE
4864 of files from matching the file name wildcard, EXCLUDE_FILE may be used to
4865 match all files except the ones specified in the EXCLUDE_FILE list. For
4868 EXCLUDE_FILE (*crtend.o *otherfile.o) *(.ctors)
4871 will cause all .ctors sections from all files except @file{crtend.o}
4872 and @file{otherfile.o} to be included. The EXCLUDE_FILE can also be
4873 placed inside the section list, for example:
4875 *(EXCLUDE_FILE (*crtend.o *otherfile.o) .ctors)
4878 The result of this is identically to the previous example. Supporting
4879 two syntaxes for EXCLUDE_FILE is useful if the section list contains
4880 more than one section, as described below.
4882 There are two ways to include more than one section:
4888 The difference between these is the order in which the @samp{.text} and
4889 @samp{.rdata} input sections will appear in the output section. In the
4890 first example, they will be intermingled, appearing in the same order as
4891 they are found in the linker input. In the second example, all
4892 @samp{.text} input sections will appear first, followed by all
4893 @samp{.rdata} input sections.
4895 When using EXCLUDE_FILE with more than one section, if the exclusion
4896 is within the section list then the exclusion only applies to the
4897 immediately following section, for example:
4899 *(EXCLUDE_FILE (*somefile.o) .text .rdata)
4902 will cause all @samp{.text} sections from all files except
4903 @file{somefile.o} to be included, while all @samp{.rdata} sections
4904 from all files, including @file{somefile.o}, will be included. To
4905 exclude the @samp{.rdata} sections from @file{somefile.o} the example
4906 could be modified to:
4908 *(EXCLUDE_FILE (*somefile.o) .text EXCLUDE_FILE (*somefile.o) .rdata)
4911 Alternatively, placing the EXCLUDE_FILE outside of the section list,
4912 before the input file selection, will cause the exclusion to apply for
4913 all sections. Thus the previous example can be rewritten as:
4915 EXCLUDE_FILE (*somefile.o) *(.text .rdata)
4918 You can specify a file name to include sections from a particular file.
4919 You would do this if one or more of your files contain special data that
4920 needs to be at a particular location in memory. For example:
4925 To refine the sections that are included based on the section flags
4926 of an input section, INPUT_SECTION_FLAGS may be used.
4928 Here is a simple example for using Section header flags for ELF sections:
4933 .text : @{ INPUT_SECTION_FLAGS (SHF_MERGE & SHF_STRINGS) *(.text) @}
4934 .text2 : @{ INPUT_SECTION_FLAGS (!SHF_WRITE) *(.text) @}
4939 In this example, the output section @samp{.text} will be comprised of any
4940 input section matching the name *(.text) whose section header flags
4941 @code{SHF_MERGE} and @code{SHF_STRINGS} are set. The output section
4942 @samp{.text2} will be comprised of any input section matching the name *(.text)
4943 whose section header flag @code{SHF_WRITE} is clear.
4945 You can also specify files within archives by writing a pattern
4946 matching the archive, a colon, then the pattern matching the file,
4947 with no whitespace around the colon.
4951 matches file within archive
4953 matches the whole archive
4955 matches file but not one in an archive
4958 Either one or both of @samp{archive} and @samp{file} can contain shell
4959 wildcards. On DOS based file systems, the linker will assume that a
4960 single letter followed by a colon is a drive specifier, so
4961 @samp{c:myfile.o} is a simple file specification, not @samp{myfile.o}
4962 within an archive called @samp{c}. @samp{archive:file} filespecs may
4963 also be used within an @code{EXCLUDE_FILE} list, but may not appear in
4964 other linker script contexts. For instance, you cannot extract a file
4965 from an archive by using @samp{archive:file} in an @code{INPUT}
4968 If you use a file name without a list of sections, then all sections in
4969 the input file will be included in the output section. This is not
4970 commonly done, but it may by useful on occasion. For example:
4975 When you use a file name which is not an @samp{archive:file} specifier
4976 and does not contain any wild card
4977 characters, the linker will first see if you also specified the file
4978 name on the linker command line or in an @code{INPUT} command. If you
4979 did not, the linker will attempt to open the file as an input file, as
4980 though it appeared on the command line. Note that this differs from an
4981 @code{INPUT} command, because the linker will not search for the file in
4982 the archive search path.
4984 @node Input Section Wildcards
4985 @subsubsection Input Section Wildcard Patterns
4986 @cindex input section wildcards
4987 @cindex wildcard file name patterns
4988 @cindex file name wildcard patterns
4989 @cindex section name wildcard patterns
4990 In an input section description, either the file name or the section
4991 name or both may be wildcard patterns.
4993 The file name of @samp{*} seen in many examples is a simple wildcard
4994 pattern for the file name.
4996 The wildcard patterns are like those used by the Unix shell.
5000 matches any number of characters
5002 matches any single character
5004 matches a single instance of any of the @var{chars}; the @samp{-}
5005 character may be used to specify a range of characters, as in
5006 @samp{[a-z]} to match any lower case letter
5008 quotes the following character
5011 When a file name is matched with a wildcard, the wildcard characters
5012 will not match a @samp{/} character (used to separate directory names on
5013 Unix). A pattern consisting of a single @samp{*} character is an
5014 exception; it will always match any file name, whether it contains a
5015 @samp{/} or not. In a section name, the wildcard characters will match
5016 a @samp{/} character.
5018 File name wildcard patterns only match files which are explicitly
5019 specified on the command line or in an @code{INPUT} command. The linker
5020 does not search directories to expand wildcards.
5022 If a file name matches more than one wildcard pattern, or if a file name
5023 appears explicitly and is also matched by a wildcard pattern, the linker
5024 will use the first match in the linker script. For example, this
5025 sequence of input section descriptions is probably in error, because the
5026 @file{data.o} rule will not be used:
5028 .data : @{ *(.data) @}
5029 .data1 : @{ data.o(.data) @}
5032 @cindex SORT_BY_NAME
5033 Normally, the linker will place files and sections matched by wildcards
5034 in the order in which they are seen during the link. You can change
5035 this by using the @code{SORT_BY_NAME} keyword, which appears before a wildcard
5036 pattern in parentheses (e.g., @code{SORT_BY_NAME(.text*)}). When the
5037 @code{SORT_BY_NAME} keyword is used, the linker will sort the files or sections
5038 into ascending order by name before placing them in the output file.
5040 @cindex SORT_BY_ALIGNMENT
5041 @code{SORT_BY_ALIGNMENT} is similar to @code{SORT_BY_NAME}.
5042 @code{SORT_BY_ALIGNMENT} will sort sections into descending order of
5043 alignment before placing them in the output file. Placing larger
5044 alignments before smaller alignments can reduce the amount of padding
5047 @cindex SORT_BY_INIT_PRIORITY
5048 @code{SORT_BY_INIT_PRIORITY} is also similar to @code{SORT_BY_NAME}.
5049 @code{SORT_BY_INIT_PRIORITY} will sort sections into ascending
5050 numerical order of the GCC init_priority attribute encoded in the
5051 section name before placing them in the output file. In
5052 @code{.init_array.NNNNN} and @code{.fini_array.NNNNN}, @code{NNNNN} is
5053 the init_priority. In @code{.ctors.NNNNN} and @code{.dtors.NNNNN},
5054 @code{NNNNN} is 65535 minus the init_priority.
5057 @code{SORT} is an alias for @code{SORT_BY_NAME}.
5059 When there are nested section sorting commands in linker script, there
5060 can be at most 1 level of nesting for section sorting commands.
5064 @code{SORT_BY_NAME} (@code{SORT_BY_ALIGNMENT} (wildcard section pattern)).
5065 It will sort the input sections by name first, then by alignment if two
5066 sections have the same name.
5068 @code{SORT_BY_ALIGNMENT} (@code{SORT_BY_NAME} (wildcard section pattern)).
5069 It will sort the input sections by alignment first, then by name if two
5070 sections have the same alignment.
5072 @code{SORT_BY_NAME} (@code{SORT_BY_NAME} (wildcard section pattern)) is
5073 treated the same as @code{SORT_BY_NAME} (wildcard section pattern).
5075 @code{SORT_BY_ALIGNMENT} (@code{SORT_BY_ALIGNMENT} (wildcard section pattern))
5076 is treated the same as @code{SORT_BY_ALIGNMENT} (wildcard section pattern).
5078 All other nested section sorting commands are invalid.
5081 When both command-line section sorting option and linker script
5082 section sorting command are used, section sorting command always
5083 takes precedence over the command-line option.
5085 If the section sorting command in linker script isn't nested, the
5086 command-line option will make the section sorting command to be
5087 treated as nested sorting command.
5091 @code{SORT_BY_NAME} (wildcard section pattern ) with
5092 @option{--sort-sections alignment} is equivalent to
5093 @code{SORT_BY_NAME} (@code{SORT_BY_ALIGNMENT} (wildcard section pattern)).
5095 @code{SORT_BY_ALIGNMENT} (wildcard section pattern) with
5096 @option{--sort-section name} is equivalent to
5097 @code{SORT_BY_ALIGNMENT} (@code{SORT_BY_NAME} (wildcard section pattern)).
5100 If the section sorting command in linker script is nested, the
5101 command-line option will be ignored.
5104 @code{SORT_NONE} disables section sorting by ignoring the command-line
5105 section sorting option.
5107 If you ever get confused about where input sections are going, use the
5108 @samp{-M} linker option to generate a map file. The map file shows
5109 precisely how input sections are mapped to output sections.
5111 This example shows how wildcard patterns might be used to partition
5112 files. This linker script directs the linker to place all @samp{.text}
5113 sections in @samp{.text} and all @samp{.bss} sections in @samp{.bss}.
5114 The linker will place the @samp{.data} section from all files beginning
5115 with an upper case character in @samp{.DATA}; for all other files, the
5116 linker will place the @samp{.data} section in @samp{.data}.
5120 .text : @{ *(.text) @}
5121 .DATA : @{ [A-Z]*(.data) @}
5122 .data : @{ *(.data) @}
5123 .bss : @{ *(.bss) @}
5128 @node Input Section Common
5129 @subsubsection Input Section for Common Symbols
5130 @cindex common symbol placement
5131 @cindex uninitialized data placement
5132 A special notation is needed for common symbols, because in many object
5133 file formats common symbols do not have a particular input section. The
5134 linker treats common symbols as though they are in an input section
5135 named @samp{COMMON}.
5137 You may use file names with the @samp{COMMON} section just as with any
5138 other input sections. You can use this to place common symbols from a
5139 particular input file in one section while common symbols from other
5140 input files are placed in another section.
5142 In most cases, common symbols in input files will be placed in the
5143 @samp{.bss} section in the output file. For example:
5145 .bss @{ *(.bss) *(COMMON) @}
5148 @cindex scommon section
5149 @cindex small common symbols
5150 Some object file formats have more than one type of common symbol. For
5151 example, the MIPS ELF object file format distinguishes standard common
5152 symbols and small common symbols. In this case, the linker will use a
5153 different special section name for other types of common symbols. In
5154 the case of MIPS ELF, the linker uses @samp{COMMON} for standard common
5155 symbols and @samp{.scommon} for small common symbols. This permits you
5156 to map the different types of common symbols into memory at different
5160 You will sometimes see @samp{[COMMON]} in old linker scripts. This
5161 notation is now considered obsolete. It is equivalent to
5164 @node Input Section Keep
5165 @subsubsection Input Section and Garbage Collection
5167 @cindex garbage collection
5168 When link-time garbage collection is in use (@samp{--gc-sections}),
5169 it is often useful to mark sections that should not be eliminated.
5170 This is accomplished by surrounding an input section's wildcard entry
5171 with @code{KEEP()}, as in @code{KEEP(*(.init))} or
5172 @code{KEEP(SORT_BY_NAME(*)(.ctors))}.
5174 @node Input Section Example
5175 @subsubsection Input Section Example
5176 The following example is a complete linker script. It tells the linker
5177 to read all of the sections from file @file{all.o} and place them at the
5178 start of output section @samp{outputa} which starts at location
5179 @samp{0x10000}. All of section @samp{.input1} from file @file{foo.o}
5180 follows immediately, in the same output section. All of section
5181 @samp{.input2} from @file{foo.o} goes into output section
5182 @samp{outputb}, followed by section @samp{.input1} from @file{foo1.o}.
5183 All of the remaining @samp{.input1} and @samp{.input2} sections from any
5184 files are written to output section @samp{outputc}.
5212 If an output section's name is the same as the input section's name
5213 and is representable as a C identifier, then the linker will
5214 automatically @pxref{PROVIDE} two symbols: __start_SECNAME and
5215 __stop_SECNAME, where SECNAME is the name of the section. These
5216 indicate the start address and end address of the output section
5217 respectively. Note: most section names are not representable as
5218 C identifiers because they contain a @samp{.} character.
5220 @node Output Section Data
5221 @subsection Output Section Data
5223 @cindex section data
5224 @cindex output section data
5225 @kindex BYTE(@var{expression})
5226 @kindex SHORT(@var{expression})
5227 @kindex LONG(@var{expression})
5228 @kindex QUAD(@var{expression})
5229 @kindex SQUAD(@var{expression})
5230 You can include explicit bytes of data in an output section by using
5231 @code{BYTE}, @code{SHORT}, @code{LONG}, @code{QUAD}, or @code{SQUAD} as
5232 an output section command. Each keyword is followed by an expression in
5233 parentheses providing the value to store (@pxref{Expressions}). The
5234 value of the expression is stored at the current value of the location
5237 The @code{BYTE}, @code{SHORT}, @code{LONG}, and @code{QUAD} commands
5238 store one, two, four, and eight bytes (respectively). After storing the
5239 bytes, the location counter is incremented by the number of bytes
5242 For example, this will store the byte 1 followed by the four byte value
5243 of the symbol @samp{addr}:
5249 When using a 64 bit host or target, @code{QUAD} and @code{SQUAD} are the
5250 same; they both store an 8 byte, or 64 bit, value. When both host and
5251 target are 32 bits, an expression is computed as 32 bits. In this case
5252 @code{QUAD} stores a 32 bit value zero extended to 64 bits, and
5253 @code{SQUAD} stores a 32 bit value sign extended to 64 bits.
5255 If the object file format of the output file has an explicit endianness,
5256 which is the normal case, the value will be stored in that endianness.
5257 When the object file format does not have an explicit endianness, as is
5258 true of, for example, S-records, the value will be stored in the
5259 endianness of the first input object file.
5261 Note---these commands only work inside a section description and not
5262 between them, so the following will produce an error from the linker:
5264 SECTIONS @{@ .text : @{@ *(.text) @}@ LONG(1) .data : @{@ *(.data) @}@ @}@
5266 whereas this will work:
5268 SECTIONS @{@ .text : @{@ *(.text) ; LONG(1) @}@ .data : @{@ *(.data) @}@ @}@
5271 @kindex FILL(@var{expression})
5272 @cindex holes, filling
5273 @cindex unspecified memory
5274 You may use the @code{FILL} command to set the fill pattern for the
5275 current section. It is followed by an expression in parentheses. Any
5276 otherwise unspecified regions of memory within the section (for example,
5277 gaps left due to the required alignment of input sections) are filled
5278 with the value of the expression, repeated as
5279 necessary. A @code{FILL} statement covers memory locations after the
5280 point at which it occurs in the section definition; by including more
5281 than one @code{FILL} statement, you can have different fill patterns in
5282 different parts of an output section.
5284 This example shows how to fill unspecified regions of memory with the
5290 The @code{FILL} command is similar to the @samp{=@var{fillexp}} output
5291 section attribute, but it only affects the
5292 part of the section following the @code{FILL} command, rather than the
5293 entire section. If both are used, the @code{FILL} command takes
5294 precedence. @xref{Output Section Fill}, for details on the fill
5297 @node Output Section Keywords
5298 @subsection Output Section Keywords
5299 There are a couple of keywords which can appear as output section
5303 @kindex CREATE_OBJECT_SYMBOLS
5304 @cindex input filename symbols
5305 @cindex filename symbols
5306 @item CREATE_OBJECT_SYMBOLS
5307 The command tells the linker to create a symbol for each input file.
5308 The name of each symbol will be the name of the corresponding input
5309 file. The section of each symbol will be the output section in which
5310 the @code{CREATE_OBJECT_SYMBOLS} command appears.
5312 This is conventional for the a.out object file format. It is not
5313 normally used for any other object file format.
5315 @kindex CONSTRUCTORS
5316 @cindex C++ constructors, arranging in link
5317 @cindex constructors, arranging in link
5319 When linking using the a.out object file format, the linker uses an
5320 unusual set construct to support C++ global constructors and
5321 destructors. When linking object file formats which do not support
5322 arbitrary sections, such as ECOFF and XCOFF, the linker will
5323 automatically recognize C++ global constructors and destructors by name.
5324 For these object file formats, the @code{CONSTRUCTORS} command tells the
5325 linker to place constructor information in the output section where the
5326 @code{CONSTRUCTORS} command appears. The @code{CONSTRUCTORS} command is
5327 ignored for other object file formats.
5329 The symbol @w{@code{__CTOR_LIST__}} marks the start of the global
5330 constructors, and the symbol @w{@code{__CTOR_END__}} marks the end.
5331 Similarly, @w{@code{__DTOR_LIST__}} and @w{@code{__DTOR_END__}} mark
5332 the start and end of the global destructors. The
5333 first word in the list is the number of entries, followed by the address
5334 of each constructor or destructor, followed by a zero word. The
5335 compiler must arrange to actually run the code. For these object file
5336 formats @sc{gnu} C++ normally calls constructors from a subroutine
5337 @code{__main}; a call to @code{__main} is automatically inserted into
5338 the startup code for @code{main}. @sc{gnu} C++ normally runs
5339 destructors either by using @code{atexit}, or directly from the function
5342 For object file formats such as @code{COFF} or @code{ELF} which support
5343 arbitrary section names, @sc{gnu} C++ will normally arrange to put the
5344 addresses of global constructors and destructors into the @code{.ctors}
5345 and @code{.dtors} sections. Placing the following sequence into your
5346 linker script will build the sort of table which the @sc{gnu} C++
5347 runtime code expects to see.
5351 LONG((__CTOR_END__ - __CTOR_LIST__) / 4 - 2)
5356 LONG((__DTOR_END__ - __DTOR_LIST__) / 4 - 2)
5362 If you are using the @sc{gnu} C++ support for initialization priority,
5363 which provides some control over the order in which global constructors
5364 are run, you must sort the constructors at link time to ensure that they
5365 are executed in the correct order. When using the @code{CONSTRUCTORS}
5366 command, use @samp{SORT_BY_NAME(CONSTRUCTORS)} instead. When using the
5367 @code{.ctors} and @code{.dtors} sections, use @samp{*(SORT_BY_NAME(.ctors))} and
5368 @samp{*(SORT_BY_NAME(.dtors))} instead of just @samp{*(.ctors)} and
5371 Normally the compiler and linker will handle these issues automatically,
5372 and you will not need to concern yourself with them. However, you may
5373 need to consider this if you are using C++ and writing your own linker
5378 @node Output Section Discarding
5379 @subsection Output Section Discarding
5380 @cindex discarding sections
5381 @cindex sections, discarding
5382 @cindex removing sections
5383 The linker will not normally create output sections with no contents.
5384 This is for convenience when referring to input sections that may or
5385 may not be present in any of the input files. For example:
5387 .foo : @{ *(.foo) @}
5390 will only create a @samp{.foo} section in the output file if there is a
5391 @samp{.foo} section in at least one input file, and if the input
5392 sections are not all empty. Other link script directives that allocate
5393 space in an output section will also create the output section. So
5394 too will assignments to dot even if the assignment does not create
5395 space, except for @samp{. = 0}, @samp{. = . + 0}, @samp{. = sym},
5396 @samp{. = . + sym} and @samp{. = ALIGN (. != 0, expr, 1)} when
5397 @samp{sym} is an absolute symbol of value 0 defined in the script.
