3 @c Copyright (C) 1991-2015 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{} 1991-2015 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{} 1991-2015 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.
832 @cindex push state governing input file handling
834 The @option{--push-state} allows to preserve the current state of the
835 flags which govern the input file handling so that they can all be
836 restored with one corresponding @option{--pop-state} option.
838 The option which are covered are: @option{-Bdynamic}, @option{-Bstatic},
839 @option{-dn}, @option{-dy}, @option{-call_shared}, @option{-non_shared},
840 @option{-static}, @option{-N}, @option{-n}, @option{--whole-archive},
841 @option{--no-whole-archive}, @option{-r}, @option{-Ur},
842 @option{--copy-dt-needed-entries}, @option{--no-copy-dt-needed-entries},
843 @option{--as-needed}, @option{--no-as-needed}, and @option{-a}.
845 One target for this option are specifications for @file{pkg-config}. When
846 used with the @option{--libs} option all possibly needed libraries are
847 listed and then possibly linked with all the time. It is better to return
848 something as follows:
851 -Wl,--push-state,--as-needed -libone -libtwo -Wl,--pop-state
855 @cindex pop state governing input file handling
856 Undoes the effect of --push-state, restores the previous values of the
857 flags governing input file handling.
860 @kindex --emit-relocs
861 @cindex retain relocations in final executable
864 Leave relocation sections and contents in fully linked executables.
865 Post link analysis and optimization tools may need this information in
866 order to perform correct modifications of executables. This results
867 in larger executables.
869 This option is currently only supported on ELF platforms.
871 @kindex --force-dynamic
872 @cindex forcing the creation of dynamic sections
873 @item --force-dynamic
874 Force the output file to have dynamic sections. This option is specific
878 @cindex relocatable output
880 @kindex --relocatable
883 Generate relocatable output---i.e., generate an output file that can in
884 turn serve as input to @command{ld}. This is often called @dfn{partial
885 linking}. As a side effect, in environments that support standard Unix
886 magic numbers, this option also sets the output file's magic number to
888 @c ; see @option{-N}.
889 If this option is not specified, an absolute file is produced. When
890 linking C++ programs, this option @emph{will not} resolve references to
891 constructors; to do that, use @samp{-Ur}.
893 When an input file does not have the same format as the output file,
894 partial linking is only supported if that input file does not contain any
895 relocations. Different output formats can have further restrictions; for
896 example some @code{a.out}-based formats do not support partial linking
897 with input files in other formats at all.
899 This option does the same thing as @samp{-i}.
901 @kindex -R @var{file}
902 @kindex --just-symbols=@var{file}
903 @cindex symbol-only input
904 @item -R @var{filename}
905 @itemx --just-symbols=@var{filename}
906 Read symbol names and their addresses from @var{filename}, but do not
907 relocate it or include it in the output. This allows your output file
908 to refer symbolically to absolute locations of memory defined in other
909 programs. You may use this option more than once.
911 For compatibility with other ELF linkers, if the @option{-R} option is
912 followed by a directory name, rather than a file name, it is treated as
913 the @option{-rpath} option.
917 @cindex strip all symbols
920 Omit all symbol information from the output file.
923 @kindex --strip-debug
924 @cindex strip debugger symbols
927 Omit debugger symbol information (but not all symbols) from the output file.
931 @cindex input files, displaying
934 Print the names of the input files as @command{ld} processes them.
936 @kindex -T @var{script}
937 @kindex --script=@var{script}
939 @item -T @var{scriptfile}
940 @itemx --script=@var{scriptfile}
941 Use @var{scriptfile} as the linker script. This script replaces
942 @command{ld}'s default linker script (rather than adding to it), so
943 @var{commandfile} must specify everything necessary to describe the
944 output file. @xref{Scripts}. If @var{scriptfile} does not exist in
945 the current directory, @code{ld} looks for it in the directories
946 specified by any preceding @samp{-L} options. Multiple @samp{-T}
949 @kindex -dT @var{script}
950 @kindex --default-script=@var{script}
952 @item -dT @var{scriptfile}
953 @itemx --default-script=@var{scriptfile}
954 Use @var{scriptfile} as the default linker script. @xref{Scripts}.
956 This option is similar to the @option{--script} option except that
957 processing of the script is delayed until after the rest of the
958 command line has been processed. This allows options placed after the
959 @option{--default-script} option on the command line to affect the
960 behaviour of the linker script, which can be important when the linker
961 command line cannot be directly controlled by the user. (eg because
962 the command line is being constructed by another tool, such as
965 @kindex -u @var{symbol}
966 @kindex --undefined=@var{symbol}
967 @cindex undefined symbol
968 @item -u @var{symbol}
969 @itemx --undefined=@var{symbol}
970 Force @var{symbol} to be entered in the output file as an undefined
971 symbol. Doing this may, for example, trigger linking of additional
972 modules from standard libraries. @samp{-u} may be repeated with
973 different option arguments to enter additional undefined symbols. This
974 option is equivalent to the @code{EXTERN} linker script command.
979 For anything other than C++ programs, this option is equivalent to
980 @samp{-r}: it generates relocatable output---i.e., an output file that can in
981 turn serve as input to @command{ld}. When linking C++ programs, @samp{-Ur}
982 @emph{does} resolve references to constructors, unlike @samp{-r}.
983 It does not work to use @samp{-Ur} on files that were themselves linked
984 with @samp{-Ur}; once the constructor table has been built, it cannot
985 be added to. Use @samp{-Ur} only for the last partial link, and
986 @samp{-r} for the others.
988 @kindex --unique[=@var{SECTION}]
989 @item --unique[=@var{SECTION}]
990 Creates a separate output section for every input section matching
991 @var{SECTION}, or if the optional wildcard @var{SECTION} argument is
992 missing, for every orphan input section. An orphan section is one not
993 specifically mentioned in a linker script. You may use this option
994 multiple times on the command line; It prevents the normal merging of
995 input sections with the same name, overriding output section assignments
1005 Display the version number for @command{ld}. The @option{-V} option also
1006 lists the supported emulations.
1009 @kindex --discard-all
1010 @cindex deleting local symbols
1012 @itemx --discard-all
1013 Delete all local symbols.
1016 @kindex --discard-locals
1017 @cindex local symbols, deleting
1019 @itemx --discard-locals
1020 Delete all temporary local symbols. (These symbols start with
1021 system-specific local label prefixes, typically @samp{.L} for ELF systems
1022 or @samp{L} for traditional a.out systems.)
1024 @kindex -y @var{symbol}
1025 @kindex --trace-symbol=@var{symbol}
1026 @cindex symbol tracing
1027 @item -y @var{symbol}
1028 @itemx --trace-symbol=@var{symbol}
1029 Print the name of each linked file in which @var{symbol} appears. This
1030 option may be given any number of times. On many systems it is necessary
1031 to prepend an underscore.
1033 This option is useful when you have an undefined symbol in your link but
1034 don't know where the reference is coming from.
1036 @kindex -Y @var{path}
1038 Add @var{path} to the default library search path. This option exists
1039 for Solaris compatibility.
1041 @kindex -z @var{keyword}
1042 @item -z @var{keyword}
1043 The recognized keywords are:
1047 Combines multiple reloc sections and sorts them to make dynamic symbol
1048 lookup caching possible.
1051 Disallows undefined symbols in object files. Undefined symbols in
1052 shared libraries are still allowed.
1055 Marks the object as requiring executable stack.
1058 This option is only meaningful when building a shared object. It makes
1059 the symbols defined by this shared object available for symbol resolution
1060 of subsequently loaded libraries.
1063 This option is only meaningful when building a shared object.
1064 It marks the object so that its runtime initialization will occur
1065 before the runtime initialization of any other objects brought into
1066 the process at the same time. Similarly the runtime finalization of
1067 the object will occur after the runtime finalization of any other
1071 Marks the object that its symbol table interposes before all symbols
1072 but the primary executable.
1075 When generating an executable or shared library, mark it to tell the
1076 dynamic linker to defer function call resolution to the point when
1077 the function is called (lazy binding), rather than at load time.
1078 Lazy binding is the default.
1081 Marks the object that its filters be processed immediately at
1085 Allows multiple definitions.
1088 Disables multiple reloc sections combining.
1091 Disable linker generated .dynbss variables used in place of variables
1092 defined in shared libraries. May result in dynamic text relocations.
1095 Marks the object that the search for dependencies of this object will
1096 ignore any default library search paths.
1099 Marks the object shouldn't be unloaded at runtime.
1102 Marks the object not available to @code{dlopen}.
1105 Marks the object can not be dumped by @code{dldump}.
1108 Marks the object as not requiring executable stack.
1111 Treat DT_TEXTREL in shared object as error.
1114 Don't treat DT_TEXTREL in shared object as error.
1117 Don't treat DT_TEXTREL in shared object as error.
1120 Don't create an ELF @code{PT_GNU_RELRO} segment header in the object.
1123 When generating an executable or shared library, mark it to tell the
1124 dynamic linker to resolve all symbols when the program is started, or
1125 when the shared library is linked to using dlopen, instead of
1126 deferring function call resolution to the point when the function is
1130 Marks the object may contain $ORIGIN.
1133 Create an ELF @code{PT_GNU_RELRO} segment header in the object.
1135 @item max-page-size=@var{value}
1136 Set the emulation maximum page size to @var{value}.
1138 @item common-page-size=@var{value}
1139 Set the emulation common page size to @var{value}.
1141 @item stack-size=@var{value}
1142 Specify a stack size for in an ELF @code{PT_GNU_STACK} segment.
1143 Specifying zero will override any default non-zero sized
1144 @code{PT_GNU_STACK} segment creation.
1147 Always generate BND prefix in PLT entries. Supported for Linux/x86_64.
1149 @item noextern-protected-data
1150 Don't treat protected data symbol as external when building shared
1151 library. This option overrides linker backend default. It can be used
1152 to workaround incorrect relocations against protected data symbols
1153 generated by compiler. Updates on protected data symbols by another
1154 module aren't visibile to the resulting shared library. Supported for
1159 Other keywords are ignored for Solaris compatibility.
1162 @cindex groups of archives
1163 @item -( @var{archives} -)
1164 @itemx --start-group @var{archives} --end-group
1165 The @var{archives} should be a list of archive files. They may be
1166 either explicit file names, or @samp{-l} options.
1168 The specified archives are searched repeatedly until no new undefined
1169 references are created. Normally, an archive is searched only once in
1170 the order that it is specified on the command line. If a symbol in that
1171 archive is needed to resolve an undefined symbol referred to by an
1172 object in an archive that appears later on the command line, the linker
1173 would not be able to resolve that reference. By grouping the archives,
1174 they all be searched repeatedly until all possible references are
1177 Using this option has a significant performance cost. It is best to use
1178 it only when there are unavoidable circular references between two or
1181 @kindex --accept-unknown-input-arch
1182 @kindex --no-accept-unknown-input-arch
1183 @item --accept-unknown-input-arch
1184 @itemx --no-accept-unknown-input-arch
1185 Tells the linker to accept input files whose architecture cannot be
1186 recognised. The assumption is that the user knows what they are doing
1187 and deliberately wants to link in these unknown input files. This was
1188 the default behaviour of the linker, before release 2.14. The default
1189 behaviour from release 2.14 onwards is to reject such input files, and
1190 so the @samp{--accept-unknown-input-arch} option has been added to
1191 restore the old behaviour.
1194 @kindex --no-as-needed
1196 @itemx --no-as-needed
1197 This option affects ELF DT_NEEDED tags for dynamic libraries mentioned
1198 on the command line after the @option{--as-needed} option. Normally
1199 the linker will add a DT_NEEDED tag for each dynamic library mentioned
1200 on the command line, regardless of whether the library is actually
1201 needed or not. @option{--as-needed} causes a DT_NEEDED tag to only be
1202 emitted for a library that @emph{at that point in the link} satisfies a
1203 non-weak undefined symbol reference from a regular object file or, if
1204 the library is not found in the DT_NEEDED lists of other needed libraries, a
1205 non-weak undefined symbol reference from another needed dynamic library.
1206 Object files or libraries appearing on the command line @emph{after}
1207 the library in question do not affect whether the library is seen as
1208 needed. This is similar to the rules for extraction of object files
1209 from archives. @option{--no-as-needed} restores the default behaviour.
1211 @kindex --add-needed
1212 @kindex --no-add-needed
1214 @itemx --no-add-needed
1215 These two options have been deprecated because of the similarity of
1216 their names to the @option{--as-needed} and @option{--no-as-needed}
1217 options. They have been replaced by @option{--copy-dt-needed-entries}
1218 and @option{--no-copy-dt-needed-entries}.
1220 @kindex -assert @var{keyword}
1221 @item -assert @var{keyword}
1222 This option is ignored for SunOS compatibility.
1226 @kindex -call_shared
1230 Link against dynamic libraries. This is only meaningful on platforms
1231 for which shared libraries are supported. This option is normally the
1232 default on such platforms. The different variants of this option are
1233 for compatibility with various systems. You may use this option
1234 multiple times on the command line: it affects library searching for
1235 @option{-l} options which follow it.
1239 Set the @code{DF_1_GROUP} flag in the @code{DT_FLAGS_1} entry in the dynamic
1240 section. This causes the runtime linker to handle lookups in this
1241 object and its dependencies to be performed only inside the group.
1242 @option{--unresolved-symbols=report-all} is implied. This option is
1243 only meaningful on ELF platforms which support shared libraries.
1253 Do not link against shared libraries. This is only meaningful on
1254 platforms for which shared libraries are supported. The different
1255 variants of this option are for compatibility with various systems. You
1256 may use this option multiple times on the command line: it affects
1257 library searching for @option{-l} options which follow it. This
1258 option also implies @option{--unresolved-symbols=report-all}. This
1259 option can be used with @option{-shared}. Doing so means that a
1260 shared library is being created but that all of the library's external
1261 references must be resolved by pulling in entries from static
1266 When creating a shared library, bind references to global symbols to the
1267 definition within the shared library, if any. Normally, it is possible
1268 for a program linked against a shared library to override the definition
1269 within the shared library. This option is only meaningful on ELF
1270 platforms which support shared libraries.
1272 @kindex -Bsymbolic-functions
1273 @item -Bsymbolic-functions
1274 When creating a shared library, bind references to global function
1275 symbols to the definition within the shared library, if any.
1276 This option is only meaningful on ELF platforms which support shared
1279 @kindex --dynamic-list=@var{dynamic-list-file}
1280 @item --dynamic-list=@var{dynamic-list-file}
1281 Specify the name of a dynamic list file to the linker. This is
1282 typically used when creating shared libraries to specify a list of
1283 global symbols whose references shouldn't be bound to the definition
1284 within the shared library, or creating dynamically linked executables
1285 to specify a list of symbols which should be added to the symbol table
1286 in the executable. This option is only meaningful on ELF platforms
1287 which support shared libraries.
1289 The format of the dynamic list is the same as the version node without
1290 scope and node name. See @ref{VERSION} for more information.
1292 @kindex --dynamic-list-data
1293 @item --dynamic-list-data
1294 Include all global data symbols to the dynamic list.
1296 @kindex --dynamic-list-cpp-new
1297 @item --dynamic-list-cpp-new
1298 Provide the builtin dynamic list for C++ operator new and delete. It
1299 is mainly useful for building shared libstdc++.
1301 @kindex --dynamic-list-cpp-typeinfo
1302 @item --dynamic-list-cpp-typeinfo
1303 Provide the builtin dynamic list for C++ runtime type identification.
1305 @kindex --check-sections
1306 @kindex --no-check-sections
1307 @item --check-sections
1308 @itemx --no-check-sections
1309 Asks the linker @emph{not} to check section addresses after they have
1310 been assigned to see if there are any overlaps. Normally the linker will
1311 perform this check, and if it finds any overlaps it will produce
1312 suitable error messages. The linker does know about, and does make
1313 allowances for sections in overlays. The default behaviour can be
1314 restored by using the command line switch @option{--check-sections}.
1315 Section overlap is not usually checked for relocatable links. You can
1316 force checking in that case by using the @option{--check-sections}
1319 @kindex --copy-dt-needed-entries
1320 @kindex --no-copy-dt-needed-entries
1321 @item --copy-dt-needed-entries
1322 @itemx --no-copy-dt-needed-entries
1323 This option affects the treatment of dynamic libraries referred to
1324 by DT_NEEDED tags @emph{inside} ELF dynamic libraries mentioned on the
1325 command line. Normally the linker won't add a DT_NEEDED tag to the
1326 output binary for each library mentioned in a DT_NEEDED tag in an
1327 input dynamic library. With @option{--copy-dt-needed-entries}
1328 specified on the command line however any dynamic libraries that
1329 follow it will have their DT_NEEDED entries added. The default
1330 behaviour can be restored with @option{--no-copy-dt-needed-entries}.
1332 This option also has an effect on the resolution of symbols in dynamic
1333 libraries. With @option{--copy-dt-needed-entries} dynamic libraries
1334 mentioned on the command line will be recursively searched, following
1335 their DT_NEEDED tags to other libraries, in order to resolve symbols
1336 required by the output binary. With the default setting however
1337 the searching of dynamic libraries that follow it will stop with the
1338 dynamic library itself. No DT_NEEDED links will be traversed to resolve
1341 @cindex cross reference table
1344 Output a cross reference table. If a linker map file is being
1345 generated, the cross reference table is printed to the map file.
1346 Otherwise, it is printed on the standard output.
1348 The format of the table is intentionally simple, so that it may be
1349 easily processed by a script if necessary. The symbols are printed out,
1350 sorted by name. For each symbol, a list of file names is given. If the
1351 symbol is defined, the first file listed is the location of the
1352 definition. If the symbol is defined as a common value then any files
1353 where this happens appear next. Finally any files that reference the
1356 @cindex common allocation
1357 @kindex --no-define-common
1358 @item --no-define-common
1359 This option inhibits the assignment of addresses to common symbols.
1360 The script command @code{INHIBIT_COMMON_ALLOCATION} has the same effect.
1361 @xref{Miscellaneous Commands}.
1363 The @samp{--no-define-common} option allows decoupling
1364 the decision to assign addresses to Common symbols from the choice
1365 of the output file type; otherwise a non-Relocatable output type
1366 forces assigning addresses to Common symbols.
1367 Using @samp{--no-define-common} allows Common symbols that are referenced
1368 from a shared library to be assigned addresses only in the main program.
1369 This eliminates the unused duplicate space in the shared library,
1370 and also prevents any possible confusion over resolving to the wrong
1371 duplicate when there are many dynamic modules with specialized search
1372 paths for runtime symbol resolution.
1374 @cindex symbols, from command line
1375 @kindex --defsym=@var{symbol}=@var{exp}
1376 @item --defsym=@var{symbol}=@var{expression}
1377 Create a global symbol in the output file, containing the absolute
1378 address given by @var{expression}. You may use this option as many
1379 times as necessary to define multiple symbols in the command line. A
1380 limited form of arithmetic is supported for the @var{expression} in this
1381 context: you may give a hexadecimal constant or the name of an existing
1382 symbol, or use @code{+} and @code{-} to add or subtract hexadecimal
1383 constants or symbols. If you need more elaborate expressions, consider
1384 using the linker command language from a script (@pxref{Assignments}).
1385 @emph{Note:} there should be no white space between @var{symbol}, the
1386 equals sign (``@key{=}''), and @var{expression}.
1388 @cindex demangling, from command line
1389 @kindex --demangle[=@var{style}]
1390 @kindex --no-demangle
1391 @item --demangle[=@var{style}]
1392 @itemx --no-demangle
1393 These options control whether to demangle symbol names in error messages
1394 and other output. When the linker is told to demangle, it tries to
1395 present symbol names in a readable fashion: it strips leading
1396 underscores if they are used by the object file format, and converts C++
1397 mangled symbol names into user readable names. Different compilers have
1398 different mangling styles. The optional demangling style argument can be used
1399 to choose an appropriate demangling style for your compiler. The linker will
1400 demangle by default unless the environment variable @samp{COLLECT_NO_DEMANGLE}
1401 is set. These options may be used to override the default.
1403 @cindex dynamic linker, from command line
1404 @kindex -I@var{file}
1405 @kindex --dynamic-linker=@var{file}
1407 @itemx --dynamic-linker=@var{file}
1408 Set the name of the dynamic linker. This is only meaningful when
1409 generating dynamically linked ELF executables. The default dynamic
1410 linker is normally correct; don't use this unless you know what you are
1413 @kindex --fatal-warnings
1414 @kindex --no-fatal-warnings
1415 @item --fatal-warnings
1416 @itemx --no-fatal-warnings
1417 Treat all warnings as errors. The default behaviour can be restored
1418 with the option @option{--no-fatal-warnings}.
1420 @kindex --force-exe-suffix
1421 @item --force-exe-suffix
1422 Make sure that an output file has a .exe suffix.
1424 If a successfully built fully linked output file does not have a
1425 @code{.exe} or @code{.dll} suffix, this option forces the linker to copy
1426 the output file to one of the same name with a @code{.exe} suffix. This
1427 option is useful when using unmodified Unix makefiles on a Microsoft
1428 Windows host, since some versions of Windows won't run an image unless
1429 it ends in a @code{.exe} suffix.
1431 @kindex --gc-sections
1432 @kindex --no-gc-sections
1433 @cindex garbage collection
1435 @itemx --no-gc-sections
1436 Enable garbage collection of unused input sections. It is ignored on
1437 targets that do not support this option. The default behaviour (of not
1438 performing this garbage collection) can be restored by specifying
1439 @samp{--no-gc-sections} on the command line. Note that garbage
1440 collection for COFF and PE format targets is supported, but the
1441 implementation is currently considered to be experimental.
1443 @samp{--gc-sections} decides which input sections are used by
1444 examining symbols and relocations. The section containing the entry
1445 symbol and all sections containing symbols undefined on the
1446 command-line will be kept, as will sections containing symbols
1447 referenced by dynamic objects. Note that when building shared
1448 libraries, the linker must assume that any visible symbol is
1449 referenced. Once this initial set of sections has been determined,
1450 the linker recursively marks as used any section referenced by their
1451 relocations. See @samp{--entry} and @samp{--undefined}.
1453 This option can be set when doing a partial link (enabled with option
1454 @samp{-r}). In this case the root of symbols kept must be explicitly
1455 specified either by an @samp{--entry} or @samp{--undefined} option or by
1456 a @code{ENTRY} command in the linker script.
1458 @kindex --print-gc-sections
1459 @kindex --no-print-gc-sections
1460 @cindex garbage collection
1461 @item --print-gc-sections
1462 @itemx --no-print-gc-sections
1463 List all sections removed by garbage collection. The listing is
1464 printed on stderr. This option is only effective if garbage
1465 collection has been enabled via the @samp{--gc-sections}) option. The
1466 default behaviour (of not listing the sections that are removed) can
1467 be restored by specifying @samp{--no-print-gc-sections} on the command
1470 @kindex --print-output-format
1471 @cindex output format
1472 @item --print-output-format
1473 Print the name of the default output format (perhaps influenced by
1474 other command-line options). This is the string that would appear
1475 in an @code{OUTPUT_FORMAT} linker script command (@pxref{File Commands}).
1477 @kindex --print-memory-usage
1478 @cindex memory usage
1479 @item --print-memory-usage
1480 Print used size, total size and used size of memory regions created with
1481 the @ref{MEMORY} command. This is useful on embedded targets to have a
1482 quick view of amount of free memory. The format of the output has one
1483 headline and one line per region. It is both human readable and easily
1484 parsable by tools. Here is an example of an output:
1487 Memory region Used Size Region Size %age Used
1488 ROM: 256 KB 1 MB 25.00%
1489 RAM: 32 B 2 GB 0.00%
1496 Print a summary of the command-line options on the standard output and exit.
1498 @kindex --target-help
1500 Print a summary of all target specific options on the standard output and exit.
1502 @kindex -Map=@var{mapfile}
1503 @item -Map=@var{mapfile}
1504 Print a link map to the file @var{mapfile}. See the description of the
1505 @option{-M} option, above.
1507 @cindex memory usage
1508 @kindex --no-keep-memory
1509 @item --no-keep-memory
1510 @command{ld} normally optimizes for speed over memory usage by caching the
1511 symbol tables of input files in memory. This option tells @command{ld} to
1512 instead optimize for memory usage, by rereading the symbol tables as
1513 necessary. This may be required if @command{ld} runs out of memory space
1514 while linking a large executable.
1516 @kindex --no-undefined
1518 @item --no-undefined
1520 Report unresolved symbol references from regular object files. This
1521 is done even if the linker is creating a non-symbolic shared library.
1522 The switch @option{--[no-]allow-shlib-undefined} controls the
1523 behaviour for reporting unresolved references found in shared
1524 libraries being linked in.
1526 @kindex --allow-multiple-definition
1528 @item --allow-multiple-definition
1530 Normally when a symbol is defined multiple times, the linker will
1531 report a fatal error. These options allow multiple definitions and the
1532 first definition will be used.
1534 @kindex --allow-shlib-undefined
1535 @kindex --no-allow-shlib-undefined
1536 @item --allow-shlib-undefined
1537 @itemx --no-allow-shlib-undefined
1538 Allows or disallows undefined symbols in shared libraries.
1539 This switch is similar to @option{--no-undefined} except that it
1540 determines the behaviour when the undefined symbols are in a
1541 shared library rather than a regular object file. It does not affect
1542 how undefined symbols in regular object files are handled.
1544 The default behaviour is to report errors for any undefined symbols
1545 referenced in shared libraries if the linker is being used to create
1546 an executable, but to allow them if the linker is being used to create
1549 The reasons for allowing undefined symbol references in shared
1550 libraries specified at link time are that:
1554 A shared library specified at link time may not be the same as the one
1555 that is available at load time, so the symbol might actually be
1556 resolvable at load time.
1558 There are some operating systems, eg BeOS and HPPA, where undefined
1559 symbols in shared libraries are normal.
1561 The BeOS kernel for example patches shared libraries at load time to
1562 select whichever function is most appropriate for the current
1563 architecture. This is used, for example, to dynamically select an
1564 appropriate memset function.
1567 @kindex --no-undefined-version
1568 @item --no-undefined-version
1569 Normally when a symbol has an undefined version, the linker will ignore
1570 it. This option disallows symbols with undefined version and a fatal error
1571 will be issued instead.
1573 @kindex --default-symver
1574 @item --default-symver
1575 Create and use a default symbol version (the soname) for unversioned
1578 @kindex --default-imported-symver
1579 @item --default-imported-symver
1580 Create and use a default symbol version (the soname) for unversioned
1583 @kindex --no-warn-mismatch
1584 @item --no-warn-mismatch
1585 Normally @command{ld} will give an error if you try to link together input
1586 files that are mismatched for some reason, perhaps because they have
1587 been compiled for different processors or for different endiannesses.
1588 This option tells @command{ld} that it should silently permit such possible
1589 errors. This option should only be used with care, in cases when you
1590 have taken some special action that ensures that the linker errors are
1593 @kindex --no-warn-search-mismatch
1594 @item --no-warn-search-mismatch
1595 Normally @command{ld} will give a warning if it finds an incompatible
1596 library during a library search. This option silences the warning.
1598 @kindex --no-whole-archive
1599 @item --no-whole-archive
1600 Turn off the effect of the @option{--whole-archive} option for subsequent
1603 @cindex output file after errors
1604 @kindex --noinhibit-exec
1605 @item --noinhibit-exec
1606 Retain the executable output file whenever it is still usable.
1607 Normally, the linker will not produce an output file if it encounters
1608 errors during the link process; it exits without writing an output file
1609 when it issues any error whatsoever.
1613 Only search library directories explicitly specified on the
1614 command line. Library directories specified in linker scripts
1615 (including linker scripts specified on the command line) are ignored.
1617 @ifclear SingleFormat
1618 @kindex --oformat=@var{output-format}
1619 @item --oformat=@var{output-format}
1620 @command{ld} may be configured to support more than one kind of object
1621 file. If your @command{ld} is configured this way, you can use the
1622 @samp{--oformat} option to specify the binary format for the output
1623 object file. Even when @command{ld} is configured to support alternative
1624 object formats, you don't usually need to specify this, as @command{ld}
1625 should be configured to produce as a default output format the most
1626 usual format on each machine. @var{output-format} is a text string, the
1627 name of a particular format supported by the BFD libraries. (You can
1628 list the available binary formats with @samp{objdump -i}.) The script
1629 command @code{OUTPUT_FORMAT} can also specify the output format, but
1630 this option overrides it. @xref{BFD}.
1634 @kindex --pic-executable
1636 @itemx --pic-executable
1637 @cindex position independent executables
1638 Create a position independent executable. This is currently only supported on
1639 ELF platforms. Position independent executables are similar to shared
1640 libraries in that they are relocated by the dynamic linker to the virtual
1641 address the OS chooses for them (which can vary between invocations). Like
1642 normal dynamically linked executables they can be executed and symbols
1643 defined in the executable cannot be overridden by shared libraries.
1647 This option is ignored for Linux compatibility.
1651 This option is ignored for SVR4 compatibility.
1654 @cindex synthesizing linker
1655 @cindex relaxing addressing modes
1659 An option with machine dependent effects.
1661 This option is only supported on a few targets.
1664 @xref{H8/300,,@command{ld} and the H8/300}.
1667 @xref{i960,, @command{ld} and the Intel 960 family}.
1670 @xref{Xtensa,, @command{ld} and Xtensa Processors}.
1673 @xref{M68HC11/68HC12,,@command{ld} and the 68HC11 and 68HC12}.
1676 @xref{Nios II,,@command{ld} and the Altera Nios II}.
1679 @xref{PowerPC ELF32,,@command{ld} and PowerPC 32-bit ELF Support}.
1682 On some platforms the @samp{--relax} option performs target specific,
1683 global optimizations that become possible when the linker resolves
1684 addressing in the program, such as relaxing address modes,
1685 synthesizing new instructions, selecting shorter version of current
1686 instructions, and combining constant values.
1688 On some platforms these link time global optimizations may make symbolic
1689 debugging of the resulting executable impossible.
1691 This is known to be the case for the Matsushita MN10200 and MN10300
1692 family of processors.
1696 On platforms where this is not supported, @samp{--relax} is accepted,
1700 On platforms where @samp{--relax} is accepted the option
1701 @samp{--no-relax} can be used to disable the feature.
1703 @cindex retaining specified symbols
1704 @cindex stripping all but some symbols
1705 @cindex symbols, retaining selectively
1706 @kindex --retain-symbols-file=@var{filename}
1707 @item --retain-symbols-file=@var{filename}
1708 Retain @emph{only} the symbols listed in the file @var{filename},
1709 discarding all others. @var{filename} is simply a flat file, with one
1710 symbol name per line. This option is especially useful in environments
1714 where a large global symbol table is accumulated gradually, to conserve
1717 @samp{--retain-symbols-file} does @emph{not} discard undefined symbols,
1718 or symbols needed for relocations.
1720 You may only specify @samp{--retain-symbols-file} once in the command
1721 line. It overrides @samp{-s} and @samp{-S}.
1724 @item -rpath=@var{dir}
1725 @cindex runtime library search path
1726 @kindex -rpath=@var{dir}
1727 Add a directory to the runtime library search path. This is used when
1728 linking an ELF executable with shared objects. All @option{-rpath}
1729 arguments are concatenated and passed to the runtime linker, which uses
1730 them to locate shared objects at runtime. The @option{-rpath} option is
1731 also used when locating shared objects which are needed by shared
1732 objects explicitly included in the link; see the description of the
1733 @option{-rpath-link} option. If @option{-rpath} is not used when linking an
1734 ELF executable, the contents of the environment variable
1735 @code{LD_RUN_PATH} will be used if it is defined.
1737 The @option{-rpath} option may also be used on SunOS. By default, on
1738 SunOS, the linker will form a runtime search patch out of all the
1739 @option{-L} options it is given. If a @option{-rpath} option is used, the
1740 runtime search path will be formed exclusively using the @option{-rpath}
1741 options, ignoring the @option{-L} options. This can be useful when using
1742 gcc, which adds many @option{-L} options which may be on NFS mounted
1745 For compatibility with other ELF linkers, if the @option{-R} option is
1746 followed by a directory name, rather than a file name, it is treated as
1747 the @option{-rpath} option.
