3 @c Copyright (C) 1991-2017 Free Software Foundation, Inc.
6 @include configdoc.texi
7 @c (configdoc.texi is generated by the Makefile)
13 @macro gcctabopt{body}
19 @c Configure for the generation of man pages
45 @dircategory Software development
47 * Ld: (ld). The GNU linker.
52 This file documents the @sc{gnu} linker LD
53 @ifset VERSION_PACKAGE
54 @value{VERSION_PACKAGE}
56 version @value{VERSION}.
58 Copyright @copyright{} 1991-2017 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{} 1991-2017 Free Software Foundation, Inc.
97 Permission is granted to copy, distribute and/or modify this document
98 under the terms of the GNU Free Documentation License, Version 1.3
99 or any later version published by the Free Software Foundation;
100 with no Invariant Sections, with no Front-Cover Texts, and with no
101 Back-Cover Texts. A copy of the license is included in the
102 section entitled ``GNU Free Documentation License''.
108 @c FIXME: Talk about importance of *order* of args, cmds to linker!
113 This file documents the @sc{gnu} linker ld
114 @ifset VERSION_PACKAGE
115 @value{VERSION_PACKAGE}
117 version @value{VERSION}.
119 This document is distributed under the terms of the GNU Free
120 Documentation License version 1.3. A copy of the license is included
121 in the section entitled ``GNU Free Documentation License''.
124 * Overview:: Overview
125 * Invocation:: Invocation
126 * Scripts:: Linker Scripts
128 * Machine Dependent:: Machine Dependent Features
132 * H8/300:: ld and the H8/300
135 * Renesas:: ld and other Renesas micros
138 * i960:: ld and the Intel 960 family
141 * ARM:: ld and the ARM family
144 * M68HC11/68HC12:: ld and the Motorola 68HC11 and 68HC12 families
147 * HPPA ELF32:: ld and HPPA 32-bit ELF
150 * M68K:: ld and Motorola 68K family
153 * MIPS:: ld and MIPS family
156 * PowerPC ELF32:: ld and PowerPC 32-bit ELF Support
159 * PowerPC64 ELF64:: ld and PowerPC64 64-bit ELF Support
162 * SPU ELF:: ld and SPU ELF Support
165 * TI COFF:: ld and the TI COFF
168 * Win32:: ld and WIN32 (cygwin/mingw)
171 * Xtensa:: ld and Xtensa Processors
174 @ifclear SingleFormat
177 @c Following blank line required for remaining bug in makeinfo conds/menus
179 * Reporting Bugs:: Reporting Bugs
180 * MRI:: MRI Compatible Script Files
181 * GNU Free Documentation License:: GNU Free Documentation License
182 * LD Index:: LD Index
189 @cindex @sc{gnu} linker
190 @cindex what is this?
193 @c man begin SYNOPSIS
194 ld [@b{options}] @var{objfile} @dots{}
198 ar(1), nm(1), objcopy(1), objdump(1), readelf(1) and
199 the Info entries for @file{binutils} and
204 @c man begin DESCRIPTION
206 @command{ld} combines a number of object and archive files, relocates
207 their data and ties up symbol references. Usually the last step in
208 compiling a program is to run @command{ld}.
210 @command{ld} accepts Linker Command Language files written in
211 a superset of AT&T's Link Editor Command Language syntax,
212 to provide explicit and total control over the linking process.
216 This man page does not describe the command language; see the
217 @command{ld} entry in @code{info} for full details on the command
218 language and on other aspects of the GNU linker.
221 @ifclear SingleFormat
222 This version of @command{ld} uses the general purpose BFD libraries
223 to operate on object files. This allows @command{ld} to read, combine, and
224 write object files in many different formats---for example, COFF or
225 @code{a.out}. Different formats may be linked together to produce any
226 available kind of object file. @xref{BFD}, for more information.
229 Aside from its flexibility, the @sc{gnu} linker is more helpful than other
230 linkers in providing diagnostic information. Many linkers abandon
231 execution immediately upon encountering an error; whenever possible,
232 @command{ld} continues executing, allowing you to identify other errors
233 (or, in some cases, to get an output file in spite of the error).
240 @c man begin DESCRIPTION
242 The @sc{gnu} linker @command{ld} is meant to cover a broad range of situations,
243 and to be as compatible as possible with other linkers. As a result,
244 you have many choices to control its behavior.
250 * Options:: Command Line Options
251 * Environment:: Environment Variables
255 @section Command Line Options
263 The linker supports a plethora of command-line options, but in actual
264 practice few of them are used in any particular context.
265 @cindex standard Unix system
266 For instance, a frequent use of @command{ld} is to link standard Unix
267 object files on a standard, supported Unix system. On such a system, to
268 link a file @code{hello.o}:
271 ld -o @var{output} /lib/crt0.o hello.o -lc
274 This tells @command{ld} to produce a file called @var{output} as the
275 result of linking the file @code{/lib/crt0.o} with @code{hello.o} and
276 the library @code{libc.a}, which will come from the standard search
277 directories. (See the discussion of the @samp{-l} option below.)
279 Some of the command-line options to @command{ld} may be specified at any
280 point in the command line. However, options which refer to files, such
281 as @samp{-l} or @samp{-T}, cause the file to be read at the point at
282 which the option appears in the command line, relative to the object
283 files and other file options. Repeating non-file options with a
284 different argument will either have no further effect, or override prior
285 occurrences (those further to the left on the command line) of that
286 option. Options which may be meaningfully specified more than once are
287 noted in the descriptions below.
290 Non-option arguments are object files or archives which are to be linked
291 together. They may follow, precede, or be mixed in with command-line
292 options, except that an object file argument may not be placed between
293 an option and its argument.
295 Usually the linker is invoked with at least one object file, but you can
296 specify other forms of binary input files using @samp{-l}, @samp{-R},
297 and the script command language. If @emph{no} binary input files at all
298 are specified, the linker does not produce any output, and issues the
299 message @samp{No input files}.
301 If the linker cannot recognize the format of an object file, it will
302 assume that it is a linker script. A script specified in this way
303 augments the main linker script used for the link (either the default
304 linker script or the one specified by using @samp{-T}). This feature
305 permits the linker to link against a file which appears to be an object
306 or an archive, but actually merely defines some symbol values, or uses
307 @code{INPUT} or @code{GROUP} to load other objects. Specifying a
308 script in this way merely augments the main linker script, with the
309 extra commands placed after the main script; use the @samp{-T} option
310 to replace the default linker script entirely, but note the effect of
311 the @code{INSERT} command. @xref{Scripts}.
313 For options whose names are a single letter,
314 option arguments must either follow the option letter without intervening
315 whitespace, or be given as separate arguments immediately following the
316 option that requires them.
318 For options whose names are multiple letters, either one dash or two can
319 precede the option name; for example, @samp{-trace-symbol} and
320 @samp{--trace-symbol} are equivalent. Note---there is one exception to
321 this rule. Multiple letter options that start with a lower case 'o' can
322 only be preceded by two dashes. This is to reduce confusion with the
323 @samp{-o} option. So for example @samp{-omagic} sets the output file
324 name to @samp{magic} whereas @samp{--omagic} sets the NMAGIC flag on the
327 Arguments to multiple-letter options must either be separated from the
328 option name by an equals sign, or be given as separate arguments
329 immediately following the option that requires them. For example,
330 @samp{--trace-symbol foo} and @samp{--trace-symbol=foo} are equivalent.
331 Unique abbreviations of the names of multiple-letter options are
334 Note---if the linker is being invoked indirectly, via a compiler driver
335 (e.g. @samp{gcc}) then all the linker command line options should be
336 prefixed by @samp{-Wl,} (or whatever is appropriate for the particular
337 compiler driver) like this:
340 gcc -Wl,--start-group foo.o bar.o -Wl,--end-group
343 This is important, because otherwise the compiler driver program may
344 silently drop the linker options, resulting in a bad link. Confusion
345 may also arise when passing options that require values through a
346 driver, as the use of a space between option and argument acts as
347 a separator, and causes the driver to pass only the option to the linker
348 and the argument to the compiler. In this case, it is simplest to use
349 the joined forms of both single- and multiple-letter options, such as:
352 gcc foo.o bar.o -Wl,-eENTRY -Wl,-Map=a.map
355 Here is a table of the generic command line switches accepted by the GNU
359 @include at-file.texi
361 @kindex -a @var{keyword}
362 @item -a @var{keyword}
363 This option is supported for HP/UX compatibility. The @var{keyword}
364 argument must be one of the strings @samp{archive}, @samp{shared}, or
365 @samp{default}. @samp{-aarchive} is functionally equivalent to
366 @samp{-Bstatic}, and the other two keywords are functionally equivalent
367 to @samp{-Bdynamic}. This option may be used any number of times.
369 @kindex --audit @var{AUDITLIB}
370 @item --audit @var{AUDITLIB}
371 Adds @var{AUDITLIB} to the @code{DT_AUDIT} entry of the dynamic section.
372 @var{AUDITLIB} is not checked for existence, nor will it use the DT_SONAME
373 specified in the library. If specified multiple times @code{DT_AUDIT}
374 will contain a colon separated list of audit interfaces to use. If the linker
375 finds an object with an audit entry while searching for shared libraries,
376 it will add a corresponding @code{DT_DEPAUDIT} entry in the output file.
377 This option is only meaningful on ELF platforms supporting the rtld-audit
381 @cindex architectures
382 @kindex -A @var{arch}
383 @item -A @var{architecture}
384 @kindex --architecture=@var{arch}
385 @itemx --architecture=@var{architecture}
386 In the current release of @command{ld}, this option is useful only for the
387 Intel 960 family of architectures. In that @command{ld} configuration, the
388 @var{architecture} argument identifies the particular architecture in
389 the 960 family, enabling some safeguards and modifying the
390 archive-library search path. @xref{i960,,@command{ld} and the Intel 960
391 family}, for details.
393 Future releases of @command{ld} may support similar functionality for
394 other architecture families.
397 @ifclear SingleFormat
398 @cindex binary input format
399 @kindex -b @var{format}
400 @kindex --format=@var{format}
403 @item -b @var{input-format}
404 @itemx --format=@var{input-format}
405 @command{ld} may be configured to support more than one kind of object
406 file. If your @command{ld} is configured this way, you can use the
407 @samp{-b} option to specify the binary format for input object files
408 that follow this option on the command line. Even when @command{ld} is
409 configured to support alternative object formats, you don't usually need
410 to specify this, as @command{ld} should be configured to expect as a
411 default input format the most usual format on each machine.
412 @var{input-format} is a text string, the name of a particular format
413 supported by the BFD libraries. (You can list the available binary
414 formats with @samp{objdump -i}.)
417 You may want to use this option if you are linking files with an unusual
418 binary format. You can also use @samp{-b} to switch formats explicitly (when
419 linking object files of different formats), by including
420 @samp{-b @var{input-format}} before each group of object files in a
423 The default format is taken from the environment variable
428 You can also define the input format from a script, using the command
431 see @ref{Format Commands}.
435 @kindex -c @var{MRI-cmdfile}
436 @kindex --mri-script=@var{MRI-cmdfile}
437 @cindex compatibility, MRI
438 @item -c @var{MRI-commandfile}
439 @itemx --mri-script=@var{MRI-commandfile}
440 For compatibility with linkers produced by MRI, @command{ld} accepts script
441 files written in an alternate, restricted command language, described in
443 @ref{MRI,,MRI Compatible Script Files}.
446 the MRI Compatible Script Files section of GNU ld documentation.
448 Introduce MRI script files with
449 the option @samp{-c}; use the @samp{-T} option to run linker
450 scripts written in the general-purpose @command{ld} scripting language.
451 If @var{MRI-cmdfile} does not exist, @command{ld} looks for it in the directories
452 specified by any @samp{-L} options.
454 @cindex common allocation
461 These three options are equivalent; multiple forms are supported for
462 compatibility with other linkers. They assign space to common symbols
463 even if a relocatable output file is specified (with @samp{-r}). The
464 script command @code{FORCE_COMMON_ALLOCATION} has the same effect.
465 @xref{Miscellaneous Commands}.
467 @kindex --depaudit @var{AUDITLIB}
468 @kindex -P @var{AUDITLIB}
469 @item --depaudit @var{AUDITLIB}
470 @itemx -P @var{AUDITLIB}
471 Adds @var{AUDITLIB} to the @code{DT_DEPAUDIT} entry of the dynamic section.
472 @var{AUDITLIB} is not checked for existence, nor will it use the DT_SONAME
473 specified in the library. If specified multiple times @code{DT_DEPAUDIT}
474 will contain a colon separated list of audit interfaces to use. This
475 option is only meaningful on ELF platforms supporting the rtld-audit interface.
476 The -P option is provided for Solaris compatibility.
478 @cindex entry point, from command line
479 @kindex -e @var{entry}
480 @kindex --entry=@var{entry}
482 @itemx --entry=@var{entry}
483 Use @var{entry} as the explicit symbol for beginning execution of your
484 program, rather than the default entry point. If there is no symbol
485 named @var{entry}, the linker will try to parse @var{entry} as a number,
486 and use that as the entry address (the number will be interpreted in
487 base 10; you may use a leading @samp{0x} for base 16, or a leading
488 @samp{0} for base 8). @xref{Entry Point}, for a discussion of defaults
489 and other ways of specifying the entry point.
491 @kindex --exclude-libs
492 @item --exclude-libs @var{lib},@var{lib},...
493 Specifies a list of archive libraries from which symbols should not be automatically
494 exported. The library names may be delimited by commas or colons. Specifying
495 @code{--exclude-libs ALL} excludes symbols in all archive libraries from
496 automatic export. This option is available only for the i386 PE targeted
497 port of the linker and for ELF targeted ports. For i386 PE, symbols
498 explicitly listed in a .def file are still exported, regardless of this
499 option. For ELF targeted ports, symbols affected by this option will
500 be treated as hidden.
502 @kindex --exclude-modules-for-implib
503 @item --exclude-modules-for-implib @var{module},@var{module},...
504 Specifies a list of object files or archive members, from which symbols
505 should not be automatically exported, but which should be copied wholesale
506 into the import library being generated during the link. The module names
507 may be delimited by commas or colons, and must match exactly the filenames
508 used by @command{ld} to open the files; for archive members, this is simply
509 the member name, but for object files the name listed must include and
510 match precisely any path used to specify the input file on the linker's
511 command-line. This option is available only for the i386 PE targeted port
512 of the linker. Symbols explicitly listed in a .def file are still exported,
513 regardless of this option.
515 @cindex dynamic symbol table
517 @kindex --export-dynamic
518 @kindex --no-export-dynamic
520 @itemx --export-dynamic
521 @itemx --no-export-dynamic
522 When creating a dynamically linked executable, using the @option{-E}
523 option or the @option{--export-dynamic} option causes the linker to add
524 all symbols to the dynamic symbol table. The dynamic symbol table is the
525 set of symbols which are visible from dynamic objects at run time.
527 If you do not use either of these options (or use the
528 @option{--no-export-dynamic} option to restore the default behavior), the
529 dynamic symbol table will normally contain only those symbols which are
530 referenced by some dynamic object mentioned in the link.
532 If you use @code{dlopen} to load a dynamic object which needs to refer
533 back to the symbols defined by the program, rather than some other
534 dynamic object, then you will probably need to use this option when
535 linking the program itself.
537 You can also use the dynamic list to control what symbols should
538 be added to the dynamic symbol table if the output format supports it.
539 See the description of @samp{--dynamic-list}.
541 Note that this option is specific to ELF targeted ports. PE targets
542 support a similar function to export all symbols from a DLL or EXE; see
543 the description of @samp{--export-all-symbols} below.
545 @ifclear SingleFormat
546 @cindex big-endian objects
550 Link big-endian objects. This affects the default output format.
552 @cindex little-endian objects
555 Link little-endian objects. This affects the default output format.
558 @kindex -f @var{name}
559 @kindex --auxiliary=@var{name}
561 @itemx --auxiliary=@var{name}
562 When creating an ELF shared object, set the internal DT_AUXILIARY field
563 to the specified name. This tells the dynamic linker that the symbol
564 table of the shared object should be used as an auxiliary filter on the
565 symbol table of the shared object @var{name}.
567 If you later link a program against this filter object, then, when you
568 run the program, the dynamic linker will see the DT_AUXILIARY field. If
569 the dynamic linker resolves any symbols from the filter object, it will
570 first check whether there is a definition in the shared object
571 @var{name}. If there is one, it will be used instead of the definition
572 in the filter object. The shared object @var{name} need not exist.
573 Thus the shared object @var{name} may be used to provide an alternative
574 implementation of certain functions, perhaps for debugging or for
575 machine specific performance.
577 This option may be specified more than once. The DT_AUXILIARY entries
578 will be created in the order in which they appear on the command line.
580 @kindex -F @var{name}
581 @kindex --filter=@var{name}
583 @itemx --filter=@var{name}
584 When creating an ELF shared object, set the internal DT_FILTER field to
585 the specified name. This tells the dynamic linker that the symbol table
586 of the shared object which is being created should be used as a filter
587 on the symbol table of the shared object @var{name}.
589 If you later link a program against this filter object, then, when you
590 run the program, the dynamic linker will see the DT_FILTER field. The
591 dynamic linker will resolve symbols according to the symbol table of the
592 filter object as usual, but it will actually link to the definitions
593 found in the shared object @var{name}. Thus the filter object can be
594 used to select a subset of the symbols provided by the object
597 Some older linkers used the @option{-F} option throughout a compilation
598 toolchain for specifying object-file format for both input and output
600 @ifclear SingleFormat
601 The @sc{gnu} linker uses other mechanisms for this purpose: the
602 @option{-b}, @option{--format}, @option{--oformat} options, the
603 @code{TARGET} command in linker scripts, and the @code{GNUTARGET}
604 environment variable.
606 The @sc{gnu} linker will ignore the @option{-F} option when not
607 creating an ELF shared object.
609 @cindex finalization function
610 @kindex -fini=@var{name}
611 @item -fini=@var{name}
612 When creating an ELF executable or shared object, call NAME when the
613 executable or shared object is unloaded, by setting DT_FINI to the
614 address of the function. By default, the linker uses @code{_fini} as
615 the function to call.
619 Ignored. Provided for compatibility with other tools.
621 @kindex -G @var{value}
622 @kindex --gpsize=@var{value}
625 @itemx --gpsize=@var{value}
626 Set the maximum size of objects to be optimized using the GP register to
627 @var{size}. This is only meaningful for object file formats such as
628 MIPS ELF that support putting large and small objects into different
629 sections. This is ignored for other object file formats.
631 @cindex runtime library name
632 @kindex -h @var{name}
633 @kindex -soname=@var{name}
635 @itemx -soname=@var{name}
636 When creating an ELF shared object, set the internal DT_SONAME field to
637 the specified name. When an executable is linked with a shared object
638 which has a DT_SONAME field, then when the executable is run the dynamic
639 linker will attempt to load the shared object specified by the DT_SONAME
640 field rather than the using the file name given to the linker.
643 @cindex incremental link
645 Perform an incremental link (same as option @samp{-r}).
647 @cindex initialization function
648 @kindex -init=@var{name}
649 @item -init=@var{name}
650 When creating an ELF executable or shared object, call NAME when the
651 executable or shared object is loaded, by setting DT_INIT to the address
652 of the function. By default, the linker uses @code{_init} as the
655 @cindex archive files, from cmd line
656 @kindex -l @var{namespec}
657 @kindex --library=@var{namespec}
658 @item -l @var{namespec}
659 @itemx --library=@var{namespec}
660 Add the archive or object file specified by @var{namespec} to the
661 list of files to link. This option may be used any number of times.
662 If @var{namespec} is of the form @file{:@var{filename}}, @command{ld}
663 will search the library path for a file called @var{filename}, otherwise it
664 will search the library path for a file called @file{lib@var{namespec}.a}.
666 On systems which support shared libraries, @command{ld} may also search for
667 files other than @file{lib@var{namespec}.a}. Specifically, on ELF
668 and SunOS systems, @command{ld} will search a directory for a library
669 called @file{lib@var{namespec}.so} before searching for one called
670 @file{lib@var{namespec}.a}. (By convention, a @code{.so} extension
671 indicates a shared library.) Note that this behavior does not apply
672 to @file{:@var{filename}}, which always specifies a file called
675 The linker will search an archive only once, at the location where it is
676 specified on the command line. If the archive defines a symbol which
677 was undefined in some object which appeared before the archive on the
678 command line, the linker will include the appropriate file(s) from the
679 archive. However, an undefined symbol in an object appearing later on
680 the command line will not cause the linker to search the archive again.
682 See the @option{-(} option for a way to force the linker to search
683 archives multiple times.
685 You may list the same archive multiple times on the command line.
688 This type of archive searching is standard for Unix linkers. However,
689 if you are using @command{ld} on AIX, note that it is different from the
690 behaviour of the AIX linker.
693 @cindex search directory, from cmd line
695 @kindex --library-path=@var{dir}
696 @item -L @var{searchdir}
697 @itemx --library-path=@var{searchdir}
698 Add path @var{searchdir} to the list of paths that @command{ld} will search
699 for archive libraries and @command{ld} control scripts. You may use this
700 option any number of times. The directories are searched in the order
701 in which they are specified on the command line. Directories specified
702 on the command line are searched before the default directories. All
703 @option{-L} options apply to all @option{-l} options, regardless of the
704 order in which the options appear. @option{-L} options do not affect
705 how @command{ld} searches for a linker script unless @option{-T}
708 If @var{searchdir} begins with @code{=}, then the @code{=} will be replaced
709 by the @dfn{sysroot prefix}, controlled by the @samp{--sysroot} option, or
710 specified when the linker is configured.
713 The default set of paths searched (without being specified with
714 @samp{-L}) depends on which emulation mode @command{ld} is using, and in
715 some cases also on how it was configured. @xref{Environment}.
718 The paths can also be specified in a link script with the
719 @code{SEARCH_DIR} command. Directories specified this way are searched
720 at the point in which the linker script appears in the command line.
723 @kindex -m @var{emulation}
724 @item -m @var{emulation}
725 Emulate the @var{emulation} linker. You can list the available
726 emulations with the @samp{--verbose} or @samp{-V} options.
728 If the @samp{-m} option is not used, the emulation is taken from the
729 @code{LDEMULATION} environment variable, if that is defined.
731 Otherwise, the default emulation depends upon how the linker was
739 Print a link map to the standard output. A link map provides
740 information about the link, including the following:
744 Where object files are mapped into memory.
746 How common symbols are allocated.
748 All archive members included in the link, with a mention of the symbol
749 which caused the archive member to be brought in.
751 The values assigned to symbols.
753 Note - symbols whose values are computed by an expression which
754 involves a reference to a previous value of the same symbol may not
755 have correct result displayed in the link map. This is because the
756 linker discards intermediate results and only retains the final value
757 of an expression. Under such circumstances the linker will display
758 the final value enclosed by square brackets. Thus for example a
759 linker script containing:
767 will produce the following output in the link map if the @option{-M}
772 [0x0000000c] foo = (foo * 0x4)
773 [0x0000000c] foo = (foo + 0x8)
776 See @ref{Expressions} for more information about expressions in linker
781 @cindex read-only text
786 Turn off page alignment of sections, and disable linking against shared
787 libraries. If the output format supports Unix style magic numbers,
788 mark the output as @code{NMAGIC}.
792 @cindex read/write from cmd line
796 Set the text and data sections to be readable and writable. Also, do
797 not page-align the data segment, and disable linking against shared
798 libraries. If the output format supports Unix style magic numbers,
799 mark the output as @code{OMAGIC}. Note: Although a writable text section
800 is allowed for PE-COFF targets, it does not conform to the format
801 specification published by Microsoft.
806 This option negates most of the effects of the @option{-N} option. It
807 sets the text section to be read-only, and forces the data segment to
808 be page-aligned. Note - this option does not enable linking against
809 shared libraries. Use @option{-Bdynamic} for this.
811 @kindex -o @var{output}
812 @kindex --output=@var{output}
813 @cindex naming the output file
814 @item -o @var{output}
815 @itemx --output=@var{output}
816 Use @var{output} as the name for the program produced by @command{ld}; if this
817 option is not specified, the name @file{a.out} is used by default. The
818 script command @code{OUTPUT} can also specify the output file name.
820 @kindex -O @var{level}
821 @cindex generating optimized output
823 If @var{level} is a numeric values greater than zero @command{ld} optimizes
824 the output. This might take significantly longer and therefore probably
825 should only be enabled for the final binary. At the moment this
826 option only affects ELF shared library generation. Future releases of
827 the linker may make more use of this option. Also currently there is
828 no difference in the linker's behaviour for different non-zero values
829 of this option. Again this may change with future releases.
831 @kindex -plugin @var{name}
832 @item -plugin @var{name}
833 Involve a plugin in the linking process. The @var{name} parameter is
834 the absolute filename of the plugin. Usually this parameter is
835 automatically added by the complier, when using link time
836 optimization, but users can also add their own plugins if they so
839 Note that the location of the compiler originated plugins is different
840 from the place where the @command{ar}, @command{nm} and
841 @command{ranlib} programs search for their plugins. In order for
842 those commands to make use of a compiler based plugin it must first be
843 copied into the @file{$@{libdir@}/bfd-plugins} directory. All gcc
844 based linker plugins are backward compatible, so it is sufficient to
845 just copy in the newest one.
848 @cindex push state governing input file handling
850 The @option{--push-state} allows to preserve the current state of the
851 flags which govern the input file handling so that they can all be
852 restored with one corresponding @option{--pop-state} option.
854 The option which are covered are: @option{-Bdynamic}, @option{-Bstatic},
855 @option{-dn}, @option{-dy}, @option{-call_shared}, @option{-non_shared},
856 @option{-static}, @option{-N}, @option{-n}, @option{--whole-archive},
857 @option{--no-whole-archive}, @option{-r}, @option{-Ur},
858 @option{--copy-dt-needed-entries}, @option{--no-copy-dt-needed-entries},
859 @option{--as-needed}, @option{--no-as-needed}, and @option{-a}.
861 One target for this option are specifications for @file{pkg-config}. When
862 used with the @option{--libs} option all possibly needed libraries are
863 listed and then possibly linked with all the time. It is better to return
864 something as follows:
867 -Wl,--push-state,--as-needed -libone -libtwo -Wl,--pop-state
871 @cindex pop state governing input file handling
872 Undoes the effect of --push-state, restores the previous values of the
873 flags governing input file handling.
876 @kindex --emit-relocs
877 @cindex retain relocations in final executable
880 Leave relocation sections and contents in fully linked executables.
881 Post link analysis and optimization tools may need this information in
882 order to perform correct modifications of executables. This results
883 in larger executables.
885 This option is currently only supported on ELF platforms.
887 @kindex --force-dynamic
888 @cindex forcing the creation of dynamic sections
889 @item --force-dynamic
890 Force the output file to have dynamic sections. This option is specific
894 @cindex relocatable output
896 @kindex --relocatable
899 Generate relocatable output---i.e., generate an output file that can in
900 turn serve as input to @command{ld}. This is often called @dfn{partial
901 linking}. As a side effect, in environments that support standard Unix
902 magic numbers, this option also sets the output file's magic number to
904 @c ; see @option{-N}.
905 If this option is not specified, an absolute file is produced. When
906 linking C++ programs, this option @emph{will not} resolve references to
907 constructors; to do that, use @samp{-Ur}.
909 When an input file does not have the same format as the output file,
910 partial linking is only supported if that input file does not contain any
911 relocations. Different output formats can have further restrictions; for
912 example some @code{a.out}-based formats do not support partial linking
913 with input files in other formats at all.
915 This option does the same thing as @samp{-i}.
917 @kindex -R @var{file}
918 @kindex --just-symbols=@var{file}
919 @cindex symbol-only input
920 @item -R @var{filename}
921 @itemx --just-symbols=@var{filename}
922 Read symbol names and their addresses from @var{filename}, but do not
923 relocate it or include it in the output. This allows your output file
924 to refer symbolically to absolute locations of memory defined in other
925 programs. You may use this option more than once.
927 For compatibility with other ELF linkers, if the @option{-R} option is
928 followed by a directory name, rather than a file name, it is treated as
929 the @option{-rpath} option.
933 @cindex strip all symbols
936 Omit all symbol information from the output file.
939 @kindex --strip-debug
940 @cindex strip debugger symbols
943 Omit debugger symbol information (but not all symbols) from the output file.
947 @cindex input files, displaying
950 Print the names of the input files as @command{ld} processes them.
952 @kindex -T @var{script}
953 @kindex --script=@var{script}
955 @item -T @var{scriptfile}
956 @itemx --script=@var{scriptfile}
957 Use @var{scriptfile} as the linker script. This script replaces
958 @command{ld}'s default linker script (rather than adding to it), so
959 @var{commandfile} must specify everything necessary to describe the
960 output file. @xref{Scripts}. If @var{scriptfile} does not exist in
961 the current directory, @code{ld} looks for it in the directories
962 specified by any preceding @samp{-L} options. Multiple @samp{-T}
965 @kindex -dT @var{script}
966 @kindex --default-script=@var{script}
968 @item -dT @var{scriptfile}
969 @itemx --default-script=@var{scriptfile}
970 Use @var{scriptfile} as the default linker script. @xref{Scripts}.
972 This option is similar to the @option{--script} option except that
973 processing of the script is delayed until after the rest of the
974 command line has been processed. This allows options placed after the
975 @option{--default-script} option on the command line to affect the
976 behaviour of the linker script, which can be important when the linker
977 command line cannot be directly controlled by the user. (eg because
978 the command line is being constructed by another tool, such as
981 @kindex -u @var{symbol}
982 @kindex --undefined=@var{symbol}
983 @cindex undefined symbol
984 @item -u @var{symbol}
985 @itemx --undefined=@var{symbol}
986 Force @var{symbol} to be entered in the output file as an undefined
987 symbol. Doing this may, for example, trigger linking of additional
988 modules from standard libraries. @samp{-u} may be repeated with
989 different option arguments to enter additional undefined symbols. This
990 option is equivalent to the @code{EXTERN} linker script command.
992 If this option is being used to force additional modules to be pulled
993 into the link, and if it is an error for the symbol to remain
994 undefined, then the option @option{--require-defined} should be used
997 @kindex --require-defined=@var{symbol}
998 @cindex symbols, require defined
999 @cindex defined symbol
1000 @item --require-defined=@var{symbol}
1001 Require that @var{symbol} is defined in the output file. This option
1002 is the same as option @option{--undefined} except that if @var{symbol}
1003 is not defined in the output file then the linker will issue an error
1004 and exit. The same effect can be achieved in a linker script by using
1005 @code{EXTERN}, @code{ASSERT} and @code{DEFINED} together. This option
1006 can be used multiple times to require additional symbols.
1009 @cindex constructors
1011 For anything other than C++ programs, this option is equivalent to
1012 @samp{-r}: it generates relocatable output---i.e., an output file that can in
1013 turn serve as input to @command{ld}. When linking C++ programs, @samp{-Ur}
1014 @emph{does} resolve references to constructors, unlike @samp{-r}.
1015 It does not work to use @samp{-Ur} on files that were themselves linked
1016 with @samp{-Ur}; once the constructor table has been built, it cannot
1017 be added to. Use @samp{-Ur} only for the last partial link, and
1018 @samp{-r} for the others.
1020 @kindex --orphan-handling=@var{MODE}
1021 @cindex orphan sections
1022 @cindex sections, orphan
1023 @item --orphan-handling=@var{MODE}
1024 Control how orphan sections are handled. An orphan section is one not
1025 specifically mentioned in a linker script. @xref{Orphan Sections}.
1027 @var{MODE} can have any of the following values:
1031 Orphan sections are placed into a suitable output section following
1032 the strategy described in @ref{Orphan Sections}. The option
1033 @samp{--unique} also effects how sections are placed.
1036 All orphan sections are discarded, by placing them in the
1037 @samp{/DISCARD/} section (@pxref{Output Section Discarding}).
1040 The linker will place the orphan section as for @code{place} and also
1044 The linker will exit with an error if any orphan section is found.
1047 The default if @samp{--orphan-handling} is not given is @code{place}.
1049 @kindex --unique[=@var{SECTION}]
1050 @item --unique[=@var{SECTION}]
1051 Creates a separate output section for every input section matching
1052 @var{SECTION}, or if the optional wildcard @var{SECTION} argument is
1053 missing, for every orphan input section. An orphan section is one not
1054 specifically mentioned in a linker script. You may use this option
1055 multiple times on the command line; It prevents the normal merging of
1056 input sections with the same name, overriding output section assignments
1066 Display the version number for @command{ld}. The @option{-V} option also
1067 lists the supported emulations.
1070 @kindex --discard-all
1071 @cindex deleting local symbols
1073 @itemx --discard-all
1074 Delete all local symbols.
1077 @kindex --discard-locals
1078 @cindex local symbols, deleting
1080 @itemx --discard-locals
1081 Delete all temporary local symbols. (These symbols start with
1082 system-specific local label prefixes, typically @samp{.L} for ELF systems
1083 or @samp{L} for traditional a.out systems.)
1085 @kindex -y @var{symbol}
1086 @kindex --trace-symbol=@var{symbol}
1087 @cindex symbol tracing
1088 @item -y @var{symbol}
1089 @itemx --trace-symbol=@var{symbol}
1090 Print the name of each linked file in which @var{symbol} appears. This
1091 option may be given any number of times. On many systems it is necessary
1092 to prepend an underscore.
1094 This option is useful when you have an undefined symbol in your link but
1095 don't know where the reference is coming from.
1097 @kindex -Y @var{path}
1099 Add @var{path} to the default library search path. This option exists
1100 for Solaris compatibility.
1102 @kindex -z @var{keyword}
1103 @item -z @var{keyword}
1104 The recognized keywords are:
1108 Combines multiple reloc sections and sorts them to make dynamic symbol
1109 lookup caching possible.
1112 Generate common symbols with the STT_COMMON type druing a relocatable
1116 Disallows undefined symbols in object files. Undefined symbols in
1117 shared libraries are still allowed.
1120 Marks the object as requiring executable stack.
1123 This option is only meaningful when building a shared object. It makes
1124 the symbols defined by this shared object available for symbol resolution
1125 of subsequently loaded libraries.
1128 This option is only meaningful when building a shared object.
1129 It marks the object so that its runtime initialization will occur
1130 before the runtime initialization of any other objects brought into
1131 the process at the same time. Similarly the runtime finalization of
1132 the object will occur after the runtime finalization of any other
1136 Marks the object that its symbol table interposes before all symbols
1137 but the primary executable.
1140 When generating an executable or shared library, mark it to tell the
1141 dynamic linker to defer function call resolution to the point when
1142 the function is called (lazy binding), rather than at load time.
1143 Lazy binding is the default.
1146 Marks the object that its filters be processed immediately at
1150 Allows multiple definitions.
1153 Disables multiple reloc sections combining.
1156 Generate common symbols with the STT_OBJECT type druing a relocatable
1160 Disable linker generated .dynbss variables used in place of variables
1161 defined in shared libraries. May result in dynamic text relocations.
1164 Marks the object that the search for dependencies of this object will
1165 ignore any default library search paths.
1168 Marks the object shouldn't be unloaded at runtime.
1171 Marks the object not available to @code{dlopen}.
1174 Marks the object can not be dumped by @code{dldump}.
1177 Marks the object as not requiring executable stack.
1180 Treat DT_TEXTREL in shared object as error.
1183 Don't treat DT_TEXTREL in shared object as error.
1186 Don't treat DT_TEXTREL in shared object as error.
1189 Don't create an ELF @code{PT_GNU_RELRO} segment header in the object.
1192 When generating an executable or shared library, mark it to tell the
1193 dynamic linker to resolve all symbols when the program is started, or
1194 when the shared library is linked to using dlopen, instead of
1195 deferring function call resolution to the point when the function is
1199 Marks the object may contain $ORIGIN.
1202 Create an ELF @code{PT_GNU_RELRO} segment header in the object.
1204 @item max-page-size=@var{value}
1205 Set the emulation maximum page size to @var{value}.
1207 @item common-page-size=@var{value}
1208 Set the emulation common page size to @var{value}.
1210 @item stack-size=@var{value}
1211 Specify a stack size for in an ELF @code{PT_GNU_STACK} segment.
1212 Specifying zero will override any default non-zero sized
1213 @code{PT_GNU_STACK} segment creation.
1216 Always generate BND prefix in PLT entries. Supported for Linux/x86_64.
1218 @item noextern-protected-data
1219 Don't treat protected data symbol as external when building shared
1220 library. This option overrides linker backend default. It can be used
1221 to workaround incorrect relocations against protected data symbols
1222 generated by compiler. Updates on protected data symbols by another
1223 module aren't visible to the resulting shared library. Supported for
1226 @item nodynamic-undefined-weak
1227 Don't treat undefined weak symbols as dynamic when building executable.
1228 This option overrides linker backend default. It can be used to avoid
1229 dynamic relocations against undefined weak symbols in executable.
1230 Supported for i386 and x86-64.
1232 @item noreloc-overflow
1233 Disable relocation overflow check. This can be used to disable
1234 relocation overflow check if there will be no dynamic relocation
1235 overflow at run-time. Supported for x86_64.
