3 @c Copyright 1991, 1992, 1993, 1994, 1995, 1996, 1997, 1998, 1999, 2000,
4 @c 2001, 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009, 2010, 2011
5 @c Free Software Foundation, Inc.
8 @include configdoc.texi
9 @c (configdoc.texi is generated by the Makefile)
15 @macro gcctabopt{body}
21 @c Configure for the generation of man pages
44 @dircategory Software development
46 * Ld: (ld). The GNU linker.
51 This file documents the @sc{gnu} linker LD
52 @ifset VERSION_PACKAGE
53 @value{VERSION_PACKAGE}
55 version @value{VERSION}.
57 Copyright @copyright{} 1991, 1992, 1993, 1994, 1995, 1996, 1997, 1998, 1999, 2000,
58 2001, 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009 Free Software Foundation, Inc.
60 Permission is granted to copy, distribute and/or modify this document
61 under the terms of the GNU Free Documentation License, Version 1.3
62 or any later version published by the Free Software Foundation;
63 with no Invariant Sections, with no Front-Cover Texts, and with no
64 Back-Cover Texts. A copy of the license is included in the
65 section entitled ``GNU Free Documentation License''.
69 @setchapternewpage odd
70 @settitle The GNU linker
75 @ifset VERSION_PACKAGE
76 @subtitle @value{VERSION_PACKAGE}
78 @subtitle Version @value{VERSION}
79 @author Steve Chamberlain
80 @author Ian Lance Taylor
85 \hfill Red Hat Inc\par
86 \hfill nickc\@credhat.com, doc\@redhat.com\par
87 \hfill {\it The GNU linker}\par
88 \hfill Edited by Jeffrey Osier (jeffrey\@cygnus.com)\par
90 \global\parindent=0pt % Steve likes it this way.
93 @vskip 0pt plus 1filll
94 @c man begin COPYRIGHT
95 Copyright @copyright{} 1991, 1992, 1993, 1994, 1995, 1996, 1997, 1998,
96 1999, 2000, 2001, 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009 Free
97 Software Foundation, Inc.
99 Permission is granted to copy, distribute and/or modify this document
100 under the terms of the GNU Free Documentation License, Version 1.3
101 or any later version published by the Free Software Foundation;
102 with no Invariant Sections, with no Front-Cover Texts, and with no
103 Back-Cover Texts. A copy of the license is included in the
104 section entitled ``GNU Free Documentation License''.
110 @c FIXME: Talk about importance of *order* of args, cmds to linker!
115 This file documents the @sc{gnu} linker ld
116 @ifset VERSION_PACKAGE
117 @value{VERSION_PACKAGE}
119 version @value{VERSION}.
121 This document is distributed under the terms of the GNU Free
122 Documentation License version 1.3. A copy of the license is included
123 in the section entitled ``GNU Free Documentation License''.
126 * Overview:: Overview
127 * Invocation:: Invocation
128 * Scripts:: Linker Scripts
130 * Machine Dependent:: Machine Dependent Features
134 * H8/300:: ld and the H8/300
137 * Renesas:: ld and other Renesas micros
140 * i960:: ld and the Intel 960 family
143 * ARM:: ld and the ARM family
146 * HPPA ELF32:: ld and HPPA 32-bit ELF
149 * M68HC11/68HC12:: ld and the Motorola 68HC11 and 68HC12 families
152 * M68K:: ld and Motorola 68K family
155 * PowerPC ELF32:: ld and PowerPC 32-bit ELF Support
158 * PowerPC64 ELF64:: ld and PowerPC64 64-bit ELF Support
161 * SPU ELF:: ld and SPU ELF Support
164 * TI COFF:: ld and the TI COFF
167 * Win32:: ld and WIN32 (cygwin/mingw)
170 * Xtensa:: ld and Xtensa Processors
173 @ifclear SingleFormat
176 @c Following blank line required for remaining bug in makeinfo conds/menus
178 * Reporting Bugs:: Reporting Bugs
179 * MRI:: MRI Compatible Script Files
180 * GNU Free Documentation License:: GNU Free Documentation License
181 * LD Index:: LD Index
188 @cindex @sc{gnu} linker
189 @cindex what is this?
192 @c man begin SYNOPSIS
193 ld [@b{options}] @var{objfile} @dots{}
197 ar(1), nm(1), objcopy(1), objdump(1), readelf(1) and
198 the Info entries for @file{binutils} and
203 @c man begin DESCRIPTION
205 @command{ld} combines a number of object and archive files, relocates
206 their data and ties up symbol references. Usually the last step in
207 compiling a program is to run @command{ld}.
209 @command{ld} accepts Linker Command Language files written in
210 a superset of AT&T's Link Editor Command Language syntax,
211 to provide explicit and total control over the linking process.
215 This man page does not describe the command language; see the
216 @command{ld} entry in @code{info} for full details on the command
217 language and on other aspects of the GNU linker.
220 @ifclear SingleFormat
221 This version of @command{ld} uses the general purpose BFD libraries
222 to operate on object files. This allows @command{ld} to read, combine, and
223 write object files in many different formats---for example, COFF or
224 @code{a.out}. Different formats may be linked together to produce any
225 available kind of object file. @xref{BFD}, for more information.
228 Aside from its flexibility, the @sc{gnu} linker is more helpful than other
229 linkers in providing diagnostic information. Many linkers abandon
230 execution immediately upon encountering an error; whenever possible,
231 @command{ld} continues executing, allowing you to identify other errors
232 (or, in some cases, to get an output file in spite of the error).
239 @c man begin DESCRIPTION
241 The @sc{gnu} linker @command{ld} is meant to cover a broad range of situations,
242 and to be as compatible as possible with other linkers. As a result,
243 you have many choices to control its behavior.
249 * Options:: Command Line Options
250 * Environment:: Environment Variables
254 @section Command Line Options
262 The linker supports a plethora of command-line options, but in actual
263 practice few of them are used in any particular context.
264 @cindex standard Unix system
265 For instance, a frequent use of @command{ld} is to link standard Unix
266 object files on a standard, supported Unix system. On such a system, to
267 link a file @code{hello.o}:
270 ld -o @var{output} /lib/crt0.o hello.o -lc
273 This tells @command{ld} to produce a file called @var{output} as the
274 result of linking the file @code{/lib/crt0.o} with @code{hello.o} and
275 the library @code{libc.a}, which will come from the standard search
276 directories. (See the discussion of the @samp{-l} option below.)
278 Some of the command-line options to @command{ld} may be specified at any
279 point in the command line. However, options which refer to files, such
280 as @samp{-l} or @samp{-T}, cause the file to be read at the point at
281 which the option appears in the command line, relative to the object
282 files and other file options. Repeating non-file options with a
283 different argument will either have no further effect, or override prior
284 occurrences (those further to the left on the command line) of that
285 option. Options which may be meaningfully specified more than once are
286 noted in the descriptions below.
289 Non-option arguments are object files or archives which are to be linked
290 together. They may follow, precede, or be mixed in with command-line
291 options, except that an object file argument may not be placed between
292 an option and its argument.
294 Usually the linker is invoked with at least one object file, but you can
295 specify other forms of binary input files using @samp{-l}, @samp{-R},
296 and the script command language. If @emph{no} binary input files at all
297 are specified, the linker does not produce any output, and issues the
298 message @samp{No input files}.
300 If the linker cannot recognize the format of an object file, it will
301 assume that it is a linker script. A script specified in this way
302 augments the main linker script used for the link (either the default
303 linker script or the one specified by using @samp{-T}). This feature
304 permits the linker to link against a file which appears to be an object
305 or an archive, but actually merely defines some symbol values, or uses
306 @code{INPUT} or @code{GROUP} to load other objects. Specifying a
307 script in this way merely augments the main linker script, with the
308 extra commands placed after the main script; use the @samp{-T} option
309 to replace the default linker script entirely, but note the effect of
310 the @code{INSERT} command. @xref{Scripts}.
312 For options whose names are a single letter,
313 option arguments must either follow the option letter without intervening
314 whitespace, or be given as separate arguments immediately following the
315 option that requires them.
317 For options whose names are multiple letters, either one dash or two can
318 precede the option name; for example, @samp{-trace-symbol} and
319 @samp{--trace-symbol} are equivalent. Note---there is one exception to
320 this rule. Multiple letter options that start with a lower case 'o' can
321 only be preceded by two dashes. This is to reduce confusion with the
322 @samp{-o} option. So for example @samp{-omagic} sets the output file
323 name to @samp{magic} whereas @samp{--omagic} sets the NMAGIC flag on the
326 Arguments to multiple-letter options must either be separated from the
327 option name by an equals sign, or be given as separate arguments
328 immediately following the option that requires them. For example,
329 @samp{--trace-symbol foo} and @samp{--trace-symbol=foo} are equivalent.
330 Unique abbreviations of the names of multiple-letter options are
333 Note---if the linker is being invoked indirectly, via a compiler driver
334 (e.g. @samp{gcc}) then all the linker command line options should be
335 prefixed by @samp{-Wl,} (or whatever is appropriate for the particular
336 compiler driver) like this:
339 gcc -Wl,--start-group foo.o bar.o -Wl,--end-group
342 This is important, because otherwise the compiler driver program may
343 silently drop the linker options, resulting in a bad link. Confusion
344 may also arise when passing options that require values through a
345 driver, as the use of a space between option and argument acts as
346 a separator, and causes the driver to pass only the option to the linker
347 and the argument to the compiler. In this case, it is simplest to use
348 the joined forms of both single- and multiple-letter options, such as:
351 gcc foo.o bar.o -Wl,-eENTRY -Wl,-Map=a.map
354 Here is a table of the generic command line switches accepted by the GNU
358 @include at-file.texi
360 @kindex -a @var{keyword}
361 @item -a @var{keyword}
362 This option is supported for HP/UX compatibility. The @var{keyword}
363 argument must be one of the strings @samp{archive}, @samp{shared}, or
364 @samp{default}. @samp{-aarchive} is functionally equivalent to
365 @samp{-Bstatic}, and the other two keywords are functionally equivalent
366 to @samp{-Bdynamic}. This option may be used any number of times.
368 @kindex --audit @var{AUDITLIB}
369 @item --audit @var{AUDITLIB}
370 Adds @var{AUDITLIB} to the @code{DT_AUDIT} entry of the dynamic section.
371 @var{AUDITLIB} is not checked for existence, nor will it use the DT_SONAME
372 specified in the library. If specified multiple times @code{DT_AUDIT}
373 will contain a colon separated list of audit interfaces to use. If the linker
374 finds an object with an audit entry while searching for shared libraries,
375 it will add a corresponding @code{DT_DEPAUDIT} entry in the output file.
376 This option is only meaningful on ELF platforms supporting the rtld-audit
380 @cindex architectures
381 @kindex -A @var{arch}
382 @item -A @var{architecture}
383 @kindex --architecture=@var{arch}
384 @itemx --architecture=@var{architecture}
385 In the current release of @command{ld}, this option is useful only for the
386 Intel 960 family of architectures. In that @command{ld} configuration, the
387 @var{architecture} argument identifies the particular architecture in
388 the 960 family, enabling some safeguards and modifying the
389 archive-library search path. @xref{i960,,@command{ld} and the Intel 960
390 family}, for details.
392 Future releases of @command{ld} may support similar functionality for
393 other architecture families.
396 @ifclear SingleFormat
397 @cindex binary input format
398 @kindex -b @var{format}
399 @kindex --format=@var{format}
402 @item -b @var{input-format}
403 @itemx --format=@var{input-format}
404 @command{ld} may be configured to support more than one kind of object
405 file. If your @command{ld} is configured this way, you can use the
406 @samp{-b} option to specify the binary format for input object files
407 that follow this option on the command line. Even when @command{ld} is
408 configured to support alternative object formats, you don't usually need
409 to specify this, as @command{ld} should be configured to expect as a
410 default input format the most usual format on each machine.
411 @var{input-format} is a text string, the name of a particular format
412 supported by the BFD libraries. (You can list the available binary
413 formats with @samp{objdump -i}.)
416 You may want to use this option if you are linking files with an unusual
417 binary format. You can also use @samp{-b} to switch formats explicitly (when
418 linking object files of different formats), by including
419 @samp{-b @var{input-format}} before each group of object files in a
422 The default format is taken from the environment variable
427 You can also define the input format from a script, using the command
430 see @ref{Format Commands}.
434 @kindex -c @var{MRI-cmdfile}
435 @kindex --mri-script=@var{MRI-cmdfile}
436 @cindex compatibility, MRI
437 @item -c @var{MRI-commandfile}
438 @itemx --mri-script=@var{MRI-commandfile}
439 For compatibility with linkers produced by MRI, @command{ld} accepts script
440 files written in an alternate, restricted command language, described in
442 @ref{MRI,,MRI Compatible Script Files}.
445 the MRI Compatible Script Files section of GNU ld documentation.
447 Introduce MRI script files with
448 the option @samp{-c}; use the @samp{-T} option to run linker
449 scripts written in the general-purpose @command{ld} scripting language.
450 If @var{MRI-cmdfile} does not exist, @command{ld} looks for it in the directories
451 specified by any @samp{-L} options.
453 @cindex common allocation
460 These three options are equivalent; multiple forms are supported for
461 compatibility with other linkers. They assign space to common symbols
462 even if a relocatable output file is specified (with @samp{-r}). The
463 script command @code{FORCE_COMMON_ALLOCATION} has the same effect.
464 @xref{Miscellaneous Commands}.
466 @kindex --depaudit @var{AUDITLIB}
467 @kindex -P @var{AUDITLIB}
468 @item --depaudit @var{AUDITLIB}
469 @itemx -P @var{AUDITLIB}
470 Adds @var{AUDITLIB} to the @code{DT_DEPAUDIT} entry of the dynamic section.
471 @var{AUDITLIB} is not checked for existence, nor will it use the DT_SONAME
472 specified in the library. If specified multiple times @code{DT_DEPAUDIT}
473 will contain a colon separated list of audit interfaces to use. This
474 option is only meaningful on ELF platforms supporting the rtld-audit interface.
475 The -P option is provided for Solaris compatibility.
477 @cindex entry point, from command line
478 @kindex -e @var{entry}
479 @kindex --entry=@var{entry}
481 @itemx --entry=@var{entry}
482 Use @var{entry} as the explicit symbol for beginning execution of your
483 program, rather than the default entry point. If there is no symbol
484 named @var{entry}, the linker will try to parse @var{entry} as a number,
485 and use that as the entry address (the number will be interpreted in
486 base 10; you may use a leading @samp{0x} for base 16, or a leading
487 @samp{0} for base 8). @xref{Entry Point}, for a discussion of defaults
488 and other ways of specifying the entry point.
490 @kindex --exclude-libs
491 @item --exclude-libs @var{lib},@var{lib},...
492 Specifies a list of archive libraries from which symbols should not be automatically
493 exported. The library names may be delimited by commas or colons. Specifying
494 @code{--exclude-libs ALL} excludes symbols in all archive libraries from
495 automatic export. This option is available only for the i386 PE targeted
496 port of the linker and for ELF targeted ports. For i386 PE, symbols
497 explicitly listed in a .def file are still exported, regardless of this
498 option. For ELF targeted ports, symbols affected by this option will
499 be treated as hidden.
501 @kindex --exclude-modules-for-implib
502 @item --exclude-modules-for-implib @var{module},@var{module},...
503 Specifies a list of object files or archive members, from which symbols
504 should not be automatically exported, but which should be copied wholesale
505 into the import library being generated during the link. The module names
506 may be delimited by commas or colons, and must match exactly the filenames
507 used by @command{ld} to open the files; for archive members, this is simply
508 the member name, but for object files the name listed must include and
509 match precisely any path used to specify the input file on the linker's
510 command-line. This option is available only for the i386 PE targeted port
511 of the linker. Symbols explicitly listed in a .def file are still exported,
512 regardless of this option.
514 @cindex dynamic symbol table
516 @kindex --export-dynamic
517 @kindex --no-export-dynamic
519 @itemx --export-dynamic
520 @itemx --no-export-dynamic
521 When creating a dynamically linked executable, using the @option{-E}
522 option or the @option{--export-dynamic} option causes the linker to add
523 all symbols to the dynamic symbol table. The dynamic symbol table is the
524 set of symbols which are visible from dynamic objects at run time.
526 If you do not use either of these options (or use the
527 @option{--no-export-dynamic} option to restore the default behavior), the
528 dynamic symbol table will normally contain only those symbols which are
529 referenced by some dynamic object mentioned in the link.
531 If you use @code{dlopen} to load a dynamic object which needs to refer
532 back to the symbols defined by the program, rather than some other
533 dynamic object, then you will probably need to use this option when
534 linking the program itself.
536 You can also use the dynamic list to control what symbols should
537 be added to the dynamic symbol table if the output format supports it.
538 See the description of @samp{--dynamic-list}.
540 Note that this option is specific to ELF targeted ports. PE targets
541 support a similar function to export all symbols from a DLL or EXE; see
542 the description of @samp{--export-all-symbols} below.
544 @ifclear SingleFormat
545 @cindex big-endian objects
549 Link big-endian objects. This affects the default output format.
551 @cindex little-endian objects
554 Link little-endian objects. This affects the default output format.
557 @kindex -f @var{name}
558 @kindex --auxiliary=@var{name}
560 @itemx --auxiliary=@var{name}
561 When creating an ELF shared object, set the internal DT_AUXILIARY field
562 to the specified name. This tells the dynamic linker that the symbol
563 table of the shared object should be used as an auxiliary filter on the
564 symbol table of the shared object @var{name}.
566 If you later link a program against this filter object, then, when you
567 run the program, the dynamic linker will see the DT_AUXILIARY field. If
568 the dynamic linker resolves any symbols from the filter object, it will
569 first check whether there is a definition in the shared object
570 @var{name}. If there is one, it will be used instead of the definition
571 in the filter object. The shared object @var{name} need not exist.
572 Thus the shared object @var{name} may be used to provide an alternative
573 implementation of certain functions, perhaps for debugging or for
574 machine specific performance.
576 This option may be specified more than once. The DT_AUXILIARY entries
577 will be created in the order in which they appear on the command line.
579 @kindex -F @var{name}
580 @kindex --filter=@var{name}
582 @itemx --filter=@var{name}
583 When creating an ELF shared object, set the internal DT_FILTER field to
584 the specified name. This tells the dynamic linker that the symbol table
585 of the shared object which is being created should be used as a filter
586 on the symbol table of the shared object @var{name}.
588 If you later link a program against this filter object, then, when you
589 run the program, the dynamic linker will see the DT_FILTER field. The
590 dynamic linker will resolve symbols according to the symbol table of the
591 filter object as usual, but it will actually link to the definitions
592 found in the shared object @var{name}. Thus the filter object can be
593 used to select a subset of the symbols provided by the object
596 Some older linkers used the @option{-F} option throughout a compilation
597 toolchain for specifying object-file format for both input and output
599 @ifclear SingleFormat
600 The @sc{gnu} linker uses other mechanisms for this purpose: the
601 @option{-b}, @option{--format}, @option{--oformat} options, the
602 @code{TARGET} command in linker scripts, and the @code{GNUTARGET}
603 environment variable.
605 The @sc{gnu} linker will ignore the @option{-F} option when not
606 creating an ELF shared object.
608 @cindex finalization function
609 @kindex -fini=@var{name}
610 @item -fini=@var{name}
611 When creating an ELF executable or shared object, call NAME when the
612 executable or shared object is unloaded, by setting DT_FINI to the
613 address of the function. By default, the linker uses @code{_fini} as
614 the function to call.
618 Ignored. Provided for compatibility with other tools.
620 @kindex -G @var{value}
621 @kindex --gpsize=@var{value}
624 @itemx --gpsize=@var{value}
625 Set the maximum size of objects to be optimized using the GP register to
626 @var{size}. This is only meaningful for object file formats such as
627 MIPS ECOFF which supports putting large and small objects into different
628 sections. This is ignored for other object file formats.
630 @cindex runtime library name
631 @kindex -h @var{name}
632 @kindex -soname=@var{name}
634 @itemx -soname=@var{name}
635 When creating an ELF shared object, set the internal DT_SONAME field to
636 the specified name. When an executable is linked with a shared object
637 which has a DT_SONAME field, then when the executable is run the dynamic
638 linker will attempt to load the shared object specified by the DT_SONAME
639 field rather than the using the file name given to the linker.
642 @cindex incremental link
644 Perform an incremental link (same as option @samp{-r}).
646 @cindex initialization function
647 @kindex -init=@var{name}
648 @item -init=@var{name}
649 When creating an ELF executable or shared object, call NAME when the
650 executable or shared object is loaded, by setting DT_INIT to the address
651 of the function. By default, the linker uses @code{_init} as the
654 @cindex archive files, from cmd line
655 @kindex -l @var{namespec}
656 @kindex --library=@var{namespec}
657 @item -l @var{namespec}
658 @itemx --library=@var{namespec}
659 Add the archive or object file specified by @var{namespec} to the
660 list of files to link. This option may be used any number of times.
661 If @var{namespec} is of the form @file{:@var{filename}}, @command{ld}
662 will search the library path for a file called @var{filename}, otherwise it
663 will search the library path for a file called @file{lib@var{namespec}.a}.
665 On systems which support shared libraries, @command{ld} may also search for
666 files other than @file{lib@var{namespec}.a}. Specifically, on ELF
667 and SunOS systems, @command{ld} will search a directory for a library
668 called @file{lib@var{namespec}.so} before searching for one called
669 @file{lib@var{namespec}.a}. (By convention, a @code{.so} extension
670 indicates a shared library.) Note that this behavior does not apply
671 to @file{:@var{filename}}, which always specifies a file called
674 The linker will search an archive only once, at the location where it is
675 specified on the command line. If the archive defines a symbol which
676 was undefined in some object which appeared before the archive on the
677 command line, the linker will include the appropriate file(s) from the
678 archive. However, an undefined symbol in an object appearing later on
679 the command line will not cause the linker to search the archive again.
681 See the @option{-(} option for a way to force the linker to search
682 archives multiple times.
684 You may list the same archive multiple times on the command line.
687 This type of archive searching is standard for Unix linkers. However,
688 if you are using @command{ld} on AIX, note that it is different from the
689 behaviour of the AIX linker.
692 @cindex search directory, from cmd line
694 @kindex --library-path=@var{dir}
695 @item -L @var{searchdir}
696 @itemx --library-path=@var{searchdir}
697 Add path @var{searchdir} to the list of paths that @command{ld} will search
698 for archive libraries and @command{ld} control scripts. You may use this
699 option any number of times. The directories are searched in the order
700 in which they are specified on the command line. Directories specified
701 on the command line are searched before the default directories. All
702 @option{-L} options apply to all @option{-l} options, regardless of the
703 order in which the options appear. @option{-L} options do not affect
704 how @command{ld} searches for a linker script unless @option{-T}
707 If @var{searchdir} begins with @code{=}, then the @code{=} will be replaced
708 by the @dfn{sysroot prefix}, a path specified when the linker is configured.
711 The default set of paths searched (without being specified with
712 @samp{-L}) depends on which emulation mode @command{ld} is using, and in
713 some cases also on how it was configured. @xref{Environment}.
716 The paths can also be specified in a link script with the
717 @code{SEARCH_DIR} command. Directories specified this way are searched
718 at the point in which the linker script appears in the command line.
721 @kindex -m @var{emulation}
722 @item -m @var{emulation}
723 Emulate the @var{emulation} linker. You can list the available
724 emulations with the @samp{--verbose} or @samp{-V} options.
726 If the @samp{-m} option is not used, the emulation is taken from the
727 @code{LDEMULATION} environment variable, if that is defined.
729 Otherwise, the default emulation depends upon how the linker was
737 Print a link map to the standard output. A link map provides
738 information about the link, including the following:
742 Where object files are mapped into memory.
744 How common symbols are allocated.
746 All archive members included in the link, with a mention of the symbol
747 which caused the archive member to be brought in.
749 The values assigned to symbols.
751 Note - symbols whose values are computed by an expression which
752 involves a reference to a previous value of the same symbol may not
753 have correct result displayed in the link map. This is because the
754 linker discards intermediate results and only retains the final value
755 of an expression. Under such circumstances the linker will display
756 the final value enclosed by square brackets. Thus for example a
757 linker script containing:
765 will produce the following output in the link map if the @option{-M}
770 [0x0000000c] foo = (foo * 0x4)
771 [0x0000000c] foo = (foo + 0x8)
774 See @ref{Expressions} for more information about expressions in linker
779 @cindex read-only text
784 Turn off page alignment of sections, and disable linking against shared
785 libraries. If the output format supports Unix style magic numbers,
786 mark the output as @code{NMAGIC}.
790 @cindex read/write from cmd line
794 Set the text and data sections to be readable and writable. Also, do
795 not page-align the data segment, and disable linking against shared
796 libraries. If the output format supports Unix style magic numbers,
797 mark the output as @code{OMAGIC}. Note: Although a writable text section
798 is allowed for PE-COFF targets, it does not conform to the format
799 specification published by Microsoft.
804 This option negates most of the effects of the @option{-N} option. It
805 sets the text section to be read-only, and forces the data segment to
806 be page-aligned. Note - this option does not enable linking against
807 shared libraries. Use @option{-Bdynamic} for this.
809 @kindex -o @var{output}
810 @kindex --output=@var{output}
811 @cindex naming the output file
812 @item -o @var{output}
813 @itemx --output=@var{output}
814 Use @var{output} as the name for the program produced by @command{ld}; if this
815 option is not specified, the name @file{a.out} is used by default. The
816 script command @code{OUTPUT} can also specify the output file name.
818 @kindex -O @var{level}
819 @cindex generating optimized output
821 If @var{level} is a numeric values greater than zero @command{ld} optimizes
822 the output. This might take significantly longer and therefore probably
823 should only be enabled for the final binary. At the moment this
824 option only affects ELF shared library generation. Future releases of
825 the linker may make more use of this option. Also currently there is
826 no difference in the linker's behaviour for different non-zero values
827 of this option. Again this may change with future releases.
830 @kindex --emit-relocs
831 @cindex retain relocations in final executable
834 Leave relocation sections and contents in fully linked executables.
835 Post link analysis and optimization tools may need this information in
836 order to perform correct modifications of executables. This results
837 in larger executables.
839 This option is currently only supported on ELF platforms.
841 @kindex --force-dynamic
842 @cindex forcing the creation of dynamic sections
843 @item --force-dynamic
844 Force the output file to have dynamic sections. This option is specific
848 @cindex relocatable output
850 @kindex --relocatable
853 Generate relocatable output---i.e., generate an output file that can in
854 turn serve as input to @command{ld}. This is often called @dfn{partial
855 linking}. As a side effect, in environments that support standard Unix
856 magic numbers, this option also sets the output file's magic number to
858 @c ; see @option{-N}.
859 If this option is not specified, an absolute file is produced. When
860 linking C++ programs, this option @emph{will not} resolve references to
861 constructors; to do that, use @samp{-Ur}.
863 When an input file does not have the same format as the output file,
864 partial linking is only supported if that input file does not contain any
865 relocations. Different output formats can have further restrictions; for
866 example some @code{a.out}-based formats do not support partial linking
867 with input files in other formats at all.
869 This option does the same thing as @samp{-i}.
871 @kindex -R @var{file}
872 @kindex --just-symbols=@var{file}
873 @cindex symbol-only input
874 @item -R @var{filename}
875 @itemx --just-symbols=@var{filename}
876 Read symbol names and their addresses from @var{filename}, but do not
877 relocate it or include it in the output. This allows your output file
878 to refer symbolically to absolute locations of memory defined in other
879 programs. You may use this option more than once.
881 For compatibility with other ELF linkers, if the @option{-R} option is
882 followed by a directory name, rather than a file name, it is treated as
883 the @option{-rpath} option.
887 @cindex strip all symbols
890 Omit all symbol information from the output file.
893 @kindex --strip-debug
894 @cindex strip debugger symbols
897 Omit debugger symbol information (but not all symbols) from the output file.
901 @cindex input files, displaying
904 Print the names of the input files as @command{ld} processes them.
906 @kindex -T @var{script}
907 @kindex --script=@var{script}
909 @item -T @var{scriptfile}
910 @itemx --script=@var{scriptfile}
911 Use @var{scriptfile} as the linker script. This script replaces
912 @command{ld}'s default linker script (rather than adding to it), so
913 @var{commandfile} must specify everything necessary to describe the
914 output file. @xref{Scripts}. If @var{scriptfile} does not exist in
915 the current directory, @code{ld} looks for it in the directories
916 specified by any preceding @samp{-L} options. Multiple @samp{-T}
919 @kindex -dT @var{script}
920 @kindex --default-script=@var{script}
922 @item -dT @var{scriptfile}
923 @itemx --default-script=@var{scriptfile}
924 Use @var{scriptfile} as the default linker script. @xref{Scripts}.
926 This option is similar to the @option{--script} option except that
927 processing of the script is delayed until after the rest of the
928 command line has been processed. This allows options placed after the
929 @option{--default-script} option on the command line to affect the
930 behaviour of the linker script, which can be important when the linker
931 command line cannot be directly controlled by the user. (eg because
932 the command line is being constructed by another tool, such as
935 @kindex -u @var{symbol}
936 @kindex --undefined=@var{symbol}
937 @cindex undefined symbol
938 @item -u @var{symbol}
939 @itemx --undefined=@var{symbol}
940 Force @var{symbol} to be entered in the output file as an undefined
941 symbol. Doing this may, for example, trigger linking of additional
942 modules from standard libraries. @samp{-u} may be repeated with
943 different option arguments to enter additional undefined symbols. This
944 option is equivalent to the @code{EXTERN} linker script command.
949 For anything other than C++ programs, this option is equivalent to
950 @samp{-r}: it generates relocatable output---i.e., an output file that can in
951 turn serve as input to @command{ld}. When linking C++ programs, @samp{-Ur}
952 @emph{does} resolve references to constructors, unlike @samp{-r}.
953 It does not work to use @samp{-Ur} on files that were themselves linked
954 with @samp{-Ur}; once the constructor table has been built, it cannot
955 be added to. Use @samp{-Ur} only for the last partial link, and
956 @samp{-r} for the others.
958 @kindex --unique[=@var{SECTION}]
959 @item --unique[=@var{SECTION}]
960 Creates a separate output section for every input section matching
961 @var{SECTION}, or if the optional wildcard @var{SECTION} argument is
962 missing, for every orphan input section. An orphan section is one not
963 specifically mentioned in a linker script. You may use this option
964 multiple times on the command line; It prevents the normal merging of
965 input sections with the same name, overriding output section assignments
975 Display the version number for @command{ld}. The @option{-V} option also
976 lists the supported emulations.
979 @kindex --discard-all
980 @cindex deleting local symbols
983 Delete all local symbols.
986 @kindex --discard-locals
987 @cindex local symbols, deleting
989 @itemx --discard-locals
990 Delete all temporary local symbols. (These symbols start with
991 system-specific local label prefixes, typically @samp{.L} for ELF systems
992 or @samp{L} for traditional a.out systems.)
994 @kindex -y @var{symbol}
995 @kindex --trace-symbol=@var{symbol}
996 @cindex symbol tracing
997 @item -y @var{symbol}
998 @itemx --trace-symbol=@var{symbol}
999 Print the name of each linked file in which @var{symbol} appears. This
1000 option may be given any number of times. On many systems it is necessary
1001 to prepend an underscore.
1003 This option is useful when you have an undefined symbol in your link but
1004 don't know where the reference is coming from.
1006 @kindex -Y @var{path}
1008 Add @var{path} to the default library search path. This option exists
1009 for Solaris compatibility.
1011 @kindex -z @var{keyword}
1012 @item -z @var{keyword}
1013 The recognized keywords are:
1017 Combines multiple reloc sections and sorts them to make dynamic symbol
1018 lookup caching possible.
1021 Disallows undefined symbols in object files. Undefined symbols in
1022 shared libraries are still allowed.
1025 Marks the object as requiring executable stack.
1028 This option is only meaningful when building a shared object.
1029 It marks the object so that its runtime initialization will occur
1030 before the runtime initialization of any other objects brought into
1031 the process at the same time. Similarly the runtime finalization of
1032 the object will occur after the runtime finalization of any other
1036 Marks the object that its symbol table interposes before all symbols
1037 but the primary executable.
1040 When generating an executable or shared library, mark it to tell the
1041 dynamic linker to defer function call resolution to the point when
1042 the function is called (lazy binding), rather than at load time.
1043 Lazy binding is the default.
1046 Marks the object that its filters be processed immediately at
1050 Allows multiple definitions.
1053 Disables multiple reloc sections combining.
1056 Disables production of copy relocs.
1059 Marks the object that the search for dependencies of this object will
1060 ignore any default library search paths.
1063 Marks the object shouldn't be unloaded at runtime.
1066 Marks the object not available to @code{dlopen}.
1069 Marks the object can not be dumped by @code{dldump}.
1072 Marks the object as not requiring executable stack.
1075 Don't create an ELF @code{PT_GNU_RELRO} segment header in the object.
1078 When generating an executable or shared library, mark it to tell the
1079 dynamic linker to resolve all symbols when the program is started, or
1080 when the shared library is linked to using dlopen, instead of
1081 deferring function call resolution to the point when the function is
1085 Marks the object may contain $ORIGIN.
1088 Create an ELF @code{PT_GNU_RELRO} segment header in the object.
1090 @item max-page-size=@var{value}
1091 Set the emulation maximum page size to @var{value}.
1093 @item common-page-size=@var{value}
1094 Set the emulation common page size to @var{value}.
1098 Other keywords are ignored for Solaris compatibility.
1101 @cindex groups of archives
1102 @item -( @var{archives} -)
1103 @itemx --start-group @var{archives} --end-group
1104 The @var{archives} should be a list of archive files. They may be
1105 either explicit file names, or @samp{-l} options.