1751 @cindex link-time runtime library search path
1752 @kindex -rpath-link=@var{dir}
1753 @item -rpath-link=@var{dir}
1754 When using ELF or SunOS, one shared library may require another. This
1755 happens when an @code{ld -shared} link includes a shared library as one
1758 When the linker encounters such a dependency when doing a non-shared,
1759 non-relocatable link, it will automatically try to locate the required
1760 shared library and include it in the link, if it is not included
1761 explicitly. In such a case, the @option{-rpath-link} option
1762 specifies the first set of directories to search. The
1763 @option{-rpath-link} option may specify a sequence of directory names
1764 either by specifying a list of names separated by colons, or by
1765 appearing multiple times.
1767 This option should be used with caution as it overrides the search path
1768 that may have been hard compiled into a shared library. In such a case it
1769 is possible to use unintentionally a different search path than the
1770 runtime linker would do.
1772 The linker uses the following search paths to locate required shared
1776 Any directories specified by @option{-rpath-link} options.
1778 Any directories specified by @option{-rpath} options. The difference
1779 between @option{-rpath} and @option{-rpath-link} is that directories
1780 specified by @option{-rpath} options are included in the executable and
1781 used at runtime, whereas the @option{-rpath-link} option is only effective
1782 at link time. Searching @option{-rpath} in this way is only supported
1783 by native linkers and cross linkers which have been configured with
1784 the @option{--with-sysroot} option.
1786 On an ELF system, for native linkers, if the @option{-rpath} and
1787 @option{-rpath-link} options were not used, search the contents of the
1788 environment variable @code{LD_RUN_PATH}.
1790 On SunOS, if the @option{-rpath} option was not used, search any
1791 directories specified using @option{-L} options.
1793 For a native linker, search the contents of the environment
1794 variable @code{LD_LIBRARY_PATH}.
1796 For a native ELF linker, the directories in @code{DT_RUNPATH} or
1797 @code{DT_RPATH} of a shared library are searched for shared
1798 libraries needed by it. The @code{DT_RPATH} entries are ignored if
1799 @code{DT_RUNPATH} entries exist.
1801 The default directories, normally @file{/lib} and @file{/usr/lib}.
1803 For a native linker on an ELF system, if the file @file{/etc/ld.so.conf}
1804 exists, the list of directories found in that file.
1807 If the required shared library is not found, the linker will issue a
1808 warning and continue with the link.
1815 @cindex shared libraries
1816 Create a shared library. This is currently only supported on ELF, XCOFF
1817 and SunOS platforms. On SunOS, the linker will automatically create a
1818 shared library if the @option{-e} option is not used and there are
1819 undefined symbols in the link.
1821 @kindex --sort-common
1823 @itemx --sort-common=ascending
1824 @itemx --sort-common=descending
1825 This option tells @command{ld} to sort the common symbols by alignment in
1826 ascending or descending order when it places them in the appropriate output
1827 sections. The symbol alignments considered are sixteen-byte or larger,
1828 eight-byte, four-byte, two-byte, and one-byte. This is to prevent gaps
1829 between symbols due to alignment constraints. If no sorting order is
1830 specified, then descending order is assumed.
1832 @kindex --sort-section=name
1833 @item --sort-section=name
1834 This option will apply @code{SORT_BY_NAME} to all wildcard section
1835 patterns in the linker script.
1837 @kindex --sort-section=alignment
1838 @item --sort-section=alignment
1839 This option will apply @code{SORT_BY_ALIGNMENT} to all wildcard section
1840 patterns in the linker script.
1842 @kindex --split-by-file
1843 @item --split-by-file[=@var{size}]
1844 Similar to @option{--split-by-reloc} but creates a new output section for
1845 each input file when @var{size} is reached. @var{size} defaults to a
1846 size of 1 if not given.
1848 @kindex --split-by-reloc
1849 @item --split-by-reloc[=@var{count}]
1850 Tries to creates extra sections in the output file so that no single
1851 output section in the file contains more than @var{count} relocations.
1852 This is useful when generating huge relocatable files for downloading into
1853 certain real time kernels with the COFF object file format; since COFF
1854 cannot represent more than 65535 relocations in a single section. Note
1855 that this will fail to work with object file formats which do not
1856 support arbitrary sections. The linker will not split up individual
1857 input sections for redistribution, so if a single input section contains
1858 more than @var{count} relocations one output section will contain that
1859 many relocations. @var{count} defaults to a value of 32768.
1863 Compute and display statistics about the operation of the linker, such
1864 as execution time and memory usage.
1866 @kindex --sysroot=@var{directory}
1867 @item --sysroot=@var{directory}
1868 Use @var{directory} as the location of the sysroot, overriding the
1869 configure-time default. This option is only supported by linkers
1870 that were configured using @option{--with-sysroot}.
1872 @kindex --traditional-format
1873 @cindex traditional format
1874 @item --traditional-format
1875 For some targets, the output of @command{ld} is different in some ways from
1876 the output of some existing linker. This switch requests @command{ld} to
1877 use the traditional format instead.
1880 For example, on SunOS, @command{ld} combines duplicate entries in the
1881 symbol string table. This can reduce the size of an output file with
1882 full debugging information by over 30 percent. Unfortunately, the SunOS
1883 @code{dbx} program can not read the resulting program (@code{gdb} has no
1884 trouble). The @samp{--traditional-format} switch tells @command{ld} to not
1885 combine duplicate entries.
1887 @kindex --section-start=@var{sectionname}=@var{org}
1888 @item --section-start=@var{sectionname}=@var{org}
1889 Locate a section in the output file at the absolute
1890 address given by @var{org}. You may use this option as many
1891 times as necessary to locate multiple sections in the command
1893 @var{org} must be a single hexadecimal integer;
1894 for compatibility with other linkers, you may omit the leading
1895 @samp{0x} usually associated with hexadecimal values. @emph{Note:} there
1896 should be no white space between @var{sectionname}, the equals
1897 sign (``@key{=}''), and @var{org}.
1899 @kindex -Tbss=@var{org}
1900 @kindex -Tdata=@var{org}
1901 @kindex -Ttext=@var{org}
1902 @cindex segment origins, cmd line
1903 @item -Tbss=@var{org}
1904 @itemx -Tdata=@var{org}
1905 @itemx -Ttext=@var{org}
1906 Same as @option{--section-start}, with @code{.bss}, @code{.data} or
1907 @code{.text} as the @var{sectionname}.
1909 @kindex -Ttext-segment=@var{org}
1910 @item -Ttext-segment=@var{org}
1911 @cindex text segment origin, cmd line
1912 When creating an ELF executable, it will set the address of the first
1913 byte of the text segment.
1915 @kindex -Trodata-segment=@var{org}
1916 @item -Trodata-segment=@var{org}
1917 @cindex rodata segment origin, cmd line
1918 When creating an ELF executable or shared object for a target where
1919 the read-only data is in its own segment separate from the executable
1920 text, it will set the address of the first byte of the read-only data segment.
1922 @kindex -Tldata-segment=@var{org}
1923 @item -Tldata-segment=@var{org}
1924 @cindex ldata segment origin, cmd line
1925 When creating an ELF executable or shared object for x86-64 medium memory
1926 model, it will set the address of the first byte of the ldata segment.
1928 @kindex --unresolved-symbols
1929 @item --unresolved-symbols=@var{method}
1930 Determine how to handle unresolved symbols. There are four possible
1931 values for @samp{method}:
1935 Do not report any unresolved symbols.
1938 Report all unresolved symbols. This is the default.
1940 @item ignore-in-object-files
1941 Report unresolved symbols that are contained in shared libraries, but
1942 ignore them if they come from regular object files.
1944 @item ignore-in-shared-libs
1945 Report unresolved symbols that come from regular object files, but
1946 ignore them if they come from shared libraries. This can be useful
1947 when creating a dynamic binary and it is known that all the shared
1948 libraries that it should be referencing are included on the linker's
1952 The behaviour for shared libraries on their own can also be controlled
1953 by the @option{--[no-]allow-shlib-undefined} option.
1955 Normally the linker will generate an error message for each reported
1956 unresolved symbol but the option @option{--warn-unresolved-symbols}
1957 can change this to a warning.
1959 @kindex --verbose[=@var{NUMBER}]
1960 @cindex verbose[=@var{NUMBER}]
1962 @itemx --verbose[=@var{NUMBER}]
1963 Display the version number for @command{ld} and list the linker emulations
1964 supported. Display which input files can and cannot be opened. Display
1965 the linker script being used by the linker. If the optional @var{NUMBER}
1966 argument > 1, plugin symbol status will also be displayed.
1968 @kindex --version-script=@var{version-scriptfile}
1969 @cindex version script, symbol versions
1970 @item --version-script=@var{version-scriptfile}
1971 Specify the name of a version script to the linker. This is typically
1972 used when creating shared libraries to specify additional information
1973 about the version hierarchy for the library being created. This option
1974 is only fully supported on ELF platforms which support shared libraries;
1975 see @ref{VERSION}. It is partially supported on PE platforms, which can
1976 use version scripts to filter symbol visibility in auto-export mode: any
1977 symbols marked @samp{local} in the version script will not be exported.
1980 @kindex --warn-common
1981 @cindex warnings, on combining symbols
1982 @cindex combining symbols, warnings on
1984 Warn when a common symbol is combined with another common symbol or with
1985 a symbol definition. Unix linkers allow this somewhat sloppy practice,
1986 but linkers on some other operating systems do not. This option allows
1987 you to find potential problems from combining global symbols.
1988 Unfortunately, some C libraries use this practice, so you may get some
1989 warnings about symbols in the libraries as well as in your programs.
1991 There are three kinds of global symbols, illustrated here by C examples:
1995 A definition, which goes in the initialized data section of the output
1999 An undefined reference, which does not allocate space.
2000 There must be either a definition or a common symbol for the
2004 A common symbol. If there are only (one or more) common symbols for a
2005 variable, it goes in the uninitialized data area of the output file.
2006 The linker merges multiple common symbols for the same variable into a
2007 single symbol. If they are of different sizes, it picks the largest
2008 size. The linker turns a common symbol into a declaration, if there is
2009 a definition of the same variable.
2012 The @samp{--warn-common} option can produce five kinds of warnings.
2013 Each warning consists of a pair of lines: the first describes the symbol
2014 just encountered, and the second describes the previous symbol
2015 encountered with the same name. One or both of the two symbols will be
2020 Turning a common symbol into a reference, because there is already a
2021 definition for the symbol.
2023 @var{file}(@var{section}): warning: common of `@var{symbol}'
2024 overridden by definition
2025 @var{file}(@var{section}): warning: defined here
2029 Turning a common symbol into a reference, because a later definition for
2030 the symbol is encountered. This is the same as the previous case,
2031 except that the symbols are encountered in a different order.
2033 @var{file}(@var{section}): warning: definition of `@var{symbol}'
2035 @var{file}(@var{section}): warning: common is here
2039 Merging a common symbol with a previous same-sized common symbol.
2041 @var{file}(@var{section}): warning: multiple common
2043 @var{file}(@var{section}): warning: previous common is here
2047 Merging a common symbol with a previous larger common symbol.
2049 @var{file}(@var{section}): warning: common of `@var{symbol}'
2050 overridden by larger common
2051 @var{file}(@var{section}): warning: larger common is here
2055 Merging a common symbol with a previous smaller common symbol. This is
2056 the same as the previous case, except that the symbols are
2057 encountered in a different order.
2059 @var{file}(@var{section}): warning: common of `@var{symbol}'
2060 overriding smaller common
2061 @var{file}(@var{section}): warning: smaller common is here
2065 @kindex --warn-constructors
2066 @item --warn-constructors
2067 Warn if any global constructors are used. This is only useful for a few
2068 object file formats. For formats like COFF or ELF, the linker can not
2069 detect the use of global constructors.
2071 @kindex --warn-multiple-gp
2072 @item --warn-multiple-gp
2073 Warn if multiple global pointer values are required in the output file.
2074 This is only meaningful for certain processors, such as the Alpha.
2075 Specifically, some processors put large-valued constants in a special
2076 section. A special register (the global pointer) points into the middle
2077 of this section, so that constants can be loaded efficiently via a
2078 base-register relative addressing mode. Since the offset in
2079 base-register relative mode is fixed and relatively small (e.g., 16
2080 bits), this limits the maximum size of the constant pool. Thus, in
2081 large programs, it is often necessary to use multiple global pointer
2082 values in order to be able to address all possible constants. This
2083 option causes a warning to be issued whenever this case occurs.
2086 @cindex warnings, on undefined symbols
2087 @cindex undefined symbols, warnings on
2089 Only warn once for each undefined symbol, rather than once per module
2092 @kindex --warn-orphan
2093 @kindex --no-warn-orphan
2094 @cindex warnings, on orphan sections
2095 @cindex orphan sections, warnings on
2097 The @option{--warn-orphan} option tells the linker to generate a
2098 warning message whenever it has to place an orphan section into the
2099 output file. @xref{Orphan Sections}. The @option{--no-warn-orphan}
2100 option restores the default behaviour of just silently placing these
2103 @kindex --warn-section-align
2104 @cindex warnings, on section alignment
2105 @cindex section alignment, warnings on
2106 @item --warn-section-align
2107 Warn if the address of an output section is changed because of
2108 alignment. Typically, the alignment will be set by an input section.
2109 The address will only be changed if it not explicitly specified; that
2110 is, if the @code{SECTIONS} command does not specify a start address for
2111 the section (@pxref{SECTIONS}).
2113 @kindex --warn-shared-textrel
2114 @item --warn-shared-textrel
2115 Warn if the linker adds a DT_TEXTREL to a shared object.
2117 @kindex --warn-alternate-em
2118 @item --warn-alternate-em
2119 Warn if an object has alternate ELF machine code.
2121 @kindex --warn-unresolved-symbols
2122 @item --warn-unresolved-symbols
2123 If the linker is going to report an unresolved symbol (see the option
2124 @option{--unresolved-symbols}) it will normally generate an error.
2125 This option makes it generate a warning instead.
2127 @kindex --error-unresolved-symbols
2128 @item --error-unresolved-symbols
2129 This restores the linker's default behaviour of generating errors when
2130 it is reporting unresolved symbols.
2132 @kindex --whole-archive
2133 @cindex including an entire archive
2134 @item --whole-archive
2135 For each archive mentioned on the command line after the
2136 @option{--whole-archive} option, include every object file in the archive
2137 in the link, rather than searching the archive for the required object
2138 files. This is normally used to turn an archive file into a shared
2139 library, forcing every object to be included in the resulting shared
2140 library. This option may be used more than once.
2142 Two notes when using this option from gcc: First, gcc doesn't know
2143 about this option, so you have to use @option{-Wl,-whole-archive}.
2144 Second, don't forget to use @option{-Wl,-no-whole-archive} after your
2145 list of archives, because gcc will add its own list of archives to
2146 your link and you may not want this flag to affect those as well.
2148 @kindex --wrap=@var{symbol}
2149 @item --wrap=@var{symbol}
2150 Use a wrapper function for @var{symbol}. Any undefined reference to
2151 @var{symbol} will be resolved to @code{__wrap_@var{symbol}}. Any
2152 undefined reference to @code{__real_@var{symbol}} will be resolved to
2155 This can be used to provide a wrapper for a system function. The
2156 wrapper function should be called @code{__wrap_@var{symbol}}. If it
2157 wishes to call the system function, it should call
2158 @code{__real_@var{symbol}}.
2160 Here is a trivial example:
2164 __wrap_malloc (size_t c)
2166 printf ("malloc called with %zu\n", c);
2167 return __real_malloc (c);
2171 If you link other code with this file using @option{--wrap malloc}, then
2172 all calls to @code{malloc} will call the function @code{__wrap_malloc}
2173 instead. The call to @code{__real_malloc} in @code{__wrap_malloc} will
2174 call the real @code{malloc} function.
2176 You may wish to provide a @code{__real_malloc} function as well, so that
2177 links without the @option{--wrap} option will succeed. If you do this,
2178 you should not put the definition of @code{__real_malloc} in the same
2179 file as @code{__wrap_malloc}; if you do, the assembler may resolve the
2180 call before the linker has a chance to wrap it to @code{malloc}.
2182 @kindex --eh-frame-hdr
2183 @item --eh-frame-hdr
2184 Request creation of @code{.eh_frame_hdr} section and ELF
2185 @code{PT_GNU_EH_FRAME} segment header.
2187 @kindex --ld-generated-unwind-info
2188 @item --no-ld-generated-unwind-info
2189 Request creation of @code{.eh_frame} unwind info for linker
2190 generated code sections like PLT. This option is on by default
2191 if linker generated unwind info is supported.
2193 @kindex --enable-new-dtags
2194 @kindex --disable-new-dtags
2195 @item --enable-new-dtags
2196 @itemx --disable-new-dtags
2197 This linker can create the new dynamic tags in ELF. But the older ELF
2198 systems may not understand them. If you specify
2199 @option{--enable-new-dtags}, the new dynamic tags will be created as needed
2200 and older dynamic tags will be omitted.
2201 If you specify @option{--disable-new-dtags}, no new dynamic tags will be
2202 created. By default, the new dynamic tags are not created. Note that
2203 those options are only available for ELF systems.
2205 @kindex --hash-size=@var{number}
2206 @item --hash-size=@var{number}
2207 Set the default size of the linker's hash tables to a prime number
2208 close to @var{number}. Increasing this value can reduce the length of
2209 time it takes the linker to perform its tasks, at the expense of
2210 increasing the linker's memory requirements. Similarly reducing this
2211 value can reduce the memory requirements at the expense of speed.
2213 @kindex --hash-style=@var{style}
2214 @item --hash-style=@var{style}
2215 Set the type of linker's hash table(s). @var{style} can be either
2216 @code{sysv} for classic ELF @code{.hash} section, @code{gnu} for
2217 new style GNU @code{.gnu.hash} section or @code{both} for both
2218 the classic ELF @code{.hash} and new style GNU @code{.gnu.hash}
2219 hash tables. The default is @code{sysv}.
2221 @kindex --compress-debug-sections=none
2222 @kindex --compress-debug-sections=zlib
2223 @kindex --compress-debug-sections=zlib-gnu
2224 @kindex --compress-debug-sections=zlib-gabi
2225 @item --compress-debug-sections=none
2226 @itemx --compress-debug-sections=zlib
2227 @itemx --compress-debug-sections=zlib-gnu
2228 @itemx --compress-debug-sections=zlib-gabi
2229 On ELF platforms , these options control how DWARF debug sections are
2230 compressed using zlib. @option{--compress-debug-sections=none} doesn't
2231 compress DWARF debug sections.
2232 @option{--compress-debug-sections=zlib-gnu} compresses DWARF debug
2233 sections and rename debug section names to begin with @samp{.zdebug}
2234 instead of @samp{.debug}. @option{--compress-debug-sections=zlib}
2235 and @option{--compress-debug-sections=zlib-gabi}
2236 compress DWARF debug sections with SHF_COMPRESSED from the ELF ABI.
2238 @kindex --reduce-memory-overheads
2239 @item --reduce-memory-overheads
2240 This option reduces memory requirements at ld runtime, at the expense of
2241 linking speed. This was introduced to select the old O(n^2) algorithm
2242 for link map file generation, rather than the new O(n) algorithm which uses
2243 about 40% more memory for symbol storage.
2245 Another effect of the switch is to set the default hash table size to
2246 1021, which again saves memory at the cost of lengthening the linker's
2247 run time. This is not done however if the @option{--hash-size} switch
2250 The @option{--reduce-memory-overheads} switch may be also be used to
2251 enable other tradeoffs in future versions of the linker.
2254 @kindex --build-id=@var{style}
2256 @itemx --build-id=@var{style}
2257 Request the creation of a @code{.note.gnu.build-id} ELF note section
2258 or a @code{.build-id} COFF section. The contents of the note are
2259 unique bits identifying this linked file. @var{style} can be
2260 @code{uuid} to use 128 random bits, @code{sha1} to use a 160-bit
2261 @sc{SHA1} hash on the normative parts of the output contents,
2262 @code{md5} to use a 128-bit @sc{MD5} hash on the normative parts of
2263 the output contents, or @code{0x@var{hexstring}} to use a chosen bit
2264 string specified as an even number of hexadecimal digits (@code{-} and
2265 @code{:} characters between digit pairs are ignored). If @var{style}
2266 is omitted, @code{sha1} is used.
2268 The @code{md5} and @code{sha1} styles produces an identifier
2269 that is always the same in an identical output file, but will be
2270 unique among all nonidentical output files. It is not intended
2271 to be compared as a checksum for the file's contents. A linked
2272 file may be changed later by other tools, but the build ID bit
2273 string identifying the original linked file does not change.
2275 Passing @code{none} for @var{style} disables the setting from any
2276 @code{--build-id} options earlier on the command line.
2281 @subsection Options Specific to i386 PE Targets
2283 @c man begin OPTIONS
2285 The i386 PE linker supports the @option{-shared} option, which causes
2286 the output to be a dynamically linked library (DLL) instead of a
2287 normal executable. You should name the output @code{*.dll} when you
2288 use this option. In addition, the linker fully supports the standard
2289 @code{*.def} files, which may be specified on the linker command line
2290 like an object file (in fact, it should precede archives it exports
2291 symbols from, to ensure that they get linked in, just like a normal
2294 In addition to the options common to all targets, the i386 PE linker
2295 support additional command line options that are specific to the i386
2296 PE target. Options that take values may be separated from their
2297 values by either a space or an equals sign.
2301 @kindex --add-stdcall-alias
2302 @item --add-stdcall-alias
2303 If given, symbols with a stdcall suffix (@@@var{nn}) will be exported
2304 as-is and also with the suffix stripped.
2305 [This option is specific to the i386 PE targeted port of the linker]
2308 @item --base-file @var{file}
2309 Use @var{file} as the name of a file in which to save the base
2310 addresses of all the relocations needed for generating DLLs with
2312 [This is an i386 PE specific option]
2316 Create a DLL instead of a regular executable. You may also use
2317 @option{-shared} or specify a @code{LIBRARY} in a given @code{.def}
2319 [This option is specific to the i386 PE targeted port of the linker]
2321 @kindex --enable-long-section-names
2322 @kindex --disable-long-section-names
2323 @item --enable-long-section-names
2324 @itemx --disable-long-section-names
2325 The PE variants of the Coff object format add an extension that permits
2326 the use of section names longer than eight characters, the normal limit
2327 for Coff. By default, these names are only allowed in object files, as
2328 fully-linked executable images do not carry the Coff string table required
2329 to support the longer names. As a GNU extension, it is possible to
2330 allow their use in executable images as well, or to (probably pointlessly!)
2331 disallow it in object files, by using these two options. Executable images
2332 generated with these long section names are slightly non-standard, carrying
2333 as they do a string table, and may generate confusing output when examined
2334 with non-GNU PE-aware tools, such as file viewers and dumpers. However,
2335 GDB relies on the use of PE long section names to find Dwarf-2 debug
2336 information sections in an executable image at runtime, and so if neither
2337 option is specified on the command-line, @command{ld} will enable long
2338 section names, overriding the default and technically correct behaviour,
2339 when it finds the presence of debug information while linking an executable
2340 image and not stripping symbols.
2341 [This option is valid for all PE targeted ports of the linker]
2343 @kindex --enable-stdcall-fixup
2344 @kindex --disable-stdcall-fixup
2345 @item --enable-stdcall-fixup
2346 @itemx --disable-stdcall-fixup
2347 If the link finds a symbol that it cannot resolve, it will attempt to
2348 do ``fuzzy linking'' by looking for another defined symbol that differs
2349 only in the format of the symbol name (cdecl vs stdcall) and will
2350 resolve that symbol by linking to the match. For example, the
2351 undefined symbol @code{_foo} might be linked to the function
2352 @code{_foo@@12}, or the undefined symbol @code{_bar@@16} might be linked
2353 to the function @code{_bar}. When the linker does this, it prints a
2354 warning, since it normally should have failed to link, but sometimes
2355 import libraries generated from third-party dlls may need this feature
2356 to be usable. If you specify @option{--enable-stdcall-fixup}, this
2357 feature is fully enabled and warnings are not printed. If you specify
2358 @option{--disable-stdcall-fixup}, this feature is disabled and such
2359 mismatches are considered to be errors.
2360 [This option is specific to the i386 PE targeted port of the linker]
2362 @kindex --leading-underscore
2363 @kindex --no-leading-underscore
2364 @item --leading-underscore
2365 @itemx --no-leading-underscore
2366 For most targets default symbol-prefix is an underscore and is defined
2367 in target's description. By this option it is possible to
2368 disable/enable the default underscore symbol-prefix.
2370 @cindex DLLs, creating
2371 @kindex --export-all-symbols
2372 @item --export-all-symbols
2373 If given, all global symbols in the objects used to build a DLL will
2374 be exported by the DLL. Note that this is the default if there
2375 otherwise wouldn't be any exported symbols. When symbols are
2376 explicitly exported via DEF files or implicitly exported via function
2377 attributes, the default is to not export anything else unless this
2378 option is given. Note that the symbols @code{DllMain@@12},
2379 @code{DllEntryPoint@@0}, @code{DllMainCRTStartup@@12}, and
2380 @code{impure_ptr} will not be automatically
2381 exported. Also, symbols imported from other DLLs will not be
2382 re-exported, nor will symbols specifying the DLL's internal layout
2383 such as those beginning with @code{_head_} or ending with
2384 @code{_iname}. In addition, no symbols from @code{libgcc},
2385 @code{libstd++}, @code{libmingw32}, or @code{crtX.o} will be exported.
2386 Symbols whose names begin with @code{__rtti_} or @code{__builtin_} will
2387 not be exported, to help with C++ DLLs. Finally, there is an
2388 extensive list of cygwin-private symbols that are not exported
2389 (obviously, this applies on when building DLLs for cygwin targets).
2390 These cygwin-excludes are: @code{_cygwin_dll_entry@@12},
2391 @code{_cygwin_crt0_common@@8}, @code{_cygwin_noncygwin_dll_entry@@12},
2392 @code{_fmode}, @code{_impure_ptr}, @code{cygwin_attach_dll},
2393 @code{cygwin_premain0}, @code{cygwin_premain1}, @code{cygwin_premain2},
2394 @code{cygwin_premain3}, and @code{environ}.
2395 [This option is specific to the i386 PE targeted port of the linker]
2397 @kindex --exclude-symbols
2398 @item --exclude-symbols @var{symbol},@var{symbol},...
2399 Specifies a list of symbols which should not be automatically
2400 exported. The symbol names may be delimited by commas or colons.
2401 [This option is specific to the i386 PE targeted port of the linker]
2403 @kindex --exclude-all-symbols
2404 @item --exclude-all-symbols
2405 Specifies no symbols should be automatically exported.
2406 [This option is specific to the i386 PE targeted port of the linker]
2408 @kindex --file-alignment
2409 @item --file-alignment
2410 Specify the file alignment. Sections in the file will always begin at
2411 file offsets which are multiples of this number. This defaults to
2413 [This option is specific to the i386 PE targeted port of the linker]
2417 @item --heap @var{reserve}
2418 @itemx --heap @var{reserve},@var{commit}
2419 Specify the number of bytes of memory to reserve (and optionally commit)
2420 to be used as heap for this program. The default is 1MB reserved, 4K
2422 [This option is specific to the i386 PE targeted port of the linker]
2425 @kindex --image-base
2426 @item --image-base @var{value}
2427 Use @var{value} as the base address of your program or dll. This is
2428 the lowest memory location that will be used when your program or dll
2429 is loaded. To reduce the need to relocate and improve performance of
2430 your dlls, each should have a unique base address and not overlap any
2431 other dlls. The default is 0x400000 for executables, and 0x10000000
2433 [This option is specific to the i386 PE targeted port of the linker]
2437 If given, the stdcall suffixes (@@@var{nn}) will be stripped from
2438 symbols before they are exported.
2439 [This option is specific to the i386 PE targeted port of the linker]
2441 @kindex --large-address-aware
2442 @item --large-address-aware
2443 If given, the appropriate bit in the ``Characteristics'' field of the COFF
2444 header is set to indicate that this executable supports virtual addresses
2445 greater than 2 gigabytes. This should be used in conjunction with the /3GB
2446 or /USERVA=@var{value} megabytes switch in the ``[operating systems]''
2447 section of the BOOT.INI. Otherwise, this bit has no effect.
2448 [This option is specific to PE targeted ports of the linker]
2450 @kindex --disable-large-address-aware
2451 @item --disable-large-address-aware
2452 Reverts the effect of a previous @samp{--large-address-aware} option.
2453 This is useful if @samp{--large-address-aware} is always set by the compiler
2454 driver (e.g. Cygwin gcc) and the executable does not support virtual
2455 addresses greater than 2 gigabytes.
2456 [This option is specific to PE targeted ports of the linker]
2458 @kindex --major-image-version
2459 @item --major-image-version @var{value}
2460 Sets the major number of the ``image version''. Defaults to 1.
2461 [This option is specific to the i386 PE targeted port of the linker]
2463 @kindex --major-os-version
2464 @item --major-os-version @var{value}
2465 Sets the major number of the ``os version''. Defaults to 4.
2466 [This option is specific to the i386 PE targeted port of the linker]
2468 @kindex --major-subsystem-version
2469 @item --major-subsystem-version @var{value}
2470 Sets the major number of the ``subsystem version''. Defaults to 4.
2471 [This option is specific to the i386 PE targeted port of the linker]
2473 @kindex --minor-image-version
2474 @item --minor-image-version @var{value}
2475 Sets the minor number of the ``image version''. Defaults to 0.
2476 [This option is specific to the i386 PE targeted port of the linker]
2478 @kindex --minor-os-version
2479 @item --minor-os-version @var{value}
2480 Sets the minor number of the ``os version''. Defaults to 0.
2481 [This option is specific to the i386 PE targeted port of the linker]
2483 @kindex --minor-subsystem-version
2484 @item --minor-subsystem-version @var{value}
2485 Sets the minor number of the ``subsystem version''. Defaults to 0.
2486 [This option is specific to the i386 PE targeted port of the linker]
2488 @cindex DEF files, creating
2489 @cindex DLLs, creating
2490 @kindex --output-def
2491 @item --output-def @var{file}
2492 The linker will create the file @var{file} which will contain a DEF
2493 file corresponding to the DLL the linker is generating. This DEF file
2494 (which should be called @code{*.def}) may be used to create an import
2495 library with @code{dlltool} or may be used as a reference to
2496 automatically or implicitly exported symbols.
2497 [This option is specific to the i386 PE targeted port of the linker]
2499 @cindex DLLs, creating
2500 @kindex --out-implib
2501 @item --out-implib @var{file}
2502 The linker will create the file @var{file} which will contain an
2503 import lib corresponding to the DLL the linker is generating. This
2504 import lib (which should be called @code{*.dll.a} or @code{*.a}
2505 may be used to link clients against the generated DLL; this behaviour
2506 makes it possible to skip a separate @code{dlltool} import library
2508 [This option is specific to the i386 PE targeted port of the linker]
2510 @kindex --enable-auto-image-base
2511 @item --enable-auto-image-base
2512 @itemx --enable-auto-image-base=@var{value}
2513 Automatically choose the image base for DLLs, optionally starting with base
2514 @var{value}, unless one is specified using the @code{--image-base} argument.
2515 By using a hash generated from the dllname to create unique image bases
2516 for each DLL, in-memory collisions and relocations which can delay program
2517 execution are avoided.
2518 [This option is specific to the i386 PE targeted port of the linker]
2520 @kindex --disable-auto-image-base
2521 @item --disable-auto-image-base
2522 Do not automatically generate a unique image base. If there is no
2523 user-specified image base (@code{--image-base}) then use the platform
2525 [This option is specific to the i386 PE targeted port of the linker]
2527 @cindex DLLs, linking to
2528 @kindex --dll-search-prefix
2529 @item --dll-search-prefix @var{string}
2530 When linking dynamically to a dll without an import library,
2531 search for @code{<string><basename>.dll} in preference to
2532 @code{lib<basename>.dll}. This behaviour allows easy distinction
2533 between DLLs built for the various "subplatforms": native, cygwin,
2534 uwin, pw, etc. For instance, cygwin DLLs typically use
2535 @code{--dll-search-prefix=cyg}.