1237 @item call-nop=prefix-addr
1238 @itemx call-nop=prefix-nop
1239 @itemx call-nop=suffix-nop
1240 @itemx call-nop=prefix-@var{byte}
1241 @itemx call-nop=suffix-@var{byte}
1242 Specify the 1-byte @code{NOP} padding when transforming indirect call
1243 to a locally defined function, foo, via its GOT slot.
1244 @option{call-nop=prefix-addr} generates @code{0x67 call foo}.
1245 @option{call-nop=prefix-nop} generates @code{0x90 call foo}.
1246 @option{call-nop=suffix-nop} generates @code{call foo 0x90}.
1247 @option{call-nop=prefix-@var{byte}} generates @code{@var{byte} call foo}.
1248 @option{call-nop=suffix-@var{byte}} generates @code{call foo @var{byte}}.
1249 Supported for i386 and x86_64.
1253 Other keywords are ignored for Solaris compatibility.
1256 @cindex groups of archives
1257 @item -( @var{archives} -)
1258 @itemx --start-group @var{archives} --end-group
1259 The @var{archives} should be a list of archive files. They may be
1260 either explicit file names, or @samp{-l} options.
1262 The specified archives are searched repeatedly until no new undefined
1263 references are created. Normally, an archive is searched only once in
1264 the order that it is specified on the command line. If a symbol in that
1265 archive is needed to resolve an undefined symbol referred to by an
1266 object in an archive that appears later on the command line, the linker
1267 would not be able to resolve that reference. By grouping the archives,
1268 they all be searched repeatedly until all possible references are
1271 Using this option has a significant performance cost. It is best to use
1272 it only when there are unavoidable circular references between two or
1275 @kindex --accept-unknown-input-arch
1276 @kindex --no-accept-unknown-input-arch
1277 @item --accept-unknown-input-arch
1278 @itemx --no-accept-unknown-input-arch
1279 Tells the linker to accept input files whose architecture cannot be
1280 recognised. The assumption is that the user knows what they are doing
1281 and deliberately wants to link in these unknown input files. This was
1282 the default behaviour of the linker, before release 2.14. The default
1283 behaviour from release 2.14 onwards is to reject such input files, and
1284 so the @samp{--accept-unknown-input-arch} option has been added to
1285 restore the old behaviour.
1288 @kindex --no-as-needed
1290 @itemx --no-as-needed
1291 This option affects ELF DT_NEEDED tags for dynamic libraries mentioned
1292 on the command line after the @option{--as-needed} option. Normally
1293 the linker will add a DT_NEEDED tag for each dynamic library mentioned
1294 on the command line, regardless of whether the library is actually
1295 needed or not. @option{--as-needed} causes a DT_NEEDED tag to only be
1296 emitted for a library that @emph{at that point in the link} satisfies a
1297 non-weak undefined symbol reference from a regular object file or, if
1298 the library is not found in the DT_NEEDED lists of other needed libraries, a
1299 non-weak undefined symbol reference from another needed dynamic library.
1300 Object files or libraries appearing on the command line @emph{after}
1301 the library in question do not affect whether the library is seen as
1302 needed. This is similar to the rules for extraction of object files
1303 from archives. @option{--no-as-needed} restores the default behaviour.
1305 @kindex --add-needed
1306 @kindex --no-add-needed
1308 @itemx --no-add-needed
1309 These two options have been deprecated because of the similarity of
1310 their names to the @option{--as-needed} and @option{--no-as-needed}
1311 options. They have been replaced by @option{--copy-dt-needed-entries}
1312 and @option{--no-copy-dt-needed-entries}.
1314 @kindex -assert @var{keyword}
1315 @item -assert @var{keyword}
1316 This option is ignored for SunOS compatibility.
1320 @kindex -call_shared
1324 Link against dynamic libraries. This is only meaningful on platforms
1325 for which shared libraries are supported. This option is normally the
1326 default on such platforms. The different variants of this option are
1327 for compatibility with various systems. You may use this option
1328 multiple times on the command line: it affects library searching for
1329 @option{-l} options which follow it.
1333 Set the @code{DF_1_GROUP} flag in the @code{DT_FLAGS_1} entry in the dynamic
1334 section. This causes the runtime linker to handle lookups in this
1335 object and its dependencies to be performed only inside the group.
1336 @option{--unresolved-symbols=report-all} is implied. This option is
1337 only meaningful on ELF platforms which support shared libraries.
1347 Do not link against shared libraries. This is only meaningful on
1348 platforms for which shared libraries are supported. The different
1349 variants of this option are for compatibility with various systems. You
1350 may use this option multiple times on the command line: it affects
1351 library searching for @option{-l} options which follow it. This
1352 option also implies @option{--unresolved-symbols=report-all}. This
1353 option can be used with @option{-shared}. Doing so means that a
1354 shared library is being created but that all of the library's external
1355 references must be resolved by pulling in entries from static
1360 When creating a shared library, bind references to global symbols to the
1361 definition within the shared library, if any. Normally, it is possible
1362 for a program linked against a shared library to override the definition
1363 within the shared library. This option can also be used with the
1364 @option{--export-dynamic} option, when creating a position independent
1365 executable, to bind references to global symbols to the definition within
1366 the executable. This option is only meaningful on ELF platforms which
1367 support shared libraries and position independent executables.
1369 @kindex -Bsymbolic-functions
1370 @item -Bsymbolic-functions
1371 When creating a shared library, bind references to global function
1372 symbols to the definition within the shared library, if any.
1373 This option can also be used with the @option{--export-dynamic} option,
1374 when creating a position independent executable, to bind references
1375 to global function symbols to the definition within the executable.
1376 This option is only meaningful on ELF platforms which support shared
1377 libraries and position independent executables.
1379 @kindex --dynamic-list=@var{dynamic-list-file}
1380 @item --dynamic-list=@var{dynamic-list-file}
1381 Specify the name of a dynamic list file to the linker. This is
1382 typically used when creating shared libraries to specify a list of
1383 global symbols whose references shouldn't be bound to the definition
1384 within the shared library, or creating dynamically linked executables
1385 to specify a list of symbols which should be added to the symbol table
1386 in the executable. This option is only meaningful on ELF platforms
1387 which support shared libraries.
1389 The format of the dynamic list is the same as the version node without
1390 scope and node name. See @ref{VERSION} for more information.
1392 @kindex --dynamic-list-data
1393 @item --dynamic-list-data
1394 Include all global data symbols to the dynamic list.
1396 @kindex --dynamic-list-cpp-new
1397 @item --dynamic-list-cpp-new
1398 Provide the builtin dynamic list for C++ operator new and delete. It
1399 is mainly useful for building shared libstdc++.
1401 @kindex --dynamic-list-cpp-typeinfo
1402 @item --dynamic-list-cpp-typeinfo
1403 Provide the builtin dynamic list for C++ runtime type identification.
1405 @kindex --check-sections
1406 @kindex --no-check-sections
1407 @item --check-sections
1408 @itemx --no-check-sections
1409 Asks the linker @emph{not} to check section addresses after they have
1410 been assigned to see if there are any overlaps. Normally the linker will
1411 perform this check, and if it finds any overlaps it will produce
1412 suitable error messages. The linker does know about, and does make
1413 allowances for sections in overlays. The default behaviour can be
1414 restored by using the command line switch @option{--check-sections}.
1415 Section overlap is not usually checked for relocatable links. You can
1416 force checking in that case by using the @option{--check-sections}
1419 @kindex --copy-dt-needed-entries
1420 @kindex --no-copy-dt-needed-entries
1421 @item --copy-dt-needed-entries
1422 @itemx --no-copy-dt-needed-entries
1423 This option affects the treatment of dynamic libraries referred to
1424 by DT_NEEDED tags @emph{inside} ELF dynamic libraries mentioned on the
1425 command line. Normally the linker won't add a DT_NEEDED tag to the
1426 output binary for each library mentioned in a DT_NEEDED tag in an
1427 input dynamic library. With @option{--copy-dt-needed-entries}
1428 specified on the command line however any dynamic libraries that
1429 follow it will have their DT_NEEDED entries added. The default
1430 behaviour can be restored with @option{--no-copy-dt-needed-entries}.
1432 This option also has an effect on the resolution of symbols in dynamic
1433 libraries. With @option{--copy-dt-needed-entries} dynamic libraries
1434 mentioned on the command line will be recursively searched, following
1435 their DT_NEEDED tags to other libraries, in order to resolve symbols
1436 required by the output binary. With the default setting however
1437 the searching of dynamic libraries that follow it will stop with the
1438 dynamic library itself. No DT_NEEDED links will be traversed to resolve
1441 @cindex cross reference table
1444 Output a cross reference table. If a linker map file is being
1445 generated, the cross reference table is printed to the map file.
1446 Otherwise, it is printed on the standard output.
1448 The format of the table is intentionally simple, so that it may be
1449 easily processed by a script if necessary. The symbols are printed out,
1450 sorted by name. For each symbol, a list of file names is given. If the
1451 symbol is defined, the first file listed is the location of the
1452 definition. If the symbol is defined as a common value then any files
1453 where this happens appear next. Finally any files that reference the
1456 @cindex common allocation
1457 @kindex --no-define-common
1458 @item --no-define-common
1459 This option inhibits the assignment of addresses to common symbols.
1460 The script command @code{INHIBIT_COMMON_ALLOCATION} has the same effect.
1461 @xref{Miscellaneous Commands}.
1463 The @samp{--no-define-common} option allows decoupling
1464 the decision to assign addresses to Common symbols from the choice
1465 of the output file type; otherwise a non-Relocatable output type
1466 forces assigning addresses to Common symbols.
1467 Using @samp{--no-define-common} allows Common symbols that are referenced
1468 from a shared library to be assigned addresses only in the main program.
1469 This eliminates the unused duplicate space in the shared library,
1470 and also prevents any possible confusion over resolving to the wrong
1471 duplicate when there are many dynamic modules with specialized search
1472 paths for runtime symbol resolution.
1474 @cindex symbols, from command line
1475 @kindex --defsym=@var{symbol}=@var{exp}
1476 @item --defsym=@var{symbol}=@var{expression}
1477 Create a global symbol in the output file, containing the absolute
1478 address given by @var{expression}. You may use this option as many
1479 times as necessary to define multiple symbols in the command line. A
1480 limited form of arithmetic is supported for the @var{expression} in this
1481 context: you may give a hexadecimal constant or the name of an existing
1482 symbol, or use @code{+} and @code{-} to add or subtract hexadecimal
1483 constants or symbols. If you need more elaborate expressions, consider
1484 using the linker command language from a script (@pxref{Assignments}).
1485 @emph{Note:} there should be no white space between @var{symbol}, the
1486 equals sign (``@key{=}''), and @var{expression}.
1488 @cindex demangling, from command line
1489 @kindex --demangle[=@var{style}]
1490 @kindex --no-demangle
1491 @item --demangle[=@var{style}]
1492 @itemx --no-demangle
1493 These options control whether to demangle symbol names in error messages
1494 and other output. When the linker is told to demangle, it tries to
1495 present symbol names in a readable fashion: it strips leading
1496 underscores if they are used by the object file format, and converts C++
1497 mangled symbol names into user readable names. Different compilers have
1498 different mangling styles. The optional demangling style argument can be used
1499 to choose an appropriate demangling style for your compiler. The linker will
1500 demangle by default unless the environment variable @samp{COLLECT_NO_DEMANGLE}
1501 is set. These options may be used to override the default.
1503 @cindex dynamic linker, from command line
1504 @kindex -I@var{file}
1505 @kindex --dynamic-linker=@var{file}
1507 @itemx --dynamic-linker=@var{file}
1508 Set the name of the dynamic linker. This is only meaningful when
1509 generating dynamically linked ELF executables. The default dynamic
1510 linker is normally correct; don't use this unless you know what you are
1513 @kindex --no-dynamic-linker
1514 @item --no-dynamic-linker
1515 When producing an executable file, omit the request for a dynamic
1516 linker to be used at load-time. This is only meaningful for ELF
1517 executables that contain dynamic relocations, and usually requires
1518 entry point code that is capable of processing these relocations.
1520 @kindex --fatal-warnings
1521 @kindex --no-fatal-warnings
1522 @item --fatal-warnings
1523 @itemx --no-fatal-warnings
1524 Treat all warnings as errors. The default behaviour can be restored
1525 with the option @option{--no-fatal-warnings}.
1527 @kindex --force-exe-suffix
1528 @item --force-exe-suffix
1529 Make sure that an output file has a .exe suffix.
1531 If a successfully built fully linked output file does not have a
1532 @code{.exe} or @code{.dll} suffix, this option forces the linker to copy
1533 the output file to one of the same name with a @code{.exe} suffix. This
1534 option is useful when using unmodified Unix makefiles on a Microsoft
1535 Windows host, since some versions of Windows won't run an image unless
1536 it ends in a @code{.exe} suffix.
1538 @kindex --gc-sections
1539 @kindex --no-gc-sections
1540 @cindex garbage collection
1542 @itemx --no-gc-sections
1543 Enable garbage collection of unused input sections. It is ignored on
1544 targets that do not support this option. The default behaviour (of not
1545 performing this garbage collection) can be restored by specifying
1546 @samp{--no-gc-sections} on the command line. Note that garbage
1547 collection for COFF and PE format targets is supported, but the
1548 implementation is currently considered to be experimental.
1550 @samp{--gc-sections} decides which input sections are used by
1551 examining symbols and relocations. The section containing the entry
1552 symbol and all sections containing symbols undefined on the
1553 command-line will be kept, as will sections containing symbols
1554 referenced by dynamic objects. Note that when building shared
1555 libraries, the linker must assume that any visible symbol is
1556 referenced. Once this initial set of sections has been determined,
1557 the linker recursively marks as used any section referenced by their
1558 relocations. See @samp{--entry} and @samp{--undefined}.
1560 This option can be set when doing a partial link (enabled with option
1561 @samp{-r}). In this case the root of symbols kept must be explicitly
1562 specified either by an @samp{--entry} or @samp{--undefined} option or by
1563 a @code{ENTRY} command in the linker script.
1565 @kindex --print-gc-sections
1566 @kindex --no-print-gc-sections
1567 @cindex garbage collection
1568 @item --print-gc-sections
1569 @itemx --no-print-gc-sections
1570 List all sections removed by garbage collection. The listing is
1571 printed on stderr. This option is only effective if garbage
1572 collection has been enabled via the @samp{--gc-sections}) option. The
1573 default behaviour (of not listing the sections that are removed) can
1574 be restored by specifying @samp{--no-print-gc-sections} on the command
1577 @kindex --gc-keep-exported
1578 @cindex garbage collection
1579 @item --gc-keep-exported
1580 When @samp{--gc-sections} is enabled, this option prevents garbage
1581 collection of unused input sections that contain global symbols having
1582 default or protected visibility. This option is intended to be used for
1583 executables where unreferenced sections would otherwise be garbage
1584 collected regardless of the external visibility of contained symbols.
1585 Note that this option has no effect when linking shared objects since
1586 it is already the default behaviour. This option is only supported for
1589 @kindex --print-output-format
1590 @cindex output format
1591 @item --print-output-format
1592 Print the name of the default output format (perhaps influenced by
1593 other command-line options). This is the string that would appear
1594 in an @code{OUTPUT_FORMAT} linker script command (@pxref{File Commands}).
1596 @kindex --print-memory-usage
1597 @cindex memory usage
1598 @item --print-memory-usage
1599 Print used size, total size and used size of memory regions created with
1600 the @ref{MEMORY} command. This is useful on embedded targets to have a
1601 quick view of amount of free memory. The format of the output has one
1602 headline and one line per region. It is both human readable and easily
1603 parsable by tools. Here is an example of an output:
1606 Memory region Used Size Region Size %age Used
1607 ROM: 256 KB 1 MB 25.00%
1608 RAM: 32 B 2 GB 0.00%
1615 Print a summary of the command-line options on the standard output and exit.
1617 @kindex --target-help
1619 Print a summary of all target specific options on the standard output and exit.
1621 @kindex -Map=@var{mapfile}
1622 @item -Map=@var{mapfile}
1623 Print a link map to the file @var{mapfile}. See the description of the
1624 @option{-M} option, above.
1626 @cindex memory usage
1627 @kindex --no-keep-memory
1628 @item --no-keep-memory
1629 @command{ld} normally optimizes for speed over memory usage by caching the
1630 symbol tables of input files in memory. This option tells @command{ld} to
1631 instead optimize for memory usage, by rereading the symbol tables as
1632 necessary. This may be required if @command{ld} runs out of memory space
1633 while linking a large executable.
1635 @kindex --no-undefined
1637 @item --no-undefined
1639 Report unresolved symbol references from regular object files. This
1640 is done even if the linker is creating a non-symbolic shared library.
1641 The switch @option{--[no-]allow-shlib-undefined} controls the
1642 behaviour for reporting unresolved references found in shared
1643 libraries being linked in.
1645 @kindex --allow-multiple-definition
1647 @item --allow-multiple-definition
1649 Normally when a symbol is defined multiple times, the linker will
1650 report a fatal error. These options allow multiple definitions and the
1651 first definition will be used.
1653 @kindex --allow-shlib-undefined
1654 @kindex --no-allow-shlib-undefined
1655 @item --allow-shlib-undefined
1656 @itemx --no-allow-shlib-undefined
1657 Allows or disallows undefined symbols in shared libraries.
1658 This switch is similar to @option{--no-undefined} except that it
1659 determines the behaviour when the undefined symbols are in a
1660 shared library rather than a regular object file. It does not affect
1661 how undefined symbols in regular object files are handled.
1663 The default behaviour is to report errors for any undefined symbols
1664 referenced in shared libraries if the linker is being used to create
1665 an executable, but to allow them if the linker is being used to create
1668 The reasons for allowing undefined symbol references in shared
1669 libraries specified at link time are that:
1673 A shared library specified at link time may not be the same as the one
1674 that is available at load time, so the symbol might actually be
1675 resolvable at load time.
1677 There are some operating systems, eg BeOS and HPPA, where undefined
1678 symbols in shared libraries are normal.
1680 The BeOS kernel for example patches shared libraries at load time to
1681 select whichever function is most appropriate for the current
1682 architecture. This is used, for example, to dynamically select an
1683 appropriate memset function.
1686 @kindex --no-undefined-version
1687 @item --no-undefined-version
1688 Normally when a symbol has an undefined version, the linker will ignore
1689 it. This option disallows symbols with undefined version and a fatal error
1690 will be issued instead.
1692 @kindex --default-symver
1693 @item --default-symver
1694 Create and use a default symbol version (the soname) for unversioned
1697 @kindex --default-imported-symver
1698 @item --default-imported-symver
1699 Create and use a default symbol version (the soname) for unversioned
1702 @kindex --no-warn-mismatch
1703 @item --no-warn-mismatch
1704 Normally @command{ld} will give an error if you try to link together input
1705 files that are mismatched for some reason, perhaps because they have
1706 been compiled for different processors or for different endiannesses.
1707 This option tells @command{ld} that it should silently permit such possible
1708 errors. This option should only be used with care, in cases when you
1709 have taken some special action that ensures that the linker errors are
1712 @kindex --no-warn-search-mismatch
1713 @item --no-warn-search-mismatch
1714 Normally @command{ld} will give a warning if it finds an incompatible
1715 library during a library search. This option silences the warning.
1717 @kindex --no-whole-archive
1718 @item --no-whole-archive
1719 Turn off the effect of the @option{--whole-archive} option for subsequent
1722 @cindex output file after errors
1723 @kindex --noinhibit-exec
1724 @item --noinhibit-exec
1725 Retain the executable output file whenever it is still usable.
1726 Normally, the linker will not produce an output file if it encounters
1727 errors during the link process; it exits without writing an output file
1728 when it issues any error whatsoever.
1732 Only search library directories explicitly specified on the
1733 command line. Library directories specified in linker scripts
1734 (including linker scripts specified on the command line) are ignored.
1736 @ifclear SingleFormat
1737 @kindex --oformat=@var{output-format}
1738 @item --oformat=@var{output-format}
1739 @command{ld} may be configured to support more than one kind of object
1740 file. If your @command{ld} is configured this way, you can use the
1741 @samp{--oformat} option to specify the binary format for the output
1742 object file. Even when @command{ld} is configured to support alternative
1743 object formats, you don't usually need to specify this, as @command{ld}
1744 should be configured to produce as a default output format the most
1745 usual format on each machine. @var{output-format} is a text string, the
1746 name of a particular format supported by the BFD libraries. (You can
1747 list the available binary formats with @samp{objdump -i}.) The script
1748 command @code{OUTPUT_FORMAT} can also specify the output format, but
1749 this option overrides it. @xref{BFD}.
1752 @kindex --out-implib
1753 @item --out-implib @var{file}
1754 Create an import library in @var{file} corresponding to the executable
1755 the linker is generating (eg. a DLL or ELF program). This import
1756 library (which should be called @code{*.dll.a} or @code{*.a} for DLLs)
1757 may be used to link clients against the generated executable; this
1758 behaviour makes it possible to skip a separate import library creation
1759 step (eg. @code{dlltool} for DLLs). This option is only available for
1760 the i386 PE and ELF targetted ports of the linker.
1763 @kindex --pic-executable
1765 @itemx --pic-executable
1766 @cindex position independent executables
1767 Create a position independent executable. This is currently only supported on
1768 ELF platforms. Position independent executables are similar to shared
1769 libraries in that they are relocated by the dynamic linker to the virtual
1770 address the OS chooses for them (which can vary between invocations). Like
1771 normal dynamically linked executables they can be executed and symbols
1772 defined in the executable cannot be overridden by shared libraries.
1776 This option is ignored for Linux compatibility.
1780 This option is ignored for SVR4 compatibility.
1783 @cindex synthesizing linker
1784 @cindex relaxing addressing modes
1788 An option with machine dependent effects.
1790 This option is only supported on a few targets.
1793 @xref{H8/300,,@command{ld} and the H8/300}.
1796 @xref{i960,, @command{ld} and the Intel 960 family}.
1799 @xref{Xtensa,, @command{ld} and Xtensa Processors}.
1802 @xref{M68HC11/68HC12,,@command{ld} and the 68HC11 and 68HC12}.
1805 @xref{Nios II,,@command{ld} and the Altera Nios II}.
1808 @xref{PowerPC ELF32,,@command{ld} and PowerPC 32-bit ELF Support}.
1811 On some platforms the @samp{--relax} option performs target specific,
1812 global optimizations that become possible when the linker resolves
1813 addressing in the program, such as relaxing address modes,
1814 synthesizing new instructions, selecting shorter version of current
1815 instructions, and combining constant values.
1817 On some platforms these link time global optimizations may make symbolic
1818 debugging of the resulting executable impossible.
1820 This is known to be the case for the Matsushita MN10200 and MN10300
1821 family of processors.
1825 On platforms where this is not supported, @samp{--relax} is accepted,
1829 On platforms where @samp{--relax} is accepted the option
1830 @samp{--no-relax} can be used to disable the feature.
1832 @cindex retaining specified symbols
1833 @cindex stripping all but some symbols
1834 @cindex symbols, retaining selectively
1835 @kindex --retain-symbols-file=@var{filename}
1836 @item --retain-symbols-file=@var{filename}
1837 Retain @emph{only} the symbols listed in the file @var{filename},
1838 discarding all others. @var{filename} is simply a flat file, with one
1839 symbol name per line. This option is especially useful in environments
1843 where a large global symbol table is accumulated gradually, to conserve
1846 @samp{--retain-symbols-file} does @emph{not} discard undefined symbols,
1847 or symbols needed for relocations.
1849 You may only specify @samp{--retain-symbols-file} once in the command
1850 line. It overrides @samp{-s} and @samp{-S}.
1853 @item -rpath=@var{dir}
1854 @cindex runtime library search path
1855 @kindex -rpath=@var{dir}
1856 Add a directory to the runtime library search path. This is used when
1857 linking an ELF executable with shared objects. All @option{-rpath}
1858 arguments are concatenated and passed to the runtime linker, which uses
1859 them to locate shared objects at runtime. The @option{-rpath} option is
1860 also used when locating shared objects which are needed by shared
1861 objects explicitly included in the link; see the description of the
1862 @option{-rpath-link} option. If @option{-rpath} is not used when linking an
1863 ELF executable, the contents of the environment variable
1864 @code{LD_RUN_PATH} will be used if it is defined.
1866 The @option{-rpath} option may also be used on SunOS. By default, on
1867 SunOS, the linker will form a runtime search path out of all the
1868 @option{-L} options it is given. If a @option{-rpath} option is used, the
1869 runtime search path will be formed exclusively using the @option{-rpath}
1870 options, ignoring the @option{-L} options. This can be useful when using
1871 gcc, which adds many @option{-L} options which may be on NFS mounted
1874 For compatibility with other ELF linkers, if the @option{-R} option is
1875 followed by a directory name, rather than a file name, it is treated as
1876 the @option{-rpath} option.
1880 @cindex link-time runtime library search path
1881 @kindex -rpath-link=@var{dir}
1882 @item -rpath-link=@var{dir}
1883 When using ELF or SunOS, one shared library may require another. This
1884 happens when an @code{ld -shared} link includes a shared library as one
1887 When the linker encounters such a dependency when doing a non-shared,
1888 non-relocatable link, it will automatically try to locate the required
1889 shared library and include it in the link, if it is not included
1890 explicitly. In such a case, the @option{-rpath-link} option
1891 specifies the first set of directories to search. The
1892 @option{-rpath-link} option may specify a sequence of directory names
1893 either by specifying a list of names separated by colons, or by
1894 appearing multiple times.
1896 The tokens @var{$ORIGIN} and @var{$LIB} can appear in these search
1897 directories. They will be replaced by the full path to the directory
1898 containing the program or shared object in the case of @var{$ORIGIN}
1899 and either @samp{lib} - for 32-bit binaries - or @samp{lib64} - for
1900 64-bit binaries - in the case of @var{$LIB}.
1902 The alternative form of these tokens - @var{$@{ORIGIN@}} and
1903 @var{$@{LIB@}} can also be used. The token @var{$PLATFORM} is not
1906 This option should be used with caution as it overrides the search path
1907 that may have been hard compiled into a shared library. In such a case it
1908 is possible to use unintentionally a different search path than the
1909 runtime linker would do.
1911 The linker uses the following search paths to locate required shared
1915 Any directories specified by @option{-rpath-link} options.
1917 Any directories specified by @option{-rpath} options. The difference
1918 between @option{-rpath} and @option{-rpath-link} is that directories
1919 specified by @option{-rpath} options are included in the executable and
1920 used at runtime, whereas the @option{-rpath-link} option is only effective
1921 at link time. Searching @option{-rpath} in this way is only supported
1922 by native linkers and cross linkers which have been configured with
1923 the @option{--with-sysroot} option.
1925 On an ELF system, for native linkers, if the @option{-rpath} and
1926 @option{-rpath-link} options were not used, search the contents of the
1927 environment variable @code{LD_RUN_PATH}.
1929 On SunOS, if the @option{-rpath} option was not used, search any
1930 directories specified using @option{-L} options.
1932 For a native linker, search the contents of the environment
1933 variable @code{LD_LIBRARY_PATH}.
1935 For a native ELF linker, the directories in @code{DT_RUNPATH} or
1936 @code{DT_RPATH} of a shared library are searched for shared
1937 libraries needed by it. The @code{DT_RPATH} entries are ignored if
1938 @code{DT_RUNPATH} entries exist.
1940 The default directories, normally @file{/lib} and @file{/usr/lib}.
1942 For a native linker on an ELF system, if the file @file{/etc/ld.so.conf}
1943 exists, the list of directories found in that file.
1946 If the required shared library is not found, the linker will issue a
1947 warning and continue with the link.
1954 @cindex shared libraries
1955 Create a shared library. This is currently only supported on ELF, XCOFF
1956 and SunOS platforms. On SunOS, the linker will automatically create a
1957 shared library if the @option{-e} option is not used and there are
1958 undefined symbols in the link.
1960 @kindex --sort-common
1962 @itemx --sort-common=ascending
1963 @itemx --sort-common=descending
1964 This option tells @command{ld} to sort the common symbols by alignment in
1965 ascending or descending order when it places them in the appropriate output
1966 sections. The symbol alignments considered are sixteen-byte or larger,
1967 eight-byte, four-byte, two-byte, and one-byte. This is to prevent gaps
1968 between symbols due to alignment constraints. If no sorting order is
1969 specified, then descending order is assumed.
1971 @kindex --sort-section=name
1972 @item --sort-section=name
1973 This option will apply @code{SORT_BY_NAME} to all wildcard section
1974 patterns in the linker script.
1976 @kindex --sort-section=alignment
1977 @item --sort-section=alignment
1978 This option will apply @code{SORT_BY_ALIGNMENT} to all wildcard section
1979 patterns in the linker script.
1981 @kindex --split-by-file
1982 @item --split-by-file[=@var{size}]
1983 Similar to @option{--split-by-reloc} but creates a new output section for
1984 each input file when @var{size} is reached. @var{size} defaults to a
1985 size of 1 if not given.
1987 @kindex --split-by-reloc
1988 @item --split-by-reloc[=@var{count}]
1989 Tries to creates extra sections in the output file so that no single
1990 output section in the file contains more than @var{count} relocations.
1991 This is useful when generating huge relocatable files for downloading into
1992 certain real time kernels with the COFF object file format; since COFF
1993 cannot represent more than 65535 relocations in a single section. Note
1994 that this will fail to work with object file formats which do not
1995 support arbitrary sections. The linker will not split up individual
1996 input sections for redistribution, so if a single input section contains
1997 more than @var{count} relocations one output section will contain that
1998 many relocations. @var{count} defaults to a value of 32768.
2002 Compute and display statistics about the operation of the linker, such
2003 as execution time and memory usage.
2005 @kindex --sysroot=@var{directory}
2006 @item --sysroot=@var{directory}
2007 Use @var{directory} as the location of the sysroot, overriding the
2008 configure-time default. This option is only supported by linkers
2009 that were configured using @option{--with-sysroot}.
2011 @kindex --traditional-format
2012 @cindex traditional format
2013 @item --traditional-format
2014 For some targets, the output of @command{ld} is different in some ways from
2015 the output of some existing linker. This switch requests @command{ld} to
2016 use the traditional format instead.
2019 For example, on SunOS, @command{ld} combines duplicate entries in the
2020 symbol string table. This can reduce the size of an output file with
2021 full debugging information by over 30 percent. Unfortunately, the SunOS
2022 @code{dbx} program can not read the resulting program (@code{gdb} has no
2023 trouble). The @samp{--traditional-format} switch tells @command{ld} to not
2024 combine duplicate entries.
2026 @kindex --section-start=@var{sectionname}=@var{org}
2027 @item --section-start=@var{sectionname}=@var{org}
2028 Locate a section in the output file at the absolute
2029 address given by @var{org}. You may use this option as many
2030 times as necessary to locate multiple sections in the command
2032 @var{org} must be a single hexadecimal integer;
2033 for compatibility with other linkers, you may omit the leading
2034 @samp{0x} usually associated with hexadecimal values. @emph{Note:} there
2035 should be no white space between @var{sectionname}, the equals
2036 sign (``@key{=}''), and @var{org}.
2038 @kindex -Tbss=@var{org}
2039 @kindex -Tdata=@var{org}
2040 @kindex -Ttext=@var{org}
2041 @cindex segment origins, cmd line
2042 @item -Tbss=@var{org}
2043 @itemx -Tdata=@var{org}
2044 @itemx -Ttext=@var{org}
2045 Same as @option{--section-start}, with @code{.bss}, @code{.data} or
2046 @code{.text} as the @var{sectionname}.
2048 @kindex -Ttext-segment=@var{org}
2049 @item -Ttext-segment=@var{org}
2050 @cindex text segment origin, cmd line
2051 When creating an ELF executable, it will set the address of the first
2052 byte of the text segment.
2054 @kindex -Trodata-segment=@var{org}
2055 @item -Trodata-segment=@var{org}
2056 @cindex rodata segment origin, cmd line
2057 When creating an ELF executable or shared object for a target where
2058 the read-only data is in its own segment separate from the executable
2059 text, it will set the address of the first byte of the read-only data segment.
2061 @kindex -Tldata-segment=@var{org}
2062 @item -Tldata-segment=@var{org}
2063 @cindex ldata segment origin, cmd line
2064 When creating an ELF executable or shared object for x86-64 medium memory
2065 model, it will set the address of the first byte of the ldata segment.
2067 @kindex --unresolved-symbols
2068 @item --unresolved-symbols=@var{method}
2069 Determine how to handle unresolved symbols. There are four possible
2070 values for @samp{method}:
2074 Do not report any unresolved symbols.
2077 Report all unresolved symbols. This is the default.
2079 @item ignore-in-object-files
2080 Report unresolved symbols that are contained in shared libraries, but
2081 ignore them if they come from regular object files.
2083 @item ignore-in-shared-libs
2084 Report unresolved symbols that come from regular object files, but
2085 ignore them if they come from shared libraries. This can be useful
2086 when creating a dynamic binary and it is known that all the shared
2087 libraries that it should be referencing are included on the linker's
2091 The behaviour for shared libraries on their own can also be controlled
2092 by the @option{--[no-]allow-shlib-undefined} option.
2094 Normally the linker will generate an error message for each reported
2095 unresolved symbol but the option @option{--warn-unresolved-symbols}
2096 can change this to a warning.
2098 @kindex --verbose[=@var{NUMBER}]
2099 @cindex verbose[=@var{NUMBER}]
2101 @itemx --verbose[=@var{NUMBER}]
2102 Display the version number for @command{ld} and list the linker emulations
2103 supported. Display which input files can and cannot be opened. Display
2104 the linker script being used by the linker. If the optional @var{NUMBER}
2105 argument > 1, plugin symbol status will also be displayed.
2107 @kindex --version-script=@var{version-scriptfile}
2108 @cindex version script, symbol versions
2109 @item --version-script=@var{version-scriptfile}
2110 Specify the name of a version script to the linker. This is typically
2111 used when creating shared libraries to specify additional information
2112 about the version hierarchy for the library being created. This option
2113 is only fully supported on ELF platforms which support shared libraries;
2114 see @ref{VERSION}. It is partially supported on PE platforms, which can
2115 use version scripts to filter symbol visibility in auto-export mode: any
2116 symbols marked @samp{local} in the version script will not be exported.
2119 @kindex --warn-common
2120 @cindex warnings, on combining symbols
2121 @cindex combining symbols, warnings on
2123 Warn when a common symbol is combined with another common symbol or with
2124 a symbol definition. Unix linkers allow this somewhat sloppy practice,
2125 but linkers on some other operating systems do not. This option allows
2126 you to find potential problems from combining global symbols.
2127 Unfortunately, some C libraries use this practice, so you may get some
2128 warnings about symbols in the libraries as well as in your programs.
2130 There are three kinds of global symbols, illustrated here by C examples:
2134 A definition, which goes in the initialized data section of the output
2138 An undefined reference, which does not allocate space.
2139 There must be either a definition or a common symbol for the
2143 A common symbol. If there are only (one or more) common symbols for a
2144 variable, it goes in the uninitialized data area of the output file.
2145 The linker merges multiple common symbols for the same variable into a
2146 single symbol. If they are of different sizes, it picks the largest
2147 size. The linker turns a common symbol into a declaration, if there is
2148 a definition of the same variable.
2151 The @samp{--warn-common} option can produce five kinds of warnings.
2152 Each warning consists of a pair of lines: the first describes the symbol
2153 just encountered, and the second describes the previous symbol
2154 encountered with the same name. One or both of the two symbols will be
2159 Turning a common symbol into a reference, because there is already a
2160 definition for the symbol.
2162 @var{file}(@var{section}): warning: common of `@var{symbol}'
2163 overridden by definition
2164 @var{file}(@var{section}): warning: defined here
2168 Turning a common symbol into a reference, because a later definition for
2169 the symbol is encountered. This is the same as the previous case,
2170 except that the symbols are encountered in a different order.
2172 @var{file}(@var{section}): warning: definition of `@var{symbol}'
2174 @var{file}(@var{section}): warning: common is here
2178 Merging a common symbol with a previous same-sized common symbol.
2180 @var{file}(@var{section}): warning: multiple common
2182 @var{file}(@var{section}): warning: previous common is here
2186 Merging a common symbol with a previous larger common symbol.
2188 @var{file}(@var{section}): warning: common of `@var{symbol}'
2189 overridden by larger common
2190 @var{file}(@var{section}): warning: larger common is here
2194 Merging a common symbol with a previous smaller common symbol. This is
2195 the same as the previous case, except that the symbols are
2196 encountered in a different order.
2198 @var{file}(@var{section}): warning: common of `@var{symbol}'
2199 overriding smaller common
2200 @var{file}(@var{section}): warning: smaller common is here
2204 @kindex --warn-constructors
2205 @item --warn-constructors
2206 Warn if any global constructors are used. This is only useful for a few
2207 object file formats. For formats like COFF or ELF, the linker can not
2208 detect the use of global constructors.
