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.
976 If this option is being used to force additional modules to be pulled
977 into the link, and if it is an error for the symbol to remain
978 undefined, then the option @option{--require-defined} should be used
981 @kindex --require-defined=@var{symbol}
982 @cindex symbols, require defined
983 @cindex defined symbol
984 @item --require-defined=@var{symbol}
985 Require that @var{symbol} is defined in the output file. This option
986 is the same as option @option{--undefined} except that if @var{symbol}
987 is not defined in the output file then the linker will issue an error
988 and exit. The same effect can be achieved in a linker script by using
989 @code{EXTERN}, @code{ASSERT} and @code{DEFINED} together. This option
990 can be used multiple times to require additional symbols.
995 For anything other than C++ programs, this option is equivalent to
996 @samp{-r}: it generates relocatable output---i.e., an output file that can in
997 turn serve as input to @command{ld}. When linking C++ programs, @samp{-Ur}
998 @emph{does} resolve references to constructors, unlike @samp{-r}.
999 It does not work to use @samp{-Ur} on files that were themselves linked
1000 with @samp{-Ur}; once the constructor table has been built, it cannot
1001 be added to. Use @samp{-Ur} only for the last partial link, and
1002 @samp{-r} for the others.
1004 @kindex --orphan-handling=@var{MODE}
1005 @cindex orphan sections
1006 @cindex sections, orphan
1007 @item --orphan-handling=@var{MODE}
1008 Control how orphan sections are handled. An orphan section is one not
1009 specifically mentioned in a linker script. @xref{Orphan Sections}.
1011 @var{MODE} can have any of the following values:
1015 Orphan sections are placed into a suitable output section following
1016 the strategy described in @ref{Orphan Sections}. The option
1017 @samp{--unique} also effects how sections are placed.
1020 All orphan sections are discarded, by placing them in the
1021 @samp{/DISCARD/} section (@pxref{Output Section Discarding}).
1024 The linker will place the orphan section as for @code{place} and also
1028 The linker will exit with an error if any orphan section is found.
1031 The default if @samp{--orphan-handling} is not given is @code{place}.
1033 @kindex --unique[=@var{SECTION}]
1034 @item --unique[=@var{SECTION}]
1035 Creates a separate output section for every input section matching
1036 @var{SECTION}, or if the optional wildcard @var{SECTION} argument is
1037 missing, for every orphan input section. An orphan section is one not
1038 specifically mentioned in a linker script. You may use this option
1039 multiple times on the command line; It prevents the normal merging of
1040 input sections with the same name, overriding output section assignments
1050 Display the version number for @command{ld}. The @option{-V} option also
1051 lists the supported emulations.
1054 @kindex --discard-all
1055 @cindex deleting local symbols
1057 @itemx --discard-all
1058 Delete all local symbols.
1061 @kindex --discard-locals
1062 @cindex local symbols, deleting
1064 @itemx --discard-locals
1065 Delete all temporary local symbols. (These symbols start with
1066 system-specific local label prefixes, typically @samp{.L} for ELF systems
1067 or @samp{L} for traditional a.out systems.)
1069 @kindex -y @var{symbol}
1070 @kindex --trace-symbol=@var{symbol}
1071 @cindex symbol tracing
1072 @item -y @var{symbol}
1073 @itemx --trace-symbol=@var{symbol}
1074 Print the name of each linked file in which @var{symbol} appears. This
1075 option may be given any number of times. On many systems it is necessary
1076 to prepend an underscore.
1078 This option is useful when you have an undefined symbol in your link but
1079 don't know where the reference is coming from.
1081 @kindex -Y @var{path}
1083 Add @var{path} to the default library search path. This option exists
1084 for Solaris compatibility.
1086 @kindex -z @var{keyword}
1087 @item -z @var{keyword}
1088 The recognized keywords are:
1092 Combines multiple reloc sections and sorts them to make dynamic symbol
1093 lookup caching possible.
1096 Disallows undefined symbols in object files. Undefined symbols in
1097 shared libraries are still allowed.
1100 Marks the object as requiring executable stack.
1103 This option is only meaningful when building a shared object. It makes
1104 the symbols defined by this shared object available for symbol resolution
1105 of subsequently loaded libraries.
1108 This option is only meaningful when building a shared object.
1109 It marks the object so that its runtime initialization will occur
1110 before the runtime initialization of any other objects brought into
1111 the process at the same time. Similarly the runtime finalization of
1112 the object will occur after the runtime finalization of any other
1116 Marks the object that its symbol table interposes before all symbols
1117 but the primary executable.
1120 When generating an executable or shared library, mark it to tell the
1121 dynamic linker to defer function call resolution to the point when
1122 the function is called (lazy binding), rather than at load time.
1123 Lazy binding is the default.
1126 Marks the object that its filters be processed immediately at
1130 Allows multiple definitions.
1133 Disables multiple reloc sections combining.
1136 Disable linker generated .dynbss variables used in place of variables
1137 defined in shared libraries. May result in dynamic text relocations.
1140 Marks the object that the search for dependencies of this object will
1141 ignore any default library search paths.
1144 Marks the object shouldn't be unloaded at runtime.
1147 Marks the object not available to @code{dlopen}.
1150 Marks the object can not be dumped by @code{dldump}.
1153 Marks the object as not requiring executable stack.
1156 Treat DT_TEXTREL in shared object as error.
1159 Don't treat DT_TEXTREL in shared object as error.
1162 Don't treat DT_TEXTREL in shared object as error.
1165 Don't create an ELF @code{PT_GNU_RELRO} segment header in the object.
1168 When generating an executable or shared library, mark it to tell the
1169 dynamic linker to resolve all symbols when the program is started, or
1170 when the shared library is linked to using dlopen, instead of
1171 deferring function call resolution to the point when the function is
1175 Marks the object may contain $ORIGIN.
1178 Create an ELF @code{PT_GNU_RELRO} segment header in the object.
1180 @item max-page-size=@var{value}
1181 Set the emulation maximum page size to @var{value}.
1183 @item common-page-size=@var{value}
1184 Set the emulation common page size to @var{value}.
1186 @item stack-size=@var{value}
1187 Specify a stack size for in an ELF @code{PT_GNU_STACK} segment.
1188 Specifying zero will override any default non-zero sized
1189 @code{PT_GNU_STACK} segment creation.
1192 Always generate BND prefix in PLT entries. Supported for Linux/x86_64.
1194 @item noextern-protected-data
1195 Don't treat protected data symbol as external when building shared
1196 library. This option overrides linker backend default. It can be used
1197 to workaround incorrect relocations against protected data symbols
1198 generated by compiler. Updates on protected data symbols by another
1199 module aren't visibile to the resulting shared library. Supported for
1204 Other keywords are ignored for Solaris compatibility.
1207 @cindex groups of archives
1208 @item -( @var{archives} -)
1209 @itemx --start-group @var{archives} --end-group
1210 The @var{archives} should be a list of archive files. They may be
1211 either explicit file names, or @samp{-l} options.
1213 The specified archives are searched repeatedly until no new undefined
1214 references are created. Normally, an archive is searched only once in
1215 the order that it is specified on the command line. If a symbol in that
1216 archive is needed to resolve an undefined symbol referred to by an
1217 object in an archive that appears later on the command line, the linker
1218 would not be able to resolve that reference. By grouping the archives,
1219 they all be searched repeatedly until all possible references are
1222 Using this option has a significant performance cost. It is best to use
1223 it only when there are unavoidable circular references between two or
1226 @kindex --accept-unknown-input-arch
1227 @kindex --no-accept-unknown-input-arch
1228 @item --accept-unknown-input-arch
1229 @itemx --no-accept-unknown-input-arch
1230 Tells the linker to accept input files whose architecture cannot be
1231 recognised. The assumption is that the user knows what they are doing
1232 and deliberately wants to link in these unknown input files. This was
1233 the default behaviour of the linker, before release 2.14. The default
1234 behaviour from release 2.14 onwards is to reject such input files, and
1235 so the @samp{--accept-unknown-input-arch} option has been added to
1236 restore the old behaviour.
1239 @kindex --no-as-needed
1241 @itemx --no-as-needed
1242 This option affects ELF DT_NEEDED tags for dynamic libraries mentioned
1243 on the command line after the @option{--as-needed} option. Normally
1244 the linker will add a DT_NEEDED tag for each dynamic library mentioned
1245 on the command line, regardless of whether the library is actually
1246 needed or not. @option{--as-needed} causes a DT_NEEDED tag to only be
1247 emitted for a library that @emph{at that point in the link} satisfies a
1248 non-weak undefined symbol reference from a regular object file or, if
1249 the library is not found in the DT_NEEDED lists of other needed libraries, a
1250 non-weak undefined symbol reference from another needed dynamic library.
1251 Object files or libraries appearing on the command line @emph{after}
1252 the library in question do not affect whether the library is seen as
1253 needed. This is similar to the rules for extraction of object files
1254 from archives. @option{--no-as-needed} restores the default behaviour.