5398 This allows you to force output of an empty section with @samp{. = .}.
5400 The linker will ignore address assignments (@pxref{Output Section Address})
5401 on discarded output sections, except when the linker script defines
5402 symbols in the output section. In that case the linker will obey
5403 the address assignments, possibly advancing dot even though the
5404 section is discarded.
5407 The special output section name @samp{/DISCARD/} may be used to discard
5408 input sections. Any input sections which are assigned to an output
5409 section named @samp{/DISCARD/} are not included in the output file.
5411 This can be used to discard input sections marked with the ELF flag
5412 @code{SHF_GNU_RETAIN}, which would otherwise have been saved from linker
5415 Note, sections that match the @samp{/DISCARD/} output section will be
5416 discarded even if they are in an ELF section group which has other
5417 members which are not being discarded. This is deliberate.
5418 Discarding takes precedence over grouping.
5420 @node Output Section Attributes
5421 @subsection Output Section Attributes
5422 @cindex output section attributes
5423 We showed above that the full description of an output section looked
5428 @var{section} [@var{address}] [(@var{type})] :
5430 [ALIGN(@var{section_align}) | ALIGN_WITH_INPUT]
5431 [SUBALIGN(@var{subsection_align})]
5434 @var{output-section-command}
5435 @var{output-section-command}
5437 @} [>@var{region}] [AT>@var{lma_region}] [:@var{phdr} :@var{phdr} @dots{}] [=@var{fillexp}]
5441 We've already described @var{section}, @var{address}, and
5442 @var{output-section-command}. In this section we will describe the
5443 remaining section attributes.
5446 * Output Section Type:: Output section type
5447 * Output Section LMA:: Output section LMA
5448 * Forced Output Alignment:: Forced Output Alignment
5449 * Forced Input Alignment:: Forced Input Alignment
5450 * Output Section Constraint:: Output section constraint
5451 * Output Section Region:: Output section region
5452 * Output Section Phdr:: Output section phdr
5453 * Output Section Fill:: Output section fill
5456 @node Output Section Type
5457 @subsubsection Output Section Type
5458 Each output section may have a type. The type is a keyword in
5459 parentheses. The following types are defined:
5463 The section should be marked as not loadable, so that it will not be
5464 loaded into memory when the program is run.
5469 These type names are supported for backward compatibility, and are
5470 rarely used. They all have the same effect: the section should be
5471 marked as not allocatable, so that no memory is allocated for the
5472 section when the program is run.
5476 @cindex prevent unnecessary loading
5477 @cindex loading, preventing
5478 The linker normally sets the attributes of an output section based on
5479 the input sections which map into it. You can override this by using
5480 the section type. For example, in the script sample below, the
5481 @samp{ROM} section is addressed at memory location @samp{0} and does not
5482 need to be loaded when the program is run.
5486 ROM 0 (NOLOAD) : @{ @dots{} @}
5492 @node Output Section LMA
5493 @subsubsection Output Section LMA
5494 @kindex AT>@var{lma_region}
5495 @kindex AT(@var{lma})
5496 @cindex load address
5497 @cindex section load address
5498 Every section has a virtual address (VMA) and a load address (LMA); see
5499 @ref{Basic Script Concepts}. The virtual address is specified by the
5500 @pxref{Output Section Address} described earlier. The load address is
5501 specified by the @code{AT} or @code{AT>} keywords. Specifying a load
5502 address is optional.
5504 The @code{AT} keyword takes an expression as an argument. This
5505 specifies the exact load address of the section. The @code{AT>} keyword
5506 takes the name of a memory region as an argument. @xref{MEMORY}. The
5507 load address of the section is set to the next free address in the
5508 region, aligned to the section's alignment requirements.
5510 If neither @code{AT} nor @code{AT>} is specified for an allocatable
5511 section, the linker will use the following heuristic to determine the
5516 If the section has a specific VMA address, then this is used as
5517 the LMA address as well.
5520 If the section is not allocatable then its LMA is set to its VMA.
5523 Otherwise if a memory region can be found that is compatible
5524 with the current section, and this region contains at least one
5525 section, then the LMA is set so the difference between the
5526 VMA and LMA is the same as the difference between the VMA and LMA of
5527 the last section in the located region.
5530 If no memory regions have been declared then a default region
5531 that covers the entire address space is used in the previous step.
5534 If no suitable region could be found, or there was no previous
5535 section then the LMA is set equal to the VMA.
5538 @cindex ROM initialized data
5539 @cindex initialized data in ROM
5540 This feature is designed to make it easy to build a ROM image. For
5541 example, the following linker script creates three output sections: one
5542 called @samp{.text}, which starts at @code{0x1000}, one called
5543 @samp{.mdata}, which is loaded at the end of the @samp{.text} section
5544 even though its VMA is @code{0x2000}, and one called @samp{.bss} to hold
5545 uninitialized data at address @code{0x3000}. The symbol @code{_data} is
5546 defined with the value @code{0x2000}, which shows that the location
5547 counter holds the VMA value, not the LMA value.
5553 .text 0x1000 : @{ *(.text) _etext = . ; @}
5555 AT ( ADDR (.text) + SIZEOF (.text) )
5556 @{ _data = . ; *(.data); _edata = . ; @}
5558 @{ _bstart = . ; *(.bss) *(COMMON) ; _bend = . ;@}
5563 The run-time initialization code for use with a program generated with
5564 this linker script would include something like the following, to copy
5565 the initialized data from the ROM image to its runtime address. Notice
5566 how this code takes advantage of the symbols defined by the linker
5571 extern char _etext, _data, _edata, _bstart, _bend;
5572 char *src = &_etext;
5575 /* ROM has data at end of text; copy it. */
5576 while (dst < &_edata)
5580 for (dst = &_bstart; dst< &_bend; dst++)
5585 @node Forced Output Alignment
5586 @subsubsection Forced Output Alignment
5587 @kindex ALIGN(@var{section_align})
5588 @cindex forcing output section alignment
5589 @cindex output section alignment
5590 You can increase an output section's alignment by using ALIGN. As an
5591 alternative you can enforce that the difference between the VMA and LMA remains
5592 intact throughout this output section with the ALIGN_WITH_INPUT attribute.
5594 @node Forced Input Alignment
5595 @subsubsection Forced Input Alignment
5596 @kindex SUBALIGN(@var{subsection_align})
5597 @cindex forcing input section alignment
5598 @cindex input section alignment
5599 You can force input section alignment within an output section by using
5600 SUBALIGN. The value specified overrides any alignment given by input
5601 sections, whether larger or smaller.
5603 @node Output Section Constraint
5604 @subsubsection Output Section Constraint
5607 @cindex constraints on output sections
5608 You can specify that an output section should only be created if all
5609 of its input sections are read-only or all of its input sections are
5610 read-write by using the keyword @code{ONLY_IF_RO} and
5611 @code{ONLY_IF_RW} respectively.
5613 @node Output Section Region
5614 @subsubsection Output Section Region
5615 @kindex >@var{region}
5616 @cindex section, assigning to memory region
5617 @cindex memory regions and sections
5618 You can assign a section to a previously defined region of memory by
5619 using @samp{>@var{region}}. @xref{MEMORY}.
5621 Here is a simple example:
5624 MEMORY @{ rom : ORIGIN = 0x1000, LENGTH = 0x1000 @}
5625 SECTIONS @{ ROM : @{ *(.text) @} >rom @}
5629 @node Output Section Phdr
5630 @subsubsection Output Section Phdr
5632 @cindex section, assigning to program header
5633 @cindex program headers and sections
5634 You can assign a section to a previously defined program segment by
5635 using @samp{:@var{phdr}}. @xref{PHDRS}. If a section is assigned to
5636 one or more segments, then all subsequent allocated sections will be
5637 assigned to those segments as well, unless they use an explicitly
5638 @code{:@var{phdr}} modifier. You can use @code{:NONE} to tell the
5639 linker to not put the section in any segment at all.
5641 Here is a simple example:
5644 PHDRS @{ text PT_LOAD ; @}
5645 SECTIONS @{ .text : @{ *(.text) @} :text @}
5649 @node Output Section Fill
5650 @subsubsection Output Section Fill
5651 @kindex =@var{fillexp}
5652 @cindex section fill pattern
5653 @cindex fill pattern, entire section
5654 You can set the fill pattern for an entire section by using
5655 @samp{=@var{fillexp}}. @var{fillexp} is an expression
5656 (@pxref{Expressions}). Any otherwise unspecified regions of memory
5657 within the output section (for example, gaps left due to the required
5658 alignment of input sections) will be filled with the value, repeated as
5659 necessary. If the fill expression is a simple hex number, ie. a string
5660 of hex digit starting with @samp{0x} and without a trailing @samp{k} or @samp{M}, then
5661 an arbitrarily long sequence of hex digits can be used to specify the
5662 fill pattern; Leading zeros become part of the pattern too. For all
5663 other cases, including extra parentheses or a unary @code{+}, the fill
5664 pattern is the four least significant bytes of the value of the
5665 expression. In all cases, the number is big-endian.
5667 You can also change the fill value with a @code{FILL} command in the
5668 output section commands; (@pxref{Output Section Data}).
5670 Here is a simple example:
5673 SECTIONS @{ .text : @{ *(.text) @} =0x90909090 @}
5677 @node Overlay Description
5678 @subsection Overlay Description
5681 An overlay description provides an easy way to describe sections which
5682 are to be loaded as part of a single memory image but are to be run at
5683 the same memory address. At run time, some sort of overlay manager will
5684 copy the overlaid sections in and out of the runtime memory address as
5685 required, perhaps by simply manipulating addressing bits. This approach
5686 can be useful, for example, when a certain region of memory is faster
5689 Overlays are described using the @code{OVERLAY} command. The
5690 @code{OVERLAY} command is used within a @code{SECTIONS} command, like an
5691 output section description. The full syntax of the @code{OVERLAY}
5692 command is as follows:
5695 OVERLAY [@var{start}] : [NOCROSSREFS] [AT ( @var{ldaddr} )]
5699 @var{output-section-command}
5700 @var{output-section-command}
5702 @} [:@var{phdr}@dots{}] [=@var{fill}]
5705 @var{output-section-command}
5706 @var{output-section-command}
5708 @} [:@var{phdr}@dots{}] [=@var{fill}]
5710 @} [>@var{region}] [:@var{phdr}@dots{}] [=@var{fill}] [,]
5714 Everything is optional except @code{OVERLAY} (a keyword), and each
5715 section must have a name (@var{secname1} and @var{secname2} above). The
5716 section definitions within the @code{OVERLAY} construct are identical to
5717 those within the general @code{SECTIONS} construct (@pxref{SECTIONS}),
5718 except that no addresses and no memory regions may be defined for
5719 sections within an @code{OVERLAY}.
5721 The comma at the end may be required if a @var{fill} is used and
5722 the next @var{sections-command} looks like a continuation of the expression.
5724 The sections are all defined with the same starting address. The load
5725 addresses of the sections are arranged such that they are consecutive in
5726 memory starting at the load address used for the @code{OVERLAY} as a
5727 whole (as with normal section definitions, the load address is optional,
5728 and defaults to the start address; the start address is also optional,
5729 and defaults to the current value of the location counter).
5731 If the @code{NOCROSSREFS} keyword is used, and there are any
5732 references among the sections, the linker will report an error. Since
5733 the sections all run at the same address, it normally does not make
5734 sense for one section to refer directly to another.
5735 @xref{Miscellaneous Commands, NOCROSSREFS}.
5737 For each section within the @code{OVERLAY}, the linker automatically
5738 provides two symbols. The symbol @code{__load_start_@var{secname}} is
5739 defined as the starting load address of the section. The symbol
5740 @code{__load_stop_@var{secname}} is defined as the final load address of
5741 the section. Any characters within @var{secname} which are not legal
5742 within C identifiers are removed. C (or assembler) code may use these
5743 symbols to move the overlaid sections around as necessary.
5745 At the end of the overlay, the value of the location counter is set to
5746 the start address of the overlay plus the size of the largest section.
5748 Here is an example. Remember that this would appear inside a
5749 @code{SECTIONS} construct.
5752 OVERLAY 0x1000 : AT (0x4000)
5754 .text0 @{ o1/*.o(.text) @}
5755 .text1 @{ o2/*.o(.text) @}
5760 This will define both @samp{.text0} and @samp{.text1} to start at
5761 address 0x1000. @samp{.text0} will be loaded at address 0x4000, and
5762 @samp{.text1} will be loaded immediately after @samp{.text0}. The
5763 following symbols will be defined if referenced: @code{__load_start_text0},
5764 @code{__load_stop_text0}, @code{__load_start_text1},
5765 @code{__load_stop_text1}.
5767 C code to copy overlay @code{.text1} into the overlay area might look
5772 extern char __load_start_text1, __load_stop_text1;
5773 memcpy ((char *) 0x1000, &__load_start_text1,
5774 &__load_stop_text1 - &__load_start_text1);
5778 Note that the @code{OVERLAY} command is just syntactic sugar, since
5779 everything it does can be done using the more basic commands. The above
5780 example could have been written identically as follows.
5784 .text0 0x1000 : AT (0x4000) @{ o1/*.o(.text) @}
5785 PROVIDE (__load_start_text0 = LOADADDR (.text0));
5786 PROVIDE (__load_stop_text0 = LOADADDR (.text0) + SIZEOF (.text0));
5787 .text1 0x1000 : AT (0x4000 + SIZEOF (.text0)) @{ o2/*.o(.text) @}
5788 PROVIDE (__load_start_text1 = LOADADDR (.text1));
5789 PROVIDE (__load_stop_text1 = LOADADDR (.text1) + SIZEOF (.text1));
5790 . = 0x1000 + MAX (SIZEOF (.text0), SIZEOF (.text1));
5795 @section MEMORY Command
5797 @cindex memory regions
5798 @cindex regions of memory
5799 @cindex allocating memory
5800 @cindex discontinuous memory
5801 The linker's default configuration permits allocation of all available
5802 memory. You can override this by using the @code{MEMORY} command.
5804 The @code{MEMORY} command describes the location and size of blocks of
5805 memory in the target. You can use it to describe which memory regions
5806 may be used by the linker, and which memory regions it must avoid. You
5807 can then assign sections to particular memory regions. The linker will
5808 set section addresses based on the memory regions, and will warn about
5809 regions that become too full. The linker will not shuffle sections
5810 around to fit into the available regions.
5812 A linker script may contain many uses of the @code{MEMORY} command,
5813 however, all memory blocks defined are treated as if they were
5814 specified inside a single @code{MEMORY} command. The syntax for
5820 @var{name} [(@var{attr})] : ORIGIN = @var{origin}, LENGTH = @var{len}
5826 The @var{name} is a name used in the linker script to refer to the
5827 region. The region name has no meaning outside of the linker script.
5828 Region names are stored in a separate name space, and will not conflict
5829 with symbol names, file names, or section names. Each memory region
5830 must have a distinct name within the @code{MEMORY} command. However you can
5831 add later alias names to existing memory regions with the @ref{REGION_ALIAS}
5834 @cindex memory region attributes
5835 The @var{attr} string is an optional list of attributes that specify
5836 whether to use a particular memory region for an input section which is
5837 not explicitly mapped in the linker script. As described in
5838 @ref{SECTIONS}, if you do not specify an output section for some input
5839 section, the linker will create an output section with the same name as
5840 the input section. If you define region attributes, the linker will use
5841 them to select the memory region for the output section that it creates.
5843 The @var{attr} string must consist only of the following characters:
5858 Invert the sense of any of the attributes that follow
5861 If an unmapped section matches any of the listed attributes other than
5862 @samp{!}, it will be placed in the memory region. The @samp{!}
5863 attribute reverses the test for the characters that follow, so that an
5864 unmapped section will be placed in the memory region only if it does
5865 not match any of the attributes listed afterwards. Thus an attribute
5866 string of @samp{RW!X} will match any unmapped section that has either
5867 or both of the @samp{R} and @samp{W} attributes, but only as long as
5868 the section does not also have the @samp{X} attribute.
5873 The @var{origin} is an numerical expression for the start address of
5874 the memory region. The expression must evaluate to a constant and it
5875 cannot involve any symbols. The keyword @code{ORIGIN} may be
5876 abbreviated to @code{org} or @code{o} (but not, for example,
5882 The @var{len} is an expression for the size in bytes of the memory
5883 region. As with the @var{origin} expression, the expression must
5884 be numerical only and must evaluate to a constant. The keyword
5885 @code{LENGTH} may be abbreviated to @code{len} or @code{l}.
5887 In the following example, we specify that there are two memory regions
5888 available for allocation: one starting at @samp{0} for 256 kilobytes,
5889 and the other starting at @samp{0x40000000} for four megabytes. The
5890 linker will place into the @samp{rom} memory region every section which
5891 is not explicitly mapped into a memory region, and is either read-only
5892 or executable. The linker will place other sections which are not
5893 explicitly mapped into a memory region into the @samp{ram} memory
5900 rom (rx) : ORIGIN = 0, LENGTH = 256K
5901 ram (!rx) : org = 0x40000000, l = 4M
5906 Once you define a memory region, you can direct the linker to place
5907 specific output sections into that memory region by using the
5908 @samp{>@var{region}} output section attribute. For example, if you have
5909 a memory region named @samp{mem}, you would use @samp{>mem} in the
5910 output section definition. @xref{Output Section Region}. If no address
5911 was specified for the output section, the linker will set the address to
5912 the next available address within the memory region. If the combined
5913 output sections directed to a memory region are too large for the
5914 region, the linker will issue an error message.
5916 It is possible to access the origin and length of a memory in an
5917 expression via the @code{ORIGIN(@var{memory})} and
5918 @code{LENGTH(@var{memory})} functions:
5922 _fstack = ORIGIN(ram) + LENGTH(ram) - 4;
5927 @section PHDRS Command
5929 @cindex program headers
5930 @cindex ELF program headers
5931 @cindex program segments
5932 @cindex segments, ELF
5933 The ELF object file format uses @dfn{program headers}, also knows as
5934 @dfn{segments}. The program headers describe how the program should be
5935 loaded into memory. You can print them out by using the @code{objdump}
5936 program with the @samp{-p} option.
5938 When you run an ELF program on a native ELF system, the system loader
5939 reads the program headers in order to figure out how to load the
5940 program. This will only work if the program headers are set correctly.
5941 This manual does not describe the details of how the system loader
5942 interprets program headers; for more information, see the ELF ABI.
5944 The linker will create reasonable program headers by default. However,
5945 in some cases, you may need to specify the program headers more
5946 precisely. You may use the @code{PHDRS} command for this purpose. When
5947 the linker sees the @code{PHDRS} command in the linker script, it will
5948 not create any program headers other than the ones specified.
5950 The linker only pays attention to the @code{PHDRS} command when
5951 generating an ELF output file. In other cases, the linker will simply
5952 ignore @code{PHDRS}.
5954 This is the syntax of the @code{PHDRS} command. The words @code{PHDRS},
5955 @code{FILEHDR}, @code{AT}, and @code{FLAGS} are keywords.
5961 @var{name} @var{type} [ FILEHDR ] [ PHDRS ] [ AT ( @var{address} ) ]
5962 [ FLAGS ( @var{flags} ) ] ;
5967 The @var{name} is used only for reference in the @code{SECTIONS} command
5968 of the linker script. It is not put into the output file. Program
5969 header names are stored in a separate name space, and will not conflict
5970 with symbol names, file names, or section names. Each program header
5971 must have a distinct name. The headers are processed in order and it
5972 is usual for them to map to sections in ascending load address order.