2210 @kindex --warn-multiple-gp
2211 @item --warn-multiple-gp
2212 Warn if multiple global pointer values are required in the output file.
2213 This is only meaningful for certain processors, such as the Alpha.
2214 Specifically, some processors put large-valued constants in a special
2215 section. A special register (the global pointer) points into the middle
2216 of this section, so that constants can be loaded efficiently via a
2217 base-register relative addressing mode. Since the offset in
2218 base-register relative mode is fixed and relatively small (e.g., 16
2219 bits), this limits the maximum size of the constant pool. Thus, in
2220 large programs, it is often necessary to use multiple global pointer
2221 values in order to be able to address all possible constants. This
2222 option causes a warning to be issued whenever this case occurs.
2225 @cindex warnings, on undefined symbols
2226 @cindex undefined symbols, warnings on
2228 Only warn once for each undefined symbol, rather than once per module
2231 @kindex --warn-section-align
2232 @cindex warnings, on section alignment
2233 @cindex section alignment, warnings on
2234 @item --warn-section-align
2235 Warn if the address of an output section is changed because of
2236 alignment. Typically, the alignment will be set by an input section.
2237 The address will only be changed if it not explicitly specified; that
2238 is, if the @code{SECTIONS} command does not specify a start address for
2239 the section (@pxref{SECTIONS}).
2241 @kindex --warn-shared-textrel
2242 @item --warn-shared-textrel
2243 Warn if the linker adds a DT_TEXTREL to a shared object.
2245 @kindex --warn-alternate-em
2246 @item --warn-alternate-em
2247 Warn if an object has alternate ELF machine code.
2249 @kindex --warn-unresolved-symbols
2250 @item --warn-unresolved-symbols
2251 If the linker is going to report an unresolved symbol (see the option
2252 @option{--unresolved-symbols}) it will normally generate an error.
2253 This option makes it generate a warning instead.
2255 @kindex --error-unresolved-symbols
2256 @item --error-unresolved-symbols
2257 This restores the linker's default behaviour of generating errors when
2258 it is reporting unresolved symbols.
2260 @kindex --whole-archive
2261 @cindex including an entire archive
2262 @item --whole-archive
2263 For each archive mentioned on the command line after the
2264 @option{--whole-archive} option, include every object file in the archive
2265 in the link, rather than searching the archive for the required object
2266 files. This is normally used to turn an archive file into a shared
2267 library, forcing every object to be included in the resulting shared
2268 library. This option may be used more than once.
2270 Two notes when using this option from gcc: First, gcc doesn't know
2271 about this option, so you have to use @option{-Wl,-whole-archive}.
2272 Second, don't forget to use @option{-Wl,-no-whole-archive} after your
2273 list of archives, because gcc will add its own list of archives to
2274 your link and you may not want this flag to affect those as well.
2276 @kindex --wrap=@var{symbol}
2277 @item --wrap=@var{symbol}
2278 Use a wrapper function for @var{symbol}. Any undefined reference to
2279 @var{symbol} will be resolved to @code{__wrap_@var{symbol}}. Any
2280 undefined reference to @code{__real_@var{symbol}} will be resolved to
2283 This can be used to provide a wrapper for a system function. The
2284 wrapper function should be called @code{__wrap_@var{symbol}}. If it
2285 wishes to call the system function, it should call
2286 @code{__real_@var{symbol}}.
2288 Here is a trivial example:
2292 __wrap_malloc (size_t c)
2294 printf ("malloc called with %zu\n", c);
2295 return __real_malloc (c);
2299 If you link other code with this file using @option{--wrap malloc}, then
2300 all calls to @code{malloc} will call the function @code{__wrap_malloc}
2301 instead. The call to @code{__real_malloc} in @code{__wrap_malloc} will
2302 call the real @code{malloc} function.
2304 You may wish to provide a @code{__real_malloc} function as well, so that
2305 links without the @option{--wrap} option will succeed. If you do this,
2306 you should not put the definition of @code{__real_malloc} in the same
2307 file as @code{__wrap_malloc}; if you do, the assembler may resolve the
2308 call before the linker has a chance to wrap it to @code{malloc}.
2310 @kindex --eh-frame-hdr
2311 @kindex --no-eh-frame-hdr
2312 @item --eh-frame-hdr
2313 @itemx --no-eh-frame-hdr
2314 Request (@option{--eh-frame-hdr}) or suppress
2315 (@option{--no-eh-frame-hdr}) the creation of @code{.eh_frame_hdr}
2316 section and ELF @code{PT_GNU_EH_FRAME} segment header.
2318 @kindex --ld-generated-unwind-info
2319 @item --no-ld-generated-unwind-info
2320 Request creation of @code{.eh_frame} unwind info for linker
2321 generated code sections like PLT. This option is on by default
2322 if linker generated unwind info is supported.
2324 @kindex --enable-new-dtags
2325 @kindex --disable-new-dtags
2326 @item --enable-new-dtags
2327 @itemx --disable-new-dtags
2328 This linker can create the new dynamic tags in ELF. But the older ELF
2329 systems may not understand them. If you specify
2330 @option{--enable-new-dtags}, the new dynamic tags will be created as needed
2331 and older dynamic tags will be omitted.
2332 If you specify @option{--disable-new-dtags}, no new dynamic tags will be
2333 created. By default, the new dynamic tags are not created. Note that
2334 those options are only available for ELF systems.
2336 @kindex --hash-size=@var{number}
2337 @item --hash-size=@var{number}
2338 Set the default size of the linker's hash tables to a prime number
2339 close to @var{number}. Increasing this value can reduce the length of
2340 time it takes the linker to perform its tasks, at the expense of
2341 increasing the linker's memory requirements. Similarly reducing this
2342 value can reduce the memory requirements at the expense of speed.
2344 @kindex --hash-style=@var{style}
2345 @item --hash-style=@var{style}
2346 Set the type of linker's hash table(s). @var{style} can be either
2347 @code{sysv} for classic ELF @code{.hash} section, @code{gnu} for
2348 new style GNU @code{.gnu.hash} section or @code{both} for both
2349 the classic ELF @code{.hash} and new style GNU @code{.gnu.hash}
2350 hash tables. The default is @code{sysv}.
2352 @kindex --compress-debug-sections=none
2353 @kindex --compress-debug-sections=zlib
2354 @kindex --compress-debug-sections=zlib-gnu
2355 @kindex --compress-debug-sections=zlib-gabi
2356 @item --compress-debug-sections=none
2357 @itemx --compress-debug-sections=zlib
2358 @itemx --compress-debug-sections=zlib-gnu
2359 @itemx --compress-debug-sections=zlib-gabi
2360 On ELF platforms, these options control how DWARF debug sections are
2361 compressed using zlib.
2363 @option{--compress-debug-sections=none} doesn't compress DWARF debug
2364 sections. @option{--compress-debug-sections=zlib-gnu} compresses
2365 DWARF debug sections and renames them to begin with @samp{.zdebug}
2366 instead of @samp{.debug}. @option{--compress-debug-sections=zlib-gabi}
2367 also compresses DWARF debug sections, but rather than renaming them it
2368 sets the SHF_COMPRESSED flag in the sections' headers.
2370 The @option{--compress-debug-sections=zlib} option is an alias for
2371 @option{--compress-debug-sections=zlib-gabi}.
2373 Note that this option overrides any compression in input debug
2374 sections, so if a binary is linked with @option{--compress-debug-sections=none}
2375 for example, then any compressed debug sections in input files will be
2376 uncompressed before they are copied into the output binary.
2378 The default compression behaviour varies depending upon the target
2379 involved and the configure options used to build the toolchain. The
2380 default can be determined by examining the output from the linker's
2381 @option{--help} option.
2383 @kindex --reduce-memory-overheads
2384 @item --reduce-memory-overheads
2385 This option reduces memory requirements at ld runtime, at the expense of
2386 linking speed. This was introduced to select the old O(n^2) algorithm
2387 for link map file generation, rather than the new O(n) algorithm which uses
2388 about 40% more memory for symbol storage.
2390 Another effect of the switch is to set the default hash table size to
2391 1021, which again saves memory at the cost of lengthening the linker's
2392 run time. This is not done however if the @option{--hash-size} switch
2395 The @option{--reduce-memory-overheads} switch may be also be used to
2396 enable other tradeoffs in future versions of the linker.
2399 @kindex --build-id=@var{style}
2401 @itemx --build-id=@var{style}
2402 Request the creation of a @code{.note.gnu.build-id} ELF note section
2403 or a @code{.buildid} COFF section. The contents of the note are
2404 unique bits identifying this linked file. @var{style} can be
2405 @code{uuid} to use 128 random bits, @code{sha1} to use a 160-bit
2406 @sc{SHA1} hash on the normative parts of the output contents,
2407 @code{md5} to use a 128-bit @sc{MD5} hash on the normative parts of
2408 the output contents, or @code{0x@var{hexstring}} to use a chosen bit
2409 string specified as an even number of hexadecimal digits (@code{-} and
2410 @code{:} characters between digit pairs are ignored). If @var{style}
2411 is omitted, @code{sha1} is used.
2413 The @code{md5} and @code{sha1} styles produces an identifier
2414 that is always the same in an identical output file, but will be
2415 unique among all nonidentical output files. It is not intended
2416 to be compared as a checksum for the file's contents. A linked
2417 file may be changed later by other tools, but the build ID bit
2418 string identifying the original linked file does not change.
2420 Passing @code{none} for @var{style} disables the setting from any
2421 @code{--build-id} options earlier on the command line.
2426 @subsection Options Specific to i386 PE Targets
2428 @c man begin OPTIONS
2430 The i386 PE linker supports the @option{-shared} option, which causes
2431 the output to be a dynamically linked library (DLL) instead of a
2432 normal executable. You should name the output @code{*.dll} when you
2433 use this option. In addition, the linker fully supports the standard
2434 @code{*.def} files, which may be specified on the linker command line
2435 like an object file (in fact, it should precede archives it exports
2436 symbols from, to ensure that they get linked in, just like a normal
2439 In addition to the options common to all targets, the i386 PE linker
2440 support additional command line options that are specific to the i386
2441 PE target. Options that take values may be separated from their
2442 values by either a space or an equals sign.
2446 @kindex --add-stdcall-alias
2447 @item --add-stdcall-alias
2448 If given, symbols with a stdcall suffix (@@@var{nn}) will be exported
2449 as-is and also with the suffix stripped.
2450 [This option is specific to the i386 PE targeted port of the linker]
2453 @item --base-file @var{file}
2454 Use @var{file} as the name of a file in which to save the base
2455 addresses of all the relocations needed for generating DLLs with
2457 [This is an i386 PE specific option]
2461 Create a DLL instead of a regular executable. You may also use
2462 @option{-shared} or specify a @code{LIBRARY} in a given @code{.def}
2464 [This option is specific to the i386 PE targeted port of the linker]
2466 @kindex --enable-long-section-names
2467 @kindex --disable-long-section-names
2468 @item --enable-long-section-names
2469 @itemx --disable-long-section-names
2470 The PE variants of the COFF object format add an extension that permits
2471 the use of section names longer than eight characters, the normal limit
2472 for COFF. By default, these names are only allowed in object files, as
2473 fully-linked executable images do not carry the COFF string table required
2474 to support the longer names. As a GNU extension, it is possible to
2475 allow their use in executable images as well, or to (probably pointlessly!)
2476 disallow it in object files, by using these two options. Executable images
2477 generated with these long section names are slightly non-standard, carrying
2478 as they do a string table, and may generate confusing output when examined
2479 with non-GNU PE-aware tools, such as file viewers and dumpers. However,
2480 GDB relies on the use of PE long section names to find Dwarf-2 debug
2481 information sections in an executable image at runtime, and so if neither
2482 option is specified on the command-line, @command{ld} will enable long
2483 section names, overriding the default and technically correct behaviour,
2484 when it finds the presence of debug information while linking an executable
2485 image and not stripping symbols.
2486 [This option is valid for all PE targeted ports of the linker]
2488 @kindex --enable-stdcall-fixup
2489 @kindex --disable-stdcall-fixup
2490 @item --enable-stdcall-fixup
2491 @itemx --disable-stdcall-fixup
2492 If the link finds a symbol that it cannot resolve, it will attempt to
2493 do ``fuzzy linking'' by looking for another defined symbol that differs
2494 only in the format of the symbol name (cdecl vs stdcall) and will
2495 resolve that symbol by linking to the match. For example, the
2496 undefined symbol @code{_foo} might be linked to the function
2497 @code{_foo@@12}, or the undefined symbol @code{_bar@@16} might be linked
2498 to the function @code{_bar}. When the linker does this, it prints a
2499 warning, since it normally should have failed to link, but sometimes
2500 import libraries generated from third-party dlls may need this feature
2501 to be usable. If you specify @option{--enable-stdcall-fixup}, this
2502 feature is fully enabled and warnings are not printed. If you specify
2503 @option{--disable-stdcall-fixup}, this feature is disabled and such
2504 mismatches are considered to be errors.
2505 [This option is specific to the i386 PE targeted port of the linker]
2507 @kindex --leading-underscore
2508 @kindex --no-leading-underscore
2509 @item --leading-underscore
2510 @itemx --no-leading-underscore
2511 For most targets default symbol-prefix is an underscore and is defined
2512 in target's description. By this option it is possible to
2513 disable/enable the default underscore symbol-prefix.
2515 @cindex DLLs, creating
2516 @kindex --export-all-symbols
2517 @item --export-all-symbols
2518 If given, all global symbols in the objects used to build a DLL will
2519 be exported by the DLL. Note that this is the default if there
2520 otherwise wouldn't be any exported symbols. When symbols are
2521 explicitly exported via DEF files or implicitly exported via function
2522 attributes, the default is to not export anything else unless this
2523 option is given. Note that the symbols @code{DllMain@@12},
2524 @code{DllEntryPoint@@0}, @code{DllMainCRTStartup@@12}, and
2525 @code{impure_ptr} will not be automatically
2526 exported. Also, symbols imported from other DLLs will not be
2527 re-exported, nor will symbols specifying the DLL's internal layout
2528 such as those beginning with @code{_head_} or ending with
2529 @code{_iname}. In addition, no symbols from @code{libgcc},
2530 @code{libstd++}, @code{libmingw32}, or @code{crtX.o} will be exported.
2531 Symbols whose names begin with @code{__rtti_} or @code{__builtin_} will
2532 not be exported, to help with C++ DLLs. Finally, there is an
2533 extensive list of cygwin-private symbols that are not exported
2534 (obviously, this applies on when building DLLs for cygwin targets).
2535 These cygwin-excludes are: @code{_cygwin_dll_entry@@12},
2536 @code{_cygwin_crt0_common@@8}, @code{_cygwin_noncygwin_dll_entry@@12},
2537 @code{_fmode}, @code{_impure_ptr}, @code{cygwin_attach_dll},
2538 @code{cygwin_premain0}, @code{cygwin_premain1}, @code{cygwin_premain2},
2539 @code{cygwin_premain3}, and @code{environ}.
2540 [This option is specific to the i386 PE targeted port of the linker]
2542 @kindex --exclude-symbols
2543 @item --exclude-symbols @var{symbol},@var{symbol},...
2544 Specifies a list of symbols which should not be automatically
2545 exported. The symbol names may be delimited by commas or colons.
2546 [This option is specific to the i386 PE targeted port of the linker]
2548 @kindex --exclude-all-symbols
2549 @item --exclude-all-symbols
2550 Specifies no symbols should be automatically exported.
2551 [This option is specific to the i386 PE targeted port of the linker]
2553 @kindex --file-alignment
2554 @item --file-alignment
2555 Specify the file alignment. Sections in the file will always begin at
2556 file offsets which are multiples of this number. This defaults to
2558 [This option is specific to the i386 PE targeted port of the linker]
2562 @item --heap @var{reserve}
2563 @itemx --heap @var{reserve},@var{commit}
2564 Specify the number of bytes of memory to reserve (and optionally commit)
2565 to be used as heap for this program. The default is 1MB reserved, 4K
2567 [This option is specific to the i386 PE targeted port of the linker]
2570 @kindex --image-base
2571 @item --image-base @var{value}
2572 Use @var{value} as the base address of your program or dll. This is
2573 the lowest memory location that will be used when your program or dll
2574 is loaded. To reduce the need to relocate and improve performance of
2575 your dlls, each should have a unique base address and not overlap any
2576 other dlls. The default is 0x400000 for executables, and 0x10000000
2578 [This option is specific to the i386 PE targeted port of the linker]
2582 If given, the stdcall suffixes (@@@var{nn}) will be stripped from
2583 symbols before they are exported.
2584 [This option is specific to the i386 PE targeted port of the linker]
2586 @kindex --large-address-aware
2587 @item --large-address-aware
2588 If given, the appropriate bit in the ``Characteristics'' field of the COFF
2589 header is set to indicate that this executable supports virtual addresses
2590 greater than 2 gigabytes. This should be used in conjunction with the /3GB
2591 or /USERVA=@var{value} megabytes switch in the ``[operating systems]''
2592 section of the BOOT.INI. Otherwise, this bit has no effect.
2593 [This option is specific to PE targeted ports of the linker]
2595 @kindex --disable-large-address-aware
2596 @item --disable-large-address-aware
2597 Reverts the effect of a previous @samp{--large-address-aware} option.
2598 This is useful if @samp{--large-address-aware} is always set by the compiler
2599 driver (e.g. Cygwin gcc) and the executable does not support virtual
2600 addresses greater than 2 gigabytes.
2601 [This option is specific to PE targeted ports of the linker]
2603 @kindex --major-image-version
2604 @item --major-image-version @var{value}
2605 Sets the major number of the ``image version''. Defaults to 1.
2606 [This option is specific to the i386 PE targeted port of the linker]
2608 @kindex --major-os-version
2609 @item --major-os-version @var{value}
2610 Sets the major number of the ``os version''. Defaults to 4.
2611 [This option is specific to the i386 PE targeted port of the linker]
2613 @kindex --major-subsystem-version
2614 @item --major-subsystem-version @var{value}
2615 Sets the major number of the ``subsystem version''. Defaults to 4.
2616 [This option is specific to the i386 PE targeted port of the linker]
2618 @kindex --minor-image-version
2619 @item --minor-image-version @var{value}
2620 Sets the minor number of the ``image version''. Defaults to 0.
2621 [This option is specific to the i386 PE targeted port of the linker]
2623 @kindex --minor-os-version
2624 @item --minor-os-version @var{value}
2625 Sets the minor number of the ``os version''. Defaults to 0.
2626 [This option is specific to the i386 PE targeted port of the linker]
2628 @kindex --minor-subsystem-version
2629 @item --minor-subsystem-version @var{value}
2630 Sets the minor number of the ``subsystem version''. Defaults to 0.
2631 [This option is specific to the i386 PE targeted port of the linker]
2633 @cindex DEF files, creating
2634 @cindex DLLs, creating
2635 @kindex --output-def
2636 @item --output-def @var{file}
2637 The linker will create the file @var{file} which will contain a DEF
2638 file corresponding to the DLL the linker is generating. This DEF file
2639 (which should be called @code{*.def}) may be used to create an import
2640 library with @code{dlltool} or may be used as a reference to
2641 automatically or implicitly exported symbols.
2642 [This option is specific to the i386 PE targeted port of the linker]
2644 @cindex DLLs, creating
2645 @kindex --enable-auto-image-base
2646 @item --enable-auto-image-base
2647 @itemx --enable-auto-image-base=@var{value}
2648 Automatically choose the image base for DLLs, optionally starting with base
2649 @var{value}, unless one is specified using the @code{--image-base} argument.
2650 By using a hash generated from the dllname to create unique image bases
2651 for each DLL, in-memory collisions and relocations which can delay program
2652 execution are avoided.
2653 [This option is specific to the i386 PE targeted port of the linker]
2655 @kindex --disable-auto-image-base
2656 @item --disable-auto-image-base
2657 Do not automatically generate a unique image base. If there is no
2658 user-specified image base (@code{--image-base}) then use the platform
2660 [This option is specific to the i386 PE targeted port of the linker]
2662 @cindex DLLs, linking to
2663 @kindex --dll-search-prefix
2664 @item --dll-search-prefix @var{string}
2665 When linking dynamically to a dll without an import library,
2666 search for @code{<string><basename>.dll} in preference to
2667 @code{lib<basename>.dll}. This behaviour allows easy distinction
2668 between DLLs built for the various "subplatforms": native, cygwin,
2669 uwin, pw, etc. For instance, cygwin DLLs typically use
2670 @code{--dll-search-prefix=cyg}.
2671 [This option is specific to the i386 PE targeted port of the linker]
2673 @kindex --enable-auto-import
2674 @item --enable-auto-import
2675 Do sophisticated linking of @code{_symbol} to @code{__imp__symbol} for
2676 DATA imports from DLLs, and create the necessary thunking symbols when
2677 building the import libraries with those DATA exports. Note: Use of the
2678 'auto-import' extension will cause the text section of the image file
2679 to be made writable. This does not conform to the PE-COFF format
2680 specification published by Microsoft.
2682 Note - use of the 'auto-import' extension will also cause read only
2683 data which would normally be placed into the .rdata section to be
2684 placed into the .data section instead. This is in order to work
2685 around a problem with consts that is described here:
2686 http://www.cygwin.com/ml/cygwin/2004-09/msg01101.html
2688 Using 'auto-import' generally will 'just work' -- but sometimes you may
2691 "variable '<var>' can't be auto-imported. Please read the
2692 documentation for ld's @code{--enable-auto-import} for details."
2694 This message occurs when some (sub)expression accesses an address
2695 ultimately given by the sum of two constants (Win32 import tables only
2696 allow one). Instances where this may occur include accesses to member
2697 fields of struct variables imported from a DLL, as well as using a
2698 constant index into an array variable imported from a DLL. Any
2699 multiword variable (arrays, structs, long long, etc) may trigger
2700 this error condition. However, regardless of the exact data type
2701 of the offending exported variable, ld will always detect it, issue
2702 the warning, and exit.
2704 There are several ways to address this difficulty, regardless of the
2705 data type of the exported variable:
2707 One way is to use --enable-runtime-pseudo-reloc switch. This leaves the task
2708 of adjusting references in your client code for runtime environment, so
2709 this method works only when runtime environment supports this feature.
2711 A second solution is to force one of the 'constants' to be a variable --
2712 that is, unknown and un-optimizable at compile time. For arrays,
2713 there are two possibilities: a) make the indexee (the array's address)
2714 a variable, or b) make the 'constant' index a variable. Thus:
2717 extern type extern_array[];
2719 @{ volatile type *t=extern_array; t[1] @}
2725 extern type extern_array[];
2727 @{ volatile int t=1; extern_array[t] @}
2730 For structs (and most other multiword data types) the only option
2731 is to make the struct itself (or the long long, or the ...) variable:
2734 extern struct s extern_struct;
2735 extern_struct.field -->
2736 @{ volatile struct s *t=&extern_struct; t->field @}
2742 extern long long extern_ll;
2744 @{ volatile long long * local_ll=&extern_ll; *local_ll @}
2747 A third method of dealing with this difficulty is to abandon
2748 'auto-import' for the offending symbol and mark it with
2749 @code{__declspec(dllimport)}. However, in practice that
2750 requires using compile-time #defines to indicate whether you are
2751 building a DLL, building client code that will link to the DLL, or
2752 merely building/linking to a static library. In making the choice
2753 between the various methods of resolving the 'direct address with
2754 constant offset' problem, you should consider typical real-world usage:
2762 void main(int argc, char **argv)@{
2763 printf("%d\n",arr[1]);
2773 void main(int argc, char **argv)@{
2774 /* This workaround is for win32 and cygwin; do not "optimize" */
2775 volatile int *parr = arr;
2776 printf("%d\n",parr[1]);
2783 /* Note: auto-export is assumed (no __declspec(dllexport)) */
2784 #if (defined(_WIN32) || defined(__CYGWIN__)) && \
2785 !(defined(FOO_BUILD_DLL) || defined(FOO_STATIC))
2786 #define FOO_IMPORT __declspec(dllimport)
2790 extern FOO_IMPORT int arr[];
2793 void main(int argc, char **argv)@{
2794 printf("%d\n",arr[1]);
2798 A fourth way to avoid this problem is to re-code your
2799 library to use a functional interface rather than a data interface
2800 for the offending variables (e.g. set_foo() and get_foo() accessor
2802 [This option is specific to the i386 PE targeted port of the linker]
2804 @kindex --disable-auto-import
2805 @item --disable-auto-import
2806 Do not attempt to do sophisticated linking of @code{_symbol} to
2807 @code{__imp__symbol} for DATA imports from DLLs.
2808 [This option is specific to the i386 PE targeted port of the linker]
2810 @kindex --enable-runtime-pseudo-reloc
2811 @item --enable-runtime-pseudo-reloc
2812 If your code contains expressions described in --enable-auto-import section,
2813 that is, DATA imports from DLL with non-zero offset, this switch will create
2814 a vector of 'runtime pseudo relocations' which can be used by runtime
2815 environment to adjust references to such data in your client code.
2816 [This option is specific to the i386 PE targeted port of the linker]
2818 @kindex --disable-runtime-pseudo-reloc
2819 @item --disable-runtime-pseudo-reloc
2820 Do not create pseudo relocations for non-zero offset DATA imports from
2822 [This option is specific to the i386 PE targeted port of the linker]
2824 @kindex --enable-extra-pe-debug
2825 @item --enable-extra-pe-debug
2826 Show additional debug info related to auto-import symbol thunking.
2827 [This option is specific to the i386 PE targeted port of the linker]
2829 @kindex --section-alignment
2830 @item --section-alignment
2831 Sets the section alignment. Sections in memory will always begin at
2832 addresses which are a multiple of this number. Defaults to 0x1000.
2833 [This option is specific to the i386 PE targeted port of the linker]
2837 @item --stack @var{reserve}
2838 @itemx --stack @var{reserve},@var{commit}
2839 Specify the number of bytes of memory to reserve (and optionally commit)
2840 to be used as stack for this program. The default is 2MB reserved, 4K
2842 [This option is specific to the i386 PE targeted port of the linker]
2845 @item --subsystem @var{which}
2846 @itemx --subsystem @var{which}:@var{major}
2847 @itemx --subsystem @var{which}:@var{major}.@var{minor}
2848 Specifies the subsystem under which your program will execute. The
2849 legal values for @var{which} are @code{native}, @code{windows},
2850 @code{console}, @code{posix}, and @code{xbox}. You may optionally set
2851 the subsystem version also. Numeric values are also accepted for
2853 [This option is specific to the i386 PE targeted port of the linker]
2855 The following options set flags in the @code{DllCharacteristics} field
2856 of the PE file header:
2857 [These options are specific to PE targeted ports of the linker]
2859 @kindex --high-entropy-va
2860 @item --high-entropy-va
2861 Image is compatible with 64-bit address space layout randomization
2864 @kindex --dynamicbase
2866 The image base address may be relocated using address space layout
2867 randomization (ASLR). This feature was introduced with MS Windows
2868 Vista for i386 PE targets.
2870 @kindex --forceinteg
2872 Code integrity checks are enforced.
2876 The image is compatible with the Data Execution Prevention.
2877 This feature was introduced with MS Windows XP SP2 for i386 PE targets.
2879 @kindex --no-isolation
2880 @item --no-isolation
2881 Although the image understands isolation, do not isolate the image.
2885 The image does not use SEH. No SE handler may be called from
2890 Do not bind this image.
2894 The driver uses the MS Windows Driver Model.
2898 The image is Terminal Server aware.
2900 @kindex --insert-timestamp
2901 @item --insert-timestamp
2902 @itemx --no-insert-timestamp
2903 Insert a real timestamp into the image. This is the default behaviour
2904 as it matches legacy code and it means that the image will work with
2905 other, proprietary tools. The problem with this default is that it
2906 will result in slightly different images being produced each time the
2907 same sources are linked. The option @option{--no-insert-timestamp}
2908 can be used to insert a zero value for the timestamp, this ensuring
2909 that binaries produced from identical sources will compare
2916 @subsection Options specific to C6X uClinux targets
2918 @c man begin OPTIONS
2920 The C6X uClinux target uses a binary format called DSBT to support shared
2921 libraries. Each shared library in the system needs to have a unique index;
2922 all executables use an index of 0.
2927 @item --dsbt-size @var{size}
2928 This option sets the number of entries in the DSBT of the current executable
2929 or shared library to @var{size}. The default is to create a table with 64
2932 @kindex --dsbt-index
2933 @item --dsbt-index @var{index}
2934 This option sets the DSBT index of the current executable or shared library
2935 to @var{index}. The default is 0, which is appropriate for generating
2936 executables. If a shared library is generated with a DSBT index of 0, the
2937 @code{R_C6000_DSBT_INDEX} relocs are copied into the output file.
2939 @kindex --no-merge-exidx-entries
2940 The @samp{--no-merge-exidx-entries} switch disables the merging of adjacent
2941 exidx entries in frame unwind info.
2949 @subsection Options specific to Motorola 68HC11 and 68HC12 targets
2951 @c man begin OPTIONS
2953 The 68HC11 and 68HC12 linkers support specific options to control the
2954 memory bank switching mapping and trampoline code generation.
2958 @kindex --no-trampoline
2959 @item --no-trampoline
2960 This option disables the generation of trampoline. By default a trampoline
2961 is generated for each far function which is called using a @code{jsr}
2962 instruction (this happens when a pointer to a far function is taken).
2964 @kindex --bank-window
2965 @item --bank-window @var{name}
2966 This option indicates to the linker the name of the memory region in
2967 the @samp{MEMORY} specification that describes the memory bank window.
2968 The definition of such region is then used by the linker to compute
2969 paging and addresses within the memory window.
2977 @subsection Options specific to Motorola 68K target
2979 @c man begin OPTIONS
2981 The following options are supported to control handling of GOT generation
2982 when linking for 68K targets.
2987 @item --got=@var{type}
2988 This option tells the linker which GOT generation scheme to use.
2989 @var{type} should be one of @samp{single}, @samp{negative},
2990 @samp{multigot} or @samp{target}. For more information refer to the
2991 Info entry for @file{ld}.
2999 @subsection Options specific to MIPS targets
3001 @c man begin OPTIONS
3003 The following options are supported to control microMIPS instruction
3004 generation and branch relocation checks for ISA mode transitions when
3005 linking for MIPS targets.
3013 These options control the choice of microMIPS instructions used in code
3014 generated by the linker, such as that in the PLT or lazy binding stubs,
3015 or in relaxation. If @samp{--insn32} is used, then the linker only uses
3016 32-bit instruction encodings. By default or if @samp{--no-insn32} is
3017 used, all instruction encodings are used, including 16-bit ones where
3020 @kindex --ignore-branch-isa
3021 @item --ignore-branch-isa
3022 @kindex --no-ignore-branch-isa
3023 @itemx --no-ignore-branch-isa
3024 These options control branch relocation checks for invalid ISA mode
3025 transitions. If @samp{--ignore-branch-isa} is used, then the linker
3026 accepts any branch relocations and any ISA mode transition required
3027 is lost in relocation calculation, except for some cases of @code{BAL}
3028 instructions which meet relaxation conditions and are converted to
3029 equivalent @code{JALX} instructions as the associated relocation is
3030 calculated. By default or if @samp{--no-ignore-branch-isa} is used
3031 a check is made causing the loss of an ISA mode transition to produce
3041 @section Environment Variables
3043 @c man begin ENVIRONMENT
3045 You can change the behaviour of @command{ld} with the environment variables
3046 @ifclear SingleFormat
3049 @code{LDEMULATION} and @code{COLLECT_NO_DEMANGLE}.
3051 @ifclear SingleFormat
3053 @cindex default input format
3054 @code{GNUTARGET} determines the input-file object format if you don't
3055 use @samp{-b} (or its synonym @samp{--format}). Its value should be one
3056 of the BFD names for an input format (@pxref{BFD}). If there is no
3057 @code{GNUTARGET} in the environment, @command{ld} uses the natural format
3058 of the target. If @code{GNUTARGET} is set to @code{default} then BFD
3059 attempts to discover the input format by examining binary input files;
3060 this method often succeeds, but there are potential ambiguities, since
3061 there is no method of ensuring that the magic number used to specify
3062 object-file formats is unique. However, the configuration procedure for
3063 BFD on each system places the conventional format for that system first
3064 in the search-list, so ambiguities are resolved in favor of convention.
3068 @cindex default emulation
3069 @cindex emulation, default
3070 @code{LDEMULATION} determines the default emulation if you don't use the
3071 @samp{-m} option. The emulation can affect various aspects of linker
3072 behaviour, particularly the default linker script. You can list the
3073 available emulations with the @samp{--verbose} or @samp{-V} options. If
3074 the @samp{-m} option is not used, and the @code{LDEMULATION} environment
3075 variable is not defined, the default emulation depends upon how the
3076 linker was configured.
3078 @kindex COLLECT_NO_DEMANGLE
3079 @cindex demangling, default
3080 Normally, the linker will default to demangling symbols. However, if
3081 @code{COLLECT_NO_DEMANGLE} is set in the environment, then it will
3082 default to not demangling symbols. This environment variable is used in
3083 a similar fashion by the @code{gcc} linker wrapper program. The default
3084 may be overridden by the @samp{--demangle} and @samp{--no-demangle}
3091 @chapter Linker Scripts
3094 @cindex linker scripts
3095 @cindex command files
3096 Every link is controlled by a @dfn{linker script}. This script is
3097 written in the linker command language.
3099 The main purpose of the linker script is to describe how the sections in
3100 the input files should be mapped into the output file, and to control
3101 the memory layout of the output file. Most linker scripts do nothing
3102 more than this. However, when necessary, the linker script can also
3103 direct the linker to perform many other operations, using the commands
3106 The linker always uses a linker script. If you do not supply one
3107 yourself, the linker will use a default script that is compiled into the
3108 linker executable. You can use the @samp{--verbose} command line option
3109 to display the default linker script. Certain command line options,
3110 such as @samp{-r} or @samp{-N}, will affect the default linker script.
3112 You may supply your own linker script by using the @samp{-T} command
3113 line option. When you do this, your linker script will replace the
3114 default linker script.
3116 You may also use linker scripts implicitly by naming them as input files
3117 to the linker, as though they were files to be linked. @xref{Implicit
3121 * Basic Script Concepts:: Basic Linker Script Concepts
3122 * Script Format:: Linker Script Format
3123 * Simple Example:: Simple Linker Script Example
3124 * Simple Commands:: Simple Linker Script Commands
3125 * Assignments:: Assigning Values to Symbols
3126 * SECTIONS:: SECTIONS Command
3127 * MEMORY:: MEMORY Command
3128 * PHDRS:: PHDRS Command
3129 * VERSION:: VERSION Command
3130 * Expressions:: Expressions in Linker Scripts
3131 * Implicit Linker Scripts:: Implicit Linker Scripts
3134 @node Basic Script Concepts
3135 @section Basic Linker Script Concepts
3136 @cindex linker script concepts
3137 We need to define some basic concepts and vocabulary in order to
3138 describe the linker script language.
3140 The linker combines input files into a single output file. The output
3141 file and each input file are in a special data format known as an
3142 @dfn{object file format}. Each file is called an @dfn{object file}.
3143 The output file is often called an @dfn{executable}, but for our
3144 purposes we will also call it an object file. Each object file has,
3145 among other things, a list of @dfn{sections}. We sometimes refer to a
3146 section in an input file as an @dfn{input section}; similarly, a section
3147 in the output file is an @dfn{output section}.
3149 Each section in an object file has a name and a size. Most sections
3150 also have an associated block of data, known as the @dfn{section
3151 contents}. A section may be marked as @dfn{loadable}, which means that
3152 the contents should be loaded into memory when the output file is run.
3153 A section with no contents may be @dfn{allocatable}, which means that an
3154 area in memory should be set aside, but nothing in particular should be
3155 loaded there (in some cases this memory must be zeroed out). A section
3156 which is neither loadable nor allocatable typically contains some sort
3157 of debugging information.
3159 Every loadable or allocatable output section has two addresses. The
3160 first is the @dfn{VMA}, or virtual memory address. This is the address
3161 the section will have when the output file is run. The second is the
3162 @dfn{LMA}, or load memory address. This is the address at which the
3163 section will be loaded. In most cases the two addresses will be the
3164 same. An example of when they might be different is when a data section
3165 is loaded into ROM, and then copied into RAM when the program starts up
3166 (this technique is often used to initialize global variables in a ROM
3167 based system). In this case the ROM address would be the LMA, and the
3168 RAM address would be the VMA.
3170 You can see the sections in an object file by using the @code{objdump}
3171 program with the @samp{-h} option.
3173 Every object file also has a list of @dfn{symbols}, known as the
3174 @dfn{symbol table}. A symbol may be defined or undefined. Each symbol
3175 has a name, and each defined symbol has an address, among other
3176 information. If you compile a C or C++ program into an object file, you
3177 will get a defined symbol for every defined function and global or
3178 static variable. Every undefined function or global variable which is
3179 referenced in the input file will become an undefined symbol.