1256 @kindex --add-needed
1257 @kindex --no-add-needed
1259 @itemx --no-add-needed
1260 These two options have been deprecated because of the similarity of
1261 their names to the @option{--as-needed} and @option{--no-as-needed}
1262 options. They have been replaced by @option{--copy-dt-needed-entries}
1263 and @option{--no-copy-dt-needed-entries}.
1265 @kindex -assert @var{keyword}
1266 @item -assert @var{keyword}
1267 This option is ignored for SunOS compatibility.
1271 @kindex -call_shared
1275 Link against dynamic libraries. This is only meaningful on platforms
1276 for which shared libraries are supported. This option is normally the
1277 default on such platforms. The different variants of this option are
1278 for compatibility with various systems. You may use this option
1279 multiple times on the command line: it affects library searching for
1280 @option{-l} options which follow it.
1284 Set the @code{DF_1_GROUP} flag in the @code{DT_FLAGS_1} entry in the dynamic
1285 section. This causes the runtime linker to handle lookups in this
1286 object and its dependencies to be performed only inside the group.
1287 @option{--unresolved-symbols=report-all} is implied. This option is
1288 only meaningful on ELF platforms which support shared libraries.
1298 Do not link against shared libraries. This is only meaningful on
1299 platforms for which shared libraries are supported. The different
1300 variants of this option are for compatibility with various systems. You
1301 may use this option multiple times on the command line: it affects
1302 library searching for @option{-l} options which follow it. This
1303 option also implies @option{--unresolved-symbols=report-all}. This
1304 option can be used with @option{-shared}. Doing so means that a
1305 shared library is being created but that all of the library's external
1306 references must be resolved by pulling in entries from static
1311 When creating a shared library, bind references to global symbols to the
1312 definition within the shared library, if any. Normally, it is possible
1313 for a program linked against a shared library to override the definition
1314 within the shared library. This option is only meaningful on ELF
1315 platforms which support shared libraries.
1317 @kindex -Bsymbolic-functions
1318 @item -Bsymbolic-functions
1319 When creating a shared library, bind references to global function
1320 symbols to the definition within the shared library, if any.
1321 This option is only meaningful on ELF platforms which support shared
1324 @kindex --dynamic-list=@var{dynamic-list-file}
1325 @item --dynamic-list=@var{dynamic-list-file}
1326 Specify the name of a dynamic list file to the linker. This is
1327 typically used when creating shared libraries to specify a list of
1328 global symbols whose references shouldn't be bound to the definition
1329 within the shared library, or creating dynamically linked executables
1330 to specify a list of symbols which should be added to the symbol table
1331 in the executable. This option is only meaningful on ELF platforms
1332 which support shared libraries.
1334 The format of the dynamic list is the same as the version node without
1335 scope and node name. See @ref{VERSION} for more information.
1337 @kindex --dynamic-list-data
1338 @item --dynamic-list-data
1339 Include all global data symbols to the dynamic list.
1341 @kindex --dynamic-list-cpp-new
1342 @item --dynamic-list-cpp-new
1343 Provide the builtin dynamic list for C++ operator new and delete. It
1344 is mainly useful for building shared libstdc++.
1346 @kindex --dynamic-list-cpp-typeinfo
1347 @item --dynamic-list-cpp-typeinfo
1348 Provide the builtin dynamic list for C++ runtime type identification.
1350 @kindex --check-sections
1351 @kindex --no-check-sections
1352 @item --check-sections
1353 @itemx --no-check-sections
1354 Asks the linker @emph{not} to check section addresses after they have
1355 been assigned to see if there are any overlaps. Normally the linker will
1356 perform this check, and if it finds any overlaps it will produce
1357 suitable error messages. The linker does know about, and does make
1358 allowances for sections in overlays. The default behaviour can be
1359 restored by using the command line switch @option{--check-sections}.
1360 Section overlap is not usually checked for relocatable links. You can
1361 force checking in that case by using the @option{--check-sections}
1364 @kindex --copy-dt-needed-entries
1365 @kindex --no-copy-dt-needed-entries
1366 @item --copy-dt-needed-entries
1367 @itemx --no-copy-dt-needed-entries
1368 This option affects the treatment of dynamic libraries referred to
1369 by DT_NEEDED tags @emph{inside} ELF dynamic libraries mentioned on the
1370 command line. Normally the linker won't add a DT_NEEDED tag to the
1371 output binary for each library mentioned in a DT_NEEDED tag in an
1372 input dynamic library. With @option{--copy-dt-needed-entries}
1373 specified on the command line however any dynamic libraries that
1374 follow it will have their DT_NEEDED entries added. The default
1375 behaviour can be restored with @option{--no-copy-dt-needed-entries}.
1377 This option also has an effect on the resolution of symbols in dynamic
1378 libraries. With @option{--copy-dt-needed-entries} dynamic libraries
1379 mentioned on the command line will be recursively searched, following
1380 their DT_NEEDED tags to other libraries, in order to resolve symbols
1381 required by the output binary. With the default setting however
1382 the searching of dynamic libraries that follow it will stop with the
1383 dynamic library itself. No DT_NEEDED links will be traversed to resolve
1386 @cindex cross reference table
1389 Output a cross reference table. If a linker map file is being
1390 generated, the cross reference table is printed to the map file.
1391 Otherwise, it is printed on the standard output.
1393 The format of the table is intentionally simple, so that it may be
1394 easily processed by a script if necessary. The symbols are printed out,
1395 sorted by name. For each symbol, a list of file names is given. If the
1396 symbol is defined, the first file listed is the location of the
1397 definition. If the symbol is defined as a common value then any files
1398 where this happens appear next. Finally any files that reference the
1401 @cindex common allocation
1402 @kindex --no-define-common
1403 @item --no-define-common
1404 This option inhibits the assignment of addresses to common symbols.
1405 The script command @code{INHIBIT_COMMON_ALLOCATION} has the same effect.
1406 @xref{Miscellaneous Commands}.
1408 The @samp{--no-define-common} option allows decoupling
1409 the decision to assign addresses to Common symbols from the choice
1410 of the output file type; otherwise a non-Relocatable output type
1411 forces assigning addresses to Common symbols.
1412 Using @samp{--no-define-common} allows Common symbols that are referenced
1413 from a shared library to be assigned addresses only in the main program.
1414 This eliminates the unused duplicate space in the shared library,
1415 and also prevents any possible confusion over resolving to the wrong
1416 duplicate when there are many dynamic modules with specialized search
1417 paths for runtime symbol resolution.
1419 @cindex symbols, from command line
1420 @kindex --defsym=@var{symbol}=@var{exp}
1421 @item --defsym=@var{symbol}=@var{expression}
1422 Create a global symbol in the output file, containing the absolute
1423 address given by @var{expression}. You may use this option as many
1424 times as necessary to define multiple symbols in the command line. A
1425 limited form of arithmetic is supported for the @var{expression} in this
1426 context: you may give a hexadecimal constant or the name of an existing
1427 symbol, or use @code{+} and @code{-} to add or subtract hexadecimal
1428 constants or symbols. If you need more elaborate expressions, consider
1429 using the linker command language from a script (@pxref{Assignments}).
1430 @emph{Note:} there should be no white space between @var{symbol}, the
1431 equals sign (``@key{=}''), and @var{expression}.
1433 @cindex demangling, from command line
1434 @kindex --demangle[=@var{style}]
1435 @kindex --no-demangle
1436 @item --demangle[=@var{style}]
1437 @itemx --no-demangle
1438 These options control whether to demangle symbol names in error messages
1439 and other output. When the linker is told to demangle, it tries to
1440 present symbol names in a readable fashion: it strips leading
1441 underscores if they are used by the object file format, and converts C++
1442 mangled symbol names into user readable names. Different compilers have
1443 different mangling styles. The optional demangling style argument can be used
1444 to choose an appropriate demangling style for your compiler. The linker will
1445 demangle by default unless the environment variable @samp{COLLECT_NO_DEMANGLE}
1446 is set. These options may be used to override the default.
1448 @cindex dynamic linker, from command line
1449 @kindex -I@var{file}
1450 @kindex --dynamic-linker=@var{file}
1452 @itemx --dynamic-linker=@var{file}
1453 Set the name of the dynamic linker. This is only meaningful when
1454 generating dynamically linked ELF executables. The default dynamic
1455 linker is normally correct; don't use this unless you know what you are
1458 @kindex --no-dynamic-linker
1459 @item --no-dynamic-linker
1460 When producing an executable file, omit the request for a dynamic
1461 linker to be used at load-time. This is only meaningful for ELF
1462 executables that contain dynamic relocations, and usually requires
1463 entry point code that is capable of processing these relocations.
1465 @kindex --fatal-warnings
1466 @kindex --no-fatal-warnings
1467 @item --fatal-warnings
1468 @itemx --no-fatal-warnings
1469 Treat all warnings as errors. The default behaviour can be restored
1470 with the option @option{--no-fatal-warnings}.