2536 [This option is specific to the i386 PE targeted port of the linker]
2538 @kindex --enable-auto-import
2539 @item --enable-auto-import
2540 Do sophisticated linking of @code{_symbol} to @code{__imp__symbol} for
2541 DATA imports from DLLs, and create the necessary thunking symbols when
2542 building the import libraries with those DATA exports. Note: Use of the
2543 'auto-import' extension will cause the text section of the image file
2544 to be made writable. This does not conform to the PE-COFF format
2545 specification published by Microsoft.
2547 Note - use of the 'auto-import' extension will also cause read only
2548 data which would normally be placed into the .rdata section to be
2549 placed into the .data section instead. This is in order to work
2550 around a problem with consts that is described here:
2551 http://www.cygwin.com/ml/cygwin/2004-09/msg01101.html
2553 Using 'auto-import' generally will 'just work' -- but sometimes you may
2556 "variable '<var>' can't be auto-imported. Please read the
2557 documentation for ld's @code{--enable-auto-import} for details."
2559 This message occurs when some (sub)expression accesses an address
2560 ultimately given by the sum of two constants (Win32 import tables only
2561 allow one). Instances where this may occur include accesses to member
2562 fields of struct variables imported from a DLL, as well as using a
2563 constant index into an array variable imported from a DLL. Any
2564 multiword variable (arrays, structs, long long, etc) may trigger
2565 this error condition. However, regardless of the exact data type
2566 of the offending exported variable, ld will always detect it, issue
2567 the warning, and exit.
2569 There are several ways to address this difficulty, regardless of the
2570 data type of the exported variable:
2572 One way is to use --enable-runtime-pseudo-reloc switch. This leaves the task
2573 of adjusting references in your client code for runtime environment, so
2574 this method works only when runtime environment supports this feature.
2576 A second solution is to force one of the 'constants' to be a variable --
2577 that is, unknown and un-optimizable at compile time. For arrays,
2578 there are two possibilities: a) make the indexee (the array's address)
2579 a variable, or b) make the 'constant' index a variable. Thus:
2582 extern type extern_array[];
2584 @{ volatile type *t=extern_array; t[1] @}
2590 extern type extern_array[];
2592 @{ volatile int t=1; extern_array[t] @}
2595 For structs (and most other multiword data types) the only option
2596 is to make the struct itself (or the long long, or the ...) variable:
2599 extern struct s extern_struct;
2600 extern_struct.field -->
2601 @{ volatile struct s *t=&extern_struct; t->field @}
2607 extern long long extern_ll;
2609 @{ volatile long long * local_ll=&extern_ll; *local_ll @}
2612 A third method of dealing with this difficulty is to abandon
2613 'auto-import' for the offending symbol and mark it with
2614 @code{__declspec(dllimport)}. However, in practice that
2615 requires using compile-time #defines to indicate whether you are
2616 building a DLL, building client code that will link to the DLL, or
2617 merely building/linking to a static library. In making the choice
2618 between the various methods of resolving the 'direct address with
2619 constant offset' problem, you should consider typical real-world usage:
2627 void main(int argc, char **argv)@{
2628 printf("%d\n",arr[1]);
2638 void main(int argc, char **argv)@{
2639 /* This workaround is for win32 and cygwin; do not "optimize" */
2640 volatile int *parr = arr;
2641 printf("%d\n",parr[1]);
2648 /* Note: auto-export is assumed (no __declspec(dllexport)) */
2649 #if (defined(_WIN32) || defined(__CYGWIN__)) && \
2650 !(defined(FOO_BUILD_DLL) || defined(FOO_STATIC))
2651 #define FOO_IMPORT __declspec(dllimport)
2655 extern FOO_IMPORT int arr[];
2658 void main(int argc, char **argv)@{
2659 printf("%d\n",arr[1]);
2663 A fourth way to avoid this problem is to re-code your
2664 library to use a functional interface rather than a data interface
2665 for the offending variables (e.g. set_foo() and get_foo() accessor
2667 [This option is specific to the i386 PE targeted port of the linker]
2669 @kindex --disable-auto-import
2670 @item --disable-auto-import
2671 Do not attempt to do sophisticated linking of @code{_symbol} to
2672 @code{__imp__symbol} for DATA imports from DLLs.
2673 [This option is specific to the i386 PE targeted port of the linker]
2675 @kindex --enable-runtime-pseudo-reloc
2676 @item --enable-runtime-pseudo-reloc
2677 If your code contains expressions described in --enable-auto-import section,
2678 that is, DATA imports from DLL with non-zero offset, this switch will create
2679 a vector of 'runtime pseudo relocations' which can be used by runtime
2680 environment to adjust references to such data in your client code.
2681 [This option is specific to the i386 PE targeted port of the linker]
2683 @kindex --disable-runtime-pseudo-reloc
2684 @item --disable-runtime-pseudo-reloc
2685 Do not create pseudo relocations for non-zero offset DATA imports from
2687 [This option is specific to the i386 PE targeted port of the linker]
2689 @kindex --enable-extra-pe-debug
2690 @item --enable-extra-pe-debug
2691 Show additional debug info related to auto-import symbol thunking.
2692 [This option is specific to the i386 PE targeted port of the linker]
2694 @kindex --section-alignment
2695 @item --section-alignment
2696 Sets the section alignment. Sections in memory will always begin at
2697 addresses which are a multiple of this number. Defaults to 0x1000.
2698 [This option is specific to the i386 PE targeted port of the linker]
2702 @item --stack @var{reserve}
2703 @itemx --stack @var{reserve},@var{commit}
2704 Specify the number of bytes of memory to reserve (and optionally commit)
2705 to be used as stack for this program. The default is 2MB reserved, 4K
2707 [This option is specific to the i386 PE targeted port of the linker]
2710 @item --subsystem @var{which}
2711 @itemx --subsystem @var{which}:@var{major}
2712 @itemx --subsystem @var{which}:@var{major}.@var{minor}
2713 Specifies the subsystem under which your program will execute. The
2714 legal values for @var{which} are @code{native}, @code{windows},
2715 @code{console}, @code{posix}, and @code{xbox}. You may optionally set
2716 the subsystem version also. Numeric values are also accepted for
2718 [This option is specific to the i386 PE targeted port of the linker]
2720 The following options set flags in the @code{DllCharacteristics} field
2721 of the PE file header:
2722 [These options are specific to PE targeted ports of the linker]
2724 @kindex --high-entropy-va
2725 @item --high-entropy-va
2726 Image is compatible with 64-bit address space layout randomization
2729 @kindex --dynamicbase
2731 The image base address may be relocated using address space layout
2732 randomization (ASLR). This feature was introduced with MS Windows
2733 Vista for i386 PE targets.
2735 @kindex --forceinteg
2737 Code integrity checks are enforced.
2741 The image is compatible with the Data Execution Prevention.
2742 This feature was introduced with MS Windows XP SP2 for i386 PE targets.
2744 @kindex --no-isolation
2745 @item --no-isolation
2746 Although the image understands isolation, do not isolate the image.
2750 The image does not use SEH. No SE handler may be called from
2755 Do not bind this image.
2759 The driver uses the MS Windows Driver Model.
2763 The image is Terminal Server aware.
2765 @kindex --insert-timestamp
2766 @item --insert-timestamp
2767 @itemx --no-insert-timestamp
2768 Insert a real timestamp into the image. This is the default behaviour
2769 as it matches legacy code and it means that the image will work with
2770 other, proprietary tools. The problem with this default is that it
2771 will result in slightly different images being produced each tiem the
2772 same sources are linked. The option @option{--no-insert-timestamp}
2773 can be used to insert a zero value for the timestamp, this ensuring
2774 that binaries produced from indentical sources will compare
2781 @subsection Options specific to C6X uClinux targets
2783 @c man begin OPTIONS
2785 The C6X uClinux target uses a binary format called DSBT to support shared
2786 libraries. Each shared library in the system needs to have a unique index;
2787 all executables use an index of 0.
2792 @item --dsbt-size @var{size}
2793 This option sets the number of entires in the DSBT of the current executable
2794 or shared library to @var{size}. The default is to create a table with 64
2797 @kindex --dsbt-index
2798 @item --dsbt-index @var{index}
2799 This option sets the DSBT index of the current executable or shared library
2800 to @var{index}. The default is 0, which is appropriate for generating
2801 executables. If a shared library is generated with a DSBT index of 0, the
2802 @code{R_C6000_DSBT_INDEX} relocs are copied into the output file.
2804 @kindex --no-merge-exidx-entries
2805 The @samp{--no-merge-exidx-entries} switch disables the merging of adjacent
2806 exidx entries in frame unwind info.
2814 @subsection Options specific to Motorola 68HC11 and 68HC12 targets
2816 @c man begin OPTIONS
2818 The 68HC11 and 68HC12 linkers support specific options to control the
2819 memory bank switching mapping and trampoline code generation.
2823 @kindex --no-trampoline
2824 @item --no-trampoline
2825 This option disables the generation of trampoline. By default a trampoline
2826 is generated for each far function which is called using a @code{jsr}
2827 instruction (this happens when a pointer to a far function is taken).
2829 @kindex --bank-window
2830 @item --bank-window @var{name}
2831 This option indicates to the linker the name of the memory region in
2832 the @samp{MEMORY} specification that describes the memory bank window.
2833 The definition of such region is then used by the linker to compute
2834 paging and addresses within the memory window.
2842 @subsection Options specific to Motorola 68K target
2844 @c man begin OPTIONS
2846 The following options are supported to control handling of GOT generation
2847 when linking for 68K targets.
2852 @item --got=@var{type}
2853 This option tells the linker which GOT generation scheme to use.
2854 @var{type} should be one of @samp{single}, @samp{negative},
2855 @samp{multigot} or @samp{target}. For more information refer to the
2856 Info entry for @file{ld}.
2864 @subsection Options specific to MIPS targets
2866 @c man begin OPTIONS
2868 The following options are supported to control microMIPS instruction
2869 generation when linking for MIPS targets.
2877 These options control the choice of microMIPS instructions used in code
2878 generated by the linker, such as that in the PLT or lazy binding stubs,
2879 or in relaxation. If @samp{--insn32} is used, then the linker only uses
2880 32-bit instruction encodings. By default or if @samp{--no-insn32} is
2881 used, all instruction encodings are used, including 16-bit ones where
2891 @section Environment Variables
2893 @c man begin ENVIRONMENT
2895 You can change the behaviour of @command{ld} with the environment variables
2896 @ifclear SingleFormat
2899 @code{LDEMULATION} and @code{COLLECT_NO_DEMANGLE}.
2901 @ifclear SingleFormat
2903 @cindex default input format
2904 @code{GNUTARGET} determines the input-file object format if you don't
2905 use @samp{-b} (or its synonym @samp{--format}). Its value should be one
2906 of the BFD names for an input format (@pxref{BFD}). If there is no
2907 @code{GNUTARGET} in the environment, @command{ld} uses the natural format
2908 of the target. If @code{GNUTARGET} is set to @code{default} then BFD
2909 attempts to discover the input format by examining binary input files;
2910 this method often succeeds, but there are potential ambiguities, since
2911 there is no method of ensuring that the magic number used to specify
2912 object-file formats is unique. However, the configuration procedure for
2913 BFD on each system places the conventional format for that system first
2914 in the search-list, so ambiguities are resolved in favor of convention.
2918 @cindex default emulation
2919 @cindex emulation, default
2920 @code{LDEMULATION} determines the default emulation if you don't use the
2921 @samp{-m} option. The emulation can affect various aspects of linker
2922 behaviour, particularly the default linker script. You can list the
2923 available emulations with the @samp{--verbose} or @samp{-V} options. If
2924 the @samp{-m} option is not used, and the @code{LDEMULATION} environment
2925 variable is not defined, the default emulation depends upon how the
2926 linker was configured.
2928 @kindex COLLECT_NO_DEMANGLE
2929 @cindex demangling, default
2930 Normally, the linker will default to demangling symbols. However, if
2931 @code{COLLECT_NO_DEMANGLE} is set in the environment, then it will
2932 default to not demangling symbols. This environment variable is used in
2933 a similar fashion by the @code{gcc} linker wrapper program. The default
2934 may be overridden by the @samp{--demangle} and @samp{--no-demangle}
2941 @chapter Linker Scripts
2944 @cindex linker scripts
2945 @cindex command files
2946 Every link is controlled by a @dfn{linker script}. This script is
2947 written in the linker command language.
2949 The main purpose of the linker script is to describe how the sections in
2950 the input files should be mapped into the output file, and to control
2951 the memory layout of the output file. Most linker scripts do nothing
2952 more than this. However, when necessary, the linker script can also
2953 direct the linker to perform many other operations, using the commands
2956 The linker always uses a linker script. If you do not supply one
2957 yourself, the linker will use a default script that is compiled into the
2958 linker executable. You can use the @samp{--verbose} command line option
2959 to display the default linker script. Certain command line options,
2960 such as @samp{-r} or @samp{-N}, will affect the default linker script.
2962 You may supply your own linker script by using the @samp{-T} command
2963 line option. When you do this, your linker script will replace the
2964 default linker script.
2966 You may also use linker scripts implicitly by naming them as input files
2967 to the linker, as though they were files to be linked. @xref{Implicit
2971 * Basic Script Concepts:: Basic Linker Script Concepts
2972 * Script Format:: Linker Script Format
2973 * Simple Example:: Simple Linker Script Example
2974 * Simple Commands:: Simple Linker Script Commands
2975 * Assignments:: Assigning Values to Symbols
2976 * SECTIONS:: SECTIONS Command
2977 * MEMORY:: MEMORY Command
2978 * PHDRS:: PHDRS Command
2979 * VERSION:: VERSION Command
2980 * Expressions:: Expressions in Linker Scripts
2981 * Implicit Linker Scripts:: Implicit Linker Scripts
2984 @node Basic Script Concepts
2985 @section Basic Linker Script Concepts
2986 @cindex linker script concepts
2987 We need to define some basic concepts and vocabulary in order to
2988 describe the linker script language.
2990 The linker combines input files into a single output file. The output
2991 file and each input file are in a special data format known as an
2992 @dfn{object file format}. Each file is called an @dfn{object file}.
2993 The output file is often called an @dfn{executable}, but for our
2994 purposes we will also call it an object file. Each object file has,
2995 among other things, a list of @dfn{sections}. We sometimes refer to a
2996 section in an input file as an @dfn{input section}; similarly, a section
2997 in the output file is an @dfn{output section}.
2999 Each section in an object file has a name and a size. Most sections
3000 also have an associated block of data, known as the @dfn{section
3001 contents}. A section may be marked as @dfn{loadable}, which means that
3002 the contents should be loaded into memory when the output file is run.
3003 A section with no contents may be @dfn{allocatable}, which means that an
3004 area in memory should be set aside, but nothing in particular should be
3005 loaded there (in some cases this memory must be zeroed out). A section
3006 which is neither loadable nor allocatable typically contains some sort
3007 of debugging information.
3009 Every loadable or allocatable output section has two addresses. The
3010 first is the @dfn{VMA}, or virtual memory address. This is the address
3011 the section will have when the output file is run. The second is the
3012 @dfn{LMA}, or load memory address. This is the address at which the
3013 section will be loaded. In most cases the two addresses will be the
3014 same. An example of when they might be different is when a data section
3015 is loaded into ROM, and then copied into RAM when the program starts up
3016 (this technique is often used to initialize global variables in a ROM
3017 based system). In this case the ROM address would be the LMA, and the
3018 RAM address would be the VMA.
3020 You can see the sections in an object file by using the @code{objdump}
3021 program with the @samp{-h} option.
3023 Every object file also has a list of @dfn{symbols}, known as the
3024 @dfn{symbol table}. A symbol may be defined or undefined. Each symbol
3025 has a name, and each defined symbol has an address, among other
3026 information. If you compile a C or C++ program into an object file, you
3027 will get a defined symbol for every defined function and global or
3028 static variable. Every undefined function or global variable which is
3029 referenced in the input file will become an undefined symbol.
3031 You can see the symbols in an object file by using the @code{nm}
3032 program, or by using the @code{objdump} program with the @samp{-t}
3036 @section Linker Script Format
3037 @cindex linker script format
3038 Linker scripts are text files.
3040 You write a linker script as a series of commands. Each command is
3041 either a keyword, possibly followed by arguments, or an assignment to a
3042 symbol. You may separate commands using semicolons. Whitespace is
3045 Strings such as file or format names can normally be entered directly.
3046 If the file name contains a character such as a comma which would
3047 otherwise serve to separate file names, you may put the file name in
3048 double quotes. There is no way to use a double quote character in a
3051 You may include comments in linker scripts just as in C, delimited by
3052 @samp{/*} and @samp{*/}. As in C, comments are syntactically equivalent
3055 @node Simple Example
3056 @section Simple Linker Script Example
3057 @cindex linker script example
3058 @cindex example of linker script
3059 Many linker scripts are fairly simple.
3061 The simplest possible linker script has just one command:
3062 @samp{SECTIONS}. You use the @samp{SECTIONS} command to describe the
3063 memory layout of the output file.
3065 The @samp{SECTIONS} command is a powerful command. Here we will
3066 describe a simple use of it. Let's assume your program consists only of
3067 code, initialized data, and uninitialized data. These will be in the
3068 @samp{.text}, @samp{.data}, and @samp{.bss} sections, respectively.
3069 Let's assume further that these are the only sections which appear in
3072 For this example, let's say that the code should be loaded at address
3073 0x10000, and that the data should start at address 0x8000000. Here is a
3074 linker script which will do that:
3079 .text : @{ *(.text) @}
3081 .data : @{ *(.data) @}
3082 .bss : @{ *(.bss) @}
3086 You write the @samp{SECTIONS} command as the keyword @samp{SECTIONS},
3087 followed by a series of symbol assignments and output section
3088 descriptions enclosed in curly braces.
3090 The first line inside the @samp{SECTIONS} command of the above example
3091 sets the value of the special symbol @samp{.}, which is the location
3092 counter. If you do not specify the address of an output section in some
3093 other way (other ways are described later), the address is set from the
3094 current value of the location counter. The location counter is then
3095 incremented by the size of the output section. At the start of the
3096 @samp{SECTIONS} command, the location counter has the value @samp{0}.
3098 The second line defines an output section, @samp{.text}. The colon is
3099 required syntax which may be ignored for now. Within the curly braces
3100 after the output section name, you list the names of the input sections
3101 which should be placed into this output section. The @samp{*} is a
3102 wildcard which matches any file name. The expression @samp{*(.text)}
3103 means all @samp{.text} input sections in all input files.
3105 Since the location counter is @samp{0x10000} when the output section
3106 @samp{.text} is defined, the linker will set the address of the
3107 @samp{.text} section in the output file to be @samp{0x10000}.
3109 The remaining lines define the @samp{.data} and @samp{.bss} sections in
3110 the output file. The linker will place the @samp{.data} output section
3111 at address @samp{0x8000000}. After the linker places the @samp{.data}
3112 output section, the value of the location counter will be
3113 @samp{0x8000000} plus the size of the @samp{.data} output section. The
3114 effect is that the linker will place the @samp{.bss} output section
3115 immediately after the @samp{.data} output section in memory.
3117 The linker will ensure that each output section has the required
3118 alignment, by increasing the location counter if necessary. In this
3119 example, the specified addresses for the @samp{.text} and @samp{.data}
3120 sections will probably satisfy any alignment constraints, but the linker
3121 may have to create a small gap between the @samp{.data} and @samp{.bss}
3124 That's it! That's a simple and complete linker script.
3126 @node Simple Commands
3127 @section Simple Linker Script Commands
3128 @cindex linker script simple commands
3129 In this section we describe the simple linker script commands.
3132 * Entry Point:: Setting the entry point
3133 * File Commands:: Commands dealing with files
3134 @ifclear SingleFormat
3135 * Format Commands:: Commands dealing with object file formats
3138 * REGION_ALIAS:: Assign alias names to memory regions
3139 * Miscellaneous Commands:: Other linker script commands
3143 @subsection Setting the Entry Point
3144 @kindex ENTRY(@var{symbol})
3145 @cindex start of execution
3146 @cindex first instruction
3148 The first instruction to execute in a program is called the @dfn{entry
3149 point}. You can use the @code{ENTRY} linker script command to set the
3150 entry point. The argument is a symbol name:
3155 There are several ways to set the entry point. The linker will set the
3156 entry point by trying each of the following methods in order, and
3157 stopping when one of them succeeds:
3160 the @samp{-e} @var{entry} command-line option;
3162 the @code{ENTRY(@var{symbol})} command in a linker script;
3164 the value of a target specific symbol, if it is defined; For many
3165 targets this is @code{start}, but PE and BeOS based systems for example
3166 check a list of possible entry symbols, matching the first one found.
3168 the address of the first byte of the @samp{.text} section, if present;
3170 The address @code{0}.
3174 @subsection Commands Dealing with Files
3175 @cindex linker script file commands
3176 Several linker script commands deal with files.
3179 @item INCLUDE @var{filename}
3180 @kindex INCLUDE @var{filename}
3181 @cindex including a linker script
3182 Include the linker script @var{filename} at this point. The file will
3183 be searched for in the current directory, and in any directory specified
3184 with the @option{-L} option. You can nest calls to @code{INCLUDE} up to
3187 You can place @code{INCLUDE} directives at the top level, in @code{MEMORY} or
3188 @code{SECTIONS} commands, or in output section descriptions.
3190 @item INPUT(@var{file}, @var{file}, @dots{})
3191 @itemx INPUT(@var{file} @var{file} @dots{})
3192 @kindex INPUT(@var{files})
3193 @cindex input files in linker scripts
3194 @cindex input object files in linker scripts
3195 @cindex linker script input object files
3196 The @code{INPUT} command directs the linker to include the named files
3197 in the link, as though they were named on the command line.
3199 For example, if you always want to include @file{subr.o} any time you do
3200 a link, but you can't be bothered to put it on every link command line,
3201 then you can put @samp{INPUT (subr.o)} in your linker script.
3203 In fact, if you like, you can list all of your input files in the linker
3204 script, and then invoke the linker with nothing but a @samp{-T} option.
3206 In case a @dfn{sysroot prefix} is configured, and the filename starts
3207 with the @samp{/} character, and the script being processed was
3208 located inside the @dfn{sysroot prefix}, the filename will be looked
3209 for in the @dfn{sysroot prefix}. Otherwise, the linker will try to
3210 open the file in the current directory. If it is not found, the
3211 linker will search through the archive library search path.
3212 The @dfn{sysroot prefix} can also be forced by specifying @code{=}
3213 as the first character in the filename path. See also the
3214 description of @samp{-L} in @ref{Options,,Command Line Options}.
3216 If you use @samp{INPUT (-l@var{file})}, @command{ld} will transform the
3217 name to @code{lib@var{file}.a}, as with the command line argument
3220 When you use the @code{INPUT} command in an implicit linker script, the
3221 files will be included in the link at the point at which the linker
3222 script file is included. This can affect archive searching.
3224 @item GROUP(@var{file}, @var{file}, @dots{})
3225 @itemx GROUP(@var{file} @var{file} @dots{})
3226 @kindex GROUP(@var{files})
3227 @cindex grouping input files
3228 The @code{GROUP} command is like @code{INPUT}, except that the named
3229 files should all be archives, and they are searched repeatedly until no
3230 new undefined references are created. See the description of @samp{-(}
3231 in @ref{Options,,Command Line Options}.
3233 @item AS_NEEDED(@var{file}, @var{file}, @dots{})
3234 @itemx AS_NEEDED(@var{file} @var{file} @dots{})
3235 @kindex AS_NEEDED(@var{files})
3236 This construct can appear only inside of the @code{INPUT} or @code{GROUP}
3237 commands, among other filenames. The files listed will be handled
3238 as if they appear directly in the @code{INPUT} or @code{GROUP} commands,
3239 with the exception of ELF shared libraries, that will be added only
3240 when they are actually needed. This construct essentially enables
3241 @option{--as-needed} option for all the files listed inside of it
3242 and restores previous @option{--as-needed} resp. @option{--no-as-needed}
3245 @item OUTPUT(@var{filename})
3246 @kindex OUTPUT(@var{filename})
3247 @cindex output file name in linker script
3248 The @code{OUTPUT} command names the output file. Using
3249 @code{OUTPUT(@var{filename})} in the linker script is exactly like using
3250 @samp{-o @var{filename}} on the command line (@pxref{Options,,Command
3251 Line Options}). If both are used, the command line option takes
3254 You can use the @code{OUTPUT} command to define a default name for the
3255 output file other than the usual default of @file{a.out}.
3257 @item SEARCH_DIR(@var{path})
3258 @kindex SEARCH_DIR(@var{path})
3259 @cindex library search path in linker script
3260 @cindex archive search path in linker script
3261 @cindex search path in linker script
3262 The @code{SEARCH_DIR} command adds @var{path} to the list of paths where
3263 @command{ld} looks for archive libraries. Using
3264 @code{SEARCH_DIR(@var{path})} is exactly like using @samp{-L @var{path}}
3265 on the command line (@pxref{Options,,Command Line Options}). If both
3266 are used, then the linker will search both paths. Paths specified using
3267 the command line option are searched first.
3269 @item STARTUP(@var{filename})
3270 @kindex STARTUP(@var{filename})
3271 @cindex first input file
3272 The @code{STARTUP} command is just like the @code{INPUT} command, except
3273 that @var{filename} will become the first input file to be linked, as
3274 though it were specified first on the command line. This may be useful
3275 when using a system in which the entry point is always the start of the
3279 @ifclear SingleFormat
3280 @node Format Commands
3281 @subsection Commands Dealing with Object File Formats
3282 A couple of linker script commands deal with object file formats.
3285 @item OUTPUT_FORMAT(@var{bfdname})
3286 @itemx OUTPUT_FORMAT(@var{default}, @var{big}, @var{little})
3287 @kindex OUTPUT_FORMAT(@var{bfdname})
3288 @cindex output file format in linker script
3289 The @code{OUTPUT_FORMAT} command names the BFD format to use for the
3290 output file (@pxref{BFD}). Using @code{OUTPUT_FORMAT(@var{bfdname})} is
3291 exactly like using @samp{--oformat @var{bfdname}} on the command line
3292 (@pxref{Options,,Command Line Options}). If both are used, the command
3293 line option takes precedence.
3295 You can use @code{OUTPUT_FORMAT} with three arguments to use different
3296 formats based on the @samp{-EB} and @samp{-EL} command line options.
3297 This permits the linker script to set the output format based on the
3300 If neither @samp{-EB} nor @samp{-EL} are used, then the output format
3301 will be the first argument, @var{default}. If @samp{-EB} is used, the
3302 output format will be the second argument, @var{big}. If @samp{-EL} is
3303 used, the output format will be the third argument, @var{little}.
3305 For example, the default linker script for the MIPS ELF target uses this
3308 OUTPUT_FORMAT(elf32-bigmips, elf32-bigmips, elf32-littlemips)
3310 This says that the default format for the output file is
3311 @samp{elf32-bigmips}, but if the user uses the @samp{-EL} command line
3312 option, the output file will be created in the @samp{elf32-littlemips}
3315 @item TARGET(@var{bfdname})
3316 @kindex TARGET(@var{bfdname})
3317 @cindex input file format in linker script
3318 The @code{TARGET} command names the BFD format to use when reading input
3319 files. It affects subsequent @code{INPUT} and @code{GROUP} commands.
3320 This command is like using @samp{-b @var{bfdname}} on the command line
3321 (@pxref{Options,,Command Line Options}). If the @code{TARGET} command
3322 is used but @code{OUTPUT_FORMAT} is not, then the last @code{TARGET}
3323 command is also used to set the format for the output file. @xref{BFD}.
3328 @subsection Assign alias names to memory regions
3329 @kindex REGION_ALIAS(@var{alias}, @var{region})
3330 @cindex region alias
3331 @cindex region names
3333 Alias names can be added to existing memory regions created with the
3334 @ref{MEMORY} command. Each name corresponds to at most one memory region.
3337 REGION_ALIAS(@var{alias}, @var{region})
3340 The @code{REGION_ALIAS} function creates an alias name @var{alias} for the
3341 memory region @var{region}. This allows a flexible mapping of output sections
3342 to memory regions. An example follows.
3344 Suppose we have an application for embedded systems which come with various
3345 memory storage devices. All have a general purpose, volatile memory @code{RAM}
3346 that allows code execution or data storage. Some may have a read-only,
3347 non-volatile memory @code{ROM} that allows code execution and read-only data
3348 access. The last variant is a read-only, non-volatile memory @code{ROM2} with
3349 read-only data access and no code execution capability. We have four output
3354 @code{.text} program code;
3356 @code{.rodata} read-only data;
3358 @code{.data} read-write initialized data;
3360 @code{.bss} read-write zero initialized data.
3363 The goal is to provide a linker command file that contains a system independent
3364 part defining the output sections and a system dependent part mapping the
3365 output sections to the memory regions available on the system. Our embedded
3366 systems come with three different memory setups @code{A}, @code{B} and
3368 @multitable @columnfractions .25 .25 .25 .25
3369 @item Section @tab Variant A @tab Variant B @tab Variant C
3370 @item .text @tab RAM @tab ROM @tab ROM
3371 @item .rodata @tab RAM @tab ROM @tab ROM2
3372 @item .data @tab RAM @tab RAM/ROM @tab RAM/ROM2
3373 @item .bss @tab RAM @tab RAM @tab RAM
3375 The notation @code{RAM/ROM} or @code{RAM/ROM2} means that this section is
3376 loaded into region @code{ROM} or @code{ROM2} respectively. Please note that
3377 the load address of the @code{.data} section starts in all three variants at
3378 the end of the @code{.rodata} section.
3380 The base linker script that deals with the output sections follows. It
3381 includes the system dependent @code{linkcmds.memory} file that describes the
3384 INCLUDE linkcmds.memory
3397 .data : AT (rodata_end)
3402 data_size = SIZEOF(.data);
3403 data_load_start = LOADADDR(.data);
3411 Now we need three different @code{linkcmds.memory} files to define memory
3412 regions and alias names. The content of @code{linkcmds.memory} for the three
3413 variants @code{A}, @code{B} and @code{C}:
3416 Here everything goes into the @code{RAM}.
3420 RAM : ORIGIN = 0, LENGTH = 4M
3423 REGION_ALIAS("REGION_TEXT", RAM);
3424 REGION_ALIAS("REGION_RODATA", RAM);
3425 REGION_ALIAS("REGION_DATA", RAM);
3426 REGION_ALIAS("REGION_BSS", RAM);
3429 Program code and read-only data go into the @code{ROM}. Read-write data goes
3430 into the @code{RAM}. An image of the initialized data is loaded into the
3431 @code{ROM} and will be copied during system start into the @code{RAM}.
3435 ROM : ORIGIN = 0, LENGTH = 3M
3436 RAM : ORIGIN = 0x10000000, LENGTH = 1M
3439 REGION_ALIAS("REGION_TEXT", ROM);
3440 REGION_ALIAS("REGION_RODATA", ROM);
3441 REGION_ALIAS("REGION_DATA", RAM);
3442 REGION_ALIAS("REGION_BSS", RAM);
3445 Program code goes into the @code{ROM}. Read-only data goes into the
3446 @code{ROM2}. Read-write data goes into the @code{RAM}. An image of the
3447 initialized data is loaded into the @code{ROM2} and will be copied during
3448 system start into the @code{RAM}.
3452 ROM : ORIGIN = 0, LENGTH = 2M
3453 ROM2 : ORIGIN = 0x10000000, LENGTH = 1M
3454 RAM : ORIGIN = 0x20000000, LENGTH = 1M
3457 REGION_ALIAS("REGION_TEXT", ROM);
3458 REGION_ALIAS("REGION_RODATA", ROM2);
3459 REGION_ALIAS("REGION_DATA", RAM);
3460 REGION_ALIAS("REGION_BSS", RAM);
3464 It is possible to write a common system initialization routine to copy the
3465 @code{.data} section from @code{ROM} or @code{ROM2} into the @code{RAM} if
3470 extern char data_start [];
3471 extern char data_size [];
3472 extern char data_load_start [];
3474 void copy_data(void)
3476 if (data_start != data_load_start)
3478 memcpy(data_start, data_load_start, (size_t) data_size);
3483 @node Miscellaneous Commands
3484 @subsection Other Linker Script Commands
3485 There are a few other linker scripts commands.
3488 @item ASSERT(@var{exp}, @var{message})
3490 @cindex assertion in linker script
3491 Ensure that @var{exp} is non-zero. If it is zero, then exit the linker
3492 with an error code, and print @var{message}.