3181 You can see the symbols in an object file by using the @code{nm}
3182 program, or by using the @code{objdump} program with the @samp{-t}
3186 @section Linker Script Format
3187 @cindex linker script format
3188 Linker scripts are text files.
3190 You write a linker script as a series of commands. Each command is
3191 either a keyword, possibly followed by arguments, or an assignment to a
3192 symbol. You may separate commands using semicolons. Whitespace is
3195 Strings such as file or format names can normally be entered directly.
3196 If the file name contains a character such as a comma which would
3197 otherwise serve to separate file names, you may put the file name in
3198 double quotes. There is no way to use a double quote character in a
3201 You may include comments in linker scripts just as in C, delimited by
3202 @samp{/*} and @samp{*/}. As in C, comments are syntactically equivalent
3205 @node Simple Example
3206 @section Simple Linker Script Example
3207 @cindex linker script example
3208 @cindex example of linker script
3209 Many linker scripts are fairly simple.
3211 The simplest possible linker script has just one command:
3212 @samp{SECTIONS}. You use the @samp{SECTIONS} command to describe the
3213 memory layout of the output file.
3215 The @samp{SECTIONS} command is a powerful command. Here we will
3216 describe a simple use of it. Let's assume your program consists only of
3217 code, initialized data, and uninitialized data. These will be in the
3218 @samp{.text}, @samp{.data}, and @samp{.bss} sections, respectively.
3219 Let's assume further that these are the only sections which appear in
3222 For this example, let's say that the code should be loaded at address
3223 0x10000, and that the data should start at address 0x8000000. Here is a
3224 linker script which will do that:
3229 .text : @{ *(.text) @}
3231 .data : @{ *(.data) @}
3232 .bss : @{ *(.bss) @}
3236 You write the @samp{SECTIONS} command as the keyword @samp{SECTIONS},
3237 followed by a series of symbol assignments and output section
3238 descriptions enclosed in curly braces.
3240 The first line inside the @samp{SECTIONS} command of the above example
3241 sets the value of the special symbol @samp{.}, which is the location
3242 counter. If you do not specify the address of an output section in some
3243 other way (other ways are described later), the address is set from the
3244 current value of the location counter. The location counter is then
3245 incremented by the size of the output section. At the start of the
3246 @samp{SECTIONS} command, the location counter has the value @samp{0}.
3248 The second line defines an output section, @samp{.text}. The colon is
3249 required syntax which may be ignored for now. Within the curly braces
3250 after the output section name, you list the names of the input sections
3251 which should be placed into this output section. The @samp{*} is a
3252 wildcard which matches any file name. The expression @samp{*(.text)}
3253 means all @samp{.text} input sections in all input files.
3255 Since the location counter is @samp{0x10000} when the output section
3256 @samp{.text} is defined, the linker will set the address of the
3257 @samp{.text} section in the output file to be @samp{0x10000}.
3259 The remaining lines define the @samp{.data} and @samp{.bss} sections in
3260 the output file. The linker will place the @samp{.data} output section
3261 at address @samp{0x8000000}. After the linker places the @samp{.data}
3262 output section, the value of the location counter will be
3263 @samp{0x8000000} plus the size of the @samp{.data} output section. The
3264 effect is that the linker will place the @samp{.bss} output section
3265 immediately after the @samp{.data} output section in memory.
3267 The linker will ensure that each output section has the required
3268 alignment, by increasing the location counter if necessary. In this
3269 example, the specified addresses for the @samp{.text} and @samp{.data}
3270 sections will probably satisfy any alignment constraints, but the linker
3271 may have to create a small gap between the @samp{.data} and @samp{.bss}
3274 That's it! That's a simple and complete linker script.
3276 @node Simple Commands
3277 @section Simple Linker Script Commands
3278 @cindex linker script simple commands
3279 In this section we describe the simple linker script commands.
3282 * Entry Point:: Setting the entry point
3283 * File Commands:: Commands dealing with files
3284 @ifclear SingleFormat
3285 * Format Commands:: Commands dealing with object file formats
3288 * REGION_ALIAS:: Assign alias names to memory regions
3289 * Miscellaneous Commands:: Other linker script commands
3293 @subsection Setting the Entry Point
3294 @kindex ENTRY(@var{symbol})
3295 @cindex start of execution
3296 @cindex first instruction
3298 The first instruction to execute in a program is called the @dfn{entry
3299 point}. You can use the @code{ENTRY} linker script command to set the
3300 entry point. The argument is a symbol name:
3305 There are several ways to set the entry point. The linker will set the
3306 entry point by trying each of the following methods in order, and
3307 stopping when one of them succeeds:
3310 the @samp{-e} @var{entry} command-line option;
3312 the @code{ENTRY(@var{symbol})} command in a linker script;
3314 the value of a target specific symbol, if it is defined; For many
3315 targets this is @code{start}, but PE and BeOS based systems for example
3316 check a list of possible entry symbols, matching the first one found.
3318 the address of the first byte of the @samp{.text} section, if present;
3320 The address @code{0}.
3324 @subsection Commands Dealing with Files
3325 @cindex linker script file commands
3326 Several linker script commands deal with files.
3329 @item INCLUDE @var{filename}
3330 @kindex INCLUDE @var{filename}
3331 @cindex including a linker script
3332 Include the linker script @var{filename} at this point. The file will
3333 be searched for in the current directory, and in any directory specified
3334 with the @option{-L} option. You can nest calls to @code{INCLUDE} up to
3337 You can place @code{INCLUDE} directives at the top level, in @code{MEMORY} or
3338 @code{SECTIONS} commands, or in output section descriptions.
3340 @item INPUT(@var{file}, @var{file}, @dots{})
3341 @itemx INPUT(@var{file} @var{file} @dots{})
3342 @kindex INPUT(@var{files})
3343 @cindex input files in linker scripts
3344 @cindex input object files in linker scripts
3345 @cindex linker script input object files
3346 The @code{INPUT} command directs the linker to include the named files
3347 in the link, as though they were named on the command line.
3349 For example, if you always want to include @file{subr.o} any time you do
3350 a link, but you can't be bothered to put it on every link command line,
3351 then you can put @samp{INPUT (subr.o)} in your linker script.
3353 In fact, if you like, you can list all of your input files in the linker
3354 script, and then invoke the linker with nothing but a @samp{-T} option.
3356 In case a @dfn{sysroot prefix} is configured, and the filename starts
3357 with the @samp{/} character, and the script being processed was
3358 located inside the @dfn{sysroot prefix}, the filename will be looked
3359 for in the @dfn{sysroot prefix}. Otherwise, the linker will try to
3360 open the file in the current directory. If it is not found, the
3361 linker will search through the archive library search path.
3362 The @dfn{sysroot prefix} can also be forced by specifying @code{=}
3363 as the first character in the filename path. See also the
3364 description of @samp{-L} in @ref{Options,,Command Line Options}.
3366 If you use @samp{INPUT (-l@var{file})}, @command{ld} will transform the
3367 name to @code{lib@var{file}.a}, as with the command line argument
3370 When you use the @code{INPUT} command in an implicit linker script, the
3371 files will be included in the link at the point at which the linker
3372 script file is included. This can affect archive searching.
3374 @item GROUP(@var{file}, @var{file}, @dots{})
3375 @itemx GROUP(@var{file} @var{file} @dots{})
3376 @kindex GROUP(@var{files})
3377 @cindex grouping input files
3378 The @code{GROUP} command is like @code{INPUT}, except that the named
3379 files should all be archives, and they are searched repeatedly until no
3380 new undefined references are created. See the description of @samp{-(}
3381 in @ref{Options,,Command Line Options}.
3383 @item AS_NEEDED(@var{file}, @var{file}, @dots{})
3384 @itemx AS_NEEDED(@var{file} @var{file} @dots{})
3385 @kindex AS_NEEDED(@var{files})
3386 This construct can appear only inside of the @code{INPUT} or @code{GROUP}
3387 commands, among other filenames. The files listed will be handled
3388 as if they appear directly in the @code{INPUT} or @code{GROUP} commands,
3389 with the exception of ELF shared libraries, that will be added only
3390 when they are actually needed. This construct essentially enables
3391 @option{--as-needed} option for all the files listed inside of it
3392 and restores previous @option{--as-needed} resp. @option{--no-as-needed}
3395 @item OUTPUT(@var{filename})
3396 @kindex OUTPUT(@var{filename})
3397 @cindex output file name in linker script
3398 The @code{OUTPUT} command names the output file. Using
3399 @code{OUTPUT(@var{filename})} in the linker script is exactly like using
3400 @samp{-o @var{filename}} on the command line (@pxref{Options,,Command
3401 Line Options}). If both are used, the command line option takes
3404 You can use the @code{OUTPUT} command to define a default name for the
3405 output file other than the usual default of @file{a.out}.
3407 @item SEARCH_DIR(@var{path})
3408 @kindex SEARCH_DIR(@var{path})
3409 @cindex library search path in linker script
3410 @cindex archive search path in linker script
3411 @cindex search path in linker script
3412 The @code{SEARCH_DIR} command adds @var{path} to the list of paths where
3413 @command{ld} looks for archive libraries. Using
3414 @code{SEARCH_DIR(@var{path})} is exactly like using @samp{-L @var{path}}
3415 on the command line (@pxref{Options,,Command Line Options}). If both
3416 are used, then the linker will search both paths. Paths specified using
3417 the command line option are searched first.
3419 @item STARTUP(@var{filename})
3420 @kindex STARTUP(@var{filename})
3421 @cindex first input file
3422 The @code{STARTUP} command is just like the @code{INPUT} command, except
3423 that @var{filename} will become the first input file to be linked, as
3424 though it were specified first on the command line. This may be useful
3425 when using a system in which the entry point is always the start of the
3429 @ifclear SingleFormat
3430 @node Format Commands
3431 @subsection Commands Dealing with Object File Formats
3432 A couple of linker script commands deal with object file formats.
3435 @item OUTPUT_FORMAT(@var{bfdname})
3436 @itemx OUTPUT_FORMAT(@var{default}, @var{big}, @var{little})
3437 @kindex OUTPUT_FORMAT(@var{bfdname})
3438 @cindex output file format in linker script
3439 The @code{OUTPUT_FORMAT} command names the BFD format to use for the
3440 output file (@pxref{BFD}). Using @code{OUTPUT_FORMAT(@var{bfdname})} is
3441 exactly like using @samp{--oformat @var{bfdname}} on the command line
3442 (@pxref{Options,,Command Line Options}). If both are used, the command
3443 line option takes precedence.
3445 You can use @code{OUTPUT_FORMAT} with three arguments to use different
3446 formats based on the @samp{-EB} and @samp{-EL} command line options.
3447 This permits the linker script to set the output format based on the
3450 If neither @samp{-EB} nor @samp{-EL} are used, then the output format
3451 will be the first argument, @var{default}. If @samp{-EB} is used, the
3452 output format will be the second argument, @var{big}. If @samp{-EL} is
3453 used, the output format will be the third argument, @var{little}.
3455 For example, the default linker script for the MIPS ELF target uses this
3458 OUTPUT_FORMAT(elf32-bigmips, elf32-bigmips, elf32-littlemips)
3460 This says that the default format for the output file is
3461 @samp{elf32-bigmips}, but if the user uses the @samp{-EL} command line
3462 option, the output file will be created in the @samp{elf32-littlemips}
3465 @item TARGET(@var{bfdname})
3466 @kindex TARGET(@var{bfdname})
3467 @cindex input file format in linker script
3468 The @code{TARGET} command names the BFD format to use when reading input
3469 files. It affects subsequent @code{INPUT} and @code{GROUP} commands.
3470 This command is like using @samp{-b @var{bfdname}} on the command line
3471 (@pxref{Options,,Command Line Options}). If the @code{TARGET} command
3472 is used but @code{OUTPUT_FORMAT} is not, then the last @code{TARGET}
3473 command is also used to set the format for the output file. @xref{BFD}.
3478 @subsection Assign alias names to memory regions
3479 @kindex REGION_ALIAS(@var{alias}, @var{region})
3480 @cindex region alias
3481 @cindex region names
3483 Alias names can be added to existing memory regions created with the
3484 @ref{MEMORY} command. Each name corresponds to at most one memory region.
3487 REGION_ALIAS(@var{alias}, @var{region})
3490 The @code{REGION_ALIAS} function creates an alias name @var{alias} for the
3491 memory region @var{region}. This allows a flexible mapping of output sections
3492 to memory regions. An example follows.
3494 Suppose we have an application for embedded systems which come with various
3495 memory storage devices. All have a general purpose, volatile memory @code{RAM}
3496 that allows code execution or data storage. Some may have a read-only,
3497 non-volatile memory @code{ROM} that allows code execution and read-only data
3498 access. The last variant is a read-only, non-volatile memory @code{ROM2} with
3499 read-only data access and no code execution capability. We have four output
3504 @code{.text} program code;
3506 @code{.rodata} read-only data;
3508 @code{.data} read-write initialized data;
3510 @code{.bss} read-write zero initialized data.
3513 The goal is to provide a linker command file that contains a system independent
3514 part defining the output sections and a system dependent part mapping the
3515 output sections to the memory regions available on the system. Our embedded
3516 systems come with three different memory setups @code{A}, @code{B} and
3518 @multitable @columnfractions .25 .25 .25 .25
3519 @item Section @tab Variant A @tab Variant B @tab Variant C
3520 @item .text @tab RAM @tab ROM @tab ROM
3521 @item .rodata @tab RAM @tab ROM @tab ROM2
3522 @item .data @tab RAM @tab RAM/ROM @tab RAM/ROM2
3523 @item .bss @tab RAM @tab RAM @tab RAM
3525 The notation @code{RAM/ROM} or @code{RAM/ROM2} means that this section is
3526 loaded into region @code{ROM} or @code{ROM2} respectively. Please note that
3527 the load address of the @code{.data} section starts in all three variants at
3528 the end of the @code{.rodata} section.
3530 The base linker script that deals with the output sections follows. It
3531 includes the system dependent @code{linkcmds.memory} file that describes the
3534 INCLUDE linkcmds.memory
3547 .data : AT (rodata_end)
3552 data_size = SIZEOF(.data);
3553 data_load_start = LOADADDR(.data);
3561 Now we need three different @code{linkcmds.memory} files to define memory
3562 regions and alias names. The content of @code{linkcmds.memory} for the three
3563 variants @code{A}, @code{B} and @code{C}:
3566 Here everything goes into the @code{RAM}.
3570 RAM : ORIGIN = 0, LENGTH = 4M
3573 REGION_ALIAS("REGION_TEXT", RAM);
3574 REGION_ALIAS("REGION_RODATA", RAM);
3575 REGION_ALIAS("REGION_DATA", RAM);
3576 REGION_ALIAS("REGION_BSS", RAM);
3579 Program code and read-only data go into the @code{ROM}. Read-write data goes
3580 into the @code{RAM}. An image of the initialized data is loaded into the
3581 @code{ROM} and will be copied during system start into the @code{RAM}.
3585 ROM : ORIGIN = 0, LENGTH = 3M
3586 RAM : ORIGIN = 0x10000000, LENGTH = 1M
3589 REGION_ALIAS("REGION_TEXT", ROM);
3590 REGION_ALIAS("REGION_RODATA", ROM);
3591 REGION_ALIAS("REGION_DATA", RAM);
3592 REGION_ALIAS("REGION_BSS", RAM);
3595 Program code goes into the @code{ROM}. Read-only data goes into the
3596 @code{ROM2}. Read-write data goes into the @code{RAM}. An image of the
3597 initialized data is loaded into the @code{ROM2} and will be copied during
3598 system start into the @code{RAM}.
3602 ROM : ORIGIN = 0, LENGTH = 2M
3603 ROM2 : ORIGIN = 0x10000000, LENGTH = 1M
3604 RAM : ORIGIN = 0x20000000, LENGTH = 1M
3607 REGION_ALIAS("REGION_TEXT", ROM);
3608 REGION_ALIAS("REGION_RODATA", ROM2);
3609 REGION_ALIAS("REGION_DATA", RAM);
3610 REGION_ALIAS("REGION_BSS", RAM);
3614 It is possible to write a common system initialization routine to copy the
3615 @code{.data} section from @code{ROM} or @code{ROM2} into the @code{RAM} if
3620 extern char data_start [];
3621 extern char data_size [];
3622 extern char data_load_start [];
3624 void copy_data(void)
3626 if (data_start != data_load_start)
3628 memcpy(data_start, data_load_start, (size_t) data_size);
3633 @node Miscellaneous Commands
3634 @subsection Other Linker Script Commands
3635 There are a few other linker scripts commands.
3638 @item ASSERT(@var{exp}, @var{message})
3640 @cindex assertion in linker script
3641 Ensure that @var{exp} is non-zero. If it is zero, then exit the linker
3642 with an error code, and print @var{message}.
3644 Note that assertions are checked before the final stages of linking
3645 take place. This means that expressions involving symbols PROVIDEd
3646 inside section definitions will fail if the user has not set values
3647 for those symbols. The only exception to this rule is PROVIDEd
3648 symbols that just reference dot. Thus an assertion like this:
3653 PROVIDE (__stack = .);
3654 PROVIDE (__stack_size = 0x100);
3655 ASSERT ((__stack > (_end + __stack_size)), "Error: No room left for the stack");
3659 will fail if @code{__stack_size} is not defined elsewhere. Symbols
3660 PROVIDEd outside of section definitions are evaluated earlier, so they
3661 can be used inside ASSERTions. Thus:
3664 PROVIDE (__stack_size = 0x100);
3667 PROVIDE (__stack = .);
3668 ASSERT ((__stack > (_end + __stack_size)), "Error: No room left for the stack");
3674 @item EXTERN(@var{symbol} @var{symbol} @dots{})
3676 @cindex undefined symbol in linker script
3677 Force @var{symbol} to be entered in the output file as an undefined
3678 symbol. Doing this may, for example, trigger linking of additional
3679 modules from standard libraries. You may list several @var{symbol}s for
3680 each @code{EXTERN}, and you may use @code{EXTERN} multiple times. This
3681 command has the same effect as the @samp{-u} command-line option.
3683 @item FORCE_COMMON_ALLOCATION
3684 @kindex FORCE_COMMON_ALLOCATION
3685 @cindex common allocation in linker script
3686 This command has the same effect as the @samp{-d} command-line option:
3687 to make @command{ld} assign space to common symbols even if a relocatable
3688 output file is specified (@samp{-r}).
3690 @item INHIBIT_COMMON_ALLOCATION
3691 @kindex INHIBIT_COMMON_ALLOCATION
3692 @cindex common allocation in linker script
3693 This command has the same effect as the @samp{--no-define-common}
3694 command-line option: to make @code{ld} omit the assignment of addresses
3695 to common symbols even for a non-relocatable output file.
3697 @item INSERT [ AFTER | BEFORE ] @var{output_section}
3699 @cindex insert user script into default script
3700 This command is typically used in a script specified by @samp{-T} to
3701 augment the default @code{SECTIONS} with, for example, overlays. It
3702 inserts all prior linker script statements after (or before)
3703 @var{output_section}, and also causes @samp{-T} to not override the
3704 default linker script. The exact insertion point is as for orphan
3705 sections. @xref{Location Counter}. The insertion happens after the
3706 linker has mapped input sections to output sections. Prior to the
3707 insertion, since @samp{-T} scripts are parsed before the default
3708 linker script, statements in the @samp{-T} script occur before the
3709 default linker script statements in the internal linker representation
3710 of the script. In particular, input section assignments will be made
3711 to @samp{-T} output sections before those in the default script. Here
3712 is an example of how a @samp{-T} script using @code{INSERT} might look:
3719 .ov1 @{ ov1*(.text) @}
3720 .ov2 @{ ov2*(.text) @}
3726 @item NOCROSSREFS(@var{section} @var{section} @dots{})
3727 @kindex NOCROSSREFS(@var{sections})
3728 @cindex cross references
3729 This command may be used to tell @command{ld} to issue an error about any
3730 references among certain output sections.
3732 In certain types of programs, particularly on embedded systems when
3733 using overlays, when one section is loaded into memory, another section
3734 will not be. Any direct references between the two sections would be
3735 errors. For example, it would be an error if code in one section called
3736 a function defined in the other section.
3738 The @code{NOCROSSREFS} command takes a list of output section names. If
3739 @command{ld} detects any cross references between the sections, it reports
3740 an error and returns a non-zero exit status. Note that the
3741 @code{NOCROSSREFS} command uses output section names, not input section
3744 @item NOCROSSREFS_TO(@var{tosection} @var{fromsection} @dots{})
3745 @kindex NOCROSSREFS_TO(@var{tosection} @var{fromsections})
3746 @cindex cross references
3747 This command may be used to tell @command{ld} to issue an error about any
3748 references to one section from a list of other sections.
3750 The @code{NOCROSSREFS} command is useful when ensuring that two or more
3751 output sections are entirely independent but there are situations where
3752 a one-way dependency is needed. For example, in a multi-core application
3753 there may be shared code that can be called from each core but for safety
3754 must never call back.
3756 The @code{NOCROSSREFS_TO} command takes a list of output section names.
3757 The first section can not be referenced from any of the other sections.
3758 If @command{ld} detects any references to the first section from any of
3759 the other sections, it reports an error and returns a non-zero exit
3760 status. Note that the @code{NOCROSSREFS_TO} command uses output section
3761 names, not input section names.
3763 @ifclear SingleFormat
3764 @item OUTPUT_ARCH(@var{bfdarch})
3765 @kindex OUTPUT_ARCH(@var{bfdarch})
3766 @cindex machine architecture
3767 @cindex architecture
3768 Specify a particular output machine architecture. The argument is one
3769 of the names used by the BFD library (@pxref{BFD}). You can see the
3770 architecture of an object file by using the @code{objdump} program with
3771 the @samp{-f} option.
3774 @item LD_FEATURE(@var{string})
3775 @kindex LD_FEATURE(@var{string})
3776 This command may be used to modify @command{ld} behavior. If
3777 @var{string} is @code{"SANE_EXPR"} then absolute symbols and numbers
3778 in a script are simply treated as numbers everywhere.
3779 @xref{Expression Section}.
3783 @section Assigning Values to Symbols
3784 @cindex assignment in scripts
3785 @cindex symbol definition, scripts
3786 @cindex variables, defining
3787 You may assign a value to a symbol in a linker script. This will define
3788 the symbol and place it into the symbol table with a global scope.
3791 * Simple Assignments:: Simple Assignments
3794 * PROVIDE_HIDDEN:: PROVIDE_HIDDEN
3795 * Source Code Reference:: How to use a linker script defined symbol in source code
3798 @node Simple Assignments
3799 @subsection Simple Assignments
3801 You may assign to a symbol using any of the C assignment operators:
3804 @item @var{symbol} = @var{expression} ;
3805 @itemx @var{symbol} += @var{expression} ;
3806 @itemx @var{symbol} -= @var{expression} ;
3807 @itemx @var{symbol} *= @var{expression} ;
3808 @itemx @var{symbol} /= @var{expression} ;
3809 @itemx @var{symbol} <<= @var{expression} ;
3810 @itemx @var{symbol} >>= @var{expression} ;
3811 @itemx @var{symbol} &= @var{expression} ;
3812 @itemx @var{symbol} |= @var{expression} ;
3815 The first case will define @var{symbol} to the value of
3816 @var{expression}. In the other cases, @var{symbol} must already be
3817 defined, and the value will be adjusted accordingly.
3819 The special symbol name @samp{.} indicates the location counter. You
3820 may only use this within a @code{SECTIONS} command. @xref{Location Counter}.
3822 The semicolon after @var{expression} is required.
3824 Expressions are defined below; see @ref{Expressions}.
3826 You may write symbol assignments as commands in their own right, or as
3827 statements within a @code{SECTIONS} command, or as part of an output
3828 section description in a @code{SECTIONS} command.
3830 The section of the symbol will be set from the section of the
3831 expression; for more information, see @ref{Expression Section}.
3833 Here is an example showing the three different places that symbol
3834 assignments may be used:
3845 _bdata = (. + 3) & ~ 3;
3846 .data : @{ *(.data) @}
3850 In this example, the symbol @samp{floating_point} will be defined as
3851 zero. The symbol @samp{_etext} will be defined as the address following
3852 the last @samp{.text} input section. The symbol @samp{_bdata} will be
3853 defined as the address following the @samp{.text} output section aligned
3854 upward to a 4 byte boundary.
3859 For ELF targeted ports, define a symbol that will be hidden and won't be
3860 exported. The syntax is @code{HIDDEN(@var{symbol} = @var{expression})}.
3862 Here is the example from @ref{Simple Assignments}, rewritten to use
3866 HIDDEN(floating_point = 0);
3874 HIDDEN(_bdata = (. + 3) & ~ 3);
3875 .data : @{ *(.data) @}
3879 In this case none of the three symbols will be visible outside this module.
3884 In some cases, it is desirable for a linker script to define a symbol
3885 only if it is referenced and is not defined by any object included in
3886 the link. For example, traditional linkers defined the symbol
3887 @samp{etext}. However, ANSI C requires that the user be able to use
3888 @samp{etext} as a function name without encountering an error. The
3889 @code{PROVIDE} keyword may be used to define a symbol, such as
3890 @samp{etext}, only if it is referenced but not defined. The syntax is
3891 @code{PROVIDE(@var{symbol} = @var{expression})}.
3893 Here is an example of using @code{PROVIDE} to define @samp{etext}:
3906 In this example, if the program defines @samp{_etext} (with a leading
3907 underscore), the linker will give a multiple definition error. If, on
3908 the other hand, the program defines @samp{etext} (with no leading
3909 underscore), the linker will silently use the definition in the program.
3910 If the program references @samp{etext} but does not define it, the
3911 linker will use the definition in the linker script.
3913 @node PROVIDE_HIDDEN
3914 @subsection PROVIDE_HIDDEN
3915 @cindex PROVIDE_HIDDEN
3916 Similar to @code{PROVIDE}. For ELF targeted ports, the symbol will be
3917 hidden and won't be exported.
3919 @node Source Code Reference
3920 @subsection Source Code Reference
3922 Accessing a linker script defined variable from source code is not
3923 intuitive. In particular a linker script symbol is not equivalent to
3924 a variable declaration in a high level language, it is instead a
3925 symbol that does not have a value.
3927 Before going further, it is important to note that compilers often
3928 transform names in the source code into different names when they are
3929 stored in the symbol table. For example, Fortran compilers commonly
3930 prepend or append an underscore, and C++ performs extensive @samp{name
3931 mangling}. Therefore there might be a discrepancy between the name
3932 of a variable as it is used in source code and the name of the same
3933 variable as it is defined in a linker script. For example in C a
3934 linker script variable might be referred to as:
3940 But in the linker script it might be defined as:
3946 In the remaining examples however it is assumed that no name
3947 transformation has taken place.
3949 When a symbol is declared in a high level language such as C, two
3950 things happen. The first is that the compiler reserves enough space
3951 in the program's memory to hold the @emph{value} of the symbol. The
3952 second is that the compiler creates an entry in the program's symbol
3953 table which holds the symbol's @emph{address}. ie the symbol table
3954 contains the address of the block of memory holding the symbol's
3955 value. So for example the following C declaration, at file scope:
3961 creates an entry called @samp{foo} in the symbol table. This entry
3962 holds the address of an @samp{int} sized block of memory where the
3963 number 1000 is initially stored.
3965 When a program references a symbol the compiler generates code that
3966 first accesses the symbol table to find the address of the symbol's
3967 memory block and then code to read the value from that memory block.
3974 looks up the symbol @samp{foo} in the symbol table, gets the address
3975 associated with this symbol and then writes the value 1 into that
3982 looks up the symbol @samp{foo} in the symbol table, gets its address
3983 and then copies this address into the block of memory associated with
3984 the variable @samp{a}.
3986 Linker scripts symbol declarations, by contrast, create an entry in
3987 the symbol table but do not assign any memory to them. Thus they are
3988 an address without a value. So for example the linker script definition:
3994 creates an entry in the symbol table called @samp{foo} which holds
3995 the address of memory location 1000, but nothing special is stored at
3996 address 1000. This means that you cannot access the @emph{value} of a
3997 linker script defined symbol - it has no value - all you can do is
3998 access the @emph{address} of a linker script defined symbol.
4000 Hence when you are using a linker script defined symbol in source code
4001 you should always take the address of the symbol, and never attempt to
4002 use its value. For example suppose you want to copy the contents of a
4003 section of memory called .ROM into a section called .FLASH and the
4004 linker script contains these declarations:
4008 start_of_ROM = .ROM;
4009 end_of_ROM = .ROM + sizeof (.ROM);
4010 start_of_FLASH = .FLASH;
4014 Then the C source code to perform the copy would be:
4018 extern char start_of_ROM, end_of_ROM, start_of_FLASH;
4020 memcpy (& start_of_FLASH, & start_of_ROM, & end_of_ROM - & start_of_ROM);
4024 Note the use of the @samp{&} operators. These are correct.
4025 Alternatively the symbols can be treated as the names of vectors or
4026 arrays and then the code will again work as expected:
4030 extern char start_of_ROM[], end_of_ROM[], start_of_FLASH[];
4032 memcpy (start_of_FLASH, start_of_ROM, end_of_ROM - start_of_ROM);
4036 Note how using this method does not require the use of @samp{&}
4040 @section SECTIONS Command
4042 The @code{SECTIONS} command tells the linker how to map input sections
4043 into output sections, and how to place the output sections in memory.
4045 The format of the @code{SECTIONS} command is:
4049 @var{sections-command}
4050 @var{sections-command}
4055 Each @var{sections-command} may of be one of the following:
4059 an @code{ENTRY} command (@pxref{Entry Point,,Entry command})
4061 a symbol assignment (@pxref{Assignments})
4063 an output section description
4065 an overlay description
4068 The @code{ENTRY} command and symbol assignments are permitted inside the
4069 @code{SECTIONS} command for convenience in using the location counter in
4070 those commands. This can also make the linker script easier to
4071 understand because you can use those commands at meaningful points in
4072 the layout of the output file.
4074 Output section descriptions and overlay descriptions are described
4077 If you do not use a @code{SECTIONS} command in your linker script, the
4078 linker will place each input section into an identically named output
4079 section in the order that the sections are first encountered in the
4080 input files. If all input sections are present in the first file, for
4081 example, the order of sections in the output file will match the order
4082 in the first input file. The first section will be at address zero.
4085 * Output Section Description:: Output section description
4086 * Output Section Name:: Output section name
4087 * Output Section Address:: Output section address
4088 * Input Section:: Input section description
4089 * Output Section Data:: Output section data
4090 * Output Section Keywords:: Output section keywords
4091 * Output Section Discarding:: Output section discarding
4092 * Output Section Attributes:: Output section attributes
4093 * Overlay Description:: Overlay description
4096 @node Output Section Description
4097 @subsection Output Section Description
4098 The full description of an output section looks like this:
4101 @var{section} [@var{address}] [(@var{type})] :
4103 [ALIGN(@var{section_align}) | ALIGN_WITH_INPUT]
4104 [SUBALIGN(@var{subsection_align})]
4107 @var{output-section-command}
4108 @var{output-section-command}
4110 @} [>@var{region}] [AT>@var{lma_region}] [:@var{phdr} :@var{phdr} @dots{}] [=@var{fillexp}] [,]
4114 Most output sections do not use most of the optional section attributes.
4116 The whitespace around @var{section} is required, so that the section
4117 name is unambiguous. The colon and the curly braces are also required.
4118 The comma at the end may be required if a @var{fillexp} is used and
4119 the next @var{sections-command} looks like a continuation of the expression.
4120 The line breaks and other white space are optional.
4122 Each @var{output-section-command} may be one of the following:
4126 a symbol assignment (@pxref{Assignments})
4128 an input section description (@pxref{Input Section})
4130 data values to include directly (@pxref{Output Section Data})
4132 a special output section keyword (@pxref{Output Section Keywords})
4135 @node Output Section Name
4136 @subsection Output Section Name
4137 @cindex name, section
4138 @cindex section name
4139 The name of the output section is @var{section}. @var{section} must
4140 meet the constraints of your output format. In formats which only
4141 support a limited number of sections, such as @code{a.out}, the name
4142 must be one of the names supported by the format (@code{a.out}, for
4143 example, allows only @samp{.text}, @samp{.data} or @samp{.bss}). If the
4144 output format supports any number of sections, but with numbers and not
4145 names (as is the case for Oasys), the name should be supplied as a
4146 quoted numeric string. A section name may consist of any sequence of
4147 characters, but a name which contains any unusual characters such as
4148 commas must be quoted.
4150 The output section name @samp{/DISCARD/} is special; @ref{Output Section
4153 @node Output Section Address
4154 @subsection Output Section Address
4155 @cindex address, section
4156 @cindex section address
4157 The @var{address} is an expression for the VMA (the virtual memory
4158 address) of the output section. This address is optional, but if it
4159 is provided then the output address will be set exactly as specified.
4161 If the output address is not specified then one will be chosen for the
4162 section, based on the heuristic below. This address will be adjusted
4163 to fit the alignment requirement of the output section. The
4164 alignment requirement is the strictest alignment of any input section
4165 contained within the output section.
4167 The output section address heuristic is as follows:
4171 If an output memory @var{region} is set for the section then it
4172 is added to this region and its address will be the next free address
4176 If the MEMORY command has been used to create a list of memory
4177 regions then the first region which has attributes compatible with the
4178 section is selected to contain it. The section's output address will
4179 be the next free address in that region; @ref{MEMORY}.
4182 If no memory regions were specified, or none match the section then
4183 the output address will be based on the current value of the location
4191 .text . : @{ *(.text) @}
4198 .text : @{ *(.text) @}
4202 are subtly different. The first will set the address of the
4203 @samp{.text} output section to the current value of the location
4204 counter. The second will set it to the current value of the location
4205 counter aligned to the strictest alignment of any of the @samp{.text}
4208 The @var{address} may be an arbitrary expression; @ref{Expressions}.
4209 For example, if you want to align the section on a 0x10 byte boundary,
4210 so that the lowest four bits of the section address are zero, you could
4211 do something like this:
4213 .text ALIGN(0x10) : @{ *(.text) @}
4216 This works because @code{ALIGN} returns the current location counter
4217 aligned upward to the specified value.
4219 Specifying @var{address} for a section will change the value of the
4220 location counter, provided that the section is non-empty. (Empty
4221 sections are ignored).
4224 @subsection Input Section Description
4225 @cindex input sections
4226 @cindex mapping input sections to output sections
4227 The most common output section command is an input section description.
4229 The input section description is the most basic linker script operation.
4230 You use output sections to tell the linker how to lay out your program
4231 in memory. You use input section descriptions to tell the linker how to
4232 map the input files into your memory layout.
4235 * Input Section Basics:: Input section basics
4236 * Input Section Wildcards:: Input section wildcard patterns
4237 * Input Section Common:: Input section for common symbols
4238 * Input Section Keep:: Input section and garbage collection
4239 * Input Section Example:: Input section example
4242 @node Input Section Basics
4243 @subsubsection Input Section Basics
4244 @cindex input section basics
4245 An input section description consists of a file name optionally followed
4246 by a list of section names in parentheses.
4248 The file name and the section name may be wildcard patterns, which we
4249 describe further below (@pxref{Input Section Wildcards}).
4251 The most common input section description is to include all input
4252 sections with a particular name in the output section. For example, to
4253 include all input @samp{.text} sections, you would write:
4258 Here the @samp{*} is a wildcard which matches any file name. To exclude a list
4259 @cindex EXCLUDE_FILE
4260 of files from matching the file name wildcard, EXCLUDE_FILE may be used to
4261 match all files except the ones specified in the EXCLUDE_FILE list. For
4264 EXCLUDE_FILE (*crtend.o *otherfile.o) *(.ctors)
4267 will cause all .ctors sections from all files except @file{crtend.o}
4268 and @file{otherfile.o} to be included. The EXCLUDE_FILE can also be
4269 placed inside the section list, for example:
4271 *(EXCLUDE_FILE (*crtend.o *otherfile.o) .ctors)
4274 The result of this is identically to the previous example. Supporting
4275 two syntaxes for EXCLUDE_FILE is useful if the section list contains
4276 more than one section, as described below.
4278 There are two ways to include more than one section:
4284 The difference between these is the order in which the @samp{.text} and
4285 @samp{.rdata} input sections will appear in the output section. In the
4286 first example, they will be intermingled, appearing in the same order as
4287 they are found in the linker input. In the second example, all
4288 @samp{.text} input sections will appear first, followed by all
4289 @samp{.rdata} input sections.
4291 When using EXCLUDE_FILE with more than one section, if the exclusion
4292 is within the section list then the exclusion only applies to the
4293 immediately following section, for example:
4295 *(EXCLUDE_FILE (*somefile.o) .text .rdata)
4298 will cause all @samp{.text} sections from all files except
4299 @file{somefile.o} to be included, while all @samp{.rdata} sections
4300 from all files, including @file{somefile.o}, will be included. To
4301 exclude the @samp{.rdata} sections from @file{somefile.o} the example
4302 could be modified to:
4304 *(EXCLUDE_FILE (*somefile.o) .text EXCLUDE_FILE (*somefile.o) .rdata)
4307 Alternatively, placing the EXCLUDE_FILE outside of the section list,
4308 before the input file selection, will cause the exclusion to apply for
4309 all sections. Thus the previous example can be rewritten as:
4311 EXCLUDE_FILE (*somefile.o) *(.text .rdata)
4314 You can specify a file name to include sections from a particular file.
4315 You would do this if one or more of your files contain special data that
4316 needs to be at a particular location in memory. For example:
4321 To refine the sections that are included based on the section flags
4322 of an input section, INPUT_SECTION_FLAGS may be used.