1472 @kindex --force-exe-suffix
1473 @item --force-exe-suffix
1474 Make sure that an output file has a .exe suffix.
1476 If a successfully built fully linked output file does not have a
1477 @code{.exe} or @code{.dll} suffix, this option forces the linker to copy
1478 the output file to one of the same name with a @code{.exe} suffix. This
1479 option is useful when using unmodified Unix makefiles on a Microsoft
1480 Windows host, since some versions of Windows won't run an image unless
1481 it ends in a @code{.exe} suffix.
1483 @kindex --gc-sections
1484 @kindex --no-gc-sections
1485 @cindex garbage collection
1487 @itemx --no-gc-sections
1488 Enable garbage collection of unused input sections. It is ignored on
1489 targets that do not support this option. The default behaviour (of not
1490 performing this garbage collection) can be restored by specifying
1491 @samp{--no-gc-sections} on the command line. Note that garbage
1492 collection for COFF and PE format targets is supported, but the
1493 implementation is currently considered to be experimental.
1495 @samp{--gc-sections} decides which input sections are used by
1496 examining symbols and relocations. The section containing the entry
1497 symbol and all sections containing symbols undefined on the
1498 command-line will be kept, as will sections containing symbols
1499 referenced by dynamic objects. Note that when building shared
1500 libraries, the linker must assume that any visible symbol is
1501 referenced. Once this initial set of sections has been determined,
1502 the linker recursively marks as used any section referenced by their
1503 relocations. See @samp{--entry} and @samp{--undefined}.
1505 This option can be set when doing a partial link (enabled with option
1506 @samp{-r}). In this case the root of symbols kept must be explicitly
1507 specified either by an @samp{--entry} or @samp{--undefined} option or by
1508 a @code{ENTRY} command in the linker script.
1510 @kindex --print-gc-sections
1511 @kindex --no-print-gc-sections
1512 @cindex garbage collection
1513 @item --print-gc-sections
1514 @itemx --no-print-gc-sections
1515 List all sections removed by garbage collection. The listing is
1516 printed on stderr. This option is only effective if garbage
1517 collection has been enabled via the @samp{--gc-sections}) option. The
1518 default behaviour (of not listing the sections that are removed) can
1519 be restored by specifying @samp{--no-print-gc-sections} on the command
1522 @kindex --print-output-format
1523 @cindex output format
1524 @item --print-output-format
1525 Print the name of the default output format (perhaps influenced by
1526 other command-line options). This is the string that would appear
1527 in an @code{OUTPUT_FORMAT} linker script command (@pxref{File Commands}).
1529 @kindex --print-memory-usage
1530 @cindex memory usage
1531 @item --print-memory-usage
1532 Print used size, total size and used size of memory regions created with
1533 the @ref{MEMORY} command. This is useful on embedded targets to have a
1534 quick view of amount of free memory. The format of the output has one
1535 headline and one line per region. It is both human readable and easily
1536 parsable by tools. Here is an example of an output:
1539 Memory region Used Size Region Size %age Used
1540 ROM: 256 KB 1 MB 25.00%
1541 RAM: 32 B 2 GB 0.00%
1548 Print a summary of the command-line options on the standard output and exit.
1550 @kindex --target-help
1552 Print a summary of all target specific options on the standard output and exit.
1554 @kindex -Map=@var{mapfile}
1555 @item -Map=@var{mapfile}
1556 Print a link map to the file @var{mapfile}. See the description of the
1557 @option{-M} option, above.
1559 @cindex memory usage
1560 @kindex --no-keep-memory
1561 @item --no-keep-memory
1562 @command{ld} normally optimizes for speed over memory usage by caching the
1563 symbol tables of input files in memory. This option tells @command{ld} to
1564 instead optimize for memory usage, by rereading the symbol tables as
1565 necessary. This may be required if @command{ld} runs out of memory space
1566 while linking a large executable.
1568 @kindex --no-undefined
1570 @item --no-undefined
1572 Report unresolved symbol references from regular object files. This
1573 is done even if the linker is creating a non-symbolic shared library.
1574 The switch @option{--[no-]allow-shlib-undefined} controls the
1575 behaviour for reporting unresolved references found in shared
1576 libraries being linked in.
1578 @kindex --allow-multiple-definition
1580 @item --allow-multiple-definition
1582 Normally when a symbol is defined multiple times, the linker will
1583 report a fatal error. These options allow multiple definitions and the
1584 first definition will be used.
1586 @kindex --allow-shlib-undefined
1587 @kindex --no-allow-shlib-undefined
1588 @item --allow-shlib-undefined
1589 @itemx --no-allow-shlib-undefined
1590 Allows or disallows undefined symbols in shared libraries.
1591 This switch is similar to @option{--no-undefined} except that it
1592 determines the behaviour when the undefined symbols are in a
1593 shared library rather than a regular object file. It does not affect
1594 how undefined symbols in regular object files are handled.
1596 The default behaviour is to report errors for any undefined symbols
1597 referenced in shared libraries if the linker is being used to create
1598 an executable, but to allow them if the linker is being used to create
1601 The reasons for allowing undefined symbol references in shared
1602 libraries specified at link time are that:
1606 A shared library specified at link time may not be the same as the one
1607 that is available at load time, so the symbol might actually be
1608 resolvable at load time.
1610 There are some operating systems, eg BeOS and HPPA, where undefined
1611 symbols in shared libraries are normal.
1613 The BeOS kernel for example patches shared libraries at load time to
1614 select whichever function is most appropriate for the current
1615 architecture. This is used, for example, to dynamically select an
1616 appropriate memset function.
1619 @kindex --no-undefined-version
1620 @item --no-undefined-version
1621 Normally when a symbol has an undefined version, the linker will ignore
1622 it. This option disallows symbols with undefined version and a fatal error
1623 will be issued instead.
1625 @kindex --default-symver
1626 @item --default-symver
1627 Create and use a default symbol version (the soname) for unversioned
1630 @kindex --default-imported-symver
1631 @item --default-imported-symver
1632 Create and use a default symbol version (the soname) for unversioned
1635 @kindex --no-warn-mismatch
1636 @item --no-warn-mismatch
1637 Normally @command{ld} will give an error if you try to link together input
1638 files that are mismatched for some reason, perhaps because they have
1639 been compiled for different processors or for different endiannesses.
1640 This option tells @command{ld} that it should silently permit such possible
1641 errors. This option should only be used with care, in cases when you
1642 have taken some special action that ensures that the linker errors are
1645 @kindex --no-warn-search-mismatch
1646 @item --no-warn-search-mismatch
1647 Normally @command{ld} will give a warning if it finds an incompatible
1648 library during a library search. This option silences the warning.
1650 @kindex --no-whole-archive
1651 @item --no-whole-archive
1652 Turn off the effect of the @option{--whole-archive} option for subsequent
1655 @cindex output file after errors
1656 @kindex --noinhibit-exec
1657 @item --noinhibit-exec
1658 Retain the executable output file whenever it is still usable.
1659 Normally, the linker will not produce an output file if it encounters
1660 errors during the link process; it exits without writing an output file
1661 when it issues any error whatsoever.
1665 Only search library directories explicitly specified on the
1666 command line. Library directories specified in linker scripts
1667 (including linker scripts specified on the command line) are ignored.
1669 @ifclear SingleFormat
1670 @kindex --oformat=@var{output-format}
1671 @item --oformat=@var{output-format}
1672 @command{ld} may be configured to support more than one kind of object
1673 file. If your @command{ld} is configured this way, you can use the
1674 @samp{--oformat} option to specify the binary format for the output
1675 object file. Even when @command{ld} is configured to support alternative
1676 object formats, you don't usually need to specify this, as @command{ld}
1677 should be configured to produce as a default output format the most
1678 usual format on each machine. @var{output-format} is a text string, the
1679 name of a particular format supported by the BFD libraries. (You can
1680 list the available binary formats with @samp{objdump -i}.) The script
1681 command @code{OUTPUT_FORMAT} can also specify the output format, but
1682 this option overrides it. @xref{BFD}.
1686 @kindex --pic-executable
1688 @itemx --pic-executable
1689 @cindex position independent executables
1690 Create a position independent executable. This is currently only supported on
1691 ELF platforms. Position independent executables are similar to shared
1692 libraries in that they are relocated by the dynamic linker to the virtual
1693 address the OS chooses for them (which can vary between invocations). Like
1694 normal dynamically linked executables they can be executed and symbols
1695 defined in the executable cannot be overridden by shared libraries.
1699 This option is ignored for Linux compatibility.
1703 This option is ignored for SVR4 compatibility.
1706 @cindex synthesizing linker
1707 @cindex relaxing addressing modes
1711 An option with machine dependent effects.
1713 This option is only supported on a few targets.
1716 @xref{H8/300,,@command{ld} and the H8/300}.
1719 @xref{i960,, @command{ld} and the Intel 960 family}.