3494 Note that assertions are checked before the final stages of linking
3495 take place. This means that expressions involving symbols PROVIDEd
3496 inside section definitions will fail if the user has not set values
3497 for those symbols. The only exception to this rule is PROVIDEd
3498 symbols that just reference dot. Thus an assertion like this:
3503 PROVIDE (__stack = .);
3504 PROVIDE (__stack_size = 0x100);
3505 ASSERT ((__stack > (_end + __stack_size)), "Error: No room left for the stack");
3509 will fail if @code{__stack_size} is not defined elsewhere. Symbols
3510 PROVIDEd outside of section definitions are evaluated earlier, so they
3511 can be used inside ASSERTions. Thus:
3514 PROVIDE (__stack_size = 0x100);
3517 PROVIDE (__stack = .);
3518 ASSERT ((__stack > (_end + __stack_size)), "Error: No room left for the stack");
3524 @item EXTERN(@var{symbol} @var{symbol} @dots{})
3526 @cindex undefined symbol in linker script
3527 Force @var{symbol} to be entered in the output file as an undefined
3528 symbol. Doing this may, for example, trigger linking of additional
3529 modules from standard libraries. You may list several @var{symbol}s for
3530 each @code{EXTERN}, and you may use @code{EXTERN} multiple times. This
3531 command has the same effect as the @samp{-u} command-line option.
3533 @item FORCE_COMMON_ALLOCATION
3534 @kindex FORCE_COMMON_ALLOCATION
3535 @cindex common allocation in linker script
3536 This command has the same effect as the @samp{-d} command-line option:
3537 to make @command{ld} assign space to common symbols even if a relocatable
3538 output file is specified (@samp{-r}).
3540 @item INHIBIT_COMMON_ALLOCATION
3541 @kindex INHIBIT_COMMON_ALLOCATION
3542 @cindex common allocation in linker script
3543 This command has the same effect as the @samp{--no-define-common}
3544 command-line option: to make @code{ld} omit the assignment of addresses
3545 to common symbols even for a non-relocatable output file.
3547 @item INSERT [ AFTER | BEFORE ] @var{output_section}
3549 @cindex insert user script into default script
3550 This command is typically used in a script specified by @samp{-T} to
3551 augment the default @code{SECTIONS} with, for example, overlays. It
3552 inserts all prior linker script statements after (or before)
3553 @var{output_section}, and also causes @samp{-T} to not override the
3554 default linker script. The exact insertion point is as for orphan
3555 sections. @xref{Location Counter}. The insertion happens after the
3556 linker has mapped input sections to output sections. Prior to the
3557 insertion, since @samp{-T} scripts are parsed before the default
3558 linker script, statements in the @samp{-T} script occur before the
3559 default linker script statements in the internal linker representation
3560 of the script. In particular, input section assignments will be made
3561 to @samp{-T} output sections before those in the default script. Here
3562 is an example of how a @samp{-T} script using @code{INSERT} might look:
3569 .ov1 @{ ov1*(.text) @}
3570 .ov2 @{ ov2*(.text) @}
3576 @item NOCROSSREFS(@var{section} @var{section} @dots{})
3577 @kindex NOCROSSREFS(@var{sections})
3578 @cindex cross references
3579 This command may be used to tell @command{ld} to issue an error about any
3580 references among certain output sections.
3582 In certain types of programs, particularly on embedded systems when
3583 using overlays, when one section is loaded into memory, another section
3584 will not be. Any direct references between the two sections would be
3585 errors. For example, it would be an error if code in one section called
3586 a function defined in the other section.
3588 The @code{NOCROSSREFS} command takes a list of output section names. If
3589 @command{ld} detects any cross references between the sections, it reports
3590 an error and returns a non-zero exit status. Note that the
3591 @code{NOCROSSREFS} command uses output section names, not input section
3594 @ifclear SingleFormat
3595 @item OUTPUT_ARCH(@var{bfdarch})
3596 @kindex OUTPUT_ARCH(@var{bfdarch})
3597 @cindex machine architecture
3598 @cindex architecture
3599 Specify a particular output machine architecture. The argument is one
3600 of the names used by the BFD library (@pxref{BFD}). You can see the
3601 architecture of an object file by using the @code{objdump} program with
3602 the @samp{-f} option.
3605 @item LD_FEATURE(@var{string})
3606 @kindex LD_FEATURE(@var{string})
3607 This command may be used to modify @command{ld} behavior. If
3608 @var{string} is @code{"SANE_EXPR"} then absolute symbols and numbers
3609 in a script are simply treated as numbers everywhere.
3610 @xref{Expression Section}.
3614 @section Assigning Values to Symbols
3615 @cindex assignment in scripts
3616 @cindex symbol definition, scripts
3617 @cindex variables, defining
3618 You may assign a value to a symbol in a linker script. This will define
3619 the symbol and place it into the symbol table with a global scope.
3622 * Simple Assignments:: Simple Assignments
3625 * PROVIDE_HIDDEN:: PROVIDE_HIDDEN
3626 * Source Code Reference:: How to use a linker script defined symbol in source code
3629 @node Simple Assignments
3630 @subsection Simple Assignments
3632 You may assign to a symbol using any of the C assignment operators:
3635 @item @var{symbol} = @var{expression} ;
3636 @itemx @var{symbol} += @var{expression} ;
3637 @itemx @var{symbol} -= @var{expression} ;
3638 @itemx @var{symbol} *= @var{expression} ;
3639 @itemx @var{symbol} /= @var{expression} ;
3640 @itemx @var{symbol} <<= @var{expression} ;
3641 @itemx @var{symbol} >>= @var{expression} ;
3642 @itemx @var{symbol} &= @var{expression} ;
3643 @itemx @var{symbol} |= @var{expression} ;
3646 The first case will define @var{symbol} to the value of
3647 @var{expression}. In the other cases, @var{symbol} must already be
3648 defined, and the value will be adjusted accordingly.
3650 The special symbol name @samp{.} indicates the location counter. You
3651 may only use this within a @code{SECTIONS} command. @xref{Location Counter}.
3653 The semicolon after @var{expression} is required.
3655 Expressions are defined below; see @ref{Expressions}.
3657 You may write symbol assignments as commands in their own right, or as
3658 statements within a @code{SECTIONS} command, or as part of an output
3659 section description in a @code{SECTIONS} command.
3661 The section of the symbol will be set from the section of the
3662 expression; for more information, see @ref{Expression Section}.
3664 Here is an example showing the three different places that symbol
3665 assignments may be used:
3676 _bdata = (. + 3) & ~ 3;
3677 .data : @{ *(.data) @}
3681 In this example, the symbol @samp{floating_point} will be defined as
3682 zero. The symbol @samp{_etext} will be defined as the address following
3683 the last @samp{.text} input section. The symbol @samp{_bdata} will be
3684 defined as the address following the @samp{.text} output section aligned
3685 upward to a 4 byte boundary.
3690 For ELF targeted ports, define a symbol that will be hidden and won't be
3691 exported. The syntax is @code{HIDDEN(@var{symbol} = @var{expression})}.
3693 Here is the example from @ref{Simple Assignments}, rewritten to use
3697 HIDDEN(floating_point = 0);
3705 HIDDEN(_bdata = (. + 3) & ~ 3);
3706 .data : @{ *(.data) @}
3710 In this case none of the three symbols will be visible outside this module.
3715 In some cases, it is desirable for a linker script to define a symbol
3716 only if it is referenced and is not defined by any object included in
3717 the link. For example, traditional linkers defined the symbol
3718 @samp{etext}. However, ANSI C requires that the user be able to use
3719 @samp{etext} as a function name without encountering an error. The
3720 @code{PROVIDE} keyword may be used to define a symbol, such as
3721 @samp{etext}, only if it is referenced but not defined. The syntax is
3722 @code{PROVIDE(@var{symbol} = @var{expression})}.
3724 Here is an example of using @code{PROVIDE} to define @samp{etext}:
3737 In this example, if the program defines @samp{_etext} (with a leading
3738 underscore), the linker will give a multiple definition error. If, on
3739 the other hand, the program defines @samp{etext} (with no leading
3740 underscore), the linker will silently use the definition in the program.
3741 If the program references @samp{etext} but does not define it, the
3742 linker will use the definition in the linker script.
3744 @node PROVIDE_HIDDEN
3745 @subsection PROVIDE_HIDDEN
3746 @cindex PROVIDE_HIDDEN
3747 Similar to @code{PROVIDE}. For ELF targeted ports, the symbol will be
3748 hidden and won't be exported.
3750 @node Source Code Reference
3751 @subsection Source Code Reference
3753 Accessing a linker script defined variable from source code is not
3754 intuitive. In particular a linker script symbol is not equivalent to
3755 a variable declaration in a high level language, it is instead a
3756 symbol that does not have a value.
3758 Before going further, it is important to note that compilers often
3759 transform names in the source code into different names when they are
3760 stored in the symbol table. For example, Fortran compilers commonly
3761 prepend or append an underscore, and C++ performs extensive @samp{name
3762 mangling}. Therefore there might be a discrepancy between the name
3763 of a variable as it is used in source code and the name of the same
3764 variable as it is defined in a linker script. For example in C a
3765 linker script variable might be referred to as:
3771 But in the linker script it might be defined as:
3777 In the remaining examples however it is assumed that no name
3778 transformation has taken place.
3780 When a symbol is declared in a high level language such as C, two
3781 things happen. The first is that the compiler reserves enough space
3782 in the program's memory to hold the @emph{value} of the symbol. The
3783 second is that the compiler creates an entry in the program's symbol
3784 table which holds the symbol's @emph{address}. ie the symbol table
3785 contains the address of the block of memory holding the symbol's
3786 value. So for example the following C declaration, at file scope:
3792 creates an entry called @samp{foo} in the symbol table. This entry
3793 holds the address of an @samp{int} sized block of memory where the
3794 number 1000 is initially stored.
3796 When a program references a symbol the compiler generates code that
3797 first accesses the symbol table to find the address of the symbol's
3798 memory block and then code to read the value from that memory block.
3805 looks up the symbol @samp{foo} in the symbol table, gets the address
3806 associated with this symbol and then writes the value 1 into that
3813 looks up the symbol @samp{foo} in the symbol table, gets its address
3814 and then copies this address into the block of memory associated with
3815 the variable @samp{a}.
3817 Linker scripts symbol declarations, by contrast, create an entry in
3818 the symbol table but do not assign any memory to them. Thus they are
3819 an address without a value. So for example the linker script definition:
3825 creates an entry in the symbol table called @samp{foo} which holds
3826 the address of memory location 1000, but nothing special is stored at
3827 address 1000. This means that you cannot access the @emph{value} of a
3828 linker script defined symbol - it has no value - all you can do is
3829 access the @emph{address} of a linker script defined symbol.
3831 Hence when you are using a linker script defined symbol in source code
3832 you should always take the address of the symbol, and never attempt to
3833 use its value. For example suppose you want to copy the contents of a
3834 section of memory called .ROM into a section called .FLASH and the
3835 linker script contains these declarations:
3839 start_of_ROM = .ROM;
3840 end_of_ROM = .ROM + sizeof (.ROM) - 1;
3841 start_of_FLASH = .FLASH;
3845 Then the C source code to perform the copy would be:
3849 extern char start_of_ROM, end_of_ROM, start_of_FLASH;
3851 memcpy (& start_of_FLASH, & start_of_ROM, & end_of_ROM - & start_of_ROM);
3855 Note the use of the @samp{&} operators. These are correct.
3858 @section SECTIONS Command
3860 The @code{SECTIONS} command tells the linker how to map input sections
3861 into output sections, and how to place the output sections in memory.
3863 The format of the @code{SECTIONS} command is:
3867 @var{sections-command}
3868 @var{sections-command}
3873 Each @var{sections-command} may of be one of the following:
3877 an @code{ENTRY} command (@pxref{Entry Point,,Entry command})
3879 a symbol assignment (@pxref{Assignments})
3881 an output section description
3883 an overlay description
3886 The @code{ENTRY} command and symbol assignments are permitted inside the
3887 @code{SECTIONS} command for convenience in using the location counter in
3888 those commands. This can also make the linker script easier to
3889 understand because you can use those commands at meaningful points in
3890 the layout of the output file.
3892 Output section descriptions and overlay descriptions are described
3895 If you do not use a @code{SECTIONS} command in your linker script, the
3896 linker will place each input section into an identically named output
3897 section in the order that the sections are first encountered in the
3898 input files. If all input sections are present in the first file, for
3899 example, the order of sections in the output file will match the order
3900 in the first input file. The first section will be at address zero.
3903 * Output Section Description:: Output section description
3904 * Output Section Name:: Output section name
3905 * Output Section Address:: Output section address
3906 * Input Section:: Input section description
3907 * Output Section Data:: Output section data
3908 * Output Section Keywords:: Output section keywords
3909 * Output Section Discarding:: Output section discarding
3910 * Output Section Attributes:: Output section attributes
3911 * Overlay Description:: Overlay description
3914 @node Output Section Description
3915 @subsection Output Section Description
3916 The full description of an output section looks like this:
3919 @var{section} [@var{address}] [(@var{type})] :
3921 [ALIGN(@var{section_align}) | ALIGN_WITH_INPUT]
3922 [SUBALIGN(@var{subsection_align})]
3925 @var{output-section-command}
3926 @var{output-section-command}
3928 @} [>@var{region}] [AT>@var{lma_region}] [:@var{phdr} :@var{phdr} @dots{}] [=@var{fillexp}] [,]
3932 Most output sections do not use most of the optional section attributes.
3934 The whitespace around @var{section} is required, so that the section
3935 name is unambiguous. The colon and the curly braces are also required.
3936 The comma at the end may be required if a @var{fillexp} is used and
3937 the next @var{sections-command} looks like a continuation of the expression.
3938 The line breaks and other white space are optional.
3940 Each @var{output-section-command} may be one of the following:
3944 a symbol assignment (@pxref{Assignments})
3946 an input section description (@pxref{Input Section})
3948 data values to include directly (@pxref{Output Section Data})
3950 a special output section keyword (@pxref{Output Section Keywords})
3953 @node Output Section Name
3954 @subsection Output Section Name
3955 @cindex name, section
3956 @cindex section name
3957 The name of the output section is @var{section}. @var{section} must
3958 meet the constraints of your output format. In formats which only
3959 support a limited number of sections, such as @code{a.out}, the name
3960 must be one of the names supported by the format (@code{a.out}, for
3961 example, allows only @samp{.text}, @samp{.data} or @samp{.bss}). If the
3962 output format supports any number of sections, but with numbers and not
3963 names (as is the case for Oasys), the name should be supplied as a
3964 quoted numeric string. A section name may consist of any sequence of
3965 characters, but a name which contains any unusual characters such as
3966 commas must be quoted.
3968 The output section name @samp{/DISCARD/} is special; @ref{Output Section
3971 @node Output Section Address
3972 @subsection Output Section Address
3973 @cindex address, section
3974 @cindex section address
3975 The @var{address} is an expression for the VMA (the virtual memory
3976 address) of the output section. This address is optional, but if it
3977 is provided then the output address will be set exactly as specified.
3979 If the output address is not specified then one will be chosen for the
3980 section, based on the heuristic below. This address will be adjusted
3981 to fit the alignment requirement of the output section. The
3982 alignment requirement is the strictest alignment of any input section
3983 contained within the output section.
3985 The output section address heuristic is as follows:
3989 If an output memory @var{region} is set for the section then it
3990 is added to this region and its address will be the next free address
3994 If the MEMORY command has been used to create a list of memory
3995 regions then the first region which has attributes compatible with the
3996 section is selected to contain it. The section's output address will
3997 be the next free address in that region; @ref{MEMORY}.
4000 If no memory regions were specified, or none match the section then
4001 the output address will be based on the current value of the location
4009 .text . : @{ *(.text) @}
4016 .text : @{ *(.text) @}
4020 are subtly different. The first will set the address of the
4021 @samp{.text} output section to the current value of the location
4022 counter. The second will set it to the current value of the location
4023 counter aligned to the strictest alignment of any of the @samp{.text}
4026 The @var{address} may be an arbitrary expression; @ref{Expressions}.
4027 For example, if you want to align the section on a 0x10 byte boundary,
4028 so that the lowest four bits of the section address are zero, you could
4029 do something like this:
4031 .text ALIGN(0x10) : @{ *(.text) @}
4034 This works because @code{ALIGN} returns the current location counter
4035 aligned upward to the specified value.
4037 Specifying @var{address} for a section will change the value of the
4038 location counter, provided that the section is non-empty. (Empty
4039 sections are ignored).
4042 @subsection Input Section Description
4043 @cindex input sections
4044 @cindex mapping input sections to output sections
4045 The most common output section command is an input section description.
4047 The input section description is the most basic linker script operation.
4048 You use output sections to tell the linker how to lay out your program
4049 in memory. You use input section descriptions to tell the linker how to
4050 map the input files into your memory layout.
4053 * Input Section Basics:: Input section basics
4054 * Input Section Wildcards:: Input section wildcard patterns
4055 * Input Section Common:: Input section for common symbols
4056 * Input Section Keep:: Input section and garbage collection
4057 * Input Section Example:: Input section example
4060 @node Input Section Basics
4061 @subsubsection Input Section Basics
4062 @cindex input section basics
4063 An input section description consists of a file name optionally followed
4064 by a list of section names in parentheses.
4066 The file name and the section name may be wildcard patterns, which we
4067 describe further below (@pxref{Input Section Wildcards}).
4069 The most common input section description is to include all input
4070 sections with a particular name in the output section. For example, to
4071 include all input @samp{.text} sections, you would write:
4076 Here the @samp{*} is a wildcard which matches any file name. To exclude a list
4077 of files from matching the file name wildcard, EXCLUDE_FILE may be used to
4078 match all files except the ones specified in the EXCLUDE_FILE list. For
4081 *(EXCLUDE_FILE (*crtend.o *otherfile.o) .ctors)
4083 will cause all .ctors sections from all files except @file{crtend.o} and
4084 @file{otherfile.o} to be included.
4086 There are two ways to include more than one section:
4092 The difference between these is the order in which the @samp{.text} and
4093 @samp{.rdata} input sections will appear in the output section. In the
4094 first example, they will be intermingled, appearing in the same order as
4095 they are found in the linker input. In the second example, all
4096 @samp{.text} input sections will appear first, followed by all
4097 @samp{.rdata} input sections.
4099 You can specify a file name to include sections from a particular file.
4100 You would do this if one or more of your files contain special data that
4101 needs to be at a particular location in memory. For example:
4106 To refine the sections that are included based on the section flags
4107 of an input section, INPUT_SECTION_FLAGS may be used.
4109 Here is a simple example for using Section header flags for ELF sections:
4114 .text : @{ INPUT_SECTION_FLAGS (SHF_MERGE & SHF_STRINGS) *(.text) @}
4115 .text2 : @{ INPUT_SECTION_FLAGS (!SHF_WRITE) *(.text) @}
4120 In this example, the output section @samp{.text} will be comprised of any
4121 input section matching the name *(.text) whose section header flags
4122 @code{SHF_MERGE} and @code{SHF_STRINGS} are set. The output section
4123 @samp{.text2} will be comprised of any input section matching the name *(.text)
4124 whose section header flag @code{SHF_WRITE} is clear.
4126 You can also specify files within archives by writing a pattern
4127 matching the archive, a colon, then the pattern matching the file,
4128 with no whitespace around the colon.
4132 matches file within archive
4134 matches the whole archive
4136 matches file but not one in an archive
4139 Either one or both of @samp{archive} and @samp{file} can contain shell
4140 wildcards. On DOS based file systems, the linker will assume that a
4141 single letter followed by a colon is a drive specifier, so
4142 @samp{c:myfile.o} is a simple file specification, not @samp{myfile.o}
4143 within an archive called @samp{c}. @samp{archive:file} filespecs may
4144 also be used within an @code{EXCLUDE_FILE} list, but may not appear in
4145 other linker script contexts. For instance, you cannot extract a file
4146 from an archive by using @samp{archive:file} in an @code{INPUT}
4149 If you use a file name without a list of sections, then all sections in
4150 the input file will be included in the output section. This is not
4151 commonly done, but it may by useful on occasion. For example:
4156 When you use a file name which is not an @samp{archive:file} specifier
4157 and does not contain any wild card
4158 characters, the linker will first see if you also specified the file
4159 name on the linker command line or in an @code{INPUT} command. If you
4160 did not, the linker will attempt to open the file as an input file, as
4161 though it appeared on the command line. Note that this differs from an
4162 @code{INPUT} command, because the linker will not search for the file in
4163 the archive search path.
4165 @node Input Section Wildcards
4166 @subsubsection Input Section Wildcard Patterns
4167 @cindex input section wildcards
4168 @cindex wildcard file name patterns
4169 @cindex file name wildcard patterns
4170 @cindex section name wildcard patterns
4171 In an input section description, either the file name or the section
4172 name or both may be wildcard patterns.
4174 The file name of @samp{*} seen in many examples is a simple wildcard
4175 pattern for the file name.
4177 The wildcard patterns are like those used by the Unix shell.
4181 matches any number of characters
4183 matches any single character
4185 matches a single instance of any of the @var{chars}; the @samp{-}
4186 character may be used to specify a range of characters, as in
4187 @samp{[a-z]} to match any lower case letter
4189 quotes the following character
4192 When a file name is matched with a wildcard, the wildcard characters
4193 will not match a @samp{/} character (used to separate directory names on
4194 Unix). A pattern consisting of a single @samp{*} character is an
4195 exception; it will always match any file name, whether it contains a
4196 @samp{/} or not. In a section name, the wildcard characters will match
4197 a @samp{/} character.
4199 File name wildcard patterns only match files which are explicitly
4200 specified on the command line or in an @code{INPUT} command. The linker
4201 does not search directories to expand wildcards.
4203 If a file name matches more than one wildcard pattern, or if a file name
4204 appears explicitly and is also matched by a wildcard pattern, the linker
4205 will use the first match in the linker script. For example, this
4206 sequence of input section descriptions is probably in error, because the
4207 @file{data.o} rule will not be used:
4209 .data : @{ *(.data) @}
4210 .data1 : @{ data.o(.data) @}
4213 @cindex SORT_BY_NAME
4214 Normally, the linker will place files and sections matched by wildcards
4215 in the order in which they are seen during the link. You can change
4216 this by using the @code{SORT_BY_NAME} keyword, which appears before a wildcard
4217 pattern in parentheses (e.g., @code{SORT_BY_NAME(.text*)}). When the
4218 @code{SORT_BY_NAME} keyword is used, the linker will sort the files or sections
4219 into ascending order by name before placing them in the output file.
4221 @cindex SORT_BY_ALIGNMENT
4222 @code{SORT_BY_ALIGNMENT} is very similar to @code{SORT_BY_NAME}. The
4223 difference is @code{SORT_BY_ALIGNMENT} will sort sections into
4224 descending order by alignment before placing them in the output file.
4225 Larger alignments are placed before smaller alignments in order to
4226 reduce the amount of padding necessary.
4228 @cindex SORT_BY_INIT_PRIORITY
4229 @code{SORT_BY_INIT_PRIORITY} is very similar to @code{SORT_BY_NAME}. The
4230 difference is @code{SORT_BY_INIT_PRIORITY} will sort sections into
4231 ascending order by numerical value of the GCC init_priority attribute
4232 encoded in the section name before placing them in the output file.
4235 @code{SORT} is an alias for @code{SORT_BY_NAME}.
4237 When there are nested section sorting commands in linker script, there
4238 can be at most 1 level of nesting for section sorting commands.
4242 @code{SORT_BY_NAME} (@code{SORT_BY_ALIGNMENT} (wildcard section pattern)).
4243 It will sort the input sections by name first, then by alignment if two
4244 sections have the same name.
4246 @code{SORT_BY_ALIGNMENT} (@code{SORT_BY_NAME} (wildcard section pattern)).
4247 It will sort the input sections by alignment first, then by name if two
4248 sections have the same alignment.
4250 @code{SORT_BY_NAME} (@code{SORT_BY_NAME} (wildcard section pattern)) is
4251 treated the same as @code{SORT_BY_NAME} (wildcard section pattern).
4253 @code{SORT_BY_ALIGNMENT} (@code{SORT_BY_ALIGNMENT} (wildcard section pattern))
4254 is treated the same as @code{SORT_BY_ALIGNMENT} (wildcard section pattern).
4256 All other nested section sorting commands are invalid.
4259 When both command line section sorting option and linker script
4260 section sorting command are used, section sorting command always
4261 takes precedence over the command line option.
4263 If the section sorting command in linker script isn't nested, the
4264 command line option will make the section sorting command to be
4265 treated as nested sorting command.
4269 @code{SORT_BY_NAME} (wildcard section pattern ) with
4270 @option{--sort-sections alignment} is equivalent to
4271 @code{SORT_BY_NAME} (@code{SORT_BY_ALIGNMENT} (wildcard section pattern)).
4273 @code{SORT_BY_ALIGNMENT} (wildcard section pattern) with
4274 @option{--sort-section name} is equivalent to
4275 @code{SORT_BY_ALIGNMENT} (@code{SORT_BY_NAME} (wildcard section pattern)).
4278 If the section sorting command in linker script is nested, the
4279 command line option will be ignored.
4282 @code{SORT_NONE} disables section sorting by ignoring the command line
4283 section sorting option.
4285 If you ever get confused about where input sections are going, use the
4286 @samp{-M} linker option to generate a map file. The map file shows
4287 precisely how input sections are mapped to output sections.
4289 This example shows how wildcard patterns might be used to partition
4290 files. This linker script directs the linker to place all @samp{.text}
4291 sections in @samp{.text} and all @samp{.bss} sections in @samp{.bss}.
4292 The linker will place the @samp{.data} section from all files beginning
4293 with an upper case character in @samp{.DATA}; for all other files, the
4294 linker will place the @samp{.data} section in @samp{.data}.
4298 .text : @{ *(.text) @}
4299 .DATA : @{ [A-Z]*(.data) @}
4300 .data : @{ *(.data) @}
4301 .bss : @{ *(.bss) @}
4306 @node Input Section Common
4307 @subsubsection Input Section for Common Symbols
4308 @cindex common symbol placement
4309 @cindex uninitialized data placement
4310 A special notation is needed for common symbols, because in many object
4311 file formats common symbols do not have a particular input section. The
4312 linker treats common symbols as though they are in an input section
4313 named @samp{COMMON}.
4315 You may use file names with the @samp{COMMON} section just as with any
4316 other input sections. You can use this to place common symbols from a
4317 particular input file in one section while common symbols from other
4318 input files are placed in another section.
4320 In most cases, common symbols in input files will be placed in the
4321 @samp{.bss} section in the output file. For example:
4323 .bss @{ *(.bss) *(COMMON) @}
4326 @cindex scommon section
4327 @cindex small common symbols
4328 Some object file formats have more than one type of common symbol. For
4329 example, the MIPS ELF object file format distinguishes standard common
4330 symbols and small common symbols. In this case, the linker will use a
4331 different special section name for other types of common symbols. In
4332 the case of MIPS ELF, the linker uses @samp{COMMON} for standard common
4333 symbols and @samp{.scommon} for small common symbols. This permits you
4334 to map the different types of common symbols into memory at different
4338 You will sometimes see @samp{[COMMON]} in old linker scripts. This
4339 notation is now considered obsolete. It is equivalent to
4342 @node Input Section Keep
4343 @subsubsection Input Section and Garbage Collection
4345 @cindex garbage collection
4346 When link-time garbage collection is in use (@samp{--gc-sections}),
4347 it is often useful to mark sections that should not be eliminated.
4348 This is accomplished by surrounding an input section's wildcard entry
4349 with @code{KEEP()}, as in @code{KEEP(*(.init))} or
4350 @code{KEEP(SORT_BY_NAME(*)(.ctors))}.
4352 @node Input Section Example
4353 @subsubsection Input Section Example
4354 The following example is a complete linker script. It tells the linker
4355 to read all of the sections from file @file{all.o} and place them at the
4356 start of output section @samp{outputa} which starts at location
4357 @samp{0x10000}. All of section @samp{.input1} from file @file{foo.o}
4358 follows immediately, in the same output section. All of section
4359 @samp{.input2} from @file{foo.o} goes into output section
4360 @samp{outputb}, followed by section @samp{.input1} from @file{foo1.o}.
4361 All of the remaining @samp{.input1} and @samp{.input2} sections from any
4362 files are written to output section @samp{outputc}.
4390 @node Output Section Data
4391 @subsection Output Section Data
4393 @cindex section data
4394 @cindex output section data
4395 @kindex BYTE(@var{expression})
4396 @kindex SHORT(@var{expression})
4397 @kindex LONG(@var{expression})
4398 @kindex QUAD(@var{expression})
4399 @kindex SQUAD(@var{expression})
4400 You can include explicit bytes of data in an output section by using
4401 @code{BYTE}, @code{SHORT}, @code{LONG}, @code{QUAD}, or @code{SQUAD} as
4402 an output section command. Each keyword is followed by an expression in
4403 parentheses providing the value to store (@pxref{Expressions}). The
4404 value of the expression is stored at the current value of the location
4407 The @code{BYTE}, @code{SHORT}, @code{LONG}, and @code{QUAD} commands
4408 store one, two, four, and eight bytes (respectively). After storing the
4409 bytes, the location counter is incremented by the number of bytes
4412 For example, this will store the byte 1 followed by the four byte value
4413 of the symbol @samp{addr}:
4419 When using a 64 bit host or target, @code{QUAD} and @code{SQUAD} are the
4420 same; they both store an 8 byte, or 64 bit, value. When both host and
4421 target are 32 bits, an expression is computed as 32 bits. In this case
4422 @code{QUAD} stores a 32 bit value zero extended to 64 bits, and
4423 @code{SQUAD} stores a 32 bit value sign extended to 64 bits.
4425 If the object file format of the output file has an explicit endianness,
4426 which is the normal case, the value will be stored in that endianness.
4427 When the object file format does not have an explicit endianness, as is
4428 true of, for example, S-records, the value will be stored in the
4429 endianness of the first input object file.
4431 Note---these commands only work inside a section description and not
4432 between them, so the following will produce an error from the linker:
4434 SECTIONS @{@ .text : @{@ *(.text) @}@ LONG(1) .data : @{@ *(.data) @}@ @}@
4436 whereas this will work:
4438 SECTIONS @{@ .text : @{@ *(.text) ; LONG(1) @}@ .data : @{@ *(.data) @}@ @}@
4441 @kindex FILL(@var{expression})
4442 @cindex holes, filling
4443 @cindex unspecified memory
4444 You may use the @code{FILL} command to set the fill pattern for the
4445 current section. It is followed by an expression in parentheses. Any
4446 otherwise unspecified regions of memory within the section (for example,
4447 gaps left due to the required alignment of input sections) are filled
4448 with the value of the expression, repeated as
4449 necessary. A @code{FILL} statement covers memory locations after the
4450 point at which it occurs in the section definition; by including more
4451 than one @code{FILL} statement, you can have different fill patterns in
4452 different parts of an output section.
4454 This example shows how to fill unspecified regions of memory with the
4460 The @code{FILL} command is similar to the @samp{=@var{fillexp}} output
4461 section attribute, but it only affects the
4462 part of the section following the @code{FILL} command, rather than the
4463 entire section. If both are used, the @code{FILL} command takes
4464 precedence. @xref{Output Section Fill}, for details on the fill
4467 @node Output Section Keywords
4468 @subsection Output Section Keywords
4469 There are a couple of keywords which can appear as output section
4473 @kindex CREATE_OBJECT_SYMBOLS
4474 @cindex input filename symbols
4475 @cindex filename symbols
4476 @item CREATE_OBJECT_SYMBOLS
4477 The command tells the linker to create a symbol for each input file.
4478 The name of each symbol will be the name of the corresponding input
4479 file. The section of each symbol will be the output section in which
4480 the @code{CREATE_OBJECT_SYMBOLS} command appears.
4482 This is conventional for the a.out object file format. It is not
4483 normally used for any other object file format.