4324 Here is a simple example for using Section header flags for ELF sections:
4329 .text : @{ INPUT_SECTION_FLAGS (SHF_MERGE & SHF_STRINGS) *(.text) @}
4330 .text2 : @{ INPUT_SECTION_FLAGS (!SHF_WRITE) *(.text) @}
4335 In this example, the output section @samp{.text} will be comprised of any
4336 input section matching the name *(.text) whose section header flags
4337 @code{SHF_MERGE} and @code{SHF_STRINGS} are set. The output section
4338 @samp{.text2} will be comprised of any input section matching the name *(.text)
4339 whose section header flag @code{SHF_WRITE} is clear.
4341 You can also specify files within archives by writing a pattern
4342 matching the archive, a colon, then the pattern matching the file,
4343 with no whitespace around the colon.
4347 matches file within archive
4349 matches the whole archive
4351 matches file but not one in an archive
4354 Either one or both of @samp{archive} and @samp{file} can contain shell
4355 wildcards. On DOS based file systems, the linker will assume that a
4356 single letter followed by a colon is a drive specifier, so
4357 @samp{c:myfile.o} is a simple file specification, not @samp{myfile.o}
4358 within an archive called @samp{c}. @samp{archive:file} filespecs may
4359 also be used within an @code{EXCLUDE_FILE} list, but may not appear in
4360 other linker script contexts. For instance, you cannot extract a file
4361 from an archive by using @samp{archive:file} in an @code{INPUT}
4364 If you use a file name without a list of sections, then all sections in
4365 the input file will be included in the output section. This is not
4366 commonly done, but it may by useful on occasion. For example:
4371 When you use a file name which is not an @samp{archive:file} specifier
4372 and does not contain any wild card
4373 characters, the linker will first see if you also specified the file
4374 name on the linker command line or in an @code{INPUT} command. If you
4375 did not, the linker will attempt to open the file as an input file, as
4376 though it appeared on the command line. Note that this differs from an
4377 @code{INPUT} command, because the linker will not search for the file in
4378 the archive search path.
4380 @node Input Section Wildcards
4381 @subsubsection Input Section Wildcard Patterns
4382 @cindex input section wildcards
4383 @cindex wildcard file name patterns
4384 @cindex file name wildcard patterns
4385 @cindex section name wildcard patterns
4386 In an input section description, either the file name or the section
4387 name or both may be wildcard patterns.
4389 The file name of @samp{*} seen in many examples is a simple wildcard
4390 pattern for the file name.
4392 The wildcard patterns are like those used by the Unix shell.
4396 matches any number of characters
4398 matches any single character
4400 matches a single instance of any of the @var{chars}; the @samp{-}
4401 character may be used to specify a range of characters, as in
4402 @samp{[a-z]} to match any lower case letter
4404 quotes the following character
4407 When a file name is matched with a wildcard, the wildcard characters
4408 will not match a @samp{/} character (used to separate directory names on
4409 Unix). A pattern consisting of a single @samp{*} character is an
4410 exception; it will always match any file name, whether it contains a
4411 @samp{/} or not. In a section name, the wildcard characters will match
4412 a @samp{/} character.
4414 File name wildcard patterns only match files which are explicitly
4415 specified on the command line or in an @code{INPUT} command. The linker
4416 does not search directories to expand wildcards.
4418 If a file name matches more than one wildcard pattern, or if a file name
4419 appears explicitly and is also matched by a wildcard pattern, the linker
4420 will use the first match in the linker script. For example, this
4421 sequence of input section descriptions is probably in error, because the
4422 @file{data.o} rule will not be used:
4424 .data : @{ *(.data) @}
4425 .data1 : @{ data.o(.data) @}
4428 @cindex SORT_BY_NAME
4429 Normally, the linker will place files and sections matched by wildcards
4430 in the order in which they are seen during the link. You can change
4431 this by using the @code{SORT_BY_NAME} keyword, which appears before a wildcard
4432 pattern in parentheses (e.g., @code{SORT_BY_NAME(.text*)}). When the
4433 @code{SORT_BY_NAME} keyword is used, the linker will sort the files or sections
4434 into ascending order by name before placing them in the output file.
4436 @cindex SORT_BY_ALIGNMENT
4437 @code{SORT_BY_ALIGNMENT} is very similar to @code{SORT_BY_NAME}. The
4438 difference is @code{SORT_BY_ALIGNMENT} will sort sections into
4439 descending order by alignment before placing them in the output file.
4440 Larger alignments are placed before smaller alignments in order to
4441 reduce the amount of padding necessary.
4443 @cindex SORT_BY_INIT_PRIORITY
4444 @code{SORT_BY_INIT_PRIORITY} is very similar to @code{SORT_BY_NAME}. The
4445 difference is @code{SORT_BY_INIT_PRIORITY} will sort sections into
4446 ascending order by numerical value of the GCC init_priority attribute
4447 encoded in the section name before placing them in the output file.
4450 @code{SORT} is an alias for @code{SORT_BY_NAME}.
4452 When there are nested section sorting commands in linker script, there
4453 can be at most 1 level of nesting for section sorting commands.
4457 @code{SORT_BY_NAME} (@code{SORT_BY_ALIGNMENT} (wildcard section pattern)).
4458 It will sort the input sections by name first, then by alignment if two
4459 sections have the same name.
4461 @code{SORT_BY_ALIGNMENT} (@code{SORT_BY_NAME} (wildcard section pattern)).
4462 It will sort the input sections by alignment first, then by name if two
4463 sections have the same alignment.
4465 @code{SORT_BY_NAME} (@code{SORT_BY_NAME} (wildcard section pattern)) is
4466 treated the same as @code{SORT_BY_NAME} (wildcard section pattern).
4468 @code{SORT_BY_ALIGNMENT} (@code{SORT_BY_ALIGNMENT} (wildcard section pattern))
4469 is treated the same as @code{SORT_BY_ALIGNMENT} (wildcard section pattern).
4471 All other nested section sorting commands are invalid.
4474 When both command line section sorting option and linker script
4475 section sorting command are used, section sorting command always
4476 takes precedence over the command line option.
4478 If the section sorting command in linker script isn't nested, the
4479 command line option will make the section sorting command to be
4480 treated as nested sorting command.
4484 @code{SORT_BY_NAME} (wildcard section pattern ) with
4485 @option{--sort-sections alignment} is equivalent to
4486 @code{SORT_BY_NAME} (@code{SORT_BY_ALIGNMENT} (wildcard section pattern)).
4488 @code{SORT_BY_ALIGNMENT} (wildcard section pattern) with
4489 @option{--sort-section name} is equivalent to
4490 @code{SORT_BY_ALIGNMENT} (@code{SORT_BY_NAME} (wildcard section pattern)).
4493 If the section sorting command in linker script is nested, the
4494 command line option will be ignored.
4497 @code{SORT_NONE} disables section sorting by ignoring the command line
4498 section sorting option.
4500 If you ever get confused about where input sections are going, use the
4501 @samp{-M} linker option to generate a map file. The map file shows
4502 precisely how input sections are mapped to output sections.
4504 This example shows how wildcard patterns might be used to partition
4505 files. This linker script directs the linker to place all @samp{.text}
4506 sections in @samp{.text} and all @samp{.bss} sections in @samp{.bss}.
4507 The linker will place the @samp{.data} section from all files beginning
4508 with an upper case character in @samp{.DATA}; for all other files, the
4509 linker will place the @samp{.data} section in @samp{.data}.
4513 .text : @{ *(.text) @}
4514 .DATA : @{ [A-Z]*(.data) @}
4515 .data : @{ *(.data) @}
4516 .bss : @{ *(.bss) @}
4521 @node Input Section Common
4522 @subsubsection Input Section for Common Symbols
4523 @cindex common symbol placement
4524 @cindex uninitialized data placement
4525 A special notation is needed for common symbols, because in many object
4526 file formats common symbols do not have a particular input section. The
4527 linker treats common symbols as though they are in an input section
4528 named @samp{COMMON}.
4530 You may use file names with the @samp{COMMON} section just as with any
4531 other input sections. You can use this to place common symbols from a
4532 particular input file in one section while common symbols from other
4533 input files are placed in another section.
4535 In most cases, common symbols in input files will be placed in the
4536 @samp{.bss} section in the output file. For example:
4538 .bss @{ *(.bss) *(COMMON) @}
4541 @cindex scommon section
4542 @cindex small common symbols
4543 Some object file formats have more than one type of common symbol. For
4544 example, the MIPS ELF object file format distinguishes standard common
4545 symbols and small common symbols. In this case, the linker will use a
4546 different special section name for other types of common symbols. In
4547 the case of MIPS ELF, the linker uses @samp{COMMON} for standard common
4548 symbols and @samp{.scommon} for small common symbols. This permits you
4549 to map the different types of common symbols into memory at different
4553 You will sometimes see @samp{[COMMON]} in old linker scripts. This
4554 notation is now considered obsolete. It is equivalent to
4557 @node Input Section Keep
4558 @subsubsection Input Section and Garbage Collection
4560 @cindex garbage collection
4561 When link-time garbage collection is in use (@samp{--gc-sections}),
4562 it is often useful to mark sections that should not be eliminated.
4563 This is accomplished by surrounding an input section's wildcard entry
4564 with @code{KEEP()}, as in @code{KEEP(*(.init))} or
4565 @code{KEEP(SORT_BY_NAME(*)(.ctors))}.
4567 @node Input Section Example
4568 @subsubsection Input Section Example
4569 The following example is a complete linker script. It tells the linker
4570 to read all of the sections from file @file{all.o} and place them at the
4571 start of output section @samp{outputa} which starts at location
4572 @samp{0x10000}. All of section @samp{.input1} from file @file{foo.o}
4573 follows immediately, in the same output section. All of section
4574 @samp{.input2} from @file{foo.o} goes into output section
4575 @samp{outputb}, followed by section @samp{.input1} from @file{foo1.o}.
4576 All of the remaining @samp{.input1} and @samp{.input2} sections from any
4577 files are written to output section @samp{outputc}.
4605 @node Output Section Data
4606 @subsection Output Section Data
4608 @cindex section data
4609 @cindex output section data
4610 @kindex BYTE(@var{expression})
4611 @kindex SHORT(@var{expression})
4612 @kindex LONG(@var{expression})
4613 @kindex QUAD(@var{expression})
4614 @kindex SQUAD(@var{expression})
4615 You can include explicit bytes of data in an output section by using
4616 @code{BYTE}, @code{SHORT}, @code{LONG}, @code{QUAD}, or @code{SQUAD} as
4617 an output section command. Each keyword is followed by an expression in
4618 parentheses providing the value to store (@pxref{Expressions}). The
4619 value of the expression is stored at the current value of the location
4622 The @code{BYTE}, @code{SHORT}, @code{LONG}, and @code{QUAD} commands
4623 store one, two, four, and eight bytes (respectively). After storing the
4624 bytes, the location counter is incremented by the number of bytes
4627 For example, this will store the byte 1 followed by the four byte value
4628 of the symbol @samp{addr}:
4634 When using a 64 bit host or target, @code{QUAD} and @code{SQUAD} are the
4635 same; they both store an 8 byte, or 64 bit, value. When both host and
4636 target are 32 bits, an expression is computed as 32 bits. In this case
4637 @code{QUAD} stores a 32 bit value zero extended to 64 bits, and
4638 @code{SQUAD} stores a 32 bit value sign extended to 64 bits.
4640 If the object file format of the output file has an explicit endianness,
4641 which is the normal case, the value will be stored in that endianness.
4642 When the object file format does not have an explicit endianness, as is
4643 true of, for example, S-records, the value will be stored in the
4644 endianness of the first input object file.
4646 Note---these commands only work inside a section description and not
4647 between them, so the following will produce an error from the linker:
4649 SECTIONS @{@ .text : @{@ *(.text) @}@ LONG(1) .data : @{@ *(.data) @}@ @}@
4651 whereas this will work:
4653 SECTIONS @{@ .text : @{@ *(.text) ; LONG(1) @}@ .data : @{@ *(.data) @}@ @}@
4656 @kindex FILL(@var{expression})
4657 @cindex holes, filling
4658 @cindex unspecified memory
4659 You may use the @code{FILL} command to set the fill pattern for the
4660 current section. It is followed by an expression in parentheses. Any
4661 otherwise unspecified regions of memory within the section (for example,
4662 gaps left due to the required alignment of input sections) are filled
4663 with the value of the expression, repeated as
4664 necessary. A @code{FILL} statement covers memory locations after the
4665 point at which it occurs in the section definition; by including more
4666 than one @code{FILL} statement, you can have different fill patterns in
4667 different parts of an output section.
4669 This example shows how to fill unspecified regions of memory with the
4675 The @code{FILL} command is similar to the @samp{=@var{fillexp}} output
4676 section attribute, but it only affects the
4677 part of the section following the @code{FILL} command, rather than the
4678 entire section. If both are used, the @code{FILL} command takes
4679 precedence. @xref{Output Section Fill}, for details on the fill
4682 @node Output Section Keywords
4683 @subsection Output Section Keywords
4684 There are a couple of keywords which can appear as output section
4688 @kindex CREATE_OBJECT_SYMBOLS
4689 @cindex input filename symbols
4690 @cindex filename symbols
4691 @item CREATE_OBJECT_SYMBOLS
4692 The command tells the linker to create a symbol for each input file.
4693 The name of each symbol will be the name of the corresponding input
4694 file. The section of each symbol will be the output section in which
4695 the @code{CREATE_OBJECT_SYMBOLS} command appears.
4697 This is conventional for the a.out object file format. It is not
4698 normally used for any other object file format.
4700 @kindex CONSTRUCTORS
4701 @cindex C++ constructors, arranging in link
4702 @cindex constructors, arranging in link
4704 When linking using the a.out object file format, the linker uses an
4705 unusual set construct to support C++ global constructors and
4706 destructors. When linking object file formats which do not support
4707 arbitrary sections, such as ECOFF and XCOFF, the linker will
4708 automatically recognize C++ global constructors and destructors by name.
4709 For these object file formats, the @code{CONSTRUCTORS} command tells the
4710 linker to place constructor information in the output section where the
4711 @code{CONSTRUCTORS} command appears. The @code{CONSTRUCTORS} command is
4712 ignored for other object file formats.
4714 The symbol @w{@code{__CTOR_LIST__}} marks the start of the global
4715 constructors, and the symbol @w{@code{__CTOR_END__}} marks the end.
4716 Similarly, @w{@code{__DTOR_LIST__}} and @w{@code{__DTOR_END__}} mark
4717 the start and end of the global destructors. The
4718 first word in the list is the number of entries, followed by the address
4719 of each constructor or destructor, followed by a zero word. The
4720 compiler must arrange to actually run the code. For these object file
4721 formats @sc{gnu} C++ normally calls constructors from a subroutine
4722 @code{__main}; a call to @code{__main} is automatically inserted into
4723 the startup code for @code{main}. @sc{gnu} C++ normally runs
4724 destructors either by using @code{atexit}, or directly from the function
4727 For object file formats such as @code{COFF} or @code{ELF} which support
4728 arbitrary section names, @sc{gnu} C++ will normally arrange to put the
4729 addresses of global constructors and destructors into the @code{.ctors}
4730 and @code{.dtors} sections. Placing the following sequence into your
4731 linker script will build the sort of table which the @sc{gnu} C++
4732 runtime code expects to see.
4736 LONG((__CTOR_END__ - __CTOR_LIST__) / 4 - 2)
4741 LONG((__DTOR_END__ - __DTOR_LIST__) / 4 - 2)
4747 If you are using the @sc{gnu} C++ support for initialization priority,
4748 which provides some control over the order in which global constructors
4749 are run, you must sort the constructors at link time to ensure that they
4750 are executed in the correct order. When using the @code{CONSTRUCTORS}
4751 command, use @samp{SORT_BY_NAME(CONSTRUCTORS)} instead. When using the
4752 @code{.ctors} and @code{.dtors} sections, use @samp{*(SORT_BY_NAME(.ctors))} and
4753 @samp{*(SORT_BY_NAME(.dtors))} instead of just @samp{*(.ctors)} and
4756 Normally the compiler and linker will handle these issues automatically,
4757 and you will not need to concern yourself with them. However, you may
4758 need to consider this if you are using C++ and writing your own linker
4763 @node Output Section Discarding
4764 @subsection Output Section Discarding
4765 @cindex discarding sections
4766 @cindex sections, discarding
4767 @cindex removing sections
4768 The linker will not normally create output sections with no contents.
4769 This is for convenience when referring to input sections that may or
4770 may not be present in any of the input files. For example:
4772 .foo : @{ *(.foo) @}
4775 will only create a @samp{.foo} section in the output file if there is a
4776 @samp{.foo} section in at least one input file, and if the input
4777 sections are not all empty. Other link script directives that allocate
4778 space in an output section will also create the output section. So
4779 too will assignments to dot even if the assignment does not create
4780 space, except for @samp{. = 0}, @samp{. = . + 0}, @samp{. = sym},
4781 @samp{. = . + sym} and @samp{. = ALIGN (. != 0, expr, 1)} when
4782 @samp{sym} is an absolute symbol of value 0 defined in the script.
4783 This allows you to force output of an empty section with @samp{. = .}.
4785 The linker will ignore address assignments (@pxref{Output Section Address})
4786 on discarded output sections, except when the linker script defines
4787 symbols in the output section. In that case the linker will obey
4788 the address assignments, possibly advancing dot even though the
4789 section is discarded.
4792 The special output section name @samp{/DISCARD/} may be used to discard
4793 input sections. Any input sections which are assigned to an output
4794 section named @samp{/DISCARD/} are not included in the output file.
4796 @node Output Section Attributes
4797 @subsection Output Section Attributes
4798 @cindex output section attributes
4799 We showed above that the full description of an output section looked
4804 @var{section} [@var{address}] [(@var{type})] :
4806 [ALIGN(@var{section_align})]
4807 [SUBALIGN(@var{subsection_align})]
4810 @var{output-section-command}
4811 @var{output-section-command}
4813 @} [>@var{region}] [AT>@var{lma_region}] [:@var{phdr} :@var{phdr} @dots{}] [=@var{fillexp}]
4817 We've already described @var{section}, @var{address}, and
4818 @var{output-section-command}. In this section we will describe the
4819 remaining section attributes.
4822 * Output Section Type:: Output section type
4823 * Output Section LMA:: Output section LMA
4824 * Forced Output Alignment:: Forced Output Alignment
4825 * Forced Input Alignment:: Forced Input Alignment
4826 * Output Section Constraint:: Output section constraint
4827 * Output Section Region:: Output section region
4828 * Output Section Phdr:: Output section phdr
4829 * Output Section Fill:: Output section fill
4832 @node Output Section Type
4833 @subsubsection Output Section Type
4834 Each output section may have a type. The type is a keyword in
4835 parentheses. The following types are defined:
4839 The section should be marked as not loadable, so that it will not be
4840 loaded into memory when the program is run.
4845 These type names are supported for backward compatibility, and are
4846 rarely used. They all have the same effect: the section should be
4847 marked as not allocatable, so that no memory is allocated for the
4848 section when the program is run.
4852 @cindex prevent unnecessary loading
4853 @cindex loading, preventing
4854 The linker normally sets the attributes of an output section based on
4855 the input sections which map into it. You can override this by using
4856 the section type. For example, in the script sample below, the
4857 @samp{ROM} section is addressed at memory location @samp{0} and does not
4858 need to be loaded when the program is run.
4862 ROM 0 (NOLOAD) : @{ @dots{} @}
4868 @node Output Section LMA
4869 @subsubsection Output Section LMA
4870 @kindex AT>@var{lma_region}
4871 @kindex AT(@var{lma})
4872 @cindex load address
4873 @cindex section load address
4874 Every section has a virtual address (VMA) and a load address (LMA); see
4875 @ref{Basic Script Concepts}. The virtual address is specified by the
4876 @pxref{Output Section Address} described earlier. The load address is
4877 specified by the @code{AT} or @code{AT>} keywords. Specifying a load
4878 address is optional.
4880 The @code{AT} keyword takes an expression as an argument. This
4881 specifies the exact load address of the section. The @code{AT>} keyword
4882 takes the name of a memory region as an argument. @xref{MEMORY}. The
4883 load address of the section is set to the next free address in the
4884 region, aligned to the section's alignment requirements.
4886 If neither @code{AT} nor @code{AT>} is specified for an allocatable
4887 section, the linker will use the following heuristic to determine the
4892 If the section has a specific VMA address, then this is used as
4893 the LMA address as well.
4896 If the section is not allocatable then its LMA is set to its VMA.
4899 Otherwise if a memory region can be found that is compatible
4900 with the current section, and this region contains at least one
4901 section, then the LMA is set so the difference between the
4902 VMA and LMA is the same as the difference between the VMA and LMA of
4903 the last section in the located region.
4906 If no memory regions have been declared then a default region
4907 that covers the entire address space is used in the previous step.
4910 If no suitable region could be found, or there was no previous
4911 section then the LMA is set equal to the VMA.
4914 @cindex ROM initialized data
4915 @cindex initialized data in ROM
4916 This feature is designed to make it easy to build a ROM image. For
4917 example, the following linker script creates three output sections: one
4918 called @samp{.text}, which starts at @code{0x1000}, one called
4919 @samp{.mdata}, which is loaded at the end of the @samp{.text} section
4920 even though its VMA is @code{0x2000}, and one called @samp{.bss} to hold
4921 uninitialized data at address @code{0x3000}. The symbol @code{_data} is
4922 defined with the value @code{0x2000}, which shows that the location
4923 counter holds the VMA value, not the LMA value.
4929 .text 0x1000 : @{ *(.text) _etext = . ; @}
4931 AT ( ADDR (.text) + SIZEOF (.text) )
4932 @{ _data = . ; *(.data); _edata = . ; @}
4934 @{ _bstart = . ; *(.bss) *(COMMON) ; _bend = . ;@}
4939 The run-time initialization code for use with a program generated with
4940 this linker script would include something like the following, to copy
4941 the initialized data from the ROM image to its runtime address. Notice
4942 how this code takes advantage of the symbols defined by the linker
4947 extern char _etext, _data, _edata, _bstart, _bend;
4948 char *src = &_etext;
4951 /* ROM has data at end of text; copy it. */
4952 while (dst < &_edata)
4956 for (dst = &_bstart; dst< &_bend; dst++)
4961 @node Forced Output Alignment
4962 @subsubsection Forced Output Alignment
4963 @kindex ALIGN(@var{section_align})
4964 @cindex forcing output section alignment
4965 @cindex output section alignment
4966 You can increase an output section's alignment by using ALIGN. As an
4967 alternative you can enforce that the difference between the VMA and LMA remains
4968 intact throughout this output section with the ALIGN_WITH_INPUT attribute.
4970 @node Forced Input Alignment
4971 @subsubsection Forced Input Alignment
4972 @kindex SUBALIGN(@var{subsection_align})
4973 @cindex forcing input section alignment
4974 @cindex input section alignment
4975 You can force input section alignment within an output section by using
4976 SUBALIGN. The value specified overrides any alignment given by input
4977 sections, whether larger or smaller.
4979 @node Output Section Constraint
4980 @subsubsection Output Section Constraint
4983 @cindex constraints on output sections
4984 You can specify that an output section should only be created if all
4985 of its input sections are read-only or all of its input sections are
4986 read-write by using the keyword @code{ONLY_IF_RO} and
4987 @code{ONLY_IF_RW} respectively.
4989 @node Output Section Region
4990 @subsubsection Output Section Region
4991 @kindex >@var{region}
4992 @cindex section, assigning to memory region
4993 @cindex memory regions and sections
4994 You can assign a section to a previously defined region of memory by
4995 using @samp{>@var{region}}. @xref{MEMORY}.
4997 Here is a simple example:
5000 MEMORY @{ rom : ORIGIN = 0x1000, LENGTH = 0x1000 @}
5001 SECTIONS @{ ROM : @{ *(.text) @} >rom @}
5005 @node Output Section Phdr
5006 @subsubsection Output Section Phdr
5008 @cindex section, assigning to program header
5009 @cindex program headers and sections
5010 You can assign a section to a previously defined program segment by
5011 using @samp{:@var{phdr}}. @xref{PHDRS}. If a section is assigned to
5012 one or more segments, then all subsequent allocated sections will be
5013 assigned to those segments as well, unless they use an explicitly
5014 @code{:@var{phdr}} modifier. You can use @code{:NONE} to tell the
5015 linker to not put the section in any segment at all.
5017 Here is a simple example:
5020 PHDRS @{ text PT_LOAD ; @}
5021 SECTIONS @{ .text : @{ *(.text) @} :text @}
5025 @node Output Section Fill
5026 @subsubsection Output Section Fill
5027 @kindex =@var{fillexp}
5028 @cindex section fill pattern
5029 @cindex fill pattern, entire section
5030 You can set the fill pattern for an entire section by using
5031 @samp{=@var{fillexp}}. @var{fillexp} is an expression
5032 (@pxref{Expressions}). Any otherwise unspecified regions of memory
5033 within the output section (for example, gaps left due to the required
5034 alignment of input sections) will be filled with the value, repeated as
5035 necessary. If the fill expression is a simple hex number, ie. a string
5036 of hex digit starting with @samp{0x} and without a trailing @samp{k} or @samp{M}, then
5037 an arbitrarily long sequence of hex digits can be used to specify the
5038 fill pattern; Leading zeros become part of the pattern too. For all
5039 other cases, including extra parentheses or a unary @code{+}, the fill
5040 pattern is the four least significant bytes of the value of the
5041 expression. In all cases, the number is big-endian.
5043 You can also change the fill value with a @code{FILL} command in the
5044 output section commands; (@pxref{Output Section Data}).
5046 Here is a simple example:
5049 SECTIONS @{ .text : @{ *(.text) @} =0x90909090 @}
5053 @node Overlay Description
5054 @subsection Overlay Description
5057 An overlay description provides an easy way to describe sections which
5058 are to be loaded as part of a single memory image but are to be run at
5059 the same memory address. At run time, some sort of overlay manager will
5060 copy the overlaid sections in and out of the runtime memory address as
5061 required, perhaps by simply manipulating addressing bits. This approach
5062 can be useful, for example, when a certain region of memory is faster
5065 Overlays are described using the @code{OVERLAY} command. The
5066 @code{OVERLAY} command is used within a @code{SECTIONS} command, like an
5067 output section description. The full syntax of the @code{OVERLAY}
5068 command is as follows:
5071 OVERLAY [@var{start}] : [NOCROSSREFS] [AT ( @var{ldaddr} )]
5075 @var{output-section-command}
5076 @var{output-section-command}
5078 @} [:@var{phdr}@dots{}] [=@var{fill}]
5081 @var{output-section-command}
5082 @var{output-section-command}
5084 @} [:@var{phdr}@dots{}] [=@var{fill}]
5086 @} [>@var{region}] [:@var{phdr}@dots{}] [=@var{fill}] [,]
5090 Everything is optional except @code{OVERLAY} (a keyword), and each
5091 section must have a name (@var{secname1} and @var{secname2} above). The
5092 section definitions within the @code{OVERLAY} construct are identical to
5093 those within the general @code{SECTIONS} construct (@pxref{SECTIONS}),
5094 except that no addresses and no memory regions may be defined for
5095 sections within an @code{OVERLAY}.
5097 The comma at the end may be required if a @var{fill} is used and
5098 the next @var{sections-command} looks like a continuation of the expression.
5100 The sections are all defined with the same starting address. The load
5101 addresses of the sections are arranged such that they are consecutive in
5102 memory starting at the load address used for the @code{OVERLAY} as a
5103 whole (as with normal section definitions, the load address is optional,
5104 and defaults to the start address; the start address is also optional,
5105 and defaults to the current value of the location counter).
5107 If the @code{NOCROSSREFS} keyword is used, and there are any
5108 references among the sections, the linker will report an error. Since
5109 the sections all run at the same address, it normally does not make
5110 sense for one section to refer directly to another.
5111 @xref{Miscellaneous Commands, NOCROSSREFS}.
5113 For each section within the @code{OVERLAY}, the linker automatically
5114 provides two symbols. The symbol @code{__load_start_@var{secname}} is
5115 defined as the starting load address of the section. The symbol
5116 @code{__load_stop_@var{secname}} is defined as the final load address of
5117 the section. Any characters within @var{secname} which are not legal
5118 within C identifiers are removed. C (or assembler) code may use these
5119 symbols to move the overlaid sections around as necessary.
5121 At the end of the overlay, the value of the location counter is set to
5122 the start address of the overlay plus the size of the largest section.
5124 Here is an example. Remember that this would appear inside a
5125 @code{SECTIONS} construct.
5128 OVERLAY 0x1000 : AT (0x4000)
5130 .text0 @{ o1/*.o(.text) @}
5131 .text1 @{ o2/*.o(.text) @}
5136 This will define both @samp{.text0} and @samp{.text1} to start at
5137 address 0x1000. @samp{.text0} will be loaded at address 0x4000, and
5138 @samp{.text1} will be loaded immediately after @samp{.text0}. The
5139 following symbols will be defined if referenced: @code{__load_start_text0},
5140 @code{__load_stop_text0}, @code{__load_start_text1},
5141 @code{__load_stop_text1}.
5143 C code to copy overlay @code{.text1} into the overlay area might look
5148 extern char __load_start_text1, __load_stop_text1;
5149 memcpy ((char *) 0x1000, &__load_start_text1,
5150 &__load_stop_text1 - &__load_start_text1);
5154 Note that the @code{OVERLAY} command is just syntactic sugar, since
5155 everything it does can be done using the more basic commands. The above
5156 example could have been written identically as follows.
5160 .text0 0x1000 : AT (0x4000) @{ o1/*.o(.text) @}
5161 PROVIDE (__load_start_text0 = LOADADDR (.text0));
5162 PROVIDE (__load_stop_text0 = LOADADDR (.text0) + SIZEOF (.text0));
5163 .text1 0x1000 : AT (0x4000 + SIZEOF (.text0)) @{ o2/*.o(.text) @}
5164 PROVIDE (__load_start_text1 = LOADADDR (.text1));
5165 PROVIDE (__load_stop_text1 = LOADADDR (.text1) + SIZEOF (.text1));
5166 . = 0x1000 + MAX (SIZEOF (.text0), SIZEOF (.text1));
5171 @section MEMORY Command
5173 @cindex memory regions
5174 @cindex regions of memory
5175 @cindex allocating memory
5176 @cindex discontinuous memory
5177 The linker's default configuration permits allocation of all available
5178 memory. You can override this by using the @code{MEMORY} command.
5180 The @code{MEMORY} command describes the location and size of blocks of
5181 memory in the target. You can use it to describe which memory regions
5182 may be used by the linker, and which memory regions it must avoid. You
5183 can then assign sections to particular memory regions. The linker will
5184 set section addresses based on the memory regions, and will warn about
5185 regions that become too full. The linker will not shuffle sections
5186 around to fit into the available regions.
5188 A linker script may contain many uses of the @code{MEMORY} command,
5189 however, all memory blocks defined are treated as if they were
5190 specified inside a single @code{MEMORY} command. The syntax for
5196 @var{name} [(@var{attr})] : ORIGIN = @var{origin}, LENGTH = @var{len}
5202 The @var{name} is a name used in the linker script to refer to the
5203 region. The region name has no meaning outside of the linker script.
5204 Region names are stored in a separate name space, and will not conflict
5205 with symbol names, file names, or section names. Each memory region
5206 must have a distinct name within the @code{MEMORY} command. However you can
5207 add later alias names to existing memory regions with the @ref{REGION_ALIAS}
5210 @cindex memory region attributes
5211 The @var{attr} string is an optional list of attributes that specify
5212 whether to use a particular memory region for an input section which is
5213 not explicitly mapped in the linker script. As described in
5214 @ref{SECTIONS}, if you do not specify an output section for some input
5215 section, the linker will create an output section with the same name as
5216 the input section. If you define region attributes, the linker will use
5217 them to select the memory region for the output section that it creates.
5219 The @var{attr} string must consist only of the following characters:
5234 Invert the sense of any of the attributes that follow
5237 If a unmapped section matches any of the listed attributes other than
5238 @samp{!}, it will be placed in the memory region. The @samp{!}
5239 attribute reverses this test, so that an unmapped section will be placed
5240 in the memory region only if it does not match any of the listed
5246 The @var{origin} is an numerical expression for the start address of
5247 the memory region. The expression must evaluate to a constant and it
5248 cannot involve any symbols. The keyword @code{ORIGIN} may be
5249 abbreviated to @code{org} or @code{o} (but not, for example,
5255 The @var{len} is an expression for the size in bytes of the memory
5256 region. As with the @var{origin} expression, the expression must
5257 be numerical only and must evaluate to a constant. The keyword
5258 @code{LENGTH} may be abbreviated to @code{len} or @code{l}.
5260 In the following example, we specify that there are two memory regions
5261 available for allocation: one starting at @samp{0} for 256 kilobytes,
5262 and the other starting at @samp{0x40000000} for four megabytes. The
5263 linker will place into the @samp{rom} memory region every section which
5264 is not explicitly mapped into a memory region, and is either read-only
5265 or executable. The linker will place other sections which are not
5266 explicitly mapped into a memory region into the @samp{ram} memory
5273 rom (rx) : ORIGIN = 0, LENGTH = 256K
5274 ram (!rx) : org = 0x40000000, l = 4M
5279 Once you define a memory region, you can direct the linker to place
5280 specific output sections into that memory region by using the
5281 @samp{>@var{region}} output section attribute. For example, if you have
5282 a memory region named @samp{mem}, you would use @samp{>mem} in the
5283 output section definition. @xref{Output Section Region}. If no address
5284 was specified for the output section, the linker will set the address to
5285 the next available address within the memory region. If the combined
5286 output sections directed to a memory region are too large for the
5287 region, the linker will issue an error message.
5289 It is possible to access the origin and length of a memory in an
5290 expression via the @code{ORIGIN(@var{memory})} and
5291 @code{LENGTH(@var{memory})} functions:
5295 _fstack = ORIGIN(ram) + LENGTH(ram) - 4;
5300 @section PHDRS Command
5302 @cindex program headers
5303 @cindex ELF program headers
5304 @cindex program segments
5305 @cindex segments, ELF
5306 The ELF object file format uses @dfn{program headers}, also knows as
5307 @dfn{segments}. The program headers describe how the program should be
5308 loaded into memory. You can print them out by using the @code{objdump}
5309 program with the @samp{-p} option.
5311 When you run an ELF program on a native ELF system, the system loader
5312 reads the program headers in order to figure out how to load the
5313 program. This will only work if the program headers are set correctly.
5314 This manual does not describe the details of how the system loader
5315 interprets program headers; for more information, see the ELF ABI.
5317 The linker will create reasonable program headers by default. However,
5318 in some cases, you may need to specify the program headers more
5319 precisely. You may use the @code{PHDRS} command for this purpose. When
5320 the linker sees the @code{PHDRS} command in the linker script, it will
5321 not create any program headers other than the ones specified.
5323 The linker only pays attention to the @code{PHDRS} command when
5324 generating an ELF output file. In other cases, the linker will simply
5325 ignore @code{PHDRS}.
5327 This is the syntax of the @code{PHDRS} command. The words @code{PHDRS},
5328 @code{FILEHDR}, @code{AT}, and @code{FLAGS} are keywords.
5334 @var{name} @var{type} [ FILEHDR ] [ PHDRS ] [ AT ( @var{address} ) ]
5335 [ FLAGS ( @var{flags} ) ] ;
5340 The @var{name} is used only for reference in the @code{SECTIONS} command
5341 of the linker script. It is not put into the output file. Program
5342 header names are stored in a separate name space, and will not conflict
5343 with symbol names, file names, or section names. Each program header
5344 must have a distinct name. The headers are processed in order and it
5345 is usual for them to map to sections in ascending load address order.
5347 Certain program header types describe segments of memory which the
5348 system loader will load from the file. In the linker script, you
5349 specify the contents of these segments by placing allocatable output
5350 sections in the segments. You use the @samp{:@var{phdr}} output section
5351 attribute to place a section in a particular segment. @xref{Output
5354 It is normal to put certain sections in more than one segment. This
5355 merely implies that one segment of memory contains another. You may
5356 repeat @samp{:@var{phdr}}, using it once for each segment which should
5357 contain the section.
5359 If you place a section in one or more segments using @samp{:@var{phdr}},
5360 then the linker will place all subsequent allocatable sections which do
5361 not specify @samp{:@var{phdr}} in the same segments. This is for
5362 convenience, since generally a whole set of contiguous sections will be
5363 placed in a single segment. You can use @code{:NONE} to override the
5364 default segment and tell the linker to not put the section in any
5369 You may use the @code{FILEHDR} and @code{PHDRS} keywords after
5370 the program header type to further describe the contents of the segment.