1722 @xref{Xtensa,, @command{ld} and Xtensa Processors}.
1725 @xref{M68HC11/68HC12,,@command{ld} and the 68HC11 and 68HC12}.
1728 @xref{Nios II,,@command{ld} and the Altera Nios II}.
1731 @xref{PowerPC ELF32,,@command{ld} and PowerPC 32-bit ELF Support}.
1734 On some platforms the @samp{--relax} option performs target specific,
1735 global optimizations that become possible when the linker resolves
1736 addressing in the program, such as relaxing address modes,
1737 synthesizing new instructions, selecting shorter version of current
1738 instructions, and combining constant values.
1740 On some platforms these link time global optimizations may make symbolic
1741 debugging of the resulting executable impossible.
1743 This is known to be the case for the Matsushita MN10200 and MN10300
1744 family of processors.
1748 On platforms where this is not supported, @samp{--relax} is accepted,
1752 On platforms where @samp{--relax} is accepted the option
1753 @samp{--no-relax} can be used to disable the feature.
1755 @cindex retaining specified symbols
1756 @cindex stripping all but some symbols
1757 @cindex symbols, retaining selectively
1758 @kindex --retain-symbols-file=@var{filename}
1759 @item --retain-symbols-file=@var{filename}
1760 Retain @emph{only} the symbols listed in the file @var{filename},
1761 discarding all others. @var{filename} is simply a flat file, with one
1762 symbol name per line. This option is especially useful in environments
1766 where a large global symbol table is accumulated gradually, to conserve
1769 @samp{--retain-symbols-file} does @emph{not} discard undefined symbols,
1770 or symbols needed for relocations.
1772 You may only specify @samp{--retain-symbols-file} once in the command
1773 line. It overrides @samp{-s} and @samp{-S}.
1776 @item -rpath=@var{dir}
1777 @cindex runtime library search path
1778 @kindex -rpath=@var{dir}
1779 Add a directory to the runtime library search path. This is used when
1780 linking an ELF executable with shared objects. All @option{-rpath}
1781 arguments are concatenated and passed to the runtime linker, which uses
1782 them to locate shared objects at runtime. The @option{-rpath} option is
1783 also used when locating shared objects which are needed by shared
1784 objects explicitly included in the link; see the description of the
1785 @option{-rpath-link} option. If @option{-rpath} is not used when linking an
1786 ELF executable, the contents of the environment variable
1787 @code{LD_RUN_PATH} will be used if it is defined.
1789 The @option{-rpath} option may also be used on SunOS. By default, on
1790 SunOS, the linker will form a runtime search patch out of all the
1791 @option{-L} options it is given. If a @option{-rpath} option is used, the
1792 runtime search path will be formed exclusively using the @option{-rpath}
1793 options, ignoring the @option{-L} options. This can be useful when using
1794 gcc, which adds many @option{-L} options which may be on NFS mounted
1797 For compatibility with other ELF linkers, if the @option{-R} option is
1798 followed by a directory name, rather than a file name, it is treated as
1799 the @option{-rpath} option.
1803 @cindex link-time runtime library search path
1804 @kindex -rpath-link=@var{dir}
1805 @item -rpath-link=@var{dir}
1806 When using ELF or SunOS, one shared library may require another. This
1807 happens when an @code{ld -shared} link includes a shared library as one
1810 When the linker encounters such a dependency when doing a non-shared,
1811 non-relocatable link, it will automatically try to locate the required
1812 shared library and include it in the link, if it is not included
1813 explicitly. In such a case, the @option{-rpath-link} option
1814 specifies the first set of directories to search. The
1815 @option{-rpath-link} option may specify a sequence of directory names
1816 either by specifying a list of names separated by colons, or by
1817 appearing multiple times.
1819 This option should be used with caution as it overrides the search path
1820 that may have been hard compiled into a shared library. In such a case it
1821 is possible to use unintentionally a different search path than the
1822 runtime linker would do.
1824 The linker uses the following search paths to locate required shared
1828 Any directories specified by @option{-rpath-link} options.
1830 Any directories specified by @option{-rpath} options. The difference
1831 between @option{-rpath} and @option{-rpath-link} is that directories
1832 specified by @option{-rpath} options are included in the executable and
1833 used at runtime, whereas the @option{-rpath-link} option is only effective
1834 at link time. Searching @option{-rpath} in this way is only supported
1835 by native linkers and cross linkers which have been configured with
1836 the @option{--with-sysroot} option.
1838 On an ELF system, for native linkers, if the @option{-rpath} and
1839 @option{-rpath-link} options were not used, search the contents of the
1840 environment variable @code{LD_RUN_PATH}.
1842 On SunOS, if the @option{-rpath} option was not used, search any
1843 directories specified using @option{-L} options.
1845 For a native linker, search the contents of the environment
1846 variable @code{LD_LIBRARY_PATH}.
1848 For a native ELF linker, the directories in @code{DT_RUNPATH} or
1849 @code{DT_RPATH} of a shared library are searched for shared
1850 libraries needed by it. The @code{DT_RPATH} entries are ignored if
1851 @code{DT_RUNPATH} entries exist.
1853 The default directories, normally @file{/lib} and @file{/usr/lib}.
1855 For a native linker on an ELF system, if the file @file{/etc/ld.so.conf}
1856 exists, the list of directories found in that file.
1859 If the required shared library is not found, the linker will issue a
1860 warning and continue with the link.
1867 @cindex shared libraries
1868 Create a shared library. This is currently only supported on ELF, XCOFF
1869 and SunOS platforms. On SunOS, the linker will automatically create a
1870 shared library if the @option{-e} option is not used and there are
1871 undefined symbols in the link.
1873 @kindex --sort-common
1875 @itemx --sort-common=ascending
1876 @itemx --sort-common=descending
1877 This option tells @command{ld} to sort the common symbols by alignment in
1878 ascending or descending order when it places them in the appropriate output
1879 sections. The symbol alignments considered are sixteen-byte or larger,
1880 eight-byte, four-byte, two-byte, and one-byte. This is to prevent gaps
1881 between symbols due to alignment constraints. If no sorting order is
1882 specified, then descending order is assumed.
1884 @kindex --sort-section=name
1885 @item --sort-section=name
1886 This option will apply @code{SORT_BY_NAME} to all wildcard section
1887 patterns in the linker script.
1889 @kindex --sort-section=alignment
1890 @item --sort-section=alignment
1891 This option will apply @code{SORT_BY_ALIGNMENT} to all wildcard section
1892 patterns in the linker script.
1894 @kindex --split-by-file
1895 @item --split-by-file[=@var{size}]
1896 Similar to @option{--split-by-reloc} but creates a new output section for
1897 each input file when @var{size} is reached. @var{size} defaults to a
1898 size of 1 if not given.
1900 @kindex --split-by-reloc
1901 @item --split-by-reloc[=@var{count}]
1902 Tries to creates extra sections in the output file so that no single
1903 output section in the file contains more than @var{count} relocations.
1904 This is useful when generating huge relocatable files for downloading into
1905 certain real time kernels with the COFF object file format; since COFF
1906 cannot represent more than 65535 relocations in a single section. Note
1907 that this will fail to work with object file formats which do not
1908 support arbitrary sections. The linker will not split up individual
1909 input sections for redistribution, so if a single input section contains
1910 more than @var{count} relocations one output section will contain that
1911 many relocations. @var{count} defaults to a value of 32768.
1915 Compute and display statistics about the operation of the linker, such
1916 as execution time and memory usage.
1918 @kindex --sysroot=@var{directory}
1919 @item --sysroot=@var{directory}
1920 Use @var{directory} as the location of the sysroot, overriding the
1921 configure-time default. This option is only supported by linkers
1922 that were configured using @option{--with-sysroot}.
1924 @kindex --traditional-format
1925 @cindex traditional format
1926 @item --traditional-format
1927 For some targets, the output of @command{ld} is different in some ways from
1928 the output of some existing linker. This switch requests @command{ld} to
1929 use the traditional format instead.
1932 For example, on SunOS, @command{ld} combines duplicate entries in the
1933 symbol string table. This can reduce the size of an output file with
1934 full debugging information by over 30 percent. Unfortunately, the SunOS
1935 @code{dbx} program can not read the resulting program (@code{gdb} has no
1936 trouble). The @samp{--traditional-format} switch tells @command{ld} to not
1937 combine duplicate entries.
1939 @kindex --section-start=@var{sectionname}=@var{org}
1940 @item --section-start=@var{sectionname}=@var{org}
1941 Locate a section in the output file at the absolute
1942 address given by @var{org}. You may use this option as many
1943 times as necessary to locate multiple sections in the command
1945 @var{org} must be a single hexadecimal integer;
1946 for compatibility with other linkers, you may omit the leading
1947 @samp{0x} usually associated with hexadecimal values. @emph{Note:} there
1948 should be no white space between @var{sectionname}, the equals
1949 sign (``@key{=}''), and @var{org}.