4485 @kindex CONSTRUCTORS
4486 @cindex C++ constructors, arranging in link
4487 @cindex constructors, arranging in link
4489 When linking using the a.out object file format, the linker uses an
4490 unusual set construct to support C++ global constructors and
4491 destructors. When linking object file formats which do not support
4492 arbitrary sections, such as ECOFF and XCOFF, the linker will
4493 automatically recognize C++ global constructors and destructors by name.
4494 For these object file formats, the @code{CONSTRUCTORS} command tells the
4495 linker to place constructor information in the output section where the
4496 @code{CONSTRUCTORS} command appears. The @code{CONSTRUCTORS} command is
4497 ignored for other object file formats.
4499 The symbol @w{@code{__CTOR_LIST__}} marks the start of the global
4500 constructors, and the symbol @w{@code{__CTOR_END__}} marks the end.
4501 Similarly, @w{@code{__DTOR_LIST__}} and @w{@code{__DTOR_END__}} mark
4502 the start and end of the global destructors. The
4503 first word in the list is the number of entries, followed by the address
4504 of each constructor or destructor, followed by a zero word. The
4505 compiler must arrange to actually run the code. For these object file
4506 formats @sc{gnu} C++ normally calls constructors from a subroutine
4507 @code{__main}; a call to @code{__main} is automatically inserted into
4508 the startup code for @code{main}. @sc{gnu} C++ normally runs
4509 destructors either by using @code{atexit}, or directly from the function
4512 For object file formats such as @code{COFF} or @code{ELF} which support
4513 arbitrary section names, @sc{gnu} C++ will normally arrange to put the
4514 addresses of global constructors and destructors into the @code{.ctors}
4515 and @code{.dtors} sections. Placing the following sequence into your
4516 linker script will build the sort of table which the @sc{gnu} C++
4517 runtime code expects to see.
4521 LONG((__CTOR_END__ - __CTOR_LIST__) / 4 - 2)
4526 LONG((__DTOR_END__ - __DTOR_LIST__) / 4 - 2)
4532 If you are using the @sc{gnu} C++ support for initialization priority,
4533 which provides some control over the order in which global constructors
4534 are run, you must sort the constructors at link time to ensure that they
4535 are executed in the correct order. When using the @code{CONSTRUCTORS}
4536 command, use @samp{SORT_BY_NAME(CONSTRUCTORS)} instead. When using the
4537 @code{.ctors} and @code{.dtors} sections, use @samp{*(SORT_BY_NAME(.ctors))} and
4538 @samp{*(SORT_BY_NAME(.dtors))} instead of just @samp{*(.ctors)} and
4541 Normally the compiler and linker will handle these issues automatically,
4542 and you will not need to concern yourself with them. However, you may
4543 need to consider this if you are using C++ and writing your own linker
4548 @node Output Section Discarding
4549 @subsection Output Section Discarding
4550 @cindex discarding sections
4551 @cindex sections, discarding
4552 @cindex removing sections
4553 The linker will not normally create output sections with no contents.
4554 This is for convenience when referring to input sections that may or
4555 may not be present in any of the input files. For example:
4557 .foo : @{ *(.foo) @}
4560 will only create a @samp{.foo} section in the output file if there is a
4561 @samp{.foo} section in at least one input file, and if the input
4562 sections are not all empty. Other link script directives that allocate
4563 space in an output section will also create the output section. So
4564 too will assignments to dot even if the assignment does not create
4565 space, except for @samp{. = 0}, @samp{. = . + 0}, @samp{. = sym},
4566 @samp{. = . + sym} and @samp{. = ALIGN (. != 0, expr, 1)} when
4567 @samp{sym} is an absolute symbol of value 0 defined in the script.
4568 This allows you to force output of an empty section with @samp{. = .}.
4570 The linker will ignore address assignments (@pxref{Output Section Address})
4571 on discarded output sections, except when the linker script defines
4572 symbols in the output section. In that case the linker will obey
4573 the address assignments, possibly advancing dot even though the
4574 section is discarded.
4577 The special output section name @samp{/DISCARD/} may be used to discard
4578 input sections. Any input sections which are assigned to an output
4579 section named @samp{/DISCARD/} are not included in the output file.
4581 @node Output Section Attributes
4582 @subsection Output Section Attributes
4583 @cindex output section attributes
4584 We showed above that the full description of an output section looked
4589 @var{section} [@var{address}] [(@var{type})] :
4591 [ALIGN(@var{section_align})]
4592 [SUBALIGN(@var{subsection_align})]
4595 @var{output-section-command}
4596 @var{output-section-command}
4598 @} [>@var{region}] [AT>@var{lma_region}] [:@var{phdr} :@var{phdr} @dots{}] [=@var{fillexp}]
4602 We've already described @var{section}, @var{address}, and
4603 @var{output-section-command}. In this section we will describe the
4604 remaining section attributes.
4607 * Output Section Type:: Output section type
4608 * Output Section LMA:: Output section LMA
4609 * Forced Output Alignment:: Forced Output Alignment
4610 * Forced Input Alignment:: Forced Input Alignment
4611 * Output Section Constraint:: Output section constraint
4612 * Output Section Region:: Output section region
4613 * Output Section Phdr:: Output section phdr
4614 * Output Section Fill:: Output section fill
4617 @node Output Section Type
4618 @subsubsection Output Section Type
4619 Each output section may have a type. The type is a keyword in
4620 parentheses. The following types are defined:
4624 The section should be marked as not loadable, so that it will not be
4625 loaded into memory when the program is run.
4630 These type names are supported for backward compatibility, and are
4631 rarely used. They all have the same effect: the section should be
4632 marked as not allocatable, so that no memory is allocated for the
4633 section when the program is run.
4637 @cindex prevent unnecessary loading
4638 @cindex loading, preventing
4639 The linker normally sets the attributes of an output section based on
4640 the input sections which map into it. You can override this by using
4641 the section type. For example, in the script sample below, the
4642 @samp{ROM} section is addressed at memory location @samp{0} and does not
4643 need to be loaded when the program is run.
4647 ROM 0 (NOLOAD) : @{ @dots{} @}
4653 @node Output Section LMA
4654 @subsubsection Output Section LMA
4655 @kindex AT>@var{lma_region}
4656 @kindex AT(@var{lma})
4657 @cindex load address
4658 @cindex section load address
4659 Every section has a virtual address (VMA) and a load address (LMA); see
4660 @ref{Basic Script Concepts}. The virtual address is specified by the
4661 @pxref{Output Section Address} described earlier. The load address is
4662 specified by the @code{AT} or @code{AT>} keywords. Specifying a load
4663 address is optional.
4665 The @code{AT} keyword takes an expression as an argument. This
4666 specifies the exact load address of the section. The @code{AT>} keyword
4667 takes the name of a memory region as an argument. @xref{MEMORY}. The
4668 load address of the section is set to the next free address in the
4669 region, aligned to the section's alignment requirements.
4671 If neither @code{AT} nor @code{AT>} is specified for an allocatable
4672 section, the linker will use the following heuristic to determine the
4677 If the section has a specific VMA address, then this is used as
4678 the LMA address as well.
4681 If the section is not allocatable then its LMA is set to its VMA.
4684 Otherwise if a memory region can be found that is compatible
4685 with the current section, and this region contains at least one
4686 section, then the LMA is set so the difference between the
4687 VMA and LMA is the same as the difference between the VMA and LMA of
4688 the last section in the located region.
4691 If no memory regions have been declared then a default region
4692 that covers the entire address space is used in the previous step.
4695 If no suitable region could be found, or there was no previous
4696 section then the LMA is set equal to the VMA.
4699 @cindex ROM initialized data
4700 @cindex initialized data in ROM
4701 This feature is designed to make it easy to build a ROM image. For
4702 example, the following linker script creates three output sections: one
4703 called @samp{.text}, which starts at @code{0x1000}, one called
4704 @samp{.mdata}, which is loaded at the end of the @samp{.text} section
4705 even though its VMA is @code{0x2000}, and one called @samp{.bss} to hold
4706 uninitialized data at address @code{0x3000}. The symbol @code{_data} is
4707 defined with the value @code{0x2000}, which shows that the location
4708 counter holds the VMA value, not the LMA value.
4714 .text 0x1000 : @{ *(.text) _etext = . ; @}
4716 AT ( ADDR (.text) + SIZEOF (.text) )
4717 @{ _data = . ; *(.data); _edata = . ; @}
4719 @{ _bstart = . ; *(.bss) *(COMMON) ; _bend = . ;@}
4724 The run-time initialization code for use with a program generated with
4725 this linker script would include something like the following, to copy
4726 the initialized data from the ROM image to its runtime address. Notice
4727 how this code takes advantage of the symbols defined by the linker
4732 extern char _etext, _data, _edata, _bstart, _bend;
4733 char *src = &_etext;
4736 /* ROM has data at end of text; copy it. */
4737 while (dst < &_edata)
4741 for (dst = &_bstart; dst< &_bend; dst++)
4746 @node Forced Output Alignment
4747 @subsubsection Forced Output Alignment
4748 @kindex ALIGN(@var{section_align})
4749 @cindex forcing output section alignment
4750 @cindex output section alignment
4751 You can increase an output section's alignment by using ALIGN. As an
4752 alternative you can enforce that the difference between the VMA and LMA remains
4753 intact throughout this output section with the ALIGN_WITH_INPUT attribute.
4755 @node Forced Input Alignment
4756 @subsubsection Forced Input Alignment
4757 @kindex SUBALIGN(@var{subsection_align})
4758 @cindex forcing input section alignment
4759 @cindex input section alignment
4760 You can force input section alignment within an output section by using
4761 SUBALIGN. The value specified overrides any alignment given by input
4762 sections, whether larger or smaller.
4764 @node Output Section Constraint
4765 @subsubsection Output Section Constraint
4768 @cindex constraints on output sections
4769 You can specify that an output section should only be created if all
4770 of its input sections are read-only or all of its input sections are
4771 read-write by using the keyword @code{ONLY_IF_RO} and
4772 @code{ONLY_IF_RW} respectively.
4774 @node Output Section Region
4775 @subsubsection Output Section Region
4776 @kindex >@var{region}
4777 @cindex section, assigning to memory region
4778 @cindex memory regions and sections
4779 You can assign a section to a previously defined region of memory by
4780 using @samp{>@var{region}}. @xref{MEMORY}.
4782 Here is a simple example:
4785 MEMORY @{ rom : ORIGIN = 0x1000, LENGTH = 0x1000 @}
4786 SECTIONS @{ ROM : @{ *(.text) @} >rom @}
4790 @node Output Section Phdr
4791 @subsubsection Output Section Phdr
4793 @cindex section, assigning to program header
4794 @cindex program headers and sections
4795 You can assign a section to a previously defined program segment by
4796 using @samp{:@var{phdr}}. @xref{PHDRS}. If a section is assigned to
4797 one or more segments, then all subsequent allocated sections will be
4798 assigned to those segments as well, unless they use an explicitly
4799 @code{:@var{phdr}} modifier. You can use @code{:NONE} to tell the
4800 linker to not put the section in any segment at all.
4802 Here is a simple example:
4805 PHDRS @{ text PT_LOAD ; @}
4806 SECTIONS @{ .text : @{ *(.text) @} :text @}
4810 @node Output Section Fill
4811 @subsubsection Output Section Fill
4812 @kindex =@var{fillexp}
4813 @cindex section fill pattern
4814 @cindex fill pattern, entire section
4815 You can set the fill pattern for an entire section by using
4816 @samp{=@var{fillexp}}. @var{fillexp} is an expression
4817 (@pxref{Expressions}). Any otherwise unspecified regions of memory
4818 within the output section (for example, gaps left due to the required
4819 alignment of input sections) will be filled with the value, repeated as
4820 necessary. If the fill expression is a simple hex number, ie. a string
4821 of hex digit starting with @samp{0x} and without a trailing @samp{k} or @samp{M}, then
4822 an arbitrarily long sequence of hex digits can be used to specify the
4823 fill pattern; Leading zeros become part of the pattern too. For all
4824 other cases, including extra parentheses or a unary @code{+}, the fill
4825 pattern is the four least significant bytes of the value of the
4826 expression. In all cases, the number is big-endian.
4828 You can also change the fill value with a @code{FILL} command in the
4829 output section commands; (@pxref{Output Section Data}).
4831 Here is a simple example:
4834 SECTIONS @{ .text : @{ *(.text) @} =0x90909090 @}
4838 @node Overlay Description
4839 @subsection Overlay Description
4842 An overlay description provides an easy way to describe sections which
4843 are to be loaded as part of a single memory image but are to be run at
4844 the same memory address. At run time, some sort of overlay manager will
4845 copy the overlaid sections in and out of the runtime memory address as
4846 required, perhaps by simply manipulating addressing bits. This approach
4847 can be useful, for example, when a certain region of memory is faster
4850 Overlays are described using the @code{OVERLAY} command. The
4851 @code{OVERLAY} command is used within a @code{SECTIONS} command, like an
4852 output section description. The full syntax of the @code{OVERLAY}
4853 command is as follows:
4856 OVERLAY [@var{start}] : [NOCROSSREFS] [AT ( @var{ldaddr} )]
4860 @var{output-section-command}
4861 @var{output-section-command}
4863 @} [:@var{phdr}@dots{}] [=@var{fill}]
4866 @var{output-section-command}
4867 @var{output-section-command}
4869 @} [:@var{phdr}@dots{}] [=@var{fill}]
4871 @} [>@var{region}] [:@var{phdr}@dots{}] [=@var{fill}] [,]
4875 Everything is optional except @code{OVERLAY} (a keyword), and each
4876 section must have a name (@var{secname1} and @var{secname2} above). The
4877 section definitions within the @code{OVERLAY} construct are identical to
4878 those within the general @code{SECTIONS} construct (@pxref{SECTIONS}),
4879 except that no addresses and no memory regions may be defined for
4880 sections within an @code{OVERLAY}.
4882 The comma at the end may be required if a @var{fill} is used and
4883 the next @var{sections-command} looks like a continuation of the expression.
4885 The sections are all defined with the same starting address. The load
4886 addresses of the sections are arranged such that they are consecutive in
4887 memory starting at the load address used for the @code{OVERLAY} as a
4888 whole (as with normal section definitions, the load address is optional,
4889 and defaults to the start address; the start address is also optional,
4890 and defaults to the current value of the location counter).
4892 If the @code{NOCROSSREFS} keyword is used, and there are any
4893 references among the sections, the linker will report an error. Since
4894 the sections all run at the same address, it normally does not make
4895 sense for one section to refer directly to another.
4896 @xref{Miscellaneous Commands, NOCROSSREFS}.
4898 For each section within the @code{OVERLAY}, the linker automatically
4899 provides two symbols. The symbol @code{__load_start_@var{secname}} is
4900 defined as the starting load address of the section. The symbol
4901 @code{__load_stop_@var{secname}} is defined as the final load address of
4902 the section. Any characters within @var{secname} which are not legal
4903 within C identifiers are removed. C (or assembler) code may use these
4904 symbols to move the overlaid sections around as necessary.
4906 At the end of the overlay, the value of the location counter is set to
4907 the start address of the overlay plus the size of the largest section.
4909 Here is an example. Remember that this would appear inside a
4910 @code{SECTIONS} construct.
4913 OVERLAY 0x1000 : AT (0x4000)
4915 .text0 @{ o1/*.o(.text) @}
4916 .text1 @{ o2/*.o(.text) @}
4921 This will define both @samp{.text0} and @samp{.text1} to start at
4922 address 0x1000. @samp{.text0} will be loaded at address 0x4000, and
4923 @samp{.text1} will be loaded immediately after @samp{.text0}. The
4924 following symbols will be defined if referenced: @code{__load_start_text0},
4925 @code{__load_stop_text0}, @code{__load_start_text1},
4926 @code{__load_stop_text1}.
4928 C code to copy overlay @code{.text1} into the overlay area might look
4933 extern char __load_start_text1, __load_stop_text1;
4934 memcpy ((char *) 0x1000, &__load_start_text1,
4935 &__load_stop_text1 - &__load_start_text1);
4939 Note that the @code{OVERLAY} command is just syntactic sugar, since
4940 everything it does can be done using the more basic commands. The above
4941 example could have been written identically as follows.
4945 .text0 0x1000 : AT (0x4000) @{ o1/*.o(.text) @}
4946 PROVIDE (__load_start_text0 = LOADADDR (.text0));
4947 PROVIDE (__load_stop_text0 = LOADADDR (.text0) + SIZEOF (.text0));
4948 .text1 0x1000 : AT (0x4000 + SIZEOF (.text0)) @{ o2/*.o(.text) @}
4949 PROVIDE (__load_start_text1 = LOADADDR (.text1));
4950 PROVIDE (__load_stop_text1 = LOADADDR (.text1) + SIZEOF (.text1));
4951 . = 0x1000 + MAX (SIZEOF (.text0), SIZEOF (.text1));
4956 @section MEMORY Command
4958 @cindex memory regions
4959 @cindex regions of memory
4960 @cindex allocating memory
4961 @cindex discontinuous memory
4962 The linker's default configuration permits allocation of all available
4963 memory. You can override this by using the @code{MEMORY} command.
4965 The @code{MEMORY} command describes the location and size of blocks of
4966 memory in the target. You can use it to describe which memory regions
4967 may be used by the linker, and which memory regions it must avoid. You
4968 can then assign sections to particular memory regions. The linker will
4969 set section addresses based on the memory regions, and will warn about
4970 regions that become too full. The linker will not shuffle sections
4971 around to fit into the available regions.
4973 A linker script may contain many uses of the @code{MEMORY} command,
4974 however, all memory blocks defined are treated as if they were
4975 specified inside a single @code{MEMORY} command. The syntax for
4981 @var{name} [(@var{attr})] : ORIGIN = @var{origin}, LENGTH = @var{len}
4987 The @var{name} is a name used in the linker script to refer to the
4988 region. The region name has no meaning outside of the linker script.
4989 Region names are stored in a separate name space, and will not conflict
4990 with symbol names, file names, or section names. Each memory region
4991 must have a distinct name within the @code{MEMORY} command. However you can
4992 add later alias names to existing memory regions with the @ref{REGION_ALIAS}
4995 @cindex memory region attributes
4996 The @var{attr} string is an optional list of attributes that specify
4997 whether to use a particular memory region for an input section which is
4998 not explicitly mapped in the linker script. As described in
4999 @ref{SECTIONS}, if you do not specify an output section for some input
5000 section, the linker will create an output section with the same name as
5001 the input section. If you define region attributes, the linker will use
5002 them to select the memory region for the output section that it creates.
5004 The @var{attr} string must consist only of the following characters:
5019 Invert the sense of any of the attributes that follow
5022 If a unmapped section matches any of the listed attributes other than
5023 @samp{!}, it will be placed in the memory region. The @samp{!}
5024 attribute reverses this test, so that an unmapped section will be placed
5025 in the memory region only if it does not match any of the listed
5031 The @var{origin} is an numerical expression for the start address of
5032 the memory region. The expression must evaluate to a constant and it
5033 cannot involve any symbols. The keyword @code{ORIGIN} may be
5034 abbreviated to @code{org} or @code{o} (but not, for example,
5040 The @var{len} is an expression for the size in bytes of the memory
5041 region. As with the @var{origin} expression, the expression must
5042 be numerical only and must evaluate to a constant. The keyword
5043 @code{LENGTH} may be abbreviated to @code{len} or @code{l}.
5045 In the following example, we specify that there are two memory regions
5046 available for allocation: one starting at @samp{0} for 256 kilobytes,
5047 and the other starting at @samp{0x40000000} for four megabytes. The
5048 linker will place into the @samp{rom} memory region every section which
5049 is not explicitly mapped into a memory region, and is either read-only
5050 or executable. The linker will place other sections which are not
5051 explicitly mapped into a memory region into the @samp{ram} memory
5058 rom (rx) : ORIGIN = 0, LENGTH = 256K
5059 ram (!rx) : org = 0x40000000, l = 4M
5064 Once you define a memory region, you can direct the linker to place
5065 specific output sections into that memory region by using the
5066 @samp{>@var{region}} output section attribute. For example, if you have
5067 a memory region named @samp{mem}, you would use @samp{>mem} in the
5068 output section definition. @xref{Output Section Region}. If no address
5069 was specified for the output section, the linker will set the address to
5070 the next available address within the memory region. If the combined
5071 output sections directed to a memory region are too large for the
5072 region, the linker will issue an error message.
5074 It is possible to access the origin and length of a memory in an
5075 expression via the @code{ORIGIN(@var{memory})} and
5076 @code{LENGTH(@var{memory})} functions:
5080 _fstack = ORIGIN(ram) + LENGTH(ram) - 4;
5085 @section PHDRS Command
5087 @cindex program headers
5088 @cindex ELF program headers
5089 @cindex program segments
5090 @cindex segments, ELF
5091 The ELF object file format uses @dfn{program headers}, also knows as
5092 @dfn{segments}. The program headers describe how the program should be
5093 loaded into memory. You can print them out by using the @code{objdump}
5094 program with the @samp{-p} option.
5096 When you run an ELF program on a native ELF system, the system loader
5097 reads the program headers in order to figure out how to load the
5098 program. This will only work if the program headers are set correctly.
5099 This manual does not describe the details of how the system loader
5100 interprets program headers; for more information, see the ELF ABI.
5102 The linker will create reasonable program headers by default. However,
5103 in some cases, you may need to specify the program headers more
5104 precisely. You may use the @code{PHDRS} command for this purpose. When
5105 the linker sees the @code{PHDRS} command in the linker script, it will
5106 not create any program headers other than the ones specified.
5108 The linker only pays attention to the @code{PHDRS} command when
5109 generating an ELF output file. In other cases, the linker will simply
5110 ignore @code{PHDRS}.
5112 This is the syntax of the @code{PHDRS} command. The words @code{PHDRS},
5113 @code{FILEHDR}, @code{AT}, and @code{FLAGS} are keywords.
5119 @var{name} @var{type} [ FILEHDR ] [ PHDRS ] [ AT ( @var{address} ) ]
5120 [ FLAGS ( @var{flags} ) ] ;
5125 The @var{name} is used only for reference in the @code{SECTIONS} command
5126 of the linker script. It is not put into the output file. Program
5127 header names are stored in a separate name space, and will not conflict
5128 with symbol names, file names, or section names. Each program header
5129 must have a distinct name. The headers are processed in order and it
5130 is usual for them to map to sections in ascending load address order.
5132 Certain program header types describe segments of memory which the
5133 system loader will load from the file. In the linker script, you
5134 specify the contents of these segments by placing allocatable output
5135 sections in the segments. You use the @samp{:@var{phdr}} output section
5136 attribute to place a section in a particular segment. @xref{Output
5139 It is normal to put certain sections in more than one segment. This
5140 merely implies that one segment of memory contains another. You may
5141 repeat @samp{:@var{phdr}}, using it once for each segment which should
5142 contain the section.
5144 If you place a section in one or more segments using @samp{:@var{phdr}},
5145 then the linker will place all subsequent allocatable sections which do
5146 not specify @samp{:@var{phdr}} in the same segments. This is for
5147 convenience, since generally a whole set of contiguous sections will be
5148 placed in a single segment. You can use @code{:NONE} to override the
5149 default segment and tell the linker to not put the section in any
5154 You may use the @code{FILEHDR} and @code{PHDRS} keywords after
5155 the program header type to further describe the contents of the segment.
5156 The @code{FILEHDR} keyword means that the segment should include the ELF
5157 file header. The @code{PHDRS} keyword means that the segment should
5158 include the ELF program headers themselves. If applied to a loadable
5159 segment (@code{PT_LOAD}), all prior loadable segments must have one of
5162 The @var{type} may be one of the following. The numbers indicate the
5163 value of the keyword.
5166 @item @code{PT_NULL} (0)
5167 Indicates an unused program header.
5169 @item @code{PT_LOAD} (1)
5170 Indicates that this program header describes a segment to be loaded from
5173 @item @code{PT_DYNAMIC} (2)
5174 Indicates a segment where dynamic linking information can be found.
5176 @item @code{PT_INTERP} (3)
5177 Indicates a segment where the name of the program interpreter may be
5180 @item @code{PT_NOTE} (4)
5181 Indicates a segment holding note information.
5183 @item @code{PT_SHLIB} (5)
5184 A reserved program header type, defined but not specified by the ELF
5187 @item @code{PT_PHDR} (6)
5188 Indicates a segment where the program headers may be found.
5190 @item @var{expression}
5191 An expression giving the numeric type of the program header. This may
5192 be used for types not defined above.
5195 You can specify that a segment should be loaded at a particular address
5196 in memory by using an @code{AT} expression. This is identical to the
5197 @code{AT} command used as an output section attribute (@pxref{Output
5198 Section LMA}). The @code{AT} command for a program header overrides the
5199 output section attribute.
5201 The linker will normally set the segment flags based on the sections
5202 which comprise the segment. You may use the @code{FLAGS} keyword to
5203 explicitly specify the segment flags. The value of @var{flags} must be
5204 an integer. It is used to set the @code{p_flags} field of the program
5207 Here is an example of @code{PHDRS}. This shows a typical set of program
5208 headers used on a native ELF system.
5214 headers PT_PHDR PHDRS ;
5216 text PT_LOAD FILEHDR PHDRS ;
5218 dynamic PT_DYNAMIC ;
5224 .interp : @{ *(.interp) @} :text :interp
5225 .text : @{ *(.text) @} :text
5226 .rodata : @{ *(.rodata) @} /* defaults to :text */
5228 . = . + 0x1000; /* move to a new page in memory */
5229 .data : @{ *(.data) @} :data
5230 .dynamic : @{ *(.dynamic) @} :data :dynamic
5237 @section VERSION Command
5238 @kindex VERSION @{script text@}
5239 @cindex symbol versions
5240 @cindex version script
5241 @cindex versions of symbols
5242 The linker supports symbol versions when using ELF. Symbol versions are
5243 only useful when using shared libraries. The dynamic linker can use
5244 symbol versions to select a specific version of a function when it runs
5245 a program that may have been linked against an earlier version of the
5248 You can include a version script directly in the main linker script, or
5249 you can supply the version script as an implicit linker script. You can
5250 also use the @samp{--version-script} linker option.
5252 The syntax of the @code{VERSION} command is simply
5254 VERSION @{ version-script-commands @}
5257 The format of the version script commands is identical to that used by
5258 Sun's linker in Solaris 2.5. The version script defines a tree of
5259 version nodes. You specify the node names and interdependencies in the
5260 version script. You can specify which symbols are bound to which
5261 version nodes, and you can reduce a specified set of symbols to local
5262 scope so that they are not globally visible outside of the shared
5265 The easiest way to demonstrate the version script language is with a few
5291 This example version script defines three version nodes. The first
5292 version node defined is @samp{VERS_1.1}; it has no other dependencies.
5293 The script binds the symbol @samp{foo1} to @samp{VERS_1.1}. It reduces
5294 a number of symbols to local scope so that they are not visible outside
5295 of the shared library; this is done using wildcard patterns, so that any
5296 symbol whose name begins with @samp{old}, @samp{original}, or @samp{new}
5297 is matched. The wildcard patterns available are the same as those used
5298 in the shell when matching filenames (also known as ``globbing'').
5299 However, if you specify the symbol name inside double quotes, then the
5300 name is treated as literal, rather than as a glob pattern.
5302 Next, the version script defines node @samp{VERS_1.2}. This node
5303 depends upon @samp{VERS_1.1}. The script binds the symbol @samp{foo2}
5304 to the version node @samp{VERS_1.2}.
5306 Finally, the version script defines node @samp{VERS_2.0}. This node
5307 depends upon @samp{VERS_1.2}. The scripts binds the symbols @samp{bar1}
5308 and @samp{bar2} are bound to the version node @samp{VERS_2.0}.
5310 When the linker finds a symbol defined in a library which is not
5311 specifically bound to a version node, it will effectively bind it to an
5312 unspecified base version of the library. You can bind all otherwise
5313 unspecified symbols to a given version node by using @samp{global: *;}
5314 somewhere in the version script. Note that it's slightly crazy to use
5315 wildcards in a global spec except on the last version node. Global
5316 wildcards elsewhere run the risk of accidentally adding symbols to the
5317 set exported for an old version. That's wrong since older versions
5318 ought to have a fixed set of symbols.
5320 The names of the version nodes have no specific meaning other than what
5321 they might suggest to the person reading them. The @samp{2.0} version
5322 could just as well have appeared in between @samp{1.1} and @samp{1.2}.
5323 However, this would be a confusing way to write a version script.
5325 Node name can be omitted, provided it is the only version node
5326 in the version script. Such version script doesn't assign any versions to
5327 symbols, only selects which symbols will be globally visible out and which
5331 @{ global: foo; bar; local: *; @};
5334 When you link an application against a shared library that has versioned
5335 symbols, the application itself knows which version of each symbol it
5336 requires, and it also knows which version nodes it needs from each
5337 shared library it is linked against. Thus at runtime, the dynamic
5338 loader can make a quick check to make sure that the libraries you have
5339 linked against do in fact supply all of the version nodes that the
5340 application will need to resolve all of the dynamic symbols. In this
5341 way it is possible for the dynamic linker to know with certainty that
5342 all external symbols that it needs will be resolvable without having to
5343 search for each symbol reference.
5345 The symbol versioning is in effect a much more sophisticated way of
5346 doing minor version checking that SunOS does. The fundamental problem
5347 that is being addressed here is that typically references to external
5348 functions are bound on an as-needed basis, and are not all bound when
5349 the application starts up. If a shared library is out of date, a
5350 required interface may be missing; when the application tries to use
5351 that interface, it may suddenly and unexpectedly fail. With symbol
5352 versioning, the user will get a warning when they start their program if
5353 the libraries being used with the application are too old.
5355 There are several GNU extensions to Sun's versioning approach. The
5356 first of these is the ability to bind a symbol to a version node in the
5357 source file where the symbol is defined instead of in the versioning
5358 script. This was done mainly to reduce the burden on the library
5359 maintainer. You can do this by putting something like:
5361 __asm__(".symver original_foo,foo@@VERS_1.1");
5364 in the C source file. This renames the function @samp{original_foo} to
5365 be an alias for @samp{foo} bound to the version node @samp{VERS_1.1}.
5366 The @samp{local:} directive can be used to prevent the symbol
5367 @samp{original_foo} from being exported. A @samp{.symver} directive
5368 takes precedence over a version script.
5370 The second GNU extension is to allow multiple versions of the same
5371 function to appear in a given shared library. In this way you can make
5372 an incompatible change to an interface without increasing the major
5373 version number of the shared library, while still allowing applications
5374 linked against the old interface to continue to function.
5376 To do this, you must use multiple @samp{.symver} directives in the
5377 source file. Here is an example:
5380 __asm__(".symver original_foo,foo@@");
5381 __asm__(".symver old_foo,foo@@VERS_1.1");
5382 __asm__(".symver old_foo1,foo@@VERS_1.2");
5383 __asm__(".symver new_foo,foo@@@@VERS_2.0");
5386 In this example, @samp{foo@@} represents the symbol @samp{foo} bound to the
5387 unspecified base version of the symbol. The source file that contains this
5388 example would define 4 C functions: @samp{original_foo}, @samp{old_foo},
5389 @samp{old_foo1}, and @samp{new_foo}.
5391 When you have multiple definitions of a given symbol, there needs to be
5392 some way to specify a default version to which external references to
5393 this symbol will be bound. You can do this with the
5394 @samp{foo@@@@VERS_2.0} type of @samp{.symver} directive. You can only
5395 declare one version of a symbol as the default in this manner; otherwise
5396 you would effectively have multiple definitions of the same symbol.
5398 If you wish to bind a reference to a specific version of the symbol
5399 within the shared library, you can use the aliases of convenience
5400 (i.e., @samp{old_foo}), or you can use the @samp{.symver} directive to
5401 specifically bind to an external version of the function in question.
5403 You can also specify the language in the version script:
5406 VERSION extern "lang" @{ version-script-commands @}
5409 The supported @samp{lang}s are @samp{C}, @samp{C++}, and @samp{Java}.
5410 The linker will iterate over the list of symbols at the link time and
5411 demangle them according to @samp{lang} before matching them to the
5412 patterns specified in @samp{version-script-commands}. The default
5413 @samp{lang} is @samp{C}.