5371 The @code{FILEHDR} keyword means that the segment should include the ELF
5372 file header. The @code{PHDRS} keyword means that the segment should
5373 include the ELF program headers themselves. If applied to a loadable
5374 segment (@code{PT_LOAD}), all prior loadable segments must have one of
5377 The @var{type} may be one of the following. The numbers indicate the
5378 value of the keyword.
5381 @item @code{PT_NULL} (0)
5382 Indicates an unused program header.
5384 @item @code{PT_LOAD} (1)
5385 Indicates that this program header describes a segment to be loaded from
5388 @item @code{PT_DYNAMIC} (2)
5389 Indicates a segment where dynamic linking information can be found.
5391 @item @code{PT_INTERP} (3)
5392 Indicates a segment where the name of the program interpreter may be
5395 @item @code{PT_NOTE} (4)
5396 Indicates a segment holding note information.
5398 @item @code{PT_SHLIB} (5)
5399 A reserved program header type, defined but not specified by the ELF
5402 @item @code{PT_PHDR} (6)
5403 Indicates a segment where the program headers may be found.
5405 @item @code{PT_TLS} (7)
5406 Indicates a segment containing thread local storage.
5408 @item @var{expression}
5409 An expression giving the numeric type of the program header. This may
5410 be used for types not defined above.
5413 You can specify that a segment should be loaded at a particular address
5414 in memory by using an @code{AT} expression. This is identical to the
5415 @code{AT} command used as an output section attribute (@pxref{Output
5416 Section LMA}). The @code{AT} command for a program header overrides the
5417 output section attribute.
5419 The linker will normally set the segment flags based on the sections
5420 which comprise the segment. You may use the @code{FLAGS} keyword to
5421 explicitly specify the segment flags. The value of @var{flags} must be
5422 an integer. It is used to set the @code{p_flags} field of the program
5425 Here is an example of @code{PHDRS}. This shows a typical set of program
5426 headers used on a native ELF system.
5432 headers PT_PHDR PHDRS ;
5434 text PT_LOAD FILEHDR PHDRS ;
5436 dynamic PT_DYNAMIC ;
5442 .interp : @{ *(.interp) @} :text :interp
5443 .text : @{ *(.text) @} :text
5444 .rodata : @{ *(.rodata) @} /* defaults to :text */
5446 . = . + 0x1000; /* move to a new page in memory */
5447 .data : @{ *(.data) @} :data
5448 .dynamic : @{ *(.dynamic) @} :data :dynamic
5455 @section VERSION Command
5456 @kindex VERSION @{script text@}
5457 @cindex symbol versions
5458 @cindex version script
5459 @cindex versions of symbols
5460 The linker supports symbol versions when using ELF. Symbol versions are
5461 only useful when using shared libraries. The dynamic linker can use
5462 symbol versions to select a specific version of a function when it runs
5463 a program that may have been linked against an earlier version of the
5466 You can include a version script directly in the main linker script, or
5467 you can supply the version script as an implicit linker script. You can
5468 also use the @samp{--version-script} linker option.
5470 The syntax of the @code{VERSION} command is simply
5472 VERSION @{ version-script-commands @}
5475 The format of the version script commands is identical to that used by
5476 Sun's linker in Solaris 2.5. The version script defines a tree of
5477 version nodes. You specify the node names and interdependencies in the
5478 version script. You can specify which symbols are bound to which
5479 version nodes, and you can reduce a specified set of symbols to local
5480 scope so that they are not globally visible outside of the shared
5483 The easiest way to demonstrate the version script language is with a few
5509 This example version script defines three version nodes. The first
5510 version node defined is @samp{VERS_1.1}; it has no other dependencies.
5511 The script binds the symbol @samp{foo1} to @samp{VERS_1.1}. It reduces
5512 a number of symbols to local scope so that they are not visible outside
5513 of the shared library; this is done using wildcard patterns, so that any
5514 symbol whose name begins with @samp{old}, @samp{original}, or @samp{new}
5515 is matched. The wildcard patterns available are the same as those used
5516 in the shell when matching filenames (also known as ``globbing'').
5517 However, if you specify the symbol name inside double quotes, then the
5518 name is treated as literal, rather than as a glob pattern.
5520 Next, the version script defines node @samp{VERS_1.2}. This node
5521 depends upon @samp{VERS_1.1}. The script binds the symbol @samp{foo2}
5522 to the version node @samp{VERS_1.2}.
5524 Finally, the version script defines node @samp{VERS_2.0}. This node
5525 depends upon @samp{VERS_1.2}. The scripts binds the symbols @samp{bar1}
5526 and @samp{bar2} are bound to the version node @samp{VERS_2.0}.
5528 When the linker finds a symbol defined in a library which is not
5529 specifically bound to a version node, it will effectively bind it to an
5530 unspecified base version of the library. You can bind all otherwise
5531 unspecified symbols to a given version node by using @samp{global: *;}
5532 somewhere in the version script. Note that it's slightly crazy to use
5533 wildcards in a global spec except on the last version node. Global
5534 wildcards elsewhere run the risk of accidentally adding symbols to the
5535 set exported for an old version. That's wrong since older versions
5536 ought to have a fixed set of symbols.
5538 The names of the version nodes have no specific meaning other than what
5539 they might suggest to the person reading them. The @samp{2.0} version
5540 could just as well have appeared in between @samp{1.1} and @samp{1.2}.
5541 However, this would be a confusing way to write a version script.
5543 Node name can be omitted, provided it is the only version node
5544 in the version script. Such version script doesn't assign any versions to
5545 symbols, only selects which symbols will be globally visible out and which
5549 @{ global: foo; bar; local: *; @};
5552 When you link an application against a shared library that has versioned
5553 symbols, the application itself knows which version of each symbol it
5554 requires, and it also knows which version nodes it needs from each
5555 shared library it is linked against. Thus at runtime, the dynamic
5556 loader can make a quick check to make sure that the libraries you have
5557 linked against do in fact supply all of the version nodes that the
5558 application will need to resolve all of the dynamic symbols. In this
5559 way it is possible for the dynamic linker to know with certainty that
5560 all external symbols that it needs will be resolvable without having to
5561 search for each symbol reference.
5563 The symbol versioning is in effect a much more sophisticated way of
5564 doing minor version checking that SunOS does. The fundamental problem
5565 that is being addressed here is that typically references to external
5566 functions are bound on an as-needed basis, and are not all bound when
5567 the application starts up. If a shared library is out of date, a
5568 required interface may be missing; when the application tries to use
5569 that interface, it may suddenly and unexpectedly fail. With symbol
5570 versioning, the user will get a warning when they start their program if
5571 the libraries being used with the application are too old.
5573 There are several GNU extensions to Sun's versioning approach. The
5574 first of these is the ability to bind a symbol to a version node in the
5575 source file where the symbol is defined instead of in the versioning
5576 script. This was done mainly to reduce the burden on the library
5577 maintainer. You can do this by putting something like:
5579 __asm__(".symver original_foo,foo@@VERS_1.1");
5582 in the C source file. This renames the function @samp{original_foo} to
5583 be an alias for @samp{foo} bound to the version node @samp{VERS_1.1}.
5584 The @samp{local:} directive can be used to prevent the symbol
5585 @samp{original_foo} from being exported. A @samp{.symver} directive
5586 takes precedence over a version script.
5588 The second GNU extension is to allow multiple versions of the same
5589 function to appear in a given shared library. In this way you can make
5590 an incompatible change to an interface without increasing the major
5591 version number of the shared library, while still allowing applications
5592 linked against the old interface to continue to function.
5594 To do this, you must use multiple @samp{.symver} directives in the
5595 source file. Here is an example:
5598 __asm__(".symver original_foo,foo@@");
5599 __asm__(".symver old_foo,foo@@VERS_1.1");
5600 __asm__(".symver old_foo1,foo@@VERS_1.2");
5601 __asm__(".symver new_foo,foo@@@@VERS_2.0");
5604 In this example, @samp{foo@@} represents the symbol @samp{foo} bound to the
5605 unspecified base version of the symbol. The source file that contains this
5606 example would define 4 C functions: @samp{original_foo}, @samp{old_foo},
5607 @samp{old_foo1}, and @samp{new_foo}.
5609 When you have multiple definitions of a given symbol, there needs to be
5610 some way to specify a default version to which external references to
5611 this symbol will be bound. You can do this with the
5612 @samp{foo@@@@VERS_2.0} type of @samp{.symver} directive. You can only
5613 declare one version of a symbol as the default in this manner; otherwise
5614 you would effectively have multiple definitions of the same symbol.
5616 If you wish to bind a reference to a specific version of the symbol
5617 within the shared library, you can use the aliases of convenience
5618 (i.e., @samp{old_foo}), or you can use the @samp{.symver} directive to
5619 specifically bind to an external version of the function in question.
5621 You can also specify the language in the version script:
5624 VERSION extern "lang" @{ version-script-commands @}
5627 The supported @samp{lang}s are @samp{C}, @samp{C++}, and @samp{Java}.
5628 The linker will iterate over the list of symbols at the link time and
5629 demangle them according to @samp{lang} before matching them to the
5630 patterns specified in @samp{version-script-commands}. The default
5631 @samp{lang} is @samp{C}.
5633 Demangled names may contains spaces and other special characters. As
5634 described above, you can use a glob pattern to match demangled names,
5635 or you can use a double-quoted string to match the string exactly. In
5636 the latter case, be aware that minor differences (such as differing
5637 whitespace) between the version script and the demangler output will
5638 cause a mismatch. As the exact string generated by the demangler
5639 might change in the future, even if the mangled name does not, you
5640 should check that all of your version directives are behaving as you
5641 expect when you upgrade.
5644 @section Expressions in Linker Scripts
5647 The syntax for expressions in the linker script language is identical to
5648 that of C expressions. All expressions are evaluated as integers. All
5649 expressions are evaluated in the same size, which is 32 bits if both the
5650 host and target are 32 bits, and is otherwise 64 bits.
5652 You can use and set symbol values in expressions.
5654 The linker defines several special purpose builtin functions for use in
5658 * Constants:: Constants
5659 * Symbolic Constants:: Symbolic constants
5660 * Symbols:: Symbol Names
5661 * Orphan Sections:: Orphan Sections
5662 * Location Counter:: The Location Counter
5663 * Operators:: Operators
5664 * Evaluation:: Evaluation
5665 * Expression Section:: The Section of an Expression
5666 * Builtin Functions:: Builtin Functions
5670 @subsection Constants
5671 @cindex integer notation
5672 @cindex constants in linker scripts
5673 All constants are integers.
5675 As in C, the linker considers an integer beginning with @samp{0} to be
5676 octal, and an integer beginning with @samp{0x} or @samp{0X} to be
5677 hexadecimal. Alternatively the linker accepts suffixes of @samp{h} or
5678 @samp{H} for hexadecimal, @samp{o} or @samp{O} for octal, @samp{b} or
5679 @samp{B} for binary and @samp{d} or @samp{D} for decimal. Any integer
5680 value without a prefix or a suffix is considered to be decimal.
5682 @cindex scaled integers
5683 @cindex K and M integer suffixes
5684 @cindex M and K integer suffixes
5685 @cindex suffixes for integers
5686 @cindex integer suffixes
5687 In addition, you can use the suffixes @code{K} and @code{M} to scale a
5691 @c END TEXI2ROFF-KILL
5692 @code{1024} or @code{1024*1024}
5696 ${\rm 1024}$ or ${\rm 1024}^2$
5698 @c END TEXI2ROFF-KILL
5699 respectively. For example, the following
5700 all refer to the same quantity:
5709 Note - the @code{K} and @code{M} suffixes cannot be used in
5710 conjunction with the base suffixes mentioned above.
5712 @node Symbolic Constants
5713 @subsection Symbolic Constants
5714 @cindex symbolic constants
5716 It is possible to refer to target specific constants via the use of
5717 the @code{CONSTANT(@var{name})} operator, where @var{name} is one of:
5722 The target's maximum page size.
5724 @item COMMONPAGESIZE
5725 @kindex COMMONPAGESIZE
5726 The target's default page size.
5732 .text ALIGN (CONSTANT (MAXPAGESIZE)) : @{ *(.text) @}
5735 will create a text section aligned to the largest page boundary
5736 supported by the target.
5739 @subsection Symbol Names
5740 @cindex symbol names
5742 @cindex quoted symbol names
5744 Unless quoted, symbol names start with a letter, underscore, or period
5745 and may include letters, digits, underscores, periods, and hyphens.
5746 Unquoted symbol names must not conflict with any keywords. You can
5747 specify a symbol which contains odd characters or has the same name as a
5748 keyword by surrounding the symbol name in double quotes:
5751 "with a space" = "also with a space" + 10;
5754 Since symbols can contain many non-alphabetic characters, it is safest
5755 to delimit symbols with spaces. For example, @samp{A-B} is one symbol,
5756 whereas @samp{A - B} is an expression involving subtraction.
5758 @node Orphan Sections
5759 @subsection Orphan Sections
5761 Orphan sections are sections present in the input files which
5762 are not explicitly placed into the output file by the linker
5763 script. The linker will still copy these sections into the
5764 output file, but it has to guess as to where they should be
5765 placed. The linker uses a simple heuristic to do this. It
5766 attempts to place orphan sections after non-orphan sections of the
5767 same attribute, such as code vs data, loadable vs non-loadable, etc.
5768 If there is not enough room to do this then it places
5769 at the end of the file.
5771 For ELF targets, the attribute of the section includes section type as
5772 well as section flag.
5774 The command line options @samp{--orphan-handling} and @samp{--unique}
5775 (@pxref{Options,,Command Line Options}) can be used to control which
5776 output sections an orphan is placed in.
5778 If an orphaned section's name is representable as a C identifier then
5779 the linker will automatically @pxref{PROVIDE} two symbols:
5780 __start_SECNAME and __stop_SECNAME, where SECNAME is the name of the
5781 section. These indicate the start address and end address of the
5782 orphaned section respectively. Note: most section names are not
5783 representable as C identifiers because they contain a @samp{.}
5786 @node Location Counter
5787 @subsection The Location Counter
5790 @cindex location counter
5791 @cindex current output location
5792 The special linker variable @dfn{dot} @samp{.} always contains the
5793 current output location counter. Since the @code{.} always refers to a
5794 location in an output section, it may only appear in an expression
5795 within a @code{SECTIONS} command. The @code{.} symbol may appear
5796 anywhere that an ordinary symbol is allowed in an expression.
5799 Assigning a value to @code{.} will cause the location counter to be
5800 moved. This may be used to create holes in the output section. The
5801 location counter may not be moved backwards inside an output section,
5802 and may not be moved backwards outside of an output section if so
5803 doing creates areas with overlapping LMAs.
5819 In the previous example, the @samp{.text} section from @file{file1} is
5820 located at the beginning of the output section @samp{output}. It is
5821 followed by a 1000 byte gap. Then the @samp{.text} section from
5822 @file{file2} appears, also with a 1000 byte gap following before the
5823 @samp{.text} section from @file{file3}. The notation @samp{= 0x12345678}
5824 specifies what data to write in the gaps (@pxref{Output Section Fill}).
5826 @cindex dot inside sections
5827 Note: @code{.} actually refers to the byte offset from the start of the
5828 current containing object. Normally this is the @code{SECTIONS}
5829 statement, whose start address is 0, hence @code{.} can be used as an
5830 absolute address. If @code{.} is used inside a section description
5831 however, it refers to the byte offset from the start of that section,
5832 not an absolute address. Thus in a script like this:
5850 The @samp{.text} section will be assigned a starting address of 0x100
5851 and a size of exactly 0x200 bytes, even if there is not enough data in
5852 the @samp{.text} input sections to fill this area. (If there is too
5853 much data, an error will be produced because this would be an attempt to
5854 move @code{.} backwards). The @samp{.data} section will start at 0x500
5855 and it will have an extra 0x600 bytes worth of space after the end of
5856 the values from the @samp{.data} input sections and before the end of
5857 the @samp{.data} output section itself.
5859 @cindex dot outside sections
5860 Setting symbols to the value of the location counter outside of an
5861 output section statement can result in unexpected values if the linker
5862 needs to place orphan sections. For example, given the following:
5868 .text: @{ *(.text) @}
5872 .data: @{ *(.data) @}
5877 If the linker needs to place some input section, e.g. @code{.rodata},
5878 not mentioned in the script, it might choose to place that section
5879 between @code{.text} and @code{.data}. You might think the linker
5880 should place @code{.rodata} on the blank line in the above script, but
5881 blank lines are of no particular significance to the linker. As well,
5882 the linker doesn't associate the above symbol names with their
5883 sections. Instead, it assumes that all assignments or other
5884 statements belong to the previous output section, except for the
5885 special case of an assignment to @code{.}. I.e., the linker will
5886 place the orphan @code{.rodata} section as if the script was written
5893 .text: @{ *(.text) @}
5897 .rodata: @{ *(.rodata) @}
5898 .data: @{ *(.data) @}
5903 This may or may not be the script author's intention for the value of
5904 @code{start_of_data}. One way to influence the orphan section
5905 placement is to assign the location counter to itself, as the linker
5906 assumes that an assignment to @code{.} is setting the start address of
5907 a following output section and thus should be grouped with that
5908 section. So you could write:
5914 .text: @{ *(.text) @}
5919 .data: @{ *(.data) @}
5924 Now, the orphan @code{.rodata} section will be placed between
5925 @code{end_of_text} and @code{start_of_data}.
5929 @subsection Operators
5930 @cindex operators for arithmetic
5931 @cindex arithmetic operators
5932 @cindex precedence in expressions
5933 The linker recognizes the standard C set of arithmetic operators, with
5934 the standard bindings and precedence levels:
5937 @c END TEXI2ROFF-KILL
5939 precedence associativity Operators Notes
5945 5 left == != > < <= >=
5951 11 right &= += -= *= /= (2)
5955 (1) Prefix operators
5956 (2) @xref{Assignments}.
5960 \vskip \baselineskip
5961 %"lispnarrowing" is the extra indent used generally for smallexample
5962 \hskip\lispnarrowing\vbox{\offinterlineskip
5965 {\vrule#&\strut\hfil\ #\ \hfil&\vrule#&\strut\hfil\ #\ \hfil&\vrule#&\strut\hfil\ {\tt #}\ \hfil&\vrule#\cr
5966 height2pt&\omit&&\omit&&\omit&\cr
5967 &Precedence&& Associativity &&{\rm Operators}&\cr
5968 height2pt&\omit&&\omit&&\omit&\cr
5970 height2pt&\omit&&\omit&&\omit&\cr
5972 % '176 is tilde, '~' in tt font
5973 &1&&left&&\qquad- \char'176\ !\qquad\dag&\cr
5974 &2&&left&&* / \%&\cr
5977 &5&&left&&== != > < <= >=&\cr
5980 &8&&left&&{\&\&}&\cr
5983 &11&&right&&\qquad\&= += -= *= /=\qquad\ddag&\cr
5985 height2pt&\omit&&\omit&&\omit&\cr}
5990 @obeylines@parskip=0pt@parindent=0pt
5991 @dag@quad Prefix operators.
5992 @ddag@quad @xref{Assignments}.
5995 @c END TEXI2ROFF-KILL
5998 @subsection Evaluation
5999 @cindex lazy evaluation
6000 @cindex expression evaluation order
6001 The linker evaluates expressions lazily. It only computes the value of
6002 an expression when absolutely necessary.
6004 The linker needs some information, such as the value of the start
6005 address of the first section, and the origins and lengths of memory
6006 regions, in order to do any linking at all. These values are computed
6007 as soon as possible when the linker reads in the linker script.
6009 However, other values (such as symbol values) are not known or needed
6010 until after storage allocation. Such values are evaluated later, when
6011 other information (such as the sizes of output sections) is available
6012 for use in the symbol assignment expression.
6014 The sizes of sections cannot be known until after allocation, so
6015 assignments dependent upon these are not performed until after
6018 Some expressions, such as those depending upon the location counter
6019 @samp{.}, must be evaluated during section allocation.
6021 If the result of an expression is required, but the value is not
6022 available, then an error results. For example, a script like the
6028 .text 9+this_isnt_constant :
6034 will cause the error message @samp{non constant expression for initial
6037 @node Expression Section
6038 @subsection The Section of an Expression
6039 @cindex expression sections
6040 @cindex absolute expressions
6041 @cindex relative expressions
6042 @cindex absolute and relocatable symbols
6043 @cindex relocatable and absolute symbols
6044 @cindex symbols, relocatable and absolute
6045 Addresses and symbols may be section relative, or absolute. A section
6046 relative symbol is relocatable. If you request relocatable output
6047 using the @samp{-r} option, a further link operation may change the
6048 value of a section relative symbol. On the other hand, an absolute
6049 symbol will retain the same value throughout any further link
6052 Some terms in linker expressions are addresses. This is true of
6053 section relative symbols and for builtin functions that return an
6054 address, such as @code{ADDR}, @code{LOADADDR}, @code{ORIGIN} and
6055 @code{SEGMENT_START}. Other terms are simply numbers, or are builtin
6056 functions that return a non-address value, such as @code{LENGTH}.
6057 One complication is that unless you set @code{LD_FEATURE ("SANE_EXPR")}
6058 (@pxref{Miscellaneous Commands}), numbers and absolute symbols are treated
6059 differently depending on their location, for compatibility with older
6060 versions of @code{ld}. Expressions appearing outside an output
6061 section definition treat all numbers as absolute addresses.
6062 Expressions appearing inside an output section definition treat
6063 absolute symbols as numbers. If @code{LD_FEATURE ("SANE_EXPR")} is
6064 given, then absolute symbols and numbers are simply treated as numbers
6067 In the following simple example,
6074 __executable_start = 0x100;
6078 __data_start = 0x10;
6086 both @code{.} and @code{__executable_start} are set to the absolute
6087 address 0x100 in the first two assignments, then both @code{.} and
6088 @code{__data_start} are set to 0x10 relative to the @code{.data}
6089 section in the second two assignments.
6091 For expressions involving numbers, relative addresses and absolute
6092 addresses, ld follows these rules to evaluate terms:
6096 Unary operations on an absolute address or number, and binary
6097 operations on two absolute addresses or two numbers, or between one
6098 absolute address and a number, apply the operator to the value(s).
6100 Unary operations on a relative address, and binary operations on two
6101 relative addresses in the same section or between one relative address
6102 and a number, apply the operator to the offset part of the address(es).
6104 Other binary operations, that is, between two relative addresses not
6105 in the same section, or between a relative address and an absolute
6106 address, first convert any non-absolute term to an absolute address
6107 before applying the operator.
6110 The result section of each sub-expression is as follows:
6114 An operation involving only numbers results in a number.
6116 The result of comparisons, @samp{&&} and @samp{||} is also a number.
6118 The result of other binary arithmetic and logical operations on two
6119 relative addresses in the same section or two absolute addresses
6120 (after above conversions) is also a number when
6121 @code{LD_FEATURE ("SANE_EXPR")} or inside an output section definition
6122 but an absolute address otherwise.
6124 The result of other operations on relative addresses or one
6125 relative address and a number, is a relative address in the same
6126 section as the relative operand(s).
6128 The result of other operations on absolute addresses (after above
6129 conversions) is an absolute address.
6132 You can use the builtin function @code{ABSOLUTE} to force an expression
6133 to be absolute when it would otherwise be relative. For example, to
6134 create an absolute symbol set to the address of the end of the output
6135 section @samp{.data}:
6139 .data : @{ *(.data) _edata = ABSOLUTE(.); @}
6143 If @samp{ABSOLUTE} were not used, @samp{_edata} would be relative to the
6144 @samp{.data} section.
6146 Using @code{LOADADDR} also forces an expression absolute, since this
6147 particular builtin function returns an absolute address.
6149 @node Builtin Functions
6150 @subsection Builtin Functions
6151 @cindex functions in expressions
6152 The linker script language includes a number of builtin functions for
6153 use in linker script expressions.
6156 @item ABSOLUTE(@var{exp})
6157 @kindex ABSOLUTE(@var{exp})
6158 @cindex expression, absolute
6159 Return the absolute (non-relocatable, as opposed to non-negative) value
6160 of the expression @var{exp}. Primarily useful to assign an absolute
6161 value to a symbol within a section definition, where symbol values are
6162 normally section relative. @xref{Expression Section}.
6164 @item ADDR(@var{section})
6165 @kindex ADDR(@var{section})
6166 @cindex section address in expression
6167 Return the address (VMA) of the named @var{section}. Your
6168 script must previously have defined the location of that section. In
6169 the following example, @code{start_of_output_1}, @code{symbol_1} and
6170 @code{symbol_2} are assigned equivalent values, except that
6171 @code{symbol_1} will be relative to the @code{.output1} section while
6172 the other two will be absolute:
6178 start_of_output_1 = ABSOLUTE(.);
6183 symbol_1 = ADDR(.output1);
6184 symbol_2 = start_of_output_1;
6190 @item ALIGN(@var{align})
6191 @itemx ALIGN(@var{exp},@var{align})
6192 @kindex ALIGN(@var{align})
6193 @kindex ALIGN(@var{exp},@var{align})
6194 @cindex round up location counter
6195 @cindex align location counter
6196 @cindex round up expression
6197 @cindex align expression
6198 Return the location counter (@code{.}) or arbitrary expression aligned
6199 to the next @var{align} boundary. The single operand @code{ALIGN}
6200 doesn't change the value of the location counter---it just does
6201 arithmetic on it. The two operand @code{ALIGN} allows an arbitrary
6202 expression to be aligned upwards (@code{ALIGN(@var{align})} is
6203 equivalent to @code{ALIGN(ABSOLUTE(.), @var{align})}).
6205 Here is an example which aligns the output @code{.data} section to the
6206 next @code{0x2000} byte boundary after the preceding section and sets a
6207 variable within the section to the next @code{0x8000} boundary after the
6212 .data ALIGN(0x2000): @{
6214 variable = ALIGN(0x8000);
6220 The first use of @code{ALIGN} in this example specifies the location of
6221 a section because it is used as the optional @var{address} attribute of
6222 a section definition (@pxref{Output Section Address}). The second use
6223 of @code{ALIGN} is used to defines the value of a symbol.
6225 The builtin function @code{NEXT} is closely related to @code{ALIGN}.
6227 @item ALIGNOF(@var{section})
6228 @kindex ALIGNOF(@var{section})
6229 @cindex section alignment
6230 Return the alignment in bytes of the named @var{section}, if that section has
6231 been allocated. If the section has not been allocated when this is
6232 evaluated, the linker will report an error. In the following example,
6233 the alignment of the @code{.output} section is stored as the first
6234 value in that section.
6239 LONG (ALIGNOF (.output))
6246 @item BLOCK(@var{exp})
6247 @kindex BLOCK(@var{exp})
6248 This is a synonym for @code{ALIGN}, for compatibility with older linker
6249 scripts. It is most often seen when setting the address of an output
6252 @item DATA_SEGMENT_ALIGN(@var{maxpagesize}, @var{commonpagesize})
6253 @kindex DATA_SEGMENT_ALIGN(@var{maxpagesize}, @var{commonpagesize})
6254 This is equivalent to either
6256 (ALIGN(@var{maxpagesize}) + (. & (@var{maxpagesize} - 1)))
6260 (ALIGN(@var{maxpagesize})
6261 + ((. + @var{commonpagesize} - 1) & (@var{maxpagesize} - @var{commonpagesize})))
6264 depending on whether the latter uses fewer @var{commonpagesize} sized pages
6265 for the data segment (area between the result of this expression and
6266 @code{DATA_SEGMENT_END}) than the former or not.
6267 If the latter form is used, it means @var{commonpagesize} bytes of runtime
6268 memory will be saved at the expense of up to @var{commonpagesize} wasted
6269 bytes in the on-disk file.
6271 This expression can only be used directly in @code{SECTIONS} commands, not in
6272 any output section descriptions and only once in the linker script.
6273 @var{commonpagesize} should be less or equal to @var{maxpagesize} and should
6274 be the system page size the object wants to be optimized for (while still
6275 working on system page sizes up to @var{maxpagesize}).
6280 . = DATA_SEGMENT_ALIGN(0x10000, 0x2000);
6283 @item DATA_SEGMENT_END(@var{exp})
6284 @kindex DATA_SEGMENT_END(@var{exp})
6285 This defines the end of data segment for @code{DATA_SEGMENT_ALIGN}
6286 evaluation purposes.
6289 . = DATA_SEGMENT_END(.);
6292 @item DATA_SEGMENT_RELRO_END(@var{offset}, @var{exp})
6293 @kindex DATA_SEGMENT_RELRO_END(@var{offset}, @var{exp})
6294 This defines the end of the @code{PT_GNU_RELRO} segment when
6295 @samp{-z relro} option is used.
6296 When @samp{-z relro} option is not present, @code{DATA_SEGMENT_RELRO_END}
6297 does nothing, otherwise @code{DATA_SEGMENT_ALIGN} is padded so that
6298 @var{exp} + @var{offset} is aligned to the most commonly used page
6299 boundary for particular target. If present in the linker script,
6300 it must always come in between @code{DATA_SEGMENT_ALIGN} and
6301 @code{DATA_SEGMENT_END}. Evaluates to the second argument plus any
6302 padding needed at the end of the @code{PT_GNU_RELRO} segment due to
6306 . = DATA_SEGMENT_RELRO_END(24, .);
6309 @item DEFINED(@var{symbol})
6310 @kindex DEFINED(@var{symbol})
6311 @cindex symbol defaults
6312 Return 1 if @var{symbol} is in the linker global symbol table and is
6313 defined before the statement using DEFINED in the script, otherwise
6314 return 0. You can use this function to provide
6315 default values for symbols. For example, the following script fragment
6316 shows how to set a global symbol @samp{begin} to the first location in
6317 the @samp{.text} section---but if a symbol called @samp{begin} already
6318 existed, its value is preserved:
6324 begin = DEFINED(begin) ? begin : . ;
6332 @item LENGTH(@var{memory})
6333 @kindex LENGTH(@var{memory})
6334 Return the length of the memory region named @var{memory}.
6336 @item LOADADDR(@var{section})
6337 @kindex LOADADDR(@var{section})
6338 @cindex section load address in expression
6339 Return the absolute LMA of the named @var{section}. (@pxref{Output
6342 @item LOG2CEIL(@var{exp})
6343 @kindex LOG2CEIL(@var{exp})
6344 Return the binary logarithm of @var{exp} rounded towards infinity.
6345 @code{LOG2CEIL(0)} returns 0.
6348 @item MAX(@var{exp1}, @var{exp2})
6349 Returns the maximum of @var{exp1} and @var{exp2}.
6352 @item MIN(@var{exp1}, @var{exp2})
6353 Returns the minimum of @var{exp1} and @var{exp2}.
6355 @item NEXT(@var{exp})
6356 @kindex NEXT(@var{exp})
6357 @cindex unallocated address, next
6358 Return the next unallocated address that is a multiple of @var{exp}.
6359 This function is closely related to @code{ALIGN(@var{exp})}; unless you
6360 use the @code{MEMORY} command to define discontinuous memory for the
6361 output file, the two functions are equivalent.
6363 @item ORIGIN(@var{memory})
6364 @kindex ORIGIN(@var{memory})
6365 Return the origin of the memory region named @var{memory}.
6367 @item SEGMENT_START(@var{segment}, @var{default})
6368 @kindex SEGMENT_START(@var{segment}, @var{default})
6369 Return the base address of the named @var{segment}. If an explicit
6370 value has already been given for this segment (with a command-line
6371 @samp{-T} option) then that value will be returned otherwise the value
6372 will be @var{default}. At present, the @samp{-T} command-line option
6373 can only be used to set the base address for the ``text'', ``data'', and
6374 ``bss'' sections, but you can use @code{SEGMENT_START} with any segment
6377 @item SIZEOF(@var{section})
6378 @kindex SIZEOF(@var{section})
6379 @cindex section size
6380 Return the size in bytes of the named @var{section}, if that section has
6381 been allocated. If the section has not been allocated when this is
6382 evaluated, the linker will report an error. In the following example,
6383 @code{symbol_1} and @code{symbol_2} are assigned identical values:
6392 symbol_1 = .end - .start ;
6393 symbol_2 = SIZEOF(.output);
6398 @item SIZEOF_HEADERS
6399 @itemx sizeof_headers
6400 @kindex SIZEOF_HEADERS
6402 Return the size in bytes of the output file's headers. This is
6403 information which appears at the start of the output file. You can use
6404 this number when setting the start address of the first section, if you
6405 choose, to facilitate paging.
6407 @cindex not enough room for program headers
6408 @cindex program headers, not enough room
6409 When producing an ELF output file, if the linker script uses the
6410 @code{SIZEOF_HEADERS} builtin function, the linker must compute the
6411 number of program headers before it has determined all the section
6412 addresses and sizes. If the linker later discovers that it needs
6413 additional program headers, it will report an error @samp{not enough
6414 room for program headers}. To avoid this error, you must avoid using
6415 the @code{SIZEOF_HEADERS} function, or you must rework your linker
6416 script to avoid forcing the linker to use additional program headers, or
6417 you must define the program headers yourself using the @code{PHDRS}
6418 command (@pxref{PHDRS}).
6421 @node Implicit Linker Scripts
6422 @section Implicit Linker Scripts
6423 @cindex implicit linker scripts
6424 If you specify a linker input file which the linker can not recognize as
6425 an object file or an archive file, it will try to read the file as a
6426 linker script. If the file can not be parsed as a linker script, the
6427 linker will report an error.
6429 An implicit linker script will not replace the default linker script.
6431 Typically an implicit linker script would contain only symbol
6432 assignments, or the @code{INPUT}, @code{GROUP}, or @code{VERSION}
6435 Any input files read because of an implicit linker script will be read
6436 at the position in the command line where the implicit linker script was
6437 read. This can affect archive searching.
6440 @node Machine Dependent
6441 @chapter Machine Dependent Features
6443 @cindex machine dependencies
6444 @command{ld} has additional features on some platforms; the following
6445 sections describe them. Machines where @command{ld} has no additional
6446 functionality are not listed.
6450 * H8/300:: @command{ld} and the H8/300
6453 * i960:: @command{ld} and the Intel 960 family
6456 * M68HC11/68HC12:: @code{ld} and the Motorola 68HC11 and 68HC12 families
6459 * ARM:: @command{ld} and the ARM family
6462 * HPPA ELF32:: @command{ld} and HPPA 32-bit ELF
6465 * M68K:: @command{ld} and the Motorola 68K family
6468 * MIPS:: @command{ld} and the MIPS family
6471 * MMIX:: @command{ld} and MMIX
6474 * MSP430:: @command{ld} and MSP430
6477 * NDS32:: @command{ld} and NDS32
6480 * Nios II:: @command{ld} and the Altera Nios II
6483 * PowerPC ELF32:: @command{ld} and PowerPC 32-bit ELF Support
6486 * PowerPC64 ELF64:: @command{ld} and PowerPC64 64-bit ELF Support
6489 * SPU ELF:: @command{ld} and SPU ELF Support
6492 * TI COFF:: @command{ld} and TI COFF
6495 * WIN32:: @command{ld} and WIN32 (cygwin/mingw)
6498 * Xtensa:: @command{ld} and Xtensa Processors
6509 @section @command{ld} and the H8/300
6511 @cindex H8/300 support
6512 For the H8/300, @command{ld} can perform these global optimizations when
6513 you specify the @samp{--relax} command-line option.
6516 @cindex relaxing on H8/300
6517 @item relaxing address modes
6518 @command{ld} finds all @code{jsr} and @code{jmp} instructions whose
6519 targets are within eight bits, and turns them into eight-bit
6520 program-counter relative @code{bsr} and @code{bra} instructions,
6523 @cindex synthesizing on H8/300
6524 @item synthesizing instructions
6525 @c FIXME: specifically mov.b, or any mov instructions really? -> mov.b only, at least on H8, H8H, H8S
6526 @command{ld} finds all @code{mov.b} instructions which use the
6527 sixteen-bit absolute address form, but refer to the top
6528 page of memory, and changes them to use the eight-bit address form.
6529 (That is: the linker turns @samp{mov.b @code{@@}@var{aa}:16} into
6530 @samp{mov.b @code{@@}@var{aa}:8} whenever the address @var{aa} is in the
6531 top page of memory).
6533 @command{ld} finds all @code{mov} instructions which use the register
6534 indirect with 32-bit displacement addressing mode, but use a small
6535 displacement inside 16-bit displacement range, and changes them to use
6536 the 16-bit displacement form. (That is: the linker turns @samp{mov.b
6537 @code{@@}@var{d}:32,ERx} into @samp{mov.b @code{@@}@var{d}:16,ERx}
6538 whenever the displacement @var{d} is in the 16 bit signed integer
6539 range. Only implemented in ELF-format ld).
6541 @item bit manipulation instructions
6542 @command{ld} finds all bit manipulation instructions like @code{band, bclr,
6543 biand, bild, bior, bist, bixor, bld, bnot, bor, bset, bst, btst, bxor}
6544 which use 32 bit and 16 bit absolute address form, but refer to the top
6545 page of memory, and changes them to use the 8 bit address form.
6546 (That is: the linker turns @samp{bset #xx:3,@code{@@}@var{aa}:32} into
6547 @samp{bset #xx:3,@code{@@}@var{aa}:8} whenever the address @var{aa} is in
6548 the top page of memory).