1951 @kindex -Tbss=@var{org}
1952 @kindex -Tdata=@var{org}
1953 @kindex -Ttext=@var{org}
1954 @cindex segment origins, cmd line
1955 @item -Tbss=@var{org}
1956 @itemx -Tdata=@var{org}
1957 @itemx -Ttext=@var{org}
1958 Same as @option{--section-start}, with @code{.bss}, @code{.data} or
1959 @code{.text} as the @var{sectionname}.
1961 @kindex -Ttext-segment=@var{org}
1962 @item -Ttext-segment=@var{org}
1963 @cindex text segment origin, cmd line
1964 When creating an ELF executable, it will set the address of the first
1965 byte of the text segment.
1967 @kindex -Trodata-segment=@var{org}
1968 @item -Trodata-segment=@var{org}
1969 @cindex rodata segment origin, cmd line
1970 When creating an ELF executable or shared object for a target where
1971 the read-only data is in its own segment separate from the executable
1972 text, it will set the address of the first byte of the read-only data segment.
1974 @kindex -Tldata-segment=@var{org}
1975 @item -Tldata-segment=@var{org}
1976 @cindex ldata segment origin, cmd line
1977 When creating an ELF executable or shared object for x86-64 medium memory
1978 model, it will set the address of the first byte of the ldata segment.
1980 @kindex --unresolved-symbols
1981 @item --unresolved-symbols=@var{method}
1982 Determine how to handle unresolved symbols. There are four possible
1983 values for @samp{method}:
1987 Do not report any unresolved symbols.
1990 Report all unresolved symbols. This is the default.
1992 @item ignore-in-object-files
1993 Report unresolved symbols that are contained in shared libraries, but
1994 ignore them if they come from regular object files.
1996 @item ignore-in-shared-libs
1997 Report unresolved symbols that come from regular object files, but
1998 ignore them if they come from shared libraries. This can be useful
1999 when creating a dynamic binary and it is known that all the shared
2000 libraries that it should be referencing are included on the linker's
2004 The behaviour for shared libraries on their own can also be controlled
2005 by the @option{--[no-]allow-shlib-undefined} option.
2007 Normally the linker will generate an error message for each reported
2008 unresolved symbol but the option @option{--warn-unresolved-symbols}
2009 can change this to a warning.
2011 @kindex --verbose[=@var{NUMBER}]
2012 @cindex verbose[=@var{NUMBER}]
2014 @itemx --verbose[=@var{NUMBER}]
2015 Display the version number for @command{ld} and list the linker emulations
2016 supported. Display which input files can and cannot be opened. Display
2017 the linker script being used by the linker. If the optional @var{NUMBER}
2018 argument > 1, plugin symbol status will also be displayed.
2020 @kindex --version-script=@var{version-scriptfile}
2021 @cindex version script, symbol versions
2022 @item --version-script=@var{version-scriptfile}
2023 Specify the name of a version script to the linker. This is typically
2024 used when creating shared libraries to specify additional information
2025 about the version hierarchy for the library being created. This option
2026 is only fully supported on ELF platforms which support shared libraries;
2027 see @ref{VERSION}. It is partially supported on PE platforms, which can
2028 use version scripts to filter symbol visibility in auto-export mode: any
2029 symbols marked @samp{local} in the version script will not be exported.
2032 @kindex --warn-common
2033 @cindex warnings, on combining symbols
2034 @cindex combining symbols, warnings on
2036 Warn when a common symbol is combined with another common symbol or with
2037 a symbol definition. Unix linkers allow this somewhat sloppy practice,
2038 but linkers on some other operating systems do not. This option allows
2039 you to find potential problems from combining global symbols.
2040 Unfortunately, some C libraries use this practice, so you may get some
2041 warnings about symbols in the libraries as well as in your programs.
2043 There are three kinds of global symbols, illustrated here by C examples:
2047 A definition, which goes in the initialized data section of the output
2051 An undefined reference, which does not allocate space.
2052 There must be either a definition or a common symbol for the
2056 A common symbol. If there are only (one or more) common symbols for a
2057 variable, it goes in the uninitialized data area of the output file.
2058 The linker merges multiple common symbols for the same variable into a
2059 single symbol. If they are of different sizes, it picks the largest
2060 size. The linker turns a common symbol into a declaration, if there is
2061 a definition of the same variable.
2064 The @samp{--warn-common} option can produce five kinds of warnings.
2065 Each warning consists of a pair of lines: the first describes the symbol
2066 just encountered, and the second describes the previous symbol
2067 encountered with the same name. One or both of the two symbols will be
2072 Turning a common symbol into a reference, because there is already a
2073 definition for the symbol.
2075 @var{file}(@var{section}): warning: common of `@var{symbol}'
2076 overridden by definition
2077 @var{file}(@var{section}): warning: defined here
2081 Turning a common symbol into a reference, because a later definition for
2082 the symbol is encountered. This is the same as the previous case,
2083 except that the symbols are encountered in a different order.
2085 @var{file}(@var{section}): warning: definition of `@var{symbol}'
2087 @var{file}(@var{section}): warning: common is here
2091 Merging a common symbol with a previous same-sized common symbol.
2093 @var{file}(@var{section}): warning: multiple common
2095 @var{file}(@var{section}): warning: previous common is here
2099 Merging a common symbol with a previous larger common symbol.
2101 @var{file}(@var{section}): warning: common of `@var{symbol}'
2102 overridden by larger common
2103 @var{file}(@var{section}): warning: larger common is here
2107 Merging a common symbol with a previous smaller common symbol. This is
2108 the same as the previous case, except that the symbols are
2109 encountered in a different order.
2111 @var{file}(@var{section}): warning: common of `@var{symbol}'
2112 overriding smaller common
2113 @var{file}(@var{section}): warning: smaller common is here
2117 @kindex --warn-constructors
2118 @item --warn-constructors
2119 Warn if any global constructors are used. This is only useful for a few
2120 object file formats. For formats like COFF or ELF, the linker can not
2121 detect the use of global constructors.
2123 @kindex --warn-multiple-gp
2124 @item --warn-multiple-gp
2125 Warn if multiple global pointer values are required in the output file.
2126 This is only meaningful for certain processors, such as the Alpha.
2127 Specifically, some processors put large-valued constants in a special
2128 section. A special register (the global pointer) points into the middle
2129 of this section, so that constants can be loaded efficiently via a
2130 base-register relative addressing mode. Since the offset in
2131 base-register relative mode is fixed and relatively small (e.g., 16
2132 bits), this limits the maximum size of the constant pool. Thus, in
2133 large programs, it is often necessary to use multiple global pointer
2134 values in order to be able to address all possible constants. This
2135 option causes a warning to be issued whenever this case occurs.
2138 @cindex warnings, on undefined symbols
2139 @cindex undefined symbols, warnings on
2141 Only warn once for each undefined symbol, rather than once per module
2144 @kindex --warn-section-align
2145 @cindex warnings, on section alignment
2146 @cindex section alignment, warnings on
2147 @item --warn-section-align
2148 Warn if the address of an output section is changed because of
2149 alignment. Typically, the alignment will be set by an input section.
2150 The address will only be changed if it not explicitly specified; that
2151 is, if the @code{SECTIONS} command does not specify a start address for
2152 the section (@pxref{SECTIONS}).
2154 @kindex --warn-shared-textrel
2155 @item --warn-shared-textrel
2156 Warn if the linker adds a DT_TEXTREL to a shared object.
2158 @kindex --warn-alternate-em
2159 @item --warn-alternate-em
2160 Warn if an object has alternate ELF machine code.
2162 @kindex --warn-unresolved-symbols
2163 @item --warn-unresolved-symbols
2164 If the linker is going to report an unresolved symbol (see the option
2165 @option{--unresolved-symbols}) it will normally generate an error.
2166 This option makes it generate a warning instead.
2168 @kindex --error-unresolved-symbols
2169 @item --error-unresolved-symbols
2170 This restores the linker's default behaviour of generating errors when
2171 it is reporting unresolved symbols.
2173 @kindex --whole-archive
2174 @cindex including an entire archive
2175 @item --whole-archive
2176 For each archive mentioned on the command line after the
2177 @option{--whole-archive} option, include every object file in the archive
2178 in the link, rather than searching the archive for the required object
2179 files. This is normally used to turn an archive file into a shared
2180 library, forcing every object to be included in the resulting shared
2181 library. This option may be used more than once.
2183 Two notes when using this option from gcc: First, gcc doesn't know
2184 about this option, so you have to use @option{-Wl,-whole-archive}.
2185 Second, don't forget to use @option{-Wl,-no-whole-archive} after your
2186 list of archives, because gcc will add its own list of archives to
2187 your link and you may not want this flag to affect those as well.