5974 Certain program header types describe segments of memory which the
5975 system loader will load from the file. In the linker script, you
5976 specify the contents of these segments by placing allocatable output
5977 sections in the segments. You use the @samp{:@var{phdr}} output section
5978 attribute to place a section in a particular segment. @xref{Output
5981 It is normal to put certain sections in more than one segment. This
5982 merely implies that one segment of memory contains another. You may
5983 repeat @samp{:@var{phdr}}, using it once for each segment which should
5984 contain the section.
5986 If you place a section in one or more segments using @samp{:@var{phdr}},
5987 then the linker will place all subsequent allocatable sections which do
5988 not specify @samp{:@var{phdr}} in the same segments. This is for
5989 convenience, since generally a whole set of contiguous sections will be
5990 placed in a single segment. You can use @code{:NONE} to override the
5991 default segment and tell the linker to not put the section in any
5996 You may use the @code{FILEHDR} and @code{PHDRS} keywords after
5997 the program header type to further describe the contents of the segment.
5998 The @code{FILEHDR} keyword means that the segment should include the ELF
5999 file header. The @code{PHDRS} keyword means that the segment should
6000 include the ELF program headers themselves. If applied to a loadable
6001 segment (@code{PT_LOAD}), all prior loadable segments must have one of
6004 The @var{type} may be one of the following. The numbers indicate the
6005 value of the keyword.
6008 @item @code{PT_NULL} (0)
6009 Indicates an unused program header.
6011 @item @code{PT_LOAD} (1)
6012 Indicates that this program header describes a segment to be loaded from
6015 @item @code{PT_DYNAMIC} (2)
6016 Indicates a segment where dynamic linking information can be found.
6018 @item @code{PT_INTERP} (3)
6019 Indicates a segment where the name of the program interpreter may be
6022 @item @code{PT_NOTE} (4)
6023 Indicates a segment holding note information.
6025 @item @code{PT_SHLIB} (5)
6026 A reserved program header type, defined but not specified by the ELF
6029 @item @code{PT_PHDR} (6)
6030 Indicates a segment where the program headers may be found.
6032 @item @code{PT_TLS} (7)
6033 Indicates a segment containing thread local storage.
6035 @item @var{expression}
6036 An expression giving the numeric type of the program header. This may
6037 be used for types not defined above.
6040 You can specify that a segment should be loaded at a particular address
6041 in memory by using an @code{AT} expression. This is identical to the
6042 @code{AT} command used as an output section attribute (@pxref{Output
6043 Section LMA}). The @code{AT} command for a program header overrides the
6044 output section attribute.
6046 The linker will normally set the segment flags based on the sections
6047 which comprise the segment. You may use the @code{FLAGS} keyword to
6048 explicitly specify the segment flags. The value of @var{flags} must be
6049 an integer. It is used to set the @code{p_flags} field of the program
6052 Here is an example of @code{PHDRS}. This shows a typical set of program
6053 headers used on a native ELF system.
6059 headers PT_PHDR PHDRS ;
6061 text PT_LOAD FILEHDR PHDRS ;
6063 dynamic PT_DYNAMIC ;
6069 .interp : @{ *(.interp) @} :text :interp
6070 .text : @{ *(.text) @} :text
6071 .rodata : @{ *(.rodata) @} /* defaults to :text */
6073 . = . + 0x1000; /* move to a new page in memory */
6074 .data : @{ *(.data) @} :data
6075 .dynamic : @{ *(.dynamic) @} :data :dynamic
6082 @section VERSION Command
6083 @kindex VERSION @{script text@}
6084 @cindex symbol versions
6085 @cindex version script
6086 @cindex versions of symbols
6087 The linker supports symbol versions when using ELF. Symbol versions are
6088 only useful when using shared libraries. The dynamic linker can use
6089 symbol versions to select a specific version of a function when it runs
6090 a program that may have been linked against an earlier version of the
6093 You can include a version script directly in the main linker script, or
6094 you can supply the version script as an implicit linker script. You can
6095 also use the @samp{--version-script} linker option.
6097 The syntax of the @code{VERSION} command is simply
6099 VERSION @{ version-script-commands @}
6102 The format of the version script commands is identical to that used by
6103 Sun's linker in Solaris 2.5. The version script defines a tree of
6104 version nodes. You specify the node names and interdependencies in the
6105 version script. You can specify which symbols are bound to which
6106 version nodes, and you can reduce a specified set of symbols to local
6107 scope so that they are not globally visible outside of the shared
6110 The easiest way to demonstrate the version script language is with a few
6136 This example version script defines three version nodes. The first
6137 version node defined is @samp{VERS_1.1}; it has no other dependencies.
6138 The script binds the symbol @samp{foo1} to @samp{VERS_1.1}. It reduces
6139 a number of symbols to local scope so that they are not visible outside
6140 of the shared library; this is done using wildcard patterns, so that any
6141 symbol whose name begins with @samp{old}, @samp{original}, or @samp{new}
6142 is matched. The wildcard patterns available are the same as those used
6143 in the shell when matching filenames (also known as ``globbing'').
6144 However, if you specify the symbol name inside double quotes, then the
6145 name is treated as literal, rather than as a glob pattern.
6147 Next, the version script defines node @samp{VERS_1.2}. This node
6148 depends upon @samp{VERS_1.1}. The script binds the symbol @samp{foo2}
6149 to the version node @samp{VERS_1.2}.
6151 Finally, the version script defines node @samp{VERS_2.0}. This node
6152 depends upon @samp{VERS_1.2}. The scripts binds the symbols @samp{bar1}
6153 and @samp{bar2} are bound to the version node @samp{VERS_2.0}.
6155 When the linker finds a symbol defined in a library which is not
6156 specifically bound to a version node, it will effectively bind it to an
6157 unspecified base version of the library. You can bind all otherwise
6158 unspecified symbols to a given version node by using @samp{global: *;}
6159 somewhere in the version script. Note that it's slightly crazy to use
6160 wildcards in a global spec except on the last version node. Global
6161 wildcards elsewhere run the risk of accidentally adding symbols to the
6162 set exported for an old version. That's wrong since older versions
6163 ought to have a fixed set of symbols.
6165 The names of the version nodes have no specific meaning other than what
6166 they might suggest to the person reading them. The @samp{2.0} version
6167 could just as well have appeared in between @samp{1.1} and @samp{1.2}.
6168 However, this would be a confusing way to write a version script.
6170 Node name can be omitted, provided it is the only version node
6171 in the version script. Such version script doesn't assign any versions to
6172 symbols, only selects which symbols will be globally visible out and which
6176 @{ global: foo; bar; local: *; @};
6179 When you link an application against a shared library that has versioned
6180 symbols, the application itself knows which version of each symbol it
6181 requires, and it also knows which version nodes it needs from each
6182 shared library it is linked against. Thus at runtime, the dynamic
6183 loader can make a quick check to make sure that the libraries you have
6184 linked against do in fact supply all of the version nodes that the
6185 application will need to resolve all of the dynamic symbols. In this
6186 way it is possible for the dynamic linker to know with certainty that
6187 all external symbols that it needs will be resolvable without having to
6188 search for each symbol reference.
6190 The symbol versioning is in effect a much more sophisticated way of
6191 doing minor version checking that SunOS does. The fundamental problem
6192 that is being addressed here is that typically references to external
6193 functions are bound on an as-needed basis, and are not all bound when
6194 the application starts up. If a shared library is out of date, a
6195 required interface may be missing; when the application tries to use
6196 that interface, it may suddenly and unexpectedly fail. With symbol
6197 versioning, the user will get a warning when they start their program if
6198 the libraries being used with the application are too old.
6200 There are several GNU extensions to Sun's versioning approach. The
6201 first of these is the ability to bind a symbol to a version node in the
6202 source file where the symbol is defined instead of in the versioning
6203 script. This was done mainly to reduce the burden on the library
6204 maintainer. You can do this by putting something like:
6206 __asm__(".symver original_foo,foo@@VERS_1.1");
6209 in the C source file. This renames the function @samp{original_foo} to
6210 be an alias for @samp{foo} bound to the version node @samp{VERS_1.1}.
6211 The @samp{local:} directive can be used to prevent the symbol
6212 @samp{original_foo} from being exported. A @samp{.symver} directive
6213 takes precedence over a version script.
6215 The second GNU extension is to allow multiple versions of the same
6216 function to appear in a given shared library. In this way you can make
6217 an incompatible change to an interface without increasing the major
6218 version number of the shared library, while still allowing applications
6219 linked against the old interface to continue to function.
6221 To do this, you must use multiple @samp{.symver} directives in the
6222 source file. Here is an example:
6225 __asm__(".symver original_foo,foo@@");
6226 __asm__(".symver old_foo,foo@@VERS_1.1");
6227 __asm__(".symver old_foo1,foo@@VERS_1.2");
6228 __asm__(".symver new_foo,foo@@@@VERS_2.0");
6231 In this example, @samp{foo@@} represents the symbol @samp{foo} bound to the
6232 unspecified base version of the symbol. The source file that contains this
6233 example would define 4 C functions: @samp{original_foo}, @samp{old_foo},
6234 @samp{old_foo1}, and @samp{new_foo}.
6236 When you have multiple definitions of a given symbol, there needs to be
6237 some way to specify a default version to which external references to
6238 this symbol will be bound. You can do this with the
6239 @samp{foo@@@@VERS_2.0} type of @samp{.symver} directive. You can only
6240 declare one version of a symbol as the default in this manner; otherwise
6241 you would effectively have multiple definitions of the same symbol.
6243 If you wish to bind a reference to a specific version of the symbol
6244 within the shared library, you can use the aliases of convenience
6245 (i.e., @samp{old_foo}), or you can use the @samp{.symver} directive to
6246 specifically bind to an external version of the function in question.
6248 You can also specify the language in the version script:
6251 VERSION extern "lang" @{ version-script-commands @}
6254 The supported @samp{lang}s are @samp{C}, @samp{C++}, and @samp{Java}.
6255 The linker will iterate over the list of symbols at the link time and
6256 demangle them according to @samp{lang} before matching them to the
6257 patterns specified in @samp{version-script-commands}. The default
6258 @samp{lang} is @samp{C}.
6260 Demangled names may contains spaces and other special characters. As
6261 described above, you can use a glob pattern to match demangled names,
6262 or you can use a double-quoted string to match the string exactly. In
6263 the latter case, be aware that minor differences (such as differing
6264 whitespace) between the version script and the demangler output will
6265 cause a mismatch. As the exact string generated by the demangler
6266 might change in the future, even if the mangled name does not, you
6267 should check that all of your version directives are behaving as you
6268 expect when you upgrade.
6271 @section Expressions in Linker Scripts
6274 The syntax for expressions in the linker script language is identical to
6275 that of C expressions. All expressions are evaluated as integers. All
6276 expressions are evaluated in the same size, which is 32 bits if both the
6277 host and target are 32 bits, and is otherwise 64 bits.
6279 You can use and set symbol values in expressions.
6281 The linker defines several special purpose builtin functions for use in
6285 * Constants:: Constants
6286 * Symbolic Constants:: Symbolic constants
6287 * Symbols:: Symbol Names
6288 * Orphan Sections:: Orphan Sections
6289 * Location Counter:: The Location Counter
6290 * Operators:: Operators
6291 * Evaluation:: Evaluation
6292 * Expression Section:: The Section of an Expression
6293 * Builtin Functions:: Builtin Functions
6297 @subsection Constants
6298 @cindex integer notation
6299 @cindex constants in linker scripts
6300 All constants are integers.
6302 As in C, the linker considers an integer beginning with @samp{0} to be
6303 octal, and an integer beginning with @samp{0x} or @samp{0X} to be
6304 hexadecimal. Alternatively the linker accepts suffixes of @samp{h} or
6305 @samp{H} for hexadecimal, @samp{o} or @samp{O} for octal, @samp{b} or
6306 @samp{B} for binary and @samp{d} or @samp{D} for decimal. Any integer
6307 value without a prefix or a suffix is considered to be decimal.
6309 @cindex scaled integers
6310 @cindex K and M integer suffixes
6311 @cindex M and K integer suffixes
6312 @cindex suffixes for integers
6313 @cindex integer suffixes
6314 In addition, you can use the suffixes @code{K} and @code{M} to scale a
6318 @c END TEXI2ROFF-KILL
6319 @code{1024} or @code{1024*1024}
6323 ${\rm 1024}$ or ${\rm 1024}^2$
6325 @c END TEXI2ROFF-KILL
6326 respectively. For example, the following
6327 all refer to the same quantity:
6336 Note - the @code{K} and @code{M} suffixes cannot be used in
6337 conjunction with the base suffixes mentioned above.
6339 @node Symbolic Constants
6340 @subsection Symbolic Constants
6341 @cindex symbolic constants
6343 It is possible to refer to target-specific constants via the use of
6344 the @code{CONSTANT(@var{name})} operator, where @var{name} is one of:
6349 The target's maximum page size.
6351 @item COMMONPAGESIZE
6352 @kindex COMMONPAGESIZE
6353 The target's default page size.
6359 .text ALIGN (CONSTANT (MAXPAGESIZE)) : @{ *(.text) @}
6362 will create a text section aligned to the largest page boundary
6363 supported by the target.
6366 @subsection Symbol Names
6367 @cindex symbol names
6369 @cindex quoted symbol names
6371 Unless quoted, symbol names start with a letter, underscore, or period
6372 and may include letters, digits, underscores, periods, and hyphens.
6373 Unquoted symbol names must not conflict with any keywords. You can
6374 specify a symbol which contains odd characters or has the same name as a
6375 keyword by surrounding the symbol name in double quotes:
6378 "with a space" = "also with a space" + 10;
6381 Since symbols can contain many non-alphabetic characters, it is safest
6382 to delimit symbols with spaces. For example, @samp{A-B} is one symbol,
6383 whereas @samp{A - B} is an expression involving subtraction.
6385 @node Orphan Sections
6386 @subsection Orphan Sections
6388 Orphan sections are sections present in the input files which
6389 are not explicitly placed into the output file by the linker
6390 script. The linker will still copy these sections into the
6391 output file by either finding, or creating a suitable output section
6392 in which to place the orphaned input section.
6394 If the name of an orphaned input section exactly matches the name of
6395 an existing output section, then the orphaned input section will be
6396 placed at the end of that output section.
6398 If there is no output section with a matching name then new output
6399 sections will be created. Each new output section will have the same
6400 name as the orphan section placed within it. If there are multiple
6401 orphan sections with the same name, these will all be combined into
6402 one new output section.
6404 If new output sections are created to hold orphaned input sections,
6405 then the linker must decide where to place these new output sections
6406 in relation to existing output sections. On most modern targets, the
6407 linker attempts to place orphan sections after sections of the same
6408 attribute, such as code vs data, loadable vs non-loadable, etc. If no
6409 sections with matching attributes are found, or your target lacks this
6410 support, the orphan section is placed at the end of the file.
6412 The command-line options @samp{--orphan-handling} and @samp{--unique}
6413 (@pxref{Options,,Command-line Options}) can be used to control which
6414 output sections an orphan is placed in.
6416 @node Location Counter
6417 @subsection The Location Counter
6420 @cindex location counter
6421 @cindex current output location
6422 The special linker variable @dfn{dot} @samp{.} always contains the
6423 current output location counter. Since the @code{.} always refers to a
6424 location in an output section, it may only appear in an expression
6425 within a @code{SECTIONS} command. The @code{.} symbol may appear
6426 anywhere that an ordinary symbol is allowed in an expression.
6429 Assigning a value to @code{.} will cause the location counter to be
6430 moved. This may be used to create holes in the output section. The
6431 location counter may not be moved backwards inside an output section,
6432 and may not be moved backwards outside of an output section if so
6433 doing creates areas with overlapping LMAs.
6449 In the previous example, the @samp{.text} section from @file{file1} is
6450 located at the beginning of the output section @samp{output}. It is
6451 followed by a 1000 byte gap. Then the @samp{.text} section from
6452 @file{file2} appears, also with a 1000 byte gap following before the
6453 @samp{.text} section from @file{file3}. The notation @samp{= 0x12345678}
6454 specifies what data to write in the gaps (@pxref{Output Section Fill}).
6456 @cindex dot inside sections
6457 Note: @code{.} actually refers to the byte offset from the start of the
6458 current containing object. Normally this is the @code{SECTIONS}
6459 statement, whose start address is 0, hence @code{.} can be used as an
6460 absolute address. If @code{.} is used inside a section description
6461 however, it refers to the byte offset from the start of that section,
6462 not an absolute address. Thus in a script like this:
6480 The @samp{.text} section will be assigned a starting address of 0x100
6481 and a size of exactly 0x200 bytes, even if there is not enough data in
6482 the @samp{.text} input sections to fill this area. (If there is too
6483 much data, an error will be produced because this would be an attempt to
6484 move @code{.} backwards). The @samp{.data} section will start at 0x500
6485 and it will have an extra 0x600 bytes worth of space after the end of
6486 the values from the @samp{.data} input sections and before the end of
6487 the @samp{.data} output section itself.
6489 @cindex dot outside sections
6490 Setting symbols to the value of the location counter outside of an
6491 output section statement can result in unexpected values if the linker
6492 needs to place orphan sections. For example, given the following:
6498 .text: @{ *(.text) @}
6502 .data: @{ *(.data) @}
6507 If the linker needs to place some input section, e.g. @code{.rodata},
6508 not mentioned in the script, it might choose to place that section
6509 between @code{.text} and @code{.data}. You might think the linker
6510 should place @code{.rodata} on the blank line in the above script, but
6511 blank lines are of no particular significance to the linker. As well,
6512 the linker doesn't associate the above symbol names with their
6513 sections. Instead, it assumes that all assignments or other
6514 statements belong to the previous output section, except for the
6515 special case of an assignment to @code{.}. I.e., the linker will
6516 place the orphan @code{.rodata} section as if the script was written
6523 .text: @{ *(.text) @}
6527 .rodata: @{ *(.rodata) @}
6528 .data: @{ *(.data) @}
6533 This may or may not be the script author's intention for the value of
6534 @code{start_of_data}. One way to influence the orphan section
6535 placement is to assign the location counter to itself, as the linker
6536 assumes that an assignment to @code{.} is setting the start address of
6537 a following output section and thus should be grouped with that
6538 section. So you could write:
6544 .text: @{ *(.text) @}
6549 .data: @{ *(.data) @}
6554 Now, the orphan @code{.rodata} section will be placed between
6555 @code{end_of_text} and @code{start_of_data}.
6559 @subsection Operators
6560 @cindex operators for arithmetic
6561 @cindex arithmetic operators
6562 @cindex precedence in expressions
6563 The linker recognizes the standard C set of arithmetic operators, with
6564 the standard bindings and precedence levels:
6567 @c END TEXI2ROFF-KILL
6569 precedence associativity Operators Notes
6575 5 left == != > < <= >=
6581 11 right &= += -= *= /= (2)
6585 (1) Prefix operators
6586 (2) @xref{Assignments}.