5415 Demangled names may contains spaces and other special characters. As
5416 described above, you can use a glob pattern to match demangled names,
5417 or you can use a double-quoted string to match the string exactly. In
5418 the latter case, be aware that minor differences (such as differing
5419 whitespace) between the version script and the demangler output will
5420 cause a mismatch. As the exact string generated by the demangler
5421 might change in the future, even if the mangled name does not, you
5422 should check that all of your version directives are behaving as you
5423 expect when you upgrade.
5426 @section Expressions in Linker Scripts
5429 The syntax for expressions in the linker script language is identical to
5430 that of C expressions. All expressions are evaluated as integers. All
5431 expressions are evaluated in the same size, which is 32 bits if both the
5432 host and target are 32 bits, and is otherwise 64 bits.
5434 You can use and set symbol values in expressions.
5436 The linker defines several special purpose builtin functions for use in
5440 * Constants:: Constants
5441 * Symbolic Constants:: Symbolic constants
5442 * Symbols:: Symbol Names
5443 * Orphan Sections:: Orphan Sections
5444 * Location Counter:: The Location Counter
5445 * Operators:: Operators
5446 * Evaluation:: Evaluation
5447 * Expression Section:: The Section of an Expression
5448 * Builtin Functions:: Builtin Functions
5452 @subsection Constants
5453 @cindex integer notation
5454 @cindex constants in linker scripts
5455 All constants are integers.
5457 As in C, the linker considers an integer beginning with @samp{0} to be
5458 octal, and an integer beginning with @samp{0x} or @samp{0X} to be
5459 hexadecimal. Alternatively the linker accepts suffixes of @samp{h} or
5460 @samp{H} for hexadecimal, @samp{o} or @samp{O} for octal, @samp{b} or
5461 @samp{B} for binary and @samp{d} or @samp{D} for decimal. Any integer
5462 value without a prefix or a suffix is considered to be decimal.
5464 @cindex scaled integers
5465 @cindex K and M integer suffixes
5466 @cindex M and K integer suffixes
5467 @cindex suffixes for integers
5468 @cindex integer suffixes
5469 In addition, you can use the suffixes @code{K} and @code{M} to scale a
5473 @c END TEXI2ROFF-KILL
5474 @code{1024} or @code{1024*1024}
5478 ${\rm 1024}$ or ${\rm 1024}^2$
5480 @c END TEXI2ROFF-KILL
5481 respectively. For example, the following
5482 all refer to the same quantity:
5491 Note - the @code{K} and @code{M} suffixes cannot be used in
5492 conjunction with the base suffixes mentioned above.
5494 @node Symbolic Constants
5495 @subsection Symbolic Constants
5496 @cindex symbolic constants
5498 It is possible to refer to target specific constants via the use of
5499 the @code{CONSTANT(@var{name})} operator, where @var{name} is one of:
5504 The target's maximum page size.
5506 @item COMMONPAGESIZE
5507 @kindex COMMONPAGESIZE
5508 The target's default page size.
5514 .text ALIGN (CONSTANT (MAXPAGESIZE)) : @{ *(.text) @}
5517 will create a text section aligned to the largest page boundary
5518 supported by the target.
5521 @subsection Symbol Names
5522 @cindex symbol names
5524 @cindex quoted symbol names
5526 Unless quoted, symbol names start with a letter, underscore, or period
5527 and may include letters, digits, underscores, periods, and hyphens.
5528 Unquoted symbol names must not conflict with any keywords. You can
5529 specify a symbol which contains odd characters or has the same name as a
5530 keyword by surrounding the symbol name in double quotes:
5533 "with a space" = "also with a space" + 10;
5536 Since symbols can contain many non-alphabetic characters, it is safest
5537 to delimit symbols with spaces. For example, @samp{A-B} is one symbol,
5538 whereas @samp{A - B} is an expression involving subtraction.
5540 @node Orphan Sections
5541 @subsection Orphan Sections
5543 Orphan sections are sections present in the input files which
5544 are not explicitly placed into the output file by the linker
5545 script. The linker will still copy these sections into the
5546 output file, but it has to guess as to where they should be
5547 placed. The linker uses a simple heuristic to do this. It
5548 attempts to place orphan sections after non-orphan sections of the
5549 same attribute, such as code vs data, loadable vs non-loadable, etc.
5550 If there is not enough room to do this then it places
5551 at the end of the file.
5553 For ELF targets, the attribute of the section includes section type as
5554 well as section flag.
5556 If an orphaned section's name is representable as a C identifier then
5557 the linker will automatically @pxref{PROVIDE} two symbols:
5558 __start_SECNAME and __stop_SECNAME, where SECNAME is the name of the
5559 section. These indicate the start address and end address of the
5560 orphaned section respectively. Note: most section names are not
5561 representable as C identifiers because they contain a @samp{.}
5564 @node Location Counter
5565 @subsection The Location Counter
5568 @cindex location counter
5569 @cindex current output location
5570 The special linker variable @dfn{dot} @samp{.} always contains the
5571 current output location counter. Since the @code{.} always refers to a
5572 location in an output section, it may only appear in an expression
5573 within a @code{SECTIONS} command. The @code{.} symbol may appear
5574 anywhere that an ordinary symbol is allowed in an expression.
5577 Assigning a value to @code{.} will cause the location counter to be
5578 moved. This may be used to create holes in the output section. The
5579 location counter may not be moved backwards inside an output section,
5580 and may not be moved backwards outside of an output section if so
5581 doing creates areas with overlapping LMAs.
5597 In the previous example, the @samp{.text} section from @file{file1} is
5598 located at the beginning of the output section @samp{output}. It is
5599 followed by a 1000 byte gap. Then the @samp{.text} section from
5600 @file{file2} appears, also with a 1000 byte gap following before the
5601 @samp{.text} section from @file{file3}. The notation @samp{= 0x12345678}
5602 specifies what data to write in the gaps (@pxref{Output Section Fill}).
5604 @cindex dot inside sections
5605 Note: @code{.} actually refers to the byte offset from the start of the
5606 current containing object. Normally this is the @code{SECTIONS}
5607 statement, whose start address is 0, hence @code{.} can be used as an
5608 absolute address. If @code{.} is used inside a section description
5609 however, it refers to the byte offset from the start of that section,
5610 not an absolute address. Thus in a script like this:
5628 The @samp{.text} section will be assigned a starting address of 0x100
5629 and a size of exactly 0x200 bytes, even if there is not enough data in
5630 the @samp{.text} input sections to fill this area. (If there is too
5631 much data, an error will be produced because this would be an attempt to
5632 move @code{.} backwards). The @samp{.data} section will start at 0x500
5633 and it will have an extra 0x600 bytes worth of space after the end of
5634 the values from the @samp{.data} input sections and before the end of
5635 the @samp{.data} output section itself.
5637 @cindex dot outside sections
5638 Setting symbols to the value of the location counter outside of an
5639 output section statement can result in unexpected values if the linker
5640 needs to place orphan sections. For example, given the following:
5646 .text: @{ *(.text) @}
5650 .data: @{ *(.data) @}
5655 If the linker needs to place some input section, e.g. @code{.rodata},
5656 not mentioned in the script, it might choose to place that section
5657 between @code{.text} and @code{.data}. You might think the linker
5658 should place @code{.rodata} on the blank line in the above script, but
5659 blank lines are of no particular significance to the linker. As well,
5660 the linker doesn't associate the above symbol names with their
5661 sections. Instead, it assumes that all assignments or other
5662 statements belong to the previous output section, except for the
5663 special case of an assignment to @code{.}. I.e., the linker will
5664 place the orphan @code{.rodata} section as if the script was written
5671 .text: @{ *(.text) @}
5675 .rodata: @{ *(.rodata) @}
5676 .data: @{ *(.data) @}
5681 This may or may not be the script author's intention for the value of
5682 @code{start_of_data}. One way to influence the orphan section
5683 placement is to assign the location counter to itself, as the linker
5684 assumes that an assignment to @code{.} is setting the start address of
5685 a following output section and thus should be grouped with that
5686 section. So you could write:
5692 .text: @{ *(.text) @}
5697 .data: @{ *(.data) @}
5702 Now, the orphan @code{.rodata} section will be placed between
5703 @code{end_of_text} and @code{start_of_data}.
5707 @subsection Operators
5708 @cindex operators for arithmetic
5709 @cindex arithmetic operators
5710 @cindex precedence in expressions
5711 The linker recognizes the standard C set of arithmetic operators, with
5712 the standard bindings and precedence levels:
5715 @c END TEXI2ROFF-KILL
5717 precedence associativity Operators Notes
5723 5 left == != > < <= >=
5729 11 right &= += -= *= /= (2)
5733 (1) Prefix operators
5734 (2) @xref{Assignments}.
5738 \vskip \baselineskip
5739 %"lispnarrowing" is the extra indent used generally for smallexample
5740 \hskip\lispnarrowing\vbox{\offinterlineskip
5743 {\vrule#&\strut\hfil\ #\ \hfil&\vrule#&\strut\hfil\ #\ \hfil&\vrule#&\strut\hfil\ {\tt #}\ \hfil&\vrule#\cr
5744 height2pt&\omit&&\omit&&\omit&\cr
5745 &Precedence&& Associativity &&{\rm Operators}&\cr
5746 height2pt&\omit&&\omit&&\omit&\cr
5748 height2pt&\omit&&\omit&&\omit&\cr
5750 % '176 is tilde, '~' in tt font
5751 &1&&left&&\qquad- \char'176\ !\qquad\dag&\cr
5752 &2&&left&&* / \%&\cr
5755 &5&&left&&== != > < <= >=&\cr
5758 &8&&left&&{\&\&}&\cr
5761 &11&&right&&\qquad\&= += -= *= /=\qquad\ddag&\cr
5763 height2pt&\omit&&\omit&&\omit&\cr}
5768 @obeylines@parskip=0pt@parindent=0pt
5769 @dag@quad Prefix operators.
5770 @ddag@quad @xref{Assignments}.
5773 @c END TEXI2ROFF-KILL
5776 @subsection Evaluation
5777 @cindex lazy evaluation
5778 @cindex expression evaluation order
5779 The linker evaluates expressions lazily. It only computes the value of
5780 an expression when absolutely necessary.
5782 The linker needs some information, such as the value of the start
5783 address of the first section, and the origins and lengths of memory
5784 regions, in order to do any linking at all. These values are computed
5785 as soon as possible when the linker reads in the linker script.
5787 However, other values (such as symbol values) are not known or needed
5788 until after storage allocation. Such values are evaluated later, when
5789 other information (such as the sizes of output sections) is available
5790 for use in the symbol assignment expression.
5792 The sizes of sections cannot be known until after allocation, so
5793 assignments dependent upon these are not performed until after
5796 Some expressions, such as those depending upon the location counter
5797 @samp{.}, must be evaluated during section allocation.
5799 If the result of an expression is required, but the value is not
5800 available, then an error results. For example, a script like the
5806 .text 9+this_isnt_constant :
5812 will cause the error message @samp{non constant expression for initial
5815 @node Expression Section
5816 @subsection The Section of an Expression
5817 @cindex expression sections
5818 @cindex absolute expressions
5819 @cindex relative expressions
5820 @cindex absolute and relocatable symbols
5821 @cindex relocatable and absolute symbols
5822 @cindex symbols, relocatable and absolute
5823 Addresses and symbols may be section relative, or absolute. A section
5824 relative symbol is relocatable. If you request relocatable output
5825 using the @samp{-r} option, a further link operation may change the
5826 value of a section relative symbol. On the other hand, an absolute
5827 symbol will retain the same value throughout any further link
5830 Some terms in linker expressions are addresses. This is true of
5831 section relative symbols and for builtin functions that return an
5832 address, such as @code{ADDR}, @code{LOADADDR}, @code{ORIGIN} and
5833 @code{SEGMENT_START}. Other terms are simply numbers, or are builtin
5834 functions that return a non-address value, such as @code{LENGTH}.
5835 One complication is that unless you set @code{LD_FEATURE ("SANE_EXPR")}
5836 (@pxref{Miscellaneous Commands}), numbers and absolute symbols are treated
5837 differently depending on their location, for compatibility with older
5838 versions of @code{ld}. Expressions appearing outside an output
5839 section definition treat all numbers as absolute addresses.
5840 Expressions appearing inside an output section definition treat
5841 absolute symbols as numbers. If @code{LD_FEATURE ("SANE_EXPR")} is
5842 given, then absolute symbols and numbers are simply treated as numbers
5845 In the following simple example,
5852 __executable_start = 0x100;
5856 __data_start = 0x10;
5864 both @code{.} and @code{__executable_start} are set to the absolute
5865 address 0x100 in the first two assignments, then both @code{.} and
5866 @code{__data_start} are set to 0x10 relative to the @code{.data}
5867 section in the second two assignments.
5869 For expressions involving numbers, relative addresses and absolute
5870 addresses, ld follows these rules to evaluate terms:
5874 Unary operations on an absolute address or number, and binary
5875 operations on two absolute addresses or two numbers, or between one
5876 absolute address and a number, apply the operator to the value(s).
5878 Unary operations on a relative address, and binary operations on two
5879 relative addresses in the same section or between one relative address
5880 and a number, apply the operator to the offset part of the address(es).
5882 Other binary operations, that is, between two relative addresses not
5883 in the same section, or between a relative address and an absolute
5884 address, first convert any non-absolute term to an absolute address
5885 before applying the operator.
5888 The result section of each sub-expression is as follows:
5892 An operation involving only numbers results in a number.
5894 The result of comparisons, @samp{&&} and @samp{||} is also a number.
5896 The result of other binary arithmetic and logical operations on two
5897 relative addresses in the same section or two absolute addresses
5898 (after above conversions) is also a number.
5900 The result of other operations on relative addresses or one
5901 relative address and a number, is a relative address in the same
5902 section as the relative operand(s).
5904 The result of other operations on absolute addresses (after above
5905 conversions) is an absolute address.
5908 You can use the builtin function @code{ABSOLUTE} to force an expression
5909 to be absolute when it would otherwise be relative. For example, to
5910 create an absolute symbol set to the address of the end of the output
5911 section @samp{.data}:
5915 .data : @{ *(.data) _edata = ABSOLUTE(.); @}
5919 If @samp{ABSOLUTE} were not used, @samp{_edata} would be relative to the
5920 @samp{.data} section.
5922 Using @code{LOADADDR} also forces an expression absolute, since this
5923 particular builtin function returns an absolute address.
5925 @node Builtin Functions
5926 @subsection Builtin Functions
5927 @cindex functions in expressions
5928 The linker script language includes a number of builtin functions for
5929 use in linker script expressions.
5932 @item ABSOLUTE(@var{exp})
5933 @kindex ABSOLUTE(@var{exp})
5934 @cindex expression, absolute
5935 Return the absolute (non-relocatable, as opposed to non-negative) value
5936 of the expression @var{exp}. Primarily useful to assign an absolute
5937 value to a symbol within a section definition, where symbol values are
5938 normally section relative. @xref{Expression Section}.
5940 @item ADDR(@var{section})
5941 @kindex ADDR(@var{section})
5942 @cindex section address in expression
5943 Return the address (VMA) of the named @var{section}. Your
5944 script must previously have defined the location of that section. In
5945 the following example, @code{start_of_output_1}, @code{symbol_1} and
5946 @code{symbol_2} are assigned equivalent values, except that
5947 @code{symbol_1} will be relative to the @code{.output1} section while
5948 the other two will be absolute:
5954 start_of_output_1 = ABSOLUTE(.);
5959 symbol_1 = ADDR(.output1);
5960 symbol_2 = start_of_output_1;
5966 @item ALIGN(@var{align})
5967 @itemx ALIGN(@var{exp},@var{align})
5968 @kindex ALIGN(@var{align})
5969 @kindex ALIGN(@var{exp},@var{align})
5970 @cindex round up location counter
5971 @cindex align location counter
5972 @cindex round up expression
5973 @cindex align expression
5974 Return the location counter (@code{.}) or arbitrary expression aligned
5975 to the next @var{align} boundary. The single operand @code{ALIGN}
5976 doesn't change the value of the location counter---it just does
5977 arithmetic on it. The two operand @code{ALIGN} allows an arbitrary
5978 expression to be aligned upwards (@code{ALIGN(@var{align})} is
5979 equivalent to @code{ALIGN(., @var{align})}).
5981 Here is an example which aligns the output @code{.data} section to the
5982 next @code{0x2000} byte boundary after the preceding section and sets a
5983 variable within the section to the next @code{0x8000} boundary after the
5988 .data ALIGN(0x2000): @{
5990 variable = ALIGN(0x8000);
5996 The first use of @code{ALIGN} in this example specifies the location of
5997 a section because it is used as the optional @var{address} attribute of
5998 a section definition (@pxref{Output Section Address}). The second use
5999 of @code{ALIGN} is used to defines the value of a symbol.
6001 The builtin function @code{NEXT} is closely related to @code{ALIGN}.
6003 @item ALIGNOF(@var{section})
6004 @kindex ALIGNOF(@var{section})
6005 @cindex section alignment
6006 Return the alignment in bytes of the named @var{section}, if that section has
6007 been allocated. If the section has not been allocated when this is
6008 evaluated, the linker will report an error. In the following example,
6009 the alignment of the @code{.output} section is stored as the first
6010 value in that section.
6015 LONG (ALIGNOF (.output))
6022 @item BLOCK(@var{exp})
6023 @kindex BLOCK(@var{exp})
6024 This is a synonym for @code{ALIGN}, for compatibility with older linker
6025 scripts. It is most often seen when setting the address of an output
6028 @item DATA_SEGMENT_ALIGN(@var{maxpagesize}, @var{commonpagesize})
6029 @kindex DATA_SEGMENT_ALIGN(@var{maxpagesize}, @var{commonpagesize})
6030 This is equivalent to either
6032 (ALIGN(@var{maxpagesize}) + (. & (@var{maxpagesize} - 1)))
6036 (ALIGN(@var{maxpagesize}) + (. & (@var{maxpagesize} - @var{commonpagesize})))
6039 depending on whether the latter uses fewer @var{commonpagesize} sized pages
6040 for the data segment (area between the result of this expression and
6041 @code{DATA_SEGMENT_END}) than the former or not.
6042 If the latter form is used, it means @var{commonpagesize} bytes of runtime
6043 memory will be saved at the expense of up to @var{commonpagesize} wasted
6044 bytes in the on-disk file.
6046 This expression can only be used directly in @code{SECTIONS} commands, not in
6047 any output section descriptions and only once in the linker script.
6048 @var{commonpagesize} should be less or equal to @var{maxpagesize} and should
6049 be the system page size the object wants to be optimized for (while still
6050 working on system page sizes up to @var{maxpagesize}).
6055 . = DATA_SEGMENT_ALIGN(0x10000, 0x2000);
6058 @item DATA_SEGMENT_END(@var{exp})
6059 @kindex DATA_SEGMENT_END(@var{exp})
6060 This defines the end of data segment for @code{DATA_SEGMENT_ALIGN}
6061 evaluation purposes.
6064 . = DATA_SEGMENT_END(.);
6067 @item DATA_SEGMENT_RELRO_END(@var{offset}, @var{exp})
6068 @kindex DATA_SEGMENT_RELRO_END(@var{offset}, @var{exp})
6069 This defines the end of the @code{PT_GNU_RELRO} segment when
6070 @samp{-z relro} option is used.
6071 When @samp{-z relro} option is not present, @code{DATA_SEGMENT_RELRO_END}
6072 does nothing, otherwise @code{DATA_SEGMENT_ALIGN} is padded so that
6073 @var{exp} + @var{offset} is aligned to the most commonly used page
6074 boundary for particular target. If present in the linker script,
6075 it must always come in between @code{DATA_SEGMENT_ALIGN} and
6076 @code{DATA_SEGMENT_END}. Evaluates to the second argument plus any
6077 padding needed at the end of the @code{PT_GNU_RELRO} segment due to
6081 . = DATA_SEGMENT_RELRO_END(24, .);
6084 @item DEFINED(@var{symbol})
6085 @kindex DEFINED(@var{symbol})
6086 @cindex symbol defaults
6087 Return 1 if @var{symbol} is in the linker global symbol table and is
6088 defined before the statement using DEFINED in the script, otherwise
6089 return 0. You can use this function to provide
6090 default values for symbols. For example, the following script fragment
6091 shows how to set a global symbol @samp{begin} to the first location in
6092 the @samp{.text} section---but if a symbol called @samp{begin} already
6093 existed, its value is preserved:
6099 begin = DEFINED(begin) ? begin : . ;
6107 @item LENGTH(@var{memory})
6108 @kindex LENGTH(@var{memory})
6109 Return the length of the memory region named @var{memory}.
6111 @item LOADADDR(@var{section})
6112 @kindex LOADADDR(@var{section})
6113 @cindex section load address in expression
6114 Return the absolute LMA of the named @var{section}. (@pxref{Output
6117 @item LOG2CEIL(@var{exp})
6118 @kindex LOG2CEIL(@var{exp})
6119 Return the binary logarithm of @var{exp} rounded towards infinity.
6120 @code{LOG2CEIL(0)} returns 0.
6123 @item MAX(@var{exp1}, @var{exp2})
6124 Returns the maximum of @var{exp1} and @var{exp2}.
6127 @item MIN(@var{exp1}, @var{exp2})
6128 Returns the minimum of @var{exp1} and @var{exp2}.
6130 @item NEXT(@var{exp})
6131 @kindex NEXT(@var{exp})
6132 @cindex unallocated address, next
6133 Return the next unallocated address that is a multiple of @var{exp}.
6134 This function is closely related to @code{ALIGN(@var{exp})}; unless you
6135 use the @code{MEMORY} command to define discontinuous memory for the
6136 output file, the two functions are equivalent.
6138 @item ORIGIN(@var{memory})
6139 @kindex ORIGIN(@var{memory})
6140 Return the origin of the memory region named @var{memory}.
6142 @item SEGMENT_START(@var{segment}, @var{default})
6143 @kindex SEGMENT_START(@var{segment}, @var{default})
6144 Return the base address of the named @var{segment}. If an explicit
6145 value has already been given for this segment (with a command-line
6146 @samp{-T} option) then that value will be returned otherwise the value
6147 will be @var{default}. At present, the @samp{-T} command-line option
6148 can only be used to set the base address for the ``text'', ``data'', and
6149 ``bss'' sections, but you can use @code{SEGMENT_START} with any segment
6152 @item SIZEOF(@var{section})
6153 @kindex SIZEOF(@var{section})
6154 @cindex section size
6155 Return the size in bytes of the named @var{section}, if that section has
6156 been allocated. If the section has not been allocated when this is
6157 evaluated, the linker will report an error. In the following example,
6158 @code{symbol_1} and @code{symbol_2} are assigned identical values:
6167 symbol_1 = .end - .start ;
6168 symbol_2 = SIZEOF(.output);
6173 @item SIZEOF_HEADERS
6174 @itemx sizeof_headers
6175 @kindex SIZEOF_HEADERS
6177 Return the size in bytes of the output file's headers. This is
6178 information which appears at the start of the output file. You can use
6179 this number when setting the start address of the first section, if you
6180 choose, to facilitate paging.
6182 @cindex not enough room for program headers
6183 @cindex program headers, not enough room
6184 When producing an ELF output file, if the linker script uses the
6185 @code{SIZEOF_HEADERS} builtin function, the linker must compute the
6186 number of program headers before it has determined all the section
6187 addresses and sizes. If the linker later discovers that it needs
6188 additional program headers, it will report an error @samp{not enough
6189 room for program headers}. To avoid this error, you must avoid using
6190 the @code{SIZEOF_HEADERS} function, or you must rework your linker
6191 script to avoid forcing the linker to use additional program headers, or
6192 you must define the program headers yourself using the @code{PHDRS}
6193 command (@pxref{PHDRS}).
6196 @node Implicit Linker Scripts
6197 @section Implicit Linker Scripts
6198 @cindex implicit linker scripts
6199 If you specify a linker input file which the linker can not recognize as
6200 an object file or an archive file, it will try to read the file as a
6201 linker script. If the file can not be parsed as a linker script, the
6202 linker will report an error.
6204 An implicit linker script will not replace the default linker script.
6206 Typically an implicit linker script would contain only symbol
6207 assignments, or the @code{INPUT}, @code{GROUP}, or @code{VERSION}
6210 Any input files read because of an implicit linker script will be read
6211 at the position in the command line where the implicit linker script was
6212 read. This can affect archive searching.
6215 @node Machine Dependent
6216 @chapter Machine Dependent Features
6218 @cindex machine dependencies
6219 @command{ld} has additional features on some platforms; the following
6220 sections describe them. Machines where @command{ld} has no additional
6221 functionality are not listed.
6225 * H8/300:: @command{ld} and the H8/300
6228 * i960:: @command{ld} and the Intel 960 family
6231 * M68HC11/68HC12:: @code{ld} and the Motorola 68HC11 and 68HC12 families
6234 * ARM:: @command{ld} and the ARM family
6237 * HPPA ELF32:: @command{ld} and HPPA 32-bit ELF
6240 * M68K:: @command{ld} and the Motorola 68K family
6243 * MIPS:: @command{ld} and the MIPS family
6246 * MMIX:: @command{ld} and MMIX
6249 * MSP430:: @command{ld} and MSP430
6252 * NDS32:: @command{ld} and NDS32
6255 * Nios II:: @command{ld} and the Altera Nios II
6258 * PowerPC ELF32:: @command{ld} and PowerPC 32-bit ELF Support
6261 * PowerPC64 ELF64:: @command{ld} and PowerPC64 64-bit ELF Support
6264 * SPU ELF:: @command{ld} and SPU ELF Support
6267 * TI COFF:: @command{ld} and TI COFF
6270 * WIN32:: @command{ld} and WIN32 (cygwin/mingw)
6273 * Xtensa:: @command{ld} and Xtensa Processors
6284 @section @command{ld} and the H8/300
6286 @cindex H8/300 support
6287 For the H8/300, @command{ld} can perform these global optimizations when
6288 you specify the @samp{--relax} command-line option.
6291 @cindex relaxing on H8/300
6292 @item relaxing address modes
6293 @command{ld} finds all @code{jsr} and @code{jmp} instructions whose
6294 targets are within eight bits, and turns them into eight-bit
6295 program-counter relative @code{bsr} and @code{bra} instructions,
6298 @cindex synthesizing on H8/300
6299 @item synthesizing instructions
6300 @c FIXME: specifically mov.b, or any mov instructions really? -> mov.b only, at least on H8, H8H, H8S
6301 @command{ld} finds all @code{mov.b} instructions which use the
6302 sixteen-bit absolute address form, but refer to the top
6303 page of memory, and changes them to use the eight-bit address form.
6304 (That is: the linker turns @samp{mov.b @code{@@}@var{aa}:16} into
6305 @samp{mov.b @code{@@}@var{aa}:8} whenever the address @var{aa} is in the
6306 top page of memory).
6308 @command{ld} finds all @code{mov} instructions which use the register
6309 indirect with 32-bit displacement addressing mode, but use a small
6310 displacement inside 16-bit displacement range, and changes them to use
6311 the 16-bit displacement form. (That is: the linker turns @samp{mov.b
6312 @code{@@}@var{d}:32,ERx} into @samp{mov.b @code{@@}@var{d}:16,ERx}
6313 whenever the displacement @var{d} is in the 16 bit signed integer
6314 range. Only implemented in ELF-format ld).
6316 @item bit manipulation instructions
6317 @command{ld} finds all bit manipulation instructions like @code{band, bclr,
6318 biand, bild, bior, bist, bixor, bld, bnot, bor, bset, bst, btst, bxor}
6319 which use 32 bit and 16 bit absolute address form, but refer to the top
6320 page of memory, and changes them to use the 8 bit address form.
6321 (That is: the linker turns @samp{bset #xx:3,@code{@@}@var{aa}:32} into
6322 @samp{bset #xx:3,@code{@@}@var{aa}:8} whenever the address @var{aa} is in
6323 the top page of memory).
6325 @item system control instructions
6326 @command{ld} finds all @code{ldc.w, stc.w} instructions which use the
6327 32 bit absolute address form, but refer to the top page of memory, and
6328 changes them to use 16 bit address form.
6329 (That is: the linker turns @samp{ldc.w @code{@@}@var{aa}:32,ccr} into
6330 @samp{ldc.w @code{@@}@var{aa}:16,ccr} whenever the address @var{aa} is in
6331 the top page of memory).
6341 @c This stuff is pointless to say unless you're especially concerned
6342 @c with Renesas chips; don't enable it for generic case, please.
6344 @chapter @command{ld} and Other Renesas Chips
6346 @command{ld} also supports the Renesas (formerly Hitachi) H8/300H,
6347 H8/500, and SH chips. No special features, commands, or command-line
6348 options are required for these chips.
6358 @section @command{ld} and the Intel 960 Family
6360 @cindex i960 support
6362 You can use the @samp{-A@var{architecture}} command line option to
6363 specify one of the two-letter names identifying members of the 960
6364 family; the option specifies the desired output target, and warns of any
6365 incompatible instructions in the input files. It also modifies the
6366 linker's search strategy for archive libraries, to support the use of
6367 libraries specific to each particular architecture, by including in the
6368 search loop names suffixed with the string identifying the architecture.
6370 For example, if your @command{ld} command line included @w{@samp{-ACA}} as
6371 well as @w{@samp{-ltry}}, the linker would look (in its built-in search
6372 paths, and in any paths you specify with @samp{-L}) for a library with
6385 The first two possibilities would be considered in any event; the last
6386 two are due to the use of @w{@samp{-ACA}}.
6388 You can meaningfully use @samp{-A} more than once on a command line, since
6389 the 960 architecture family allows combination of target architectures; each
6390 use will add another pair of name variants to search for when @w{@samp{-l}}
6391 specifies a library.
6393 @cindex @option{--relax} on i960
6394 @cindex relaxing on i960
6395 @command{ld} supports the @samp{--relax} option for the i960 family. If
6396 you specify @samp{--relax}, @command{ld} finds all @code{balx} and
6397 @code{calx} instructions whose targets are within 24 bits, and turns
6398 them into 24-bit program-counter relative @code{bal} and @code{cal}
6399 instructions, respectively. @command{ld} also turns @code{cal}
6400 instructions into @code{bal} instructions when it determines that the
6401 target subroutine is a leaf routine (that is, the target subroutine does
6402 not itself call any subroutines).
6419 @node M68HC11/68HC12
6420 @section @command{ld} and the Motorola 68HC11 and 68HC12 families
6422 @cindex M68HC11 and 68HC12 support
6424 @subsection Linker Relaxation
6426 For the Motorola 68HC11, @command{ld} can perform these global
6427 optimizations when you specify the @samp{--relax} command-line option.
6430 @cindex relaxing on M68HC11
6431 @item relaxing address modes
6432 @command{ld} finds all @code{jsr} and @code{jmp} instructions whose
6433 targets are within eight bits, and turns them into eight-bit
6434 program-counter relative @code{bsr} and @code{bra} instructions,
6437 @command{ld} also looks at all 16-bit extended addressing modes and
6438 transforms them in a direct addressing mode when the address is in
6439 page 0 (between 0 and 0x0ff).
6441 @item relaxing gcc instruction group
6442 When @command{gcc} is called with @option{-mrelax}, it can emit group
6443 of instructions that the linker can optimize to use a 68HC11 direct
6444 addressing mode. These instructions consists of @code{bclr} or
6445 @code{bset} instructions.
6449 @subsection Trampoline Generation
6451 @cindex trampoline generation on M68HC11
6452 @cindex trampoline generation on M68HC12
6453 For 68HC11 and 68HC12, @command{ld} can generate trampoline code to
6454 call a far function using a normal @code{jsr} instruction. The linker
6455 will also change the relocation to some far function to use the
6456 trampoline address instead of the function address. This is typically the
6457 case when a pointer to a function is taken. The pointer will in fact
6458 point to the function trampoline.