2189 @kindex --wrap=@var{symbol}
2190 @item --wrap=@var{symbol}
2191 Use a wrapper function for @var{symbol}. Any undefined reference to
2192 @var{symbol} will be resolved to @code{__wrap_@var{symbol}}. Any
2193 undefined reference to @code{__real_@var{symbol}} will be resolved to
2196 This can be used to provide a wrapper for a system function. The
2197 wrapper function should be called @code{__wrap_@var{symbol}}. If it
2198 wishes to call the system function, it should call
2199 @code{__real_@var{symbol}}.
2201 Here is a trivial example:
2205 __wrap_malloc (size_t c)
2207 printf ("malloc called with %zu\n", c);
2208 return __real_malloc (c);
2212 If you link other code with this file using @option{--wrap malloc}, then
2213 all calls to @code{malloc} will call the function @code{__wrap_malloc}
2214 instead. The call to @code{__real_malloc} in @code{__wrap_malloc} will
2215 call the real @code{malloc} function.
2217 You may wish to provide a @code{__real_malloc} function as well, so that
2218 links without the @option{--wrap} option will succeed. If you do this,
2219 you should not put the definition of @code{__real_malloc} in the same
2220 file as @code{__wrap_malloc}; if you do, the assembler may resolve the
2221 call before the linker has a chance to wrap it to @code{malloc}.
2223 @kindex --eh-frame-hdr
2224 @item --eh-frame-hdr
2225 Request creation of @code{.eh_frame_hdr} section and ELF
2226 @code{PT_GNU_EH_FRAME} segment header.
2228 @kindex --ld-generated-unwind-info
2229 @item --no-ld-generated-unwind-info
2230 Request creation of @code{.eh_frame} unwind info for linker
2231 generated code sections like PLT. This option is on by default
2232 if linker generated unwind info is supported.
2234 @kindex --enable-new-dtags
2235 @kindex --disable-new-dtags
2236 @item --enable-new-dtags
2237 @itemx --disable-new-dtags
2238 This linker can create the new dynamic tags in ELF. But the older ELF
2239 systems may not understand them. If you specify
2240 @option{--enable-new-dtags}, the new dynamic tags will be created as needed
2241 and older dynamic tags will be omitted.
2242 If you specify @option{--disable-new-dtags}, no new dynamic tags will be
2243 created. By default, the new dynamic tags are not created. Note that
2244 those options are only available for ELF systems.
2246 @kindex --hash-size=@var{number}
2247 @item --hash-size=@var{number}
2248 Set the default size of the linker's hash tables to a prime number
2249 close to @var{number}. Increasing this value can reduce the length of
2250 time it takes the linker to perform its tasks, at the expense of
2251 increasing the linker's memory requirements. Similarly reducing this
2252 value can reduce the memory requirements at the expense of speed.
2254 @kindex --hash-style=@var{style}
2255 @item --hash-style=@var{style}
2256 Set the type of linker's hash table(s). @var{style} can be either
2257 @code{sysv} for classic ELF @code{.hash} section, @code{gnu} for
2258 new style GNU @code{.gnu.hash} section or @code{both} for both
2259 the classic ELF @code{.hash} and new style GNU @code{.gnu.hash}
2260 hash tables. The default is @code{sysv}.
2262 @kindex --compress-debug-sections=none
2263 @kindex --compress-debug-sections=zlib
2264 @kindex --compress-debug-sections=zlib-gnu
2265 @kindex --compress-debug-sections=zlib-gabi
2266 @item --compress-debug-sections=none
2267 @itemx --compress-debug-sections=zlib
2268 @itemx --compress-debug-sections=zlib-gnu
2269 @itemx --compress-debug-sections=zlib-gabi
2270 On ELF platforms , these options control how DWARF debug sections are
2271 compressed using zlib. @option{--compress-debug-sections=none} doesn't
2272 compress DWARF debug sections.
2273 @option{--compress-debug-sections=zlib-gnu} compresses DWARF debug
2274 sections and rename debug section names to begin with @samp{.zdebug}
2275 instead of @samp{.debug}. @option{--compress-debug-sections=zlib}
2276 and @option{--compress-debug-sections=zlib-gabi}
2277 compress DWARF debug sections with SHF_COMPRESSED from the ELF ABI.
2279 @kindex --reduce-memory-overheads
2280 @item --reduce-memory-overheads
2281 This option reduces memory requirements at ld runtime, at the expense of
2282 linking speed. This was introduced to select the old O(n^2) algorithm
2283 for link map file generation, rather than the new O(n) algorithm which uses
2284 about 40% more memory for symbol storage.
2286 Another effect of the switch is to set the default hash table size to
2287 1021, which again saves memory at the cost of lengthening the linker's
2288 run time. This is not done however if the @option{--hash-size} switch
2291 The @option{--reduce-memory-overheads} switch may be also be used to
2292 enable other tradeoffs in future versions of the linker.
2295 @kindex --build-id=@var{style}
2297 @itemx --build-id=@var{style}
2298 Request the creation of a @code{.note.gnu.build-id} ELF note section
2299 or a @code{.buildid} COFF section. The contents of the note are
2300 unique bits identifying this linked file. @var{style} can be
2301 @code{uuid} to use 128 random bits, @code{sha1} to use a 160-bit
2302 @sc{SHA1} hash on the normative parts of the output contents,
2303 @code{md5} to use a 128-bit @sc{MD5} hash on the normative parts of
2304 the output contents, or @code{0x@var{hexstring}} to use a chosen bit
2305 string specified as an even number of hexadecimal digits (@code{-} and
2306 @code{:} characters between digit pairs are ignored). If @var{style}
2307 is omitted, @code{sha1} is used.
2309 The @code{md5} and @code{sha1} styles produces an identifier
2310 that is always the same in an identical output file, but will be
2311 unique among all nonidentical output files. It is not intended
2312 to be compared as a checksum for the file's contents. A linked
2313 file may be changed later by other tools, but the build ID bit
2314 string identifying the original linked file does not change.
2316 Passing @code{none} for @var{style} disables the setting from any
2317 @code{--build-id} options earlier on the command line.
2322 @subsection Options Specific to i386 PE Targets
2324 @c man begin OPTIONS
2326 The i386 PE linker supports the @option{-shared} option, which causes
2327 the output to be a dynamically linked library (DLL) instead of a
2328 normal executable. You should name the output @code{*.dll} when you
2329 use this option. In addition, the linker fully supports the standard
2330 @code{*.def} files, which may be specified on the linker command line
2331 like an object file (in fact, it should precede archives it exports
2332 symbols from, to ensure that they get linked in, just like a normal
2335 In addition to the options common to all targets, the i386 PE linker
2336 support additional command line options that are specific to the i386
2337 PE target. Options that take values may be separated from their
2338 values by either a space or an equals sign.
2342 @kindex --add-stdcall-alias
2343 @item --add-stdcall-alias
2344 If given, symbols with a stdcall suffix (@@@var{nn}) will be exported
2345 as-is and also with the suffix stripped.
2346 [This option is specific to the i386 PE targeted port of the linker]
2349 @item --base-file @var{file}
2350 Use @var{file} as the name of a file in which to save the base
2351 addresses of all the relocations needed for generating DLLs with
2353 [This is an i386 PE specific option]
2357 Create a DLL instead of a regular executable. You may also use
2358 @option{-shared} or specify a @code{LIBRARY} in a given @code{.def}
2360 [This option is specific to the i386 PE targeted port of the linker]
2362 @kindex --enable-long-section-names
2363 @kindex --disable-long-section-names
2364 @item --enable-long-section-names
2365 @itemx --disable-long-section-names
2366 The PE variants of the Coff object format add an extension that permits
2367 the use of section names longer than eight characters, the normal limit
2368 for Coff. By default, these names are only allowed in object files, as
2369 fully-linked executable images do not carry the Coff string table required
2370 to support the longer names. As a GNU extension, it is possible to
2371 allow their use in executable images as well, or to (probably pointlessly!)
2372 disallow it in object files, by using these two options. Executable images
2373 generated with these long section names are slightly non-standard, carrying
2374 as they do a string table, and may generate confusing output when examined
2375 with non-GNU PE-aware tools, such as file viewers and dumpers. However,
2376 GDB relies on the use of PE long section names to find Dwarf-2 debug
2377 information sections in an executable image at runtime, and so if neither
2378 option is specified on the command-line, @command{ld} will enable long
2379 section names, overriding the default and technically correct behaviour,
2380 when it finds the presence of debug information while linking an executable
2381 image and not stripping symbols.
2382 [This option is valid for all PE targeted ports of the linker]
2384 @kindex --enable-stdcall-fixup
2385 @kindex --disable-stdcall-fixup
2386 @item --enable-stdcall-fixup
2387 @itemx --disable-stdcall-fixup
2388 If the link finds a symbol that it cannot resolve, it will attempt to
2389 do ``fuzzy linking'' by looking for another defined symbol that differs
2390 only in the format of the symbol name (cdecl vs stdcall) and will
2391 resolve that symbol by linking to the match. For example, the
2392 undefined symbol @code{_foo} might be linked to the function
2393 @code{_foo@@12}, or the undefined symbol @code{_bar@@16} might be linked
2394 to the function @code{_bar}. When the linker does this, it prints a
2395 warning, since it normally should have failed to link, but sometimes
2396 import libraries generated from third-party dlls may need this feature
2397 to be usable. If you specify @option{--enable-stdcall-fixup}, this
2398 feature is fully enabled and warnings are not printed. If you specify
2399 @option{--disable-stdcall-fixup}, this feature is disabled and such
2400 mismatches are considered to be errors.