6590 \vskip \baselineskip
6591 %"lispnarrowing" is the extra indent used generally for smallexample
6592 \hskip\lispnarrowing\vbox{\offinterlineskip
6595 {\vrule#&\strut\hfil\ #\ \hfil&\vrule#&\strut\hfil\ #\ \hfil&\vrule#&\strut\hfil\ {\tt #}\ \hfil&\vrule#\cr
6596 height2pt&\omit&&\omit&&\omit&\cr
6597 &Precedence&& Associativity &&{\rm Operators}&\cr
6598 height2pt&\omit&&\omit&&\omit&\cr
6600 height2pt&\omit&&\omit&&\omit&\cr
6602 % '176 is tilde, '~' in tt font
6603 &1&&left&&\qquad- \char'176\ !\qquad\dag&\cr
6604 &2&&left&&* / \%&\cr
6607 &5&&left&&== != > < <= >=&\cr
6610 &8&&left&&{\&\&}&\cr
6613 &11&&right&&\qquad\&= += -= *= /=\qquad\ddag&\cr
6615 height2pt&\omit&&\omit&&\omit&\cr}
6620 @obeylines@parskip=0pt@parindent=0pt
6621 @dag@quad Prefix operators.
6622 @ddag@quad @xref{Assignments}.
6625 @c END TEXI2ROFF-KILL
6628 @subsection Evaluation
6629 @cindex lazy evaluation
6630 @cindex expression evaluation order
6631 The linker evaluates expressions lazily. It only computes the value of
6632 an expression when absolutely necessary.
6634 The linker needs some information, such as the value of the start
6635 address of the first section, and the origins and lengths of memory
6636 regions, in order to do any linking at all. These values are computed
6637 as soon as possible when the linker reads in the linker script.
6639 However, other values (such as symbol values) are not known or needed
6640 until after storage allocation. Such values are evaluated later, when
6641 other information (such as the sizes of output sections) is available
6642 for use in the symbol assignment expression.
6644 The sizes of sections cannot be known until after allocation, so
6645 assignments dependent upon these are not performed until after
6648 Some expressions, such as those depending upon the location counter
6649 @samp{.}, must be evaluated during section allocation.
6651 If the result of an expression is required, but the value is not
6652 available, then an error results. For example, a script like the
6658 .text 9+this_isnt_constant :
6664 will cause the error message @samp{non constant expression for initial
6667 @node Expression Section
6668 @subsection The Section of an Expression
6669 @cindex expression sections
6670 @cindex absolute expressions
6671 @cindex relative expressions
6672 @cindex absolute and relocatable symbols
6673 @cindex relocatable and absolute symbols
6674 @cindex symbols, relocatable and absolute
6675 Addresses and symbols may be section relative, or absolute. A section
6676 relative symbol is relocatable. If you request relocatable output
6677 using the @samp{-r} option, a further link operation may change the
6678 value of a section relative symbol. On the other hand, an absolute
6679 symbol will retain the same value throughout any further link
6682 Some terms in linker expressions are addresses. This is true of
6683 section relative symbols and for builtin functions that return an
6684 address, such as @code{ADDR}, @code{LOADADDR}, @code{ORIGIN} and
6685 @code{SEGMENT_START}. Other terms are simply numbers, or are builtin
6686 functions that return a non-address value, such as @code{LENGTH}.
6687 One complication is that unless you set @code{LD_FEATURE ("SANE_EXPR")}
6688 (@pxref{Miscellaneous Commands}), numbers and absolute symbols are treated
6689 differently depending on their location, for compatibility with older
6690 versions of @code{ld}. Expressions appearing outside an output
6691 section definition treat all numbers as absolute addresses.
6692 Expressions appearing inside an output section definition treat
6693 absolute symbols as numbers. If @code{LD_FEATURE ("SANE_EXPR")} is
6694 given, then absolute symbols and numbers are simply treated as numbers
6697 In the following simple example,
6704 __executable_start = 0x100;
6708 __data_start = 0x10;
6716 both @code{.} and @code{__executable_start} are set to the absolute
6717 address 0x100 in the first two assignments, then both @code{.} and
6718 @code{__data_start} are set to 0x10 relative to the @code{.data}
6719 section in the second two assignments.
6721 For expressions involving numbers, relative addresses and absolute
6722 addresses, ld follows these rules to evaluate terms:
6726 Unary operations on an absolute address or number, and binary
6727 operations on two absolute addresses or two numbers, or between one
6728 absolute address and a number, apply the operator to the value(s).
6730 Unary operations on a relative address, and binary operations on two
6731 relative addresses in the same section or between one relative address
6732 and a number, apply the operator to the offset part of the address(es).
6734 Other binary operations, that is, between two relative addresses not
6735 in the same section, or between a relative address and an absolute
6736 address, first convert any non-absolute term to an absolute address
6737 before applying the operator.
6740 The result section of each sub-expression is as follows:
6744 An operation involving only numbers results in a number.
6746 The result of comparisons, @samp{&&} and @samp{||} is also a number.
6748 The result of other binary arithmetic and logical operations on two
6749 relative addresses in the same section or two absolute addresses
6750 (after above conversions) is also a number when
6751 @code{LD_FEATURE ("SANE_EXPR")} or inside an output section definition
6752 but an absolute address otherwise.
6754 The result of other operations on relative addresses or one
6755 relative address and a number, is a relative address in the same
6756 section as the relative operand(s).
6758 The result of other operations on absolute addresses (after above
6759 conversions) is an absolute address.
6762 You can use the builtin function @code{ABSOLUTE} to force an expression
6763 to be absolute when it would otherwise be relative. For example, to
6764 create an absolute symbol set to the address of the end of the output
6765 section @samp{.data}:
6769 .data : @{ *(.data) _edata = ABSOLUTE(.); @}
6773 If @samp{ABSOLUTE} were not used, @samp{_edata} would be relative to the
6774 @samp{.data} section.
6776 Using @code{LOADADDR} also forces an expression absolute, since this
6777 particular builtin function returns an absolute address.
6779 @node Builtin Functions
6780 @subsection Builtin Functions
6781 @cindex functions in expressions
6782 The linker script language includes a number of builtin functions for
6783 use in linker script expressions.
6786 @item ABSOLUTE(@var{exp})
6787 @kindex ABSOLUTE(@var{exp})
6788 @cindex expression, absolute
6789 Return the absolute (non-relocatable, as opposed to non-negative) value
6790 of the expression @var{exp}. Primarily useful to assign an absolute
6791 value to a symbol within a section definition, where symbol values are
6792 normally section relative. @xref{Expression Section}.
6794 @item ADDR(@var{section})
6795 @kindex ADDR(@var{section})
6796 @cindex section address in expression
6797 Return the address (VMA) of the named @var{section}. Your
6798 script must previously have defined the location of that section. In
6799 the following example, @code{start_of_output_1}, @code{symbol_1} and
6800 @code{symbol_2} are assigned equivalent values, except that
6801 @code{symbol_1} will be relative to the @code{.output1} section while
6802 the other two will be absolute:
6808 start_of_output_1 = ABSOLUTE(.);
6813 symbol_1 = ADDR(.output1);
6814 symbol_2 = start_of_output_1;
6820 @item ALIGN(@var{align})
6821 @itemx ALIGN(@var{exp},@var{align})
6822 @kindex ALIGN(@var{align})
6823 @kindex ALIGN(@var{exp},@var{align})
6824 @cindex round up location counter
6825 @cindex align location counter
6826 @cindex round up expression
6827 @cindex align expression
6828 Return the location counter (@code{.}) or arbitrary expression aligned
6829 to the next @var{align} boundary. The single operand @code{ALIGN}
6830 doesn't change the value of the location counter---it just does
6831 arithmetic on it. The two operand @code{ALIGN} allows an arbitrary
6832 expression to be aligned upwards (@code{ALIGN(@var{align})} is
6833 equivalent to @code{ALIGN(ABSOLUTE(.), @var{align})}).
6835 Here is an example which aligns the output @code{.data} section to the
6836 next @code{0x2000} byte boundary after the preceding section and sets a
6837 variable within the section to the next @code{0x8000} boundary after the
6842 .data ALIGN(0x2000): @{
6844 variable = ALIGN(0x8000);
6850 The first use of @code{ALIGN} in this example specifies the location of
6851 a section because it is used as the optional @var{address} attribute of
6852 a section definition (@pxref{Output Section Address}). The second use
6853 of @code{ALIGN} is used to defines the value of a symbol.
6855 The builtin function @code{NEXT} is closely related to @code{ALIGN}.
6857 @item ALIGNOF(@var{section})
6858 @kindex ALIGNOF(@var{section})
6859 @cindex section alignment
6860 Return the alignment in bytes of the named @var{section}, if that section has
6861 been allocated. If the section has not been allocated when this is
6862 evaluated, the linker will report an error. In the following example,
6863 the alignment of the @code{.output} section is stored as the first
6864 value in that section.
6869 LONG (ALIGNOF (.output))
6876 @item BLOCK(@var{exp})
6877 @kindex BLOCK(@var{exp})
6878 This is a synonym for @code{ALIGN}, for compatibility with older linker
6879 scripts. It is most often seen when setting the address of an output
6882 @item DATA_SEGMENT_ALIGN(@var{maxpagesize}, @var{commonpagesize})
6883 @kindex DATA_SEGMENT_ALIGN(@var{maxpagesize}, @var{commonpagesize})
6884 This is equivalent to either
6886 (ALIGN(@var{maxpagesize}) + (. & (@var{maxpagesize} - 1)))
6890 (ALIGN(@var{maxpagesize})
6891 + ((. + @var{commonpagesize} - 1) & (@var{maxpagesize} - @var{commonpagesize})))
6894 depending on whether the latter uses fewer @var{commonpagesize} sized pages
6895 for the data segment (area between the result of this expression and
6896 @code{DATA_SEGMENT_END}) than the former or not.
6897 If the latter form is used, it means @var{commonpagesize} bytes of runtime
6898 memory will be saved at the expense of up to @var{commonpagesize} wasted
6899 bytes in the on-disk file.
6901 This expression can only be used directly in @code{SECTIONS} commands, not in
6902 any output section descriptions and only once in the linker script.
6903 @var{commonpagesize} should be less or equal to @var{maxpagesize} and should
6904 be the system page size the object wants to be optimized for while still
6905 running on system page sizes up to @var{maxpagesize}. Note however
6906 that @samp{-z relro} protection will not be effective if the system
6907 page size is larger than @var{commonpagesize}.
6912 . = DATA_SEGMENT_ALIGN(0x10000, 0x2000);
6915 @item DATA_SEGMENT_END(@var{exp})
6916 @kindex DATA_SEGMENT_END(@var{exp})
6917 This defines the end of data segment for @code{DATA_SEGMENT_ALIGN}
6918 evaluation purposes.
6921 . = DATA_SEGMENT_END(.);
6924 @item DATA_SEGMENT_RELRO_END(@var{offset}, @var{exp})
6925 @kindex DATA_SEGMENT_RELRO_END(@var{offset}, @var{exp})
6926 This defines the end of the @code{PT_GNU_RELRO} segment when
6927 @samp{-z relro} option is used.
6928 When @samp{-z relro} option is not present, @code{DATA_SEGMENT_RELRO_END}
6929 does nothing, otherwise @code{DATA_SEGMENT_ALIGN} is padded so that
6930 @var{exp} + @var{offset} is aligned to the @var{commonpagesize}
6931 argument given to @code{DATA_SEGMENT_ALIGN}. If present in the linker
6932 script, it must be placed between @code{DATA_SEGMENT_ALIGN} and
6933 @code{DATA_SEGMENT_END}. Evaluates to the second argument plus any
6934 padding needed at the end of the @code{PT_GNU_RELRO} segment due to
6938 . = DATA_SEGMENT_RELRO_END(24, .);
6941 @item DEFINED(@var{symbol})
6942 @kindex DEFINED(@var{symbol})
6943 @cindex symbol defaults
6944 Return 1 if @var{symbol} is in the linker global symbol table and is
6945 defined before the statement using DEFINED in the script, otherwise
6946 return 0. You can use this function to provide
6947 default values for symbols. For example, the following script fragment
6948 shows how to set a global symbol @samp{begin} to the first location in
6949 the @samp{.text} section---but if a symbol called @samp{begin} already
6950 existed, its value is preserved:
6956 begin = DEFINED(begin) ? begin : . ;
6964 @item LENGTH(@var{memory})
6965 @kindex LENGTH(@var{memory})
6966 Return the length of the memory region named @var{memory}.
6968 @item LOADADDR(@var{section})
6969 @kindex LOADADDR(@var{section})
6970 @cindex section load address in expression
6971 Return the absolute LMA of the named @var{section}. (@pxref{Output
6974 @item LOG2CEIL(@var{exp})
6975 @kindex LOG2CEIL(@var{exp})
6976 Return the binary logarithm of @var{exp} rounded towards infinity.
6977 @code{LOG2CEIL(0)} returns 0.
6980 @item MAX(@var{exp1}, @var{exp2})
6981 Returns the maximum of @var{exp1} and @var{exp2}.
6984 @item MIN(@var{exp1}, @var{exp2})
6985 Returns the minimum of @var{exp1} and @var{exp2}.
6987 @item NEXT(@var{exp})
6988 @kindex NEXT(@var{exp})
6989 @cindex unallocated address, next
6990 Return the next unallocated address that is a multiple of @var{exp}.
6991 This function is closely related to @code{ALIGN(@var{exp})}; unless you
6992 use the @code{MEMORY} command to define discontinuous memory for the
6993 output file, the two functions are equivalent.
6995 @item ORIGIN(@var{memory})
6996 @kindex ORIGIN(@var{memory})
6997 Return the origin of the memory region named @var{memory}.
6999 @item SEGMENT_START(@var{segment}, @var{default})
7000 @kindex SEGMENT_START(@var{segment}, @var{default})
7001 Return the base address of the named @var{segment}. If an explicit
7002 value has already been given for this segment (with a command-line
7003 @samp{-T} option) then that value will be returned otherwise the value
7004 will be @var{default}. At present, the @samp{-T} command-line option
7005 can only be used to set the base address for the ``text'', ``data'', and
7006 ``bss'' sections, but you can use @code{SEGMENT_START} with any segment
7009 @item SIZEOF(@var{section})
7010 @kindex SIZEOF(@var{section})
7011 @cindex section size
7012 Return the size in bytes of the named @var{section}, if that section has
7013 been allocated. If the section has not been allocated when this is
7014 evaluated, the linker will report an error. In the following example,
7015 @code{symbol_1} and @code{symbol_2} are assigned identical values:
7024 symbol_1 = .end - .start ;
7025 symbol_2 = SIZEOF(.output);
7030 @item SIZEOF_HEADERS
7031 @itemx sizeof_headers
7032 @kindex SIZEOF_HEADERS
7034 Return the size in bytes of the output file's headers. This is
7035 information which appears at the start of the output file. You can use
7036 this number when setting the start address of the first section, if you
7037 choose, to facilitate paging.
7039 @cindex not enough room for program headers
7040 @cindex program headers, not enough room
7041 When producing an ELF output file, if the linker script uses the
7042 @code{SIZEOF_HEADERS} builtin function, the linker must compute the
7043 number of program headers before it has determined all the section
7044 addresses and sizes. If the linker later discovers that it needs
7045 additional program headers, it will report an error @samp{not enough
7046 room for program headers}. To avoid this error, you must avoid using
7047 the @code{SIZEOF_HEADERS} function, or you must rework your linker
7048 script to avoid forcing the linker to use additional program headers, or
7049 you must define the program headers yourself using the @code{PHDRS}
7050 command (@pxref{PHDRS}).
7053 @node Implicit Linker Scripts
7054 @section Implicit Linker Scripts
7055 @cindex implicit linker scripts
7056 If you specify a linker input file which the linker can not recognize as
7057 an object file or an archive file, it will try to read the file as a
7058 linker script. If the file can not be parsed as a linker script, the
7059 linker will report an error.
7061 An implicit linker script will not replace the default linker script.
7063 Typically an implicit linker script would contain only symbol
7064 assignments, or the @code{INPUT}, @code{GROUP}, or @code{VERSION}
7067 Any input files read because of an implicit linker script will be read
7068 at the position in the command line where the implicit linker script was
7069 read. This can affect archive searching.
7072 @chapter Linker Plugins
7075 @cindex linker plugins
7076 The linker can use dynamically loaded plugins to modify its behavior.
7077 For example, the link-time optimization feature that some compilers
7078 support is implemented with a linker plugin.
7080 Currently there is only one plugin shipped by default, but more may
7081 be added here later.
7084 * libdep Plugin:: Static Library Dependencies Plugin
7088 @section Static Library Dependencies Plugin
7089 @cindex static library dependencies
7090 Originally, static libraries were contained in an archive file consisting
7091 just of a collection of relocatable object files. Later they evolved to
7092 optionally include a symbol table, to assist in finding the needed objects
7093 within a library. There their evolution ended, and dynamic libraries
7096 One useful feature of dynamic libraries was that, more than just collecting
7097 multiple objects into a single file, they also included a list of their
7098 dependencies, such that one could specify just the name of a single dynamic
7099 library at link time, and all of its dependencies would be implicitly
7100 referenced as well. But static libraries lacked this feature, so if a
7101 link invocation was switched from using dynamic libraries to static
7102 libraries, the link command would usually fail unless it was rewritten to
7103 explicitly list the dependencies of the static library.
7105 The GNU @command{ar} utility now supports a @option{--record-libdeps} option
7106 to embed dependency lists into static libraries as well, and the @file{libdep}
7107 plugin may be used to read this dependency information at link time. The
7108 dependency information is stored as a single string, carrying @option{-l}
7109 and @option{-L} arguments as they would normally appear in a linker
7110 command line. As such, the information can be written with any text
7111 utility and stored into any archive, even if GNU @command{ar} is not
7112 being used to create the archive. The information is stored in an
7113 archive member named @samp{__.LIBDEP}.
7115 For example, given a library @file{libssl.a} that depends on another
7116 library @file{libcrypto.a} which may be found in @file{/usr/local/lib},
7117 the @samp{__.LIBDEP} member of @file{libssl.a} would contain
7120 -L/usr/local/lib -lcrypto
7124 @node Machine Dependent
7125 @chapter Machine Dependent Features
7127 @cindex machine dependencies
7128 @command{ld} has additional features on some platforms; the following
7129 sections describe them. Machines where @command{ld} has no additional
7130 functionality are not listed.
7134 * H8/300:: @command{ld} and the H8/300
7137 * M68HC11/68HC12:: @code{ld} and the Motorola 68HC11 and 68HC12 families
7140 * ARM:: @command{ld} and the ARM family
7143 * HPPA ELF32:: @command{ld} and HPPA 32-bit ELF
7146 * M68K:: @command{ld} and the Motorola 68K family
7149 * MIPS:: @command{ld} and the MIPS family
7152 * MMIX:: @command{ld} and MMIX
7155 * MSP430:: @command{ld} and MSP430
7158 * NDS32:: @command{ld} and NDS32
7161 * Nios II:: @command{ld} and the Altera Nios II
7164 * PowerPC ELF32:: @command{ld} and PowerPC 32-bit ELF Support
7167 * PowerPC64 ELF64:: @command{ld} and PowerPC64 64-bit ELF Support
7170 * S/390 ELF:: @command{ld} and S/390 ELF Support
7173 * SPU ELF:: @command{ld} and SPU ELF Support
7176 * TI COFF:: @command{ld} and TI COFF
7179 * WIN32:: @command{ld} and WIN32 (cygwin/mingw)
7182 * Xtensa:: @command{ld} and Xtensa Processors
7193 @section @command{ld} and the H8/300
7195 @cindex H8/300 support
7196 For the H8/300, @command{ld} can perform these global optimizations when
7197 you specify the @samp{--relax} command-line option.