1107 The specified archives are searched repeatedly until no new undefined
1108 references are created. Normally, an archive is searched only once in
1109 the order that it is specified on the command line. If a symbol in that
1110 archive is needed to resolve an undefined symbol referred to by an
1111 object in an archive that appears later on the command line, the linker
1112 would not be able to resolve that reference. By grouping the archives,
1113 they all be searched repeatedly until all possible references are
1116 Using this option has a significant performance cost. It is best to use
1117 it only when there are unavoidable circular references between two or
1120 @kindex --accept-unknown-input-arch
1121 @kindex --no-accept-unknown-input-arch
1122 @item --accept-unknown-input-arch
1123 @itemx --no-accept-unknown-input-arch
1124 Tells the linker to accept input files whose architecture cannot be
1125 recognised. The assumption is that the user knows what they are doing
1126 and deliberately wants to link in these unknown input files. This was
1127 the default behaviour of the linker, before release 2.14. The default
1128 behaviour from release 2.14 onwards is to reject such input files, and
1129 so the @samp{--accept-unknown-input-arch} option has been added to
1130 restore the old behaviour.
1133 @kindex --no-as-needed
1135 @itemx --no-as-needed
1136 This option affects ELF DT_NEEDED tags for dynamic libraries mentioned
1137 on the command line after the @option{--as-needed} option. Normally
1138 the linker will add a DT_NEEDED tag for each dynamic library mentioned
1139 on the command line, regardless of whether the library is actually
1140 needed or not. @option{--as-needed} causes a DT_NEEDED tag to only be
1141 emitted for a library that satisfies an undefined symbol reference
1142 from a regular object file or, if the library is not found in the
1143 DT_NEEDED lists of other libraries linked up to that point, an
1144 undefined symbol reference from another dynamic library.
1145 @option{--no-as-needed} restores the default behaviour.
1147 @kindex --add-needed
1148 @kindex --no-add-needed
1150 @itemx --no-add-needed
1151 These two options have been deprecated because of the similarity of
1152 their names to the @option{--as-needed} and @option{--no-as-needed}
1153 options. They have been replaced by @option{--copy-dt-needed-entries}
1154 and @option{--no-copy-dt-needed-entries}.
1156 @kindex -assert @var{keyword}
1157 @item -assert @var{keyword}
1158 This option is ignored for SunOS compatibility.
1162 @kindex -call_shared
1166 Link against dynamic libraries. This is only meaningful on platforms
1167 for which shared libraries are supported. This option is normally the
1168 default on such platforms. The different variants of this option are
1169 for compatibility with various systems. You may use this option
1170 multiple times on the command line: it affects library searching for
1171 @option{-l} options which follow it.
1175 Set the @code{DF_1_GROUP} flag in the @code{DT_FLAGS_1} entry in the dynamic
1176 section. This causes the runtime linker to handle lookups in this
1177 object and its dependencies to be performed only inside the group.
1178 @option{--unresolved-symbols=report-all} is implied. This option is
1179 only meaningful on ELF platforms which support shared libraries.
1189 Do not link against shared libraries. This is only meaningful on
1190 platforms for which shared libraries are supported. The different
1191 variants of this option are for compatibility with various systems. You
1192 may use this option multiple times on the command line: it affects
1193 library searching for @option{-l} options which follow it. This
1194 option also implies @option{--unresolved-symbols=report-all}. This
1195 option can be used with @option{-shared}. Doing so means that a
1196 shared library is being created but that all of the library's external
1197 references must be resolved by pulling in entries from static
1202 When creating a shared library, bind references to global symbols to the
1203 definition within the shared library, if any. Normally, it is possible
1204 for a program linked against a shared library to override the definition
1205 within the shared library. This option is only meaningful on ELF
1206 platforms which support shared libraries.
1208 @kindex -Bsymbolic-functions
1209 @item -Bsymbolic-functions
1210 When creating a shared library, bind references to global function
1211 symbols to the definition within the shared library, if any.
1212 This option is only meaningful on ELF platforms which support shared
1215 @kindex --dynamic-list=@var{dynamic-list-file}
1216 @item --dynamic-list=@var{dynamic-list-file}
1217 Specify the name of a dynamic list file to the linker. This is
1218 typically used when creating shared libraries to specify a list of
1219 global symbols whose references shouldn't be bound to the definition
1220 within the shared library, or creating dynamically linked executables
1221 to specify a list of symbols which should be added to the symbol table
1222 in the executable. This option is only meaningful on ELF platforms
1223 which support shared libraries.
1225 The format of the dynamic list is the same as the version node without
1226 scope and node name. See @ref{VERSION} for more information.
1228 @kindex --dynamic-list-data
1229 @item --dynamic-list-data
1230 Include all global data symbols to the dynamic list.
1232 @kindex --dynamic-list-cpp-new
1233 @item --dynamic-list-cpp-new
1234 Provide the builtin dynamic list for C++ operator new and delete. It
1235 is mainly useful for building shared libstdc++.
1237 @kindex --dynamic-list-cpp-typeinfo
1238 @item --dynamic-list-cpp-typeinfo
1239 Provide the builtin dynamic list for C++ runtime type identification.
1241 @kindex --check-sections
1242 @kindex --no-check-sections
1243 @item --check-sections
1244 @itemx --no-check-sections
1245 Asks the linker @emph{not} to check section addresses after they have
1246 been assigned to see if there are any overlaps. Normally the linker will
1247 perform this check, and if it finds any overlaps it will produce
1248 suitable error messages. The linker does know about, and does make
1249 allowances for sections in overlays. The default behaviour can be
1250 restored by using the command line switch @option{--check-sections}.
1251 Section overlap is not usually checked for relocatable links. You can
1252 force checking in that case by using the @option{--check-sections}
1255 @kindex --copy-dt-needed-entries
1256 @kindex --no-copy-dt-needed-entries
1257 @item --copy-dt-needed-entries
1258 @itemx --no-copy-dt-needed-entries
1259 This option affects the treatment of dynamic libraries referred to
1260 by DT_NEEDED tags @emph{inside} ELF dynamic libraries mentioned on the
1261 command line. Normally the linker will add a DT_NEEDED tag to the
1262 output binary for each library mentioned in a DT_NEEDED tag in an
1263 input dynamic library. With @option{--no-copy-dt-needed-entries}
1264 specified on the command line however any dynamic libraries that
1265 follow it will have their DT_NEEDED entries ignored. The default
1266 behaviour can be restored with @option{--copy-dt-needed-entries}.
1268 This option also has an effect on the resolution of symbols in dynamic
1269 libraries. With the default setting dynamic libraries mentioned on
1270 the command line will be recursively searched, following their
1271 DT_NEEDED tags to other libraries, in order to resolve symbols
1272 required by the output binary. With
1273 @option{--no-copy-dt-needed-entries} specified however the searching
1274 of dynamic libraries that follow it will stop with the dynamic
1275 library itself. No DT_NEEDED links will be traversed to resolve
1278 @cindex cross reference table
1281 Output a cross reference table. If a linker map file is being
1282 generated, the cross reference table is printed to the map file.
1283 Otherwise, it is printed on the standard output.
1285 The format of the table is intentionally simple, so that it may be
1286 easily processed by a script if necessary. The symbols are printed out,
1287 sorted by name. For each symbol, a list of file names is given. If the
1288 symbol is defined, the first file listed is the location of the
1289 definition. The remaining files contain references to the symbol.
1291 @cindex common allocation
1292 @kindex --no-define-common
1293 @item --no-define-common
1294 This option inhibits the assignment of addresses to common symbols.
1295 The script command @code{INHIBIT_COMMON_ALLOCATION} has the same effect.
1296 @xref{Miscellaneous Commands}.
1298 The @samp{--no-define-common} option allows decoupling
1299 the decision to assign addresses to Common symbols from the choice
1300 of the output file type; otherwise a non-Relocatable output type
1301 forces assigning addresses to Common symbols.
1302 Using @samp{--no-define-common} allows Common symbols that are referenced
1303 from a shared library to be assigned addresses only in the main program.
1304 This eliminates the unused duplicate space in the shared library,
1305 and also prevents any possible confusion over resolving to the wrong
1306 duplicate when there are many dynamic modules with specialized search
1307 paths for runtime symbol resolution.
1309 @cindex symbols, from command line
1310 @kindex --defsym=@var{symbol}=@var{exp}
1311 @item --defsym=@var{symbol}=@var{expression}
1312 Create a global symbol in the output file, containing the absolute
1313 address given by @var{expression}. You may use this option as many
1314 times as necessary to define multiple symbols in the command line. A
1315 limited form of arithmetic is supported for the @var{expression} in this
1316 context: you may give a hexadecimal constant or the name of an existing
1317 symbol, or use @code{+} and @code{-} to add or subtract hexadecimal
1318 constants or symbols. If you need more elaborate expressions, consider
1319 using the linker command language from a script (@pxref{Assignments,,
1320 Assignment: Symbol Definitions}). @emph{Note:} there should be no white
1321 space between @var{symbol}, the equals sign (``@key{=}''), and
1324 @cindex demangling, from command line
1325 @kindex --demangle[=@var{style}]
1326 @kindex --no-demangle
1327 @item --demangle[=@var{style}]
1328 @itemx --no-demangle
1329 These options control whether to demangle symbol names in error messages
1330 and other output. When the linker is told to demangle, it tries to
1331 present symbol names in a readable fashion: it strips leading
1332 underscores if they are used by the object file format, and converts C++
1333 mangled symbol names into user readable names. Different compilers have
1334 different mangling styles. The optional demangling style argument can be used
1335 to choose an appropriate demangling style for your compiler. The linker will
1336 demangle by default unless the environment variable @samp{COLLECT_NO_DEMANGLE}
1337 is set. These options may be used to override the default.
1339 @cindex dynamic linker, from command line
1340 @kindex -I@var{file}
1341 @kindex --dynamic-linker=@var{file}
1343 @itemx --dynamic-linker=@var{file}
1344 Set the name of the dynamic linker. This is only meaningful when
1345 generating dynamically linked ELF executables. The default dynamic
1346 linker is normally correct; don't use this unless you know what you are
1349 @kindex --fatal-warnings
1350 @kindex --no-fatal-warnings
1351 @item --fatal-warnings
1352 @itemx --no-fatal-warnings
1353 Treat all warnings as errors. The default behaviour can be restored
1354 with the option @option{--no-fatal-warnings}.
1356 @kindex --force-exe-suffix
1357 @item --force-exe-suffix
1358 Make sure that an output file has a .exe suffix.
1360 If a successfully built fully linked output file does not have a
1361 @code{.exe} or @code{.dll} suffix, this option forces the linker to copy
1362 the output file to one of the same name with a @code{.exe} suffix. This
1363 option is useful when using unmodified Unix makefiles on a Microsoft
1364 Windows host, since some versions of Windows won't run an image unless
1365 it ends in a @code{.exe} suffix.
1367 @kindex --gc-sections
1368 @kindex --no-gc-sections
1369 @cindex garbage collection
1371 @itemx --no-gc-sections
1372 Enable garbage collection of unused input sections. It is ignored on
1373 targets that do not support this option. The default behaviour (of not
1374 performing this garbage collection) can be restored by specifying
1375 @samp{--no-gc-sections} on the command line.
1377 @samp{--gc-sections} decides which input sections are used by
1378 examining symbols and relocations. The section containing the entry
1379 symbol and all sections containing symbols undefined on the
1380 command-line will be kept, as will sections containing symbols
1381 referenced by dynamic objects. Note that when building shared
1382 libraries, the linker must assume that any visible symbol is
1383 referenced. Once this initial set of sections has been determined,
1384 the linker recursively marks as used any section referenced by their
1385 relocations. See @samp{--entry} and @samp{--undefined}.
1387 This option can be set when doing a partial link (enabled with option
1388 @samp{-r}). In this case the root of symbols kept must be explicitly
1389 specified either by an @samp{--entry} or @samp{--undefined} option or by
1390 a @code{ENTRY} command in the linker script.
1392 @kindex --print-gc-sections
1393 @kindex --no-print-gc-sections
1394 @cindex garbage collection
1395 @item --print-gc-sections
1396 @itemx --no-print-gc-sections
1397 List all sections removed by garbage collection. The listing is
1398 printed on stderr. This option is only effective if garbage
1399 collection has been enabled via the @samp{--gc-sections}) option. The
1400 default behaviour (of not listing the sections that are removed) can
1401 be restored by specifying @samp{--no-print-gc-sections} on the command
1408 Print a summary of the command-line options on the standard output and exit.
1410 @kindex --target-help
1412 Print a summary of all target specific options on the standard output and exit.
1414 @kindex -Map=@var{mapfile}
1415 @item -Map=@var{mapfile}
1416 Print a link map to the file @var{mapfile}. See the description of the
1417 @option{-M} option, above.
1419 @cindex memory usage
1420 @kindex --no-keep-memory
1421 @item --no-keep-memory
1422 @command{ld} normally optimizes for speed over memory usage by caching the
1423 symbol tables of input files in memory. This option tells @command{ld} to
1424 instead optimize for memory usage, by rereading the symbol tables as
1425 necessary. This may be required if @command{ld} runs out of memory space
1426 while linking a large executable.
1428 @kindex --no-undefined
1430 @item --no-undefined
1432 Report unresolved symbol references from regular object files. This
1433 is done even if the linker is creating a non-symbolic shared library.
1434 The switch @option{--[no-]allow-shlib-undefined} controls the
1435 behaviour for reporting unresolved references found in shared
1436 libraries being linked in.
1438 @kindex --allow-multiple-definition
1440 @item --allow-multiple-definition
1442 Normally when a symbol is defined multiple times, the linker will
1443 report a fatal error. These options allow multiple definitions and the
1444 first definition will be used.
1446 @kindex --allow-shlib-undefined
1447 @kindex --no-allow-shlib-undefined
1448 @item --allow-shlib-undefined
1449 @itemx --no-allow-shlib-undefined
1450 Allows or disallows undefined symbols in shared libraries.
1451 This switch is similar to @option{--no-undefined} except that it
1452 determines the behaviour when the undefined symbols are in a
1453 shared library rather than a regular object file. It does not affect
1454 how undefined symbols in regular object files are handled.
1456 The default behaviour is to report errors for any undefined symbols
1457 referenced in shared libraries if the linker is being used to create
1458 an executable, but to allow them if the linker is being used to create
1461 The reasons for allowing undefined symbol references in shared
1462 libraries specified at link time are that:
1466 A shared library specified at link time may not be the same as the one
1467 that is available at load time, so the symbol might actually be
1468 resolvable at load time.
1470 There are some operating systems, eg BeOS and HPPA, where undefined
1471 symbols in shared libraries are normal.
1473 The BeOS kernel for example patches shared libraries at load time to
1474 select whichever function is most appropriate for the current
1475 architecture. This is used, for example, to dynamically select an
1476 appropriate memset function.
1479 @kindex --no-undefined-version
1480 @item --no-undefined-version
1481 Normally when a symbol has an undefined version, the linker will ignore
1482 it. This option disallows symbols with undefined version and a fatal error
1483 will be issued instead.
1485 @kindex --default-symver
1486 @item --default-symver
1487 Create and use a default symbol version (the soname) for unversioned
1490 @kindex --default-imported-symver
1491 @item --default-imported-symver
1492 Create and use a default symbol version (the soname) for unversioned
1495 @kindex --no-warn-mismatch
1496 @item --no-warn-mismatch
1497 Normally @command{ld} will give an error if you try to link together input
1498 files that are mismatched for some reason, perhaps because they have
1499 been compiled for different processors or for different endiannesses.
1500 This option tells @command{ld} that it should silently permit such possible
1501 errors. This option should only be used with care, in cases when you
1502 have taken some special action that ensures that the linker errors are
1505 @kindex --no-warn-search-mismatch
1506 @item --no-warn-search-mismatch
1507 Normally @command{ld} will give a warning if it finds an incompatible
1508 library during a library search. This option silences the warning.
1510 @kindex --no-whole-archive
1511 @item --no-whole-archive
1512 Turn off the effect of the @option{--whole-archive} option for subsequent
1515 @cindex output file after errors
1516 @kindex --noinhibit-exec
1517 @item --noinhibit-exec
1518 Retain the executable output file whenever it is still usable.
1519 Normally, the linker will not produce an output file if it encounters
1520 errors during the link process; it exits without writing an output file
1521 when it issues any error whatsoever.
1525 Only search library directories explicitly specified on the
1526 command line. Library directories specified in linker scripts
1527 (including linker scripts specified on the command line) are ignored.
1529 @ifclear SingleFormat
1530 @kindex --oformat=@var{output-format}
1531 @item --oformat=@var{output-format}
1532 @command{ld} may be configured to support more than one kind of object
1533 file. If your @command{ld} is configured this way, you can use the
1534 @samp{--oformat} option to specify the binary format for the output
1535 object file. Even when @command{ld} is configured to support alternative
1536 object formats, you don't usually need to specify this, as @command{ld}
1537 should be configured to produce as a default output format the most
1538 usual format on each machine. @var{output-format} is a text string, the
1539 name of a particular format supported by the BFD libraries. (You can
1540 list the available binary formats with @samp{objdump -i}.) The script
1541 command @code{OUTPUT_FORMAT} can also specify the output format, but
1542 this option overrides it. @xref{BFD}.
1546 @kindex --pic-executable
1548 @itemx --pic-executable
1549 @cindex position independent executables
1550 Create a position independent executable. This is currently only supported on
1551 ELF platforms. Position independent executables are similar to shared
1552 libraries in that they are relocated by the dynamic linker to the virtual
1553 address the OS chooses for them (which can vary between invocations). Like
1554 normal dynamically linked executables they can be executed and symbols
1555 defined in the executable cannot be overridden by shared libraries.
1559 This option is ignored for Linux compatibility.
1563 This option is ignored for SVR4 compatibility.
1566 @cindex synthesizing linker
1567 @cindex relaxing addressing modes
1571 An option with machine dependent effects.
1573 This option is only supported on a few targets.
1576 @xref{H8/300,,@command{ld} and the H8/300}.
1579 @xref{i960,, @command{ld} and the Intel 960 family}.
1582 @xref{Xtensa,, @command{ld} and Xtensa Processors}.
1585 @xref{M68HC11/68HC12,,@command{ld} and the 68HC11 and 68HC12}.
1588 @xref{PowerPC ELF32,,@command{ld} and PowerPC 32-bit ELF Support}.
1591 On some platforms the @samp{--relax} option performs target specific,
1592 global optimizations that become possible when the linker resolves
1593 addressing in the program, such as relaxing address modes,
1594 synthesizing new instructions, selecting shorter version of current
1595 instructions, and combinig constant values.
1597 On some platforms these link time global optimizations may make symbolic
1598 debugging of the resulting executable impossible.
1600 This is known to be the case for the Matsushita MN10200 and MN10300
1601 family of processors.
1605 On platforms where this is not supported, @samp{--relax} is accepted,
1609 On platforms where @samp{--relax} is accepted the option
1610 @samp{--no-relax} can be used to disable the feature.
1612 @cindex retaining specified symbols
1613 @cindex stripping all but some symbols
1614 @cindex symbols, retaining selectively
1615 @kindex --retain-symbols-file=@var{filename}
1616 @item --retain-symbols-file=@var{filename}
1617 Retain @emph{only} the symbols listed in the file @var{filename},
1618 discarding all others. @var{filename} is simply a flat file, with one
1619 symbol name per line. This option is especially useful in environments
1623 where a large global symbol table is accumulated gradually, to conserve
1626 @samp{--retain-symbols-file} does @emph{not} discard undefined symbols,
1627 or symbols needed for relocations.
1629 You may only specify @samp{--retain-symbols-file} once in the command
1630 line. It overrides @samp{-s} and @samp{-S}.
1633 @item -rpath=@var{dir}
1634 @cindex runtime library search path
1635 @kindex -rpath=@var{dir}
1636 Add a directory to the runtime library search path. This is used when
1637 linking an ELF executable with shared objects. All @option{-rpath}
1638 arguments are concatenated and passed to the runtime linker, which uses
1639 them to locate shared objects at runtime. The @option{-rpath} option is
1640 also used when locating shared objects which are needed by shared
1641 objects explicitly included in the link; see the description of the
1642 @option{-rpath-link} option. If @option{-rpath} is not used when linking an
1643 ELF executable, the contents of the environment variable
1644 @code{LD_RUN_PATH} will be used if it is defined.
1646 The @option{-rpath} option may also be used on SunOS. By default, on
1647 SunOS, the linker will form a runtime search patch out of all the
1648 @option{-L} options it is given. If a @option{-rpath} option is used, the
1649 runtime search path will be formed exclusively using the @option{-rpath}
1650 options, ignoring the @option{-L} options. This can be useful when using
1651 gcc, which adds many @option{-L} options which may be on NFS mounted
1654 For compatibility with other ELF linkers, if the @option{-R} option is
1655 followed by a directory name, rather than a file name, it is treated as
1656 the @option{-rpath} option.
1660 @cindex link-time runtime library search path
1661 @kindex -rpath-link=@var{dir}
1662 @item -rpath-link=@var{dir}
1663 When using ELF or SunOS, one shared library may require another. This
1664 happens when an @code{ld -shared} link includes a shared library as one
1667 When the linker encounters such a dependency when doing a non-shared,
1668 non-relocatable link, it will automatically try to locate the required
1669 shared library and include it in the link, if it is not included
1670 explicitly. In such a case, the @option{-rpath-link} option
1671 specifies the first set of directories to search. The
1672 @option{-rpath-link} option may specify a sequence of directory names
1673 either by specifying a list of names separated by colons, or by
1674 appearing multiple times.
1676 This option should be used with caution as it overrides the search path
1677 that may have been hard compiled into a shared library. In such a case it
1678 is possible to use unintentionally a different search path than the
1679 runtime linker would do.
1681 The linker uses the following search paths to locate required shared
1685 Any directories specified by @option{-rpath-link} options.
1687 Any directories specified by @option{-rpath} options. The difference
1688 between @option{-rpath} and @option{-rpath-link} is that directories
1689 specified by @option{-rpath} options are included in the executable and
1690 used at runtime, whereas the @option{-rpath-link} option is only effective
1691 at link time. Searching @option{-rpath} in this way is only supported
1692 by native linkers and cross linkers which have been configured with
1693 the @option{--with-sysroot} option.
1695 On an ELF system, for native linkers, if the @option{-rpath} and
1696 @option{-rpath-link} options were not used, search the contents of the
1697 environment variable @code{LD_RUN_PATH}.
1699 On SunOS, if the @option{-rpath} option was not used, search any
1700 directories specified using @option{-L} options.
1702 For a native linker, the search the contents of the environment
1703 variable @code{LD_LIBRARY_PATH}.
1705 For a native ELF linker, the directories in @code{DT_RUNPATH} or
1706 @code{DT_RPATH} of a shared library are searched for shared
1707 libraries needed by it. The @code{DT_RPATH} entries are ignored if
1708 @code{DT_RUNPATH} entries exist.
1710 The default directories, normally @file{/lib} and @file{/usr/lib}.
1712 For a native linker on an ELF system, if the file @file{/etc/ld.so.conf}
1713 exists, the list of directories found in that file.
1716 If the required shared library is not found, the linker will issue a
1717 warning and continue with the link.
1724 @cindex shared libraries
1725 Create a shared library. This is currently only supported on ELF, XCOFF
1726 and SunOS platforms. On SunOS, the linker will automatically create a
1727 shared library if the @option{-e} option is not used and there are
1728 undefined symbols in the link.
1730 @kindex --sort-common
1732 @itemx --sort-common=ascending
1733 @itemx --sort-common=descending
1734 This option tells @command{ld} to sort the common symbols by alignment in
1735 ascending or descending order when it places them in the appropriate output
1736 sections. The symbol alignments considered are sixteen-byte or larger,
1737 eight-byte, four-byte, two-byte, and one-byte. This is to prevent gaps
1738 between symbols due to alignment constraints. If no sorting order is
1739 specified, then descending order is assumed.
1741 @kindex --sort-section=name
1742 @item --sort-section=name
1743 This option will apply @code{SORT_BY_NAME} to all wildcard section
1744 patterns in the linker script.
1746 @kindex --sort-section=alignment
1747 @item --sort-section=alignment
1748 This option will apply @code{SORT_BY_ALIGNMENT} to all wildcard section
1749 patterns in the linker script.
1751 @kindex --split-by-file
1752 @item --split-by-file[=@var{size}]
1753 Similar to @option{--split-by-reloc} but creates a new output section for
1754 each input file when @var{size} is reached. @var{size} defaults to a
1755 size of 1 if not given.
1757 @kindex --split-by-reloc
1758 @item --split-by-reloc[=@var{count}]
1759 Tries to creates extra sections in the output file so that no single
1760 output section in the file contains more than @var{count} relocations.
1761 This is useful when generating huge relocatable files for downloading into
1762 certain real time kernels with the COFF object file format; since COFF
1763 cannot represent more than 65535 relocations in a single section. Note
1764 that this will fail to work with object file formats which do not
1765 support arbitrary sections. The linker will not split up individual
1766 input sections for redistribution, so if a single input section contains
1767 more than @var{count} relocations one output section will contain that
1768 many relocations. @var{count} defaults to a value of 32768.
1772 Compute and display statistics about the operation of the linker, such
1773 as execution time and memory usage.
1775 @kindex --sysroot=@var{directory}
1776 @item --sysroot=@var{directory}
1777 Use @var{directory} as the location of the sysroot, overriding the
1778 configure-time default. This option is only supported by linkers
1779 that were configured using @option{--with-sysroot}.
1781 @kindex --traditional-format
1782 @cindex traditional format
1783 @item --traditional-format
1784 For some targets, the output of @command{ld} is different in some ways from
1785 the output of some existing linker. This switch requests @command{ld} to
1786 use the traditional format instead.
1789 For example, on SunOS, @command{ld} combines duplicate entries in the
1790 symbol string table. This can reduce the size of an output file with
1791 full debugging information by over 30 percent. Unfortunately, the SunOS
1792 @code{dbx} program can not read the resulting program (@code{gdb} has no
1793 trouble). The @samp{--traditional-format} switch tells @command{ld} to not
1794 combine duplicate entries.
1796 @kindex --section-start=@var{sectionname}=@var{org}
1797 @item --section-start=@var{sectionname}=@var{org}
1798 Locate a section in the output file at the absolute
1799 address given by @var{org}. You may use this option as many
1800 times as necessary to locate multiple sections in the command
1802 @var{org} must be a single hexadecimal integer;
1803 for compatibility with other linkers, you may omit the leading
1804 @samp{0x} usually associated with hexadecimal values. @emph{Note:} there
1805 should be no white space between @var{sectionname}, the equals
1806 sign (``@key{=}''), and @var{org}.
1808 @kindex -Tbss=@var{org}
1809 @kindex -Tdata=@var{org}
1810 @kindex -Ttext=@var{org}
1811 @cindex segment origins, cmd line
1812 @item -Tbss=@var{org}
1813 @itemx -Tdata=@var{org}
1814 @itemx -Ttext=@var{org}
1815 Same as @option{--section-start}, with @code{.bss}, @code{.data} or
1816 @code{.text} as the @var{sectionname}.
1818 @kindex -Ttext-segment=@var{org}
1819 @item -Ttext-segment=@var{org}
1820 @cindex text segment origin, cmd line
1821 When creating an ELF executable or shared object, it will set the address
1822 of the first byte of the text segment.
1824 @kindex --unresolved-symbols
1825 @item --unresolved-symbols=@var{method}
1826 Determine how to handle unresolved symbols. There are four possible
1827 values for @samp{method}:
1831 Do not report any unresolved symbols.
1834 Report all unresolved symbols. This is the default.
1836 @item ignore-in-object-files
1837 Report unresolved symbols that are contained in shared libraries, but
1838 ignore them if they come from regular object files.
1840 @item ignore-in-shared-libs
1841 Report unresolved symbols that come from regular object files, but
1842 ignore them if they come from shared libraries. This can be useful
1843 when creating a dynamic binary and it is known that all the shared
1844 libraries that it should be referencing are included on the linker's
1848 The behaviour for shared libraries on their own can also be controlled
1849 by the @option{--[no-]allow-shlib-undefined} option.
1851 Normally the linker will generate an error message for each reported
1852 unresolved symbol but the option @option{--warn-unresolved-symbols}
1853 can change this to a warning.
1855 @kindex --verbose[=@var{NUMBER}]
1856 @cindex verbose[=@var{NUMBER}]
1858 @itemx --verbose[=@var{NUMBER}]
1859 Display the version number for @command{ld} and list the linker emulations
1860 supported. Display which input files can and cannot be opened. Display
1861 the linker script being used by the linker. If the optional @var{NUMBER}
1862 argument > 1, plugin symbol status will also be displayed.
1864 @kindex --version-script=@var{version-scriptfile}
1865 @cindex version script, symbol versions
1866 @item --version-script=@var{version-scriptfile}
1867 Specify the name of a version script to the linker. This is typically
1868 used when creating shared libraries to specify additional information
1869 about the version hierarchy for the library being created. This option
1870 is only fully supported on ELF platforms which support shared libraries;
1871 see @ref{VERSION}. It is partially supported on PE platforms, which can
1872 use version scripts to filter symbol visibility in auto-export mode: any
1873 symbols marked @samp{local} in the version script will not be exported.
1876 @kindex --warn-common
1877 @cindex warnings, on combining symbols
1878 @cindex combining symbols, warnings on
1880 Warn when a common symbol is combined with another common symbol or with
1881 a symbol definition. Unix linkers allow this somewhat sloppy practise,
1882 but linkers on some other operating systems do not. This option allows
1883 you to find potential problems from combining global symbols.
1884 Unfortunately, some C libraries use this practise, so you may get some
1885 warnings about symbols in the libraries as well as in your programs.
1887 There are three kinds of global symbols, illustrated here by C examples:
1891 A definition, which goes in the initialized data section of the output
1895 An undefined reference, which does not allocate space.
1896 There must be either a definition or a common symbol for the
1900 A common symbol. If there are only (one or more) common symbols for a
1901 variable, it goes in the uninitialized data area of the output file.
1902 The linker merges multiple common symbols for the same variable into a
1903 single symbol. If they are of different sizes, it picks the largest
1904 size. The linker turns a common symbol into a declaration, if there is
1905 a definition of the same variable.
1908 The @samp{--warn-common} option can produce five kinds of warnings.
1909 Each warning consists of a pair of lines: the first describes the symbol
1910 just encountered, and the second describes the previous symbol
1911 encountered with the same name. One or both of the two symbols will be
1916 Turning a common symbol into a reference, because there is already a
1917 definition for the symbol.
1919 @var{file}(@var{section}): warning: common of `@var{symbol}'
1920 overridden by definition
1921 @var{file}(@var{section}): warning: defined here
1925 Turning a common symbol into a reference, because a later definition for
1926 the symbol is encountered. This is the same as the previous case,
1927 except that the symbols are encountered in a different order.
1929 @var{file}(@var{section}): warning: definition of `@var{symbol}'
1931 @var{file}(@var{section}): warning: common is here
1935 Merging a common symbol with a previous same-sized common symbol.
1937 @var{file}(@var{section}): warning: multiple common
1939 @var{file}(@var{section}): warning: previous common is here
1943 Merging a common symbol with a previous larger common symbol.
1945 @var{file}(@var{section}): warning: common of `@var{symbol}'
1946 overridden by larger common
1947 @var{file}(@var{section}): warning: larger common is here
1951 Merging a common symbol with a previous smaller common symbol. This is
1952 the same as the previous case, except that the symbols are
1953 encountered in a different order.
1955 @var{file}(@var{section}): warning: common of `@var{symbol}'
1956 overriding smaller common
1957 @var{file}(@var{section}): warning: smaller common is here
1961 @kindex --warn-constructors
1962 @item --warn-constructors
1963 Warn if any global constructors are used. This is only useful for a few
1964 object file formats. For formats like COFF or ELF, the linker can not
1965 detect the use of global constructors.
1967 @kindex --warn-multiple-gp
1968 @item --warn-multiple-gp
1969 Warn if multiple global pointer values are required in the output file.
1970 This is only meaningful for certain processors, such as the Alpha.
1971 Specifically, some processors put large-valued constants in a special
1972 section. A special register (the global pointer) points into the middle
1973 of this section, so that constants can be loaded efficiently via a
1974 base-register relative addressing mode. Since the offset in
1975 base-register relative mode is fixed and relatively small (e.g., 16
1976 bits), this limits the maximum size of the constant pool. Thus, in
1977 large programs, it is often necessary to use multiple global pointer
1978 values in order to be able to address all possible constants. This
1979 option causes a warning to be issued whenever this case occurs.
1982 @cindex warnings, on undefined symbols
1983 @cindex undefined symbols, warnings on
1985 Only warn once for each undefined symbol, rather than once per module
1988 @kindex --warn-section-align
1989 @cindex warnings, on section alignment
1990 @cindex section alignment, warnings on
1991 @item --warn-section-align
1992 Warn if the address of an output section is changed because of
1993 alignment. Typically, the alignment will be set by an input section.