2401 [This option is specific to the i386 PE targeted port of the linker]
2403 @kindex --leading-underscore
2404 @kindex --no-leading-underscore
2405 @item --leading-underscore
2406 @itemx --no-leading-underscore
2407 For most targets default symbol-prefix is an underscore and is defined
2408 in target's description. By this option it is possible to
2409 disable/enable the default underscore symbol-prefix.
2411 @cindex DLLs, creating
2412 @kindex --export-all-symbols
2413 @item --export-all-symbols
2414 If given, all global symbols in the objects used to build a DLL will
2415 be exported by the DLL. Note that this is the default if there
2416 otherwise wouldn't be any exported symbols. When symbols are
2417 explicitly exported via DEF files or implicitly exported via function
2418 attributes, the default is to not export anything else unless this
2419 option is given. Note that the symbols @code{DllMain@@12},
2420 @code{DllEntryPoint@@0}, @code{DllMainCRTStartup@@12}, and
2421 @code{impure_ptr} will not be automatically
2422 exported. Also, symbols imported from other DLLs will not be
2423 re-exported, nor will symbols specifying the DLL's internal layout
2424 such as those beginning with @code{_head_} or ending with
2425 @code{_iname}. In addition, no symbols from @code{libgcc},
2426 @code{libstd++}, @code{libmingw32}, or @code{crtX.o} will be exported.
2427 Symbols whose names begin with @code{__rtti_} or @code{__builtin_} will
2428 not be exported, to help with C++ DLLs. Finally, there is an
2429 extensive list of cygwin-private symbols that are not exported
2430 (obviously, this applies on when building DLLs for cygwin targets).
2431 These cygwin-excludes are: @code{_cygwin_dll_entry@@12},
2432 @code{_cygwin_crt0_common@@8}, @code{_cygwin_noncygwin_dll_entry@@12},
2433 @code{_fmode}, @code{_impure_ptr}, @code{cygwin_attach_dll},
2434 @code{cygwin_premain0}, @code{cygwin_premain1}, @code{cygwin_premain2},
2435 @code{cygwin_premain3}, and @code{environ}.
2436 [This option is specific to the i386 PE targeted port of the linker]
2438 @kindex --exclude-symbols
2439 @item --exclude-symbols @var{symbol},@var{symbol},...
2440 Specifies a list of symbols which should not be automatically
2441 exported. The symbol names may be delimited by commas or colons.
2442 [This option is specific to the i386 PE targeted port of the linker]
2444 @kindex --exclude-all-symbols
2445 @item --exclude-all-symbols
2446 Specifies no symbols should be automatically exported.
2447 [This option is specific to the i386 PE targeted port of the linker]
2449 @kindex --file-alignment
2450 @item --file-alignment
2451 Specify the file alignment. Sections in the file will always begin at
2452 file offsets which are multiples of this number. This defaults to
2454 [This option is specific to the i386 PE targeted port of the linker]
2458 @item --heap @var{reserve}
2459 @itemx --heap @var{reserve},@var{commit}
2460 Specify the number of bytes of memory to reserve (and optionally commit)
2461 to be used as heap for this program. The default is 1MB reserved, 4K
2463 [This option is specific to the i386 PE targeted port of the linker]
2466 @kindex --image-base
2467 @item --image-base @var{value}
2468 Use @var{value} as the base address of your program or dll. This is
2469 the lowest memory location that will be used when your program or dll
2470 is loaded. To reduce the need to relocate and improve performance of
2471 your dlls, each should have a unique base address and not overlap any
2472 other dlls. The default is 0x400000 for executables, and 0x10000000
2474 [This option is specific to the i386 PE targeted port of the linker]
2478 If given, the stdcall suffixes (@@@var{nn}) will be stripped from
2479 symbols before they are exported.
2480 [This option is specific to the i386 PE targeted port of the linker]
2482 @kindex --large-address-aware
2483 @item --large-address-aware
2484 If given, the appropriate bit in the ``Characteristics'' field of the COFF
2485 header is set to indicate that this executable supports virtual addresses
2486 greater than 2 gigabytes. This should be used in conjunction with the /3GB
2487 or /USERVA=@var{value} megabytes switch in the ``[operating systems]''
2488 section of the BOOT.INI. Otherwise, this bit has no effect.
2489 [This option is specific to PE targeted ports of the linker]
2491 @kindex --disable-large-address-aware
2492 @item --disable-large-address-aware
2493 Reverts the effect of a previous @samp{--large-address-aware} option.
2494 This is useful if @samp{--large-address-aware} is always set by the compiler
2495 driver (e.g. Cygwin gcc) and the executable does not support virtual
2496 addresses greater than 2 gigabytes.
2497 [This option is specific to PE targeted ports of the linker]
2499 @kindex --major-image-version
2500 @item --major-image-version @var{value}
2501 Sets the major number of the ``image version''. Defaults to 1.
2502 [This option is specific to the i386 PE targeted port of the linker]
2504 @kindex --major-os-version
2505 @item --major-os-version @var{value}
2506 Sets the major number of the ``os version''. Defaults to 4.
2507 [This option is specific to the i386 PE targeted port of the linker]
2509 @kindex --major-subsystem-version
2510 @item --major-subsystem-version @var{value}
2511 Sets the major number of the ``subsystem version''. Defaults to 4.
2512 [This option is specific to the i386 PE targeted port of the linker]
2514 @kindex --minor-image-version
2515 @item --minor-image-version @var{value}
2516 Sets the minor number of the ``image version''. Defaults to 0.
2517 [This option is specific to the i386 PE targeted port of the linker]
2519 @kindex --minor-os-version
2520 @item --minor-os-version @var{value}
2521 Sets the minor number of the ``os version''. Defaults to 0.
2522 [This option is specific to the i386 PE targeted port of the linker]
2524 @kindex --minor-subsystem-version
2525 @item --minor-subsystem-version @var{value}
2526 Sets the minor number of the ``subsystem version''. Defaults to 0.
2527 [This option is specific to the i386 PE targeted port of the linker]
2529 @cindex DEF files, creating
2530 @cindex DLLs, creating
2531 @kindex --output-def
2532 @item --output-def @var{file}
2533 The linker will create the file @var{file} which will contain a DEF
2534 file corresponding to the DLL the linker is generating. This DEF file
2535 (which should be called @code{*.def}) may be used to create an import
2536 library with @code{dlltool} or may be used as a reference to
2537 automatically or implicitly exported symbols.
2538 [This option is specific to the i386 PE targeted port of the linker]
2540 @cindex DLLs, creating
2541 @kindex --out-implib
2542 @item --out-implib @var{file}
2543 The linker will create the file @var{file} which will contain an
2544 import lib corresponding to the DLL the linker is generating. This
2545 import lib (which should be called @code{*.dll.a} or @code{*.a}
2546 may be used to link clients against the generated DLL; this behaviour
2547 makes it possible to skip a separate @code{dlltool} import library
2549 [This option is specific to the i386 PE targeted port of the linker]
2551 @kindex --enable-auto-image-base
2552 @item --enable-auto-image-base
2553 @itemx --enable-auto-image-base=@var{value}
2554 Automatically choose the image base for DLLs, optionally starting with base
2555 @var{value}, unless one is specified using the @code{--image-base} argument.
2556 By using a hash generated from the dllname to create unique image bases
2557 for each DLL, in-memory collisions and relocations which can delay program
2558 execution are avoided.
2559 [This option is specific to the i386 PE targeted port of the linker]
2561 @kindex --disable-auto-image-base
2562 @item --disable-auto-image-base
2563 Do not automatically generate a unique image base. If there is no
2564 user-specified image base (@code{--image-base}) then use the platform
2566 [This option is specific to the i386 PE targeted port of the linker]
2568 @cindex DLLs, linking to
2569 @kindex --dll-search-prefix
2570 @item --dll-search-prefix @var{string}
2571 When linking dynamically to a dll without an import library,
2572 search for @code{<string><basename>.dll} in preference to
2573 @code{lib<basename>.dll}. This behaviour allows easy distinction
2574 between DLLs built for the various "subplatforms": native, cygwin,
2575 uwin, pw, etc. For instance, cygwin DLLs typically use
2576 @code{--dll-search-prefix=cyg}.