7200 @cindex relaxing on H8/300
7201 @item relaxing address modes
7202 @command{ld} finds all @code{jsr} and @code{jmp} instructions whose
7203 targets are within eight bits, and turns them into eight-bit
7204 program-counter relative @code{bsr} and @code{bra} instructions,
7207 @cindex synthesizing on H8/300
7208 @item synthesizing instructions
7209 @c FIXME: specifically mov.b, or any mov instructions really? -> mov.b only, at least on H8, H8H, H8S
7210 @command{ld} finds all @code{mov.b} instructions which use the
7211 sixteen-bit absolute address form, but refer to the top
7212 page of memory, and changes them to use the eight-bit address form.
7213 (That is: the linker turns @samp{mov.b @code{@@}@var{aa}:16} into
7214 @samp{mov.b @code{@@}@var{aa}:8} whenever the address @var{aa} is in the
7215 top page of memory).
7217 @command{ld} finds all @code{mov} instructions which use the register
7218 indirect with 32-bit displacement addressing mode, but use a small
7219 displacement inside 16-bit displacement range, and changes them to use
7220 the 16-bit displacement form. (That is: the linker turns @samp{mov.b
7221 @code{@@}@var{d}:32,ERx} into @samp{mov.b @code{@@}@var{d}:16,ERx}
7222 whenever the displacement @var{d} is in the 16 bit signed integer
7223 range. Only implemented in ELF-format ld).
7225 @item bit manipulation instructions
7226 @command{ld} finds all bit manipulation instructions like @code{band, bclr,
7227 biand, bild, bior, bist, bixor, bld, bnot, bor, bset, bst, btst, bxor}
7228 which use 32 bit and 16 bit absolute address form, but refer to the top
7229 page of memory, and changes them to use the 8 bit address form.
7230 (That is: the linker turns @samp{bset #xx:3,@code{@@}@var{aa}:32} into
7231 @samp{bset #xx:3,@code{@@}@var{aa}:8} whenever the address @var{aa} is in
7232 the top page of memory).
7234 @item system control instructions
7235 @command{ld} finds all @code{ldc.w, stc.w} instructions which use the
7236 32 bit absolute address form, but refer to the top page of memory, and
7237 changes them to use 16 bit address form.
7238 (That is: the linker turns @samp{ldc.w @code{@@}@var{aa}:32,ccr} into
7239 @samp{ldc.w @code{@@}@var{aa}:16,ccr} whenever the address @var{aa} is in
7240 the top page of memory).
7250 @c This stuff is pointless to say unless you're especially concerned
7251 @c with Renesas chips; don't enable it for generic case, please.
7253 @chapter @command{ld} and Other Renesas Chips
7255 @command{ld} also supports the Renesas (formerly Hitachi) H8/300H,
7256 H8/500, and SH chips. No special features, commands, or command-line
7257 options are required for these chips.
7271 @node M68HC11/68HC12
7272 @section @command{ld} and the Motorola 68HC11 and 68HC12 families
7274 @cindex M68HC11 and 68HC12 support
7276 @subsection Linker Relaxation
7278 For the Motorola 68HC11, @command{ld} can perform these global
7279 optimizations when you specify the @samp{--relax} command-line option.
7282 @cindex relaxing on M68HC11
7283 @item relaxing address modes
7284 @command{ld} finds all @code{jsr} and @code{jmp} instructions whose
7285 targets are within eight bits, and turns them into eight-bit
7286 program-counter relative @code{bsr} and @code{bra} instructions,
7289 @command{ld} also looks at all 16-bit extended addressing modes and
7290 transforms them in a direct addressing mode when the address is in
7291 page 0 (between 0 and 0x0ff).
7293 @item relaxing gcc instruction group
7294 When @command{gcc} is called with @option{-mrelax}, it can emit group
7295 of instructions that the linker can optimize to use a 68HC11 direct
7296 addressing mode. These instructions consists of @code{bclr} or
7297 @code{bset} instructions.
7301 @subsection Trampoline Generation
7303 @cindex trampoline generation on M68HC11
7304 @cindex trampoline generation on M68HC12
7305 For 68HC11 and 68HC12, @command{ld} can generate trampoline code to
7306 call a far function using a normal @code{jsr} instruction. The linker
7307 will also change the relocation to some far function to use the
7308 trampoline address instead of the function address. This is typically the
7309 case when a pointer to a function is taken. The pointer will in fact
7310 point to the function trampoline.
7318 @section @command{ld} and the ARM family
7320 @cindex ARM interworking support
7321 @kindex --support-old-code
7322 For the ARM, @command{ld} will generate code stubs to allow functions calls
7323 between ARM and Thumb code. These stubs only work with code that has
7324 been compiled and assembled with the @samp{-mthumb-interwork} command
7325 line option. If it is necessary to link with old ARM object files or
7326 libraries, which have not been compiled with the -mthumb-interwork
7327 option then the @samp{--support-old-code} command-line switch should be
7328 given to the linker. This will make it generate larger stub functions
7329 which will work with non-interworking aware ARM code. Note, however,
7330 the linker does not support generating stubs for function calls to
7331 non-interworking aware Thumb code.
7333 @cindex thumb entry point
7334 @cindex entry point, thumb
7335 @kindex --thumb-entry=@var{entry}
7336 The @samp{--thumb-entry} switch is a duplicate of the generic
7337 @samp{--entry} switch, in that it sets the program's starting address.
7338 But it also sets the bottom bit of the address, so that it can be
7339 branched to using a BX instruction, and the program will start
7340 executing in Thumb mode straight away.
7342 @cindex PE import table prefixing
7343 @kindex --use-nul-prefixed-import-tables
7344 The @samp{--use-nul-prefixed-import-tables} switch is specifying, that
7345 the import tables idata4 and idata5 have to be generated with a zero
7346 element prefix for import libraries. This is the old style to generate
7347 import tables. By default this option is turned off.
7351 The @samp{--be8} switch instructs @command{ld} to generate BE8 format
7352 executables. This option is only valid when linking big-endian
7353 objects - ie ones which have been assembled with the @option{-EB}
7354 option. The resulting image will contain big-endian data and
7358 @kindex --target1-rel
7359 @kindex --target1-abs
7360 The @samp{R_ARM_TARGET1} relocation is typically used for entries in the
7361 @samp{.init_array} section. It is interpreted as either @samp{R_ARM_REL32}
7362 or @samp{R_ARM_ABS32}, depending on the target. The @samp{--target1-rel}
7363 and @samp{--target1-abs} switches override the default.
7366 @kindex --target2=@var{type}
7367 The @samp{--target2=type} switch overrides the default definition of the
7368 @samp{R_ARM_TARGET2} relocation. Valid values for @samp{type}, their
7369 meanings, and target defaults are as follows:
7372 @samp{R_ARM_REL32} (arm*-*-elf, arm*-*-eabi)
7376 @samp{R_ARM_GOT_PREL} (arm*-*-linux, arm*-*-*bsd)
7381 The @samp{R_ARM_V4BX} relocation (defined by the ARM AAELF
7382 specification) enables objects compiled for the ARMv4 architecture to be
7383 interworking-safe when linked with other objects compiled for ARMv4t, but
7384 also allows pure ARMv4 binaries to be built from the same ARMv4 objects.
7386 In the latter case, the switch @option{--fix-v4bx} must be passed to the
7387 linker, which causes v4t @code{BX rM} instructions to be rewritten as
7388 @code{MOV PC,rM}, since v4 processors do not have a @code{BX} instruction.
7390 In the former case, the switch should not be used, and @samp{R_ARM_V4BX}
7391 relocations are ignored.
7393 @cindex FIX_V4BX_INTERWORKING
7394 @kindex --fix-v4bx-interworking
7395 Replace @code{BX rM} instructions identified by @samp{R_ARM_V4BX}
7396 relocations with a branch to the following veneer:
7404 This allows generation of libraries/applications that work on ARMv4 cores
7405 and are still interworking safe. Note that the above veneer clobbers the
7406 condition flags, so may cause incorrect program behavior in rare cases.
7410 The @samp{--use-blx} switch enables the linker to use ARM/Thumb
7411 BLX instructions (available on ARMv5t and above) in various
7412 situations. Currently it is used to perform calls via the PLT from Thumb
7413 code using BLX rather than using BX and a mode-switching stub before
7414 each PLT entry. This should lead to such calls executing slightly faster.
7416 @cindex VFP11_DENORM_FIX
7417 @kindex --vfp11-denorm-fix
7418 The @samp{--vfp11-denorm-fix} switch enables a link-time workaround for a
7419 bug in certain VFP11 coprocessor hardware, which sometimes allows
7420 instructions with denorm operands (which must be handled by support code)
7421 to have those operands overwritten by subsequent instructions before
7422 the support code can read the intended values.
7424 The bug may be avoided in scalar mode if you allow at least one
7425 intervening instruction between a VFP11 instruction which uses a register
7426 and another instruction which writes to the same register, or at least two
7427 intervening instructions if vector mode is in use. The bug only affects
7428 full-compliance floating-point mode: you do not need this workaround if
7429 you are using "runfast" mode. Please contact ARM for further details.
7431 If you know you are using buggy VFP11 hardware, you can
7432 enable this workaround by specifying the linker option
7433 @samp{--vfp-denorm-fix=scalar} if you are using the VFP11 scalar
7434 mode only, or @samp{--vfp-denorm-fix=vector} if you are using
7435 vector mode (the latter also works for scalar code). The default is
7436 @samp{--vfp-denorm-fix=none}.
7438 If the workaround is enabled, instructions are scanned for
7439 potentially-troublesome sequences, and a veneer is created for each
7440 such sequence which may trigger the erratum. The veneer consists of the
7441 first instruction of the sequence and a branch back to the subsequent
7442 instruction. The original instruction is then replaced with a branch to
7443 the veneer. The extra cycles required to call and return from the veneer
7444 are sufficient to avoid the erratum in both the scalar and vector cases.
7446 @cindex ARM1176 erratum workaround
7447 @kindex --fix-arm1176
7448 @kindex --no-fix-arm1176
7449 The @samp{--fix-arm1176} switch enables a link-time workaround for an erratum
7450 in certain ARM1176 processors. The workaround is enabled by default if you
7451 are targeting ARM v6 (excluding ARM v6T2) or earlier. It can be disabled
7452 unconditionally by specifying @samp{--no-fix-arm1176}.
7454 Further information is available in the ``ARM1176JZ-S and ARM1176JZF-S
7455 Programmer Advice Notice'' available on the ARM documentation website at:
7456 http://infocenter.arm.com/.
7458 @cindex STM32L4xx erratum workaround
7459 @kindex --fix-stm32l4xx-629360
7461 The @samp{--fix-stm32l4xx-629360} switch enables a link-time
7462 workaround for a bug in the bus matrix / memory controller for some of
7463 the STM32 Cortex-M4 based products (STM32L4xx). When accessing
7464 off-chip memory via the affected bus for bus reads of 9 words or more,
7465 the bus can generate corrupt data and/or abort. These are only
7466 core-initiated accesses (not DMA), and might affect any access:
7467 integer loads such as LDM, POP and floating-point loads such as VLDM,
7468 VPOP. Stores are not affected.
7470 The bug can be avoided by splitting memory accesses into the
7471 necessary chunks to keep bus reads below 8 words.
7473 The workaround is not enabled by default, this is equivalent to use
7474 @samp{--fix-stm32l4xx-629360=none}. If you know you are using buggy
7475 STM32L4xx hardware, you can enable the workaround by specifying the
7476 linker option @samp{--fix-stm32l4xx-629360}, or the equivalent
7477 @samp{--fix-stm32l4xx-629360=default}.
7479 If the workaround is enabled, instructions are scanned for
7480 potentially-troublesome sequences, and a veneer is created for each
7481 such sequence which may trigger the erratum. The veneer consists in a
7482 replacement sequence emulating the behaviour of the original one and a
7483 branch back to the subsequent instruction. The original instruction is
7484 then replaced with a branch to the veneer.
7486 The workaround does not always preserve the memory access order for
7487 the LDMDB instruction, when the instruction loads the PC.
7489 The workaround is not able to handle problematic instructions when
7490 they are in the middle of an IT block, since a branch is not allowed
7491 there. In that case, the linker reports a warning and no replacement
7494 The workaround is not able to replace problematic instructions with a
7495 PC-relative branch instruction if the @samp{.text} section is too
7496 large. In that case, when the branch that replaces the original code
7497 cannot be encoded, the linker reports a warning and no replacement
7500 @cindex NO_ENUM_SIZE_WARNING
7501 @kindex --no-enum-size-warning
7502 The @option{--no-enum-size-warning} switch prevents the linker from
7503 warning when linking object files that specify incompatible EABI
7504 enumeration size attributes. For example, with this switch enabled,
7505 linking of an object file using 32-bit enumeration values with another
7506 using enumeration values fitted into the smallest possible space will
7509 @cindex NO_WCHAR_SIZE_WARNING
7510 @kindex --no-wchar-size-warning
7511 The @option{--no-wchar-size-warning} switch prevents the linker from
7512 warning when linking object files that specify incompatible EABI
7513 @code{wchar_t} size attributes. For example, with this switch enabled,
7514 linking of an object file using 32-bit @code{wchar_t} values with another
7515 using 16-bit @code{wchar_t} values will not be diagnosed.
7518 @kindex --pic-veneer
7519 The @samp{--pic-veneer} switch makes the linker use PIC sequences for
7520 ARM/Thumb interworking veneers, even if the rest of the binary
7521 is not PIC. This avoids problems on uClinux targets where
7522 @samp{--emit-relocs} is used to generate relocatable binaries.
7524 @cindex STUB_GROUP_SIZE
7525 @kindex --stub-group-size=@var{N}
7526 The linker will automatically generate and insert small sequences of
7527 code into a linked ARM ELF executable whenever an attempt is made to
7528 perform a function call to a symbol that is too far away. The
7529 placement of these sequences of instructions - called stubs - is
7530 controlled by the command-line option @option{--stub-group-size=N}.
7531 The placement is important because a poor choice can create a need for
7532 duplicate stubs, increasing the code size. The linker will try to
7533 group stubs together in order to reduce interruptions to the flow of
7534 code, but it needs guidance as to how big these groups should be and
7535 where they should be placed.
7537 The value of @samp{N}, the parameter to the
7538 @option{--stub-group-size=} option controls where the stub groups are
7539 placed. If it is negative then all stubs are placed after the first
7540 branch that needs them. If it is positive then the stubs can be
7541 placed either before or after the branches that need them. If the
7542 value of @samp{N} is 1 (either +1 or -1) then the linker will choose
7543 exactly where to place groups of stubs, using its built in heuristics.
7544 A value of @samp{N} greater than 1 (or smaller than -1) tells the
7545 linker that a single group of stubs can service at most @samp{N} bytes
7546 from the input sections.
7548 The default, if @option{--stub-group-size=} is not specified, is
7551 Farcalls stubs insertion is fully supported for the ARM-EABI target
7552 only, because it relies on object files properties not present
7555 @cindex Cortex-A8 erratum workaround
7556 @kindex --fix-cortex-a8
7557 @kindex --no-fix-cortex-a8
7558 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}.
7560 The erratum only affects Thumb-2 code. Please contact ARM for further details.
7562 @cindex Cortex-A53 erratum 835769 workaround
7563 @kindex --fix-cortex-a53-835769
7564 @kindex --no-fix-cortex-a53-835769
7565 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}.
7567 Please contact ARM for further details.
7569 @kindex --merge-exidx-entries
7570 @kindex --no-merge-exidx-entries
7571 @cindex Merging exidx entries
7572 The @samp{--no-merge-exidx-entries} switch disables the merging of adjacent exidx entries in debuginfo.
7575 @cindex 32-bit PLT entries
7576 The @samp{--long-plt} option enables the use of 16 byte PLT entries
7577 which support up to 4Gb of code. The default is to use 12 byte PLT
7578 entries which only support 512Mb of code.
7580 @kindex --no-apply-dynamic-relocs
7581 @cindex AArch64 rela addend
7582 The @samp{--no-apply-dynamic-relocs} option makes AArch64 linker do not apply
7583 link-time values for dynamic relocations.
7585 @cindex Placement of SG veneers
7586 All SG veneers are placed in the special output section @code{.gnu.sgstubs}.
7587 Its start address must be set, either with the command-line option
7588 @samp{--section-start} or in a linker script, to indicate where to place these
7591 @kindex --cmse-implib
7592 @cindex Secure gateway import library
7593 The @samp{--cmse-implib} option requests that the import libraries
7594 specified by the @samp{--out-implib} and @samp{--in-implib} options are
7595 secure gateway import libraries, suitable for linking a non-secure
7596 executable against secure code as per ARMv8-M Security Extensions.
7598 @kindex --in-implib=@var{file}
7599 @cindex Input import library
7600 The @samp{--in-implib=file} specifies an input import library whose symbols
7601 must keep the same address in the executable being produced. A warning is
7602 given if no @samp{--out-implib} is given but new symbols have been introduced
7603 in the executable that should be listed in its import library. Otherwise, if
7604 @samp{--out-implib} is specified, the symbols are added to the output import
7605 library. A warning is also given if some symbols present in the input import
7606 library have disappeared from the executable. This option is only effective
7607 for Secure Gateway import libraries, ie. when @samp{--cmse-implib} is
7621 @section @command{ld} and HPPA 32-bit ELF Support
7622 @cindex HPPA multiple sub-space stubs
7623 @kindex --multi-subspace
7624 When generating a shared library, @command{ld} will by default generate
7625 import stubs suitable for use with a single sub-space application.
7626 The @samp{--multi-subspace} switch causes @command{ld} to generate export
7627 stubs, and different (larger) import stubs suitable for use with
7628 multiple sub-spaces.
7630 @cindex HPPA stub grouping
7631 @kindex --stub-group-size=@var{N}
7632 Long branch stubs and import/export stubs are placed by @command{ld} in
7633 stub sections located between groups of input sections.
7634 @samp{--stub-group-size} specifies the maximum size of a group of input
7635 sections handled by one stub section. Since branch offsets are signed,
7636 a stub section may serve two groups of input sections, one group before
7637 the stub section, and one group after it. However, when using
7638 conditional branches that require stubs, it may be better (for branch
7639 prediction) that stub sections only serve one group of input sections.
7640 A negative value for @samp{N} chooses this scheme, ensuring that
7641 branches to stubs always use a negative offset. Two special values of
7642 @samp{N} are recognized, @samp{1} and @samp{-1}. These both instruct
7643 @command{ld} to automatically size input section groups for the branch types
7644 detected, with the same behaviour regarding stub placement as other
7645 positive or negative values of @samp{N} respectively.
7647 Note that @samp{--stub-group-size} does not split input sections. A
7648 single input section larger than the group size specified will of course
7649 create a larger group (of one section). If input sections are too
7650 large, it may not be possible for a branch to reach its stub.
7663 @section @command{ld} and the Motorola 68K family
7665 @cindex Motorola 68K GOT generation
7666 @kindex --got=@var{type}
7667 The @samp{--got=@var{type}} option lets you choose the GOT generation scheme.
7668 The choices are @samp{single}, @samp{negative}, @samp{multigot} and
7669 @samp{target}. When @samp{target} is selected the linker chooses
7670 the default GOT generation scheme for the current target.
7671 @samp{single} tells the linker to generate a single GOT with
7672 entries only at non-negative offsets.
7673 @samp{negative} instructs the linker to generate a single GOT with
7674 entries at both negative and positive offsets. Not all environments
7676 @samp{multigot} allows the linker to generate several GOTs in the
7677 output file. All GOT references from a single input object
7678 file access the same GOT, but references from different input object
7679 files might access different GOTs. Not all environments support such GOTs.
7692 @section @command{ld} and the MIPS family
7694 @cindex MIPS microMIPS instruction choice selection
7697 The @samp{--insn32} and @samp{--no-insn32} options control the choice of
7698 microMIPS instructions used in code generated by the linker, such as that
7699 in the PLT or lazy binding stubs, or in relaxation. If @samp{--insn32} is
7700 used, then the linker only uses 32-bit instruction encodings. By default
7701 or if @samp{--no-insn32} is used, all instruction encodings are used,
7702 including 16-bit ones where possible.