6466 @section @command{ld} and the ARM family
6468 @cindex ARM interworking support
6469 @kindex --support-old-code
6470 For the ARM, @command{ld} will generate code stubs to allow functions calls
6471 between ARM and Thumb code. These stubs only work with code that has
6472 been compiled and assembled with the @samp{-mthumb-interwork} command
6473 line option. If it is necessary to link with old ARM object files or
6474 libraries, which have not been compiled with the -mthumb-interwork
6475 option then the @samp{--support-old-code} command line switch should be
6476 given to the linker. This will make it generate larger stub functions
6477 which will work with non-interworking aware ARM code. Note, however,
6478 the linker does not support generating stubs for function calls to
6479 non-interworking aware Thumb code.
6481 @cindex thumb entry point
6482 @cindex entry point, thumb
6483 @kindex --thumb-entry=@var{entry}
6484 The @samp{--thumb-entry} switch is a duplicate of the generic
6485 @samp{--entry} switch, in that it sets the program's starting address.
6486 But it also sets the bottom bit of the address, so that it can be
6487 branched to using a BX instruction, and the program will start
6488 executing in Thumb mode straight away.
6490 @cindex PE import table prefixing
6491 @kindex --use-nul-prefixed-import-tables
6492 The @samp{--use-nul-prefixed-import-tables} switch is specifying, that
6493 the import tables idata4 and idata5 have to be generated with a zero
6494 element prefix for import libraries. This is the old style to generate
6495 import tables. By default this option is turned off.
6499 The @samp{--be8} switch instructs @command{ld} to generate BE8 format
6500 executables. This option is only valid when linking big-endian
6501 objects - ie ones which have been assembled with the @option{-EB}
6502 option. The resulting image will contain big-endian data and
6506 @kindex --target1-rel
6507 @kindex --target1-abs
6508 The @samp{R_ARM_TARGET1} relocation is typically used for entries in the
6509 @samp{.init_array} section. It is interpreted as either @samp{R_ARM_REL32}
6510 or @samp{R_ARM_ABS32}, depending on the target. The @samp{--target1-rel}
6511 and @samp{--target1-abs} switches override the default.
6514 @kindex --target2=@var{type}
6515 The @samp{--target2=type} switch overrides the default definition of the
6516 @samp{R_ARM_TARGET2} relocation. Valid values for @samp{type}, their
6517 meanings, and target defaults are as follows:
6520 @samp{R_ARM_REL32} (arm*-*-elf, arm*-*-eabi)
6522 @samp{R_ARM_ABS32} (arm*-*-symbianelf)
6524 @samp{R_ARM_GOT_PREL} (arm*-*-linux, arm*-*-*bsd)
6529 The @samp{R_ARM_V4BX} relocation (defined by the ARM AAELF
6530 specification) enables objects compiled for the ARMv4 architecture to be
6531 interworking-safe when linked with other objects compiled for ARMv4t, but
6532 also allows pure ARMv4 binaries to be built from the same ARMv4 objects.
6534 In the latter case, the switch @option{--fix-v4bx} must be passed to the
6535 linker, which causes v4t @code{BX rM} instructions to be rewritten as
6536 @code{MOV PC,rM}, since v4 processors do not have a @code{BX} instruction.
6538 In the former case, the switch should not be used, and @samp{R_ARM_V4BX}
6539 relocations are ignored.
6541 @cindex FIX_V4BX_INTERWORKING
6542 @kindex --fix-v4bx-interworking
6543 Replace @code{BX rM} instructions identified by @samp{R_ARM_V4BX}
6544 relocations with a branch to the following veneer:
6552 This allows generation of libraries/applications that work on ARMv4 cores
6553 and are still interworking safe. Note that the above veneer clobbers the
6554 condition flags, so may cause incorrect program behavior in rare cases.
6558 The @samp{--use-blx} switch enables the linker to use ARM/Thumb
6559 BLX instructions (available on ARMv5t and above) in various
6560 situations. Currently it is used to perform calls via the PLT from Thumb
6561 code using BLX rather than using BX and a mode-switching stub before
6562 each PLT entry. This should lead to such calls executing slightly faster.
6564 This option is enabled implicitly for SymbianOS, so there is no need to
6565 specify it if you are using that target.
6567 @cindex VFP11_DENORM_FIX
6568 @kindex --vfp11-denorm-fix
6569 The @samp{--vfp11-denorm-fix} switch enables a link-time workaround for a
6570 bug in certain VFP11 coprocessor hardware, which sometimes allows
6571 instructions with denorm operands (which must be handled by support code)
6572 to have those operands overwritten by subsequent instructions before
6573 the support code can read the intended values.
6575 The bug may be avoided in scalar mode if you allow at least one
6576 intervening instruction between a VFP11 instruction which uses a register
6577 and another instruction which writes to the same register, or at least two
6578 intervening instructions if vector mode is in use. The bug only affects
6579 full-compliance floating-point mode: you do not need this workaround if
6580 you are using "runfast" mode. Please contact ARM for further details.
6582 If you know you are using buggy VFP11 hardware, you can
6583 enable this workaround by specifying the linker option
6584 @samp{--vfp-denorm-fix=scalar} if you are using the VFP11 scalar
6585 mode only, or @samp{--vfp-denorm-fix=vector} if you are using
6586 vector mode (the latter also works for scalar code). The default is
6587 @samp{--vfp-denorm-fix=none}.
6589 If the workaround is enabled, instructions are scanned for
6590 potentially-troublesome sequences, and a veneer is created for each
6591 such sequence which may trigger the erratum. The veneer consists of the
6592 first instruction of the sequence and a branch back to the subsequent
6593 instruction. The original instruction is then replaced with a branch to
6594 the veneer. The extra cycles required to call and return from the veneer
6595 are sufficient to avoid the erratum in both the scalar and vector cases.
6597 @cindex ARM1176 erratum workaround
6598 @kindex --fix-arm1176
6599 @kindex --no-fix-arm1176
6600 The @samp{--fix-arm1176} switch enables a link-time workaround for an erratum
6601 in certain ARM1176 processors. The workaround is enabled by default if you
6602 are targeting ARM v6 (excluding ARM v6T2) or earlier. It can be disabled
6603 unconditionally by specifying @samp{--no-fix-arm1176}.
6605 Further information is available in the ``ARM1176JZ-S and ARM1176JZF-S
6606 Programmer Advice Notice'' available on the ARM documentation website at:
6607 http://infocenter.arm.com/.
6609 @cindex NO_ENUM_SIZE_WARNING
6610 @kindex --no-enum-size-warning
6611 The @option{--no-enum-size-warning} switch prevents the linker from
6612 warning when linking object files that specify incompatible EABI
6613 enumeration size attributes. For example, with this switch enabled,
6614 linking of an object file using 32-bit enumeration values with another
6615 using enumeration values fitted into the smallest possible space will
6618 @cindex NO_WCHAR_SIZE_WARNING
6619 @kindex --no-wchar-size-warning
6620 The @option{--no-wchar-size-warning} switch prevents the linker from
6621 warning when linking object files that specify incompatible EABI
6622 @code{wchar_t} size attributes. For example, with this switch enabled,
6623 linking of an object file using 32-bit @code{wchar_t} values with another
6624 using 16-bit @code{wchar_t} values will not be diagnosed.
6627 @kindex --pic-veneer
6628 The @samp{--pic-veneer} switch makes the linker use PIC sequences for
6629 ARM/Thumb interworking veneers, even if the rest of the binary
6630 is not PIC. This avoids problems on uClinux targets where
6631 @samp{--emit-relocs} is used to generate relocatable binaries.
6633 @cindex STUB_GROUP_SIZE
6634 @kindex --stub-group-size=@var{N}
6635 The linker will automatically generate and insert small sequences of
6636 code into a linked ARM ELF executable whenever an attempt is made to
6637 perform a function call to a symbol that is too far away. The
6638 placement of these sequences of instructions - called stubs - is
6639 controlled by the command line option @option{--stub-group-size=N}.
6640 The placement is important because a poor choice can create a need for
6641 duplicate stubs, increasing the code size. The linker will try to
6642 group stubs together in order to reduce interruptions to the flow of
6643 code, but it needs guidance as to how big these groups should be and
6644 where they should be placed.
6646 The value of @samp{N}, the parameter to the
6647 @option{--stub-group-size=} option controls where the stub groups are
6648 placed. If it is negative then all stubs are placed after the first
6649 branch that needs them. If it is positive then the stubs can be
6650 placed either before or after the branches that need them. If the
6651 value of @samp{N} is 1 (either +1 or -1) then the linker will choose
6652 exactly where to place groups of stubs, using its built in heuristics.
6653 A value of @samp{N} greater than 1 (or smaller than -1) tells the
6654 linker that a single group of stubs can service at most @samp{N} bytes
6655 from the input sections.
6657 The default, if @option{--stub-group-size=} is not specified, is
6660 Farcalls stubs insertion is fully supported for the ARM-EABI target
6661 only, because it relies on object files properties not present
6664 @cindex Cortex-A8 erratum workaround
6665 @kindex --fix-cortex-a8
6666 @kindex --no-fix-cortex-a8
6667 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}.
6669 The erratum only affects Thumb-2 code. Please contact ARM for further details.
6671 @cindex Cortex-A53 erratum 835769 workaround
6672 @kindex --fix-cortex-a53-835769
6673 @kindex --no-fix-cortex-a53-835769
6674 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}.
6676 Please contact ARM for further details.
6678 @kindex --merge-exidx-entries
6679 @kindex --no-merge-exidx-entries
6680 @cindex Merging exidx entries
6681 The @samp{--no-merge-exidx-entries} switch disables the merging of adjacent exidx entries in debuginfo.
6684 @cindex 32-bit PLT entries
6685 The @samp{--long-plt} option enables the use of 16 byte PLT entries
6686 which support up to 4Gb of code. The default is to use 12 byte PLT
6687 entries which only support 512Mb of code.
6700 @section @command{ld} and HPPA 32-bit ELF Support
6701 @cindex HPPA multiple sub-space stubs
6702 @kindex --multi-subspace
6703 When generating a shared library, @command{ld} will by default generate
6704 import stubs suitable for use with a single sub-space application.
6705 The @samp{--multi-subspace} switch causes @command{ld} to generate export
6706 stubs, and different (larger) import stubs suitable for use with
6707 multiple sub-spaces.
6709 @cindex HPPA stub grouping
6710 @kindex --stub-group-size=@var{N}
6711 Long branch stubs and import/export stubs are placed by @command{ld} in
6712 stub sections located between groups of input sections.
6713 @samp{--stub-group-size} specifies the maximum size of a group of input
6714 sections handled by one stub section. Since branch offsets are signed,
6715 a stub section may serve two groups of input sections, one group before
6716 the stub section, and one group after it. However, when using
6717 conditional branches that require stubs, it may be better (for branch
6718 prediction) that stub sections only serve one group of input sections.
6719 A negative value for @samp{N} chooses this scheme, ensuring that
6720 branches to stubs always use a negative offset. Two special values of
6721 @samp{N} are recognized, @samp{1} and @samp{-1}. These both instruct
6722 @command{ld} to automatically size input section groups for the branch types
6723 detected, with the same behaviour regarding stub placement as other
6724 positive or negative values of @samp{N} respectively.
6726 Note that @samp{--stub-group-size} does not split input sections. A
6727 single input section larger than the group size specified will of course
6728 create a larger group (of one section). If input sections are too
6729 large, it may not be possible for a branch to reach its stub.
6742 @section @command{ld} and the Motorola 68K family
6744 @cindex Motorola 68K GOT generation
6745 @kindex --got=@var{type}
6746 The @samp{--got=@var{type}} option lets you choose the GOT generation scheme.
6747 The choices are @samp{single}, @samp{negative}, @samp{multigot} and
6748 @samp{target}. When @samp{target} is selected the linker chooses
6749 the default GOT generation scheme for the current target.
6750 @samp{single} tells the linker to generate a single GOT with
6751 entries only at non-negative offsets.
6752 @samp{negative} instructs the linker to generate a single GOT with
6753 entries at both negative and positive offsets. Not all environments
6755 @samp{multigot} allows the linker to generate several GOTs in the
6756 output file. All GOT references from a single input object
6757 file access the same GOT, but references from different input object
6758 files might access different GOTs. Not all environments support such GOTs.
6771 @section @command{ld} and the MIPS family
6773 @cindex MIPS microMIPS instruction choice selection
6776 The @samp{--insn32} and @samp{--no-insn32} options control the choice of
6777 microMIPS instructions used in code generated by the linker, such as that
6778 in the PLT or lazy binding stubs, or in relaxation. If @samp{--insn32} is
6779 used, then the linker only uses 32-bit instruction encodings. By default
6780 or if @samp{--no-insn32} is used, all instruction encodings are used,
6781 including 16-bit ones where possible.
6794 @section @code{ld} and MMIX
6795 For MMIX, there is a choice of generating @code{ELF} object files or
6796 @code{mmo} object files when linking. The simulator @code{mmix}
6797 understands the @code{mmo} format. The binutils @code{objcopy} utility
6798 can translate between the two formats.
6800 There is one special section, the @samp{.MMIX.reg_contents} section.
6801 Contents in this section is assumed to correspond to that of global
6802 registers, and symbols referring to it are translated to special symbols,
6803 equal to registers. In a final link, the start address of the
6804 @samp{.MMIX.reg_contents} section corresponds to the first allocated
6805 global register multiplied by 8. Register @code{$255} is not included in
6806 this section; it is always set to the program entry, which is at the
6807 symbol @code{Main} for @code{mmo} files.
6809 Global symbols with the prefix @code{__.MMIX.start.}, for example
6810 @code{__.MMIX.start..text} and @code{__.MMIX.start..data} are special.
6811 The default linker script uses these to set the default start address
6814 Initial and trailing multiples of zero-valued 32-bit words in a section,
6815 are left out from an mmo file.
6828 @section @code{ld} and MSP430
6829 For the MSP430 it is possible to select the MPU architecture. The flag @samp{-m [mpu type]}
6830 will select an appropriate linker script for selected MPU type. (To get a list of known MPUs
6831 just pass @samp{-m help} option to the linker).
6833 @cindex MSP430 extra sections
6834 The linker will recognize some extra sections which are MSP430 specific:
6837 @item @samp{.vectors}
6838 Defines a portion of ROM where interrupt vectors located.
6840 @item @samp{.bootloader}
6841 Defines the bootloader portion of the ROM (if applicable). Any code
6842 in this section will be uploaded to the MPU.
6844 @item @samp{.infomem}
6845 Defines an information memory section (if applicable). Any code in
6846 this section will be uploaded to the MPU.
6848 @item @samp{.infomemnobits}
6849 This is the same as the @samp{.infomem} section except that any code
6850 in this section will not be uploaded to the MPU.
6852 @item @samp{.noinit}
6853 Denotes a portion of RAM located above @samp{.bss} section.
6855 The last two sections are used by gcc.
6869 @section @code{ld} and NDS32
6870 @kindex relaxing on NDS32
6871 For NDS32, there are some options to select relaxation behavior. The linker
6872 relaxes objects according to these options.
6875 @item @samp{--m[no-]fp-as-gp}
6876 Disable/enable fp-as-gp relaxation.
6878 @item @samp{--mexport-symbols=FILE}
6879 Exporting symbols and their address into FILE as linker script.
6881 @item @samp{--m[no-]ex9}
6882 Disable/enable link-time EX9 relaxation.
6884 @item @samp{--mexport-ex9=FILE}
6885 Export the EX9 table after linking.
6887 @item @samp{--mimport-ex9=FILE}
6888 Import the Ex9 table for EX9 relaxation.
6890 @item @samp{--mupdate-ex9}
6891 Update the existing EX9 table.
6893 @item @samp{--mex9-limit=NUM}
6894 Maximum number of entries in the ex9 table.
6896 @item @samp{--mex9-loop-aware}
6897 Avoid generating the EX9 instruction inside the loop.
6899 @item @samp{--m[no-]ifc}
6900 Disable/enable the link-time IFC optimization.
6902 @item @samp{--mifc-loop-aware}
6903 Avoid generating the IFC instruction inside the loop.
6917 @section @command{ld} and the Altera Nios II
6918 @cindex Nios II call relaxation
6919 @kindex --relax on Nios II
6921 Call and immediate jump instructions on Nios II processors are limited to
6922 transferring control to addresses in the same 256MB memory segment,
6923 which may result in @command{ld} giving
6924 @samp{relocation truncated to fit} errors with very large programs.
6925 The command-line option @option{--relax} enables the generation of
6926 trampolines that can access the entire 32-bit address space for calls
6927 outside the normal @code{call} and @code{jmpi} address range. These
6928 trampolines are inserted at section boundaries, so may not themselves
6929 be reachable if an input section and its associated call trampolines are
6932 The @option{--relax} option is enabled by default unless @option{-r}
6933 is also specified. You can disable trampoline generation by using the
6934 @option{--no-relax} linker option. You can also disable this optimization
6935 locally by using the @samp{set .noat} directive in assembly-language
6936 source files, as the linker-inserted trampolines use the @code{at}
6937 register as a temporary.
6939 Note that the linker @option{--relax} option is independent of assembler
6940 relaxation options, and that using the GNU assembler's @option{-relax-all}
6941 option interferes with the linker's more selective call instruction relaxation.
6954 @section @command{ld} and PowerPC 32-bit ELF Support
6955 @cindex PowerPC long branches
6956 @kindex --relax on PowerPC
6957 Branches on PowerPC processors are limited to a signed 26-bit
6958 displacement, which may result in @command{ld} giving
6959 @samp{relocation truncated to fit} errors with very large programs.
6960 @samp{--relax} enables the generation of trampolines that can access
6961 the entire 32-bit address space. These trampolines are inserted at
6962 section boundaries, so may not themselves be reachable if an input
6963 section exceeds 33M in size. You may combine @samp{-r} and
6964 @samp{--relax} to add trampolines in a partial link. In that case
6965 both branches to undefined symbols and inter-section branches are also
6966 considered potentially out of range, and trampolines inserted.
6968 @cindex PowerPC ELF32 options
6973 Current PowerPC GCC accepts a @samp{-msecure-plt} option that
6974 generates code capable of using a newer PLT and GOT layout that has
6975 the security advantage of no executable section ever needing to be
6976 writable and no writable section ever being executable. PowerPC
6977 @command{ld} will generate this layout, including stubs to access the
6978 PLT, if all input files (including startup and static libraries) were
6979 compiled with @samp{-msecure-plt}. @samp{--bss-plt} forces the old
6980 BSS PLT (and GOT layout) which can give slightly better performance.
6982 @kindex --secure-plt
6984 @command{ld} will use the new PLT and GOT layout if it is linking new
6985 @samp{-fpic} or @samp{-fPIC} code, but does not do so automatically
6986 when linking non-PIC code. This option requests the new PLT and GOT
6987 layout. A warning will be given if some object file requires the old
6993 The new secure PLT and GOT are placed differently relative to other
6994 sections compared to older BSS PLT and GOT placement. The location of
6995 @code{.plt} must change because the new secure PLT is an initialized
6996 section while the old PLT is uninitialized. The reason for the
6997 @code{.got} change is more subtle: The new placement allows
6998 @code{.got} to be read-only in applications linked with
6999 @samp{-z relro -z now}. However, this placement means that
7000 @code{.sdata} cannot always be used in shared libraries, because the
7001 PowerPC ABI accesses @code{.sdata} in shared libraries from the GOT
7002 pointer. @samp{--sdata-got} forces the old GOT placement. PowerPC
7003 GCC doesn't use @code{.sdata} in shared libraries, so this option is
7004 really only useful for other compilers that may do so.
7006 @cindex PowerPC stub symbols
7007 @kindex --emit-stub-syms
7008 @item --emit-stub-syms
7009 This option causes @command{ld} to label linker stubs with a local
7010 symbol that encodes the stub type and destination.
7012 @cindex PowerPC TLS optimization
7013 @kindex --no-tls-optimize
7014 @item --no-tls-optimize
7015 PowerPC @command{ld} normally performs some optimization of code
7016 sequences used to access Thread-Local Storage. Use this option to
7017 disable the optimization.
7030 @node PowerPC64 ELF64
7031 @section @command{ld} and PowerPC64 64-bit ELF Support
7033 @cindex PowerPC64 ELF64 options
7035 @cindex PowerPC64 stub grouping
7036 @kindex --stub-group-size
7037 @item --stub-group-size
7038 Long branch stubs, PLT call stubs and TOC adjusting stubs are placed
7039 by @command{ld} in stub sections located between groups of input sections.
7040 @samp{--stub-group-size} specifies the maximum size of a group of input
7041 sections handled by one stub section. Since branch offsets are signed,
7042 a stub section may serve two groups of input sections, one group before
7043 the stub section, and one group after it. However, when using
7044 conditional branches that require stubs, it may be better (for branch
7045 prediction) that stub sections only serve one group of input sections.
7046 A negative value for @samp{N} chooses this scheme, ensuring that
7047 branches to stubs always use a negative offset. Two special values of
7048 @samp{N} are recognized, @samp{1} and @samp{-1}. These both instruct
7049 @command{ld} to automatically size input section groups for the branch types
7050 detected, with the same behaviour regarding stub placement as other
7051 positive or negative values of @samp{N} respectively.
7053 Note that @samp{--stub-group-size} does not split input sections. A
7054 single input section larger than the group size specified will of course
7055 create a larger group (of one section). If input sections are too
7056 large, it may not be possible for a branch to reach its stub.
7058 @cindex PowerPC64 stub symbols
7059 @kindex --emit-stub-syms
7060 @item --emit-stub-syms
7061 This option causes @command{ld} to label linker stubs with a local
7062 symbol that encodes the stub type and destination.
7064 @cindex PowerPC64 dot symbols
7066 @kindex --no-dotsyms
7067 @item --dotsyms, --no-dotsyms
7068 These two options control how @command{ld} interprets version patterns
7069 in a version script. Older PowerPC64 compilers emitted both a
7070 function descriptor symbol with the same name as the function, and a
7071 code entry symbol with the name prefixed by a dot (@samp{.}). To
7072 properly version a function @samp{foo}, the version script thus needs
7073 to control both @samp{foo} and @samp{.foo}. The option
7074 @samp{--dotsyms}, on by default, automatically adds the required
7075 dot-prefixed patterns. Use @samp{--no-dotsyms} to disable this
7078 @cindex PowerPC64 TLS optimization
7079 @kindex --no-tls-optimize
7080 @item --no-tls-optimize
7081 PowerPC64 @command{ld} normally performs some optimization of code
7082 sequences used to access Thread-Local Storage. Use this option to
7083 disable the optimization.
7085 @cindex PowerPC64 OPD optimization
7086 @kindex --no-opd-optimize
7087 @item --no-opd-optimize
7088 PowerPC64 @command{ld} normally removes @code{.opd} section entries
7089 corresponding to deleted link-once functions, or functions removed by
7090 the action of @samp{--gc-sections} or linker script @code{/DISCARD/}.
7091 Use this option to disable @code{.opd} optimization.
7093 @cindex PowerPC64 OPD spacing
7094 @kindex --non-overlapping-opd
7095 @item --non-overlapping-opd
7096 Some PowerPC64 compilers have an option to generate compressed
7097 @code{.opd} entries spaced 16 bytes apart, overlapping the third word,
7098 the static chain pointer (unused in C) with the first word of the next
7099 entry. This option expands such entries to the full 24 bytes.
7101 @cindex PowerPC64 TOC optimization
7102 @kindex --no-toc-optimize
7103 @item --no-toc-optimize
7104 PowerPC64 @command{ld} normally removes unused @code{.toc} section
7105 entries. Such entries are detected by examining relocations that
7106 reference the TOC in code sections. A reloc in a deleted code section
7107 marks a TOC word as unneeded, while a reloc in a kept code section
7108 marks a TOC word as needed. Since the TOC may reference itself, TOC
7109 relocs are also examined. TOC words marked as both needed and
7110 unneeded will of course be kept. TOC words without any referencing
7111 reloc are assumed to be part of a multi-word entry, and are kept or
7112 discarded as per the nearest marked preceding word. This works
7113 reliably for compiler generated code, but may be incorrect if assembly
7114 code is used to insert TOC entries. Use this option to disable the
7117 @cindex PowerPC64 multi-TOC
7118 @kindex --no-multi-toc
7119 @item --no-multi-toc
7120 If given any toc option besides @code{-mcmodel=medium} or
7121 @code{-mcmodel=large}, PowerPC64 GCC generates code for a TOC model
7123 entries are accessed with a 16-bit offset from r2. This limits the
7124 total TOC size to 64K. PowerPC64 @command{ld} extends this limit by
7125 grouping code sections such that each group uses less than 64K for its
7126 TOC entries, then inserts r2 adjusting stubs between inter-group
7127 calls. @command{ld} does not split apart input sections, so cannot
7128 help if a single input file has a @code{.toc} section that exceeds
7129 64K, most likely from linking multiple files with @command{ld -r}.
7130 Use this option to turn off this feature.
7132 @cindex PowerPC64 TOC sorting
7133 @kindex --no-toc-sort
7135 By default, @command{ld} sorts TOC sections so that those whose file
7136 happens to have a section called @code{.init} or @code{.fini} are
7137 placed first, followed by TOC sections referenced by code generated
7138 with PowerPC64 gcc's @code{-mcmodel=small}, and lastly TOC sections
7139 referenced only by code generated with PowerPC64 gcc's
7140 @code{-mcmodel=medium} or @code{-mcmodel=large} options. Doing this
7141 results in better TOC grouping for multi-TOC. Use this option to turn
7144 @cindex PowerPC64 PLT stub alignment
7146 @kindex --no-plt-align
7148 @itemx --no-plt-align
7149 Use these options to control whether individual PLT call stubs are
7150 padded so that they don't cross a 32-byte boundary, or to the
7151 specified power of two boundary when using @code{--plt-align=}. Note
7152 that this isn't alignment in the usual sense. By default PLT call
7153 stubs are packed tightly.
7155 @cindex PowerPC64 PLT call stub static chain
7156 @kindex --plt-static-chain
7157 @kindex --no-plt-static-chain
7158 @item --plt-static-chain
7159 @itemx --no-plt-static-chain
7160 Use these options to control whether PLT call stubs load the static
7161 chain pointer (r11). @code{ld} defaults to not loading the static
7162 chain since there is never any need to do so on a PLT call.
7164 @cindex PowerPC64 PLT call stub thread safety
7165 @kindex --plt-thread-safe
7166 @kindex --no-plt-thread-safe
7167 @item --plt-thread-safe
7168 @itemx --no-thread-safe
7169 With power7's weakly ordered memory model, it is possible when using
7170 lazy binding for ld.so to update a plt entry in one thread and have
7171 another thread see the individual plt entry words update in the wrong
7172 order, despite ld.so carefully writing in the correct order and using
7173 memory write barriers. To avoid this we need some sort of read
7174 barrier in the call stub, or use LD_BIND_NOW=1. By default, @code{ld}
7175 looks for calls to commonly used functions that create threads, and if
7176 seen, adds the necessary barriers. Use these options to change the
7191 @section @command{ld} and SPU ELF Support
7193 @cindex SPU ELF options
7199 This option marks an executable as a PIC plugin module.
7201 @cindex SPU overlays
7202 @kindex --no-overlays
7204 Normally, @command{ld} recognizes calls to functions within overlay
7205 regions, and redirects such calls to an overlay manager via a stub.
7206 @command{ld} also provides a built-in overlay manager. This option
7207 turns off all this special overlay handling.
7209 @cindex SPU overlay stub symbols
7210 @kindex --emit-stub-syms
7211 @item --emit-stub-syms
7212 This option causes @command{ld} to label overlay stubs with a local
7213 symbol that encodes the stub type and destination.
7215 @cindex SPU extra overlay stubs
7216 @kindex --extra-overlay-stubs
7217 @item --extra-overlay-stubs
7218 This option causes @command{ld} to add overlay call stubs on all
7219 function calls out of overlay regions. Normally stubs are not added
7220 on calls to non-overlay regions.
7222 @cindex SPU local store size
7223 @kindex --local-store=lo:hi
7224 @item --local-store=lo:hi
7225 @command{ld} usually checks that a final executable for SPU fits in
7226 the address range 0 to 256k. This option may be used to change the
7227 range. Disable the check entirely with @option{--local-store=0:0}.
7230 @kindex --stack-analysis
7231 @item --stack-analysis
7232 SPU local store space is limited. Over-allocation of stack space
7233 unnecessarily limits space available for code and data, while
7234 under-allocation results in runtime failures. If given this option,
7235 @command{ld} will provide an estimate of maximum stack usage.
7236 @command{ld} does this by examining symbols in code sections to
7237 determine the extents of functions, and looking at function prologues
7238 for stack adjusting instructions. A call-graph is created by looking
7239 for relocations on branch instructions. The graph is then searched
7240 for the maximum stack usage path. Note that this analysis does not
7241 find calls made via function pointers, and does not handle recursion
7242 and other cycles in the call graph. Stack usage may be
7243 under-estimated if your code makes such calls. Also, stack usage for
7244 dynamic allocation, e.g. alloca, will not be detected. If a link map
7245 is requested, detailed information about each function's stack usage
7246 and calls will be given.
7249 @kindex --emit-stack-syms
7250 @item --emit-stack-syms
7251 This option, if given along with @option{--stack-analysis} will result
7252 in @command{ld} emitting stack sizing symbols for each function.
7253 These take the form @code{__stack_<function_name>} for global
7254 functions, and @code{__stack_<number>_<function_name>} for static
7255 functions. @code{<number>} is the section id in hex. The value of
7256 such symbols is the stack requirement for the corresponding function.
7257 The symbol size will be zero, type @code{STT_NOTYPE}, binding
7258 @code{STB_LOCAL}, and section @code{SHN_ABS}.
7272 @section @command{ld}'s Support for Various TI COFF Versions
7273 @cindex TI COFF versions
7274 @kindex --format=@var{version}
7275 The @samp{--format} switch allows selection of one of the various
7276 TI COFF versions. The latest of this writing is 2; versions 0 and 1 are
7277 also supported. The TI COFF versions also vary in header byte-order
7278 format; @command{ld} will read any version or byte order, but the output
7279 header format depends on the default specified by the specific target.
7292 @section @command{ld} and WIN32 (cygwin/mingw)
7294 This section describes some of the win32 specific @command{ld} issues.
7295 See @ref{Options,,Command Line Options} for detailed description of the
7296 command line options mentioned here.
7299 @cindex import libraries
7300 @item import libraries
7301 The standard Windows linker creates and uses so-called import
7302 libraries, which contains information for linking to dll's. They are
7303 regular static archives and are handled as any other static
7304 archive. The cygwin and mingw ports of @command{ld} have specific
7305 support for creating such libraries provided with the
7306 @samp{--out-implib} command line option.
7308 @item exporting DLL symbols
7309 @cindex exporting DLL symbols
7310 The cygwin/mingw @command{ld} has several ways to export symbols for dll's.
7313 @item using auto-export functionality
7314 @cindex using auto-export functionality
7315 By default @command{ld} exports symbols with the auto-export functionality,
7316 which is controlled by the following command line options:
7319 @item --export-all-symbols [This is the default]
7320 @item --exclude-symbols
7321 @item --exclude-libs
7322 @item --exclude-modules-for-implib
7323 @item --version-script
7326 When auto-export is in operation, @command{ld} will export all the non-local
7327 (global and common) symbols it finds in a DLL, with the exception of a few
7328 symbols known to belong to the system's runtime and libraries. As it will
7329 often not be desirable to export all of a DLL's symbols, which may include
7330 private functions that are not part of any public interface, the command-line
7331 options listed above may be used to filter symbols out from the list for
7332 exporting. The @samp{--output-def} option can be used in order to see the
7333 final list of exported symbols with all exclusions taken into effect.
7335 If @samp{--export-all-symbols} is not given explicitly on the
7336 command line, then the default auto-export behavior will be @emph{disabled}
7337 if either of the following are true:
7340 @item A DEF file is used.
7341 @item Any symbol in any object file was marked with the __declspec(dllexport) attribute.
7344 @item using a DEF file
7345 @cindex using a DEF file
7346 Another way of exporting symbols is using a DEF file. A DEF file is
7347 an ASCII file containing definitions of symbols which should be
7348 exported when a dll is created. Usually it is named @samp{<dll
7349 name>.def} and is added as any other object file to the linker's
7350 command line. The file's name must end in @samp{.def} or @samp{.DEF}.
7353 gcc -o <output> <objectfiles> <dll name>.def
7356 Using a DEF file turns off the normal auto-export behavior, unless the
7357 @samp{--export-all-symbols} option is also used.