1994 The address will only be changed if it not explicitly specified; that
1995 is, if the @code{SECTIONS} command does not specify a start address for
1996 the section (@pxref{SECTIONS}).
1998 @kindex --warn-shared-textrel
1999 @item --warn-shared-textrel
2000 Warn if the linker adds a DT_TEXTREL to a shared object.
2002 @kindex --warn-alternate-em
2003 @item --warn-alternate-em
2004 Warn if an object has alternate ELF machine code.
2006 @kindex --warn-unresolved-symbols
2007 @item --warn-unresolved-symbols
2008 If the linker is going to report an unresolved symbol (see the option
2009 @option{--unresolved-symbols}) it will normally generate an error.
2010 This option makes it generate a warning instead.
2012 @kindex --error-unresolved-symbols
2013 @item --error-unresolved-symbols
2014 This restores the linker's default behaviour of generating errors when
2015 it is reporting unresolved symbols.
2017 @kindex --whole-archive
2018 @cindex including an entire archive
2019 @item --whole-archive
2020 For each archive mentioned on the command line after the
2021 @option{--whole-archive} option, include every object file in the archive
2022 in the link, rather than searching the archive for the required object
2023 files. This is normally used to turn an archive file into a shared
2024 library, forcing every object to be included in the resulting shared
2025 library. This option may be used more than once.
2027 Two notes when using this option from gcc: First, gcc doesn't know
2028 about this option, so you have to use @option{-Wl,-whole-archive}.
2029 Second, don't forget to use @option{-Wl,-no-whole-archive} after your
2030 list of archives, because gcc will add its own list of archives to
2031 your link and you may not want this flag to affect those as well.
2033 @kindex --wrap=@var{symbol}
2034 @item --wrap=@var{symbol}
2035 Use a wrapper function for @var{symbol}. Any undefined reference to
2036 @var{symbol} will be resolved to @code{__wrap_@var{symbol}}. Any
2037 undefined reference to @code{__real_@var{symbol}} will be resolved to
2040 This can be used to provide a wrapper for a system function. The
2041 wrapper function should be called @code{__wrap_@var{symbol}}. If it
2042 wishes to call the system function, it should call
2043 @code{__real_@var{symbol}}.
2045 Here is a trivial example:
2049 __wrap_malloc (size_t c)
2051 printf ("malloc called with %zu\n", c);
2052 return __real_malloc (c);
2056 If you link other code with this file using @option{--wrap malloc}, then
2057 all calls to @code{malloc} will call the function @code{__wrap_malloc}
2058 instead. The call to @code{__real_malloc} in @code{__wrap_malloc} will
2059 call the real @code{malloc} function.
2061 You may wish to provide a @code{__real_malloc} function as well, so that
2062 links without the @option{--wrap} option will succeed. If you do this,
2063 you should not put the definition of @code{__real_malloc} in the same
2064 file as @code{__wrap_malloc}; if you do, the assembler may resolve the
2065 call before the linker has a chance to wrap it to @code{malloc}.
2067 @kindex --eh-frame-hdr
2068 @item --eh-frame-hdr
2069 Request creation of @code{.eh_frame_hdr} section and ELF
2070 @code{PT_GNU_EH_FRAME} segment header.
2072 @kindex --ld-generated-unwind-info
2073 @item --no-ld-generated-unwind-info
2074 Request creation of @code{.eh_frame} unwind info for linker
2075 generated code sections like PLT. This option is on by default
2076 if linker generated unwind info is supported.
2078 @kindex --enable-new-dtags
2079 @kindex --disable-new-dtags
2080 @item --enable-new-dtags
2081 @itemx --disable-new-dtags
2082 This linker can create the new dynamic tags in ELF. But the older ELF
2083 systems may not understand them. If you specify
2084 @option{--enable-new-dtags}, the dynamic tags will be created as needed.
2085 If you specify @option{--disable-new-dtags}, no new dynamic tags will be
2086 created. By default, the new dynamic tags are not created. Note that
2087 those options are only available for ELF systems.
2089 @kindex --hash-size=@var{number}
2090 @item --hash-size=@var{number}
2091 Set the default size of the linker's hash tables to a prime number
2092 close to @var{number}. Increasing this value can reduce the length of
2093 time it takes the linker to perform its tasks, at the expense of
2094 increasing the linker's memory requirements. Similarly reducing this
2095 value can reduce the memory requirements at the expense of speed.
2097 @kindex --hash-style=@var{style}
2098 @item --hash-style=@var{style}
2099 Set the type of linker's hash table(s). @var{style} can be either
2100 @code{sysv} for classic ELF @code{.hash} section, @code{gnu} for
2101 new style GNU @code{.gnu.hash} section or @code{both} for both
2102 the classic ELF @code{.hash} and new style GNU @code{.gnu.hash}
2103 hash tables. The default is @code{sysv}.
2105 @kindex --reduce-memory-overheads
2106 @item --reduce-memory-overheads
2107 This option reduces memory requirements at ld runtime, at the expense of
2108 linking speed. This was introduced to select the old O(n^2) algorithm
2109 for link map file generation, rather than the new O(n) algorithm which uses
2110 about 40% more memory for symbol storage.
2112 Another effect of the switch is to set the default hash table size to
2113 1021, which again saves memory at the cost of lengthening the linker's
2114 run time. This is not done however if the @option{--hash-size} switch
2117 The @option{--reduce-memory-overheads} switch may be also be used to
2118 enable other tradeoffs in future versions of the linker.
2121 @kindex --build-id=@var{style}
2123 @itemx --build-id=@var{style}
2124 Request creation of @code{.note.gnu.build-id} ELF note section.
2125 The contents of the note are unique bits identifying this linked
2126 file. @var{style} can be @code{uuid} to use 128 random bits,
2127 @code{sha1} to use a 160-bit @sc{SHA1} hash on the normative
2128 parts of the output contents, @code{md5} to use a 128-bit
2129 @sc{MD5} hash on the normative parts of the output contents, or
2130 @code{0x@var{hexstring}} to use a chosen bit string specified as
2131 an even number of hexadecimal digits (@code{-} and @code{:}
2132 characters between digit pairs are ignored). If @var{style} is
2133 omitted, @code{sha1} is used.
2135 The @code{md5} and @code{sha1} styles produces an identifier
2136 that is always the same in an identical output file, but will be
2137 unique among all nonidentical output files. It is not intended
2138 to be compared as a checksum for the file's contents. A linked
2139 file may be changed later by other tools, but the build ID bit
2140 string identifying the original linked file does not change.
2142 Passing @code{none} for @var{style} disables the setting from any
2143 @code{--build-id} options earlier on the command line.
2148 @subsection Options Specific to i386 PE Targets
2150 @c man begin OPTIONS
2152 The i386 PE linker supports the @option{-shared} option, which causes
2153 the output to be a dynamically linked library (DLL) instead of a
2154 normal executable. You should name the output @code{*.dll} when you
2155 use this option. In addition, the linker fully supports the standard
2156 @code{*.def} files, which may be specified on the linker command line
2157 like an object file (in fact, it should precede archives it exports
2158 symbols from, to ensure that they get linked in, just like a normal
2161 In addition to the options common to all targets, the i386 PE linker
2162 support additional command line options that are specific to the i386
2163 PE target. Options that take values may be separated from their
2164 values by either a space or an equals sign.
2168 @kindex --add-stdcall-alias
2169 @item --add-stdcall-alias
2170 If given, symbols with a stdcall suffix (@@@var{nn}) will be exported
2171 as-is and also with the suffix stripped.
2172 [This option is specific to the i386 PE targeted port of the linker]
2175 @item --base-file @var{file}
2176 Use @var{file} as the name of a file in which to save the base
2177 addresses of all the relocations needed for generating DLLs with
2179 [This is an i386 PE specific option]
2183 Create a DLL instead of a regular executable. You may also use
2184 @option{-shared} or specify a @code{LIBRARY} in a given @code{.def}
2186 [This option is specific to the i386 PE targeted port of the linker]
2188 @kindex --enable-long-section-names
2189 @kindex --disable-long-section-names
2190 @item --enable-long-section-names
2191 @itemx --disable-long-section-names
2192 The PE variants of the Coff object format add an extension that permits
2193 the use of section names longer than eight characters, the normal limit
2194 for Coff. By default, these names are only allowed in object files, as
2195 fully-linked executable images do not carry the Coff string table required
2196 to support the longer names. As a GNU extension, it is possible to
2197 allow their use in executable images as well, or to (probably pointlessly!)
2198 disallow it in object files, by using these two options. Executable images
2199 generated with these long section names are slightly non-standard, carrying
2200 as they do a string table, and may generate confusing output when examined
2201 with non-GNU PE-aware tools, such as file viewers and dumpers. However,
2202 GDB relies on the use of PE long section names to find Dwarf-2 debug
2203 information sections in an executable image at runtime, and so if neither
2204 option is specified on the command-line, @command{ld} will enable long
2205 section names, overriding the default and technically correct behaviour,
2206 when it finds the presence of debug information while linking an executable
2207 image and not stripping symbols.
2208 [This option is valid for all PE targeted ports of the linker]
2210 @kindex --enable-stdcall-fixup
2211 @kindex --disable-stdcall-fixup
2212 @item --enable-stdcall-fixup
2213 @itemx --disable-stdcall-fixup
2214 If the link finds a symbol that it cannot resolve, it will attempt to
2215 do ``fuzzy linking'' by looking for another defined symbol that differs
2216 only in the format of the symbol name (cdecl vs stdcall) and will
2217 resolve that symbol by linking to the match. For example, the
2218 undefined symbol @code{_foo} might be linked to the function
2219 @code{_foo@@12}, or the undefined symbol @code{_bar@@16} might be linked
2220 to the function @code{_bar}. When the linker does this, it prints a
2221 warning, since it normally should have failed to link, but sometimes
2222 import libraries generated from third-party dlls may need this feature
2223 to be usable. If you specify @option{--enable-stdcall-fixup}, this
2224 feature is fully enabled and warnings are not printed. If you specify
2225 @option{--disable-stdcall-fixup}, this feature is disabled and such
2226 mismatches are considered to be errors.
2227 [This option is specific to the i386 PE targeted port of the linker]
2229 @kindex --leading-underscore
2230 @kindex --no-leading-underscore
2231 @item --leading-underscore
2232 @itemx --no-leading-underscore
2233 For most targets default symbol-prefix is an underscore and is defined
2234 in target's description. By this option it is possible to
2235 disable/enable the default underscore symbol-prefix.
2237 @cindex DLLs, creating
2238 @kindex --export-all-symbols
2239 @item --export-all-symbols
2240 If given, all global symbols in the objects used to build a DLL will
2241 be exported by the DLL. Note that this is the default if there
2242 otherwise wouldn't be any exported symbols. When symbols are
2243 explicitly exported via DEF files or implicitly exported via function
2244 attributes, the default is to not export anything else unless this
2245 option is given. Note that the symbols @code{DllMain@@12},
2246 @code{DllEntryPoint@@0}, @code{DllMainCRTStartup@@12}, and
2247 @code{impure_ptr} will not be automatically
2248 exported. Also, symbols imported from other DLLs will not be
2249 re-exported, nor will symbols specifying the DLL's internal layout
2250 such as those beginning with @code{_head_} or ending with
2251 @code{_iname}. In addition, no symbols from @code{libgcc},
2252 @code{libstd++}, @code{libmingw32}, or @code{crtX.o} will be exported.
2253 Symbols whose names begin with @code{__rtti_} or @code{__builtin_} will
2254 not be exported, to help with C++ DLLs. Finally, there is an
2255 extensive list of cygwin-private symbols that are not exported
2256 (obviously, this applies on when building DLLs for cygwin targets).
2257 These cygwin-excludes are: @code{_cygwin_dll_entry@@12},
2258 @code{_cygwin_crt0_common@@8}, @code{_cygwin_noncygwin_dll_entry@@12},
2259 @code{_fmode}, @code{_impure_ptr}, @code{cygwin_attach_dll},
2260 @code{cygwin_premain0}, @code{cygwin_premain1}, @code{cygwin_premain2},
2261 @code{cygwin_premain3}, and @code{environ}.
2262 [This option is specific to the i386 PE targeted port of the linker]
2264 @kindex --exclude-symbols
2265 @item --exclude-symbols @var{symbol},@var{symbol},...
2266 Specifies a list of symbols which should not be automatically
2267 exported. The symbol names may be delimited by commas or colons.
2268 [This option is specific to the i386 PE targeted port of the linker]
2270 @kindex --exclude-all-symbols
2271 @item --exclude-all-symbols
2272 Specifies no symbols should be automatically exported.
2273 [This option is specific to the i386 PE targeted port of the linker]
2275 @kindex --file-alignment
2276 @item --file-alignment
2277 Specify the file alignment. Sections in the file will always begin at
2278 file offsets which are multiples of this number. This defaults to
2280 [This option is specific to the i386 PE targeted port of the linker]
2284 @item --heap @var{reserve}
2285 @itemx --heap @var{reserve},@var{commit}
2286 Specify the number of bytes of memory to reserve (and optionally commit)
2287 to be used as heap for this program. The default is 1Mb reserved, 4K
2289 [This option is specific to the i386 PE targeted port of the linker]
2292 @kindex --image-base
2293 @item --image-base @var{value}
2294 Use @var{value} as the base address of your program or dll. This is
2295 the lowest memory location that will be used when your program or dll
2296 is loaded. To reduce the need to relocate and improve performance of
2297 your dlls, each should have a unique base address and not overlap any
2298 other dlls. The default is 0x400000 for executables, and 0x10000000
2300 [This option is specific to the i386 PE targeted port of the linker]
2304 If given, the stdcall suffixes (@@@var{nn}) will be stripped from
2305 symbols before they are exported.
2306 [This option is specific to the i386 PE targeted port of the linker]
2308 @kindex --large-address-aware
2309 @item --large-address-aware
2310 If given, the appropriate bit in the ``Characteristics'' field of the COFF
2311 header is set to indicate that this executable supports virtual addresses
2312 greater than 2 gigabytes. This should be used in conjunction with the /3GB
2313 or /USERVA=@var{value} megabytes switch in the ``[operating systems]''
2314 section of the BOOT.INI. Otherwise, this bit has no effect.
2315 [This option is specific to PE targeted ports of the linker]
2317 @kindex --major-image-version
2318 @item --major-image-version @var{value}
2319 Sets the major number of the ``image version''. Defaults to 1.
2320 [This option is specific to the i386 PE targeted port of the linker]
2322 @kindex --major-os-version
2323 @item --major-os-version @var{value}
2324 Sets the major number of the ``os version''. Defaults to 4.
2325 [This option is specific to the i386 PE targeted port of the linker]
2327 @kindex --major-subsystem-version
2328 @item --major-subsystem-version @var{value}
2329 Sets the major number of the ``subsystem version''. Defaults to 4.
2330 [This option is specific to the i386 PE targeted port of the linker]
2332 @kindex --minor-image-version
2333 @item --minor-image-version @var{value}
2334 Sets the minor number of the ``image version''. Defaults to 0.
2335 [This option is specific to the i386 PE targeted port of the linker]
2337 @kindex --minor-os-version
2338 @item --minor-os-version @var{value}
2339 Sets the minor number of the ``os version''. Defaults to 0.
2340 [This option is specific to the i386 PE targeted port of the linker]
2342 @kindex --minor-subsystem-version
2343 @item --minor-subsystem-version @var{value}
2344 Sets the minor number of the ``subsystem version''. Defaults to 0.
2345 [This option is specific to the i386 PE targeted port of the linker]
2347 @cindex DEF files, creating
2348 @cindex DLLs, creating
2349 @kindex --output-def
2350 @item --output-def @var{file}
2351 The linker will create the file @var{file} which will contain a DEF
2352 file corresponding to the DLL the linker is generating. This DEF file
2353 (which should be called @code{*.def}) may be used to create an import
2354 library with @code{dlltool} or may be used as a reference to
2355 automatically or implicitly exported symbols.
2356 [This option is specific to the i386 PE targeted port of the linker]
2358 @cindex DLLs, creating
2359 @kindex --out-implib
2360 @item --out-implib @var{file}
2361 The linker will create the file @var{file} which will contain an
2362 import lib corresponding to the DLL the linker is generating. This
2363 import lib (which should be called @code{*.dll.a} or @code{*.a}
2364 may be used to link clients against the generated DLL; this behaviour
2365 makes it possible to skip a separate @code{dlltool} import library
2367 [This option is specific to the i386 PE targeted port of the linker]
2369 @kindex --enable-auto-image-base
2370 @item --enable-auto-image-base
2371 Automatically choose the image base for DLLs, unless one is specified
2372 using the @code{--image-base} argument. By using a hash generated
2373 from the dllname to create unique image bases for each DLL, in-memory
2374 collisions and relocations which can delay program execution are
2376 [This option is specific to the i386 PE targeted port of the linker]
2378 @kindex --disable-auto-image-base
2379 @item --disable-auto-image-base
2380 Do not automatically generate a unique image base. If there is no
2381 user-specified image base (@code{--image-base}) then use the platform
2383 [This option is specific to the i386 PE targeted port of the linker]
2385 @cindex DLLs, linking to
2386 @kindex --dll-search-prefix
2387 @item --dll-search-prefix @var{string}
2388 When linking dynamically to a dll without an import library,
2389 search for @code{<string><basename>.dll} in preference to
2390 @code{lib<basename>.dll}. This behaviour allows easy distinction
2391 between DLLs built for the various "subplatforms": native, cygwin,
2392 uwin, pw, etc. For instance, cygwin DLLs typically use
2393 @code{--dll-search-prefix=cyg}.
2394 [This option is specific to the i386 PE targeted port of the linker]
2396 @kindex --enable-auto-import
2397 @item --enable-auto-import
2398 Do sophisticated linking of @code{_symbol} to @code{__imp__symbol} for
2399 DATA imports from DLLs, and create the necessary thunking symbols when
2400 building the import libraries with those DATA exports. Note: Use of the
2401 'auto-import' extension will cause the text section of the image file
2402 to be made writable. This does not conform to the PE-COFF format
2403 specification published by Microsoft.
2405 Note - use of the 'auto-import' extension will also cause read only
2406 data which would normally be placed into the .rdata section to be
2407 placed into the .data section instead. This is in order to work
2408 around a problem with consts that is described here:
2409 http://www.cygwin.com/ml/cygwin/2004-09/msg01101.html
2411 Using 'auto-import' generally will 'just work' -- but sometimes you may
2414 "variable '<var>' can't be auto-imported. Please read the
2415 documentation for ld's @code{--enable-auto-import} for details."
2417 This message occurs when some (sub)expression accesses an address
2418 ultimately given by the sum of two constants (Win32 import tables only
2419 allow one). Instances where this may occur include accesses to member
2420 fields of struct variables imported from a DLL, as well as using a
2421 constant index into an array variable imported from a DLL. Any
2422 multiword variable (arrays, structs, long long, etc) may trigger
2423 this error condition. However, regardless of the exact data type
2424 of the offending exported variable, ld will always detect it, issue
2425 the warning, and exit.
2427 There are several ways to address this difficulty, regardless of the
2428 data type of the exported variable:
2430 One way is to use --enable-runtime-pseudo-reloc switch. This leaves the task
2431 of adjusting references in your client code for runtime environment, so
2432 this method works only when runtime environment supports this feature.
2434 A second solution is to force one of the 'constants' to be a variable --
2435 that is, unknown and un-optimizable at compile time. For arrays,
2436 there are two possibilities: a) make the indexee (the array's address)
2437 a variable, or b) make the 'constant' index a variable. Thus:
2440 extern type extern_array[];
2442 @{ volatile type *t=extern_array; t[1] @}
2448 extern type extern_array[];
2450 @{ volatile int t=1; extern_array[t] @}
2453 For structs (and most other multiword data types) the only option
2454 is to make the struct itself (or the long long, or the ...) variable:
2457 extern struct s extern_struct;
2458 extern_struct.field -->
2459 @{ volatile struct s *t=&extern_struct; t->field @}
2465 extern long long extern_ll;
2467 @{ volatile long long * local_ll=&extern_ll; *local_ll @}
2470 A third method of dealing with this difficulty is to abandon
2471 'auto-import' for the offending symbol and mark it with
2472 @code{__declspec(dllimport)}. However, in practise that
2473 requires using compile-time #defines to indicate whether you are
2474 building a DLL, building client code that will link to the DLL, or
2475 merely building/linking to a static library. In making the choice
2476 between the various methods of resolving the 'direct address with
2477 constant offset' problem, you should consider typical real-world usage:
2485 void main(int argc, char **argv)@{
2486 printf("%d\n",arr[1]);
2496 void main(int argc, char **argv)@{
2497 /* This workaround is for win32 and cygwin; do not "optimize" */
2498 volatile int *parr = arr;
2499 printf("%d\n",parr[1]);
2506 /* Note: auto-export is assumed (no __declspec(dllexport)) */
2507 #if (defined(_WIN32) || defined(__CYGWIN__)) && \
2508 !(defined(FOO_BUILD_DLL) || defined(FOO_STATIC))
2509 #define FOO_IMPORT __declspec(dllimport)
2513 extern FOO_IMPORT int arr[];
2516 void main(int argc, char **argv)@{
2517 printf("%d\n",arr[1]);
2521 A fourth way to avoid this problem is to re-code your
2522 library to use a functional interface rather than a data interface
2523 for the offending variables (e.g. set_foo() and get_foo() accessor
2525 [This option is specific to the i386 PE targeted port of the linker]
2527 @kindex --disable-auto-import
2528 @item --disable-auto-import
2529 Do not attempt to do sophisticated linking of @code{_symbol} to
2530 @code{__imp__symbol} for DATA imports from DLLs.
2531 [This option is specific to the i386 PE targeted port of the linker]
2533 @kindex --enable-runtime-pseudo-reloc
2534 @item --enable-runtime-pseudo-reloc
2535 If your code contains expressions described in --enable-auto-import section,
2536 that is, DATA imports from DLL with non-zero offset, this switch will create
2537 a vector of 'runtime pseudo relocations' which can be used by runtime
2538 environment to adjust references to such data in your client code.
2539 [This option is specific to the i386 PE targeted port of the linker]
2541 @kindex --disable-runtime-pseudo-reloc
2542 @item --disable-runtime-pseudo-reloc
2543 Do not create pseudo relocations for non-zero offset DATA imports from
2544 DLLs. This is the default.
2545 [This option is specific to the i386 PE targeted port of the linker]
2547 @kindex --enable-extra-pe-debug
2548 @item --enable-extra-pe-debug
2549 Show additional debug info related to auto-import symbol thunking.
2550 [This option is specific to the i386 PE targeted port of the linker]
2552 @kindex --section-alignment
2553 @item --section-alignment
2554 Sets the section alignment. Sections in memory will always begin at
2555 addresses which are a multiple of this number. Defaults to 0x1000.
2556 [This option is specific to the i386 PE targeted port of the linker]
2560 @item --stack @var{reserve}
2561 @itemx --stack @var{reserve},@var{commit}
2562 Specify the number of bytes of memory to reserve (and optionally commit)
2563 to be used as stack for this program. The default is 2Mb reserved, 4K
2565 [This option is specific to the i386 PE targeted port of the linker]
2568 @item --subsystem @var{which}
2569 @itemx --subsystem @var{which}:@var{major}
2570 @itemx --subsystem @var{which}:@var{major}.@var{minor}
2571 Specifies the subsystem under which your program will execute. The
2572 legal values for @var{which} are @code{native}, @code{windows},
2573 @code{console}, @code{posix}, and @code{xbox}. You may optionally set
2574 the subsystem version also. Numeric values are also accepted for
2576 [This option is specific to the i386 PE targeted port of the linker]
2578 The following options set flags in the @code{DllCharacteristics} field
2579 of the PE file header:
2580 [These options are specific to PE targeted ports of the linker]
2582 @kindex --dynamicbase
2584 The image base address may be relocated using address space layout
2585 randomization (ASLR). This feature was introduced with MS Windows
2586 Vista for i386 PE targets.
2588 @kindex --forceinteg
2590 Code integrity checks are enforced.
2594 The image is compatible with the Data Execution Prevention.
2595 This feature was introduced with MS Windows XP SP2 for i386 PE targets.
2597 @kindex --no-isolation
2598 @item --no-isolation
2599 Although the image understands isolation, do not isolate the image.
2603 The image does not use SEH. No SE handler may be called from
2608 Do not bind this image.
2612 The driver uses the MS Windows Driver Model.
2616 The image is Terminal Server aware.
2623 @subsection Options specific to C6X uClinux targets
2625 @c man begin OPTIONS
2627 The C6X uClinux target uses a binary format called DSBT to support shared
2628 libraries. Each shared library in the system needs to have a unique index;
2629 all executables use an index of 0.
2634 @item --dsbt-size @var{size}
2635 This option sets the number of entires in the DSBT of the current executable
2636 or shared library to @var{size}. The default is to create a table with 64
2639 @kindex --dsbt-index
2640 @item --dsbt-index @var{index}
2641 This option sets the DSBT index of the current executable or shared library
2642 to @var{index}. The default is 0, which is appropriate for generating
2643 executables. If a shared library is generated with a DSBT index of 0, the
2644 @code{R_C6000_DSBT_INDEX} relocs are copied into the output file.
2646 @kindex --no-merge-exidx-entries
2647 The @samp{--no-merge-exidx-entries} switch disables the merging of adjacent
2648 exidx entries in frame unwind info.
2656 @subsection Options specific to Motorola 68HC11 and 68HC12 targets
2658 @c man begin OPTIONS
2660 The 68HC11 and 68HC12 linkers support specific options to control the
2661 memory bank switching mapping and trampoline code generation.
2665 @kindex --no-trampoline
2666 @item --no-trampoline
2667 This option disables the generation of trampoline. By default a trampoline
2668 is generated for each far function which is called using a @code{jsr}
2669 instruction (this happens when a pointer to a far function is taken).
2671 @kindex --bank-window
2672 @item --bank-window @var{name}
2673 This option indicates to the linker the name of the memory region in
2674 the @samp{MEMORY} specification that describes the memory bank window.
2675 The definition of such region is then used by the linker to compute
2676 paging and addresses within the memory window.
2684 @subsection Options specific to Motorola 68K target
2686 @c man begin OPTIONS
2688 The following options are supported to control handling of GOT generation
2689 when linking for 68K targets.
2694 @item --got=@var{type}
2695 This option tells the linker which GOT generation scheme to use.
2696 @var{type} should be one of @samp{single}, @samp{negative},
2697 @samp{multigot} or @samp{target}. For more information refer to the
2698 Info entry for @file{ld}.
2707 @section Environment Variables
2709 @c man begin ENVIRONMENT
2711 You can change the behaviour of @command{ld} with the environment variables
2712 @ifclear SingleFormat
2715 @code{LDEMULATION} and @code{COLLECT_NO_DEMANGLE}.
2717 @ifclear SingleFormat
2719 @cindex default input format
2720 @code{GNUTARGET} determines the input-file object format if you don't
2721 use @samp{-b} (or its synonym @samp{--format}). Its value should be one
2722 of the BFD names for an input format (@pxref{BFD}). If there is no
2723 @code{GNUTARGET} in the environment, @command{ld} uses the natural format
2724 of the target. If @code{GNUTARGET} is set to @code{default} then BFD
2725 attempts to discover the input format by examining binary input files;
2726 this method often succeeds, but there are potential ambiguities, since
2727 there is no method of ensuring that the magic number used to specify
2728 object-file formats is unique. However, the configuration procedure for
2729 BFD on each system places the conventional format for that system first
2730 in the search-list, so ambiguities are resolved in favor of convention.
2734 @cindex default emulation
2735 @cindex emulation, default
2736 @code{LDEMULATION} determines the default emulation if you don't use the
2737 @samp{-m} option. The emulation can affect various aspects of linker
2738 behaviour, particularly the default linker script. You can list the
2739 available emulations with the @samp{--verbose} or @samp{-V} options. If
2740 the @samp{-m} option is not used, and the @code{LDEMULATION} environment
2741 variable is not defined, the default emulation depends upon how the
2742 linker was configured.
2744 @kindex COLLECT_NO_DEMANGLE
2745 @cindex demangling, default
2746 Normally, the linker will default to demangling symbols. However, if
2747 @code{COLLECT_NO_DEMANGLE} is set in the environment, then it will
2748 default to not demangling symbols. This environment variable is used in
2749 a similar fashion by the @code{gcc} linker wrapper program. The default
2750 may be overridden by the @samp{--demangle} and @samp{--no-demangle}
2757 @chapter Linker Scripts
2760 @cindex linker scripts
2761 @cindex command files
2762 Every link is controlled by a @dfn{linker script}. This script is
2763 written in the linker command language.
2765 The main purpose of the linker script is to describe how the sections in
2766 the input files should be mapped into the output file, and to control
2767 the memory layout of the output file. Most linker scripts do nothing
2768 more than this. However, when necessary, the linker script can also
2769 direct the linker to perform many other operations, using the commands
2772 The linker always uses a linker script. If you do not supply one
2773 yourself, the linker will use a default script that is compiled into the
2774 linker executable. You can use the @samp{--verbose} command line option
2775 to display the default linker script. Certain command line options,
2776 such as @samp{-r} or @samp{-N}, will affect the default linker script.
2778 You may supply your own linker script by using the @samp{-T} command
2779 line option. When you do this, your linker script will replace the
2780 default linker script.
2782 You may also use linker scripts implicitly by naming them as input files
2783 to the linker, as though they were files to be linked. @xref{Implicit
2787 * Basic Script Concepts:: Basic Linker Script Concepts
2788 * Script Format:: Linker Script Format
2789 * Simple Example:: Simple Linker Script Example
2790 * Simple Commands:: Simple Linker Script Commands
2791 * Assignments:: Assigning Values to Symbols
2792 * SECTIONS:: SECTIONS Command
2793 * MEMORY:: MEMORY Command
2794 * PHDRS:: PHDRS Command
2795 * VERSION:: VERSION Command
2796 * Expressions:: Expressions in Linker Scripts
2797 * Implicit Linker Scripts:: Implicit Linker Scripts
2800 @node Basic Script Concepts
2801 @section Basic Linker Script Concepts
2802 @cindex linker script concepts
2803 We need to define some basic concepts and vocabulary in order to
2804 describe the linker script language.
2806 The linker combines input files into a single output file. The output
2807 file and each input file are in a special data format known as an
2808 @dfn{object file format}. Each file is called an @dfn{object file}.
2809 The output file is often called an @dfn{executable}, but for our
2810 purposes we will also call it an object file. Each object file has,
2811 among other things, a list of @dfn{sections}. We sometimes refer to a
2812 section in an input file as an @dfn{input section}; similarly, a section
2813 in the output file is an @dfn{output section}.
2815 Each section in an object file has a name and a size. Most sections
2816 also have an associated block of data, known as the @dfn{section
2817 contents}. A section may be marked as @dfn{loadable}, which mean that
2818 the contents should be loaded into memory when the output file is run.
2819 A section with no contents may be @dfn{allocatable}, which means that an
2820 area in memory should be set aside, but nothing in particular should be
2821 loaded there (in some cases this memory must be zeroed out). A section
2822 which is neither loadable nor allocatable typically contains some sort
2823 of debugging information.
2825 Every loadable or allocatable output section has two addresses. The
2826 first is the @dfn{VMA}, or virtual memory address. This is the address
2827 the section will have when the output file is run. The second is the
2828 @dfn{LMA}, or load memory address. This is the address at which the
2829 section will be loaded. In most cases the two addresses will be the
2830 same. An example of when they might be different is when a data section
2831 is loaded into ROM, and then copied into RAM when the program starts up
2832 (this technique is often used to initialize global variables in a ROM
2833 based system). In this case the ROM address would be the LMA, and the
2834 RAM address would be the VMA.
2836 You can see the sections in an object file by using the @code{objdump}
2837 program with the @samp{-h} option.
2839 Every object file also has a list of @dfn{symbols}, known as the
2840 @dfn{symbol table}. A symbol may be defined or undefined. Each symbol
2841 has a name, and each defined symbol has an address, among other
2842 information. If you compile a C or C++ program into an object file, you
2843 will get a defined symbol for every defined function and global or
2844 static variable. Every undefined function or global variable which is
2845 referenced in the input file will become an undefined symbol.
2847 You can see the symbols in an object file by using the @code{nm}
2848 program, or by using the @code{objdump} program with the @samp{-t}
2852 @section Linker Script Format
2853 @cindex linker script format
2854 Linker scripts are text files.
2856 You write a linker script as a series of commands. Each command is
2857 either a keyword, possibly followed by arguments, or an assignment to a
2858 symbol. You may separate commands using semicolons. Whitespace is
2861 Strings such as file or format names can normally be entered directly.
2862 If the file name contains a character such as a comma which would
2863 otherwise serve to separate file names, you may put the file name in
2864 double quotes. There is no way to use a double quote character in a
2867 You may include comments in linker scripts just as in C, delimited by
2868 @samp{/*} and @samp{*/}. As in C, comments are syntactically equivalent
2871 @node Simple Example
2872 @section Simple Linker Script Example
2873 @cindex linker script example
2874 @cindex example of linker script
2875 Many linker scripts are fairly simple.
2877 The simplest possible linker script has just one command:
2878 @samp{SECTIONS}. You use the @samp{SECTIONS} command to describe the
2879 memory layout of the output file.