7704 @cindex MIPS branch relocation check control
7705 @kindex --ignore-branch-isa
7706 @kindex --no-ignore-branch-isa
7707 The @samp{--ignore-branch-isa} and @samp{--no-ignore-branch-isa} options
7708 control branch relocation checks for invalid ISA mode transitions. If
7709 @samp{--ignore-branch-isa} is used, then the linker accepts any branch
7710 relocations and any ISA mode transition required is lost in relocation
7711 calculation, except for some cases of @code{BAL} instructions which meet
7712 relaxation conditions and are converted to equivalent @code{JALX}
7713 instructions as the associated relocation is calculated. By default
7714 or if @samp{--no-ignore-branch-isa} is used a check is made causing
7715 the loss of an ISA mode transition to produce an error.
7728 @section @code{ld} and MMIX
7729 For MMIX, there is a choice of generating @code{ELF} object files or
7730 @code{mmo} object files when linking. The simulator @code{mmix}
7731 understands the @code{mmo} format. The binutils @code{objcopy} utility
7732 can translate between the two formats.
7734 There is one special section, the @samp{.MMIX.reg_contents} section.
7735 Contents in this section is assumed to correspond to that of global
7736 registers, and symbols referring to it are translated to special symbols,
7737 equal to registers. In a final link, the start address of the
7738 @samp{.MMIX.reg_contents} section corresponds to the first allocated
7739 global register multiplied by 8. Register @code{$255} is not included in
7740 this section; it is always set to the program entry, which is at the
7741 symbol @code{Main} for @code{mmo} files.
7743 Global symbols with the prefix @code{__.MMIX.start.}, for example
7744 @code{__.MMIX.start..text} and @code{__.MMIX.start..data} are special.
7745 The default linker script uses these to set the default start address
7748 Initial and trailing multiples of zero-valued 32-bit words in a section,
7749 are left out from an mmo file.
7762 @section @code{ld} and MSP430
7763 For the MSP430 it is possible to select the MPU architecture. The flag @samp{-m [mpu type]}
7764 will select an appropriate linker script for selected MPU type. (To get a list of known MPUs
7765 just pass @samp{-m help} option to the linker).
7767 @cindex MSP430 extra sections
7768 The linker will recognize some extra sections which are MSP430 specific:
7771 @item @samp{.vectors}
7772 Defines a portion of ROM where interrupt vectors located.
7774 @item @samp{.bootloader}
7775 Defines the bootloader portion of the ROM (if applicable). Any code
7776 in this section will be uploaded to the MPU.
7778 @item @samp{.infomem}
7779 Defines an information memory section (if applicable). Any code in
7780 this section will be uploaded to the MPU.
7782 @item @samp{.infomemnobits}
7783 This is the same as the @samp{.infomem} section except that any code
7784 in this section will not be uploaded to the MPU.
7786 @item @samp{.noinit}
7787 Denotes a portion of RAM located above @samp{.bss} section.
7789 The last two sections are used by gcc.
7793 @cindex MSP430 Options
7794 @kindex --code-region
7795 @item --code-region=[either,lower,upper,none]
7796 This will transform .text* sections to [either,lower,upper].text* sections. The
7797 argument passed to GCC for -mcode-region is propagated to the linker
7800 @kindex --data-region
7801 @item --data-region=[either,lower,upper,none]
7802 This will transform .data*, .bss* and .rodata* sections to
7803 [either,lower,upper].[data,bss,rodata]* sections. The argument passed to GCC
7804 for -mdata-region is propagated to the linker using this option.
7806 @kindex --disable-sec-transformation
7807 @item --disable-sec-transformation
7808 Prevent the transformation of sections as specified by the @code{--code-region}
7809 and @code{--data-region} options.
7810 This is useful if you are compiling and linking using a single call to the GCC
7811 wrapper, and want to compile the source files using -m[code,data]-region but
7812 not transform the sections for prebuilt libraries and objects.
7826 @section @code{ld} and NDS32
7827 @kindex relaxing on NDS32
7828 For NDS32, there are some options to select relaxation behavior. The linker
7829 relaxes objects according to these options.
7832 @item @samp{--m[no-]fp-as-gp}
7833 Disable/enable fp-as-gp relaxation.
7835 @item @samp{--mexport-symbols=FILE}
7836 Exporting symbols and their address into FILE as linker script.
7838 @item @samp{--m[no-]ex9}
7839 Disable/enable link-time EX9 relaxation.
7841 @item @samp{--mexport-ex9=FILE}
7842 Export the EX9 table after linking.
7844 @item @samp{--mimport-ex9=FILE}
7845 Import the Ex9 table for EX9 relaxation.
7847 @item @samp{--mupdate-ex9}
7848 Update the existing EX9 table.
7850 @item @samp{--mex9-limit=NUM}
7851 Maximum number of entries in the ex9 table.
7853 @item @samp{--mex9-loop-aware}
7854 Avoid generating the EX9 instruction inside the loop.
7856 @item @samp{--m[no-]ifc}
7857 Disable/enable the link-time IFC optimization.
7859 @item @samp{--mifc-loop-aware}
7860 Avoid generating the IFC instruction inside the loop.
7874 @section @command{ld} and the Altera Nios II
7875 @cindex Nios II call relaxation
7876 @kindex --relax on Nios II
7878 Call and immediate jump instructions on Nios II processors are limited to
7879 transferring control to addresses in the same 256MB memory segment,
7880 which may result in @command{ld} giving
7881 @samp{relocation truncated to fit} errors with very large programs.
7882 The command-line option @option{--relax} enables the generation of
7883 trampolines that can access the entire 32-bit address space for calls
7884 outside the normal @code{call} and @code{jmpi} address range. These
7885 trampolines are inserted at section boundaries, so may not themselves
7886 be reachable if an input section and its associated call trampolines are
7889 The @option{--relax} option is enabled by default unless @option{-r}
7890 is also specified. You can disable trampoline generation by using the
7891 @option{--no-relax} linker option. You can also disable this optimization
7892 locally by using the @samp{set .noat} directive in assembly-language
7893 source files, as the linker-inserted trampolines use the @code{at}
7894 register as a temporary.
7896 Note that the linker @option{--relax} option is independent of assembler
7897 relaxation options, and that using the GNU assembler's @option{-relax-all}
7898 option interferes with the linker's more selective call instruction relaxation.
7911 @section @command{ld} and PowerPC 32-bit ELF Support
7912 @cindex PowerPC long branches
7913 @kindex --relax on PowerPC
7914 Branches on PowerPC processors are limited to a signed 26-bit
7915 displacement, which may result in @command{ld} giving
7916 @samp{relocation truncated to fit} errors with very large programs.
7917 @samp{--relax} enables the generation of trampolines that can access
7918 the entire 32-bit address space. These trampolines are inserted at
7919 section boundaries, so may not themselves be reachable if an input
7920 section exceeds 33M in size. You may combine @samp{-r} and
7921 @samp{--relax} to add trampolines in a partial link. In that case
7922 both branches to undefined symbols and inter-section branches are also
7923 considered potentially out of range, and trampolines inserted.
7925 @cindex PowerPC ELF32 options
7930 Current PowerPC GCC accepts a @samp{-msecure-plt} option that
7931 generates code capable of using a newer PLT and GOT layout that has
7932 the security advantage of no executable section ever needing to be
7933 writable and no writable section ever being executable. PowerPC
7934 @command{ld} will generate this layout, including stubs to access the
7935 PLT, if all input files (including startup and static libraries) were
7936 compiled with @samp{-msecure-plt}. @samp{--bss-plt} forces the old
7937 BSS PLT (and GOT layout) which can give slightly better performance.
7939 @kindex --secure-plt
7941 @command{ld} will use the new PLT and GOT layout if it is linking new
7942 @samp{-fpic} or @samp{-fPIC} code, but does not do so automatically
7943 when linking non-PIC code. This option requests the new PLT and GOT
7944 layout. A warning will be given if some object file requires the old
7950 The new secure PLT and GOT are placed differently relative to other
7951 sections compared to older BSS PLT and GOT placement. The location of
7952 @code{.plt} must change because the new secure PLT is an initialized
7953 section while the old PLT is uninitialized. The reason for the
7954 @code{.got} change is more subtle: The new placement allows
7955 @code{.got} to be read-only in applications linked with
7956 @samp{-z relro -z now}. However, this placement means that
7957 @code{.sdata} cannot always be used in shared libraries, because the
7958 PowerPC ABI accesses @code{.sdata} in shared libraries from the GOT
7959 pointer. @samp{--sdata-got} forces the old GOT placement. PowerPC
7960 GCC doesn't use @code{.sdata} in shared libraries, so this option is
7961 really only useful for other compilers that may do so.
7963 @cindex PowerPC stub symbols
7964 @kindex --emit-stub-syms
7965 @item --emit-stub-syms
7966 This option causes @command{ld} to label linker stubs with a local
7967 symbol that encodes the stub type and destination.
7969 @cindex PowerPC TLS optimization
7970 @kindex --no-tls-optimize
7971 @item --no-tls-optimize
7972 PowerPC @command{ld} normally performs some optimization of code
7973 sequences used to access Thread-Local Storage. Use this option to
7974 disable the optimization.
7987 @node PowerPC64 ELF64
7988 @section @command{ld} and PowerPC64 64-bit ELF Support
7990 @cindex PowerPC64 ELF64 options
7992 @cindex PowerPC64 stub grouping
7993 @kindex --stub-group-size
7994 @item --stub-group-size
7995 Long branch stubs, PLT call stubs and TOC adjusting stubs are placed
7996 by @command{ld} in stub sections located between groups of input sections.
7997 @samp{--stub-group-size} specifies the maximum size of a group of input
7998 sections handled by one stub section. Since branch offsets are signed,
7999 a stub section may serve two groups of input sections, one group before
8000 the stub section, and one group after it. However, when using
8001 conditional branches that require stubs, it may be better (for branch
8002 prediction) that stub sections only serve one group of input sections.
8003 A negative value for @samp{N} chooses this scheme, ensuring that
8004 branches to stubs always use a negative offset. Two special values of
8005 @samp{N} are recognized, @samp{1} and @samp{-1}. These both instruct
8006 @command{ld} to automatically size input section groups for the branch types
8007 detected, with the same behaviour regarding stub placement as other
8008 positive or negative values of @samp{N} respectively.
8010 Note that @samp{--stub-group-size} does not split input sections. A
8011 single input section larger than the group size specified will of course
8012 create a larger group (of one section). If input sections are too
8013 large, it may not be possible for a branch to reach its stub.
8015 @cindex PowerPC64 stub symbols
8016 @kindex --emit-stub-syms
8017 @item --emit-stub-syms
8018 This option causes @command{ld} to label linker stubs with a local
8019 symbol that encodes the stub type and destination.
8021 @cindex PowerPC64 dot symbols
8023 @kindex --no-dotsyms
8026 These two options control how @command{ld} interprets version patterns
8027 in a version script. Older PowerPC64 compilers emitted both a
8028 function descriptor symbol with the same name as the function, and a
8029 code entry symbol with the name prefixed by a dot (@samp{.}). To
8030 properly version a function @samp{foo}, the version script thus needs
8031 to control both @samp{foo} and @samp{.foo}. The option
8032 @samp{--dotsyms}, on by default, automatically adds the required
8033 dot-prefixed patterns. Use @samp{--no-dotsyms} to disable this
8036 @cindex PowerPC64 register save/restore functions
8037 @kindex --save-restore-funcs
8038 @kindex --no-save-restore-funcs
8039 @item --save-restore-funcs
8040 @itemx --no-save-restore-funcs
8041 These two options control whether PowerPC64 @command{ld} automatically
8042 provides out-of-line register save and restore functions used by
8043 @samp{-Os} code. The default is to provide any such referenced
8044 function for a normal final link, and to not do so for a relocatable
8047 @cindex PowerPC64 TLS optimization
8048 @kindex --no-tls-optimize
8049 @item --no-tls-optimize
8050 PowerPC64 @command{ld} normally performs some optimization of code
8051 sequences used to access Thread-Local Storage. Use this option to
8052 disable the optimization.
8054 @cindex PowerPC64 __tls_get_addr optimization
8055 @kindex --tls-get-addr-optimize
8056 @kindex --no-tls-get-addr-optimize
8057 @kindex --tls-get-addr-regsave
8058 @kindex --no-tls-get-addr-regsave
8059 @item --tls-get-addr-optimize
8060 @itemx --no-tls-get-addr-optimize
8061 These options control how PowerPC64 @command{ld} uses a special
8062 stub to call __tls_get_addr. PowerPC64 glibc 2.22 and later support
8063 an optimization that allows the second and subsequent calls to
8064 @code{__tls_get_addr} for a given symbol to be resolved by the special
8065 stub without calling in to glibc. By default the linker enables
8066 generation of the stub when glibc advertises the availability of
8068 Using @option{--tls-get-addr-optimize} with an older glibc won't do
8069 much besides slow down your applications, but may be useful if linking
8070 an application against an older glibc with the expectation that it
8071 will normally be used on systems having a newer glibc.
8072 @option{--tls-get-addr-regsave} forces generation of a stub that saves
8073 and restores volatile registers around the call into glibc. Normally,
8074 this is done when the linker detects a call to __tls_get_addr_desc.
8075 Such calls then go via the register saving stub to __tls_get_addr_opt.
8076 @option{--no-tls-get-addr-regsave} disables generation of the
8079 @cindex PowerPC64 OPD optimization
8080 @kindex --no-opd-optimize
8081 @item --no-opd-optimize
8082 PowerPC64 @command{ld} normally removes @code{.opd} section entries
8083 corresponding to deleted link-once functions, or functions removed by
8084 the action of @samp{--gc-sections} or linker script @code{/DISCARD/}.
8085 Use this option to disable @code{.opd} optimization.
8087 @cindex PowerPC64 OPD spacing
8088 @kindex --non-overlapping-opd
8089 @item --non-overlapping-opd
8090 Some PowerPC64 compilers have an option to generate compressed
8091 @code{.opd} entries spaced 16 bytes apart, overlapping the third word,
8092 the static chain pointer (unused in C) with the first word of the next
8093 entry. This option expands such entries to the full 24 bytes.
8095 @cindex PowerPC64 TOC optimization
8096 @kindex --no-toc-optimize
8097 @item --no-toc-optimize
8098 PowerPC64 @command{ld} normally removes unused @code{.toc} section
8099 entries. Such entries are detected by examining relocations that
8100 reference the TOC in code sections. A reloc in a deleted code section
8101 marks a TOC word as unneeded, while a reloc in a kept code section
8102 marks a TOC word as needed. Since the TOC may reference itself, TOC
8103 relocs are also examined. TOC words marked as both needed and
8104 unneeded will of course be kept. TOC words without any referencing
8105 reloc are assumed to be part of a multi-word entry, and are kept or
8106 discarded as per the nearest marked preceding word. This works
8107 reliably for compiler generated code, but may be incorrect if assembly
8108 code is used to insert TOC entries. Use this option to disable the
8111 @cindex PowerPC64 inline PLT call optimization
8112 @kindex --no-inline-optimize
8113 @item --no-inline-optimize
8114 PowerPC64 @command{ld} normally replaces inline PLT call sequences
8115 marked with @code{R_PPC64_PLTSEQ}, @code{R_PPC64_PLTCALL},
8116 @code{R_PPC64_PLT16_HA} and @code{R_PPC64_PLT16_LO_DS} relocations by
8117 a number of @code{nop}s and a direct call when the function is defined
8118 locally and can't be overridden by some other definition. This option
8119 disables that optimization.
8121 @cindex PowerPC64 multi-TOC
8122 @kindex --no-multi-toc
8123 @item --no-multi-toc
8124 If given any toc option besides @code{-mcmodel=medium} or
8125 @code{-mcmodel=large}, PowerPC64 GCC generates code for a TOC model
8127 entries are accessed with a 16-bit offset from r2. This limits the
8128 total TOC size to 64K. PowerPC64 @command{ld} extends this limit by
8129 grouping code sections such that each group uses less than 64K for its
8130 TOC entries, then inserts r2 adjusting stubs between inter-group
8131 calls. @command{ld} does not split apart input sections, so cannot
8132 help if a single input file has a @code{.toc} section that exceeds
8133 64K, most likely from linking multiple files with @command{ld -r}.
8134 Use this option to turn off this feature.
8136 @cindex PowerPC64 TOC sorting
8137 @kindex --no-toc-sort
8139 By default, @command{ld} sorts TOC sections so that those whose file
8140 happens to have a section called @code{.init} or @code{.fini} are
8141 placed first, followed by TOC sections referenced by code generated
8142 with PowerPC64 gcc's @code{-mcmodel=small}, and lastly TOC sections
8143 referenced only by code generated with PowerPC64 gcc's
8144 @code{-mcmodel=medium} or @code{-mcmodel=large} options. Doing this
8145 results in better TOC grouping for multi-TOC. Use this option to turn
8148 @cindex PowerPC64 PLT stub alignment
8150 @kindex --no-plt-align
8152 @itemx --no-plt-align
8153 Use these options to control whether individual PLT call stubs are
8154 aligned to a 32-byte boundary, or to the specified power of two
8155 boundary when using @code{--plt-align=}. A negative value may be
8156 specified to pad PLT call stubs so that they do not cross the
8157 specified power of two boundary (or the minimum number of boundaries
8158 if a PLT stub is so large that it must cross a boundary). By default
8159 PLT call stubs are aligned to 32-byte boundaries.
8161 @cindex PowerPC64 PLT call stub static chain
8162 @kindex --plt-static-chain
8163 @kindex --no-plt-static-chain
8164 @item --plt-static-chain
8165 @itemx --no-plt-static-chain
8166 Use these options to control whether PLT call stubs load the static
8167 chain pointer (r11). @code{ld} defaults to not loading the static
8168 chain since there is never any need to do so on a PLT call.
8170 @cindex PowerPC64 PLT call stub thread safety
8171 @kindex --plt-thread-safe
8172 @kindex --no-plt-thread-safe
8173 @item --plt-thread-safe
8174 @itemx --no-plt-thread-safe
8175 With power7's weakly ordered memory model, it is possible when using
8176 lazy binding for ld.so to update a plt entry in one thread and have
8177 another thread see the individual plt entry words update in the wrong
8178 order, despite ld.so carefully writing in the correct order and using
8179 memory write barriers. To avoid this we need some sort of read
8180 barrier in the call stub, or use LD_BIND_NOW=1. By default, @code{ld}
8181 looks for calls to commonly used functions that create threads, and if
8182 seen, adds the necessary barriers. Use these options to change the
8185 @cindex PowerPC64 ELFv2 PLT localentry optimization
8186 @kindex --plt-localentry
8187 @kindex --no-plt-localentry
8188 @item --plt-localentry
8189 @itemx --no-localentry
8190 ELFv2 functions with localentry:0 are those with a single entry point,
8191 ie. global entry == local entry, and that have no requirement on r2
8192 (the TOC/GOT pointer) or r12, and guarantee r2 is unchanged on return.
8193 Such an external function can be called via the PLT without saving r2
8194 or restoring it on return, avoiding a common load-hit-store for small
8195 functions. The optimization is attractive, with up to 40% reduction
8196 in execution time for a small function, but can result in symbol
8197 interposition failures. Also, minor changes in a shared library,
8198 including system libraries, can cause a function that was localentry:0
8199 to become localentry:8. This will result in a dynamic loader
8200 complaint and failure to run. The option is experimental, use with
8201 care. @option{--no-plt-localentry} is the default.