7359 Here is an example of a DEF file for a shared library called @samp{xyz.dll}:
7362 LIBRARY "xyz.dll" BASE=0x20000000
7368 another_foo = abc.dll.afoo
7374 This example defines a DLL with a non-default base address and seven
7375 symbols in the export table. The third exported symbol @code{_bar} is an
7376 alias for the second. The fourth symbol, @code{another_foo} is resolved
7377 by "forwarding" to another module and treating it as an alias for
7378 @code{afoo} exported from the DLL @samp{abc.dll}. The final symbol
7379 @code{var1} is declared to be a data object. The @samp{doo} symbol in
7380 export library is an alias of @samp{foo}, which gets the string name
7381 in export table @samp{foo2}. The @samp{eoo} symbol is an data export
7382 symbol, which gets in export table the name @samp{var1}.
7384 The optional @code{LIBRARY <name>} command indicates the @emph{internal}
7385 name of the output DLL. If @samp{<name>} does not include a suffix,
7386 the default library suffix, @samp{.DLL} is appended.
7388 When the .DEF file is used to build an application, rather than a
7389 library, the @code{NAME <name>} command should be used instead of
7390 @code{LIBRARY}. If @samp{<name>} does not include a suffix, the default
7391 executable suffix, @samp{.EXE} is appended.
7393 With either @code{LIBRARY <name>} or @code{NAME <name>} the optional
7394 specification @code{BASE = <number>} may be used to specify a
7395 non-default base address for the image.
7397 If neither @code{LIBRARY <name>} nor @code{NAME <name>} is specified,
7398 or they specify an empty string, the internal name is the same as the
7399 filename specified on the command line.
7401 The complete specification of an export symbol is:
7405 ( ( ( <name1> [ = <name2> ] )
7406 | ( <name1> = <module-name> . <external-name>))
7407 [ @@ <integer> ] [NONAME] [DATA] [CONSTANT] [PRIVATE] [== <name3>] ) *
7410 Declares @samp{<name1>} as an exported symbol from the DLL, or declares
7411 @samp{<name1>} as an exported alias for @samp{<name2>}; or declares
7412 @samp{<name1>} as a "forward" alias for the symbol
7413 @samp{<external-name>} in the DLL @samp{<module-name>}.
7414 Optionally, the symbol may be exported by the specified ordinal
7415 @samp{<integer>} alias. The optional @samp{<name3>} is the to be used
7416 string in import/export table for the symbol.
7418 The optional keywords that follow the declaration indicate:
7420 @code{NONAME}: Do not put the symbol name in the DLL's export table. It
7421 will still be exported by its ordinal alias (either the value specified
7422 by the .def specification or, otherwise, the value assigned by the
7423 linker). The symbol name, however, does remain visible in the import
7424 library (if any), unless @code{PRIVATE} is also specified.
7426 @code{DATA}: The symbol is a variable or object, rather than a function.
7427 The import lib will export only an indirect reference to @code{foo} as
7428 the symbol @code{_imp__foo} (ie, @code{foo} must be resolved as
7431 @code{CONSTANT}: Like @code{DATA}, but put the undecorated @code{foo} as
7432 well as @code{_imp__foo} into the import library. Both refer to the
7433 read-only import address table's pointer to the variable, not to the
7434 variable itself. This can be dangerous. If the user code fails to add
7435 the @code{dllimport} attribute and also fails to explicitly add the
7436 extra indirection that the use of the attribute enforces, the
7437 application will behave unexpectedly.
7439 @code{PRIVATE}: Put the symbol in the DLL's export table, but do not put
7440 it into the static import library used to resolve imports at link time. The
7441 symbol can still be imported using the @code{LoadLibrary/GetProcAddress}
7442 API at runtime or by by using the GNU ld extension of linking directly to
7443 the DLL without an import library.
7445 See ld/deffilep.y in the binutils sources for the full specification of
7446 other DEF file statements
7448 @cindex creating a DEF file
7449 While linking a shared dll, @command{ld} is able to create a DEF file
7450 with the @samp{--output-def <file>} command line option.
7452 @item Using decorations
7453 @cindex Using decorations
7454 Another way of marking symbols for export is to modify the source code
7455 itself, so that when building the DLL each symbol to be exported is
7459 __declspec(dllexport) int a_variable
7460 __declspec(dllexport) void a_function(int with_args)
7463 All such symbols will be exported from the DLL. If, however,
7464 any of the object files in the DLL contain symbols decorated in
7465 this way, then the normal auto-export behavior is disabled, unless
7466 the @samp{--export-all-symbols} option is also used.
7468 Note that object files that wish to access these symbols must @emph{not}
7469 decorate them with dllexport. Instead, they should use dllimport,
7473 __declspec(dllimport) int a_variable
7474 __declspec(dllimport) void a_function(int with_args)
7477 This complicates the structure of library header files, because
7478 when included by the library itself the header must declare the
7479 variables and functions as dllexport, but when included by client
7480 code the header must declare them as dllimport. There are a number
7481 of idioms that are typically used to do this; often client code can
7482 omit the __declspec() declaration completely. See
7483 @samp{--enable-auto-import} and @samp{automatic data imports} for more
7487 @cindex automatic data imports
7488 @item automatic data imports
7489 The standard Windows dll format supports data imports from dlls only
7490 by adding special decorations (dllimport/dllexport), which let the
7491 compiler produce specific assembler instructions to deal with this
7492 issue. This increases the effort necessary to port existing Un*x
7493 code to these platforms, especially for large
7494 c++ libraries and applications. The auto-import feature, which was
7495 initially provided by Paul Sokolovsky, allows one to omit the
7496 decorations to achieve a behavior that conforms to that on POSIX/Un*x
7497 platforms. This feature is enabled with the @samp{--enable-auto-import}
7498 command-line option, although it is enabled by default on cygwin/mingw.
7499 The @samp{--enable-auto-import} option itself now serves mainly to
7500 suppress any warnings that are ordinarily emitted when linked objects
7501 trigger the feature's use.
7503 auto-import of variables does not always work flawlessly without
7504 additional assistance. Sometimes, you will see this message
7506 "variable '<var>' can't be auto-imported. Please read the
7507 documentation for ld's @code{--enable-auto-import} for details."
7509 The @samp{--enable-auto-import} documentation explains why this error
7510 occurs, and several methods that can be used to overcome this difficulty.
7511 One of these methods is the @emph{runtime pseudo-relocs} feature, described
7514 @cindex runtime pseudo-relocation
7515 For complex variables imported from DLLs (such as structs or classes),
7516 object files typically contain a base address for the variable and an
7517 offset (@emph{addend}) within the variable--to specify a particular
7518 field or public member, for instance. Unfortunately, the runtime loader used
7519 in win32 environments is incapable of fixing these references at runtime
7520 without the additional information supplied by dllimport/dllexport decorations.
7521 The standard auto-import feature described above is unable to resolve these
7524 The @samp{--enable-runtime-pseudo-relocs} switch allows these references to
7525 be resolved without error, while leaving the task of adjusting the references
7526 themselves (with their non-zero addends) to specialized code provided by the
7527 runtime environment. Recent versions of the cygwin and mingw environments and
7528 compilers provide this runtime support; older versions do not. However, the
7529 support is only necessary on the developer's platform; the compiled result will
7530 run without error on an older system.
7532 @samp{--enable-runtime-pseudo-relocs} is not the default; it must be explicitly
7535 @cindex direct linking to a dll
7536 @item direct linking to a dll
7537 The cygwin/mingw ports of @command{ld} support the direct linking,
7538 including data symbols, to a dll without the usage of any import
7539 libraries. This is much faster and uses much less memory than does the
7540 traditional import library method, especially when linking large
7541 libraries or applications. When @command{ld} creates an import lib, each
7542 function or variable exported from the dll is stored in its own bfd, even
7543 though a single bfd could contain many exports. The overhead involved in
7544 storing, loading, and processing so many bfd's is quite large, and explains the
7545 tremendous time, memory, and storage needed to link against particularly
7546 large or complex libraries when using import libs.
7548 Linking directly to a dll uses no extra command-line switches other than
7549 @samp{-L} and @samp{-l}, because @command{ld} already searches for a number
7550 of names to match each library. All that is needed from the developer's
7551 perspective is an understanding of this search, in order to force ld to
7552 select the dll instead of an import library.
7555 For instance, when ld is called with the argument @samp{-lxxx} it will attempt
7556 to find, in the first directory of its search path,
7568 before moving on to the next directory in the search path.
7570 (*) Actually, this is not @samp{cygxxx.dll} but in fact is @samp{<prefix>xxx.dll},
7571 where @samp{<prefix>} is set by the @command{ld} option
7572 @samp{--dll-search-prefix=<prefix>}. In the case of cygwin, the standard gcc spec
7573 file includes @samp{--dll-search-prefix=cyg}, so in effect we actually search for
7576 Other win32-based unix environments, such as mingw or pw32, may use other
7577 @samp{<prefix>}es, although at present only cygwin makes use of this feature. It
7578 was originally intended to help avoid name conflicts among dll's built for the
7579 various win32/un*x environments, so that (for example) two versions of a zlib dll
7580 could coexist on the same machine.
7582 The generic cygwin/mingw path layout uses a @samp{bin} directory for
7583 applications and dll's and a @samp{lib} directory for the import
7584 libraries (using cygwin nomenclature):
7590 libxxx.dll.a (in case of dll's)
7591 libxxx.a (in case of static archive)
7594 Linking directly to a dll without using the import library can be
7597 1. Use the dll directly by adding the @samp{bin} path to the link line
7599 gcc -Wl,-verbose -o a.exe -L../bin/ -lxxx
7602 However, as the dll's often have version numbers appended to their names
7603 (@samp{cygncurses-5.dll}) this will often fail, unless one specifies
7604 @samp{-L../bin -lncurses-5} to include the version. Import libs are generally
7605 not versioned, and do not have this difficulty.
7607 2. Create a symbolic link from the dll to a file in the @samp{lib}
7608 directory according to the above mentioned search pattern. This
7609 should be used to avoid unwanted changes in the tools needed for
7613 ln -s bin/cygxxx.dll lib/[cyg|lib|]xxx.dll[.a]
7616 Then you can link without any make environment changes.
7619 gcc -Wl,-verbose -o a.exe -L../lib/ -lxxx
7622 This technique also avoids the version number problems, because the following is
7629 libxxx.dll.a -> ../bin/cygxxx-5.dll
7632 Linking directly to a dll without using an import lib will work
7633 even when auto-import features are exercised, and even when
7634 @samp{--enable-runtime-pseudo-relocs} is used.
7636 Given the improvements in speed and memory usage, one might justifiably
7637 wonder why import libraries are used at all. There are three reasons:
7639 1. Until recently, the link-directly-to-dll functionality did @emph{not}
7640 work with auto-imported data.
7642 2. Sometimes it is necessary to include pure static objects within the
7643 import library (which otherwise contains only bfd's for indirection
7644 symbols that point to the exports of a dll). Again, the import lib
7645 for the cygwin kernel makes use of this ability, and it is not
7646 possible to do this without an import lib.
7648 3. Symbol aliases can only be resolved using an import lib. This is
7649 critical when linking against OS-supplied dll's (eg, the win32 API)
7650 in which symbols are usually exported as undecorated aliases of their
7651 stdcall-decorated assembly names.
7653 So, import libs are not going away. But the ability to replace
7654 true import libs with a simple symbolic link to (or a copy of)
7655 a dll, in many cases, is a useful addition to the suite of tools
7656 binutils makes available to the win32 developer. Given the
7657 massive improvements in memory requirements during linking, storage
7658 requirements, and linking speed, we expect that many developers
7659 will soon begin to use this feature whenever possible.
7661 @item symbol aliasing
7663 @item adding additional names
7664 Sometimes, it is useful to export symbols with additional names.
7665 A symbol @samp{foo} will be exported as @samp{foo}, but it can also be
7666 exported as @samp{_foo} by using special directives in the DEF file
7667 when creating the dll. This will affect also the optional created
7668 import library. Consider the following DEF file:
7671 LIBRARY "xyz.dll" BASE=0x61000000
7678 The line @samp{_foo = foo} maps the symbol @samp{foo} to @samp{_foo}.
7680 Another method for creating a symbol alias is to create it in the
7681 source code using the "weak" attribute:
7684 void foo () @{ /* Do something. */; @}
7685 void _foo () __attribute__ ((weak, alias ("foo")));
7688 See the gcc manual for more information about attributes and weak
7691 @item renaming symbols
7692 Sometimes it is useful to rename exports. For instance, the cygwin
7693 kernel does this regularly. A symbol @samp{_foo} can be exported as
7694 @samp{foo} but not as @samp{_foo} by using special directives in the
7695 DEF file. (This will also affect the import library, if it is
7696 created). In the following example:
7699 LIBRARY "xyz.dll" BASE=0x61000000
7705 The line @samp{_foo = foo} maps the exported symbol @samp{foo} to
7709 Note: using a DEF file disables the default auto-export behavior,
7710 unless the @samp{--export-all-symbols} command line option is used.
7711 If, however, you are trying to rename symbols, then you should list
7712 @emph{all} desired exports in the DEF file, including the symbols
7713 that are not being renamed, and do @emph{not} use the
7714 @samp{--export-all-symbols} option. If you list only the
7715 renamed symbols in the DEF file, and use @samp{--export-all-symbols}
7716 to handle the other symbols, then the both the new names @emph{and}
7717 the original names for the renamed symbols will be exported.
7718 In effect, you'd be aliasing those symbols, not renaming them,
7719 which is probably not what you wanted.
7721 @cindex weak externals
7722 @item weak externals
7723 The Windows object format, PE, specifies a form of weak symbols called
7724 weak externals. When a weak symbol is linked and the symbol is not
7725 defined, the weak symbol becomes an alias for some other symbol. There
7726 are three variants of weak externals:
7728 @item Definition is searched for in objects and libraries, historically
7729 called lazy externals.
7730 @item Definition is searched for only in other objects, not in libraries.
7731 This form is not presently implemented.
7732 @item No search; the symbol is an alias. This form is not presently
7735 As a GNU extension, weak symbols that do not specify an alternate symbol
7736 are supported. If the symbol is undefined when linking, the symbol
7737 uses a default value.
7739 @cindex aligned common symbols
7740 @item aligned common symbols
7741 As a GNU extension to the PE file format, it is possible to specify the
7742 desired alignment for a common symbol. This information is conveyed from
7743 the assembler or compiler to the linker by means of GNU-specific commands
7744 carried in the object file's @samp{.drectve} section, which are recognized
7745 by @command{ld} and respected when laying out the common symbols. Native
7746 tools will be able to process object files employing this GNU extension,
7747 but will fail to respect the alignment instructions, and may issue noisy
7748 warnings about unknown linker directives.
7763 @section @code{ld} and Xtensa Processors
7765 @cindex Xtensa processors
7766 The default @command{ld} behavior for Xtensa processors is to interpret
7767 @code{SECTIONS} commands so that lists of explicitly named sections in a
7768 specification with a wildcard file will be interleaved when necessary to
7769 keep literal pools within the range of PC-relative load offsets. For
7770 example, with the command:
7782 @command{ld} may interleave some of the @code{.literal}
7783 and @code{.text} sections from different object files to ensure that the
7784 literal pools are within the range of PC-relative load offsets. A valid
7785 interleaving might place the @code{.literal} sections from an initial
7786 group of files followed by the @code{.text} sections of that group of
7787 files. Then, the @code{.literal} sections from the rest of the files
7788 and the @code{.text} sections from the rest of the files would follow.
7790 @cindex @option{--relax} on Xtensa
7791 @cindex relaxing on Xtensa
7792 Relaxation is enabled by default for the Xtensa version of @command{ld} and
7793 provides two important link-time optimizations. The first optimization
7794 is to combine identical literal values to reduce code size. A redundant
7795 literal will be removed and all the @code{L32R} instructions that use it
7796 will be changed to reference an identical literal, as long as the
7797 location of the replacement literal is within the offset range of all
7798 the @code{L32R} instructions. The second optimization is to remove
7799 unnecessary overhead from assembler-generated ``longcall'' sequences of
7800 @code{L32R}/@code{CALLX@var{n}} when the target functions are within
7801 range of direct @code{CALL@var{n}} instructions.
7803 For each of these cases where an indirect call sequence can be optimized
7804 to a direct call, the linker will change the @code{CALLX@var{n}}
7805 instruction to a @code{CALL@var{n}} instruction, remove the @code{L32R}
7806 instruction, and remove the literal referenced by the @code{L32R}
7807 instruction if it is not used for anything else. Removing the
7808 @code{L32R} instruction always reduces code size but can potentially
7809 hurt performance by changing the alignment of subsequent branch targets.
7810 By default, the linker will always preserve alignments, either by
7811 switching some instructions between 24-bit encodings and the equivalent
7812 density instructions or by inserting a no-op in place of the @code{L32R}
7813 instruction that was removed. If code size is more important than
7814 performance, the @option{--size-opt} option can be used to prevent the
7815 linker from widening density instructions or inserting no-ops, except in
7816 a few cases where no-ops are required for correctness.
7818 The following Xtensa-specific command-line options can be used to
7821 @cindex Xtensa options
7824 When optimizing indirect calls to direct calls, optimize for code size
7825 more than performance. With this option, the linker will not insert
7826 no-ops or widen density instructions to preserve branch target
7827 alignment. There may still be some cases where no-ops are required to
7828 preserve the correctness of the code.
7836 @ifclear SingleFormat
7841 @cindex object file management
7842 @cindex object formats available
7844 The linker accesses object and archive files using the BFD libraries.
7845 These libraries allow the linker to use the same routines to operate on
7846 object files whatever the object file format. A different object file
7847 format can be supported simply by creating a new BFD back end and adding
7848 it to the library. To conserve runtime memory, however, the linker and
7849 associated tools are usually configured to support only a subset of the
7850 object file formats available. You can use @code{objdump -i}
7851 (@pxref{objdump,,objdump,binutils.info,The GNU Binary Utilities}) to
7852 list all the formats available for your configuration.
7854 @cindex BFD requirements
7855 @cindex requirements for BFD
7856 As with most implementations, BFD is a compromise between
7857 several conflicting requirements. The major factor influencing
7858 BFD design was efficiency: any time used converting between
7859 formats is time which would not have been spent had BFD not
7860 been involved. This is partly offset by abstraction payback; since
7861 BFD simplifies applications and back ends, more time and care
7862 may be spent optimizing algorithms for a greater speed.
7864 One minor artifact of the BFD solution which you should bear in
7865 mind is the potential for information loss. There are two places where
7866 useful information can be lost using the BFD mechanism: during
7867 conversion and during output. @xref{BFD information loss}.
7870 * BFD outline:: How it works: an outline of BFD
7874 @section How It Works: An Outline of BFD
7875 @cindex opening object files
7876 @include bfdsumm.texi
7879 @node Reporting Bugs
7880 @chapter Reporting Bugs
7881 @cindex bugs in @command{ld}
7882 @cindex reporting bugs in @command{ld}
7884 Your bug reports play an essential role in making @command{ld} reliable.
7886 Reporting a bug may help you by bringing a solution to your problem, or
7887 it may not. But in any case the principal function of a bug report is
7888 to help the entire community by making the next version of @command{ld}
7889 work better. Bug reports are your contribution to the maintenance of
7892 In order for a bug report to serve its purpose, you must include the
7893 information that enables us to fix the bug.
7896 * Bug Criteria:: Have you found a bug?
7897 * Bug Reporting:: How to report bugs
7901 @section Have You Found a Bug?
7902 @cindex bug criteria
7904 If you are not sure whether you have found a bug, here are some guidelines:
7907 @cindex fatal signal
7908 @cindex linker crash
7909 @cindex crash of linker
7911 If the linker gets a fatal signal, for any input whatever, that is a
7912 @command{ld} bug. Reliable linkers never crash.
7914 @cindex error on valid input
7916 If @command{ld} produces an error message for valid input, that is a bug.
7918 @cindex invalid input
7920 If @command{ld} does not produce an error message for invalid input, that
7921 may be a bug. In the general case, the linker can not verify that
7922 object files are correct.
7925 If you are an experienced user of linkers, your suggestions for
7926 improvement of @command{ld} are welcome in any case.
7930 @section How to Report Bugs
7932 @cindex @command{ld} bugs, reporting
7934 A number of companies and individuals offer support for @sc{gnu}
7935 products. If you obtained @command{ld} from a support organization, we
7936 recommend you contact that organization first.
7938 You can find contact information for many support companies and
7939 individuals in the file @file{etc/SERVICE} in the @sc{gnu} Emacs
7943 Otherwise, send bug reports for @command{ld} to
7947 The fundamental principle of reporting bugs usefully is this:
7948 @strong{report all the facts}. If you are not sure whether to state a
7949 fact or leave it out, state it!
7951 Often people omit facts because they think they know what causes the
7952 problem and assume that some details do not matter. Thus, you might
7953 assume that the name of a symbol you use in an example does not
7954 matter. Well, probably it does not, but one cannot be sure. Perhaps
7955 the bug is a stray memory reference which happens to fetch from the
7956 location where that name is stored in memory; perhaps, if the name
7957 were different, the contents of that location would fool the linker
7958 into doing the right thing despite the bug. Play it safe and give a
7959 specific, complete example. That is the easiest thing for you to do,
7960 and the most helpful.
7962 Keep in mind that the purpose of a bug report is to enable us to fix
7963 the bug if it is new to us. Therefore, always write your bug reports
7964 on the assumption that the bug has not been reported previously.
7966 Sometimes people give a few sketchy facts and ask, ``Does this ring a
7967 bell?'' This cannot help us fix a bug, so it is basically useless. We
7968 respond by asking for enough details to enable us to investigate.
7969 You might as well expedite matters by sending them to begin with.
7971 To enable us to fix the bug, you should include all these things:
7975 The version of @command{ld}. @command{ld} announces it if you start it with
7976 the @samp{--version} argument.
7978 Without this, we will not know whether there is any point in looking for
7979 the bug in the current version of @command{ld}.
7982 Any patches you may have applied to the @command{ld} source, including any
7983 patches made to the @code{BFD} library.
7986 The type of machine you are using, and the operating system name and
7990 What compiler (and its version) was used to compile @command{ld}---e.g.
7994 The command arguments you gave the linker to link your example and
7995 observe the bug. To guarantee you will not omit something important,
7996 list them all. A copy of the Makefile (or the output from make) is
7999 If we were to try to guess the arguments, we would probably guess wrong
8000 and then we might not encounter the bug.
8003 A complete input file, or set of input files, that will reproduce the
8004 bug. It is generally most helpful to send the actual object files
8005 provided that they are reasonably small. Say no more than 10K. For
8006 bigger files you can either make them available by FTP or HTTP or else
8007 state that you are willing to send the object file(s) to whomever
8008 requests them. (Note - your email will be going to a mailing list, so
8009 we do not want to clog it up with large attachments). But small
8010 attachments are best.
8012 If the source files were assembled using @code{gas} or compiled using
8013 @code{gcc}, then it may be OK to send the source files rather than the
8014 object files. In this case, be sure to say exactly what version of
8015 @code{gas} or @code{gcc} was used to produce the object files. Also say
8016 how @code{gas} or @code{gcc} were configured.
8019 A description of what behavior you observe that you believe is
8020 incorrect. For example, ``It gets a fatal signal.''
8022 Of course, if the bug is that @command{ld} gets a fatal signal, then we
8023 will certainly notice it. But if the bug is incorrect output, we might
8024 not notice unless it is glaringly wrong. You might as well not give us
8025 a chance to make a mistake.
8027 Even if the problem you experience is a fatal signal, you should still
8028 say so explicitly. Suppose something strange is going on, such as, your
8029 copy of @command{ld} is out of sync, or you have encountered a bug in the
8030 C library on your system. (This has happened!) Your copy might crash
8031 and ours would not. If you told us to expect a crash, then when ours
8032 fails to crash, we would know that the bug was not happening for us. If
8033 you had not told us to expect a crash, then we would not be able to draw
8034 any conclusion from our observations.
8037 If you wish to suggest changes to the @command{ld} source, send us context
8038 diffs, as generated by @code{diff} with the @samp{-u}, @samp{-c}, or
8039 @samp{-p} option. Always send diffs from the old file to the new file.
8040 If you even discuss something in the @command{ld} source, refer to it by
8041 context, not by line number.
8043 The line numbers in our development sources will not match those in your
8044 sources. Your line numbers would convey no useful information to us.
8047 Here are some things that are not necessary:
8051 A description of the envelope of the bug.
8053 Often people who encounter a bug spend a lot of time investigating
8054 which changes to the input file will make the bug go away and which
8055 changes will not affect it.
8057 This is often time consuming and not very useful, because the way we
8058 will find the bug is by running a single example under the debugger
8059 with breakpoints, not by pure deduction from a series of examples.
8060 We recommend that you save your time for something else.
8062 Of course, if you can find a simpler example to report @emph{instead}
8063 of the original one, that is a convenience for us. Errors in the
8064 output will be easier to spot, running under the debugger will take
8065 less time, and so on.
8067 However, simplification is not vital; if you do not want to do this,
8068 report the bug anyway and send us the entire test case you used.
8071 A patch for the bug.
8073 A patch for the bug does help us if it is a good one. But do not omit
8074 the necessary information, such as the test case, on the assumption that
8075 a patch is all we need. We might see problems with your patch and decide
8076 to fix the problem another way, or we might not understand it at all.
8078 Sometimes with a program as complicated as @command{ld} it is very hard to
8079 construct an example that will make the program follow a certain path
8080 through the code. If you do not send us the example, we will not be
8081 able to construct one, so we will not be able to verify that the bug is
8084 And if we cannot understand what bug you are trying to fix, or why your
8085 patch should be an improvement, we will not install it. A test case will
8086 help us to understand.
8089 A guess about what the bug is or what it depends on.
8091 Such guesses are usually wrong. Even we cannot guess right about such
8092 things without first using the debugger to find the facts.
8096 @appendix MRI Compatible Script Files
8097 @cindex MRI compatibility
8098 To aid users making the transition to @sc{gnu} @command{ld} from the MRI
8099 linker, @command{ld} can use MRI compatible linker scripts as an
8100 alternative to the more general-purpose linker scripting language
8101 described in @ref{Scripts}. MRI compatible linker scripts have a much
8102 simpler command set than the scripting language otherwise used with
8103 @command{ld}. @sc{gnu} @command{ld} supports the most commonly used MRI
8104 linker commands; these commands are described here.
8106 In general, MRI scripts aren't of much use with the @code{a.out} object
8107 file format, since it only has three sections and MRI scripts lack some
8108 features to make use of them.
8110 You can specify a file containing an MRI-compatible script using the
8111 @samp{-c} command-line option.
8113 Each command in an MRI-compatible script occupies its own line; each
8114 command line starts with the keyword that identifies the command (though
8115 blank lines are also allowed for punctuation). If a line of an
8116 MRI-compatible script begins with an unrecognized keyword, @command{ld}
8117 issues a warning message, but continues processing the script.
8119 Lines beginning with @samp{*} are comments.
8121 You can write these commands using all upper-case letters, or all
8122 lower case; for example, @samp{chip} is the same as @samp{CHIP}.
8123 The following list shows only the upper-case form of each command.
8126 @cindex @code{ABSOLUTE} (MRI)
8127 @item ABSOLUTE @var{secname}
8128 @itemx ABSOLUTE @var{secname}, @var{secname}, @dots{} @var{secname}
8129 Normally, @command{ld} includes in the output file all sections from all
8130 the input files. However, in an MRI-compatible script, you can use the
8131 @code{ABSOLUTE} command to restrict the sections that will be present in
8132 your output program. If the @code{ABSOLUTE} command is used at all in a
8133 script, then only the sections named explicitly in @code{ABSOLUTE}
8134 commands will appear in the linker output. You can still use other
8135 input sections (whatever you select on the command line, or using
8136 @code{LOAD}) to resolve addresses in the output file.
8138 @cindex @code{ALIAS} (MRI)
8139 @item ALIAS @var{out-secname}, @var{in-secname}
8140 Use this command to place the data from input section @var{in-secname}
8141 in a section called @var{out-secname} in the linker output file.
8143 @var{in-secname} may be an integer.
8145 @cindex @code{ALIGN} (MRI)
8146 @item ALIGN @var{secname} = @var{expression}
8147 Align the section called @var{secname} to @var{expression}. The
8148 @var{expression} should be a power of two.
8150 @cindex @code{BASE} (MRI)
8151 @item BASE @var{expression}
8152 Use the value of @var{expression} as the lowest address (other than
8153 absolute addresses) in the output file.
8155 @cindex @code{CHIP} (MRI)
8156 @item CHIP @var{expression}
8157 @itemx CHIP @var{expression}, @var{expression}
8158 This command does nothing; it is accepted only for compatibility.
8160 @cindex @code{END} (MRI)
8162 This command does nothing whatever; it's only accepted for compatibility.
8164 @cindex @code{FORMAT} (MRI)
8165 @item FORMAT @var{output-format}
8166 Similar to the @code{OUTPUT_FORMAT} command in the more general linker
8167 language, but restricted to one of these output formats:
8171 S-records, if @var{output-format} is @samp{S}
8174 IEEE, if @var{output-format} is @samp{IEEE}
8177 COFF (the @samp{coff-m68k} variant in BFD), if @var{output-format} is
8181 @cindex @code{LIST} (MRI)
8182 @item LIST @var{anything}@dots{}
8183 Print (to the standard output file) a link map, as produced by the
8184 @command{ld} command-line option @samp{-M}.
8186 The keyword @code{LIST} may be followed by anything on the
8187 same line, with no change in its effect.
8189 @cindex @code{LOAD} (MRI)
8190 @item LOAD @var{filename}
8191 @itemx LOAD @var{filename}, @var{filename}, @dots{} @var{filename}
8192 Include one or more object file @var{filename} in the link; this has the
8193 same effect as specifying @var{filename} directly on the @command{ld}
8196 @cindex @code{NAME} (MRI)
8197 @item NAME @var{output-name}
8198 @var{output-name} is the name for the program produced by @command{ld}; the
8199 MRI-compatible command @code{NAME} is equivalent to the command-line
8200 option @samp{-o} or the general script language command @code{OUTPUT}.
8202 @cindex @code{ORDER} (MRI)
8203 @item ORDER @var{secname}, @var{secname}, @dots{} @var{secname}
8204 @itemx ORDER @var{secname} @var{secname} @var{secname}
8205 Normally, @command{ld} orders the sections in its output file in the
8206 order in which they first appear in the input files. In an MRI-compatible
8207 script, you can override this ordering with the @code{ORDER} command. The
8208 sections you list with @code{ORDER} will appear first in your output
8209 file, in the order specified.
8211 @cindex @code{PUBLIC} (MRI)
8212 @item PUBLIC @var{name}=@var{expression}
8213 @itemx PUBLIC @var{name},@var{expression}
8214 @itemx PUBLIC @var{name} @var{expression}
8215 Supply a value (@var{expression}) for external symbol
8216 @var{name} used in the linker input files.
8218 @cindex @code{SECT} (MRI)
8219 @item SECT @var{secname}, @var{expression}
8220 @itemx SECT @var{secname}=@var{expression}
8221 @itemx SECT @var{secname} @var{expression}
8222 You can use any of these three forms of the @code{SECT} command to
8223 specify the start address (@var{expression}) for section @var{secname}.
8224 If you have more than one @code{SECT} statement for the same
8225 @var{secname}, only the @emph{first} sets the start address.
8228 @node GNU Free Documentation License
8229 @appendix GNU Free Documentation License
8233 @unnumbered LD Index
8238 % I think something like @@colophon should be in texinfo. In the
8240 \long\def\colophon{\hbox to0pt{}\vfill
8241 \centerline{The body of this manual is set in}
8242 \centerline{\fontname\tenrm,}
8243 \centerline{with headings in {\bf\fontname\tenbf}}
8244 \centerline{and examples in {\tt\fontname\tentt}.}
8245 \centerline{{\it\fontname\tenit\/} and}
8246 \centerline{{\sl\fontname\tensl\/}}
8247 \centerline{are used for emphasis.}\vfill}
8249 % Blame: doc@@cygnus.com, 28mar91.