2881 The @samp{SECTIONS} command is a powerful command. Here we will
2882 describe a simple use of it. Let's assume your program consists only of
2883 code, initialized data, and uninitialized data. These will be in the
2884 @samp{.text}, @samp{.data}, and @samp{.bss} sections, respectively.
2885 Let's assume further that these are the only sections which appear in
2888 For this example, let's say that the code should be loaded at address
2889 0x10000, and that the data should start at address 0x8000000. Here is a
2890 linker script which will do that:
2895 .text : @{ *(.text) @}
2897 .data : @{ *(.data) @}
2898 .bss : @{ *(.bss) @}
2902 You write the @samp{SECTIONS} command as the keyword @samp{SECTIONS},
2903 followed by a series of symbol assignments and output section
2904 descriptions enclosed in curly braces.
2906 The first line inside the @samp{SECTIONS} command of the above example
2907 sets the value of the special symbol @samp{.}, which is the location
2908 counter. If you do not specify the address of an output section in some
2909 other way (other ways are described later), the address is set from the
2910 current value of the location counter. The location counter is then
2911 incremented by the size of the output section. At the start of the
2912 @samp{SECTIONS} command, the location counter has the value @samp{0}.
2914 The second line defines an output section, @samp{.text}. The colon is
2915 required syntax which may be ignored for now. Within the curly braces
2916 after the output section name, you list the names of the input sections
2917 which should be placed into this output section. The @samp{*} is a
2918 wildcard which matches any file name. The expression @samp{*(.text)}
2919 means all @samp{.text} input sections in all input files.
2921 Since the location counter is @samp{0x10000} when the output section
2922 @samp{.text} is defined, the linker will set the address of the
2923 @samp{.text} section in the output file to be @samp{0x10000}.
2925 The remaining lines define the @samp{.data} and @samp{.bss} sections in
2926 the output file. The linker will place the @samp{.data} output section
2927 at address @samp{0x8000000}. After the linker places the @samp{.data}
2928 output section, the value of the location counter will be
2929 @samp{0x8000000} plus the size of the @samp{.data} output section. The
2930 effect is that the linker will place the @samp{.bss} output section
2931 immediately after the @samp{.data} output section in memory.
2933 The linker will ensure that each output section has the required
2934 alignment, by increasing the location counter if necessary. In this
2935 example, the specified addresses for the @samp{.text} and @samp{.data}
2936 sections will probably satisfy any alignment constraints, but the linker
2937 may have to create a small gap between the @samp{.data} and @samp{.bss}
2940 That's it! That's a simple and complete linker script.
2942 @node Simple Commands
2943 @section Simple Linker Script Commands
2944 @cindex linker script simple commands
2945 In this section we describe the simple linker script commands.
2948 * Entry Point:: Setting the entry point
2949 * File Commands:: Commands dealing with files
2950 @ifclear SingleFormat
2951 * Format Commands:: Commands dealing with object file formats
2954 * REGION_ALIAS:: Assign alias names to memory regions
2955 * Miscellaneous Commands:: Other linker script commands
2959 @subsection Setting the Entry Point
2960 @kindex ENTRY(@var{symbol})
2961 @cindex start of execution
2962 @cindex first instruction
2964 The first instruction to execute in a program is called the @dfn{entry
2965 point}. You can use the @code{ENTRY} linker script command to set the
2966 entry point. The argument is a symbol name:
2971 There are several ways to set the entry point. The linker will set the
2972 entry point by trying each of the following methods in order, and
2973 stopping when one of them succeeds:
2976 the @samp{-e} @var{entry} command-line option;
2978 the @code{ENTRY(@var{symbol})} command in a linker script;
2980 the value of a target specific symbol, if it is defined; For many
2981 targets this is @code{start}, but PE and BeOS based systems for example
2982 check a list of possible entry symbols, matching the first one found.
2984 the address of the first byte of the @samp{.text} section, if present;
2986 The address @code{0}.
2990 @subsection Commands Dealing with Files
2991 @cindex linker script file commands
2992 Several linker script commands deal with files.
2995 @item INCLUDE @var{filename}
2996 @kindex INCLUDE @var{filename}
2997 @cindex including a linker script
2998 Include the linker script @var{filename} at this point. The file will
2999 be searched for in the current directory, and in any directory specified
3000 with the @option{-L} option. You can nest calls to @code{INCLUDE} up to
3003 You can place @code{INCLUDE} directives at the top level, in @code{MEMORY} or
3004 @code{SECTIONS} commands, or in output section descriptions.
3006 @item INPUT(@var{file}, @var{file}, @dots{})
3007 @itemx INPUT(@var{file} @var{file} @dots{})
3008 @kindex INPUT(@var{files})
3009 @cindex input files in linker scripts
3010 @cindex input object files in linker scripts
3011 @cindex linker script input object files
3012 The @code{INPUT} command directs the linker to include the named files
3013 in the link, as though they were named on the command line.
3015 For example, if you always want to include @file{subr.o} any time you do
3016 a link, but you can't be bothered to put it on every link command line,
3017 then you can put @samp{INPUT (subr.o)} in your linker script.
3019 In fact, if you like, you can list all of your input files in the linker
3020 script, and then invoke the linker with nothing but a @samp{-T} option.
3022 In case a @dfn{sysroot prefix} is configured, and the filename starts
3023 with the @samp{/} character, and the script being processed was
3024 located inside the @dfn{sysroot prefix}, the filename will be looked
3025 for in the @dfn{sysroot prefix}. Otherwise, the linker will try to
3026 open the file in the current directory. If it is not found, the
3027 linker will search through the archive library search path. See the
3028 description of @samp{-L} in @ref{Options,,Command Line Options}.
3030 If you use @samp{INPUT (-l@var{file})}, @command{ld} will transform the
3031 name to @code{lib@var{file}.a}, as with the command line argument
3034 When you use the @code{INPUT} command in an implicit linker script, the
3035 files will be included in the link at the point at which the linker
3036 script file is included. This can affect archive searching.
3038 @item GROUP(@var{file}, @var{file}, @dots{})
3039 @itemx GROUP(@var{file} @var{file} @dots{})
3040 @kindex GROUP(@var{files})
3041 @cindex grouping input files
3042 The @code{GROUP} command is like @code{INPUT}, except that the named
3043 files should all be archives, and they are searched repeatedly until no
3044 new undefined references are created. See the description of @samp{-(}
3045 in @ref{Options,,Command Line Options}.
3047 @item AS_NEEDED(@var{file}, @var{file}, @dots{})
3048 @itemx AS_NEEDED(@var{file} @var{file} @dots{})
3049 @kindex AS_NEEDED(@var{files})
3050 This construct can appear only inside of the @code{INPUT} or @code{GROUP}
3051 commands, among other filenames. The files listed will be handled
3052 as if they appear directly in the @code{INPUT} or @code{GROUP} commands,
3053 with the exception of ELF shared libraries, that will be added only
3054 when they are actually needed. This construct essentially enables
3055 @option{--as-needed} option for all the files listed inside of it
3056 and restores previous @option{--as-needed} resp. @option{--no-as-needed}
3059 @item OUTPUT(@var{filename})
3060 @kindex OUTPUT(@var{filename})
3061 @cindex output file name in linker script
3062 The @code{OUTPUT} command names the output file. Using
3063 @code{OUTPUT(@var{filename})} in the linker script is exactly like using
3064 @samp{-o @var{filename}} on the command line (@pxref{Options,,Command
3065 Line Options}). If both are used, the command line option takes
3068 You can use the @code{OUTPUT} command to define a default name for the
3069 output file other than the usual default of @file{a.out}.
3071 @item SEARCH_DIR(@var{path})
3072 @kindex SEARCH_DIR(@var{path})
3073 @cindex library search path in linker script
3074 @cindex archive search path in linker script
3075 @cindex search path in linker script
3076 The @code{SEARCH_DIR} command adds @var{path} to the list of paths where
3077 @command{ld} looks for archive libraries. Using
3078 @code{SEARCH_DIR(@var{path})} is exactly like using @samp{-L @var{path}}
3079 on the command line (@pxref{Options,,Command Line Options}). If both
3080 are used, then the linker will search both paths. Paths specified using
3081 the command line option are searched first.
3083 @item STARTUP(@var{filename})
3084 @kindex STARTUP(@var{filename})
3085 @cindex first input file
3086 The @code{STARTUP} command is just like the @code{INPUT} command, except
3087 that @var{filename} will become the first input file to be linked, as
3088 though it were specified first on the command line. This may be useful
3089 when using a system in which the entry point is always the start of the
3093 @ifclear SingleFormat
3094 @node Format Commands
3095 @subsection Commands Dealing with Object File Formats
3096 A couple of linker script commands deal with object file formats.
3099 @item OUTPUT_FORMAT(@var{bfdname})
3100 @itemx OUTPUT_FORMAT(@var{default}, @var{big}, @var{little})
3101 @kindex OUTPUT_FORMAT(@var{bfdname})
3102 @cindex output file format in linker script
3103 The @code{OUTPUT_FORMAT} command names the BFD format to use for the
3104 output file (@pxref{BFD}). Using @code{OUTPUT_FORMAT(@var{bfdname})} is
3105 exactly like using @samp{--oformat @var{bfdname}} on the command line
3106 (@pxref{Options,,Command Line Options}). If both are used, the command
3107 line option takes precedence.
3109 You can use @code{OUTPUT_FORMAT} with three arguments to use different
3110 formats based on the @samp{-EB} and @samp{-EL} command line options.
3111 This permits the linker script to set the output format based on the
3114 If neither @samp{-EB} nor @samp{-EL} are used, then the output format
3115 will be the first argument, @var{default}. If @samp{-EB} is used, the
3116 output format will be the second argument, @var{big}. If @samp{-EL} is
3117 used, the output format will be the third argument, @var{little}.
3119 For example, the default linker script for the MIPS ELF target uses this
3122 OUTPUT_FORMAT(elf32-bigmips, elf32-bigmips, elf32-littlemips)
3124 This says that the default format for the output file is
3125 @samp{elf32-bigmips}, but if the user uses the @samp{-EL} command line
3126 option, the output file will be created in the @samp{elf32-littlemips}
3129 @item TARGET(@var{bfdname})
3130 @kindex TARGET(@var{bfdname})
3131 @cindex input file format in linker script
3132 The @code{TARGET} command names the BFD format to use when reading input
3133 files. It affects subsequent @code{INPUT} and @code{GROUP} commands.
3134 This command is like using @samp{-b @var{bfdname}} on the command line
3135 (@pxref{Options,,Command Line Options}). If the @code{TARGET} command
3136 is used but @code{OUTPUT_FORMAT} is not, then the last @code{TARGET}
3137 command is also used to set the format for the output file. @xref{BFD}.
3142 @subsection Assign alias names to memory regions
3143 @kindex REGION_ALIAS(@var{alias}, @var{region})
3144 @cindex region alias
3145 @cindex region names
3147 Alias names can be added to existing memory regions created with the
3148 @ref{MEMORY} command. Each name corresponds to at most one memory region.
3151 REGION_ALIAS(@var{alias}, @var{region})
3154 The @code{REGION_ALIAS} function creates an alias name @var{alias} for the
3155 memory region @var{region}. This allows a flexible mapping of output sections
3156 to memory regions. An example follows.
3158 Suppose we have an application for embedded systems which come with various
3159 memory storage devices. All have a general purpose, volatile memory @code{RAM}
3160 that allows code execution or data storage. Some may have a read-only,
3161 non-volatile memory @code{ROM} that allows code execution and read-only data
3162 access. The last variant is a read-only, non-volatile memory @code{ROM2} with
3163 read-only data access and no code execution capability. We have four output
3168 @code{.text} program code;
3170 @code{.rodata} read-only data;
3172 @code{.data} read-write initialized data;
3174 @code{.bss} read-write zero initialized data.
3177 The goal is to provide a linker command file that contains a system independent
3178 part defining the output sections and a system dependent part mapping the
3179 output sections to the memory regions available on the system. Our embedded
3180 systems come with three different memory setups @code{A}, @code{B} and
3182 @multitable @columnfractions .25 .25 .25 .25
3183 @item Section @tab Variant A @tab Variant B @tab Variant C
3184 @item .text @tab RAM @tab ROM @tab ROM
3185 @item .rodata @tab RAM @tab ROM @tab ROM2
3186 @item .data @tab RAM @tab RAM/ROM @tab RAM/ROM2
3187 @item .bss @tab RAM @tab RAM @tab RAM
3189 The notation @code{RAM/ROM} or @code{RAM/ROM2} means that this section is
3190 loaded into region @code{ROM} or @code{ROM2} respectively. Please note that
3191 the load address of the @code{.data} section starts in all three variants at
3192 the end of the @code{.rodata} section.
3194 The base linker script that deals with the output sections follows. It
3195 includes the system dependent @code{linkcmds.memory} file that describes the
3198 INCLUDE linkcmds.memory
3211 .data : AT (rodata_end)
3216 data_size = SIZEOF(.data);
3217 data_load_start = LOADADDR(.data);
3225 Now we need three different @code{linkcmds.memory} files to define memory
3226 regions and alias names. The content of @code{linkcmds.memory} for the three
3227 variants @code{A}, @code{B} and @code{C}:
3230 Here everything goes into the @code{RAM}.
3234 RAM : ORIGIN = 0, LENGTH = 4M
3237 REGION_ALIAS("REGION_TEXT", RAM);
3238 REGION_ALIAS("REGION_RODATA", RAM);
3239 REGION_ALIAS("REGION_DATA", RAM);
3240 REGION_ALIAS("REGION_BSS", RAM);
3243 Program code and read-only data go into the @code{ROM}. Read-write data goes
3244 into the @code{RAM}. An image of the initialized data is loaded into the
3245 @code{ROM} and will be copied during system start into the @code{RAM}.
3249 ROM : ORIGIN = 0, LENGTH = 3M
3250 RAM : ORIGIN = 0x10000000, LENGTH = 1M
3253 REGION_ALIAS("REGION_TEXT", ROM);
3254 REGION_ALIAS("REGION_RODATA", ROM);
3255 REGION_ALIAS("REGION_DATA", RAM);
3256 REGION_ALIAS("REGION_BSS", RAM);
3259 Program code goes into the @code{ROM}. Read-only data goes into the
3260 @code{ROM2}. Read-write data goes into the @code{RAM}. An image of the
3261 initialized data is loaded into the @code{ROM2} and will be copied during
3262 system start into the @code{RAM}.
3266 ROM : ORIGIN = 0, LENGTH = 2M
3267 ROM2 : ORIGIN = 0x10000000, LENGTH = 1M
3268 RAM : ORIGIN = 0x20000000, LENGTH = 1M
3271 REGION_ALIAS("REGION_TEXT", ROM);
3272 REGION_ALIAS("REGION_RODATA", ROM2);
3273 REGION_ALIAS("REGION_DATA", RAM);
3274 REGION_ALIAS("REGION_BSS", RAM);
3278 It is possible to write a common system initialization routine to copy the
3279 @code{.data} section from @code{ROM} or @code{ROM2} into the @code{RAM} if
3284 extern char data_start [];
3285 extern char data_size [];
3286 extern char data_load_start [];
3288 void copy_data(void)
3290 if (data_start != data_load_start)
3292 memcpy(data_start, data_load_start, (size_t) data_size);
3297 @node Miscellaneous Commands
3298 @subsection Other Linker Script Commands
3299 There are a few other linker scripts commands.
3302 @item ASSERT(@var{exp}, @var{message})
3304 @cindex assertion in linker script
3305 Ensure that @var{exp} is non-zero. If it is zero, then exit the linker
3306 with an error code, and print @var{message}.
3308 @item EXTERN(@var{symbol} @var{symbol} @dots{})
3310 @cindex undefined symbol in linker script
3311 Force @var{symbol} to be entered in the output file as an undefined
3312 symbol. Doing this may, for example, trigger linking of additional
3313 modules from standard libraries. You may list several @var{symbol}s for
3314 each @code{EXTERN}, and you may use @code{EXTERN} multiple times. This
3315 command has the same effect as the @samp{-u} command-line option.
3317 @item FORCE_COMMON_ALLOCATION
3318 @kindex FORCE_COMMON_ALLOCATION
3319 @cindex common allocation in linker script
3320 This command has the same effect as the @samp{-d} command-line option:
3321 to make @command{ld} assign space to common symbols even if a relocatable
3322 output file is specified (@samp{-r}).
3324 @item INHIBIT_COMMON_ALLOCATION
3325 @kindex INHIBIT_COMMON_ALLOCATION
3326 @cindex common allocation in linker script
3327 This command has the same effect as the @samp{--no-define-common}
3328 command-line option: to make @code{ld} omit the assignment of addresses
3329 to common symbols even for a non-relocatable output file.
3331 @item INSERT [ AFTER | BEFORE ] @var{output_section}
3333 @cindex insert user script into default script
3334 This command is typically used in a script specified by @samp{-T} to
3335 augment the default @code{SECTIONS} with, for example, overlays. It
3336 inserts all prior linker script statements after (or before)
3337 @var{output_section}, and also causes @samp{-T} to not override the
3338 default linker script. The exact insertion point is as for orphan
3339 sections. @xref{Location Counter}. The insertion happens after the
3340 linker has mapped input sections to output sections. Prior to the
3341 insertion, since @samp{-T} scripts are parsed before the default
3342 linker script, statements in the @samp{-T} script occur before the
3343 default linker script statements in the internal linker representation
3344 of the script. In particular, input section assignments will be made
3345 to @samp{-T} output sections before those in the default script. Here
3346 is an example of how a @samp{-T} script using @code{INSERT} might look:
3353 .ov1 @{ ov1*(.text) @}
3354 .ov2 @{ ov2*(.text) @}
3360 @item NOCROSSREFS(@var{section} @var{section} @dots{})
3361 @kindex NOCROSSREFS(@var{sections})
3362 @cindex cross references
3363 This command may be used to tell @command{ld} to issue an error about any
3364 references among certain output sections.
3366 In certain types of programs, particularly on embedded systems when
3367 using overlays, when one section is loaded into memory, another section
3368 will not be. Any direct references between the two sections would be
3369 errors. For example, it would be an error if code in one section called
3370 a function defined in the other section.
3372 The @code{NOCROSSREFS} command takes a list of output section names. If
3373 @command{ld} detects any cross references between the sections, it reports
3374 an error and returns a non-zero exit status. Note that the
3375 @code{NOCROSSREFS} command uses output section names, not input section
3378 @ifclear SingleFormat
3379 @item OUTPUT_ARCH(@var{bfdarch})
3380 @kindex OUTPUT_ARCH(@var{bfdarch})
3381 @cindex machine architecture
3382 @cindex architecture
3383 Specify a particular output machine architecture. The argument is one
3384 of the names used by the BFD library (@pxref{BFD}). You can see the
3385 architecture of an object file by using the @code{objdump} program with
3386 the @samp{-f} option.
3389 @item LD_FEATURE(@var{string})
3390 @kindex LD_FEATURE(@var{string})
3391 This command may be used to modify @command{ld} behavior. If
3392 @var{string} is @code{"SANE_EXPR"} then absolute symbols and numbers
3393 in a script are simply treated as numbers everywhere.
3394 @xref{Expression Section}.
3398 @section Assigning Values to Symbols
3399 @cindex assignment in scripts
3400 @cindex symbol definition, scripts
3401 @cindex variables, defining
3402 You may assign a value to a symbol in a linker script. This will define
3403 the symbol and place it into the symbol table with a global scope.
3406 * Simple Assignments:: Simple Assignments
3408 * PROVIDE_HIDDEN:: PROVIDE_HIDDEN
3409 * Source Code Reference:: How to use a linker script defined symbol in source code
3412 @node Simple Assignments
3413 @subsection Simple Assignments
3415 You may assign to a symbol using any of the C assignment operators:
3418 @item @var{symbol} = @var{expression} ;
3419 @itemx @var{symbol} += @var{expression} ;
3420 @itemx @var{symbol} -= @var{expression} ;
3421 @itemx @var{symbol} *= @var{expression} ;
3422 @itemx @var{symbol} /= @var{expression} ;
3423 @itemx @var{symbol} <<= @var{expression} ;
3424 @itemx @var{symbol} >>= @var{expression} ;
3425 @itemx @var{symbol} &= @var{expression} ;
3426 @itemx @var{symbol} |= @var{expression} ;
3429 The first case will define @var{symbol} to the value of
3430 @var{expression}. In the other cases, @var{symbol} must already be
3431 defined, and the value will be adjusted accordingly.
3433 The special symbol name @samp{.} indicates the location counter. You
3434 may only use this within a @code{SECTIONS} command. @xref{Location Counter}.
3436 The semicolon after @var{expression} is required.
3438 Expressions are defined below; see @ref{Expressions}.
3440 You may write symbol assignments as commands in their own right, or as
3441 statements within a @code{SECTIONS} command, or as part of an output
3442 section description in a @code{SECTIONS} command.
3444 The section of the symbol will be set from the section of the
3445 expression; for more information, see @ref{Expression Section}.
3447 Here is an example showing the three different places that symbol
3448 assignments may be used:
3459 _bdata = (. + 3) & ~ 3;
3460 .data : @{ *(.data) @}
3464 In this example, the symbol @samp{floating_point} will be defined as
3465 zero. The symbol @samp{_etext} will be defined as the address following
3466 the last @samp{.text} input section. The symbol @samp{_bdata} will be
3467 defined as the address following the @samp{.text} output section aligned
3468 upward to a 4 byte boundary.
3473 In some cases, it is desirable for a linker script to define a symbol
3474 only if it is referenced and is not defined by any object included in
3475 the link. For example, traditional linkers defined the symbol
3476 @samp{etext}. However, ANSI C requires that the user be able to use
3477 @samp{etext} as a function name without encountering an error. The
3478 @code{PROVIDE} keyword may be used to define a symbol, such as
3479 @samp{etext}, only if it is referenced but not defined. The syntax is
3480 @code{PROVIDE(@var{symbol} = @var{expression})}.
3482 Here is an example of using @code{PROVIDE} to define @samp{etext}:
3495 In this example, if the program defines @samp{_etext} (with a leading
3496 underscore), the linker will give a multiple definition error. If, on
3497 the other hand, the program defines @samp{etext} (with no leading
3498 underscore), the linker will silently use the definition in the program.
3499 If the program references @samp{etext} but does not define it, the
3500 linker will use the definition in the linker script.
3502 @node PROVIDE_HIDDEN
3503 @subsection PROVIDE_HIDDEN
3504 @cindex PROVIDE_HIDDEN
3505 Similar to @code{PROVIDE}. For ELF targeted ports, the symbol will be
3506 hidden and won't be exported.
3508 @node Source Code Reference
3509 @subsection Source Code Reference
3511 Accessing a linker script defined variable from source code is not
3512 intuitive. In particular a linker script symbol is not equivalent to
3513 a variable declaration in a high level language, it is instead a
3514 symbol that does not have a value.
3516 Before going further, it is important to note that compilers often
3517 transform names in the source code into different names when they are
3518 stored in the symbol table. For example, Fortran compilers commonly
3519 prepend or append an underscore, and C++ performs extensive @samp{name
3520 mangling}. Therefore there might be a discrepancy between the name
3521 of a variable as it is used in source code and the name of the same
3522 variable as it is defined in a linker script. For example in C a
3523 linker script variable might be referred to as:
3529 But in the linker script it might be defined as:
3535 In the remaining examples however it is assumed that no name
3536 transformation has taken place.
3538 When a symbol is declared in a high level language such as C, two
3539 things happen. The first is that the compiler reserves enough space
3540 in the program's memory to hold the @emph{value} of the symbol. The
3541 second is that the compiler creates an entry in the program's symbol
3542 table which holds the symbol's @emph{address}. ie the symbol table
3543 contains the address of the block of memory holding the symbol's
3544 value. So for example the following C declaration, at file scope:
3550 creates a entry called @samp{foo} in the symbol table. This entry
3551 holds the address of an @samp{int} sized block of memory where the
3552 number 1000 is initially stored.
3554 When a program references a symbol the compiler generates code that
3555 first accesses the symbol table to find the address of the symbol's
3556 memory block and then code to read the value from that memory block.
3563 looks up the symbol @samp{foo} in the symbol table, gets the address
3564 associated with this symbol and then writes the value 1 into that
3571 looks up the symbol @samp{foo} in the symbol table, gets it address
3572 and then copies this address into the block of memory associated with
3573 the variable @samp{a}.
3575 Linker scripts symbol declarations, by contrast, create an entry in
3576 the symbol table but do not assign any memory to them. Thus they are
3577 an address without a value. So for example the linker script definition:
3583 creates an entry in the symbol table called @samp{foo} which holds
3584 the address of memory location 1000, but nothing special is stored at
3585 address 1000. This means that you cannot access the @emph{value} of a
3586 linker script defined symbol - it has no value - all you can do is
3587 access the @emph{address} of a linker script defined symbol.
3589 Hence when you are using a linker script defined symbol in source code
3590 you should always take the address of the symbol, and never attempt to
3591 use its value. For example suppose you want to copy the contents of a
3592 section of memory called .ROM into a section called .FLASH and the
3593 linker script contains these declarations:
3597 start_of_ROM = .ROM;
3598 end_of_ROM = .ROM + sizeof (.ROM) - 1;
3599 start_of_FLASH = .FLASH;
3603 Then the C source code to perform the copy would be:
3607 extern char start_of_ROM, end_of_ROM, start_of_FLASH;
3609 memcpy (& start_of_FLASH, & start_of_ROM, & end_of_ROM - & start_of_ROM);
3613 Note the use of the @samp{&} operators. These are correct.
3616 @section SECTIONS Command
3618 The @code{SECTIONS} command tells the linker how to map input sections
3619 into output sections, and how to place the output sections in memory.
3621 The format of the @code{SECTIONS} command is:
3625 @var{sections-command}
3626 @var{sections-command}
3631 Each @var{sections-command} may of be one of the following:
3635 an @code{ENTRY} command (@pxref{Entry Point,,Entry command})
3637 a symbol assignment (@pxref{Assignments})
3639 an output section description
3641 an overlay description
3644 The @code{ENTRY} command and symbol assignments are permitted inside the
3645 @code{SECTIONS} command for convenience in using the location counter in
3646 those commands. This can also make the linker script easier to
3647 understand because you can use those commands at meaningful points in
3648 the layout of the output file.
3650 Output section descriptions and overlay descriptions are described
3653 If you do not use a @code{SECTIONS} command in your linker script, the
3654 linker will place each input section into an identically named output
3655 section in the order that the sections are first encountered in the
3656 input files. If all input sections are present in the first file, for
3657 example, the order of sections in the output file will match the order
3658 in the first input file. The first section will be at address zero.
3661 * Output Section Description:: Output section description
3662 * Output Section Name:: Output section name
3663 * Output Section Address:: Output section address
3664 * Input Section:: Input section description
3665 * Output Section Data:: Output section data
3666 * Output Section Keywords:: Output section keywords
3667 * Output Section Discarding:: Output section discarding
3668 * Output Section Attributes:: Output section attributes
3669 * Overlay Description:: Overlay description
3672 @node Output Section Description
3673 @subsection Output Section Description
3674 The full description of an output section looks like this:
3677 @var{section} [@var{address}] [(@var{type})] :
3679 [ALIGN(@var{section_align})]
3680 [SUBALIGN(@var{subsection_align})]
3683 @var{output-section-command}
3684 @var{output-section-command}
3686 @} [>@var{region}] [AT>@var{lma_region}] [:@var{phdr} :@var{phdr} @dots{}] [=@var{fillexp}]
3690 Most output sections do not use most of the optional section attributes.
3692 The whitespace around @var{section} is required, so that the section
3693 name is unambiguous. The colon and the curly braces are also required.
3694 The line breaks and other white space are optional.
3696 Each @var{output-section-command} may be one of the following:
3700 a symbol assignment (@pxref{Assignments})
3702 an input section description (@pxref{Input Section})
3704 data values to include directly (@pxref{Output Section Data})
3706 a special output section keyword (@pxref{Output Section Keywords})
3709 @node Output Section Name
3710 @subsection Output Section Name
3711 @cindex name, section
3712 @cindex section name
3713 The name of the output section is @var{section}. @var{section} must
3714 meet the constraints of your output format. In formats which only
3715 support a limited number of sections, such as @code{a.out}, the name
3716 must be one of the names supported by the format (@code{a.out}, for
3717 example, allows only @samp{.text}, @samp{.data} or @samp{.bss}). If the
3718 output format supports any number of sections, but with numbers and not
3719 names (as is the case for Oasys), the name should be supplied as a
3720 quoted numeric string. A section name may consist of any sequence of
3721 characters, but a name which contains any unusual characters such as
3722 commas must be quoted.
3724 The output section name @samp{/DISCARD/} is special; @ref{Output Section
3727 @node Output Section Address
3728 @subsection Output Section Address
3729 @cindex address, section
3730 @cindex section address
3731 The @var{address} is an expression for the VMA (the virtual memory
3732 address) of the output section. This address is optional, but if it
3733 is provided then the output address will be set exactly as specified.
3735 If the output address is not specified then one will be chosen for the
3736 section, based on the heuristic below. This address will be adjusted
3737 to fit the alignment requirement of the output section. The
3738 alignment requirement is the strictest alignment of any input section
3739 contained within the output section.
3741 The output section address heuristic is as follows:
3745 If an output memory @var{region} is set for the section then it
3746 is added to this region and its address will be the next free address
3750 If the MEMORY command has been used to create a list of memory
3751 regions then the first region which has attributes compatible with the
3752 section is selected to contain it. The section's output address will
3753 be the next free address in that region; @ref{MEMORY}.
3756 If no memory regions were specified, or none match the section then
3757 the output address will be based on the current value of the location
3765 .text . : @{ *(.text) @}
3772 .text : @{ *(.text) @}
3776 are subtly different. The first will set the address of the
3777 @samp{.text} output section to the current value of the location
3778 counter. The second will set it to the current value of the location
3779 counter aligned to the strictest alignment of any of the @samp{.text}
3782 The @var{address} may be an arbitrary expression; @ref{Expressions}.
3783 For example, if you want to align the section on a 0x10 byte boundary,
3784 so that the lowest four bits of the section address are zero, you could
3785 do something like this:
3787 .text ALIGN(0x10) : @{ *(.text) @}
3790 This works because @code{ALIGN} returns the current location counter
3791 aligned upward to the specified value.
3793 Specifying @var{address} for a section will change the value of the
3794 location counter, provided that the section is non-empty. (Empty
3795 sections are ignored).
3798 @subsection Input Section Description
3799 @cindex input sections
3800 @cindex mapping input sections to output sections
3801 The most common output section command is an input section description.
3803 The input section description is the most basic linker script operation.
3804 You use output sections to tell the linker how to lay out your program
3805 in memory. You use input section descriptions to tell the linker how to
3806 map the input files into your memory layout.
3809 * Input Section Basics:: Input section basics
3810 * Input Section Wildcards:: Input section wildcard patterns
3811 * Input Section Common:: Input section for common symbols
3812 * Input Section Keep:: Input section and garbage collection
3813 * Input Section Example:: Input section example
3816 @node Input Section Basics
3817 @subsubsection Input Section Basics
3818 @cindex input section basics
3819 An input section description consists of a file name optionally followed
3820 by a list of section names in parentheses.
3822 The file name and the section name may be wildcard patterns, which we
3823 describe further below (@pxref{Input Section Wildcards}).
3825 The most common input section description is to include all input
3826 sections with a particular name in the output section. For example, to
3827 include all input @samp{.text} sections, you would write:
3832 Here the @samp{*} is a wildcard which matches any file name. To exclude a list
3833 of files from matching the file name wildcard, EXCLUDE_FILE may be used to
3834 match all files except the ones specified in the EXCLUDE_FILE list. For
3837 *(EXCLUDE_FILE (*crtend.o *otherfile.o) .ctors)
3839 will cause all .ctors sections from all files except @file{crtend.o} and
3840 @file{otherfile.o} to be included.
3842 There are two ways to include more than one section:
3848 The difference between these is the order in which the @samp{.text} and
3849 @samp{.rdata} input sections will appear in the output section. In the
3850 first example, they will be intermingled, appearing in the same order as
3851 they are found in the linker input. In the second example, all
3852 @samp{.text} input sections will appear first, followed by all
3853 @samp{.rdata} input sections.
3855 You can specify a file name to include sections from a particular file.
3856 You would do this if one or more of your files contain special data that
3857 needs to be at a particular location in memory. For example:
3862 To refine the sections that are included based on the section flags
3863 of an input section, INPUT_SECTION_FLAGS may be used.
3865 Here is a simple example for using Section header flags for ELF sections:
3870 .text : @{ INPUT_SECTION_FLAGS (SHF_MERGE & SHF_STRINGS) *(.text) @}
3871 .text2 : @{ INPUT_SECTION_FLAGS (!SHF_WRITE) *(.text) @}
3876 In this example, the output section @samp{.text} will be comprised of any
3877 input section matching the name *(.text) whose section header flags
3878 @code{SHF_MERGE} and @code{SHF_STRINGS} are set. The output section
3879 @samp{.text2} will be comprised of any input section matching the name *(.text)
3880 whose section header flag @code{SHF_WRITE} is clear.