6550 @item system control instructions
6551 @command{ld} finds all @code{ldc.w, stc.w} instructions which use the
6552 32 bit absolute address form, but refer to the top page of memory, and
6553 changes them to use 16 bit address form.
6554 (That is: the linker turns @samp{ldc.w @code{@@}@var{aa}:32,ccr} into
6555 @samp{ldc.w @code{@@}@var{aa}:16,ccr} whenever the address @var{aa} is in
6556 the top page of memory).
6566 @c This stuff is pointless to say unless you're especially concerned
6567 @c with Renesas chips; don't enable it for generic case, please.
6569 @chapter @command{ld} and Other Renesas Chips
6571 @command{ld} also supports the Renesas (formerly Hitachi) H8/300H,
6572 H8/500, and SH chips. No special features, commands, or command-line
6573 options are required for these chips.
6583 @section @command{ld} and the Intel 960 Family
6585 @cindex i960 support
6587 You can use the @samp{-A@var{architecture}} command line option to
6588 specify one of the two-letter names identifying members of the 960
6589 family; the option specifies the desired output target, and warns of any
6590 incompatible instructions in the input files. It also modifies the
6591 linker's search strategy for archive libraries, to support the use of
6592 libraries specific to each particular architecture, by including in the
6593 search loop names suffixed with the string identifying the architecture.
6595 For example, if your @command{ld} command line included @w{@samp{-ACA}} as
6596 well as @w{@samp{-ltry}}, the linker would look (in its built-in search
6597 paths, and in any paths you specify with @samp{-L}) for a library with
6610 The first two possibilities would be considered in any event; the last
6611 two are due to the use of @w{@samp{-ACA}}.
6613 You can meaningfully use @samp{-A} more than once on a command line, since
6614 the 960 architecture family allows combination of target architectures; each
6615 use will add another pair of name variants to search for when @w{@samp{-l}}
6616 specifies a library.
6618 @cindex @option{--relax} on i960
6619 @cindex relaxing on i960
6620 @command{ld} supports the @samp{--relax} option for the i960 family. If
6621 you specify @samp{--relax}, @command{ld} finds all @code{balx} and
6622 @code{calx} instructions whose targets are within 24 bits, and turns
6623 them into 24-bit program-counter relative @code{bal} and @code{cal}
6624 instructions, respectively. @command{ld} also turns @code{cal}
6625 instructions into @code{bal} instructions when it determines that the
6626 target subroutine is a leaf routine (that is, the target subroutine does
6627 not itself call any subroutines).
6644 @node M68HC11/68HC12
6645 @section @command{ld} and the Motorola 68HC11 and 68HC12 families
6647 @cindex M68HC11 and 68HC12 support
6649 @subsection Linker Relaxation
6651 For the Motorola 68HC11, @command{ld} can perform these global
6652 optimizations when you specify the @samp{--relax} command-line option.
6655 @cindex relaxing on M68HC11
6656 @item relaxing address modes
6657 @command{ld} finds all @code{jsr} and @code{jmp} instructions whose
6658 targets are within eight bits, and turns them into eight-bit
6659 program-counter relative @code{bsr} and @code{bra} instructions,
6662 @command{ld} also looks at all 16-bit extended addressing modes and
6663 transforms them in a direct addressing mode when the address is in
6664 page 0 (between 0 and 0x0ff).
6666 @item relaxing gcc instruction group
6667 When @command{gcc} is called with @option{-mrelax}, it can emit group
6668 of instructions that the linker can optimize to use a 68HC11 direct
6669 addressing mode. These instructions consists of @code{bclr} or
6670 @code{bset} instructions.
6674 @subsection Trampoline Generation
6676 @cindex trampoline generation on M68HC11
6677 @cindex trampoline generation on M68HC12
6678 For 68HC11 and 68HC12, @command{ld} can generate trampoline code to
6679 call a far function using a normal @code{jsr} instruction. The linker
6680 will also change the relocation to some far function to use the
6681 trampoline address instead of the function address. This is typically the
6682 case when a pointer to a function is taken. The pointer will in fact
6683 point to the function trampoline.
6691 @section @command{ld} and the ARM family
6693 @cindex ARM interworking support
6694 @kindex --support-old-code
6695 For the ARM, @command{ld} will generate code stubs to allow functions calls
6696 between ARM and Thumb code. These stubs only work with code that has
6697 been compiled and assembled with the @samp{-mthumb-interwork} command
6698 line option. If it is necessary to link with old ARM object files or
6699 libraries, which have not been compiled with the -mthumb-interwork
6700 option then the @samp{--support-old-code} command line switch should be
6701 given to the linker. This will make it generate larger stub functions
6702 which will work with non-interworking aware ARM code. Note, however,
6703 the linker does not support generating stubs for function calls to
6704 non-interworking aware Thumb code.
6706 @cindex thumb entry point
6707 @cindex entry point, thumb
6708 @kindex --thumb-entry=@var{entry}
6709 The @samp{--thumb-entry} switch is a duplicate of the generic
6710 @samp{--entry} switch, in that it sets the program's starting address.
6711 But it also sets the bottom bit of the address, so that it can be
6712 branched to using a BX instruction, and the program will start
6713 executing in Thumb mode straight away.
6715 @cindex PE import table prefixing
6716 @kindex --use-nul-prefixed-import-tables
6717 The @samp{--use-nul-prefixed-import-tables} switch is specifying, that
6718 the import tables idata4 and idata5 have to be generated with a zero
6719 element prefix for import libraries. This is the old style to generate
6720 import tables. By default this option is turned off.
6724 The @samp{--be8} switch instructs @command{ld} to generate BE8 format
6725 executables. This option is only valid when linking big-endian
6726 objects - ie ones which have been assembled with the @option{-EB}
6727 option. The resulting image will contain big-endian data and
6731 @kindex --target1-rel
6732 @kindex --target1-abs
6733 The @samp{R_ARM_TARGET1} relocation is typically used for entries in the
6734 @samp{.init_array} section. It is interpreted as either @samp{R_ARM_REL32}
6735 or @samp{R_ARM_ABS32}, depending on the target. The @samp{--target1-rel}
6736 and @samp{--target1-abs} switches override the default.
6739 @kindex --target2=@var{type}
6740 The @samp{--target2=type} switch overrides the default definition of the
6741 @samp{R_ARM_TARGET2} relocation. Valid values for @samp{type}, their
6742 meanings, and target defaults are as follows:
6745 @samp{R_ARM_REL32} (arm*-*-elf, arm*-*-eabi)
6747 @samp{R_ARM_ABS32} (arm*-*-symbianelf)
6749 @samp{R_ARM_GOT_PREL} (arm*-*-linux, arm*-*-*bsd)
6754 The @samp{R_ARM_V4BX} relocation (defined by the ARM AAELF
6755 specification) enables objects compiled for the ARMv4 architecture to be
6756 interworking-safe when linked with other objects compiled for ARMv4t, but
6757 also allows pure ARMv4 binaries to be built from the same ARMv4 objects.
6759 In the latter case, the switch @option{--fix-v4bx} must be passed to the
6760 linker, which causes v4t @code{BX rM} instructions to be rewritten as
6761 @code{MOV PC,rM}, since v4 processors do not have a @code{BX} instruction.
6763 In the former case, the switch should not be used, and @samp{R_ARM_V4BX}
6764 relocations are ignored.
6766 @cindex FIX_V4BX_INTERWORKING
6767 @kindex --fix-v4bx-interworking
6768 Replace @code{BX rM} instructions identified by @samp{R_ARM_V4BX}
6769 relocations with a branch to the following veneer:
6777 This allows generation of libraries/applications that work on ARMv4 cores
6778 and are still interworking safe. Note that the above veneer clobbers the
6779 condition flags, so may cause incorrect program behavior in rare cases.
6783 The @samp{--use-blx} switch enables the linker to use ARM/Thumb
6784 BLX instructions (available on ARMv5t and above) in various
6785 situations. Currently it is used to perform calls via the PLT from Thumb
6786 code using BLX rather than using BX and a mode-switching stub before
6787 each PLT entry. This should lead to such calls executing slightly faster.
6789 This option is enabled implicitly for SymbianOS, so there is no need to
6790 specify it if you are using that target.
6792 @cindex VFP11_DENORM_FIX
6793 @kindex --vfp11-denorm-fix
6794 The @samp{--vfp11-denorm-fix} switch enables a link-time workaround for a
6795 bug in certain VFP11 coprocessor hardware, which sometimes allows
6796 instructions with denorm operands (which must be handled by support code)
6797 to have those operands overwritten by subsequent instructions before
6798 the support code can read the intended values.
6800 The bug may be avoided in scalar mode if you allow at least one
6801 intervening instruction between a VFP11 instruction which uses a register
6802 and another instruction which writes to the same register, or at least two
6803 intervening instructions if vector mode is in use. The bug only affects
6804 full-compliance floating-point mode: you do not need this workaround if
6805 you are using "runfast" mode. Please contact ARM for further details.
6807 If you know you are using buggy VFP11 hardware, you can
6808 enable this workaround by specifying the linker option
6809 @samp{--vfp-denorm-fix=scalar} if you are using the VFP11 scalar
6810 mode only, or @samp{--vfp-denorm-fix=vector} if you are using
6811 vector mode (the latter also works for scalar code). The default is
6812 @samp{--vfp-denorm-fix=none}.
6814 If the workaround is enabled, instructions are scanned for
6815 potentially-troublesome sequences, and a veneer is created for each
6816 such sequence which may trigger the erratum. The veneer consists of the
6817 first instruction of the sequence and a branch back to the subsequent
6818 instruction. The original instruction is then replaced with a branch to
6819 the veneer. The extra cycles required to call and return from the veneer
6820 are sufficient to avoid the erratum in both the scalar and vector cases.
6822 @cindex ARM1176 erratum workaround
6823 @kindex --fix-arm1176
6824 @kindex --no-fix-arm1176
6825 The @samp{--fix-arm1176} switch enables a link-time workaround for an erratum
6826 in certain ARM1176 processors. The workaround is enabled by default if you
6827 are targeting ARM v6 (excluding ARM v6T2) or earlier. It can be disabled
6828 unconditionally by specifying @samp{--no-fix-arm1176}.
6830 Further information is available in the ``ARM1176JZ-S and ARM1176JZF-S
6831 Programmer Advice Notice'' available on the ARM documentation website at:
6832 http://infocenter.arm.com/.
6834 @cindex STM32L4xx erratum workaround
6835 @kindex --fix-stm32l4xx-629360
6837 The @samp{--fix-stm32l4xx-629360} switch enables a link-time
6838 workaround for a bug in the bus matrix / memory controller for some of
6839 the STM32 Cortex-M4 based products (STM32L4xx). When accessing
6840 off-chip memory via the affected bus for bus reads of 9 words or more,
6841 the bus can generate corrupt data and/or abort. These are only
6842 core-initiated accesses (not DMA), and might affect any access:
6843 integer loads such as LDM, POP and floating-point loads such as VLDM,
6844 VPOP. Stores are not affected.
6846 The bug can be avoided by splitting memory accesses into the
6847 necessary chunks to keep bus reads below 8 words.
6849 The workaround is not enabled by default, this is equivalent to use
6850 @samp{--fix-stm32l4xx-629360=none}. If you know you are using buggy
6851 STM32L4xx hardware, you can enable the workaround by specifying the
6852 linker option @samp{--fix-stm32l4xx-629360}, or the equivalent
6853 @samp{--fix-stm32l4xx-629360=default}.
6855 If the workaround is enabled, instructions are scanned for
6856 potentially-troublesome sequences, and a veneer is created for each
6857 such sequence which may trigger the erratum. The veneer consists in a
6858 replacement sequence emulating the behaviour of the original one and a
6859 branch back to the subsequent instruction. The original instruction is
6860 then replaced with a branch to the veneer.
6862 The workaround does not always preserve the memory access order for
6863 the LDMDB instruction, when the instruction loads the PC.
6865 The workaround is not able to handle problematic instructions when
6866 they are in the middle of an IT block, since a branch is not allowed
6867 there. In that case, the linker reports a warning and no replacement
6870 The workaround is not able to replace problematic instructions with a
6871 PC-relative branch instruction if the @samp{.text} section is too
6872 large. In that case, when the branch that replaces the original code
6873 cannot be encoded, the linker reports a warning and no replacement
6876 @cindex NO_ENUM_SIZE_WARNING
6877 @kindex --no-enum-size-warning
6878 The @option{--no-enum-size-warning} switch prevents the linker from
6879 warning when linking object files that specify incompatible EABI
6880 enumeration size attributes. For example, with this switch enabled,
6881 linking of an object file using 32-bit enumeration values with another
6882 using enumeration values fitted into the smallest possible space will
6885 @cindex NO_WCHAR_SIZE_WARNING
6886 @kindex --no-wchar-size-warning
6887 The @option{--no-wchar-size-warning} switch prevents the linker from
6888 warning when linking object files that specify incompatible EABI
6889 @code{wchar_t} size attributes. For example, with this switch enabled,
6890 linking of an object file using 32-bit @code{wchar_t} values with another
6891 using 16-bit @code{wchar_t} values will not be diagnosed.
6894 @kindex --pic-veneer
6895 The @samp{--pic-veneer} switch makes the linker use PIC sequences for
6896 ARM/Thumb interworking veneers, even if the rest of the binary
6897 is not PIC. This avoids problems on uClinux targets where
6898 @samp{--emit-relocs} is used to generate relocatable binaries.
6900 @cindex STUB_GROUP_SIZE
6901 @kindex --stub-group-size=@var{N}
6902 The linker will automatically generate and insert small sequences of
6903 code into a linked ARM ELF executable whenever an attempt is made to
6904 perform a function call to a symbol that is too far away. The
6905 placement of these sequences of instructions - called stubs - is
6906 controlled by the command line option @option{--stub-group-size=N}.
6907 The placement is important because a poor choice can create a need for
6908 duplicate stubs, increasing the code size. The linker will try to
6909 group stubs together in order to reduce interruptions to the flow of
6910 code, but it needs guidance as to how big these groups should be and
6911 where they should be placed.
6913 The value of @samp{N}, the parameter to the
6914 @option{--stub-group-size=} option controls where the stub groups are
6915 placed. If it is negative then all stubs are placed after the first
6916 branch that needs them. If it is positive then the stubs can be
6917 placed either before or after the branches that need them. If the
6918 value of @samp{N} is 1 (either +1 or -1) then the linker will choose
6919 exactly where to place groups of stubs, using its built in heuristics.
6920 A value of @samp{N} greater than 1 (or smaller than -1) tells the
6921 linker that a single group of stubs can service at most @samp{N} bytes
6922 from the input sections.
6924 The default, if @option{--stub-group-size=} is not specified, is
6927 Farcalls stubs insertion is fully supported for the ARM-EABI target
6928 only, because it relies on object files properties not present
6931 @cindex Cortex-A8 erratum workaround
6932 @kindex --fix-cortex-a8
6933 @kindex --no-fix-cortex-a8
6934 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}.
6936 The erratum only affects Thumb-2 code. Please contact ARM for further details.
6938 @cindex Cortex-A53 erratum 835769 workaround
6939 @kindex --fix-cortex-a53-835769
6940 @kindex --no-fix-cortex-a53-835769
6941 The @samp{--fix-cortex-a53-835769} switch enables a link-time workaround for erratum 835769 present on certain early revisions of Cortex-A53 processors. The workaround is disabled by default. It can be enabled by specifying @samp{--fix-cortex-a53-835769}, or disabled unconditionally by specifying @samp{--no-fix-cortex-a53-835769}.
6943 Please contact ARM for further details.
6945 @kindex --merge-exidx-entries
6946 @kindex --no-merge-exidx-entries
6947 @cindex Merging exidx entries
6948 The @samp{--no-merge-exidx-entries} switch disables the merging of adjacent exidx entries in debuginfo.
6951 @cindex 32-bit PLT entries
6952 The @samp{--long-plt} option enables the use of 16 byte PLT entries
6953 which support up to 4Gb of code. The default is to use 12 byte PLT
6954 entries which only support 512Mb of code.
6956 @kindex --no-apply-dynamic-relocs
6957 @cindex AArch64 rela addend
6958 The @samp{--no-apply-dynamic-relocs} option makes AArch64 linker do not apply
6959 link-time values for dynamic relocations.
6961 @cindex Placement of SG veneers
6962 All SG veneers are placed in the special output section @code{.gnu.sgstubs}.
6963 Its start address must be set, either with the command line option
6964 @samp{--section-start} or in a linker script, to indicate where to place these
6967 @kindex --cmse-implib
6968 @cindex Secure gateway import library
6969 The @samp{--cmse-implib} option requests that the import libraries
6970 specified by the @samp{--out-implib} and @samp{--in-implib} options are
6971 secure gateway import libraries, suitable for linking a non-secure
6972 executable against secure code as per ARMv8-M Security Extensions.
6974 @kindex --in-implib=@var{file}
6975 @cindex Input import library
6976 The @samp{--in-implib=file} specifies an input import library whose symbols
6977 must keep the same address in the executable being produced. A warning is
6978 given if no @samp{--out-implib} is given but new symbols have been introduced
6979 in the executable that should be listed in its import library. Otherwise, if
6980 @samp{--out-implib} is specified, the symbols are added to the output import
6981 library. A warning is also given if some symbols present in the input import
6982 library have disappeared from the executable. This option is only effective
6983 for Secure Gateway import libraries, ie. when @samp{--cmse-implib} is
6997 @section @command{ld} and HPPA 32-bit ELF Support
6998 @cindex HPPA multiple sub-space stubs
6999 @kindex --multi-subspace
7000 When generating a shared library, @command{ld} will by default generate
7001 import stubs suitable for use with a single sub-space application.
7002 The @samp{--multi-subspace} switch causes @command{ld} to generate export
7003 stubs, and different (larger) import stubs suitable for use with
7004 multiple sub-spaces.
7006 @cindex HPPA stub grouping
7007 @kindex --stub-group-size=@var{N}
7008 Long branch stubs and import/export stubs are placed by @command{ld} in
7009 stub sections located between groups of input sections.
7010 @samp{--stub-group-size} specifies the maximum size of a group of input
7011 sections handled by one stub section. Since branch offsets are signed,
7012 a stub section may serve two groups of input sections, one group before
7013 the stub section, and one group after it. However, when using
7014 conditional branches that require stubs, it may be better (for branch
7015 prediction) that stub sections only serve one group of input sections.
7016 A negative value for @samp{N} chooses this scheme, ensuring that
7017 branches to stubs always use a negative offset. Two special values of
7018 @samp{N} are recognized, @samp{1} and @samp{-1}. These both instruct
7019 @command{ld} to automatically size input section groups for the branch types
7020 detected, with the same behaviour regarding stub placement as other
7021 positive or negative values of @samp{N} respectively.
7023 Note that @samp{--stub-group-size} does not split input sections. A
7024 single input section larger than the group size specified will of course
7025 create a larger group (of one section). If input sections are too
7026 large, it may not be possible for a branch to reach its stub.
7039 @section @command{ld} and the Motorola 68K family
7041 @cindex Motorola 68K GOT generation
7042 @kindex --got=@var{type}
7043 The @samp{--got=@var{type}} option lets you choose the GOT generation scheme.
7044 The choices are @samp{single}, @samp{negative}, @samp{multigot} and
7045 @samp{target}. When @samp{target} is selected the linker chooses
7046 the default GOT generation scheme for the current target.
7047 @samp{single} tells the linker to generate a single GOT with
7048 entries only at non-negative offsets.
7049 @samp{negative} instructs the linker to generate a single GOT with
7050 entries at both negative and positive offsets. Not all environments
7052 @samp{multigot} allows the linker to generate several GOTs in the
7053 output file. All GOT references from a single input object
7054 file access the same GOT, but references from different input object
7055 files might access different GOTs. Not all environments support such GOTs.
7068 @section @command{ld} and the MIPS family
7070 @cindex MIPS microMIPS instruction choice selection
7073 The @samp{--insn32} and @samp{--no-insn32} options control the choice of
7074 microMIPS instructions used in code generated by the linker, such as that
7075 in the PLT or lazy binding stubs, or in relaxation. If @samp{--insn32} is
7076 used, then the linker only uses 32-bit instruction encodings. By default
7077 or if @samp{--no-insn32} is used, all instruction encodings are used,
7078 including 16-bit ones where possible.
7080 @cindex MIPS branch relocation check control
7081 @kindex --ignore-branch-isa
7082 @kindex --no-ignore-branch-isa
7083 The @samp{--ignore-branch-isa} and @samp{--no-ignore-branch-isa} options
7084 control branch relocation checks for invalid ISA mode transitions. If
7085 @samp{--ignore-branch-isa} is used, then the linker accepts any branch
7086 relocations and any ISA mode transition required is lost in relocation
7087 calculation, except for some cases of @code{BAL} instructions which meet
7088 relaxation conditions and are converted to equivalent @code{JALX}
7089 instructions as the associated relocation is calculated. By default
7090 or if @samp{--no-ignore-branch-isa} is used a check is made causing
7091 the loss of an ISA mode transition to produce an error.
7104 @section @code{ld} and MMIX
7105 For MMIX, there is a choice of generating @code{ELF} object files or
7106 @code{mmo} object files when linking. The simulator @code{mmix}
7107 understands the @code{mmo} format. The binutils @code{objcopy} utility
7108 can translate between the two formats.
7110 There is one special section, the @samp{.MMIX.reg_contents} section.
7111 Contents in this section is assumed to correspond to that of global
7112 registers, and symbols referring to it are translated to special symbols,
7113 equal to registers. In a final link, the start address of the
7114 @samp{.MMIX.reg_contents} section corresponds to the first allocated
7115 global register multiplied by 8. Register @code{$255} is not included in
7116 this section; it is always set to the program entry, which is at the
7117 symbol @code{Main} for @code{mmo} files.
7119 Global symbols with the prefix @code{__.MMIX.start.}, for example
7120 @code{__.MMIX.start..text} and @code{__.MMIX.start..data} are special.
7121 The default linker script uses these to set the default start address
7124 Initial and trailing multiples of zero-valued 32-bit words in a section,
7125 are left out from an mmo file.
7138 @section @code{ld} and MSP430
7139 For the MSP430 it is possible to select the MPU architecture. The flag @samp{-m [mpu type]}
7140 will select an appropriate linker script for selected MPU type. (To get a list of known MPUs
7141 just pass @samp{-m help} option to the linker).
7143 @cindex MSP430 extra sections
7144 The linker will recognize some extra sections which are MSP430 specific:
7147 @item @samp{.vectors}
7148 Defines a portion of ROM where interrupt vectors located.
7150 @item @samp{.bootloader}
7151 Defines the bootloader portion of the ROM (if applicable). Any code
7152 in this section will be uploaded to the MPU.
7154 @item @samp{.infomem}
7155 Defines an information memory section (if applicable). Any code in
7156 this section will be uploaded to the MPU.
7158 @item @samp{.infomemnobits}
7159 This is the same as the @samp{.infomem} section except that any code
7160 in this section will not be uploaded to the MPU.
7162 @item @samp{.noinit}
7163 Denotes a portion of RAM located above @samp{.bss} section.
7165 The last two sections are used by gcc.
7179 @section @code{ld} and NDS32
7180 @kindex relaxing on NDS32
7181 For NDS32, there are some options to select relaxation behavior. The linker
7182 relaxes objects according to these options.
7185 @item @samp{--m[no-]fp-as-gp}
7186 Disable/enable fp-as-gp relaxation.
7188 @item @samp{--mexport-symbols=FILE}
7189 Exporting symbols and their address into FILE as linker script.
7191 @item @samp{--m[no-]ex9}
7192 Disable/enable link-time EX9 relaxation.
7194 @item @samp{--mexport-ex9=FILE}
7195 Export the EX9 table after linking.
7197 @item @samp{--mimport-ex9=FILE}
7198 Import the Ex9 table for EX9 relaxation.
7200 @item @samp{--mupdate-ex9}
7201 Update the existing EX9 table.
7203 @item @samp{--mex9-limit=NUM}
7204 Maximum number of entries in the ex9 table.
7206 @item @samp{--mex9-loop-aware}
7207 Avoid generating the EX9 instruction inside the loop.
7209 @item @samp{--m[no-]ifc}
7210 Disable/enable the link-time IFC optimization.
7212 @item @samp{--mifc-loop-aware}
7213 Avoid generating the IFC instruction inside the loop.
7227 @section @command{ld} and the Altera Nios II
7228 @cindex Nios II call relaxation
7229 @kindex --relax on Nios II
7231 Call and immediate jump instructions on Nios II processors are limited to
7232 transferring control to addresses in the same 256MB memory segment,
7233 which may result in @command{ld} giving
7234 @samp{relocation truncated to fit} errors with very large programs.
7235 The command-line option @option{--relax} enables the generation of
7236 trampolines that can access the entire 32-bit address space for calls
7237 outside the normal @code{call} and @code{jmpi} address range. These
7238 trampolines are inserted at section boundaries, so may not themselves
7239 be reachable if an input section and its associated call trampolines are
7242 The @option{--relax} option is enabled by default unless @option{-r}
7243 is also specified. You can disable trampoline generation by using the
7244 @option{--no-relax} linker option. You can also disable this optimization
7245 locally by using the @samp{set .noat} directive in assembly-language
7246 source files, as the linker-inserted trampolines use the @code{at}
7247 register as a temporary.
7249 Note that the linker @option{--relax} option is independent of assembler
7250 relaxation options, and that using the GNU assembler's @option{-relax-all}
7251 option interferes with the linker's more selective call instruction relaxation.
7264 @section @command{ld} and PowerPC 32-bit ELF Support
7265 @cindex PowerPC long branches
7266 @kindex --relax on PowerPC
7267 Branches on PowerPC processors are limited to a signed 26-bit
7268 displacement, which may result in @command{ld} giving
7269 @samp{relocation truncated to fit} errors with very large programs.
7270 @samp{--relax} enables the generation of trampolines that can access
7271 the entire 32-bit address space. These trampolines are inserted at
7272 section boundaries, so may not themselves be reachable if an input
7273 section exceeds 33M in size. You may combine @samp{-r} and
7274 @samp{--relax} to add trampolines in a partial link. In that case
7275 both branches to undefined symbols and inter-section branches are also
7276 considered potentially out of range, and trampolines inserted.
7278 @cindex PowerPC ELF32 options
7283 Current PowerPC GCC accepts a @samp{-msecure-plt} option that
7284 generates code capable of using a newer PLT and GOT layout that has
7285 the security advantage of no executable section ever needing to be
7286 writable and no writable section ever being executable. PowerPC
7287 @command{ld} will generate this layout, including stubs to access the
7288 PLT, if all input files (including startup and static libraries) were
7289 compiled with @samp{-msecure-plt}. @samp{--bss-plt} forces the old
7290 BSS PLT (and GOT layout) which can give slightly better performance.
7292 @kindex --secure-plt
7294 @command{ld} will use the new PLT and GOT layout if it is linking new
7295 @samp{-fpic} or @samp{-fPIC} code, but does not do so automatically
7296 when linking non-PIC code. This option requests the new PLT and GOT
7297 layout. A warning will be given if some object file requires the old
7303 The new secure PLT and GOT are placed differently relative to other
7304 sections compared to older BSS PLT and GOT placement. The location of
7305 @code{.plt} must change because the new secure PLT is an initialized
7306 section while the old PLT is uninitialized. The reason for the
7307 @code{.got} change is more subtle: The new placement allows
7308 @code{.got} to be read-only in applications linked with
7309 @samp{-z relro -z now}. However, this placement means that
7310 @code{.sdata} cannot always be used in shared libraries, because the
7311 PowerPC ABI accesses @code{.sdata} in shared libraries from the GOT
7312 pointer. @samp{--sdata-got} forces the old GOT placement. PowerPC
7313 GCC doesn't use @code{.sdata} in shared libraries, so this option is
7314 really only useful for other compilers that may do so.
7316 @cindex PowerPC stub symbols
7317 @kindex --emit-stub-syms
7318 @item --emit-stub-syms
7319 This option causes @command{ld} to label linker stubs with a local
7320 symbol that encodes the stub type and destination.
7322 @cindex PowerPC TLS optimization
7323 @kindex --no-tls-optimize
7324 @item --no-tls-optimize
7325 PowerPC @command{ld} normally performs some optimization of code
7326 sequences used to access Thread-Local Storage. Use this option to
7327 disable the optimization.
7340 @node PowerPC64 ELF64
7341 @section @command{ld} and PowerPC64 64-bit ELF Support
7343 @cindex PowerPC64 ELF64 options
7345 @cindex PowerPC64 stub grouping
7346 @kindex --stub-group-size
7347 @item --stub-group-size
7348 Long branch stubs, PLT call stubs and TOC adjusting stubs are placed
7349 by @command{ld} in stub sections located between groups of input sections.
7350 @samp{--stub-group-size} specifies the maximum size of a group of input
7351 sections handled by one stub section. Since branch offsets are signed,
7352 a stub section may serve two groups of input sections, one group before
7353 the stub section, and one group after it. However, when using
7354 conditional branches that require stubs, it may be better (for branch
7355 prediction) that stub sections only serve one group of input sections.
7356 A negative value for @samp{N} chooses this scheme, ensuring that
7357 branches to stubs always use a negative offset. Two special values of
7358 @samp{N} are recognized, @samp{1} and @samp{-1}. These both instruct
7359 @command{ld} to automatically size input section groups for the branch types
7360 detected, with the same behaviour regarding stub placement as other
7361 positive or negative values of @samp{N} respectively.
7363 Note that @samp{--stub-group-size} does not split input sections. A
7364 single input section larger than the group size specified will of course
7365 create a larger group (of one section). If input sections are too
7366 large, it may not be possible for a branch to reach its stub.
7368 @cindex PowerPC64 stub symbols
7369 @kindex --emit-stub-syms
7370 @item --emit-stub-syms
7371 This option causes @command{ld} to label linker stubs with a local
7372 symbol that encodes the stub type and destination.
7374 @cindex PowerPC64 dot symbols
7376 @kindex --no-dotsyms
7379 These two options control how @command{ld} interprets version patterns
7380 in a version script. Older PowerPC64 compilers emitted both a
7381 function descriptor symbol with the same name as the function, and a
7382 code entry symbol with the name prefixed by a dot (@samp{.}). To
7383 properly version a function @samp{foo}, the version script thus needs
7384 to control both @samp{foo} and @samp{.foo}. The option
7385 @samp{--dotsyms}, on by default, automatically adds the required
7386 dot-prefixed patterns. Use @samp{--no-dotsyms} to disable this
7389 @cindex PowerPC64 register save/restore functions
7390 @kindex --save-restore-funcs
7391 @kindex --no-save-restore-funcs
7392 @item --save-restore-funcs
7393 @itemx --no-save-restore-funcs
7394 These two options control whether PowerPC64 @command{ld} automatically
7395 provides out-of-line register save and restore functions used by
7396 @samp{-Os} code. The default is to provide any such referenced
7397 function for a normal final link, and to not do so for a relocatable
7400 @cindex PowerPC64 TLS optimization
7401 @kindex --no-tls-optimize
7402 @item --no-tls-optimize
7403 PowerPC64 @command{ld} normally performs some optimization of code
7404 sequences used to access Thread-Local Storage. Use this option to
7405 disable the optimization.
7407 @cindex PowerPC64 __tls_get_addr optimization
7408 @kindex --tls-get-addr-optimize
7409 @kindex --no-tls-get-addr-optimize
7410 @item --tls-get-addr-optimize
7411 @itemx --no-tls-get-addr-optimize
7412 These options control whether PowerPC64 @command{ld} uses a special
7413 stub to call __tls_get_addr. PowerPC64 glibc 2.22 and later support
7414 an optimization that allows the second and subsequent calls to
7415 @code{__tls_get_addr} for a given symbol to be resolved by the special
7416 stub without calling in to glibc. By default the linker enables this
7417 option when glibc advertises the availability of __tls_get_addr_opt.
7418 Forcing this option on when using an older glibc won't do much besides
7419 slow down your applications, but may be useful if linking an
7420 application against an older glibc with the expectation that it will
7421 normally be used on systems having a newer glibc.
7423 @cindex PowerPC64 OPD optimization
7424 @kindex --no-opd-optimize
7425 @item --no-opd-optimize
7426 PowerPC64 @command{ld} normally removes @code{.opd} section entries
7427 corresponding to deleted link-once functions, or functions removed by
7428 the action of @samp{--gc-sections} or linker script @code{/DISCARD/}.
7429 Use this option to disable @code{.opd} optimization.
7431 @cindex PowerPC64 OPD spacing
7432 @kindex --non-overlapping-opd
7433 @item --non-overlapping-opd
7434 Some PowerPC64 compilers have an option to generate compressed
7435 @code{.opd} entries spaced 16 bytes apart, overlapping the third word,
7436 the static chain pointer (unused in C) with the first word of the next
7437 entry. This option expands such entries to the full 24 bytes.
7439 @cindex PowerPC64 TOC optimization
7440 @kindex --no-toc-optimize
7441 @item --no-toc-optimize
7442 PowerPC64 @command{ld} normally removes unused @code{.toc} section
7443 entries. Such entries are detected by examining relocations that
7444 reference the TOC in code sections. A reloc in a deleted code section
7445 marks a TOC word as unneeded, while a reloc in a kept code section
7446 marks a TOC word as needed. Since the TOC may reference itself, TOC
7447 relocs are also examined. TOC words marked as both needed and
7448 unneeded will of course be kept. TOC words without any referencing
7449 reloc are assumed to be part of a multi-word entry, and are kept or
7450 discarded as per the nearest marked preceding word. This works
7451 reliably for compiler generated code, but may be incorrect if assembly
7452 code is used to insert TOC entries. Use this option to disable the
7455 @cindex PowerPC64 multi-TOC
7456 @kindex --no-multi-toc
7457 @item --no-multi-toc
7458 If given any toc option besides @code{-mcmodel=medium} or
7459 @code{-mcmodel=large}, PowerPC64 GCC generates code for a TOC model
7461 entries are accessed with a 16-bit offset from r2. This limits the
7462 total TOC size to 64K. PowerPC64 @command{ld} extends this limit by
7463 grouping code sections such that each group uses less than 64K for its
7464 TOC entries, then inserts r2 adjusting stubs between inter-group
7465 calls. @command{ld} does not split apart input sections, so cannot
7466 help if a single input file has a @code{.toc} section that exceeds
7467 64K, most likely from linking multiple files with @command{ld -r}.
7468 Use this option to turn off this feature.
7470 @cindex PowerPC64 TOC sorting
7471 @kindex --no-toc-sort
7473 By default, @command{ld} sorts TOC sections so that those whose file
7474 happens to have a section called @code{.init} or @code{.fini} are
7475 placed first, followed by TOC sections referenced by code generated
7476 with PowerPC64 gcc's @code{-mcmodel=small}, and lastly TOC sections
7477 referenced only by code generated with PowerPC64 gcc's
7478 @code{-mcmodel=medium} or @code{-mcmodel=large} options. Doing this
7479 results in better TOC grouping for multi-TOC. Use this option to turn
7482 @cindex PowerPC64 PLT stub alignment
7484 @kindex --no-plt-align
7486 @itemx --no-plt-align
7487 Use these options to control whether individual PLT call stubs are
7488 padded so that they don't cross a 32-byte boundary, or to the
7489 specified power of two boundary when using @code{--plt-align=}. Note
7490 that this isn't alignment in the usual sense. By default PLT call
7491 stubs are packed tightly.
7493 @cindex PowerPC64 PLT call stub static chain
7494 @kindex --plt-static-chain
7495 @kindex --no-plt-static-chain
7496 @item --plt-static-chain
7497 @itemx --no-plt-static-chain
7498 Use these options to control whether PLT call stubs load the static
7499 chain pointer (r11). @code{ld} defaults to not loading the static
7500 chain since there is never any need to do so on a PLT call.
7502 @cindex PowerPC64 PLT call stub thread safety
7503 @kindex --plt-thread-safe
7504 @kindex --no-plt-thread-safe
7505 @item --plt-thread-safe
7506 @itemx --no-thread-safe
7507 With power7's weakly ordered memory model, it is possible when using
7508 lazy binding for ld.so to update a plt entry in one thread and have
7509 another thread see the individual plt entry words update in the wrong
7510 order, despite ld.so carefully writing in the correct order and using
7511 memory write barriers. To avoid this we need some sort of read
7512 barrier in the call stub, or use LD_BIND_NOW=1. By default, @code{ld}
7513 looks for calls to commonly used functions that create threads, and if
7514 seen, adds the necessary barriers. Use these options to change the
7529 @section @command{ld} and SPU ELF Support
7531 @cindex SPU ELF options
7537 This option marks an executable as a PIC plugin module.
7539 @cindex SPU overlays
7540 @kindex --no-overlays
7542 Normally, @command{ld} recognizes calls to functions within overlay
7543 regions, and redirects such calls to an overlay manager via a stub.
7544 @command{ld} also provides a built-in overlay manager. This option
7545 turns off all this special overlay handling.
7547 @cindex SPU overlay stub symbols
7548 @kindex --emit-stub-syms
7549 @item --emit-stub-syms
7550 This option causes @command{ld} to label overlay stubs with a local
7551 symbol that encodes the stub type and destination.