8203 @cindex PowerPC64 Power10 stubs
8204 @kindex --power10-stubs
8205 @kindex --no-power10-stubs
8206 @item --power10-stubs
8207 @itemx --no-power10-stubs
8208 When PowerPC64 @command{ld} links input object files containing
8209 relocations used on power10 prefixed instructions it normally creates
8210 linkage stubs (PLT call and long branch) using power10 instructions
8211 for @code{@@notoc} PLT calls where @code{r2} is not known. The
8212 power10 notoc stubs are smaller and faster, so are preferred for
8213 power10. @option{--power10-stubs} and @option{--no-power10-stubs}
8214 allow you to override the linker's selection of stub instructions.
8215 @option{--power10-stubs=auto} allows the user to select the default
8230 @section @command{ld} and S/390 ELF Support
8232 @cindex S/390 ELF options
8236 @kindex --s390-pgste
8238 This option marks the result file with a @code{PT_S390_PGSTE}
8239 segment. The Linux kernel is supposed to allocate 4k page tables for
8240 binaries marked that way.
8254 @section @command{ld} and SPU ELF Support
8256 @cindex SPU ELF options
8262 This option marks an executable as a PIC plugin module.
8264 @cindex SPU overlays
8265 @kindex --no-overlays
8267 Normally, @command{ld} recognizes calls to functions within overlay
8268 regions, and redirects such calls to an overlay manager via a stub.
8269 @command{ld} also provides a built-in overlay manager. This option
8270 turns off all this special overlay handling.
8272 @cindex SPU overlay stub symbols
8273 @kindex --emit-stub-syms
8274 @item --emit-stub-syms
8275 This option causes @command{ld} to label overlay stubs with a local
8276 symbol that encodes the stub type and destination.
8278 @cindex SPU extra overlay stubs
8279 @kindex --extra-overlay-stubs
8280 @item --extra-overlay-stubs
8281 This option causes @command{ld} to add overlay call stubs on all
8282 function calls out of overlay regions. Normally stubs are not added
8283 on calls to non-overlay regions.
8285 @cindex SPU local store size
8286 @kindex --local-store=lo:hi
8287 @item --local-store=lo:hi
8288 @command{ld} usually checks that a final executable for SPU fits in
8289 the address range 0 to 256k. This option may be used to change the
8290 range. Disable the check entirely with @option{--local-store=0:0}.
8293 @kindex --stack-analysis
8294 @item --stack-analysis
8295 SPU local store space is limited. Over-allocation of stack space
8296 unnecessarily limits space available for code and data, while
8297 under-allocation results in runtime failures. If given this option,
8298 @command{ld} will provide an estimate of maximum stack usage.
8299 @command{ld} does this by examining symbols in code sections to
8300 determine the extents of functions, and looking at function prologues
8301 for stack adjusting instructions. A call-graph is created by looking
8302 for relocations on branch instructions. The graph is then searched
8303 for the maximum stack usage path. Note that this analysis does not
8304 find calls made via function pointers, and does not handle recursion
8305 and other cycles in the call graph. Stack usage may be
8306 under-estimated if your code makes such calls. Also, stack usage for
8307 dynamic allocation, e.g. alloca, will not be detected. If a link map
8308 is requested, detailed information about each function's stack usage
8309 and calls will be given.
8312 @kindex --emit-stack-syms
8313 @item --emit-stack-syms
8314 This option, if given along with @option{--stack-analysis} will result
8315 in @command{ld} emitting stack sizing symbols for each function.
8316 These take the form @code{__stack_<function_name>} for global
8317 functions, and @code{__stack_<number>_<function_name>} for static
8318 functions. @code{<number>} is the section id in hex. The value of
8319 such symbols is the stack requirement for the corresponding function.
8320 The symbol size will be zero, type @code{STT_NOTYPE}, binding
8321 @code{STB_LOCAL}, and section @code{SHN_ABS}.
8335 @section @command{ld}'s Support for Various TI COFF Versions
8336 @cindex TI COFF versions
8337 @kindex --format=@var{version}
8338 The @samp{--format} switch allows selection of one of the various
8339 TI COFF versions. The latest of this writing is 2; versions 0 and 1 are
8340 also supported. The TI COFF versions also vary in header byte-order
8341 format; @command{ld} will read any version or byte order, but the output
8342 header format depends on the default specified by the specific target.
8355 @section @command{ld} and WIN32 (cygwin/mingw)
8357 This section describes some of the win32 specific @command{ld} issues.
8358 See @ref{Options,,Command-line Options} for detailed description of the
8359 command-line options mentioned here.
8362 @cindex import libraries
8363 @item import libraries
8364 The standard Windows linker creates and uses so-called import
8365 libraries, which contains information for linking to dll's. They are
8366 regular static archives and are handled as any other static
8367 archive. The cygwin and mingw ports of @command{ld} have specific
8368 support for creating such libraries provided with the
8369 @samp{--out-implib} command-line option.
8371 @item exporting DLL symbols
8372 @cindex exporting DLL symbols
8373 The cygwin/mingw @command{ld} has several ways to export symbols for dll's.
8376 @item using auto-export functionality
8377 @cindex using auto-export functionality
8378 By default @command{ld} exports symbols with the auto-export functionality,
8379 which is controlled by the following command-line options:
8382 @item --export-all-symbols [This is the default]
8383 @item --exclude-symbols
8384 @item --exclude-libs
8385 @item --exclude-modules-for-implib
8386 @item --version-script
8389 When auto-export is in operation, @command{ld} will export all the non-local
8390 (global and common) symbols it finds in a DLL, with the exception of a few
8391 symbols known to belong to the system's runtime and libraries. As it will
8392 often not be desirable to export all of a DLL's symbols, which may include
8393 private functions that are not part of any public interface, the command-line
8394 options listed above may be used to filter symbols out from the list for
8395 exporting. The @samp{--output-def} option can be used in order to see the
8396 final list of exported symbols with all exclusions taken into effect.
8398 If @samp{--export-all-symbols} is not given explicitly on the
8399 command line, then the default auto-export behavior will be @emph{disabled}
8400 if either of the following are true:
8403 @item A DEF file is used.
8404 @item Any symbol in any object file was marked with the __declspec(dllexport) attribute.
8407 @item using a DEF file
8408 @cindex using a DEF file
8409 Another way of exporting symbols is using a DEF file. A DEF file is
8410 an ASCII file containing definitions of symbols which should be
8411 exported when a dll is created. Usually it is named @samp{<dll
8412 name>.def} and is added as any other object file to the linker's
8413 command line. The file's name must end in @samp{.def} or @samp{.DEF}.
8416 gcc -o <output> <objectfiles> <dll name>.def
8419 Using a DEF file turns off the normal auto-export behavior, unless the
8420 @samp{--export-all-symbols} option is also used.
8422 Here is an example of a DEF file for a shared library called @samp{xyz.dll}:
8425 LIBRARY "xyz.dll" BASE=0x20000000
8431 another_foo = abc.dll.afoo
8437 This example defines a DLL with a non-default base address and seven
8438 symbols in the export table. The third exported symbol @code{_bar} is an
8439 alias for the second. The fourth symbol, @code{another_foo} is resolved
8440 by "forwarding" to another module and treating it as an alias for
8441 @code{afoo} exported from the DLL @samp{abc.dll}. The final symbol
8442 @code{var1} is declared to be a data object. The @samp{doo} symbol in
8443 export library is an alias of @samp{foo}, which gets the string name
8444 in export table @samp{foo2}. The @samp{eoo} symbol is an data export
8445 symbol, which gets in export table the name @samp{var1}.
8447 The optional @code{LIBRARY <name>} command indicates the @emph{internal}
8448 name of the output DLL. If @samp{<name>} does not include a suffix,
8449 the default library suffix, @samp{.DLL} is appended.
8451 When the .DEF file is used to build an application, rather than a
8452 library, the @code{NAME <name>} command should be used instead of
8453 @code{LIBRARY}. If @samp{<name>} does not include a suffix, the default
8454 executable suffix, @samp{.EXE} is appended.
8456 With either @code{LIBRARY <name>} or @code{NAME <name>} the optional
8457 specification @code{BASE = <number>} may be used to specify a
8458 non-default base address for the image.
8460 If neither @code{LIBRARY <name>} nor @code{NAME <name>} is specified,
8461 or they specify an empty string, the internal name is the same as the
8462 filename specified on the command line.
8464 The complete specification of an export symbol is:
8468 ( ( ( <name1> [ = <name2> ] )
8469 | ( <name1> = <module-name> . <external-name>))
8470 [ @@ <integer> ] [NONAME] [DATA] [CONSTANT] [PRIVATE] [== <name3>] ) *
8473 Declares @samp{<name1>} as an exported symbol from the DLL, or declares
8474 @samp{<name1>} as an exported alias for @samp{<name2>}; or declares
8475 @samp{<name1>} as a "forward" alias for the symbol
8476 @samp{<external-name>} in the DLL @samp{<module-name>}.
8477 Optionally, the symbol may be exported by the specified ordinal
8478 @samp{<integer>} alias. The optional @samp{<name3>} is the to be used
8479 string in import/export table for the symbol.
8481 The optional keywords that follow the declaration indicate:
8483 @code{NONAME}: Do not put the symbol name in the DLL's export table. It
8484 will still be exported by its ordinal alias (either the value specified
8485 by the .def specification or, otherwise, the value assigned by the
8486 linker). The symbol name, however, does remain visible in the import
8487 library (if any), unless @code{PRIVATE} is also specified.
8489 @code{DATA}: The symbol is a variable or object, rather than a function.
8490 The import lib will export only an indirect reference to @code{foo} as
8491 the symbol @code{_imp__foo} (ie, @code{foo} must be resolved as
8494 @code{CONSTANT}: Like @code{DATA}, but put the undecorated @code{foo} as
8495 well as @code{_imp__foo} into the import library. Both refer to the
8496 read-only import address table's pointer to the variable, not to the
8497 variable itself. This can be dangerous. If the user code fails to add
8498 the @code{dllimport} attribute and also fails to explicitly add the
8499 extra indirection that the use of the attribute enforces, the
8500 application will behave unexpectedly.
8502 @code{PRIVATE}: Put the symbol in the DLL's export table, but do not put
8503 it into the static import library used to resolve imports at link time. The
8504 symbol can still be imported using the @code{LoadLibrary/GetProcAddress}
8505 API at runtime or by using the GNU ld extension of linking directly to
8506 the DLL without an import library.
8508 See ld/deffilep.y in the binutils sources for the full specification of
8509 other DEF file statements
8511 @cindex creating a DEF file
8512 While linking a shared dll, @command{ld} is able to create a DEF file
8513 with the @samp{--output-def <file>} command-line option.
8515 @item Using decorations
8516 @cindex Using decorations
8517 Another way of marking symbols for export is to modify the source code
8518 itself, so that when building the DLL each symbol to be exported is
8522 __declspec(dllexport) int a_variable
8523 __declspec(dllexport) void a_function(int with_args)
8526 All such symbols will be exported from the DLL. If, however,
8527 any of the object files in the DLL contain symbols decorated in
8528 this way, then the normal auto-export behavior is disabled, unless
8529 the @samp{--export-all-symbols} option is also used.
8531 Note that object files that wish to access these symbols must @emph{not}
8532 decorate them with dllexport. Instead, they should use dllimport,
8536 __declspec(dllimport) int a_variable
8537 __declspec(dllimport) void a_function(int with_args)
8540 This complicates the structure of library header files, because
8541 when included by the library itself the header must declare the
8542 variables and functions as dllexport, but when included by client
8543 code the header must declare them as dllimport. There are a number
8544 of idioms that are typically used to do this; often client code can
8545 omit the __declspec() declaration completely. See
8546 @samp{--enable-auto-import} and @samp{automatic data imports} for more
8550 @cindex automatic data imports
8551 @item automatic data imports
8552 The standard Windows dll format supports data imports from dlls only
8553 by adding special decorations (dllimport/dllexport), which let the
8554 compiler produce specific assembler instructions to deal with this
8555 issue. This increases the effort necessary to port existing Un*x
8556 code to these platforms, especially for large
8557 c++ libraries and applications. The auto-import feature, which was
8558 initially provided by Paul Sokolovsky, allows one to omit the
8559 decorations to achieve a behavior that conforms to that on POSIX/Un*x
8560 platforms. This feature is enabled with the @samp{--enable-auto-import}
8561 command-line option, although it is enabled by default on cygwin/mingw.
8562 The @samp{--enable-auto-import} option itself now serves mainly to
8563 suppress any warnings that are ordinarily emitted when linked objects
8564 trigger the feature's use.
8566 auto-import of variables does not always work flawlessly without
8567 additional assistance. Sometimes, you will see this message
8569 "variable '<var>' can't be auto-imported. Please read the
8570 documentation for ld's @code{--enable-auto-import} for details."
8572 The @samp{--enable-auto-import} documentation explains why this error
8573 occurs, and several methods that can be used to overcome this difficulty.
8574 One of these methods is the @emph{runtime pseudo-relocs} feature, described
8577 @cindex runtime pseudo-relocation
8578 For complex variables imported from DLLs (such as structs or classes),
8579 object files typically contain a base address for the variable and an
8580 offset (@emph{addend}) within the variable--to specify a particular
8581 field or public member, for instance. Unfortunately, the runtime loader used
8582 in win32 environments is incapable of fixing these references at runtime
8583 without the additional information supplied by dllimport/dllexport decorations.
8584 The standard auto-import feature described above is unable to resolve these
8587 The @samp{--enable-runtime-pseudo-relocs} switch allows these references to
8588 be resolved without error, while leaving the task of adjusting the references
8589 themselves (with their non-zero addends) to specialized code provided by the
8590 runtime environment. Recent versions of the cygwin and mingw environments and
8591 compilers provide this runtime support; older versions do not. However, the
8592 support is only necessary on the developer's platform; the compiled result will
8593 run without error on an older system.
8595 @samp{--enable-runtime-pseudo-relocs} is not the default; it must be explicitly
8598 @cindex direct linking to a dll
8599 @item direct linking to a dll
8600 The cygwin/mingw ports of @command{ld} support the direct linking,
8601 including data symbols, to a dll without the usage of any import
8602 libraries. This is much faster and uses much less memory than does the
8603 traditional import library method, especially when linking large
8604 libraries or applications. When @command{ld} creates an import lib, each
8605 function or variable exported from the dll is stored in its own bfd, even
8606 though a single bfd could contain many exports. The overhead involved in
8607 storing, loading, and processing so many bfd's is quite large, and explains the
8608 tremendous time, memory, and storage needed to link against particularly
8609 large or complex libraries when using import libs.
8611 Linking directly to a dll uses no extra command-line switches other than
8612 @samp{-L} and @samp{-l}, because @command{ld} already searches for a number
8613 of names to match each library. All that is needed from the developer's
8614 perspective is an understanding of this search, in order to force ld to
8615 select the dll instead of an import library.
8618 For instance, when ld is called with the argument @samp{-lxxx} it will attempt
8619 to find, in the first directory of its search path,
8632 before moving on to the next directory in the search path.
8634 (*) Actually, this is not @samp{cygxxx.dll} but in fact is @samp{<prefix>xxx.dll},
8635 where @samp{<prefix>} is set by the @command{ld} option
8636 @samp{--dll-search-prefix=<prefix>}. In the case of cygwin, the standard gcc spec
8637 file includes @samp{--dll-search-prefix=cyg}, so in effect we actually search for
8640 Other win32-based unix environments, such as mingw or pw32, may use other
8641 @samp{<prefix>}es, although at present only cygwin makes use of this feature. It
8642 was originally intended to help avoid name conflicts among dll's built for the
8643 various win32/un*x environments, so that (for example) two versions of a zlib dll
8644 could coexist on the same machine.
8646 The generic cygwin/mingw path layout uses a @samp{bin} directory for
8647 applications and dll's and a @samp{lib} directory for the import
8648 libraries (using cygwin nomenclature):
8654 libxxx.dll.a (in case of dll's)
8655 libxxx.a (in case of static archive)
8658 Linking directly to a dll without using the import library can be
8661 1. Use the dll directly by adding the @samp{bin} path to the link line
8663 gcc -Wl,-verbose -o a.exe -L../bin/ -lxxx
8666 However, as the dll's often have version numbers appended to their names
8667 (@samp{cygncurses-5.dll}) this will often fail, unless one specifies
8668 @samp{-L../bin -lncurses-5} to include the version. Import libs are generally
8669 not versioned, and do not have this difficulty.
8671 2. Create a symbolic link from the dll to a file in the @samp{lib}
8672 directory according to the above mentioned search pattern. This
8673 should be used to avoid unwanted changes in the tools needed for
8677 ln -s bin/cygxxx.dll lib/[cyg|lib|]xxx.dll[.a]
8680 Then you can link without any make environment changes.
8683 gcc -Wl,-verbose -o a.exe -L../lib/ -lxxx
8686 This technique also avoids the version number problems, because the following is
8693 libxxx.dll.a -> ../bin/cygxxx-5.dll
8696 Linking directly to a dll without using an import lib will work
8697 even when auto-import features are exercised, and even when
8698 @samp{--enable-runtime-pseudo-relocs} is used.
8700 Given the improvements in speed and memory usage, one might justifiably
8701 wonder why import libraries are used at all. There are three reasons:
8703 1. Until recently, the link-directly-to-dll functionality did @emph{not}
8704 work with auto-imported data.
8706 2. Sometimes it is necessary to include pure static objects within the
8707 import library (which otherwise contains only bfd's for indirection
8708 symbols that point to the exports of a dll). Again, the import lib
8709 for the cygwin kernel makes use of this ability, and it is not
8710 possible to do this without an import lib.
8712 3. Symbol aliases can only be resolved using an import lib. This is
8713 critical when linking against OS-supplied dll's (eg, the win32 API)
8714 in which symbols are usually exported as undecorated aliases of their
8715 stdcall-decorated assembly names.
8717 So, import libs are not going away. But the ability to replace
8718 true import libs with a simple symbolic link to (or a copy of)
8719 a dll, in many cases, is a useful addition to the suite of tools
8720 binutils makes available to the win32 developer. Given the
8721 massive improvements in memory requirements during linking, storage
8722 requirements, and linking speed, we expect that many developers
8723 will soon begin to use this feature whenever possible.
8725 @item symbol aliasing
8727 @item adding additional names
8728 Sometimes, it is useful to export symbols with additional names.
8729 A symbol @samp{foo} will be exported as @samp{foo}, but it can also be
8730 exported as @samp{_foo} by using special directives in the DEF file
8731 when creating the dll. This will affect also the optional created
8732 import library. Consider the following DEF file:
8735 LIBRARY "xyz.dll" BASE=0x61000000
8742 The line @samp{_foo = foo} maps the symbol @samp{foo} to @samp{_foo}.
8744 Another method for creating a symbol alias is to create it in the
8745 source code using the "weak" attribute:
8748 void foo () @{ /* Do something. */; @}
8749 void _foo () __attribute__ ((weak, alias ("foo")));
8752 See the gcc manual for more information about attributes and weak
8755 @item renaming symbols
8756 Sometimes it is useful to rename exports. For instance, the cygwin
8757 kernel does this regularly. A symbol @samp{_foo} can be exported as
8758 @samp{foo} but not as @samp{_foo} by using special directives in the
8759 DEF file. (This will also affect the import library, if it is
8760 created). In the following example:
8763 LIBRARY "xyz.dll" BASE=0x61000000
8769 The line @samp{_foo = foo} maps the exported symbol @samp{foo} to
8773 Note: using a DEF file disables the default auto-export behavior,
8774 unless the @samp{--export-all-symbols} command-line option is used.