3882 You can also specify files within archives by writing a pattern
3883 matching the archive, a colon, then the pattern matching the file,
3884 with no whitespace around the colon.
3888 matches file within archive
3890 matches the whole archive
3892 matches file but not one in an archive
3895 Either one or both of @samp{archive} and @samp{file} can contain shell
3896 wildcards. On DOS based file systems, the linker will assume that a
3897 single letter followed by a colon is a drive specifier, so
3898 @samp{c:myfile.o} is a simple file specification, not @samp{myfile.o}
3899 within an archive called @samp{c}. @samp{archive:file} filespecs may
3900 also be used within an @code{EXCLUDE_FILE} list, but may not appear in
3901 other linker script contexts. For instance, you cannot extract a file
3902 from an archive by using @samp{archive:file} in an @code{INPUT}
3905 If you use a file name without a list of sections, then all sections in
3906 the input file will be included in the output section. This is not
3907 commonly done, but it may by useful on occasion. For example:
3912 When you use a file name which is not an @samp{archive:file} specifier
3913 and does not contain any wild card
3914 characters, the linker will first see if you also specified the file
3915 name on the linker command line or in an @code{INPUT} command. If you
3916 did not, the linker will attempt to open the file as an input file, as
3917 though it appeared on the command line. Note that this differs from an
3918 @code{INPUT} command, because the linker will not search for the file in
3919 the archive search path.
3921 @node Input Section Wildcards
3922 @subsubsection Input Section Wildcard Patterns
3923 @cindex input section wildcards
3924 @cindex wildcard file name patterns
3925 @cindex file name wildcard patterns
3926 @cindex section name wildcard patterns
3927 In an input section description, either the file name or the section
3928 name or both may be wildcard patterns.
3930 The file name of @samp{*} seen in many examples is a simple wildcard
3931 pattern for the file name.
3933 The wildcard patterns are like those used by the Unix shell.
3937 matches any number of characters
3939 matches any single character
3941 matches a single instance of any of the @var{chars}; the @samp{-}
3942 character may be used to specify a range of characters, as in
3943 @samp{[a-z]} to match any lower case letter
3945 quotes the following character
3948 When a file name is matched with a wildcard, the wildcard characters
3949 will not match a @samp{/} character (used to separate directory names on
3950 Unix). A pattern consisting of a single @samp{*} character is an
3951 exception; it will always match any file name, whether it contains a
3952 @samp{/} or not. In a section name, the wildcard characters will match
3953 a @samp{/} character.
3955 File name wildcard patterns only match files which are explicitly
3956 specified on the command line or in an @code{INPUT} command. The linker
3957 does not search directories to expand wildcards.
3959 If a file name matches more than one wildcard pattern, or if a file name
3960 appears explicitly and is also matched by a wildcard pattern, the linker
3961 will use the first match in the linker script. For example, this
3962 sequence of input section descriptions is probably in error, because the
3963 @file{data.o} rule will not be used:
3965 .data : @{ *(.data) @}
3966 .data1 : @{ data.o(.data) @}
3969 @cindex SORT_BY_NAME
3970 Normally, the linker will place files and sections matched by wildcards
3971 in the order in which they are seen during the link. You can change
3972 this by using the @code{SORT_BY_NAME} keyword, which appears before a wildcard
3973 pattern in parentheses (e.g., @code{SORT_BY_NAME(.text*)}). When the
3974 @code{SORT_BY_NAME} keyword is used, the linker will sort the files or sections
3975 into ascending order by name before placing them in the output file.
3977 @cindex SORT_BY_ALIGNMENT
3978 @code{SORT_BY_ALIGNMENT} is very similar to @code{SORT_BY_NAME}. The
3979 difference is @code{SORT_BY_ALIGNMENT} will sort sections into
3980 ascending order by alignment before placing them in the output file.
3982 @cindex SORT_BY_INIT_PRIORITY
3983 @code{SORT_BY_INIT_PRIORITY} is very similar to @code{SORT_BY_NAME}. The
3984 difference is @code{SORT_BY_INIT_PRIORITY} will sort sections into
3985 ascending order by numerical value of the GCC init_priority attribute
3986 encoded in the section name before placing them in the output file.
3989 @code{SORT} is an alias for @code{SORT_BY_NAME}.
3991 When there are nested section sorting commands in linker script, there
3992 can be at most 1 level of nesting for section sorting commands.
3996 @code{SORT_BY_NAME} (@code{SORT_BY_ALIGNMENT} (wildcard section pattern)).
3997 It will sort the input sections by name first, then by alignment if 2
3998 sections have the same name.
4000 @code{SORT_BY_ALIGNMENT} (@code{SORT_BY_NAME} (wildcard section pattern)).
4001 It will sort the input sections by alignment first, then by name if 2
4002 sections have the same alignment.
4004 @code{SORT_BY_NAME} (@code{SORT_BY_NAME} (wildcard section pattern)) is
4005 treated the same as @code{SORT_BY_NAME} (wildcard section pattern).
4007 @code{SORT_BY_ALIGNMENT} (@code{SORT_BY_ALIGNMENT} (wildcard section pattern))
4008 is treated the same as @code{SORT_BY_ALIGNMENT} (wildcard section pattern).
4010 All other nested section sorting commands are invalid.
4013 When both command line section sorting option and linker script
4014 section sorting command are used, section sorting command always
4015 takes precedence over the command line option.
4017 If the section sorting command in linker script isn't nested, the
4018 command line option will make the section sorting command to be
4019 treated as nested sorting command.
4023 @code{SORT_BY_NAME} (wildcard section pattern ) with
4024 @option{--sort-sections alignment} is equivalent to
4025 @code{SORT_BY_NAME} (@code{SORT_BY_ALIGNMENT} (wildcard section pattern)).
4027 @code{SORT_BY_ALIGNMENT} (wildcard section pattern) with
4028 @option{--sort-section name} is equivalent to
4029 @code{SORT_BY_ALIGNMENT} (@code{SORT_BY_NAME} (wildcard section pattern)).
4032 If the section sorting command in linker script is nested, the
4033 command line option will be ignored.
4035 If you ever get confused about where input sections are going, use the
4036 @samp{-M} linker option to generate a map file. The map file shows
4037 precisely how input sections are mapped to output sections.
4039 This example shows how wildcard patterns might be used to partition
4040 files. This linker script directs the linker to place all @samp{.text}
4041 sections in @samp{.text} and all @samp{.bss} sections in @samp{.bss}.
4042 The linker will place the @samp{.data} section from all files beginning
4043 with an upper case character in @samp{.DATA}; for all other files, the
4044 linker will place the @samp{.data} section in @samp{.data}.
4048 .text : @{ *(.text) @}
4049 .DATA : @{ [A-Z]*(.data) @}
4050 .data : @{ *(.data) @}
4051 .bss : @{ *(.bss) @}
4056 @node Input Section Common
4057 @subsubsection Input Section for Common Symbols
4058 @cindex common symbol placement
4059 @cindex uninitialized data placement
4060 A special notation is needed for common symbols, because in many object
4061 file formats common symbols do not have a particular input section. The
4062 linker treats common symbols as though they are in an input section
4063 named @samp{COMMON}.
4065 You may use file names with the @samp{COMMON} section just as with any
4066 other input sections. You can use this to place common symbols from a
4067 particular input file in one section while common symbols from other
4068 input files are placed in another section.
4070 In most cases, common symbols in input files will be placed in the
4071 @samp{.bss} section in the output file. For example:
4073 .bss @{ *(.bss) *(COMMON) @}
4076 @cindex scommon section
4077 @cindex small common symbols
4078 Some object file formats have more than one type of common symbol. For
4079 example, the MIPS ELF object file format distinguishes standard common
4080 symbols and small common symbols. In this case, the linker will use a
4081 different special section name for other types of common symbols. In
4082 the case of MIPS ELF, the linker uses @samp{COMMON} for standard common
4083 symbols and @samp{.scommon} for small common symbols. This permits you
4084 to map the different types of common symbols into memory at different
4088 You will sometimes see @samp{[COMMON]} in old linker scripts. This
4089 notation is now considered obsolete. It is equivalent to
4092 @node Input Section Keep
4093 @subsubsection Input Section and Garbage Collection
4095 @cindex garbage collection
4096 When link-time garbage collection is in use (@samp{--gc-sections}),
4097 it is often useful to mark sections that should not be eliminated.
4098 This is accomplished by surrounding an input section's wildcard entry
4099 with @code{KEEP()}, as in @code{KEEP(*(.init))} or
4100 @code{KEEP(SORT_BY_NAME(*)(.ctors))}.
4102 @node Input Section Example
4103 @subsubsection Input Section Example
4104 The following example is a complete linker script. It tells the linker
4105 to read all of the sections from file @file{all.o} and place them at the
4106 start of output section @samp{outputa} which starts at location
4107 @samp{0x10000}. All of section @samp{.input1} from file @file{foo.o}
4108 follows immediately, in the same output section. All of section
4109 @samp{.input2} from @file{foo.o} goes into output section
4110 @samp{outputb}, followed by section @samp{.input1} from @file{foo1.o}.
4111 All of the remaining @samp{.input1} and @samp{.input2} sections from any
4112 files are written to output section @samp{outputc}.
4140 @node Output Section Data
4141 @subsection Output Section Data
4143 @cindex section data
4144 @cindex output section data
4145 @kindex BYTE(@var{expression})
4146 @kindex SHORT(@var{expression})
4147 @kindex LONG(@var{expression})
4148 @kindex QUAD(@var{expression})
4149 @kindex SQUAD(@var{expression})
4150 You can include explicit bytes of data in an output section by using
4151 @code{BYTE}, @code{SHORT}, @code{LONG}, @code{QUAD}, or @code{SQUAD} as
4152 an output section command. Each keyword is followed by an expression in
4153 parentheses providing the value to store (@pxref{Expressions}). The
4154 value of the expression is stored at the current value of the location
4157 The @code{BYTE}, @code{SHORT}, @code{LONG}, and @code{QUAD} commands
4158 store one, two, four, and eight bytes (respectively). After storing the
4159 bytes, the location counter is incremented by the number of bytes
4162 For example, this will store the byte 1 followed by the four byte value
4163 of the symbol @samp{addr}:
4169 When using a 64 bit host or target, @code{QUAD} and @code{SQUAD} are the
4170 same; they both store an 8 byte, or 64 bit, value. When both host and
4171 target are 32 bits, an expression is computed as 32 bits. In this case
4172 @code{QUAD} stores a 32 bit value zero extended to 64 bits, and
4173 @code{SQUAD} stores a 32 bit value sign extended to 64 bits.
4175 If the object file format of the output file has an explicit endianness,
4176 which is the normal case, the value will be stored in that endianness.
4177 When the object file format does not have an explicit endianness, as is
4178 true of, for example, S-records, the value will be stored in the
4179 endianness of the first input object file.
4181 Note---these commands only work inside a section description and not
4182 between them, so the following will produce an error from the linker:
4184 SECTIONS @{@ .text : @{@ *(.text) @}@ LONG(1) .data : @{@ *(.data) @}@ @}@
4186 whereas this will work:
4188 SECTIONS @{@ .text : @{@ *(.text) ; LONG(1) @}@ .data : @{@ *(.data) @}@ @}@
4191 @kindex FILL(@var{expression})
4192 @cindex holes, filling
4193 @cindex unspecified memory
4194 You may use the @code{FILL} command to set the fill pattern for the
4195 current section. It is followed by an expression in parentheses. Any
4196 otherwise unspecified regions of memory within the section (for example,
4197 gaps left due to the required alignment of input sections) are filled
4198 with the value of the expression, repeated as
4199 necessary. A @code{FILL} statement covers memory locations after the
4200 point at which it occurs in the section definition; by including more
4201 than one @code{FILL} statement, you can have different fill patterns in
4202 different parts of an output section.
4204 This example shows how to fill unspecified regions of memory with the
4210 The @code{FILL} command is similar to the @samp{=@var{fillexp}} output
4211 section attribute, but it only affects the
4212 part of the section following the @code{FILL} command, rather than the
4213 entire section. If both are used, the @code{FILL} command takes
4214 precedence. @xref{Output Section Fill}, for details on the fill
4217 @node Output Section Keywords
4218 @subsection Output Section Keywords
4219 There are a couple of keywords which can appear as output section
4223 @kindex CREATE_OBJECT_SYMBOLS
4224 @cindex input filename symbols
4225 @cindex filename symbols
4226 @item CREATE_OBJECT_SYMBOLS
4227 The command tells the linker to create a symbol for each input file.
4228 The name of each symbol will be the name of the corresponding input
4229 file. The section of each symbol will be the output section in which
4230 the @code{CREATE_OBJECT_SYMBOLS} command appears.
4232 This is conventional for the a.out object file format. It is not
4233 normally used for any other object file format.
4235 @kindex CONSTRUCTORS
4236 @cindex C++ constructors, arranging in link
4237 @cindex constructors, arranging in link
4239 When linking using the a.out object file format, the linker uses an
4240 unusual set construct to support C++ global constructors and
4241 destructors. When linking object file formats which do not support
4242 arbitrary sections, such as ECOFF and XCOFF, the linker will
4243 automatically recognize C++ global constructors and destructors by name.
4244 For these object file formats, the @code{CONSTRUCTORS} command tells the
4245 linker to place constructor information in the output section where the
4246 @code{CONSTRUCTORS} command appears. The @code{CONSTRUCTORS} command is
4247 ignored for other object file formats.
4249 The symbol @w{@code{__CTOR_LIST__}} marks the start of the global
4250 constructors, and the symbol @w{@code{__CTOR_END__}} marks the end.
4251 Similarly, @w{@code{__DTOR_LIST__}} and @w{@code{__DTOR_END__}} mark
4252 the start and end of the global destructors. The
4253 first word in the list is the number of entries, followed by the address
4254 of each constructor or destructor, followed by a zero word. The
4255 compiler must arrange to actually run the code. For these object file
4256 formats @sc{gnu} C++ normally calls constructors from a subroutine
4257 @code{__main}; a call to @code{__main} is automatically inserted into
4258 the startup code for @code{main}. @sc{gnu} C++ normally runs
4259 destructors either by using @code{atexit}, or directly from the function
4262 For object file formats such as @code{COFF} or @code{ELF} which support
4263 arbitrary section names, @sc{gnu} C++ will normally arrange to put the
4264 addresses of global constructors and destructors into the @code{.ctors}
4265 and @code{.dtors} sections. Placing the following sequence into your
4266 linker script will build the sort of table which the @sc{gnu} C++
4267 runtime code expects to see.
4271 LONG((__CTOR_END__ - __CTOR_LIST__) / 4 - 2)
4276 LONG((__DTOR_END__ - __DTOR_LIST__) / 4 - 2)
4282 If you are using the @sc{gnu} C++ support for initialization priority,
4283 which provides some control over the order in which global constructors
4284 are run, you must sort the constructors at link time to ensure that they
4285 are executed in the correct order. When using the @code{CONSTRUCTORS}
4286 command, use @samp{SORT_BY_NAME(CONSTRUCTORS)} instead. When using the
4287 @code{.ctors} and @code{.dtors} sections, use @samp{*(SORT_BY_NAME(.ctors))} and
4288 @samp{*(SORT_BY_NAME(.dtors))} instead of just @samp{*(.ctors)} and
4291 Normally the compiler and linker will handle these issues automatically,
4292 and you will not need to concern yourself with them. However, you may
4293 need to consider this if you are using C++ and writing your own linker
4298 @node Output Section Discarding
4299 @subsection Output Section Discarding
4300 @cindex discarding sections
4301 @cindex sections, discarding
4302 @cindex removing sections
4303 The linker will not create output sections with no contents. This is
4304 for convenience when referring to input sections that may or may not
4305 be present in any of the input files. For example:
4307 .foo : @{ *(.foo) @}
4310 will only create a @samp{.foo} section in the output file if there is a
4311 @samp{.foo} section in at least one input file, and if the input
4312 sections are not all empty. Other link script directives that allocate
4313 space in an output section will also create the output section.
4315 The linker will ignore address assignments (@pxref{Output Section Address})
4316 on discarded output sections, except when the linker script defines
4317 symbols in the output section. In that case the linker will obey
4318 the address assignments, possibly advancing dot even though the
4319 section is discarded.
4322 The special output section name @samp{/DISCARD/} may be used to discard
4323 input sections. Any input sections which are assigned to an output
4324 section named @samp{/DISCARD/} are not included in the output file.
4326 @node Output Section Attributes
4327 @subsection Output Section Attributes
4328 @cindex output section attributes
4329 We showed above that the full description of an output section looked
4334 @var{section} [@var{address}] [(@var{type})] :
4336 [ALIGN(@var{section_align})]
4337 [SUBALIGN(@var{subsection_align})]
4340 @var{output-section-command}
4341 @var{output-section-command}
4343 @} [>@var{region}] [AT>@var{lma_region}] [:@var{phdr} :@var{phdr} @dots{}] [=@var{fillexp}]
4347 We've already described @var{section}, @var{address}, and
4348 @var{output-section-command}. In this section we will describe the
4349 remaining section attributes.
4352 * Output Section Type:: Output section type
4353 * Output Section LMA:: Output section LMA
4354 * Forced Output Alignment:: Forced Output Alignment
4355 * Forced Input Alignment:: Forced Input Alignment
4356 * Output Section Constraint:: Output section constraint
4357 * Output Section Region:: Output section region
4358 * Output Section Phdr:: Output section phdr
4359 * Output Section Fill:: Output section fill
4362 @node Output Section Type
4363 @subsubsection Output Section Type
4364 Each output section may have a type. The type is a keyword in
4365 parentheses. The following types are defined:
4369 The section should be marked as not loadable, so that it will not be
4370 loaded into memory when the program is run.
4375 These type names are supported for backward compatibility, and are
4376 rarely used. They all have the same effect: the section should be
4377 marked as not allocatable, so that no memory is allocated for the
4378 section when the program is run.
4382 @cindex prevent unnecessary loading
4383 @cindex loading, preventing
4384 The linker normally sets the attributes of an output section based on
4385 the input sections which map into it. You can override this by using
4386 the section type. For example, in the script sample below, the
4387 @samp{ROM} section is addressed at memory location @samp{0} and does not
4388 need to be loaded when the program is run.
4392 ROM 0 (NOLOAD) : @{ @dots{} @}
4398 @node Output Section LMA
4399 @subsubsection Output Section LMA
4400 @kindex AT>@var{lma_region}
4401 @kindex AT(@var{lma})
4402 @cindex load address
4403 @cindex section load address
4404 Every section has a virtual address (VMA) and a load address (LMA); see
4405 @ref{Basic Script Concepts}. The virtual address is specified by the
4406 @pxref{Output Section Address} described earlier. The load address is
4407 specified by the @code{AT} or @code{AT>} keywords. Specifying a load
4408 address is optional.
4410 The @code{AT} keyword takes an expression as an argument. This
4411 specifies the exact load address of the section. The @code{AT>} keyword
4412 takes the name of a memory region as an argument. @xref{MEMORY}. The
4413 load address of the section is set to the next free address in the
4414 region, aligned to the section's alignment requirements.
4416 If neither @code{AT} nor @code{AT>} is specified for an allocatable
4417 section, the linker will use the following heuristic to determine the
4422 If the section has a specific VMA address, then this is used as
4423 the LMA address as well.
4426 If the section is not allocatable then its LMA is set to its VMA.
4429 Otherwise if a memory region can be found that is compatible
4430 with the current section, and this region contains at least one
4431 section, then the LMA is set so the difference between the
4432 VMA and LMA is the same as the difference between the VMA and LMA of
4433 the last section in the located region.
4436 If no memory regions have been declared then a default region
4437 that covers the entire address space is used in the previous step.
4440 If no suitable region could be found, or there was no previous
4441 section then the LMA is set equal to the VMA.
4444 @cindex ROM initialized data
4445 @cindex initialized data in ROM
4446 This feature is designed to make it easy to build a ROM image. For
4447 example, the following linker script creates three output sections: one
4448 called @samp{.text}, which starts at @code{0x1000}, one called
4449 @samp{.mdata}, which is loaded at the end of the @samp{.text} section
4450 even though its VMA is @code{0x2000}, and one called @samp{.bss} to hold
4451 uninitialized data at address @code{0x3000}. The symbol @code{_data} is
4452 defined with the value @code{0x2000}, which shows that the location
4453 counter holds the VMA value, not the LMA value.
4459 .text 0x1000 : @{ *(.text) _etext = . ; @}
4461 AT ( ADDR (.text) + SIZEOF (.text) )
4462 @{ _data = . ; *(.data); _edata = . ; @}
4464 @{ _bstart = . ; *(.bss) *(COMMON) ; _bend = . ;@}
4469 The run-time initialization code for use with a program generated with
4470 this linker script would include something like the following, to copy
4471 the initialized data from the ROM image to its runtime address. Notice
4472 how this code takes advantage of the symbols defined by the linker
4477 extern char _etext, _data, _edata, _bstart, _bend;
4478 char *src = &_etext;
4481 /* ROM has data at end of text; copy it. */
4482 while (dst < &_edata)
4486 for (dst = &_bstart; dst< &_bend; dst++)
4491 @node Forced Output Alignment
4492 @subsubsection Forced Output Alignment
4493 @kindex ALIGN(@var{section_align})
4494 @cindex forcing output section alignment
4495 @cindex output section alignment
4496 You can increase an output section's alignment by using ALIGN.
4498 @node Forced Input Alignment
4499 @subsubsection Forced Input Alignment
4500 @kindex SUBALIGN(@var{subsection_align})
4501 @cindex forcing input section alignment
4502 @cindex input section alignment
4503 You can force input section alignment within an output section by using
4504 SUBALIGN. The value specified overrides any alignment given by input
4505 sections, whether larger or smaller.
4507 @node Output Section Constraint
4508 @subsubsection Output Section Constraint
4511 @cindex constraints on output sections
4512 You can specify that an output section should only be created if all
4513 of its input sections are read-only or all of its input sections are
4514 read-write by using the keyword @code{ONLY_IF_RO} and
4515 @code{ONLY_IF_RW} respectively.
4517 @node Output Section Region
4518 @subsubsection Output Section Region
4519 @kindex >@var{region}
4520 @cindex section, assigning to memory region
4521 @cindex memory regions and sections
4522 You can assign a section to a previously defined region of memory by
4523 using @samp{>@var{region}}. @xref{MEMORY}.
4525 Here is a simple example:
4528 MEMORY @{ rom : ORIGIN = 0x1000, LENGTH = 0x1000 @}
4529 SECTIONS @{ ROM : @{ *(.text) @} >rom @}
4533 @node Output Section Phdr
4534 @subsubsection Output Section Phdr
4536 @cindex section, assigning to program header
4537 @cindex program headers and sections
4538 You can assign a section to a previously defined program segment by
4539 using @samp{:@var{phdr}}. @xref{PHDRS}. If a section is assigned to
4540 one or more segments, then all subsequent allocated sections will be
4541 assigned to those segments as well, unless they use an explicitly
4542 @code{:@var{phdr}} modifier. You can use @code{:NONE} to tell the
4543 linker to not put the section in any segment at all.
4545 Here is a simple example:
4548 PHDRS @{ text PT_LOAD ; @}
4549 SECTIONS @{ .text : @{ *(.text) @} :text @}
4553 @node Output Section Fill
4554 @subsubsection Output Section Fill
4555 @kindex =@var{fillexp}
4556 @cindex section fill pattern
4557 @cindex fill pattern, entire section
4558 You can set the fill pattern for an entire section by using
4559 @samp{=@var{fillexp}}. @var{fillexp} is an expression
4560 (@pxref{Expressions}). Any otherwise unspecified regions of memory
4561 within the output section (for example, gaps left due to the required
4562 alignment of input sections) will be filled with the value, repeated as
4563 necessary. If the fill expression is a simple hex number, ie. a string
4564 of hex digit starting with @samp{0x} and without a trailing @samp{k} or @samp{M}, then
4565 an arbitrarily long sequence of hex digits can be used to specify the
4566 fill pattern; Leading zeros become part of the pattern too. For all
4567 other cases, including extra parentheses or a unary @code{+}, the fill
4568 pattern is the four least significant bytes of the value of the
4569 expression. In all cases, the number is big-endian.
4571 You can also change the fill value with a @code{FILL} command in the
4572 output section commands; (@pxref{Output Section Data}).
4574 Here is a simple example:
4577 SECTIONS @{ .text : @{ *(.text) @} =0x90909090 @}
4581 @node Overlay Description
4582 @subsection Overlay Description
4585 An overlay description provides an easy way to describe sections which
4586 are to be loaded as part of a single memory image but are to be run at
4587 the same memory address. At run time, some sort of overlay manager will
4588 copy the overlaid sections in and out of the runtime memory address as
4589 required, perhaps by simply manipulating addressing bits. This approach
4590 can be useful, for example, when a certain region of memory is faster
4593 Overlays are described using the @code{OVERLAY} command. The
4594 @code{OVERLAY} command is used within a @code{SECTIONS} command, like an
4595 output section description. The full syntax of the @code{OVERLAY}
4596 command is as follows:
4599 OVERLAY [@var{start}] : [NOCROSSREFS] [AT ( @var{ldaddr} )]
4603 @var{output-section-command}
4604 @var{output-section-command}
4606 @} [:@var{phdr}@dots{}] [=@var{fill}]
4609 @var{output-section-command}
4610 @var{output-section-command}
4612 @} [:@var{phdr}@dots{}] [=@var{fill}]
4614 @} [>@var{region}] [:@var{phdr}@dots{}] [=@var{fill}]
4618 Everything is optional except @code{OVERLAY} (a keyword), and each
4619 section must have a name (@var{secname1} and @var{secname2} above). The
4620 section definitions within the @code{OVERLAY} construct are identical to
4621 those within the general @code{SECTIONS} contruct (@pxref{SECTIONS}),
4622 except that no addresses and no memory regions may be defined for
4623 sections within an @code{OVERLAY}.
4625 The sections are all defined with the same starting address. The load
4626 addresses of the sections are arranged such that they are consecutive in
4627 memory starting at the load address used for the @code{OVERLAY} as a
4628 whole (as with normal section definitions, the load address is optional,
4629 and defaults to the start address; the start address is also optional,
4630 and defaults to the current value of the location counter).
4632 If the @code{NOCROSSREFS} keyword is used, and there any references
4633 among the sections, the linker will report an error. Since the sections
4634 all run at the same address, it normally does not make sense for one
4635 section to refer directly to another. @xref{Miscellaneous Commands,
4638 For each section within the @code{OVERLAY}, the linker automatically
4639 provides two symbols. The symbol @code{__load_start_@var{secname}} is
4640 defined as the starting load address of the section. The symbol
4641 @code{__load_stop_@var{secname}} is defined as the final load address of
4642 the section. Any characters within @var{secname} which are not legal
4643 within C identifiers are removed. C (or assembler) code may use these
4644 symbols to move the overlaid sections around as necessary.
4646 At the end of the overlay, the value of the location counter is set to
4647 the start address of the overlay plus the size of the largest section.
4649 Here is an example. Remember that this would appear inside a
4650 @code{SECTIONS} construct.
4653 OVERLAY 0x1000 : AT (0x4000)
4655 .text0 @{ o1/*.o(.text) @}
4656 .text1 @{ o2/*.o(.text) @}
4661 This will define both @samp{.text0} and @samp{.text1} to start at
4662 address 0x1000. @samp{.text0} will be loaded at address 0x4000, and
4663 @samp{.text1} will be loaded immediately after @samp{.text0}. The
4664 following symbols will be defined if referenced: @code{__load_start_text0},
4665 @code{__load_stop_text0}, @code{__load_start_text1},
4666 @code{__load_stop_text1}.
4668 C code to copy overlay @code{.text1} into the overlay area might look
4673 extern char __load_start_text1, __load_stop_text1;
4674 memcpy ((char *) 0x1000, &__load_start_text1,
4675 &__load_stop_text1 - &__load_start_text1);
4679 Note that the @code{OVERLAY} command is just syntactic sugar, since
4680 everything it does can be done using the more basic commands. The above
4681 example could have been written identically as follows.
4685 .text0 0x1000 : AT (0x4000) @{ o1/*.o(.text) @}
4686 PROVIDE (__load_start_text0 = LOADADDR (.text0));
4687 PROVIDE (__load_stop_text0 = LOADADDR (.text0) + SIZEOF (.text0));
4688 .text1 0x1000 : AT (0x4000 + SIZEOF (.text0)) @{ o2/*.o(.text) @}
4689 PROVIDE (__load_start_text1 = LOADADDR (.text1));
4690 PROVIDE (__load_stop_text1 = LOADADDR (.text1) + SIZEOF (.text1));
4691 . = 0x1000 + MAX (SIZEOF (.text0), SIZEOF (.text1));
4696 @section MEMORY Command
4698 @cindex memory regions
4699 @cindex regions of memory
4700 @cindex allocating memory
4701 @cindex discontinuous memory
4702 The linker's default configuration permits allocation of all available
4703 memory. You can override this by using the @code{MEMORY} command.
4705 The @code{MEMORY} command describes the location and size of blocks of
4706 memory in the target. You can use it to describe which memory regions
4707 may be used by the linker, and which memory regions it must avoid. You
4708 can then assign sections to particular memory regions. The linker will
4709 set section addresses based on the memory regions, and will warn about
4710 regions that become too full. The linker will not shuffle sections
4711 around to fit into the available regions.
4713 A linker script may contain at most one use of the @code{MEMORY}
4714 command. However, you can define as many blocks of memory within it as
4715 you wish. The syntax is:
4720 @var{name} [(@var{attr})] : ORIGIN = @var{origin}, LENGTH = @var{len}
4726 The @var{name} is a name used in the linker script to refer to the
4727 region. The region name has no meaning outside of the linker script.
4728 Region names are stored in a separate name space, and will not conflict
4729 with symbol names, file names, or section names. Each memory region
4730 must have a distinct name within the @code{MEMORY} command. However you can
4731 add later alias names to existing memory regions with the @ref{REGION_ALIAS}
4734 @cindex memory region attributes
4735 The @var{attr} string is an optional list of attributes that specify
4736 whether to use a particular memory region for an input section which is
4737 not explicitly mapped in the linker script. As described in
4738 @ref{SECTIONS}, if you do not specify an output section for some input
4739 section, the linker will create an output section with the same name as
4740 the input section. If you define region attributes, the linker will use
4741 them to select the memory region for the output section that it creates.
4743 The @var{attr} string must consist only of the following characters:
4758 Invert the sense of any of the attributes that follow
4761 If a unmapped section matches any of the listed attributes other than
4762 @samp{!}, it will be placed in the memory region. The @samp{!}
4763 attribute reverses this test, so that an unmapped section will be placed
4764 in the memory region only if it does not match any of the listed
4770 The @var{origin} is an numerical expression for the start address of
4771 the memory region. The expression must evaluate to a constant and it
4772 cannot involve any symbols. The keyword @code{ORIGIN} may be
4773 abbreviated to @code{org} or @code{o} (but not, for example,
4779 The @var{len} is an expression for the size in bytes of the memory
4780 region. As with the @var{origin} expression, the expression must
4781 be numerical only and must evaluate to a constant. The keyword
4782 @code{LENGTH} may be abbreviated to @code{len} or @code{l}.
4784 In the following example, we specify that there are two memory regions
4785 available for allocation: one starting at @samp{0} for 256 kilobytes,
4786 and the other starting at @samp{0x40000000} for four megabytes. The
4787 linker will place into the @samp{rom} memory region every section which
4788 is not explicitly mapped into a memory region, and is either read-only
4789 or executable. The linker will place other sections which are not
4790 explicitly mapped into a memory region into the @samp{ram} memory
4797 rom (rx) : ORIGIN = 0, LENGTH = 256K
4798 ram (!rx) : org = 0x40000000, l = 4M
4803 Once you define a memory region, you can direct the linker to place
4804 specific output sections into that memory region by using the
4805 @samp{>@var{region}} output section attribute. For example, if you have
4806 a memory region named @samp{mem}, you would use @samp{>mem} in the
4807 output section definition. @xref{Output Section Region}. If no address
4808 was specified for the output section, the linker will set the address to
4809 the next available address within the memory region. If the combined
4810 output sections directed to a memory region are too large for the
4811 region, the linker will issue an error message.
4813 It is possible to access the origin and length of a memory in an
4814 expression via the @code{ORIGIN(@var{memory})} and
4815 @code{LENGTH(@var{memory})} functions:
4819 _fstack = ORIGIN(ram) + LENGTH(ram) - 4;
4824 @section PHDRS Command
4826 @cindex program headers
4827 @cindex ELF program headers
4828 @cindex program segments
4829 @cindex segments, ELF
4830 The ELF object file format uses @dfn{program headers}, also knows as
4831 @dfn{segments}. The program headers describe how the program should be
4832 loaded into memory. You can print them out by using the @code{objdump}
4833 program with the @samp{-p} option.
4835 When you run an ELF program on a native ELF system, the system loader
4836 reads the program headers in order to figure out how to load the
4837 program. This will only work if the program headers are set correctly.
4838 This manual does not describe the details of how the system loader
4839 interprets program headers; for more information, see the ELF ABI.
4841 The linker will create reasonable program headers by default. However,
4842 in some cases, you may need to specify the program headers more
4843 precisely. You may use the @code{PHDRS} command for this purpose. When
4844 the linker sees the @code{PHDRS} command in the linker script, it will
4845 not create any program headers other than the ones specified.