7553 @cindex SPU extra overlay stubs
7554 @kindex --extra-overlay-stubs
7555 @item --extra-overlay-stubs
7556 This option causes @command{ld} to add overlay call stubs on all
7557 function calls out of overlay regions. Normally stubs are not added
7558 on calls to non-overlay regions.
7560 @cindex SPU local store size
7561 @kindex --local-store=lo:hi
7562 @item --local-store=lo:hi
7563 @command{ld} usually checks that a final executable for SPU fits in
7564 the address range 0 to 256k. This option may be used to change the
7565 range. Disable the check entirely with @option{--local-store=0:0}.
7568 @kindex --stack-analysis
7569 @item --stack-analysis
7570 SPU local store space is limited. Over-allocation of stack space
7571 unnecessarily limits space available for code and data, while
7572 under-allocation results in runtime failures. If given this option,
7573 @command{ld} will provide an estimate of maximum stack usage.
7574 @command{ld} does this by examining symbols in code sections to
7575 determine the extents of functions, and looking at function prologues
7576 for stack adjusting instructions. A call-graph is created by looking
7577 for relocations on branch instructions. The graph is then searched
7578 for the maximum stack usage path. Note that this analysis does not
7579 find calls made via function pointers, and does not handle recursion
7580 and other cycles in the call graph. Stack usage may be
7581 under-estimated if your code makes such calls. Also, stack usage for
7582 dynamic allocation, e.g. alloca, will not be detected. If a link map
7583 is requested, detailed information about each function's stack usage
7584 and calls will be given.
7587 @kindex --emit-stack-syms
7588 @item --emit-stack-syms
7589 This option, if given along with @option{--stack-analysis} will result
7590 in @command{ld} emitting stack sizing symbols for each function.
7591 These take the form @code{__stack_<function_name>} for global
7592 functions, and @code{__stack_<number>_<function_name>} for static
7593 functions. @code{<number>} is the section id in hex. The value of
7594 such symbols is the stack requirement for the corresponding function.
7595 The symbol size will be zero, type @code{STT_NOTYPE}, binding
7596 @code{STB_LOCAL}, and section @code{SHN_ABS}.
7610 @section @command{ld}'s Support for Various TI COFF Versions
7611 @cindex TI COFF versions
7612 @kindex --format=@var{version}
7613 The @samp{--format} switch allows selection of one of the various
7614 TI COFF versions. The latest of this writing is 2; versions 0 and 1 are
7615 also supported. The TI COFF versions also vary in header byte-order
7616 format; @command{ld} will read any version or byte order, but the output
7617 header format depends on the default specified by the specific target.
7630 @section @command{ld} and WIN32 (cygwin/mingw)
7632 This section describes some of the win32 specific @command{ld} issues.
7633 See @ref{Options,,Command Line Options} for detailed description of the
7634 command line options mentioned here.
7637 @cindex import libraries
7638 @item import libraries
7639 The standard Windows linker creates and uses so-called import
7640 libraries, which contains information for linking to dll's. They are
7641 regular static archives and are handled as any other static
7642 archive. The cygwin and mingw ports of @command{ld} have specific
7643 support for creating such libraries provided with the
7644 @samp{--out-implib} command line option.
7646 @item exporting DLL symbols
7647 @cindex exporting DLL symbols
7648 The cygwin/mingw @command{ld} has several ways to export symbols for dll's.
7651 @item using auto-export functionality
7652 @cindex using auto-export functionality
7653 By default @command{ld} exports symbols with the auto-export functionality,
7654 which is controlled by the following command line options:
7657 @item --export-all-symbols [This is the default]
7658 @item --exclude-symbols
7659 @item --exclude-libs
7660 @item --exclude-modules-for-implib
7661 @item --version-script
7664 When auto-export is in operation, @command{ld} will export all the non-local
7665 (global and common) symbols it finds in a DLL, with the exception of a few
7666 symbols known to belong to the system's runtime and libraries. As it will
7667 often not be desirable to export all of a DLL's symbols, which may include
7668 private functions that are not part of any public interface, the command-line
7669 options listed above may be used to filter symbols out from the list for
7670 exporting. The @samp{--output-def} option can be used in order to see the
7671 final list of exported symbols with all exclusions taken into effect.
7673 If @samp{--export-all-symbols} is not given explicitly on the
7674 command line, then the default auto-export behavior will be @emph{disabled}
7675 if either of the following are true:
7678 @item A DEF file is used.
7679 @item Any symbol in any object file was marked with the __declspec(dllexport) attribute.
7682 @item using a DEF file
7683 @cindex using a DEF file
7684 Another way of exporting symbols is using a DEF file. A DEF file is
7685 an ASCII file containing definitions of symbols which should be
7686 exported when a dll is created. Usually it is named @samp{<dll
7687 name>.def} and is added as any other object file to the linker's
7688 command line. The file's name must end in @samp{.def} or @samp{.DEF}.
7691 gcc -o <output> <objectfiles> <dll name>.def
7694 Using a DEF file turns off the normal auto-export behavior, unless the
7695 @samp{--export-all-symbols} option is also used.
7697 Here is an example of a DEF file for a shared library called @samp{xyz.dll}:
7700 LIBRARY "xyz.dll" BASE=0x20000000
7706 another_foo = abc.dll.afoo
7712 This example defines a DLL with a non-default base address and seven
7713 symbols in the export table. The third exported symbol @code{_bar} is an
7714 alias for the second. The fourth symbol, @code{another_foo} is resolved
7715 by "forwarding" to another module and treating it as an alias for
7716 @code{afoo} exported from the DLL @samp{abc.dll}. The final symbol
7717 @code{var1} is declared to be a data object. The @samp{doo} symbol in
7718 export library is an alias of @samp{foo}, which gets the string name
7719 in export table @samp{foo2}. The @samp{eoo} symbol is an data export
7720 symbol, which gets in export table the name @samp{var1}.
7722 The optional @code{LIBRARY <name>} command indicates the @emph{internal}
7723 name of the output DLL. If @samp{<name>} does not include a suffix,
7724 the default library suffix, @samp{.DLL} is appended.
7726 When the .DEF file is used to build an application, rather than a
7727 library, the @code{NAME <name>} command should be used instead of
7728 @code{LIBRARY}. If @samp{<name>} does not include a suffix, the default
7729 executable suffix, @samp{.EXE} is appended.
7731 With either @code{LIBRARY <name>} or @code{NAME <name>} the optional
7732 specification @code{BASE = <number>} may be used to specify a
7733 non-default base address for the image.
7735 If neither @code{LIBRARY <name>} nor @code{NAME <name>} is specified,
7736 or they specify an empty string, the internal name is the same as the
7737 filename specified on the command line.
7739 The complete specification of an export symbol is:
7743 ( ( ( <name1> [ = <name2> ] )
7744 | ( <name1> = <module-name> . <external-name>))
7745 [ @@ <integer> ] [NONAME] [DATA] [CONSTANT] [PRIVATE] [== <name3>] ) *
7748 Declares @samp{<name1>} as an exported symbol from the DLL, or declares
7749 @samp{<name1>} as an exported alias for @samp{<name2>}; or declares
7750 @samp{<name1>} as a "forward" alias for the symbol
7751 @samp{<external-name>} in the DLL @samp{<module-name>}.
7752 Optionally, the symbol may be exported by the specified ordinal
7753 @samp{<integer>} alias. The optional @samp{<name3>} is the to be used
7754 string in import/export table for the symbol.
7756 The optional keywords that follow the declaration indicate:
7758 @code{NONAME}: Do not put the symbol name in the DLL's export table. It
7759 will still be exported by its ordinal alias (either the value specified
7760 by the .def specification or, otherwise, the value assigned by the
7761 linker). The symbol name, however, does remain visible in the import
7762 library (if any), unless @code{PRIVATE} is also specified.
7764 @code{DATA}: The symbol is a variable or object, rather than a function.
7765 The import lib will export only an indirect reference to @code{foo} as
7766 the symbol @code{_imp__foo} (ie, @code{foo} must be resolved as
7769 @code{CONSTANT}: Like @code{DATA}, but put the undecorated @code{foo} as
7770 well as @code{_imp__foo} into the import library. Both refer to the
7771 read-only import address table's pointer to the variable, not to the
7772 variable itself. This can be dangerous. If the user code fails to add
7773 the @code{dllimport} attribute and also fails to explicitly add the
7774 extra indirection that the use of the attribute enforces, the
7775 application will behave unexpectedly.
7777 @code{PRIVATE}: Put the symbol in the DLL's export table, but do not put
7778 it into the static import library used to resolve imports at link time. The
7779 symbol can still be imported using the @code{LoadLibrary/GetProcAddress}
7780 API at runtime or by by using the GNU ld extension of linking directly to
7781 the DLL without an import library.
7783 See ld/deffilep.y in the binutils sources for the full specification of
7784 other DEF file statements
7786 @cindex creating a DEF file
7787 While linking a shared dll, @command{ld} is able to create a DEF file
7788 with the @samp{--output-def <file>} command line option.
7790 @item Using decorations
7791 @cindex Using decorations
7792 Another way of marking symbols for export is to modify the source code
7793 itself, so that when building the DLL each symbol to be exported is
7797 __declspec(dllexport) int a_variable
7798 __declspec(dllexport) void a_function(int with_args)
7801 All such symbols will be exported from the DLL. If, however,
7802 any of the object files in the DLL contain symbols decorated in
7803 this way, then the normal auto-export behavior is disabled, unless
7804 the @samp{--export-all-symbols} option is also used.
7806 Note that object files that wish to access these symbols must @emph{not}
7807 decorate them with dllexport. Instead, they should use dllimport,
7811 __declspec(dllimport) int a_variable
7812 __declspec(dllimport) void a_function(int with_args)
7815 This complicates the structure of library header files, because
7816 when included by the library itself the header must declare the
7817 variables and functions as dllexport, but when included by client
7818 code the header must declare them as dllimport. There are a number
7819 of idioms that are typically used to do this; often client code can
7820 omit the __declspec() declaration completely. See
7821 @samp{--enable-auto-import} and @samp{automatic data imports} for more
7825 @cindex automatic data imports
7826 @item automatic data imports
7827 The standard Windows dll format supports data imports from dlls only
7828 by adding special decorations (dllimport/dllexport), which let the
7829 compiler produce specific assembler instructions to deal with this
7830 issue. This increases the effort necessary to port existing Un*x
7831 code to these platforms, especially for large
7832 c++ libraries and applications. The auto-import feature, which was
7833 initially provided by Paul Sokolovsky, allows one to omit the
7834 decorations to achieve a behavior that conforms to that on POSIX/Un*x
7835 platforms. This feature is enabled with the @samp{--enable-auto-import}
7836 command-line option, although it is enabled by default on cygwin/mingw.
7837 The @samp{--enable-auto-import} option itself now serves mainly to
7838 suppress any warnings that are ordinarily emitted when linked objects
7839 trigger the feature's use.
7841 auto-import of variables does not always work flawlessly without
7842 additional assistance. Sometimes, you will see this message
7844 "variable '<var>' can't be auto-imported. Please read the
7845 documentation for ld's @code{--enable-auto-import} for details."
7847 The @samp{--enable-auto-import} documentation explains why this error
7848 occurs, and several methods that can be used to overcome this difficulty.
7849 One of these methods is the @emph{runtime pseudo-relocs} feature, described
7852 @cindex runtime pseudo-relocation
7853 For complex variables imported from DLLs (such as structs or classes),
7854 object files typically contain a base address for the variable and an
7855 offset (@emph{addend}) within the variable--to specify a particular
7856 field or public member, for instance. Unfortunately, the runtime loader used
7857 in win32 environments is incapable of fixing these references at runtime
7858 without the additional information supplied by dllimport/dllexport decorations.
7859 The standard auto-import feature described above is unable to resolve these
7862 The @samp{--enable-runtime-pseudo-relocs} switch allows these references to
7863 be resolved without error, while leaving the task of adjusting the references
7864 themselves (with their non-zero addends) to specialized code provided by the
7865 runtime environment. Recent versions of the cygwin and mingw environments and
7866 compilers provide this runtime support; older versions do not. However, the
7867 support is only necessary on the developer's platform; the compiled result will
7868 run without error on an older system.
7870 @samp{--enable-runtime-pseudo-relocs} is not the default; it must be explicitly
7873 @cindex direct linking to a dll
7874 @item direct linking to a dll
7875 The cygwin/mingw ports of @command{ld} support the direct linking,
7876 including data symbols, to a dll without the usage of any import
7877 libraries. This is much faster and uses much less memory than does the
7878 traditional import library method, especially when linking large
7879 libraries or applications. When @command{ld} creates an import lib, each
7880 function or variable exported from the dll is stored in its own bfd, even
7881 though a single bfd could contain many exports. The overhead involved in
7882 storing, loading, and processing so many bfd's is quite large, and explains the
7883 tremendous time, memory, and storage needed to link against particularly
7884 large or complex libraries when using import libs.
7886 Linking directly to a dll uses no extra command-line switches other than
7887 @samp{-L} and @samp{-l}, because @command{ld} already searches for a number
7888 of names to match each library. All that is needed from the developer's
7889 perspective is an understanding of this search, in order to force ld to
7890 select the dll instead of an import library.
7893 For instance, when ld is called with the argument @samp{-lxxx} it will attempt
7894 to find, in the first directory of its search path,
7906 before moving on to the next directory in the search path.
7908 (*) Actually, this is not @samp{cygxxx.dll} but in fact is @samp{<prefix>xxx.dll},
7909 where @samp{<prefix>} is set by the @command{ld} option
7910 @samp{--dll-search-prefix=<prefix>}. In the case of cygwin, the standard gcc spec
7911 file includes @samp{--dll-search-prefix=cyg}, so in effect we actually search for
7914 Other win32-based unix environments, such as mingw or pw32, may use other
7915 @samp{<prefix>}es, although at present only cygwin makes use of this feature. It
7916 was originally intended to help avoid name conflicts among dll's built for the
7917 various win32/un*x environments, so that (for example) two versions of a zlib dll
7918 could coexist on the same machine.
7920 The generic cygwin/mingw path layout uses a @samp{bin} directory for
7921 applications and dll's and a @samp{lib} directory for the import
7922 libraries (using cygwin nomenclature):
7928 libxxx.dll.a (in case of dll's)
7929 libxxx.a (in case of static archive)
7932 Linking directly to a dll without using the import library can be
7935 1. Use the dll directly by adding the @samp{bin} path to the link line
7937 gcc -Wl,-verbose -o a.exe -L../bin/ -lxxx
7940 However, as the dll's often have version numbers appended to their names
7941 (@samp{cygncurses-5.dll}) this will often fail, unless one specifies
7942 @samp{-L../bin -lncurses-5} to include the version. Import libs are generally
7943 not versioned, and do not have this difficulty.
7945 2. Create a symbolic link from the dll to a file in the @samp{lib}
7946 directory according to the above mentioned search pattern. This
7947 should be used to avoid unwanted changes in the tools needed for
7951 ln -s bin/cygxxx.dll lib/[cyg|lib|]xxx.dll[.a]
7954 Then you can link without any make environment changes.
7957 gcc -Wl,-verbose -o a.exe -L../lib/ -lxxx
7960 This technique also avoids the version number problems, because the following is
7967 libxxx.dll.a -> ../bin/cygxxx-5.dll
7970 Linking directly to a dll without using an import lib will work
7971 even when auto-import features are exercised, and even when
7972 @samp{--enable-runtime-pseudo-relocs} is used.
7974 Given the improvements in speed and memory usage, one might justifiably
7975 wonder why import libraries are used at all. There are three reasons:
7977 1. Until recently, the link-directly-to-dll functionality did @emph{not}
7978 work with auto-imported data.
7980 2. Sometimes it is necessary to include pure static objects within the
7981 import library (which otherwise contains only bfd's for indirection
7982 symbols that point to the exports of a dll). Again, the import lib
7983 for the cygwin kernel makes use of this ability, and it is not
7984 possible to do this without an import lib.
7986 3. Symbol aliases can only be resolved using an import lib. This is
7987 critical when linking against OS-supplied dll's (eg, the win32 API)
7988 in which symbols are usually exported as undecorated aliases of their
7989 stdcall-decorated assembly names.
7991 So, import libs are not going away. But the ability to replace
7992 true import libs with a simple symbolic link to (or a copy of)
7993 a dll, in many cases, is a useful addition to the suite of tools
7994 binutils makes available to the win32 developer. Given the
7995 massive improvements in memory requirements during linking, storage
7996 requirements, and linking speed, we expect that many developers
7997 will soon begin to use this feature whenever possible.
7999 @item symbol aliasing
8001 @item adding additional names
8002 Sometimes, it is useful to export symbols with additional names.
8003 A symbol @samp{foo} will be exported as @samp{foo}, but it can also be
8004 exported as @samp{_foo} by using special directives in the DEF file
8005 when creating the dll. This will affect also the optional created
8006 import library. Consider the following DEF file:
8009 LIBRARY "xyz.dll" BASE=0x61000000
8016 The line @samp{_foo = foo} maps the symbol @samp{foo} to @samp{_foo}.
8018 Another method for creating a symbol alias is to create it in the
8019 source code using the "weak" attribute:
8022 void foo () @{ /* Do something. */; @}
8023 void _foo () __attribute__ ((weak, alias ("foo")));
8026 See the gcc manual for more information about attributes and weak
8029 @item renaming symbols
8030 Sometimes it is useful to rename exports. For instance, the cygwin
8031 kernel does this regularly. A symbol @samp{_foo} can be exported as
8032 @samp{foo} but not as @samp{_foo} by using special directives in the
8033 DEF file. (This will also affect the import library, if it is
8034 created). In the following example:
8037 LIBRARY "xyz.dll" BASE=0x61000000
8043 The line @samp{_foo = foo} maps the exported symbol @samp{foo} to
8047 Note: using a DEF file disables the default auto-export behavior,
8048 unless the @samp{--export-all-symbols} command line option is used.
8049 If, however, you are trying to rename symbols, then you should list
8050 @emph{all} desired exports in the DEF file, including the symbols
8051 that are not being renamed, and do @emph{not} use the
8052 @samp{--export-all-symbols} option. If you list only the
8053 renamed symbols in the DEF file, and use @samp{--export-all-symbols}
8054 to handle the other symbols, then the both the new names @emph{and}
8055 the original names for the renamed symbols will be exported.
8056 In effect, you'd be aliasing those symbols, not renaming them,
8057 which is probably not what you wanted.
8059 @cindex weak externals
8060 @item weak externals
8061 The Windows object format, PE, specifies a form of weak symbols called
8062 weak externals. When a weak symbol is linked and the symbol is not
8063 defined, the weak symbol becomes an alias for some other symbol. There
8064 are three variants of weak externals:
8066 @item Definition is searched for in objects and libraries, historically
8067 called lazy externals.
8068 @item Definition is searched for only in other objects, not in libraries.
8069 This form is not presently implemented.
8070 @item No search; the symbol is an alias. This form is not presently
8073 As a GNU extension, weak symbols that do not specify an alternate symbol
8074 are supported. If the symbol is undefined when linking, the symbol
8075 uses a default value.
8077 @cindex aligned common symbols
8078 @item aligned common symbols
8079 As a GNU extension to the PE file format, it is possible to specify the
8080 desired alignment for a common symbol. This information is conveyed from
8081 the assembler or compiler to the linker by means of GNU-specific commands
8082 carried in the object file's @samp{.drectve} section, which are recognized
8083 by @command{ld} and respected when laying out the common symbols. Native
8084 tools will be able to process object files employing this GNU extension,
8085 but will fail to respect the alignment instructions, and may issue noisy
8086 warnings about unknown linker directives.
8101 @section @code{ld} and Xtensa Processors
8103 @cindex Xtensa processors
8104 The default @command{ld} behavior for Xtensa processors is to interpret
8105 @code{SECTIONS} commands so that lists of explicitly named sections in a
8106 specification with a wildcard file will be interleaved when necessary to
8107 keep literal pools within the range of PC-relative load offsets. For
8108 example, with the command:
8120 @command{ld} may interleave some of the @code{.literal}
8121 and @code{.text} sections from different object files to ensure that the
8122 literal pools are within the range of PC-relative load offsets. A valid
8123 interleaving might place the @code{.literal} sections from an initial
8124 group of files followed by the @code{.text} sections of that group of
8125 files. Then, the @code{.literal} sections from the rest of the files
8126 and the @code{.text} sections from the rest of the files would follow.
8128 @cindex @option{--relax} on Xtensa
8129 @cindex relaxing on Xtensa
8130 Relaxation is enabled by default for the Xtensa version of @command{ld} and
8131 provides two important link-time optimizations. The first optimization
8132 is to combine identical literal values to reduce code size. A redundant
8133 literal will be removed and all the @code{L32R} instructions that use it
8134 will be changed to reference an identical literal, as long as the
8135 location of the replacement literal is within the offset range of all
8136 the @code{L32R} instructions. The second optimization is to remove
8137 unnecessary overhead from assembler-generated ``longcall'' sequences of
8138 @code{L32R}/@code{CALLX@var{n}} when the target functions are within
8139 range of direct @code{CALL@var{n}} instructions.
8141 For each of these cases where an indirect call sequence can be optimized
8142 to a direct call, the linker will change the @code{CALLX@var{n}}
8143 instruction to a @code{CALL@var{n}} instruction, remove the @code{L32R}
8144 instruction, and remove the literal referenced by the @code{L32R}
8145 instruction if it is not used for anything else. Removing the
8146 @code{L32R} instruction always reduces code size but can potentially
8147 hurt performance by changing the alignment of subsequent branch targets.
8148 By default, the linker will always preserve alignments, either by
8149 switching some instructions between 24-bit encodings and the equivalent
8150 density instructions or by inserting a no-op in place of the @code{L32R}
8151 instruction that was removed. If code size is more important than
8152 performance, the @option{--size-opt} option can be used to prevent the
8153 linker from widening density instructions or inserting no-ops, except in
8154 a few cases where no-ops are required for correctness.
8156 The following Xtensa-specific command-line options can be used to
8159 @cindex Xtensa options
8162 When optimizing indirect calls to direct calls, optimize for code size
8163 more than performance. With this option, the linker will not insert
8164 no-ops or widen density instructions to preserve branch target
8165 alignment. There may still be some cases where no-ops are required to
8166 preserve the correctness of the code.
8174 @ifclear SingleFormat
8179 @cindex object file management
8180 @cindex object formats available
8182 The linker accesses object and archive files using the BFD libraries.
8183 These libraries allow the linker to use the same routines to operate on
8184 object files whatever the object file format. A different object file
8185 format can be supported simply by creating a new BFD back end and adding
8186 it to the library. To conserve runtime memory, however, the linker and
8187 associated tools are usually configured to support only a subset of the
8188 object file formats available. You can use @code{objdump -i}
8189 (@pxref{objdump,,objdump,binutils.info,The GNU Binary Utilities}) to
8190 list all the formats available for your configuration.
8192 @cindex BFD requirements
8193 @cindex requirements for BFD
8194 As with most implementations, BFD is a compromise between
8195 several conflicting requirements. The major factor influencing
8196 BFD design was efficiency: any time used converting between
8197 formats is time which would not have been spent had BFD not
8198 been involved. This is partly offset by abstraction payback; since
8199 BFD simplifies applications and back ends, more time and care
8200 may be spent optimizing algorithms for a greater speed.
8202 One minor artifact of the BFD solution which you should bear in
8203 mind is the potential for information loss. There are two places where
8204 useful information can be lost using the BFD mechanism: during
8205 conversion and during output. @xref{BFD information loss}.
8208 * BFD outline:: How it works: an outline of BFD
8212 @section How It Works: An Outline of BFD
8213 @cindex opening object files
8214 @include bfdsumm.texi
8217 @node Reporting Bugs
8218 @chapter Reporting Bugs
8219 @cindex bugs in @command{ld}
8220 @cindex reporting bugs in @command{ld}
8222 Your bug reports play an essential role in making @command{ld} reliable.
8224 Reporting a bug may help you by bringing a solution to your problem, or
8225 it may not. But in any case the principal function of a bug report is
8226 to help the entire community by making the next version of @command{ld}
8227 work better. Bug reports are your contribution to the maintenance of
8230 In order for a bug report to serve its purpose, you must include the
8231 information that enables us to fix the bug.
8234 * Bug Criteria:: Have you found a bug?
8235 * Bug Reporting:: How to report bugs
8239 @section Have You Found a Bug?
8240 @cindex bug criteria
8242 If you are not sure whether you have found a bug, here are some guidelines:
8245 @cindex fatal signal
8246 @cindex linker crash
8247 @cindex crash of linker
8249 If the linker gets a fatal signal, for any input whatever, that is a
8250 @command{ld} bug. Reliable linkers never crash.
8252 @cindex error on valid input
8254 If @command{ld} produces an error message for valid input, that is a bug.
8256 @cindex invalid input
8258 If @command{ld} does not produce an error message for invalid input, that
8259 may be a bug. In the general case, the linker can not verify that
8260 object files are correct.
8263 If you are an experienced user of linkers, your suggestions for
8264 improvement of @command{ld} are welcome in any case.
8268 @section How to Report Bugs
8270 @cindex @command{ld} bugs, reporting
8272 A number of companies and individuals offer support for @sc{gnu}
8273 products. If you obtained @command{ld} from a support organization, we
8274 recommend you contact that organization first.
8276 You can find contact information for many support companies and
8277 individuals in the file @file{etc/SERVICE} in the @sc{gnu} Emacs
8281 Otherwise, send bug reports for @command{ld} to
8285 The fundamental principle of reporting bugs usefully is this:
8286 @strong{report all the facts}. If you are not sure whether to state a
8287 fact or leave it out, state it!
8289 Often people omit facts because they think they know what causes the
8290 problem and assume that some details do not matter. Thus, you might
8291 assume that the name of a symbol you use in an example does not
8292 matter. Well, probably it does not, but one cannot be sure. Perhaps
8293 the bug is a stray memory reference which happens to fetch from the
8294 location where that name is stored in memory; perhaps, if the name
8295 were different, the contents of that location would fool the linker
8296 into doing the right thing despite the bug. Play it safe and give a
8297 specific, complete example. That is the easiest thing for you to do,
8298 and the most helpful.
8300 Keep in mind that the purpose of a bug report is to enable us to fix
8301 the bug if it is new to us. Therefore, always write your bug reports
8302 on the assumption that the bug has not been reported previously.
8304 Sometimes people give a few sketchy facts and ask, ``Does this ring a
8305 bell?'' This cannot help us fix a bug, so it is basically useless. We
8306 respond by asking for enough details to enable us to investigate.
8307 You might as well expedite matters by sending them to begin with.
8309 To enable us to fix the bug, you should include all these things:
8313 The version of @command{ld}. @command{ld} announces it if you start it with
8314 the @samp{--version} argument.
8316 Without this, we will not know whether there is any point in looking for
8317 the bug in the current version of @command{ld}.
8320 Any patches you may have applied to the @command{ld} source, including any
8321 patches made to the @code{BFD} library.
8324 The type of machine you are using, and the operating system name and
8328 What compiler (and its version) was used to compile @command{ld}---e.g.
8332 The command arguments you gave the linker to link your example and
8333 observe the bug. To guarantee you will not omit something important,
8334 list them all. A copy of the Makefile (or the output from make) is
8337 If we were to try to guess the arguments, we would probably guess wrong
8338 and then we might not encounter the bug.
8341 A complete input file, or set of input files, that will reproduce the
8342 bug. It is generally most helpful to send the actual object files
8343 provided that they are reasonably small. Say no more than 10K. For
8344 bigger files you can either make them available by FTP or HTTP or else
8345 state that you are willing to send the object file(s) to whomever
8346 requests them. (Note - your email will be going to a mailing list, so
8347 we do not want to clog it up with large attachments). But small
8348 attachments are best.
8350 If the source files were assembled using @code{gas} or compiled using
8351 @code{gcc}, then it may be OK to send the source files rather than the
8352 object files. In this case, be sure to say exactly what version of
8353 @code{gas} or @code{gcc} was used to produce the object files. Also say
8354 how @code{gas} or @code{gcc} were configured.
8357 A description of what behavior you observe that you believe is
8358 incorrect. For example, ``It gets a fatal signal.''
8360 Of course, if the bug is that @command{ld} gets a fatal signal, then we
8361 will certainly notice it. But if the bug is incorrect output, we might
8362 not notice unless it is glaringly wrong. You might as well not give us
8363 a chance to make a mistake.
8365 Even if the problem you experience is a fatal signal, you should still
8366 say so explicitly. Suppose something strange is going on, such as, your
8367 copy of @command{ld} is out of sync, or you have encountered a bug in the
8368 C library on your system. (This has happened!) Your copy might crash
8369 and ours would not. If you told us to expect a crash, then when ours
8370 fails to crash, we would know that the bug was not happening for us. If
8371 you had not told us to expect a crash, then we would not be able to draw
8372 any conclusion from our observations.
8375 If you wish to suggest changes to the @command{ld} source, send us context
8376 diffs, as generated by @code{diff} with the @samp{-u}, @samp{-c}, or
8377 @samp{-p} option. Always send diffs from the old file to the new file.
8378 If you even discuss something in the @command{ld} source, refer to it by
8379 context, not by line number.
8381 The line numbers in our development sources will not match those in your
8382 sources. Your line numbers would convey no useful information to us.
8385 Here are some things that are not necessary:
8389 A description of the envelope of the bug.
8391 Often people who encounter a bug spend a lot of time investigating
8392 which changes to the input file will make the bug go away and which
8393 changes will not affect it.
8395 This is often time consuming and not very useful, because the way we
8396 will find the bug is by running a single example under the debugger
8397 with breakpoints, not by pure deduction from a series of examples.
8398 We recommend that you save your time for something else.
8400 Of course, if you can find a simpler example to report @emph{instead}
8401 of the original one, that is a convenience for us. Errors in the
8402 output will be easier to spot, running under the debugger will take
8403 less time, and so on.
8405 However, simplification is not vital; if you do not want to do this,
8406 report the bug anyway and send us the entire test case you used.
8409 A patch for the bug.
8411 A patch for the bug does help us if it is a good one. But do not omit
8412 the necessary information, such as the test case, on the assumption that
8413 a patch is all we need. We might see problems with your patch and decide
8414 to fix the problem another way, or we might not understand it at all.
8416 Sometimes with a program as complicated as @command{ld} it is very hard to
8417 construct an example that will make the program follow a certain path
8418 through the code. If you do not send us the example, we will not be
8419 able to construct one, so we will not be able to verify that the bug is
8422 And if we cannot understand what bug you are trying to fix, or why your
8423 patch should be an improvement, we will not install it. A test case will
8424 help us to understand.
8427 A guess about what the bug is or what it depends on.
8429 Such guesses are usually wrong. Even we cannot guess right about such
8430 things without first using the debugger to find the facts.
8434 @appendix MRI Compatible Script Files
8435 @cindex MRI compatibility
8436 To aid users making the transition to @sc{gnu} @command{ld} from the MRI
8437 linker, @command{ld} can use MRI compatible linker scripts as an
8438 alternative to the more general-purpose linker scripting language
8439 described in @ref{Scripts}. MRI compatible linker scripts have a much
8440 simpler command set than the scripting language otherwise used with
8441 @command{ld}. @sc{gnu} @command{ld} supports the most commonly used MRI
8442 linker commands; these commands are described here.
8444 In general, MRI scripts aren't of much use with the @code{a.out} object
8445 file format, since it only has three sections and MRI scripts lack some
8446 features to make use of them.
8448 You can specify a file containing an MRI-compatible script using the
8449 @samp{-c} command-line option.
8451 Each command in an MRI-compatible script occupies its own line; each
8452 command line starts with the keyword that identifies the command (though
8453 blank lines are also allowed for punctuation). If a line of an
8454 MRI-compatible script begins with an unrecognized keyword, @command{ld}
8455 issues a warning message, but continues processing the script.
8457 Lines beginning with @samp{*} are comments.
8459 You can write these commands using all upper-case letters, or all
8460 lower case; for example, @samp{chip} is the same as @samp{CHIP}.
8461 The following list shows only the upper-case form of each command.
8464 @cindex @code{ABSOLUTE} (MRI)
8465 @item ABSOLUTE @var{secname}
8466 @itemx ABSOLUTE @var{secname}, @var{secname}, @dots{} @var{secname}
8467 Normally, @command{ld} includes in the output file all sections from all
8468 the input files. However, in an MRI-compatible script, you can use the
8469 @code{ABSOLUTE} command to restrict the sections that will be present in
8470 your output program. If the @code{ABSOLUTE} command is used at all in a
8471 script, then only the sections named explicitly in @code{ABSOLUTE}
8472 commands will appear in the linker output. You can still use other
8473 input sections (whatever you select on the command line, or using
8474 @code{LOAD}) to resolve addresses in the output file.
8476 @cindex @code{ALIAS} (MRI)
8477 @item ALIAS @var{out-secname}, @var{in-secname}
8478 Use this command to place the data from input section @var{in-secname}
8479 in a section called @var{out-secname} in the linker output file.
8481 @var{in-secname} may be an integer.
8483 @cindex @code{ALIGN} (MRI)
8484 @item ALIGN @var{secname} = @var{expression}
8485 Align the section called @var{secname} to @var{expression}. The
8486 @var{expression} should be a power of two.
8488 @cindex @code{BASE} (MRI)
8489 @item BASE @var{expression}
8490 Use the value of @var{expression} as the lowest address (other than
8491 absolute addresses) in the output file.
8493 @cindex @code{CHIP} (MRI)
8494 @item CHIP @var{expression}
8495 @itemx CHIP @var{expression}, @var{expression}
8496 This command does nothing; it is accepted only for compatibility.
8498 @cindex @code{END} (MRI)
8500 This command does nothing whatever; it's only accepted for compatibility.
8502 @cindex @code{FORMAT} (MRI)
8503 @item FORMAT @var{output-format}
8504 Similar to the @code{OUTPUT_FORMAT} command in the more general linker
8505 language, but restricted to one of these output formats:
8509 S-records, if @var{output-format} is @samp{S}
8512 IEEE, if @var{output-format} is @samp{IEEE}
8515 COFF (the @samp{coff-m68k} variant in BFD), if @var{output-format} is
8519 @cindex @code{LIST} (MRI)
8520 @item LIST @var{anything}@dots{}
8521 Print (to the standard output file) a link map, as produced by the
8522 @command{ld} command-line option @samp{-M}.
8524 The keyword @code{LIST} may be followed by anything on the
8525 same line, with no change in its effect.
8527 @cindex @code{LOAD} (MRI)
8528 @item LOAD @var{filename}
8529 @itemx LOAD @var{filename}, @var{filename}, @dots{} @var{filename}
8530 Include one or more object file @var{filename} in the link; this has the
8531 same effect as specifying @var{filename} directly on the @command{ld}
8534 @cindex @code{NAME} (MRI)
8535 @item NAME @var{output-name}
8536 @var{output-name} is the name for the program produced by @command{ld}; the
8537 MRI-compatible command @code{NAME} is equivalent to the command-line
8538 option @samp{-o} or the general script language command @code{OUTPUT}.
8540 @cindex @code{ORDER} (MRI)
8541 @item ORDER @var{secname}, @var{secname}, @dots{} @var{secname}
8542 @itemx ORDER @var{secname} @var{secname} @var{secname}
8543 Normally, @command{ld} orders the sections in its output file in the
8544 order in which they first appear in the input files. In an MRI-compatible
8545 script, you can override this ordering with the @code{ORDER} command. The
8546 sections you list with @code{ORDER} will appear first in your output
8547 file, in the order specified.
8549 @cindex @code{PUBLIC} (MRI)
8550 @item PUBLIC @var{name}=@var{expression}
8551 @itemx PUBLIC @var{name},@var{expression}
8552 @itemx PUBLIC @var{name} @var{expression}
8553 Supply a value (@var{expression}) for external symbol
8554 @var{name} used in the linker input files.
8556 @cindex @code{SECT} (MRI)
8557 @item SECT @var{secname}, @var{expression}
8558 @itemx SECT @var{secname}=@var{expression}
8559 @itemx SECT @var{secname} @var{expression}
8560 You can use any of these three forms of the @code{SECT} command to
8561 specify the start address (@var{expression}) for section @var{secname}.
8562 If you have more than one @code{SECT} statement for the same
8563 @var{secname}, only the @emph{first} sets the start address.
8566 @node GNU Free Documentation License
8567 @appendix GNU Free Documentation License
8571 @unnumbered LD Index
8576 % I think something like @@colophon should be in texinfo. In the
8578 \long\def\colophon{\hbox to0pt{}\vfill
8579 \centerline{The body of this manual is set in}
8580 \centerline{\fontname\tenrm,}
8581 \centerline{with headings in {\bf\fontname\tenbf}}
8582 \centerline{and examples in {\tt\fontname\tentt}.}
8583 \centerline{{\it\fontname\tenit\/} and}
8584 \centerline{{\sl\fontname\tensl\/}}
8585 \centerline{are used for emphasis.}\vfill}
8587 % Blame: doc@@cygnus.com, 28mar91.