8775 If, however, you are trying to rename symbols, then you should list
8776 @emph{all} desired exports in the DEF file, including the symbols
8777 that are not being renamed, and do @emph{not} use the
8778 @samp{--export-all-symbols} option. If you list only the
8779 renamed symbols in the DEF file, and use @samp{--export-all-symbols}
8780 to handle the other symbols, then the both the new names @emph{and}
8781 the original names for the renamed symbols will be exported.
8782 In effect, you'd be aliasing those symbols, not renaming them,
8783 which is probably not what you wanted.
8785 @cindex weak externals
8786 @item weak externals
8787 The Windows object format, PE, specifies a form of weak symbols called
8788 weak externals. When a weak symbol is linked and the symbol is not
8789 defined, the weak symbol becomes an alias for some other symbol. There
8790 are three variants of weak externals:
8792 @item Definition is searched for in objects and libraries, historically
8793 called lazy externals.
8794 @item Definition is searched for only in other objects, not in libraries.
8795 This form is not presently implemented.
8796 @item No search; the symbol is an alias. This form is not presently
8799 As a GNU extension, weak symbols that do not specify an alternate symbol
8800 are supported. If the symbol is undefined when linking, the symbol
8801 uses a default value.
8803 @cindex aligned common symbols
8804 @item aligned common symbols
8805 As a GNU extension to the PE file format, it is possible to specify the
8806 desired alignment for a common symbol. This information is conveyed from
8807 the assembler or compiler to the linker by means of GNU-specific commands
8808 carried in the object file's @samp{.drectve} section, which are recognized
8809 by @command{ld} and respected when laying out the common symbols. Native
8810 tools will be able to process object files employing this GNU extension,
8811 but will fail to respect the alignment instructions, and may issue noisy
8812 warnings about unknown linker directives.
8827 @section @code{ld} and Xtensa Processors
8829 @cindex Xtensa processors
8830 The default @command{ld} behavior for Xtensa processors is to interpret
8831 @code{SECTIONS} commands so that lists of explicitly named sections in a
8832 specification with a wildcard file will be interleaved when necessary to
8833 keep literal pools within the range of PC-relative load offsets. For
8834 example, with the command:
8846 @command{ld} may interleave some of the @code{.literal}
8847 and @code{.text} sections from different object files to ensure that the
8848 literal pools are within the range of PC-relative load offsets. A valid
8849 interleaving might place the @code{.literal} sections from an initial
8850 group of files followed by the @code{.text} sections of that group of
8851 files. Then, the @code{.literal} sections from the rest of the files
8852 and the @code{.text} sections from the rest of the files would follow.
8854 @cindex @option{--relax} on Xtensa
8855 @cindex relaxing on Xtensa
8856 Relaxation is enabled by default for the Xtensa version of @command{ld} and
8857 provides two important link-time optimizations. The first optimization
8858 is to combine identical literal values to reduce code size. A redundant
8859 literal will be removed and all the @code{L32R} instructions that use it
8860 will be changed to reference an identical literal, as long as the
8861 location of the replacement literal is within the offset range of all
8862 the @code{L32R} instructions. The second optimization is to remove
8863 unnecessary overhead from assembler-generated ``longcall'' sequences of
8864 @code{L32R}/@code{CALLX@var{n}} when the target functions are within
8865 range of direct @code{CALL@var{n}} instructions.
8867 For each of these cases where an indirect call sequence can be optimized
8868 to a direct call, the linker will change the @code{CALLX@var{n}}
8869 instruction to a @code{CALL@var{n}} instruction, remove the @code{L32R}
8870 instruction, and remove the literal referenced by the @code{L32R}
8871 instruction if it is not used for anything else. Removing the
8872 @code{L32R} instruction always reduces code size but can potentially
8873 hurt performance by changing the alignment of subsequent branch targets.
8874 By default, the linker will always preserve alignments, either by
8875 switching some instructions between 24-bit encodings and the equivalent
8876 density instructions or by inserting a no-op in place of the @code{L32R}
8877 instruction that was removed. If code size is more important than
8878 performance, the @option{--size-opt} option can be used to prevent the
8879 linker from widening density instructions or inserting no-ops, except in
8880 a few cases where no-ops are required for correctness.
8882 The following Xtensa-specific command-line options can be used to
8885 @cindex Xtensa options
8888 When optimizing indirect calls to direct calls, optimize for code size
8889 more than performance. With this option, the linker will not insert
8890 no-ops or widen density instructions to preserve branch target
8891 alignment. There may still be some cases where no-ops are required to
8892 preserve the correctness of the code.
8894 @item --abi-windowed
8896 Choose ABI for the output object and for the generated PLT code.
8897 PLT code inserted by the linker must match ABI of the output object
8898 because windowed and call0 ABI use incompatible function call
8900 Default ABI is chosen by the ABI tag in the @code{.xtensa.info} section
8901 of the first input object.
8902 A warning is issued if ABI tags of input objects do not match each other
8903 or the chosen output object ABI.
8911 @ifclear SingleFormat
8916 @cindex object file management
8917 @cindex object formats available
8919 The linker accesses object and archive files using the BFD libraries.
8920 These libraries allow the linker to use the same routines to operate on
8921 object files whatever the object file format. A different object file
8922 format can be supported simply by creating a new BFD back end and adding
8923 it to the library. To conserve runtime memory, however, the linker and
8924 associated tools are usually configured to support only a subset of the
8925 object file formats available. You can use @code{objdump -i}
8926 (@pxref{objdump,,objdump,binutils.info,The GNU Binary Utilities}) to
8927 list all the formats available for your configuration.
8929 @cindex BFD requirements
8930 @cindex requirements for BFD
8931 As with most implementations, BFD is a compromise between
8932 several conflicting requirements. The major factor influencing
8933 BFD design was efficiency: any time used converting between
8934 formats is time which would not have been spent had BFD not
8935 been involved. This is partly offset by abstraction payback; since
8936 BFD simplifies applications and back ends, more time and care
8937 may be spent optimizing algorithms for a greater speed.
8939 One minor artifact of the BFD solution which you should bear in
8940 mind is the potential for information loss. There are two places where
8941 useful information can be lost using the BFD mechanism: during
8942 conversion and during output. @xref{BFD information loss}.
8945 * BFD outline:: How it works: an outline of BFD
8949 @section How It Works: An Outline of BFD
8950 @cindex opening object files
8951 @include bfdsumm.texi
8954 @node Reporting Bugs
8955 @chapter Reporting Bugs
8956 @cindex bugs in @command{ld}
8957 @cindex reporting bugs in @command{ld}
8959 Your bug reports play an essential role in making @command{ld} reliable.
8961 Reporting a bug may help you by bringing a solution to your problem, or
8962 it may not. But in any case the principal function of a bug report is
8963 to help the entire community by making the next version of @command{ld}
8964 work better. Bug reports are your contribution to the maintenance of
8967 In order for a bug report to serve its purpose, you must include the
8968 information that enables us to fix the bug.
8971 * Bug Criteria:: Have you found a bug?
8972 * Bug Reporting:: How to report bugs
8976 @section Have You Found a Bug?
8977 @cindex bug criteria
8979 If you are not sure whether you have found a bug, here are some guidelines:
8982 @cindex fatal signal
8983 @cindex linker crash
8984 @cindex crash of linker
8986 If the linker gets a fatal signal, for any input whatever, that is a
8987 @command{ld} bug. Reliable linkers never crash.
8989 @cindex error on valid input
8991 If @command{ld} produces an error message for valid input, that is a bug.
8993 @cindex invalid input
8995 If @command{ld} does not produce an error message for invalid input, that
8996 may be a bug. In the general case, the linker can not verify that
8997 object files are correct.
9000 If you are an experienced user of linkers, your suggestions for
9001 improvement of @command{ld} are welcome in any case.
9005 @section How to Report Bugs
9007 @cindex @command{ld} bugs, reporting
9009 A number of companies and individuals offer support for @sc{gnu}
9010 products. If you obtained @command{ld} from a support organization, we
9011 recommend you contact that organization first.
9013 You can find contact information for many support companies and
9014 individuals in the file @file{etc/SERVICE} in the @sc{gnu} Emacs
9018 Otherwise, send bug reports for @command{ld} to
9022 The fundamental principle of reporting bugs usefully is this:
9023 @strong{report all the facts}. If you are not sure whether to state a
9024 fact or leave it out, state it!
9026 Often people omit facts because they think they know what causes the
9027 problem and assume that some details do not matter. Thus, you might
9028 assume that the name of a symbol you use in an example does not
9029 matter. Well, probably it does not, but one cannot be sure. Perhaps
9030 the bug is a stray memory reference which happens to fetch from the
9031 location where that name is stored in memory; perhaps, if the name
9032 were different, the contents of that location would fool the linker
9033 into doing the right thing despite the bug. Play it safe and give a
9034 specific, complete example. That is the easiest thing for you to do,
9035 and the most helpful.
9037 Keep in mind that the purpose of a bug report is to enable us to fix
9038 the bug if it is new to us. Therefore, always write your bug reports
9039 on the assumption that the bug has not been reported previously.
9041 Sometimes people give a few sketchy facts and ask, ``Does this ring a
9042 bell?'' This cannot help us fix a bug, so it is basically useless. We
9043 respond by asking for enough details to enable us to investigate.
9044 You might as well expedite matters by sending them to begin with.
9046 To enable us to fix the bug, you should include all these things:
9050 The version of @command{ld}. @command{ld} announces it if you start it with
9051 the @samp{--version} argument.
9053 Without this, we will not know whether there is any point in looking for
9054 the bug in the current version of @command{ld}.
9057 Any patches you may have applied to the @command{ld} source, including any
9058 patches made to the @code{BFD} library.
9061 The type of machine you are using, and the operating system name and
9065 What compiler (and its version) was used to compile @command{ld}---e.g.
9069 The command arguments you gave the linker to link your example and
9070 observe the bug. To guarantee you will not omit something important,
9071 list them all. A copy of the Makefile (or the output from make) is
9074 If we were to try to guess the arguments, we would probably guess wrong
9075 and then we might not encounter the bug.
9078 A complete input file, or set of input files, that will reproduce the
9079 bug. It is generally most helpful to send the actual object files
9080 provided that they are reasonably small. Say no more than 10K. For
9081 bigger files you can either make them available by FTP or HTTP or else
9082 state that you are willing to send the object file(s) to whomever
9083 requests them. (Note - your email will be going to a mailing list, so
9084 we do not want to clog it up with large attachments). But small
9085 attachments are best.
9087 If the source files were assembled using @code{gas} or compiled using
9088 @code{gcc}, then it may be OK to send the source files rather than the
9089 object files. In this case, be sure to say exactly what version of
9090 @code{gas} or @code{gcc} was used to produce the object files. Also say
9091 how @code{gas} or @code{gcc} were configured.
9094 A description of what behavior you observe that you believe is
9095 incorrect. For example, ``It gets a fatal signal.''
9097 Of course, if the bug is that @command{ld} gets a fatal signal, then we
9098 will certainly notice it. But if the bug is incorrect output, we might
9099 not notice unless it is glaringly wrong. You might as well not give us
9100 a chance to make a mistake.
9102 Even if the problem you experience is a fatal signal, you should still
9103 say so explicitly. Suppose something strange is going on, such as, your
9104 copy of @command{ld} is out of sync, or you have encountered a bug in the
9105 C library on your system. (This has happened!) Your copy might crash
9106 and ours would not. If you told us to expect a crash, then when ours
9107 fails to crash, we would know that the bug was not happening for us. If
9108 you had not told us to expect a crash, then we would not be able to draw
9109 any conclusion from our observations.
9112 If you wish to suggest changes to the @command{ld} source, send us context
9113 diffs, as generated by @code{diff} with the @samp{-u}, @samp{-c}, or
9114 @samp{-p} option. Always send diffs from the old file to the new file.
9115 If you even discuss something in the @command{ld} source, refer to it by
9116 context, not by line number.
9118 The line numbers in our development sources will not match those in your
9119 sources. Your line numbers would convey no useful information to us.
9122 Here are some things that are not necessary:
9126 A description of the envelope of the bug.
9128 Often people who encounter a bug spend a lot of time investigating
9129 which changes to the input file will make the bug go away and which
9130 changes will not affect it.
9132 This is often time consuming and not very useful, because the way we
9133 will find the bug is by running a single example under the debugger
9134 with breakpoints, not by pure deduction from a series of examples.
9135 We recommend that you save your time for something else.
9137 Of course, if you can find a simpler example to report @emph{instead}
9138 of the original one, that is a convenience for us. Errors in the
9139 output will be easier to spot, running under the debugger will take
9140 less time, and so on.
9142 However, simplification is not vital; if you do not want to do this,
9143 report the bug anyway and send us the entire test case you used.
9146 A patch for the bug.
9148 A patch for the bug does help us if it is a good one. But do not omit
9149 the necessary information, such as the test case, on the assumption that
9150 a patch is all we need. We might see problems with your patch and decide
9151 to fix the problem another way, or we might not understand it at all.
9153 Sometimes with a program as complicated as @command{ld} it is very hard to
9154 construct an example that will make the program follow a certain path
9155 through the code. If you do not send us the example, we will not be
9156 able to construct one, so we will not be able to verify that the bug is
9159 And if we cannot understand what bug you are trying to fix, or why your
9160 patch should be an improvement, we will not install it. A test case will
9161 help us to understand.
9164 A guess about what the bug is or what it depends on.
9166 Such guesses are usually wrong. Even we cannot guess right about such
9167 things without first using the debugger to find the facts.
9171 @appendix MRI Compatible Script Files
9172 @cindex MRI compatibility
9173 To aid users making the transition to @sc{gnu} @command{ld} from the MRI
9174 linker, @command{ld} can use MRI compatible linker scripts as an
9175 alternative to the more general-purpose linker scripting language
9176 described in @ref{Scripts}. MRI compatible linker scripts have a much
9177 simpler command set than the scripting language otherwise used with
9178 @command{ld}. @sc{gnu} @command{ld} supports the most commonly used MRI
9179 linker commands; these commands are described here.
9181 In general, MRI scripts aren't of much use with the @code{a.out} object
9182 file format, since it only has three sections and MRI scripts lack some
9183 features to make use of them.
9185 You can specify a file containing an MRI-compatible script using the
9186 @samp{-c} command-line option.
9188 Each command in an MRI-compatible script occupies its own line; each
9189 command line starts with the keyword that identifies the command (though
9190 blank lines are also allowed for punctuation). If a line of an
9191 MRI-compatible script begins with an unrecognized keyword, @command{ld}
9192 issues a warning message, but continues processing the script.
9194 Lines beginning with @samp{*} are comments.
9196 You can write these commands using all upper-case letters, or all
9197 lower case; for example, @samp{chip} is the same as @samp{CHIP}.
9198 The following list shows only the upper-case form of each command.
9201 @cindex @code{ABSOLUTE} (MRI)
9202 @item ABSOLUTE @var{secname}
9203 @itemx ABSOLUTE @var{secname}, @var{secname}, @dots{} @var{secname}
9204 Normally, @command{ld} includes in the output file all sections from all
9205 the input files. However, in an MRI-compatible script, you can use the
9206 @code{ABSOLUTE} command to restrict the sections that will be present in
9207 your output program. If the @code{ABSOLUTE} command is used at all in a
9208 script, then only the sections named explicitly in @code{ABSOLUTE}
9209 commands will appear in the linker output. You can still use other
9210 input sections (whatever you select on the command line, or using
9211 @code{LOAD}) to resolve addresses in the output file.
9213 @cindex @code{ALIAS} (MRI)
9214 @item ALIAS @var{out-secname}, @var{in-secname}
9215 Use this command to place the data from input section @var{in-secname}
9216 in a section called @var{out-secname} in the linker output file.
9218 @var{in-secname} may be an integer.
9220 @cindex @code{ALIGN} (MRI)
9221 @item ALIGN @var{secname} = @var{expression}
9222 Align the section called @var{secname} to @var{expression}. The
9223 @var{expression} should be a power of two.
9225 @cindex @code{BASE} (MRI)
9226 @item BASE @var{expression}
9227 Use the value of @var{expression} as the lowest address (other than
9228 absolute addresses) in the output file.
9230 @cindex @code{CHIP} (MRI)
9231 @item CHIP @var{expression}
9232 @itemx CHIP @var{expression}, @var{expression}
9233 This command does nothing; it is accepted only for compatibility.
9235 @cindex @code{END} (MRI)
9237 This command does nothing whatever; it's only accepted for compatibility.
9239 @cindex @code{FORMAT} (MRI)
9240 @item FORMAT @var{output-format}
9241 Similar to the @code{OUTPUT_FORMAT} command in the more general linker
9242 language, but restricted to S-records, if @var{output-format} is @samp{S}
9244 @cindex @code{LIST} (MRI)
9245 @item LIST @var{anything}@dots{}
9246 Print (to the standard output file) a link map, as produced by the
9247 @command{ld} command-line option @samp{-M}.
9249 The keyword @code{LIST} may be followed by anything on the
9250 same line, with no change in its effect.
9252 @cindex @code{LOAD} (MRI)
9253 @item LOAD @var{filename}
9254 @itemx LOAD @var{filename}, @var{filename}, @dots{} @var{filename}
9255 Include one or more object file @var{filename} in the link; this has the
9256 same effect as specifying @var{filename} directly on the @command{ld}
9259 @cindex @code{NAME} (MRI)
9260 @item NAME @var{output-name}
9261 @var{output-name} is the name for the program produced by @command{ld}; the
9262 MRI-compatible command @code{NAME} is equivalent to the command-line
9263 option @samp{-o} or the general script language command @code{OUTPUT}.
9265 @cindex @code{ORDER} (MRI)
9266 @item ORDER @var{secname}, @var{secname}, @dots{} @var{secname}
9267 @itemx ORDER @var{secname} @var{secname} @var{secname}
9268 Normally, @command{ld} orders the sections in its output file in the
9269 order in which they first appear in the input files. In an MRI-compatible
9270 script, you can override this ordering with the @code{ORDER} command. The
9271 sections you list with @code{ORDER} will appear first in your output
9272 file, in the order specified.
9274 @cindex @code{PUBLIC} (MRI)
9275 @item PUBLIC @var{name}=@var{expression}
9276 @itemx PUBLIC @var{name},@var{expression}
9277 @itemx PUBLIC @var{name} @var{expression}
9278 Supply a value (@var{expression}) for external symbol
9279 @var{name} used in the linker input files.
9281 @cindex @code{SECT} (MRI)
9282 @item SECT @var{secname}, @var{expression}
9283 @itemx SECT @var{secname}=@var{expression}
9284 @itemx SECT @var{secname} @var{expression}
9285 You can use any of these three forms of the @code{SECT} command to
9286 specify the start address (@var{expression}) for section @var{secname}.
9287 If you have more than one @code{SECT} statement for the same
9288 @var{secname}, only the @emph{first} sets the start address.
9291 @node GNU Free Documentation License
9292 @appendix GNU Free Documentation License
9296 @unnumbered LD Index
9301 % I think something like @@colophon should be in texinfo. In the
9303 \long\def\colophon{\hbox to0pt{}\vfill
9304 \centerline{The body of this manual is set in}
9305 \centerline{\fontname\tenrm,}
9306 \centerline{with headings in {\bf\fontname\tenbf}}
9307 \centerline{and examples in {\tt\fontname\tentt}.}
9308 \centerline{{\it\fontname\tenit\/} and}
9309 \centerline{{\sl\fontname\tensl\/}}
9310 \centerline{are used for emphasis.}\vfill}
9312 % Blame: doc@@cygnus.com, 28mar91.