4847 The linker only pays attention to the @code{PHDRS} command when
4848 generating an ELF output file. In other cases, the linker will simply
4849 ignore @code{PHDRS}.
4851 This is the syntax of the @code{PHDRS} command. The words @code{PHDRS},
4852 @code{FILEHDR}, @code{AT}, and @code{FLAGS} are keywords.
4858 @var{name} @var{type} [ FILEHDR ] [ PHDRS ] [ AT ( @var{address} ) ]
4859 [ FLAGS ( @var{flags} ) ] ;
4864 The @var{name} is used only for reference in the @code{SECTIONS} command
4865 of the linker script. It is not put into the output file. Program
4866 header names are stored in a separate name space, and will not conflict
4867 with symbol names, file names, or section names. Each program header
4868 must have a distinct name. The headers are processed in order and it
4869 is usual for them to map to sections in ascending load address order.
4871 Certain program header types describe segments of memory which the
4872 system loader will load from the file. In the linker script, you
4873 specify the contents of these segments by placing allocatable output
4874 sections in the segments. You use the @samp{:@var{phdr}} output section
4875 attribute to place a section in a particular segment. @xref{Output
4878 It is normal to put certain sections in more than one segment. This
4879 merely implies that one segment of memory contains another. You may
4880 repeat @samp{:@var{phdr}}, using it once for each segment which should
4881 contain the section.
4883 If you place a section in one or more segments using @samp{:@var{phdr}},
4884 then the linker will place all subsequent allocatable sections which do
4885 not specify @samp{:@var{phdr}} in the same segments. This is for
4886 convenience, since generally a whole set of contiguous sections will be
4887 placed in a single segment. You can use @code{:NONE} to override the
4888 default segment and tell the linker to not put the section in any
4893 You may use the @code{FILEHDR} and @code{PHDRS} keywords after
4894 the program header type to further describe the contents of the segment.
4895 The @code{FILEHDR} keyword means that the segment should include the ELF
4896 file header. The @code{PHDRS} keyword means that the segment should
4897 include the ELF program headers themselves. If applied to a loadable
4898 segment (@code{PT_LOAD}), all prior loadable segments must have one of
4901 The @var{type} may be one of the following. The numbers indicate the
4902 value of the keyword.
4905 @item @code{PT_NULL} (0)
4906 Indicates an unused program header.
4908 @item @code{PT_LOAD} (1)
4909 Indicates that this program header describes a segment to be loaded from
4912 @item @code{PT_DYNAMIC} (2)
4913 Indicates a segment where dynamic linking information can be found.
4915 @item @code{PT_INTERP} (3)
4916 Indicates a segment where the name of the program interpreter may be
4919 @item @code{PT_NOTE} (4)
4920 Indicates a segment holding note information.
4922 @item @code{PT_SHLIB} (5)
4923 A reserved program header type, defined but not specified by the ELF
4926 @item @code{PT_PHDR} (6)
4927 Indicates a segment where the program headers may be found.
4929 @item @var{expression}
4930 An expression giving the numeric type of the program header. This may
4931 be used for types not defined above.
4934 You can specify that a segment should be loaded at a particular address
4935 in memory by using an @code{AT} expression. This is identical to the
4936 @code{AT} command used as an output section attribute (@pxref{Output
4937 Section LMA}). The @code{AT} command for a program header overrides the
4938 output section attribute.
4940 The linker will normally set the segment flags based on the sections
4941 which comprise the segment. You may use the @code{FLAGS} keyword to
4942 explicitly specify the segment flags. The value of @var{flags} must be
4943 an integer. It is used to set the @code{p_flags} field of the program
4946 Here is an example of @code{PHDRS}. This shows a typical set of program
4947 headers used on a native ELF system.
4953 headers PT_PHDR PHDRS ;
4955 text PT_LOAD FILEHDR PHDRS ;
4957 dynamic PT_DYNAMIC ;
4963 .interp : @{ *(.interp) @} :text :interp
4964 .text : @{ *(.text) @} :text
4965 .rodata : @{ *(.rodata) @} /* defaults to :text */
4967 . = . + 0x1000; /* move to a new page in memory */
4968 .data : @{ *(.data) @} :data
4969 .dynamic : @{ *(.dynamic) @} :data :dynamic
4976 @section VERSION Command
4977 @kindex VERSION @{script text@}
4978 @cindex symbol versions
4979 @cindex version script
4980 @cindex versions of symbols
4981 The linker supports symbol versions when using ELF. Symbol versions are
4982 only useful when using shared libraries. The dynamic linker can use
4983 symbol versions to select a specific version of a function when it runs
4984 a program that may have been linked against an earlier version of the
4987 You can include a version script directly in the main linker script, or
4988 you can supply the version script as an implicit linker script. You can
4989 also use the @samp{--version-script} linker option.
4991 The syntax of the @code{VERSION} command is simply
4993 VERSION @{ version-script-commands @}
4996 The format of the version script commands is identical to that used by
4997 Sun's linker in Solaris 2.5. The version script defines a tree of
4998 version nodes. You specify the node names and interdependencies in the
4999 version script. You can specify which symbols are bound to which
5000 version nodes, and you can reduce a specified set of symbols to local
5001 scope so that they are not globally visible outside of the shared
5004 The easiest way to demonstrate the version script language is with a few
5030 This example version script defines three version nodes. The first
5031 version node defined is @samp{VERS_1.1}; it has no other dependencies.
5032 The script binds the symbol @samp{foo1} to @samp{VERS_1.1}. It reduces
5033 a number of symbols to local scope so that they are not visible outside
5034 of the shared library; this is done using wildcard patterns, so that any
5035 symbol whose name begins with @samp{old}, @samp{original}, or @samp{new}
5036 is matched. The wildcard patterns available are the same as those used
5037 in the shell when matching filenames (also known as ``globbing'').
5038 However, if you specify the symbol name inside double quotes, then the
5039 name is treated as literal, rather than as a glob pattern.
5041 Next, the version script defines node @samp{VERS_1.2}. This node
5042 depends upon @samp{VERS_1.1}. The script binds the symbol @samp{foo2}
5043 to the version node @samp{VERS_1.2}.
5045 Finally, the version script defines node @samp{VERS_2.0}. This node
5046 depends upon @samp{VERS_1.2}. The scripts binds the symbols @samp{bar1}
5047 and @samp{bar2} are bound to the version node @samp{VERS_2.0}.
5049 When the linker finds a symbol defined in a library which is not
5050 specifically bound to a version node, it will effectively bind it to an
5051 unspecified base version of the library. You can bind all otherwise
5052 unspecified symbols to a given version node by using @samp{global: *;}
5053 somewhere in the version script. Note that it's slightly crazy to use
5054 wildcards in a global spec except on the last version node. Global
5055 wildcards elsewhere run the risk of accidentally adding symbols to the
5056 set exported for an old version. That's wrong since older versions
5057 ought to have a fixed set of symbols.
5059 The names of the version nodes have no specific meaning other than what
5060 they might suggest to the person reading them. The @samp{2.0} version
5061 could just as well have appeared in between @samp{1.1} and @samp{1.2}.
5062 However, this would be a confusing way to write a version script.
5064 Node name can be omitted, provided it is the only version node
5065 in the version script. Such version script doesn't assign any versions to
5066 symbols, only selects which symbols will be globally visible out and which
5070 @{ global: foo; bar; local: *; @};
5073 When you link an application against a shared library that has versioned
5074 symbols, the application itself knows which version of each symbol it
5075 requires, and it also knows which version nodes it needs from each
5076 shared library it is linked against. Thus at runtime, the dynamic
5077 loader can make a quick check to make sure that the libraries you have
5078 linked against do in fact supply all of the version nodes that the
5079 application will need to resolve all of the dynamic symbols. In this
5080 way it is possible for the dynamic linker to know with certainty that
5081 all external symbols that it needs will be resolvable without having to
5082 search for each symbol reference.
5084 The symbol versioning is in effect a much more sophisticated way of
5085 doing minor version checking that SunOS does. The fundamental problem
5086 that is being addressed here is that typically references to external
5087 functions are bound on an as-needed basis, and are not all bound when
5088 the application starts up. If a shared library is out of date, a
5089 required interface may be missing; when the application tries to use
5090 that interface, it may suddenly and unexpectedly fail. With symbol
5091 versioning, the user will get a warning when they start their program if
5092 the libraries being used with the application are too old.
5094 There are several GNU extensions to Sun's versioning approach. The
5095 first of these is the ability to bind a symbol to a version node in the
5096 source file where the symbol is defined instead of in the versioning
5097 script. This was done mainly to reduce the burden on the library
5098 maintainer. You can do this by putting something like:
5100 __asm__(".symver original_foo,foo@@VERS_1.1");
5103 in the C source file. This renames the function @samp{original_foo} to
5104 be an alias for @samp{foo} bound to the version node @samp{VERS_1.1}.
5105 The @samp{local:} directive can be used to prevent the symbol
5106 @samp{original_foo} from being exported. A @samp{.symver} directive
5107 takes precedence over a version script.
5109 The second GNU extension is to allow multiple versions of the same
5110 function to appear in a given shared library. In this way you can make
5111 an incompatible change to an interface without increasing the major
5112 version number of the shared library, while still allowing applications
5113 linked against the old interface to continue to function.
5115 To do this, you must use multiple @samp{.symver} directives in the
5116 source file. Here is an example:
5119 __asm__(".symver original_foo,foo@@");
5120 __asm__(".symver old_foo,foo@@VERS_1.1");
5121 __asm__(".symver old_foo1,foo@@VERS_1.2");
5122 __asm__(".symver new_foo,foo@@@@VERS_2.0");
5125 In this example, @samp{foo@@} represents the symbol @samp{foo} bound to the
5126 unspecified base version of the symbol. The source file that contains this
5127 example would define 4 C functions: @samp{original_foo}, @samp{old_foo},
5128 @samp{old_foo1}, and @samp{new_foo}.
5130 When you have multiple definitions of a given symbol, there needs to be
5131 some way to specify a default version to which external references to
5132 this symbol will be bound. You can do this with the
5133 @samp{foo@@@@VERS_2.0} type of @samp{.symver} directive. You can only
5134 declare one version of a symbol as the default in this manner; otherwise
5135 you would effectively have multiple definitions of the same symbol.
5137 If you wish to bind a reference to a specific version of the symbol
5138 within the shared library, you can use the aliases of convenience
5139 (i.e., @samp{old_foo}), or you can use the @samp{.symver} directive to
5140 specifically bind to an external version of the function in question.
5142 You can also specify the language in the version script:
5145 VERSION extern "lang" @{ version-script-commands @}
5148 The supported @samp{lang}s are @samp{C}, @samp{C++}, and @samp{Java}.
5149 The linker will iterate over the list of symbols at the link time and
5150 demangle them according to @samp{lang} before matching them to the
5151 patterns specified in @samp{version-script-commands}. The default
5152 @samp{lang} is @samp{C}.
5154 Demangled names may contains spaces and other special characters. As
5155 described above, you can use a glob pattern to match demangled names,
5156 or you can use a double-quoted string to match the string exactly. In
5157 the latter case, be aware that minor differences (such as differing
5158 whitespace) between the version script and the demangler output will
5159 cause a mismatch. As the exact string generated by the demangler
5160 might change in the future, even if the mangled name does not, you
5161 should check that all of your version directives are behaving as you
5162 expect when you upgrade.
5165 @section Expressions in Linker Scripts
5168 The syntax for expressions in the linker script language is identical to
5169 that of C expressions. All expressions are evaluated as integers. All
5170 expressions are evaluated in the same size, which is 32 bits if both the
5171 host and target are 32 bits, and is otherwise 64 bits.
5173 You can use and set symbol values in expressions.
5175 The linker defines several special purpose builtin functions for use in
5179 * Constants:: Constants
5180 * Symbolic Constants:: Symbolic constants
5181 * Symbols:: Symbol Names
5182 * Orphan Sections:: Orphan Sections
5183 * Location Counter:: The Location Counter
5184 * Operators:: Operators
5185 * Evaluation:: Evaluation
5186 * Expression Section:: The Section of an Expression
5187 * Builtin Functions:: Builtin Functions
5191 @subsection Constants
5192 @cindex integer notation
5193 @cindex constants in linker scripts
5194 All constants are integers.
5196 As in C, the linker considers an integer beginning with @samp{0} to be
5197 octal, and an integer beginning with @samp{0x} or @samp{0X} to be
5198 hexadecimal. Alternatively the linker accepts suffixes of @samp{h} or
5199 @samp{H} for hexadeciaml, @samp{o} or @samp{O} for octal, @samp{b} or
5200 @samp{B} for binary and @samp{d} or @samp{D} for decimal. Any integer
5201 value without a prefix or a suffix is considered to be decimal.
5203 @cindex scaled integers
5204 @cindex K and M integer suffixes
5205 @cindex M and K integer suffixes
5206 @cindex suffixes for integers
5207 @cindex integer suffixes
5208 In addition, you can use the suffixes @code{K} and @code{M} to scale a
5212 @c END TEXI2ROFF-KILL
5213 @code{1024} or @code{1024*1024}
5217 ${\rm 1024}$ or ${\rm 1024}^2$
5219 @c END TEXI2ROFF-KILL
5220 respectively. For example, the following
5221 all refer to the same quantity:
5230 Note - the @code{K} and @code{M} suffixes cannot be used in
5231 conjunction with the base suffixes mentioned above.
5233 @node Symbolic Constants
5234 @subsection Symbolic Constants
5235 @cindex symbolic constants
5237 It is possible to refer to target specific constants via the use of
5238 the @code{CONSTANT(@var{name})} operator, where @var{name} is one of:
5243 The target's maximum page size.
5245 @item COMMONPAGESIZE
5246 @kindex COMMONPAGESIZE
5247 The target's default page size.
5253 .text ALIGN (CONSTANT (MAXPAGESIZE)) : @{ *(.text) @}
5256 will create a text section aligned to the largest page boundary
5257 supported by the target.
5260 @subsection Symbol Names
5261 @cindex symbol names
5263 @cindex quoted symbol names
5265 Unless quoted, symbol names start with a letter, underscore, or period
5266 and may include letters, digits, underscores, periods, and hyphens.
5267 Unquoted symbol names must not conflict with any keywords. You can
5268 specify a symbol which contains odd characters or has the same name as a
5269 keyword by surrounding the symbol name in double quotes:
5272 "with a space" = "also with a space" + 10;
5275 Since symbols can contain many non-alphabetic characters, it is safest
5276 to delimit symbols with spaces. For example, @samp{A-B} is one symbol,
5277 whereas @samp{A - B} is an expression involving subtraction.
5279 @node Orphan Sections
5280 @subsection Orphan Sections
5282 Orphan sections are sections present in the input files which
5283 are not explicitly placed into the output file by the linker
5284 script. The linker will still copy these sections into the
5285 output file, but it has to guess as to where they should be
5286 placed. The linker uses a simple heuristic to do this. It
5287 attempts to place orphan sections after non-orphan sections of the
5288 same attribute, such as code vs data, loadable vs non-loadable, etc.
5289 If there is not enough room to do this then it places
5290 at the end of the file.
5292 For ELF targets, the attribute of the section includes section type as
5293 well as section flag.
5295 If an orphaned section's name is representable as a C identifier then
5296 the linker will automatically @pxref{PROVIDE} two symbols:
5297 __start_SECNAME and __end_SECNAME, where SECNAME is the name of the
5298 section. These indicate the start address and end address of the
5299 orphaned section respectively. Note: most section names are not
5300 representable as C identifiers because they contain a @samp{.}
5303 @node Location Counter
5304 @subsection The Location Counter
5307 @cindex location counter
5308 @cindex current output location
5309 The special linker variable @dfn{dot} @samp{.} always contains the
5310 current output location counter. Since the @code{.} always refers to a
5311 location in an output section, it may only appear in an expression
5312 within a @code{SECTIONS} command. The @code{.} symbol may appear
5313 anywhere that an ordinary symbol is allowed in an expression.
5316 Assigning a value to @code{.} will cause the location counter to be
5317 moved. This may be used to create holes in the output section. The
5318 location counter may not be moved backwards inside an output section,
5319 and may not be moved backwards outside of an output section if so
5320 doing creates areas with overlapping LMAs.
5336 In the previous example, the @samp{.text} section from @file{file1} is
5337 located at the beginning of the output section @samp{output}. It is
5338 followed by a 1000 byte gap. Then the @samp{.text} section from
5339 @file{file2} appears, also with a 1000 byte gap following before the
5340 @samp{.text} section from @file{file3}. The notation @samp{= 0x12345678}
5341 specifies what data to write in the gaps (@pxref{Output Section Fill}).
5343 @cindex dot inside sections
5344 Note: @code{.} actually refers to the byte offset from the start of the
5345 current containing object. Normally this is the @code{SECTIONS}
5346 statement, whose start address is 0, hence @code{.} can be used as an
5347 absolute address. If @code{.} is used inside a section description
5348 however, it refers to the byte offset from the start of that section,
5349 not an absolute address. Thus in a script like this:
5367 The @samp{.text} section will be assigned a starting address of 0x100
5368 and a size of exactly 0x200 bytes, even if there is not enough data in
5369 the @samp{.text} input sections to fill this area. (If there is too
5370 much data, an error will be produced because this would be an attempt to
5371 move @code{.} backwards). The @samp{.data} section will start at 0x500
5372 and it will have an extra 0x600 bytes worth of space after the end of
5373 the values from the @samp{.data} input sections and before the end of
5374 the @samp{.data} output section itself.
5376 @cindex dot outside sections
5377 Setting symbols to the value of the location counter outside of an
5378 output section statement can result in unexpected values if the linker
5379 needs to place orphan sections. For example, given the following:
5385 .text: @{ *(.text) @}
5389 .data: @{ *(.data) @}
5394 If the linker needs to place some input section, e.g. @code{.rodata},
5395 not mentioned in the script, it might choose to place that section
5396 between @code{.text} and @code{.data}. You might think the linker
5397 should place @code{.rodata} on the blank line in the above script, but
5398 blank lines are of no particular significance to the linker. As well,
5399 the linker doesn't associate the above symbol names with their
5400 sections. Instead, it assumes that all assignments or other
5401 statements belong to the previous output section, except for the
5402 special case of an assignment to @code{.}. I.e., the linker will
5403 place the orphan @code{.rodata} section as if the script was written
5410 .text: @{ *(.text) @}
5414 .rodata: @{ *(.rodata) @}
5415 .data: @{ *(.data) @}
5420 This may or may not be the script author's intention for the value of
5421 @code{start_of_data}. One way to influence the orphan section
5422 placement is to assign the location counter to itself, as the linker
5423 assumes that an assignment to @code{.} is setting the start address of
5424 a following output section and thus should be grouped with that
5425 section. So you could write:
5431 .text: @{ *(.text) @}
5436 .data: @{ *(.data) @}
5441 Now, the orphan @code{.rodata} section will be placed between
5442 @code{end_of_text} and @code{start_of_data}.
5446 @subsection Operators
5447 @cindex operators for arithmetic
5448 @cindex arithmetic operators
5449 @cindex precedence in expressions
5450 The linker recognizes the standard C set of arithmetic operators, with
5451 the standard bindings and precedence levels:
5454 @c END TEXI2ROFF-KILL
5456 precedence associativity Operators Notes
5462 5 left == != > < <= >=
5468 11 right &= += -= *= /= (2)
5472 (1) Prefix operators
5473 (2) @xref{Assignments}.
5477 \vskip \baselineskip
5478 %"lispnarrowing" is the extra indent used generally for smallexample
5479 \hskip\lispnarrowing\vbox{\offinterlineskip
5482 {\vrule#&\strut\hfil\ #\ \hfil&\vrule#&\strut\hfil\ #\ \hfil&\vrule#&\strut\hfil\ {\tt #}\ \hfil&\vrule#\cr
5483 height2pt&\omit&&\omit&&\omit&\cr
5484 &Precedence&& Associativity &&{\rm Operators}&\cr
5485 height2pt&\omit&&\omit&&\omit&\cr
5487 height2pt&\omit&&\omit&&\omit&\cr
5489 % '176 is tilde, '~' in tt font
5490 &1&&left&&\qquad- \char'176\ !\qquad\dag&\cr
5491 &2&&left&&* / \%&\cr
5494 &5&&left&&== != > < <= >=&\cr
5497 &8&&left&&{\&\&}&\cr
5500 &11&&right&&\qquad\&= += -= *= /=\qquad\ddag&\cr
5502 height2pt&\omit&&\omit&&\omit&\cr}
5507 @obeylines@parskip=0pt@parindent=0pt
5508 @dag@quad Prefix operators.
5509 @ddag@quad @xref{Assignments}.
5512 @c END TEXI2ROFF-KILL
5515 @subsection Evaluation
5516 @cindex lazy evaluation
5517 @cindex expression evaluation order
5518 The linker evaluates expressions lazily. It only computes the value of
5519 an expression when absolutely necessary.
5521 The linker needs some information, such as the value of the start
5522 address of the first section, and the origins and lengths of memory
5523 regions, in order to do any linking at all. These values are computed
5524 as soon as possible when the linker reads in the linker script.
5526 However, other values (such as symbol values) are not known or needed
5527 until after storage allocation. Such values are evaluated later, when
5528 other information (such as the sizes of output sections) is available
5529 for use in the symbol assignment expression.
5531 The sizes of sections cannot be known until after allocation, so
5532 assignments dependent upon these are not performed until after
5535 Some expressions, such as those depending upon the location counter
5536 @samp{.}, must be evaluated during section allocation.
5538 If the result of an expression is required, but the value is not
5539 available, then an error results. For example, a script like the
5545 .text 9+this_isnt_constant :
5551 will cause the error message @samp{non constant expression for initial
5554 @node Expression Section
5555 @subsection The Section of an Expression
5556 @cindex expression sections
5557 @cindex absolute expressions
5558 @cindex relative expressions
5559 @cindex absolute and relocatable symbols
5560 @cindex relocatable and absolute symbols
5561 @cindex symbols, relocatable and absolute
5562 Addresses and symbols may be section relative, or absolute. A section
5563 relative symbol is relocatable. If you request relocatable output
5564 using the @samp{-r} option, a further link operation may change the
5565 value of a section relative symbol. On the other hand, an absolute
5566 symbol will retain the same value throughout any further link
5569 Some terms in linker expressions are addresses. This is true of
5570 section relative symbols and for builtin functions that return an
5571 address, such as @code{ADDR}, @code{LOADADDR}, @code{ORIGIN} and
5572 @code{SEGMENT_START}. Other terms are simply numbers, or are builtin
5573 functions that return a non-address value, such as @code{LENGTH}.
5574 One complication is that unless you set @code{LD_FEATURE ("SANE_EXPR")}
5575 (@pxref{Miscellaneous Commands}), numbers and absolute symbols are treated
5576 differently depending on their location, for compatibility with older
5577 versions of @code{ld}. Expressions appearing outside an output
5578 section definition treat all numbers as absolute addresses.
5579 Expressions appearing inside an output section definition treat
5580 absolute symbols as numbers. If @code{LD_FEATURE ("SANE_EXPR")} is
5581 given, then absolute symbols and numbers are simply treated as numbers
5584 In the following simple example,
5591 __executable_start = 0x100;
5595 __data_start = 0x10;
5603 both @code{.} and @code{__executable_start} are set to the absolute
5604 address 0x100 in the first two assignments, then both @code{.} and
5605 @code{__data_start} are set to 0x10 relative to the @code{.data}
5606 section in the second two assignments.
5608 For expressions involving numbers, relative addresses and absolute
5609 addresses, ld follows these rules to evaluate terms:
5613 Unary operations on a relative address, and binary operations on two
5614 relative addresses in the same section or between one relative address
5615 and a number, apply the operator to the offset part of the address(es).
5617 Unary operations on an absolute address, and binary operations on one
5618 or more absolute addresses or on two relative addresses not in the
5619 same section, first convert any non-absolute term to an absolute
5620 address before applying the operator.
5623 The result section of each sub-expression is as follows:
5627 An operation involving only numbers results in a number.
5629 The result of comparisons, @samp{&&} and @samp{||} is also a number.
5631 The result of other binary arithmetic and logical operations on two
5632 relative addresses in the same section or two absolute addresess
5633 (after above conversions) is also a number.
5635 The result of other operations on relative addresses or one
5636 relative address and a number, is a relative address in the same
5637 section as the relative operand(s).
5639 The result of other operations on absolute addresses (after above
5640 conversions) is an absolute address.
5643 You can use the builtin function @code{ABSOLUTE} to force an expression
5644 to be absolute when it would otherwise be relative. For example, to
5645 create an absolute symbol set to the address of the end of the output
5646 section @samp{.data}:
5650 .data : @{ *(.data) _edata = ABSOLUTE(.); @}
5654 If @samp{ABSOLUTE} were not used, @samp{_edata} would be relative to the
5655 @samp{.data} section.
5657 Using @code{LOADADDR} also forces an expression absolute, since this
5658 particular builtin function returns an absolute address.
5660 @node Builtin Functions
5661 @subsection Builtin Functions
5662 @cindex functions in expressions
5663 The linker script language includes a number of builtin functions for
5664 use in linker script expressions.
5667 @item ABSOLUTE(@var{exp})
5668 @kindex ABSOLUTE(@var{exp})
5669 @cindex expression, absolute
5670 Return the absolute (non-relocatable, as opposed to non-negative) value
5671 of the expression @var{exp}. Primarily useful to assign an absolute
5672 value to a symbol within a section definition, where symbol values are
5673 normally section relative. @xref{Expression Section}.
5675 @item ADDR(@var{section})
5676 @kindex ADDR(@var{section})
5677 @cindex section address in expression
5678 Return the address (VMA) of the named @var{section}. Your
5679 script must previously have defined the location of that section. In
5680 the following example, @code{start_of_output_1}, @code{symbol_1} and
5681 @code{symbol_2} are assigned equivalent values, except that
5682 @code{symbol_1} will be relative to the @code{.output1} section while
5683 the other two will be absolute:
5689 start_of_output_1 = ABSOLUTE(.);
5694 symbol_1 = ADDR(.output1);
5695 symbol_2 = start_of_output_1;
5701 @item ALIGN(@var{align})
5702 @itemx ALIGN(@var{exp},@var{align})
5703 @kindex ALIGN(@var{align})
5704 @kindex ALIGN(@var{exp},@var{align})
5705 @cindex round up location counter
5706 @cindex align location counter
5707 @cindex round up expression
5708 @cindex align expression
5709 Return the location counter (@code{.}) or arbitrary expression aligned
5710 to the next @var{align} boundary. The single operand @code{ALIGN}
5711 doesn't change the value of the location counter---it just does
5712 arithmetic on it. The two operand @code{ALIGN} allows an arbitrary
5713 expression to be aligned upwards (@code{ALIGN(@var{align})} is
5714 equivalent to @code{ALIGN(., @var{align})}).
5716 Here is an example which aligns the output @code{.data} section to the
5717 next @code{0x2000} byte boundary after the preceding section and sets a
5718 variable within the section to the next @code{0x8000} boundary after the
5723 .data ALIGN(0x2000): @{
5725 variable = ALIGN(0x8000);
5731 The first use of @code{ALIGN} in this example specifies the location of
5732 a section because it is used as the optional @var{address} attribute of
5733 a section definition (@pxref{Output Section Address}). The second use
5734 of @code{ALIGN} is used to defines the value of a symbol.
5736 The builtin function @code{NEXT} is closely related to @code{ALIGN}.
5738 @item ALIGNOF(@var{section})
5739 @kindex ALIGNOF(@var{section})
5740 @cindex section alignment
5741 Return the alignment in bytes of the named @var{section}, if that section has
5742 been allocated. If the section has not been allocated when this is
5743 evaluated, the linker will report an error. In the following example,
5744 the alignment of the @code{.output} section is stored as the first
5745 value in that section.
5750 LONG (ALIGNOF (.output))
5757 @item BLOCK(@var{exp})
5758 @kindex BLOCK(@var{exp})
5759 This is a synonym for @code{ALIGN}, for compatibility with older linker
5760 scripts. It is most often seen when setting the address of an output
5763 @item DATA_SEGMENT_ALIGN(@var{maxpagesize}, @var{commonpagesize})
5764 @kindex DATA_SEGMENT_ALIGN(@var{maxpagesize}, @var{commonpagesize})
5765 This is equivalent to either
5767 (ALIGN(@var{maxpagesize}) + (. & (@var{maxpagesize} - 1)))
5771 (ALIGN(@var{maxpagesize}) + (. & (@var{maxpagesize} - @var{commonpagesize})))
5774 depending on whether the latter uses fewer @var{commonpagesize} sized pages
5775 for the data segment (area between the result of this expression and
5776 @code{DATA_SEGMENT_END}) than the former or not.
5777 If the latter form is used, it means @var{commonpagesize} bytes of runtime
5778 memory will be saved at the expense of up to @var{commonpagesize} wasted
5779 bytes in the on-disk file.
5781 This expression can only be used directly in @code{SECTIONS} commands, not in
5782 any output section descriptions and only once in the linker script.
5783 @var{commonpagesize} should be less or equal to @var{maxpagesize} and should
5784 be the system page size the object wants to be optimized for (while still
5785 working on system page sizes up to @var{maxpagesize}).
5790 . = DATA_SEGMENT_ALIGN(0x10000, 0x2000);
5793 @item DATA_SEGMENT_END(@var{exp})
5794 @kindex DATA_SEGMENT_END(@var{exp})
5795 This defines the end of data segment for @code{DATA_SEGMENT_ALIGN}
5796 evaluation purposes.
5799 . = DATA_SEGMENT_END(.);
5802 @item DATA_SEGMENT_RELRO_END(@var{offset}, @var{exp})
5803 @kindex DATA_SEGMENT_RELRO_END(@var{offset}, @var{exp})
5804 This defines the end of the @code{PT_GNU_RELRO} segment when
5805 @samp{-z relro} option is used. Second argument is returned.
5806 When @samp{-z relro} option is not present, @code{DATA_SEGMENT_RELRO_END}
5807 does nothing, otherwise @code{DATA_SEGMENT_ALIGN} is padded so that
5808 @var{exp} + @var{offset} is aligned to the most commonly used page
5809 boundary for particular target. If present in the linker script,
5810 it must always come in between @code{DATA_SEGMENT_ALIGN} and
5811 @code{DATA_SEGMENT_END}.
5814 . = DATA_SEGMENT_RELRO_END(24, .);
5817 @item DEFINED(@var{symbol})
5818 @kindex DEFINED(@var{symbol})
5819 @cindex symbol defaults
5820 Return 1 if @var{symbol} is in the linker global symbol table and is
5821 defined before the statement using DEFINED in the script, otherwise
5822 return 0. You can use this function to provide
5823 default values for symbols. For example, the following script fragment
5824 shows how to set a global symbol @samp{begin} to the first location in
5825 the @samp{.text} section---but if a symbol called @samp{begin} already
5826 existed, its value is preserved:
5832 begin = DEFINED(begin) ? begin : . ;
5840 @item LENGTH(@var{memory})
5841 @kindex LENGTH(@var{memory})
5842 Return the length of the memory region named @var{memory}.
5844 @item LOADADDR(@var{section})
5845 @kindex LOADADDR(@var{section})
5846 @cindex section load address in expression
5847 Return the absolute LMA of the named @var{section}. (@pxref{Output
5851 @item MAX(@var{exp1}, @var{exp2})
5852 Returns the maximum of @var{exp1} and @var{exp2}.
5855 @item MIN(@var{exp1}, @var{exp2})
5856 Returns the minimum of @var{exp1} and @var{exp2}.
5858 @item NEXT(@var{exp})
5859 @kindex NEXT(@var{exp})
5860 @cindex unallocated address, next
5861 Return the next unallocated address that is a multiple of @var{exp}.
5862 This function is closely related to @code{ALIGN(@var{exp})}; unless you
5863 use the @code{MEMORY} command to define discontinuous memory for the
5864 output file, the two functions are equivalent.
5866 @item ORIGIN(@var{memory})
5867 @kindex ORIGIN(@var{memory})
5868 Return the origin of the memory region named @var{memory}.
5870 @item SEGMENT_START(@var{segment}, @var{default})
5871 @kindex SEGMENT_START(@var{segment}, @var{default})
5872 Return the base address of the named @var{segment}. If an explicit
5873 value has been given for this segment (with a command-line @samp{-T}
5874 option) that value will be returned; otherwise the value will be
5875 @var{default}. At present, the @samp{-T} command-line option can only
5876 be used to set the base address for the ``text'', ``data'', and
5877 ``bss'' sections, but you can use @code{SEGMENT_START} with any segment
5880 @item SIZEOF(@var{section})
5881 @kindex SIZEOF(@var{section})
5882 @cindex section size
5883 Return the size in bytes of the named @var{section}, if that section has
5884 been allocated. If the section has not been allocated when this is
5885 evaluated, the linker will report an error. In the following example,
5886 @code{symbol_1} and @code{symbol_2} are assigned identical values:
5895 symbol_1 = .end - .start ;
5896 symbol_2 = SIZEOF(.output);
5901 @item SIZEOF_HEADERS
5902 @itemx sizeof_headers
5903 @kindex SIZEOF_HEADERS
5905 Return the size in bytes of the output file's headers. This is
5906 information which appears at the start of the output file. You can use
5907 this number when setting the start address of the first section, if you
5908 choose, to facilitate paging.
5910 @cindex not enough room for program headers
5911 @cindex program headers, not enough room
5912 When producing an ELF output file, if the linker script uses the
5913 @code{SIZEOF_HEADERS} builtin function, the linker must compute the
5914 number of program headers before it has determined all the section
5915 addresses and sizes. If the linker later discovers that it needs
5916 additional program headers, it will report an error @samp{not enough
5917 room for program headers}. To avoid this error, you must avoid using
5918 the @code{SIZEOF_HEADERS} function, or you must rework your linker
5919 script to avoid forcing the linker to use additional program headers, or
5920 you must define the program headers yourself using the @code{PHDRS}
5921 command (@pxref{PHDRS}).
5924 @node Implicit Linker Scripts
5925 @section Implicit Linker Scripts
5926 @cindex implicit linker scripts
5927 If you specify a linker input file which the linker can not recognize as
5928 an object file or an archive file, it will try to read the file as a
5929 linker script. If the file can not be parsed as a linker script, the
5930 linker will report an error.
5932 An implicit linker script will not replace the default linker script.
5934 Typically an implicit linker script would contain only symbol
5935 assignments, or the @code{INPUT}, @code{GROUP}, or @code{VERSION}
5938 Any input files read because of an implicit linker script will be read
5939 at the position in the command line where the implicit linker script was
5940 read. This can affect archive searching.
5943 @node Machine Dependent
5944 @chapter Machine Dependent Features
5946 @cindex machine dependencies
5947 @command{ld} has additional features on some platforms; the following
5948 sections describe them. Machines where @command{ld} has no additional
5949 functionality are not listed.
5953 * H8/300:: @command{ld} and the H8/300
5956 * i960:: @command{ld} and the Intel 960 family
5959 * ARM:: @command{ld} and the ARM family
5962 * HPPA ELF32:: @command{ld} and HPPA 32-bit ELF
5965 * M68K:: @command{ld} and the Motorola 68K family
5968 * MMIX:: @command{ld} and MMIX
5971 * MSP430:: @command{ld} and MSP430
5974 * M68HC11/68HC12:: @code{ld} and the Motorola 68HC11 and 68HC12 families
5977 * PowerPC ELF32:: @command{ld} and PowerPC 32-bit ELF Support
5980 * PowerPC64 ELF64:: @command{ld} and PowerPC64 64-bit ELF Support
5983 * SPU ELF:: @command{ld} and SPU ELF Support
5986 * TI COFF:: @command{ld} and TI COFF
5989 * WIN32:: @command{ld} and WIN32 (cygwin/mingw)
5992 * Xtensa:: @command{ld} and Xtensa Processors
6003 @section @command{ld} and the H8/300
6005 @cindex H8/300 support
6006 For the H8/300, @command{ld} can perform these global optimizations when
6007 you specify the @samp{--relax} command-line option.
6010 @cindex relaxing on H8/300
6011 @item relaxing address modes
6012 @command{ld} finds all @code{jsr} and @code{jmp} instructions whose
6013 targets are within eight bits, and turns them into eight-bit
6014 program-counter relative @code{bsr} and @code{bra} instructions,
6017 @cindex synthesizing on H8/300
6018 @item synthesizing instructions
6019 @c FIXME: specifically mov.b, or any mov instructions really?
6020 @command{ld} finds all @code{mov.b} instructions which use the
6021 sixteen-bit absolute address form, but refer to the top
6022 page of memory, and changes them to use the eight-bit address form.
6023 (That is: the linker turns @samp{mov.b @code{@@}@var{aa}:16} into
6024 @samp{mov.b @code{@@}@var{aa}:8} whenever the address @var{aa} is in the
6025 top page of memory).
6027 @item bit manipulation instructions
6028 @command{ld} finds all bit manipulation instructions like @code{band, bclr,
6029 biand, bild, bior, bist, bixor, bld, bnot, bor, bset, bst, btst, bxor}
6030 which use 32 bit and 16 bit absolute address form, but refer to the top
6031 page of memory, and changes them to use the 8 bit address form.
6032 (That is: the linker turns @samp{bset #xx:3,@code{@@}@var{aa}:32} into
6033 @samp{bset #xx:3,@code{@@}@var{aa}:8} whenever the address @var{aa} is in
6034 the top page of memory).
6036 @item system control instructions
6037 @command{ld} finds all @code{ldc.w, stc.w} instructions which use the
6038 32 bit absolute address form, but refer to the top page of memory, and
6039 changes them to use 16 bit address form.
6040 (That is: the linker turns @samp{ldc.w @code{@@}@var{aa}:32,ccr} into
6041 @samp{ldc.w @code{@@}@var{aa}:16,ccr} whenever the address @var{aa} is in
6042 the top page of memory).
6052 @c This stuff is pointless to say unless you're especially concerned
6053 @c with Renesas chips; don't enable it for generic case, please.
6055 @chapter @command{ld} and Other Renesas Chips
6057 @command{ld} also supports the Renesas (formerly Hitachi) H8/300H,
6058 H8/500, and SH chips. No special features, commands, or command-line
6059 options are required for these chips.
6069 @section @command{ld} and the Intel 960 Family
6071 @cindex i960 support
6073 You can use the @samp{-A@var{architecture}} command line option to
6074 specify one of the two-letter names identifying members of the 960
6075 family; the option specifies the desired output target, and warns of any
6076 incompatible instructions in the input files. It also modifies the
6077 linker's search strategy for archive libraries, to support the use of
6078 libraries specific to each particular architecture, by including in the
6079 search loop names suffixed with the string identifying the architecture.
6081 For example, if your @command{ld} command line included @w{@samp{-ACA}} as
6082 well as @w{@samp{-ltry}}, the linker would look (in its built-in search
6083 paths, and in any paths you specify with @samp{-L}) for a library with
6096 The first two possibilities would be considered in any event; the last
6097 two are due to the use of @w{@samp{-ACA}}.
6099 You can meaningfully use @samp{-A} more than once on a command line, since
6100 the 960 architecture family allows combination of target architectures; each
6101 use will add another pair of name variants to search for when @w{@samp{-l}}
6102 specifies a library.
6104 @cindex @option{--relax} on i960
6105 @cindex relaxing on i960
6106 @command{ld} supports the @samp{--relax} option for the i960 family. If
6107 you specify @samp{--relax}, @command{ld} finds all @code{balx} and
6108 @code{calx} instructions whose targets are within 24 bits, and turns
6109 them into 24-bit program-counter relative @code{bal} and @code{cal}
6110 instructions, respectively. @command{ld} also turns @code{cal}
6111 instructions into @code{bal} instructions when it determines that the
6112 target subroutine is a leaf routine (that is, the target subroutine does
6113 not itself call any subroutines).
6115 @cindex Cortex-A8 erratum workaround
6116 @kindex --fix-cortex-a8
6117 @kindex --no-fix-cortex-a8
6118 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}.
6120 The erratum only affects Thumb-2 code. Please contact ARM for further details.
6122 @kindex --merge-exidx-entries
6123 @kindex --no-merge-exidx-entries
6124 The @samp{--no-merge-exidx-entries} switch disables the merging of adjacent exidx entries in debuginfo.
6141 @node M68HC11/68HC12
6142 @section @command{ld} and the Motorola 68HC11 and 68HC12 families
6144 @cindex M68HC11 and 68HC12 support
6146 @subsection Linker Relaxation
6148 For the Motorola 68HC11, @command{ld} can perform these global
6149 optimizations when you specify the @samp{--relax} command-line option.
6152 @cindex relaxing on M68HC11
6153 @item relaxing address modes
6154 @command{ld} finds all @code{jsr} and @code{jmp} instructions whose
6155 targets are within eight bits, and turns them into eight-bit
6156 program-counter relative @code{bsr} and @code{bra} instructions,
6159 @command{ld} also looks at all 16-bit extended addressing modes and
6160 transforms them in a direct addressing mode when the address is in
6161 page 0 (between 0 and 0x0ff).
6163 @item relaxing gcc instruction group
6164 When @command{gcc} is called with @option{-mrelax}, it can emit group
6165 of instructions that the linker can optimize to use a 68HC11 direct
6166 addressing mode. These instructions consists of @code{bclr} or
6167 @code{bset} instructions.
6171 @subsection Trampoline Generation
6173 @cindex trampoline generation on M68HC11
6174 @cindex trampoline generation on M68HC12
6175 For 68HC11 and 68HC12, @command{ld} can generate trampoline code to
6176 call a far function using a normal @code{jsr} instruction. The linker
6177 will also change the relocation to some far function to use the
6178 trampoline address instead of the function address. This is typically the
6179 case when a pointer to a function is taken. The pointer will in fact
6180 point to the function trampoline.
6188 @section @command{ld} and the ARM family
6190 @cindex ARM interworking support
6191 @kindex --support-old-code
6192 For the ARM, @command{ld} will generate code stubs to allow functions calls
6193 between ARM and Thumb code. These stubs only work with code that has
6194 been compiled and assembled with the @samp{-mthumb-interwork} command
6195 line option. If it is necessary to link with old ARM object files or
6196 libraries, which have not been compiled with the -mthumb-interwork
6197 option then the @samp{--support-old-code} command line switch should be
6198 given to the linker. This will make it generate larger stub functions
6199 which will work with non-interworking aware ARM code. Note, however,
6200 the linker does not support generating stubs for function calls to
6201 non-interworking aware Thumb code.
6203 @cindex thumb entry point
6204 @cindex entry point, thumb
6205 @kindex --thumb-entry=@var{entry}
6206 The @samp{--thumb-entry} switch is a duplicate of the generic
6207 @samp{--entry} switch, in that it sets the program's starting address.
6208 But it also sets the bottom bit of the address, so that it can be
6209 branched to using a BX instruction, and the program will start
6210 executing in Thumb mode straight away.
6212 @cindex PE import table prefixing
6213 @kindex --use-nul-prefixed-import-tables
6214 The @samp{--use-nul-prefixed-import-tables} switch is specifying, that
6215 the import tables idata4 and idata5 have to be generated with a zero
6216 elememt prefix for import libraries. This is the old style to generate
6217 import tables. By default this option is turned off.
6221 The @samp{--be8} switch instructs @command{ld} to generate BE8 format
6222 executables. This option is only valid when linking big-endian objects.
6223 The resulting image will contain big-endian data and little-endian code.
6226 @kindex --target1-rel
6227 @kindex --target1-abs
6228 The @samp{R_ARM_TARGET1} relocation is typically used for entries in the
6229 @samp{.init_array} section. It is interpreted as either @samp{R_ARM_REL32}
6230 or @samp{R_ARM_ABS32}, depending on the target. The @samp{--target1-rel}
6231 and @samp{--target1-abs} switches override the default.
6234 @kindex --target2=@var{type}
6235 The @samp{--target2=type} switch overrides the default definition of the
6236 @samp{R_ARM_TARGET2} relocation. Valid values for @samp{type}, their
6237 meanings, and target defaults are as follows:
6240 @samp{R_ARM_REL32} (arm*-*-elf, arm*-*-eabi)
6242 @samp{R_ARM_ABS32} (arm*-*-symbianelf)
6244 @samp{R_ARM_GOT_PREL} (arm*-*-linux, arm*-*-*bsd)
6249 The @samp{R_ARM_V4BX} relocation (defined by the ARM AAELF
6250 specification) enables objects compiled for the ARMv4 architecture to be
6251 interworking-safe when linked with other objects compiled for ARMv4t, but
6252 also allows pure ARMv4 binaries to be built from the same ARMv4 objects.
6254 In the latter case, the switch @option{--fix-v4bx} must be passed to the
6255 linker, which causes v4t @code{BX rM} instructions to be rewritten as
6256 @code{MOV PC,rM}, since v4 processors do not have a @code{BX} instruction.
6258 In the former case, the switch should not be used, and @samp{R_ARM_V4BX}
6259 relocations are ignored.
6261 @cindex FIX_V4BX_INTERWORKING
6262 @kindex --fix-v4bx-interworking
6263 Replace @code{BX rM} instructions identified by @samp{R_ARM_V4BX}
6264 relocations with a branch to the following veneer:
6272 This allows generation of libraries/applications that work on ARMv4 cores
6273 and are still interworking safe. Note that the above veneer clobbers the
6274 condition flags, so may cause incorrect progrm behavior in rare cases.
6278 The @samp{--use-blx} switch enables the linker to use ARM/Thumb
6279 BLX instructions (available on ARMv5t and above) in various
6280 situations. Currently it is used to perform calls via the PLT from Thumb
6281 code using BLX rather than using BX and a mode-switching stub before
6282 each PLT entry. This should lead to such calls executing slightly faster.
6284 This option is enabled implicitly for SymbianOS, so there is no need to
6285 specify it if you are using that target.
6287 @cindex VFP11_DENORM_FIX
6288 @kindex --vfp11-denorm-fix
6289 The @samp{--vfp11-denorm-fix} switch enables a link-time workaround for a
6290 bug in certain VFP11 coprocessor hardware, which sometimes allows
6291 instructions with denorm operands (which must be handled by support code)
6292 to have those operands overwritten by subsequent instructions before
6293 the support code can read the intended values.
6295 The bug may be avoided in scalar mode if you allow at least one
6296 intervening instruction between a VFP11 instruction which uses a register
6297 and another instruction which writes to the same register, or at least two
6298 intervening instructions if vector mode is in use. The bug only affects
6299 full-compliance floating-point mode: you do not need this workaround if
6300 you are using "runfast" mode. Please contact ARM for further details.
6302 If you know you are using buggy VFP11 hardware, you can
6303 enable this workaround by specifying the linker option
6304 @samp{--vfp-denorm-fix=scalar} if you are using the VFP11 scalar
6305 mode only, or @samp{--vfp-denorm-fix=vector} if you are using
6306 vector mode (the latter also works for scalar code). The default is
6307 @samp{--vfp-denorm-fix=none}.
6309 If the workaround is enabled, instructions are scanned for
6310 potentially-troublesome sequences, and a veneer is created for each
6311 such sequence which may trigger the erratum. The veneer consists of the
6312 first instruction of the sequence and a branch back to the subsequent
6313 instruction. The original instruction is then replaced with a branch to
6314 the veneer. The extra cycles required to call and return from the veneer
6315 are sufficient to avoid the erratum in both the scalar and vector cases.
6317 @cindex NO_ENUM_SIZE_WARNING
6318 @kindex --no-enum-size-warning
6319 The @option{--no-enum-size-warning} switch prevents the linker from
6320 warning when linking object files that specify incompatible EABI
6321 enumeration size attributes. For example, with this switch enabled,
6322 linking of an object file using 32-bit enumeration values with another
6323 using enumeration values fitted into the smallest possible space will
6326 @cindex NO_WCHAR_SIZE_WARNING
6327 @kindex --no-wchar-size-warning
6328 The @option{--no-wchar-size-warning} switch prevents the linker from
6329 warning when linking object files that specify incompatible EABI
6330 @code{wchar_t} size attributes. For example, with this switch enabled,
6331 linking of an object file using 32-bit @code{wchar_t} values with another
6332 using 16-bit @code{wchar_t} values will not be diagnosed.
6335 @kindex --pic-veneer
6336 The @samp{--pic-veneer} switch makes the linker use PIC sequences for
6337 ARM/Thumb interworking veneers, even if the rest of the binary
6338 is not PIC. This avoids problems on uClinux targets where
6339 @samp{--emit-relocs} is used to generate relocatable binaries.
6341 @cindex STUB_GROUP_SIZE
6342 @kindex --stub-group-size=@var{N}
6343 The linker will automatically generate and insert small sequences of
6344 code into a linked ARM ELF executable whenever an attempt is made to
6345 perform a function call to a symbol that is too far away. The
6346 placement of these sequences of instructions - called stubs - is
6347 controlled by the command line option @option{--stub-group-size=N}.
6348 The placement is important because a poor choice can create a need for
6349 duplicate stubs, increasing the code sizw. The linker will try to
6350 group stubs together in order to reduce interruptions to the flow of
6351 code, but it needs guidance as to how big these groups should be and
6352 where they should be placed.
6354 The value of @samp{N}, the parameter to the
6355 @option{--stub-group-size=} option controls where the stub groups are
6356 placed. If it is negative then all stubs are placed after the first
6357 branch that needs them. If it is positive then the stubs can be
6358 placed either before or after the branches that need them. If the
6359 value of @samp{N} is 1 (either +1 or -1) then the linker will choose
6360 exactly where to place groups of stubs, using its built in heuristics.
6361 A value of @samp{N} greater than 1 (or smaller than -1) tells the
6362 linker that a single group of stubs can service at most @samp{N} bytes
6363 from the input sections.
6365 The default, if @option{--stub-group-size=} is not specified, is
6368 Farcalls stubs insertion is fully supported for the ARM-EABI target
6369 only, because it relies on object files properties not present
6383 @section @command{ld} and HPPA 32-bit ELF Support
6384 @cindex HPPA multiple sub-space stubs
6385 @kindex --multi-subspace
6386 When generating a shared library, @command{ld} will by default generate
6387 import stubs suitable for use with a single sub-space application.
6388 The @samp{--multi-subspace} switch causes @command{ld} to generate export
6389 stubs, and different (larger) import stubs suitable for use with
6390 multiple sub-spaces.
6392 @cindex HPPA stub grouping
6393 @kindex --stub-group-size=@var{N}
6394 Long branch stubs and import/export stubs are placed by @command{ld} in
6395 stub sections located between groups of input sections.
6396 @samp{--stub-group-size} specifies the maximum size of a group of input
6397 sections handled by one stub section. Since branch offsets are signed,
6398 a stub section may serve two groups of input sections, one group before
6399 the stub section, and one group after it. However, when using
6400 conditional branches that require stubs, it may be better (for branch
6401 prediction) that stub sections only serve one group of input sections.
6402 A negative value for @samp{N} chooses this scheme, ensuring that
6403 branches to stubs always use a negative offset. Two special values of
6404 @samp{N} are recognized, @samp{1} and @samp{-1}. These both instruct
6405 @command{ld} to automatically size input section groups for the branch types
6406 detected, with the same behaviour regarding stub placement as other
6407 positive or negative values of @samp{N} respectively.
6409 Note that @samp{--stub-group-size} does not split input sections. A
6410 single input section larger than the group size specified will of course
6411 create a larger group (of one section). If input sections are too
6412 large, it may not be possible for a branch to reach its stub.
6425 @section @command{ld} and the Motorola 68K family
6427 @cindex Motorola 68K GOT generation
6428 @kindex --got=@var{type}
6429 The @samp{--got=@var{type}} option lets you choose the GOT generation scheme.
6430 The choices are @samp{single}, @samp{negative}, @samp{multigot} and
6431 @samp{target}. When @samp{target} is selected the linker chooses
6432 the default GOT generation scheme for the current target.
6433 @samp{single} tells the linker to generate a single GOT with
6434 entries only at non-negative offsets.
6435 @samp{negative} instructs the linker to generate a single GOT with
6436 entries at both negative and positive offsets. Not all environments
6438 @samp{multigot} allows the linker to generate several GOTs in the
6439 output file. All GOT references from a single input object
6440 file access the same GOT, but references from different input object
6441 files might access different GOTs. Not all environments support such GOTs.
6454 @section @code{ld} and MMIX
6455 For MMIX, there is a choice of generating @code{ELF} object files or
6456 @code{mmo} object files when linking. The simulator @code{mmix}
6457 understands the @code{mmo} format. The binutils @code{objcopy} utility
6458 can translate between the two formats.
6460 There is one special section, the @samp{.MMIX.reg_contents} section.
6461 Contents in this section is assumed to correspond to that of global
6462 registers, and symbols referring to it are translated to special symbols,
6463 equal to registers. In a final link, the start address of the
6464 @samp{.MMIX.reg_contents} section corresponds to the first allocated
6465 global register multiplied by 8. Register @code{$255} is not included in
6466 this section; it is always set to the program entry, which is at the
6467 symbol @code{Main} for @code{mmo} files.
6469 Global symbols with the prefix @code{__.MMIX.start.}, for example
6470 @code{__.MMIX.start..text} and @code{__.MMIX.start..data} are special.
6471 The default linker script uses these to set the default start address
6474 Initial and trailing multiples of zero-valued 32-bit words in a section,
6475 are left out from an mmo file.
6488 @section @code{ld} and MSP430
6489 For the MSP430 it is possible to select the MPU architecture. The flag @samp{-m [mpu type]}
6490 will select an appropriate linker script for selected MPU type. (To get a list of known MPUs
6491 just pass @samp{-m help} option to the linker).
6493 @cindex MSP430 extra sections
6494 The linker will recognize some extra sections which are MSP430 specific:
6497 @item @samp{.vectors}
6498 Defines a portion of ROM where interrupt vectors located.
6500 @item @samp{.bootloader}
6501 Defines the bootloader portion of the ROM (if applicable). Any code
6502 in this section will be uploaded to the MPU.
6504 @item @samp{.infomem}
6505 Defines an information memory section (if applicable). Any code in
6506 this section will be uploaded to the MPU.
6508 @item @samp{.infomemnobits}
6509 This is the same as the @samp{.infomem} section except that any code
6510 in this section will not be uploaded to the MPU.
6512 @item @samp{.noinit}
6513 Denotes a portion of RAM located above @samp{.bss} section.
6515 The last two sections are used by gcc.
6529 @section @command{ld} and PowerPC 32-bit ELF Support
6530 @cindex PowerPC long branches
6531 @kindex --relax on PowerPC
6532 Branches on PowerPC processors are limited to a signed 26-bit
6533 displacement, which may result in @command{ld} giving
6534 @samp{relocation truncated to fit} errors with very large programs.
6535 @samp{--relax} enables the generation of trampolines that can access
6536 the entire 32-bit address space. These trampolines are inserted at
6537 section boundaries, so may not themselves be reachable if an input
6538 section exceeds 33M in size. You may combine @samp{-r} and
6539 @samp{--relax} to add trampolines in a partial link. In that case
6540 both branches to undefined symbols and inter-section branches are also
6541 considered potentially out of range, and trampolines inserted.
6543 @cindex PowerPC ELF32 options
6548 Current PowerPC GCC accepts a @samp{-msecure-plt} option that
6549 generates code capable of using a newer PLT and GOT layout that has
6550 the security advantage of no executable section ever needing to be
6551 writable and no writable section ever being executable. PowerPC
6552 @command{ld} will generate this layout, including stubs to access the
6553 PLT, if all input files (including startup and static libraries) were
6554 compiled with @samp{-msecure-plt}. @samp{--bss-plt} forces the old
6555 BSS PLT (and GOT layout) which can give slightly better performance.
6557 @kindex --secure-plt
6559 @command{ld} will use the new PLT and GOT layout if it is linking new
6560 @samp{-fpic} or @samp{-fPIC} code, but does not do so automatically
6561 when linking non-PIC code. This option requests the new PLT and GOT
6562 layout. A warning will be given if some object file requires the old
6568 The new secure PLT and GOT are placed differently relative to other
6569 sections compared to older BSS PLT and GOT placement. The location of
6570 @code{.plt} must change because the new secure PLT is an initialized
6571 section while the old PLT is uninitialized. The reason for the
6572 @code{.got} change is more subtle: The new placement allows
6573 @code{.got} to be read-only in applications linked with
6574 @samp{-z relro -z now}. However, this placement means that
6575 @code{.sdata} cannot always be used in shared libraries, because the
6576 PowerPC ABI accesses @code{.sdata} in shared libraries from the GOT
6577 pointer. @samp{--sdata-got} forces the old GOT placement. PowerPC
6578 GCC doesn't use @code{.sdata} in shared libraries, so this option is
6579 really only useful for other compilers that may do so.
6581 @cindex PowerPC stub symbols
6582 @kindex --emit-stub-syms
6583 @item --emit-stub-syms
6584 This option causes @command{ld} to label linker stubs with a local
6585 symbol that encodes the stub type and destination.
6587 @cindex PowerPC TLS optimization
6588 @kindex --no-tls-optimize
6589 @item --no-tls-optimize
6590 PowerPC @command{ld} normally performs some optimization of code
6591 sequences used to access Thread-Local Storage. Use this option to
6592 disable the optimization.
6605 @node PowerPC64 ELF64
6606 @section @command{ld} and PowerPC64 64-bit ELF Support
6608 @cindex PowerPC64 ELF64 options
6610 @cindex PowerPC64 stub grouping
6611 @kindex --stub-group-size
6612 @item --stub-group-size
6613 Long branch stubs, PLT call stubs and TOC adjusting stubs are placed
6614 by @command{ld} in stub sections located between groups of input sections.
6615 @samp{--stub-group-size} specifies the maximum size of a group of input
6616 sections handled by one stub section. Since branch offsets are signed,
6617 a stub section may serve two groups of input sections, one group before
6618 the stub section, and one group after it. However, when using
6619 conditional branches that require stubs, it may be better (for branch
6620 prediction) that stub sections only serve one group of input sections.
6621 A negative value for @samp{N} chooses this scheme, ensuring that
6622 branches to stubs always use a negative offset. Two special values of
6623 @samp{N} are recognized, @samp{1} and @samp{-1}. These both instruct
6624 @command{ld} to automatically size input section groups for the branch types
6625 detected, with the same behaviour regarding stub placement as other
6626 positive or negative values of @samp{N} respectively.
6628 Note that @samp{--stub-group-size} does not split input sections. A
6629 single input section larger than the group size specified will of course
6630 create a larger group (of one section). If input sections are too
6631 large, it may not be possible for a branch to reach its stub.
6633 @cindex PowerPC64 stub symbols
6634 @kindex --emit-stub-syms
6635 @item --emit-stub-syms
6636 This option causes @command{ld} to label linker stubs with a local
6637 symbol that encodes the stub type and destination.
6639 @cindex PowerPC64 dot symbols
6641 @kindex --no-dotsyms
6642 @item --dotsyms, --no-dotsyms
6643 These two options control how @command{ld} interprets version patterns
6644 in a version script. Older PowerPC64 compilers emitted both a
6645 function descriptor symbol with the same name as the function, and a
6646 code entry symbol with the name prefixed by a dot (@samp{.}). To
6647 properly version a function @samp{foo}, the version script thus needs
6648 to control both @samp{foo} and @samp{.foo}. The option
6649 @samp{--dotsyms}, on by default, automatically adds the required
6650 dot-prefixed patterns. Use @samp{--no-dotsyms} to disable this
6653 @cindex PowerPC64 TLS optimization
6654 @kindex --no-tls-optimize
6655 @item --no-tls-optimize
6656 PowerPC64 @command{ld} normally performs some optimization of code
6657 sequences used to access Thread-Local Storage. Use this option to
6658 disable the optimization.
6660 @cindex PowerPC64 OPD optimization
6661 @kindex --no-opd-optimize
6662 @item --no-opd-optimize
6663 PowerPC64 @command{ld} normally removes @code{.opd} section entries
6664 corresponding to deleted link-once functions, or functions removed by
6665 the action of @samp{--gc-sections} or linker script @code{/DISCARD/}.
6666 Use this option to disable @code{.opd} optimization.
6668 @cindex PowerPC64 OPD spacing
6669 @kindex --non-overlapping-opd
6670 @item --non-overlapping-opd
6671 Some PowerPC64 compilers have an option to generate compressed
6672 @code{.opd} entries spaced 16 bytes apart, overlapping the third word,
6673 the static chain pointer (unused in C) with the first word of the next
6674 entry. This option expands such entries to the full 24 bytes.
6676 @cindex PowerPC64 TOC optimization
6677 @kindex --no-toc-optimize
6678 @item --no-toc-optimize
6679 PowerPC64 @command{ld} normally removes unused @code{.toc} section
6680 entries. Such entries are detected by examining relocations that
6681 reference the TOC in code sections. A reloc in a deleted code section
6682 marks a TOC word as unneeded, while a reloc in a kept code section
6683 marks a TOC word as needed. Since the TOC may reference itself, TOC
6684 relocs are also examined. TOC words marked as both needed and
6685 unneeded will of course be kept. TOC words without any referencing
6686 reloc are assumed to be part of a multi-word entry, and are kept or
6687 discarded as per the nearest marked preceding word. This works
6688 reliably for compiler generated code, but may be incorrect if assembly
6689 code is used to insert TOC entries. Use this option to disable the
6692 @cindex PowerPC64 multi-TOC
6693 @kindex --no-multi-toc
6694 @item --no-multi-toc
6695 By default, PowerPC64 GCC generates code for a TOC model where TOC
6696 entries are accessed with a 16-bit offset from r2. This limits the
6697 total TOC size to 64K. PowerPC64 @command{ld} extends this limit by
6698 grouping code sections such that each group uses less than 64K for its
6699 TOC entries, then inserts r2 adjusting stubs between inter-group
6700 calls. @command{ld} does not split apart input sections, so cannot
6701 help if a single input file has a @code{.toc} section that exceeds
6702 64K, most likely from linking multiple files with @command{ld -r}.
6703 Use this option to turn off this feature.
6717 @section @command{ld} and SPU ELF Support
6719 @cindex SPU ELF options
6725 This option marks an executable as a PIC plugin module.
6727 @cindex SPU overlays
6728 @kindex --no-overlays
6730 Normally, @command{ld} recognizes calls to functions within overlay
6731 regions, and redirects such calls to an overlay manager via a stub.
6732 @command{ld} also provides a built-in overlay manager. This option
6733 turns off all this special overlay handling.
6735 @cindex SPU overlay stub symbols
6736 @kindex --emit-stub-syms
6737 @item --emit-stub-syms
6738 This option causes @command{ld} to label overlay stubs with a local
6739 symbol that encodes the stub type and destination.
6741 @cindex SPU extra overlay stubs
6742 @kindex --extra-overlay-stubs
6743 @item --extra-overlay-stubs
6744 This option causes @command{ld} to add overlay call stubs on all
6745 function calls out of overlay regions. Normally stubs are not added
6746 on calls to non-overlay regions.
6748 @cindex SPU local store size
6749 @kindex --local-store=lo:hi
6750 @item --local-store=lo:hi
6751 @command{ld} usually checks that a final executable for SPU fits in
6752 the address range 0 to 256k. This option may be used to change the
6753 range. Disable the check entirely with @option{--local-store=0:0}.
6756 @kindex --stack-analysis
6757 @item --stack-analysis
6758 SPU local store space is limited. Over-allocation of stack space
6759 unnecessarily limits space available for code and data, while
6760 under-allocation results in runtime failures. If given this option,
6761 @command{ld} will provide an estimate of maximum stack usage.
6762 @command{ld} does this by examining symbols in code sections to
6763 determine the extents of functions, and looking at function prologues
6764 for stack adjusting instructions. A call-graph is created by looking
6765 for relocations on branch instructions. The graph is then searched
6766 for the maximum stack usage path. Note that this analysis does not
6767 find calls made via function pointers, and does not handle recursion
6768 and other cycles in the call graph. Stack usage may be
6769 under-estimated if your code makes such calls. Also, stack usage for
6770 dynamic allocation, e.g. alloca, will not be detected. If a link map
6771 is requested, detailed information about each function's stack usage
6772 and calls will be given.
6775 @kindex --emit-stack-syms
6776 @item --emit-stack-syms
6777 This option, if given along with @option{--stack-analysis} will result
6778 in @command{ld} emitting stack sizing symbols for each function.
6779 These take the form @code{__stack_<function_name>} for global
6780 functions, and @code{__stack_<number>_<function_name>} for static
6781 functions. @code{<number>} is the section id in hex. The value of
6782 such symbols is the stack requirement for the corresponding function.
6783 The symbol size will be zero, type @code{STT_NOTYPE}, binding
6784 @code{STB_LOCAL}, and section @code{SHN_ABS}.
6798 @section @command{ld}'s Support for Various TI COFF Versions
6799 @cindex TI COFF versions
6800 @kindex --format=@var{version}
6801 The @samp{--format} switch allows selection of one of the various
6802 TI COFF versions. The latest of this writing is 2; versions 0 and 1 are
6803 also supported. The TI COFF versions also vary in header byte-order
6804 format; @command{ld} will read any version or byte order, but the output
6805 header format depends on the default specified by the specific target.
6818 @section @command{ld} and WIN32 (cygwin/mingw)
6820 This section describes some of the win32 specific @command{ld} issues.
6821 See @ref{Options,,Command Line Options} for detailed description of the
6822 command line options mentioned here.
6825 @cindex import libraries
6826 @item import libraries
6827 The standard Windows linker creates and uses so-called import
6828 libraries, which contains information for linking to dll's. They are
6829 regular static archives and are handled as any other static
6830 archive. The cygwin and mingw ports of @command{ld} have specific
6831 support for creating such libraries provided with the
6832 @samp{--out-implib} command line option.
6834 @item exporting DLL symbols
6835 @cindex exporting DLL symbols
6836 The cygwin/mingw @command{ld} has several ways to export symbols for dll's.
6839 @item using auto-export functionality
6840 @cindex using auto-export functionality
6841 By default @command{ld} exports symbols with the auto-export functionality,
6842 which is controlled by the following command line options:
6845 @item --export-all-symbols [This is the default]
6846 @item --exclude-symbols
6847 @item --exclude-libs
6848 @item --exclude-modules-for-implib
6849 @item --version-script
6852 When auto-export is in operation, @command{ld} will export all the non-local
6853 (global and common) symbols it finds in a DLL, with the exception of a few
6854 symbols known to belong to the system's runtime and libraries. As it will
6855 often not be desirable to export all of a DLL's symbols, which may include
6856 private functions that are not part of any public interface, the command-line
6857 options listed above may be used to filter symbols out from the list for
6858 exporting. The @samp{--output-def} option can be used in order to see the
6859 final list of exported symbols with all exclusions taken into effect.
6861 If @samp{--export-all-symbols} is not given explicitly on the
6862 command line, then the default auto-export behavior will be @emph{disabled}
6863 if either of the following are true:
6866 @item A DEF file is used.
6867 @item Any symbol in any object file was marked with the __declspec(dllexport) attribute.
6870 @item using a DEF file
6871 @cindex using a DEF file
6872 Another way of exporting symbols is using a DEF file. A DEF file is
6873 an ASCII file containing definitions of symbols which should be
6874 exported when a dll is created. Usually it is named @samp{<dll
6875 name>.def} and is added as any other object file to the linker's
6876 command line. The file's name must end in @samp{.def} or @samp{.DEF}.
6879 gcc -o <output> <objectfiles> <dll name>.def
6882 Using a DEF file turns off the normal auto-export behavior, unless the
6883 @samp{--export-all-symbols} option is also used.
6885 Here is an example of a DEF file for a shared library called @samp{xyz.dll}:
6888 LIBRARY "xyz.dll" BASE=0x20000000
6894 another_foo = abc.dll.afoo
6900 This example defines a DLL with a non-default base address and seven
6901 symbols in the export table. The third exported symbol @code{_bar} is an
6902 alias for the second. The fourth symbol, @code{another_foo} is resolved
6903 by "forwarding" to another module and treating it as an alias for
6904 @code{afoo} exported from the DLL @samp{abc.dll}. The final symbol
6905 @code{var1} is declared to be a data object. The @samp{doo} symbol in
6906 export library is an alias of @samp{foo}, which gets the string name
6907 in export table @samp{foo2}. The @samp{eoo} symbol is an data export
6908 symbol, which gets in export table the name @samp{var1}.
6910 The optional @code{LIBRARY <name>} command indicates the @emph{internal}
6911 name of the output DLL. If @samp{<name>} does not include a suffix,
6912 the default library suffix, @samp{.DLL} is appended.
6914 When the .DEF file is used to build an application, rather than a
6915 library, the @code{NAME <name>} command should be used instead of
6916 @code{LIBRARY}. If @samp{<name>} does not include a suffix, the default
6917 executable suffix, @samp{.EXE} is appended.
6919 With either @code{LIBRARY <name>} or @code{NAME <name>} the optional
6920 specification @code{BASE = <number>} may be used to specify a
6921 non-default base address for the image.
6923 If neither @code{LIBRARY <name>} nor @code{NAME <name>} is specified,
6924 or they specify an empty string, the internal name is the same as the
6925 filename specified on the command line.
6927 The complete specification of an export symbol is:
6931 ( ( ( <name1> [ = <name2> ] )
6932 | ( <name1> = <module-name> . <external-name>))
6933 [ @@ <integer> ] [NONAME] [DATA] [CONSTANT] [PRIVATE] [== <name3>] ) *
6936 Declares @samp{<name1>} as an exported symbol from the DLL, or declares
6937 @samp{<name1>} as an exported alias for @samp{<name2>}; or declares
6938 @samp{<name1>} as a "forward" alias for the symbol
6939 @samp{<external-name>} in the DLL @samp{<module-name>}.
6940 Optionally, the symbol may be exported by the specified ordinal
6941 @samp{<integer>} alias. The optional @samp{<name3>} is the to be used
6942 string in import/export table for the symbol.
6944 The optional keywords that follow the declaration indicate:
6946 @code{NONAME}: Do not put the symbol name in the DLL's export table. It
6947 will still be exported by its ordinal alias (either the value specified
6948 by the .def specification or, otherwise, the value assigned by the
6949 linker). The symbol name, however, does remain visible in the import
6950 library (if any), unless @code{PRIVATE} is also specified.
6952 @code{DATA}: The symbol is a variable or object, rather than a function.
6953 The import lib will export only an indirect reference to @code{foo} as
6954 the symbol @code{_imp__foo} (ie, @code{foo} must be resolved as
6957 @code{CONSTANT}: Like @code{DATA}, but put the undecorated @code{foo} as
6958 well as @code{_imp__foo} into the import library. Both refer to the
6959 read-only import address table's pointer to the variable, not to the
6960 variable itself. This can be dangerous. If the user code fails to add
6961 the @code{dllimport} attribute and also fails to explicitly add the
6962 extra indirection that the use of the attribute enforces, the
6963 application will behave unexpectedly.
6965 @code{PRIVATE}: Put the symbol in the DLL's export table, but do not put
6966 it into the static import library used to resolve imports at link time. The
6967 symbol can still be imported using the @code{LoadLibrary/GetProcAddress}
6968 API at runtime or by by using the GNU ld extension of linking directly to
6969 the DLL without an import library.
6971 See ld/deffilep.y in the binutils sources for the full specification of
6972 other DEF file statements
6974 @cindex creating a DEF file
6975 While linking a shared dll, @command{ld} is able to create a DEF file
6976 with the @samp{--output-def <file>} command line option.
6978 @item Using decorations
6979 @cindex Using decorations
6980 Another way of marking symbols for export is to modify the source code
6981 itself, so that when building the DLL each symbol to be exported is
6985 __declspec(dllexport) int a_variable
6986 __declspec(dllexport) void a_function(int with_args)
6989 All such symbols will be exported from the DLL. If, however,
6990 any of the object files in the DLL contain symbols decorated in
6991 this way, then the normal auto-export behavior is disabled, unless
6992 the @samp{--export-all-symbols} option is also used.
6994 Note that object files that wish to access these symbols must @emph{not}
6995 decorate them with dllexport. Instead, they should use dllimport,
6999 __declspec(dllimport) int a_variable
7000 __declspec(dllimport) void a_function(int with_args)
7003 This complicates the structure of library header files, because
7004 when included by the library itself the header must declare the
7005 variables and functions as dllexport, but when included by client
7006 code the header must declare them as dllimport. There are a number
7007 of idioms that are typically used to do this; often client code can
7008 omit the __declspec() declaration completely. See
7009 @samp{--enable-auto-import} and @samp{automatic data imports} for more
7013 @cindex automatic data imports
7014 @item automatic data imports
7015 The standard Windows dll format supports data imports from dlls only
7016 by adding special decorations (dllimport/dllexport), which let the
7017 compiler produce specific assembler instructions to deal with this
7018 issue. This increases the effort necessary to port existing Un*x
7019 code to these platforms, especially for large
7020 c++ libraries and applications. The auto-import feature, which was
7021 initially provided by Paul Sokolovsky, allows one to omit the
7022 decorations to achieve a behavior that conforms to that on POSIX/Un*x
7023 platforms. This feature is enabled with the @samp{--enable-auto-import}
7024 command-line option, although it is enabled by default on cygwin/mingw.
7025 The @samp{--enable-auto-import} option itself now serves mainly to
7026 suppress any warnings that are ordinarily emitted when linked objects
7027 trigger the feature's use.
7029 auto-import of variables does not always work flawlessly without
7030 additional assistance. Sometimes, you will see this message
7032 "variable '<var>' can't be auto-imported. Please read the
7033 documentation for ld's @code{--enable-auto-import} for details."
7035 The @samp{--enable-auto-import} documentation explains why this error
7036 occurs, and several methods that can be used to overcome this difficulty.
7037 One of these methods is the @emph{runtime pseudo-relocs} feature, described
7040 @cindex runtime pseudo-relocation
7041 For complex variables imported from DLLs (such as structs or classes),
7042 object files typically contain a base address for the variable and an
7043 offset (@emph{addend}) within the variable--to specify a particular
7044 field or public member, for instance. Unfortunately, the runtime loader used
7045 in win32 environments is incapable of fixing these references at runtime
7046 without the additional information supplied by dllimport/dllexport decorations.
7047 The standard auto-import feature described above is unable to resolve these
7050 The @samp{--enable-runtime-pseudo-relocs} switch allows these references to
7051 be resolved without error, while leaving the task of adjusting the references
7052 themselves (with their non-zero addends) to specialized code provided by the
7053 runtime environment. Recent versions of the cygwin and mingw environments and
7054 compilers provide this runtime support; older versions do not. However, the
7055 support is only necessary on the developer's platform; the compiled result will
7056 run without error on an older system.
7058 @samp{--enable-runtime-pseudo-relocs} is not the default; it must be explicitly
7061 @cindex direct linking to a dll
7062 @item direct linking to a dll
7063 The cygwin/mingw ports of @command{ld} support the direct linking,
7064 including data symbols, to a dll without the usage of any import
7065 libraries. This is much faster and uses much less memory than does the
7066 traditional import library method, especially when linking large
7067 libraries or applications. When @command{ld} creates an import lib, each
7068 function or variable exported from the dll is stored in its own bfd, even
7069 though a single bfd could contain many exports. The overhead involved in
7070 storing, loading, and processing so many bfd's is quite large, and explains the
7071 tremendous time, memory, and storage needed to link against particularly
7072 large or complex libraries when using import libs.
7074 Linking directly to a dll uses no extra command-line switches other than
7075 @samp{-L} and @samp{-l}, because @command{ld} already searches for a number
7076 of names to match each library. All that is needed from the developer's
7077 perspective is an understanding of this search, in order to force ld to
7078 select the dll instead of an import library.
7081 For instance, when ld is called with the argument @samp{-lxxx} it will attempt
7082 to find, in the first directory of its search path,
7094 before moving on to the next directory in the search path.
7096 (*) Actually, this is not @samp{cygxxx.dll} but in fact is @samp{<prefix>xxx.dll},
7097 where @samp{<prefix>} is set by the @command{ld} option
7098 @samp{--dll-search-prefix=<prefix>}. In the case of cygwin, the standard gcc spec
7099 file includes @samp{--dll-search-prefix=cyg}, so in effect we actually search for
7102 Other win32-based unix environments, such as mingw or pw32, may use other
7103 @samp{<prefix>}es, although at present only cygwin makes use of this feature. It
7104 was originally intended to help avoid name conflicts among dll's built for the
7105 various win32/un*x environments, so that (for example) two versions of a zlib dll
7106 could coexist on the same machine.
7108 The generic cygwin/mingw path layout uses a @samp{bin} directory for
7109 applications and dll's and a @samp{lib} directory for the import
7110 libraries (using cygwin nomenclature):
7116 libxxx.dll.a (in case of dll's)
7117 libxxx.a (in case of static archive)
7120 Linking directly to a dll without using the import library can be
7123 1. Use the dll directly by adding the @samp{bin} path to the link line
7125 gcc -Wl,-verbose -o a.exe -L../bin/ -lxxx
7128 However, as the dll's often have version numbers appended to their names
7129 (@samp{cygncurses-5.dll}) this will often fail, unless one specifies
7130 @samp{-L../bin -lncurses-5} to include the version. Import libs are generally
7131 not versioned, and do not have this difficulty.
7133 2. Create a symbolic link from the dll to a file in the @samp{lib}
7134 directory according to the above mentioned search pattern. This
7135 should be used to avoid unwanted changes in the tools needed for
7139 ln -s bin/cygxxx.dll lib/[cyg|lib|]xxx.dll[.a]
7142 Then you can link without any make environment changes.
7145 gcc -Wl,-verbose -o a.exe -L../lib/ -lxxx
7148 This technique also avoids the version number problems, because the following is
7155 libxxx.dll.a -> ../bin/cygxxx-5.dll
7158 Linking directly to a dll without using an import lib will work
7159 even when auto-import features are exercised, and even when
7160 @samp{--enable-runtime-pseudo-relocs} is used.
7162 Given the improvements in speed and memory usage, one might justifiably
7163 wonder why import libraries are used at all. There are three reasons:
7165 1. Until recently, the link-directly-to-dll functionality did @emph{not}
7166 work with auto-imported data.
7168 2. Sometimes it is necessary to include pure static objects within the
7169 import library (which otherwise contains only bfd's for indirection
7170 symbols that point to the exports of a dll). Again, the import lib
7171 for the cygwin kernel makes use of this ability, and it is not
7172 possible to do this without an import lib.
7174 3. Symbol aliases can only be resolved using an import lib. This is
7175 critical when linking against OS-supplied dll's (eg, the win32 API)
7176 in which symbols are usually exported as undecorated aliases of their
7177 stdcall-decorated assembly names.
7179 So, import libs are not going away. But the ability to replace
7180 true import libs with a simple symbolic link to (or a copy of)
7181 a dll, in many cases, is a useful addition to the suite of tools
7182 binutils makes available to the win32 developer. Given the
7183 massive improvements in memory requirements during linking, storage
7184 requirements, and linking speed, we expect that many developers
7185 will soon begin to use this feature whenever possible.
7187 @item symbol aliasing
7189 @item adding additional names
7190 Sometimes, it is useful to export symbols with additional names.
7191 A symbol @samp{foo} will be exported as @samp{foo}, but it can also be
7192 exported as @samp{_foo} by using special directives in the DEF file
7193 when creating the dll. This will affect also the optional created
7194 import library. Consider the following DEF file:
7197 LIBRARY "xyz.dll" BASE=0x61000000
7204 The line @samp{_foo = foo} maps the symbol @samp{foo} to @samp{_foo}.
7206 Another method for creating a symbol alias is to create it in the
7207 source code using the "weak" attribute:
7210 void foo () @{ /* Do something. */; @}
7211 void _foo () __attribute__ ((weak, alias ("foo")));
7214 See the gcc manual for more information about attributes and weak
7217 @item renaming symbols
7218 Sometimes it is useful to rename exports. For instance, the cygwin
7219 kernel does this regularly. A symbol @samp{_foo} can be exported as
7220 @samp{foo} but not as @samp{_foo} by using special directives in the
7221 DEF file. (This will also affect the import library, if it is
7222 created). In the following example:
7225 LIBRARY "xyz.dll" BASE=0x61000000
7231 The line @samp{_foo = foo} maps the exported symbol @samp{foo} to
7235 Note: using a DEF file disables the default auto-export behavior,
7236 unless the @samp{--export-all-symbols} command line option is used.
7237 If, however, you are trying to rename symbols, then you should list
7238 @emph{all} desired exports in the DEF file, including the symbols
7239 that are not being renamed, and do @emph{not} use the
7240 @samp{--export-all-symbols} option. If you list only the
7241 renamed symbols in the DEF file, and use @samp{--export-all-symbols}
7242 to handle the other symbols, then the both the new names @emph{and}
7243 the original names for the renamed symbols will be exported.
7244 In effect, you'd be aliasing those symbols, not renaming them,
7245 which is probably not what you wanted.
7247 @cindex weak externals
7248 @item weak externals
7249 The Windows object format, PE, specifies a form of weak symbols called
7250 weak externals. When a weak symbol is linked and the symbol is not
7251 defined, the weak symbol becomes an alias for some other symbol. There
7252 are three variants of weak externals:
7254 @item Definition is searched for in objects and libraries, historically
7255 called lazy externals.
7256 @item Definition is searched for only in other objects, not in libraries.
7257 This form is not presently implemented.
7258 @item No search; the symbol is an alias. This form is not presently
7261 As a GNU extension, weak symbols that do not specify an alternate symbol
7262 are supported. If the symbol is undefined when linking, the symbol
7263 uses a default value.
7265 @cindex aligned common symbols
7266 @item aligned common symbols
7267 As a GNU extension to the PE file format, it is possible to specify the
7268 desired alignment for a common symbol. This information is conveyed from
7269 the assembler or compiler to the linker by means of GNU-specific commands
7270 carried in the object file's @samp{.drectve} section, which are recognized
7271 by @command{ld} and respected when laying out the common symbols. Native
7272 tools will be able to process object files employing this GNU extension,
7273 but will fail to respect the alignment instructions, and may issue noisy
7274 warnings about unknown linker directives.
7288 @section @code{ld} and Xtensa Processors
7290 @cindex Xtensa processors
7291 The default @command{ld} behavior for Xtensa processors is to interpret
7292 @code{SECTIONS} commands so that lists of explicitly named sections in a
7293 specification with a wildcard file will be interleaved when necessary to
7294 keep literal pools within the range of PC-relative load offsets. For
7295 example, with the command:
7307 @command{ld} may interleave some of the @code{.literal}
7308 and @code{.text} sections from different object files to ensure that the
7309 literal pools are within the range of PC-relative load offsets. A valid
7310 interleaving might place the @code{.literal} sections from an initial
7311 group of files followed by the @code{.text} sections of that group of
7312 files. Then, the @code{.literal} sections from the rest of the files
7313 and the @code{.text} sections from the rest of the files would follow.
7315 @cindex @option{--relax} on Xtensa
7316 @cindex relaxing on Xtensa
7317 Relaxation is enabled by default for the Xtensa version of @command{ld} and
7318 provides two important link-time optimizations. The first optimization
7319 is to combine identical literal values to reduce code size. A redundant
7320 literal will be removed and all the @code{L32R} instructions that use it
7321 will be changed to reference an identical literal, as long as the
7322 location of the replacement literal is within the offset range of all
7323 the @code{L32R} instructions. The second optimization is to remove
7324 unnecessary overhead from assembler-generated ``longcall'' sequences of
7325 @code{L32R}/@code{CALLX@var{n}} when the target functions are within
7326 range of direct @code{CALL@var{n}} instructions.
7328 For each of these cases where an indirect call sequence can be optimized
7329 to a direct call, the linker will change the @code{CALLX@var{n}}
7330 instruction to a @code{CALL@var{n}} instruction, remove the @code{L32R}
7331 instruction, and remove the literal referenced by the @code{L32R}
7332 instruction if it is not used for anything else. Removing the
7333 @code{L32R} instruction always reduces code size but can potentially
7334 hurt performance by changing the alignment of subsequent branch targets.
7335 By default, the linker will always preserve alignments, either by
7336 switching some instructions between 24-bit encodings and the equivalent
7337 density instructions or by inserting a no-op in place of the @code{L32R}
7338 instruction that was removed. If code size is more important than
7339 performance, the @option{--size-opt} option can be used to prevent the
7340 linker from widening density instructions or inserting no-ops, except in
7341 a few cases where no-ops are required for correctness.
7343 The following Xtensa-specific command-line options can be used to
7346 @cindex Xtensa options
7349 When optimizing indirect calls to direct calls, optimize for code size
7350 more than performance. With this option, the linker will not insert
7351 no-ops or widen density instructions to preserve branch target
7352 alignment. There may still be some cases where no-ops are required to
7353 preserve the correctness of the code.
7361 @ifclear SingleFormat
7366 @cindex object file management
7367 @cindex object formats available
7369 The linker accesses object and archive files using the BFD libraries.
7370 These libraries allow the linker to use the same routines to operate on
7371 object files whatever the object file format. A different object file
7372 format can be supported simply by creating a new BFD back end and adding
7373 it to the library. To conserve runtime memory, however, the linker and
7374 associated tools are usually configured to support only a subset of the
7375 object file formats available. You can use @code{objdump -i}
7376 (@pxref{objdump,,objdump,binutils.info,The GNU Binary Utilities}) to
7377 list all the formats available for your configuration.
7379 @cindex BFD requirements
7380 @cindex requirements for BFD
7381 As with most implementations, BFD is a compromise between
7382 several conflicting requirements. The major factor influencing
7383 BFD design was efficiency: any time used converting between
7384 formats is time which would not have been spent had BFD not
7385 been involved. This is partly offset by abstraction payback; since
7386 BFD simplifies applications and back ends, more time and care
7387 may be spent optimizing algorithms for a greater speed.
7389 One minor artifact of the BFD solution which you should bear in
7390 mind is the potential for information loss. There are two places where
7391 useful information can be lost using the BFD mechanism: during
7392 conversion and during output. @xref{BFD information loss}.
7395 * BFD outline:: How it works: an outline of BFD
7399 @section How It Works: An Outline of BFD
7400 @cindex opening object files
7401 @include bfdsumm.texi
7404 @node Reporting Bugs
7405 @chapter Reporting Bugs
7406 @cindex bugs in @command{ld}
7407 @cindex reporting bugs in @command{ld}
7409 Your bug reports play an essential role in making @command{ld} reliable.
7411 Reporting a bug may help you by bringing a solution to your problem, or
7412 it may not. But in any case the principal function of a bug report is
7413 to help the entire community by making the next version of @command{ld}
7414 work better. Bug reports are your contribution to the maintenance of
7417 In order for a bug report to serve its purpose, you must include the
7418 information that enables us to fix the bug.
7421 * Bug Criteria:: Have you found a bug?
7422 * Bug Reporting:: How to report bugs
7426 @section Have You Found a Bug?
7427 @cindex bug criteria
7429 If you are not sure whether you have found a bug, here are some guidelines:
7432 @cindex fatal signal
7433 @cindex linker crash
7434 @cindex crash of linker
7436 If the linker gets a fatal signal, for any input whatever, that is a
7437 @command{ld} bug. Reliable linkers never crash.
7439 @cindex error on valid input
7441 If @command{ld} produces an error message for valid input, that is a bug.
7443 @cindex invalid input
7445 If @command{ld} does not produce an error message for invalid input, that
7446 may be a bug. In the general case, the linker can not verify that
7447 object files are correct.
7450 If you are an experienced user of linkers, your suggestions for
7451 improvement of @command{ld} are welcome in any case.
7455 @section How to Report Bugs
7457 @cindex @command{ld} bugs, reporting
7459 A number of companies and individuals offer support for @sc{gnu}
7460 products. If you obtained @command{ld} from a support organization, we
7461 recommend you contact that organization first.
7463 You can find contact information for many support companies and
7464 individuals in the file @file{etc/SERVICE} in the @sc{gnu} Emacs
7468 Otherwise, send bug reports for @command{ld} to
7472 The fundamental principle of reporting bugs usefully is this:
7473 @strong{report all the facts}. If you are not sure whether to state a
7474 fact or leave it out, state it!
7476 Often people omit facts because they think they know what causes the
7477 problem and assume that some details do not matter. Thus, you might
7478 assume that the name of a symbol you use in an example does not
7479 matter. Well, probably it does not, but one cannot be sure. Perhaps
7480 the bug is a stray memory reference which happens to fetch from the
7481 location where that name is stored in memory; perhaps, if the name
7482 were different, the contents of that location would fool the linker
7483 into doing the right thing despite the bug. Play it safe and give a
7484 specific, complete example. That is the easiest thing for you to do,
7485 and the most helpful.
7487 Keep in mind that the purpose of a bug report is to enable us to fix
7488 the bug if it is new to us. Therefore, always write your bug reports
7489 on the assumption that the bug has not been reported previously.
7491 Sometimes people give a few sketchy facts and ask, ``Does this ring a
7492 bell?'' This cannot help us fix a bug, so it is basically useless. We
7493 respond by asking for enough details to enable us to investigate.
7494 You might as well expedite matters by sending them to begin with.
7496 To enable us to fix the bug, you should include all these things:
7500 The version of @command{ld}. @command{ld} announces it if you start it with
7501 the @samp{--version} argument.
7503 Without this, we will not know whether there is any point in looking for
7504 the bug in the current version of @command{ld}.
7507 Any patches you may have applied to the @command{ld} source, including any
7508 patches made to the @code{BFD} library.
7511 The type of machine you are using, and the operating system name and
7515 What compiler (and its version) was used to compile @command{ld}---e.g.
7519 The command arguments you gave the linker to link your example and
7520 observe the bug. To guarantee you will not omit something important,
7521 list them all. A copy of the Makefile (or the output from make) is
7524 If we were to try to guess the arguments, we would probably guess wrong
7525 and then we might not encounter the bug.
7528 A complete input file, or set of input files, that will reproduce the
7529 bug. It is generally most helpful to send the actual object files
7530 provided that they are reasonably small. Say no more than 10K. For
7531 bigger files you can either make them available by FTP or HTTP or else
7532 state that you are willing to send the object file(s) to whomever
7533 requests them. (Note - your email will be going to a mailing list, so
7534 we do not want to clog it up with large attachments). But small
7535 attachments are best.
7537 If the source files were assembled using @code{gas} or compiled using
7538 @code{gcc}, then it may be OK to send the source files rather than the
7539 object files. In this case, be sure to say exactly what version of
7540 @code{gas} or @code{gcc} was used to produce the object files. Also say
7541 how @code{gas} or @code{gcc} were configured.
7544 A description of what behavior you observe that you believe is
7545 incorrect. For example, ``It gets a fatal signal.''
7547 Of course, if the bug is that @command{ld} gets a fatal signal, then we
7548 will certainly notice it. But if the bug is incorrect output, we might
7549 not notice unless it is glaringly wrong. You might as well not give us
7550 a chance to make a mistake.
7552 Even if the problem you experience is a fatal signal, you should still
7553 say so explicitly. Suppose something strange is going on, such as, your
7554 copy of @command{ld} is out of sync, or you have encountered a bug in the
7555 C library on your system. (This has happened!) Your copy might crash
7556 and ours would not. If you told us to expect a crash, then when ours
7557 fails to crash, we would know that the bug was not happening for us. If
7558 you had not told us to expect a crash, then we would not be able to draw
7559 any conclusion from our observations.
7562 If you wish to suggest changes to the @command{ld} source, send us context
7563 diffs, as generated by @code{diff} with the @samp{-u}, @samp{-c}, or
7564 @samp{-p} option. Always send diffs from the old file to the new file.
7565 If you even discuss something in the @command{ld} source, refer to it by
7566 context, not by line number.
7568 The line numbers in our development sources will not match those in your
7569 sources. Your line numbers would convey no useful information to us.
7572 Here are some things that are not necessary:
7576 A description of the envelope of the bug.
7578 Often people who encounter a bug spend a lot of time investigating
7579 which changes to the input file will make the bug go away and which
7580 changes will not affect it.
7582 This is often time consuming and not very useful, because the way we
7583 will find the bug is by running a single example under the debugger
7584 with breakpoints, not by pure deduction from a series of examples.
7585 We recommend that you save your time for something else.
7587 Of course, if you can find a simpler example to report @emph{instead}
7588 of the original one, that is a convenience for us. Errors in the
7589 output will be easier to spot, running under the debugger will take
7590 less time, and so on.
7592 However, simplification is not vital; if you do not want to do this,
7593 report the bug anyway and send us the entire test case you used.
7596 A patch for the bug.
7598 A patch for the bug does help us if it is a good one. But do not omit
7599 the necessary information, such as the test case, on the assumption that
7600 a patch is all we need. We might see problems with your patch and decide
7601 to fix the problem another way, or we might not understand it at all.
7603 Sometimes with a program as complicated as @command{ld} it is very hard to
7604 construct an example that will make the program follow a certain path
7605 through the code. If you do not send us the example, we will not be
7606 able to construct one, so we will not be able to verify that the bug is
7609 And if we cannot understand what bug you are trying to fix, or why your
7610 patch should be an improvement, we will not install it. A test case will
7611 help us to understand.
7614 A guess about what the bug is or what it depends on.
7616 Such guesses are usually wrong. Even we cannot guess right about such
7617 things without first using the debugger to find the facts.
7621 @appendix MRI Compatible Script Files
7622 @cindex MRI compatibility
7623 To aid users making the transition to @sc{gnu} @command{ld} from the MRI
7624 linker, @command{ld} can use MRI compatible linker scripts as an
7625 alternative to the more general-purpose linker scripting language
7626 described in @ref{Scripts}. MRI compatible linker scripts have a much
7627 simpler command set than the scripting language otherwise used with
7628 @command{ld}. @sc{gnu} @command{ld} supports the most commonly used MRI
7629 linker commands; these commands are described here.
7631 In general, MRI scripts aren't of much use with the @code{a.out} object
7632 file format, since it only has three sections and MRI scripts lack some
7633 features to make use of them.
7635 You can specify a file containing an MRI-compatible script using the
7636 @samp{-c} command-line option.
7638 Each command in an MRI-compatible script occupies its own line; each
7639 command line starts with the keyword that identifies the command (though
7640 blank lines are also allowed for punctuation). If a line of an
7641 MRI-compatible script begins with an unrecognized keyword, @command{ld}
7642 issues a warning message, but continues processing the script.
7644 Lines beginning with @samp{*} are comments.
7646 You can write these commands using all upper-case letters, or all
7647 lower case; for example, @samp{chip} is the same as @samp{CHIP}.
7648 The following list shows only the upper-case form of each command.
7651 @cindex @code{ABSOLUTE} (MRI)
7652 @item ABSOLUTE @var{secname}
7653 @itemx ABSOLUTE @var{secname}, @var{secname}, @dots{} @var{secname}
7654 Normally, @command{ld} includes in the output file all sections from all
7655 the input files. However, in an MRI-compatible script, you can use the
7656 @code{ABSOLUTE} command to restrict the sections that will be present in
7657 your output program. If the @code{ABSOLUTE} command is used at all in a
7658 script, then only the sections named explicitly in @code{ABSOLUTE}
7659 commands will appear in the linker output. You can still use other
7660 input sections (whatever you select on the command line, or using
7661 @code{LOAD}) to resolve addresses in the output file.
7663 @cindex @code{ALIAS} (MRI)
7664 @item ALIAS @var{out-secname}, @var{in-secname}
7665 Use this command to place the data from input section @var{in-secname}
7666 in a section called @var{out-secname} in the linker output file.
7668 @var{in-secname} may be an integer.
7670 @cindex @code{ALIGN} (MRI)
7671 @item ALIGN @var{secname} = @var{expression}
7672 Align the section called @var{secname} to @var{expression}. The
7673 @var{expression} should be a power of two.
7675 @cindex @code{BASE} (MRI)
7676 @item BASE @var{expression}
7677 Use the value of @var{expression} as the lowest address (other than
7678 absolute addresses) in the output file.
7680 @cindex @code{CHIP} (MRI)
7681 @item CHIP @var{expression}
7682 @itemx CHIP @var{expression}, @var{expression}
7683 This command does nothing; it is accepted only for compatibility.
7685 @cindex @code{END} (MRI)
7687 This command does nothing whatever; it's only accepted for compatibility.
7689 @cindex @code{FORMAT} (MRI)
7690 @item FORMAT @var{output-format}
7691 Similar to the @code{OUTPUT_FORMAT} command in the more general linker
7692 language, but restricted to one of these output formats:
7696 S-records, if @var{output-format} is @samp{S}
7699 IEEE, if @var{output-format} is @samp{IEEE}
7702 COFF (the @samp{coff-m68k} variant in BFD), if @var{output-format} is
7706 @cindex @code{LIST} (MRI)
7707 @item LIST @var{anything}@dots{}
7708 Print (to the standard output file) a link map, as produced by the
7709 @command{ld} command-line option @samp{-M}.
7711 The keyword @code{LIST} may be followed by anything on the
7712 same line, with no change in its effect.
7714 @cindex @code{LOAD} (MRI)
7715 @item LOAD @var{filename}
7716 @itemx LOAD @var{filename}, @var{filename}, @dots{} @var{filename}
7717 Include one or more object file @var{filename} in the link; this has the
7718 same effect as specifying @var{filename} directly on the @command{ld}
7721 @cindex @code{NAME} (MRI)
7722 @item NAME @var{output-name}
7723 @var{output-name} is the name for the program produced by @command{ld}; the
7724 MRI-compatible command @code{NAME} is equivalent to the command-line
7725 option @samp{-o} or the general script language command @code{OUTPUT}.
7727 @cindex @code{ORDER} (MRI)
7728 @item ORDER @var{secname}, @var{secname}, @dots{} @var{secname}
7729 @itemx ORDER @var{secname} @var{secname} @var{secname}
7730 Normally, @command{ld} orders the sections in its output file in the
7731 order in which they first appear in the input files. In an MRI-compatible
7732 script, you can override this ordering with the @code{ORDER} command. The
7733 sections you list with @code{ORDER} will appear first in your output
7734 file, in the order specified.
7736 @cindex @code{PUBLIC} (MRI)
7737 @item PUBLIC @var{name}=@var{expression}
7738 @itemx PUBLIC @var{name},@var{expression}
7739 @itemx PUBLIC @var{name} @var{expression}
7740 Supply a value (@var{expression}) for external symbol
7741 @var{name} used in the linker input files.
7743 @cindex @code{SECT} (MRI)
7744 @item SECT @var{secname}, @var{expression}
7745 @itemx SECT @var{secname}=@var{expression}
7746 @itemx SECT @var{secname} @var{expression}
7747 You can use any of these three forms of the @code{SECT} command to
7748 specify the start address (@var{expression}) for section @var{secname}.
7749 If you have more than one @code{SECT} statement for the same
7750 @var{secname}, only the @emph{first} sets the start address.
7753 @node GNU Free Documentation License
7754 @appendix GNU Free Documentation License
7758 @unnumbered LD Index
7763 % I think something like @colophon should be in texinfo. In the
7765 \long\def\colophon{\hbox to0pt{}\vfill
7766 \centerline{The body of this manual is set in}
7767 \centerline{\fontname\tenrm,}
7768 \centerline{with headings in {\bf\fontname\tenbf}}
7769 \centerline{and examples in {\tt\fontname\tentt}.}
7770 \centerline{{\it\fontname\tenit\/} and}
7771 \centerline{{\sl\fontname\tensl\/}}
7772 \centerline{are used for emphasis.}\vfill}
7774 % Blame: doc@cygnus.com, 28mar91.