2577 [This option is specific to the i386 PE targeted port of the linker]
2579 @kindex --enable-auto-import
2580 @item --enable-auto-import
2581 Do sophisticated linking of @code{_symbol} to @code{__imp__symbol} for
2582 DATA imports from DLLs, and create the necessary thunking symbols when
2583 building the import libraries with those DATA exports. Note: Use of the
2584 'auto-import' extension will cause the text section of the image file
2585 to be made writable. This does not conform to the PE-COFF format
2586 specification published by Microsoft.
2588 Note - use of the 'auto-import' extension will also cause read only
2589 data which would normally be placed into the .rdata section to be
2590 placed into the .data section instead. This is in order to work
2591 around a problem with consts that is described here:
2592 http://www.cygwin.com/ml/cygwin/2004-09/msg01101.html
2594 Using 'auto-import' generally will 'just work' -- but sometimes you may
2597 "variable '<var>' can't be auto-imported. Please read the
2598 documentation for ld's @code{--enable-auto-import} for details."
2600 This message occurs when some (sub)expression accesses an address
2601 ultimately given by the sum of two constants (Win32 import tables only
2602 allow one). Instances where this may occur include accesses to member
2603 fields of struct variables imported from a DLL, as well as using a
2604 constant index into an array variable imported from a DLL. Any
2605 multiword variable (arrays, structs, long long, etc) may trigger
2606 this error condition. However, regardless of the exact data type
2607 of the offending exported variable, ld will always detect it, issue
2608 the warning, and exit.
2610 There are several ways to address this difficulty, regardless of the
2611 data type of the exported variable:
2613 One way is to use --enable-runtime-pseudo-reloc switch. This leaves the task
2614 of adjusting references in your client code for runtime environment, so
2615 this method works only when runtime environment supports this feature.
2617 A second solution is to force one of the 'constants' to be a variable --
2618 that is, unknown and un-optimizable at compile time. For arrays,
2619 there are two possibilities: a) make the indexee (the array's address)
2620 a variable, or b) make the 'constant' index a variable. Thus:
2623 extern type extern_array[];
2625 @{ volatile type *t=extern_array; t[1] @}
2631 extern type extern_array[];
2633 @{ volatile int t=1; extern_array[t] @}
2636 For structs (and most other multiword data types) the only option
2637 is to make the struct itself (or the long long, or the ...) variable:
2640 extern struct s extern_struct;
2641 extern_struct.field -->
2642 @{ volatile struct s *t=&extern_struct; t->field @}
2648 extern long long extern_ll;
2650 @{ volatile long long * local_ll=&extern_ll; *local_ll @}
2653 A third method of dealing with this difficulty is to abandon
2654 'auto-import' for the offending symbol and mark it with
2655 @code{__declspec(dllimport)}. However, in practice that
2656 requires using compile-time #defines to indicate whether you are
2657 building a DLL, building client code that will link to the DLL, or
2658 merely building/linking to a static library. In making the choice
2659 between the various methods of resolving the 'direct address with
2660 constant offset' problem, you should consider typical real-world usage:
2668 void main(int argc, char **argv)@{
2669 printf("%d\n",arr[1]);
2679 void main(int argc, char **argv)@{
2680 /* This workaround is for win32 and cygwin; do not "optimize" */
2681 volatile int *parr = arr;
2682 printf("%d\n",parr[1]);
2689 /* Note: auto-export is assumed (no __declspec(dllexport)) */
2690 #if (defined(_WIN32) || defined(__CYGWIN__)) && \
2691 !(defined(FOO_BUILD_DLL) || defined(FOO_STATIC))
2692 #define FOO_IMPORT __declspec(dllimport)
2696 extern FOO_IMPORT int arr[];
2699 void main(int argc, char **argv)@{
2700 printf("%d\n",arr[1]);
2704 A fourth way to avoid this problem is to re-code your
2705 library to use a functional interface rather than a data interface
2706 for the offending variables (e.g. set_foo() and get_foo() accessor
2708 [This option is specific to the i386 PE targeted port of the linker]
2710 @kindex --disable-auto-import
2711 @item --disable-auto-import
2712 Do not attempt to do sophisticated linking of @code{_symbol} to
2713 @code{__imp__symbol} for DATA imports from DLLs.
2714 [This option is specific to the i386 PE targeted port of the linker]
2716 @kindex --enable-runtime-pseudo-reloc
2717 @item --enable-runtime-pseudo-reloc
2718 If your code contains expressions described in --enable-auto-import section,
2719 that is, DATA imports from DLL with non-zero offset, this switch will create
2720 a vector of 'runtime pseudo relocations' which can be used by runtime
2721 environment to adjust references to such data in your client code.
2722 [This option is specific to the i386 PE targeted port of the linker]
2724 @kindex --disable-runtime-pseudo-reloc
2725 @item --disable-runtime-pseudo-reloc
2726 Do not create pseudo relocations for non-zero offset DATA imports from
2728 [This option is specific to the i386 PE targeted port of the linker]
2730 @kindex --enable-extra-pe-debug
2731 @item --enable-extra-pe-debug
2732 Show additional debug info related to auto-import symbol thunking.
2733 [This option is specific to the i386 PE targeted port of the linker]
2735 @kindex --section-alignment
2736 @item --section-alignment
2737 Sets the section alignment. Sections in memory will always begin at
2738 addresses which are a multiple of this number. Defaults to 0x1000.
2739 [This option is specific to the i386 PE targeted port of the linker]
2743 @item --stack @var{reserve}
2744 @itemx --stack @var{reserve},@var{commit}
2745 Specify the number of bytes of memory to reserve (and optionally commit)
2746 to be used as stack for this program. The default is 2MB reserved, 4K
2748 [This option is specific to the i386 PE targeted port of the linker]
2751 @item --subsystem @var{which}
2752 @itemx --subsystem @var{which}:@var{major}
2753 @itemx --subsystem @var{which}:@var{major}.@var{minor}
2754 Specifies the subsystem under which your program will execute. The
2755 legal values for @var{which} are @code{native}, @code{windows},
2756 @code{console}, @code{posix}, and @code{xbox}. You may optionally set
2757 the subsystem version also. Numeric values are also accepted for
2759 [This option is specific to the i386 PE targeted port of the linker]
2761 The following options set flags in the @code{DllCharacteristics} field
2762 of the PE file header:
2763 [These options are specific to PE targeted ports of the linker]
2765 @kindex --high-entropy-va
2766 @item --high-entropy-va
2767 Image is compatible with 64-bit address space layout randomization
2770 @kindex --dynamicbase
2772 The image base address may be relocated using address space layout
2773 randomization (ASLR). This feature was introduced with MS Windows
2774 Vista for i386 PE targets.
2776 @kindex --forceinteg
2778 Code integrity checks are enforced.
2782 The image is compatible with the Data Execution Prevention.
2783 This feature was introduced with MS Windows XP SP2 for i386 PE targets.
2785 @kindex --no-isolation
2786 @item --no-isolation
2787 Although the image understands isolation, do not isolate the image.
2791 The image does not use SEH. No SE handler may be called from
2796 Do not bind this image.
2800 The driver uses the MS Windows Driver Model.
2804 The image is Terminal Server aware.
2806 @kindex --insert-timestamp
2807 @item --insert-timestamp
2808 @itemx --no-insert-timestamp
2809 Insert a real timestamp into the image. This is the default behaviour
2810 as it matches legacy code and it means that the image will work with
2811 other, proprietary tools. The problem with this default is that it
2812 will result in slightly different images being produced each tiem the
2813 same sources are linked. The option @option{--no-insert-timestamp}
2814 can be used to insert a zero value for the timestamp, this ensuring
2815 that binaries produced from indentical sources will compare
2822 @subsection Options specific to C6X uClinux targets
2824 @c man begin OPTIONS
2826 The C6X uClinux target uses a binary format called DSBT to support shared
2827 libraries. Each shared library in the system needs to have a unique index;
2828 all executables use an index of 0.
2833 @item --dsbt-size @var{size}
2834 This option sets the number of entires in the DSBT of the current executable
2835 or shared library to @var{size}. The default is to create a table with 64
2838 @kindex --dsbt-index
2839 @item --dsbt-index @var{index}
2840 This option sets the DSBT index of the current executable or shared library
2841 to @var{index}. The default is 0, which is appropriate for generating
2842 executables. If a shared library is generated with a DSBT index of 0, the
2843 @code{R_C6000_DSBT_INDEX} relocs are copied into the output file.
2845 @kindex --no-merge-exidx-entries
2846 The @samp{--no-merge-exidx-entries} switch disables the merging of adjacent
2847 exidx entries in frame unwind info.
2855 @subsection Options specific to Motorola 68HC11 and 68HC12 targets
2857 @c man begin OPTIONS
2859 The 68HC11 and 68HC12 linkers support specific options to control the
2860 memory bank switching mapping and trampoline code generation.
2864 @kindex --no-trampoline
2865 @item --no-trampoline
2866 This option disables the generation of trampoline. By default a trampoline
2867 is generated for each far function which is called using a @code{jsr}
2868 instruction (this happens when a pointer to a far function is taken).
2870 @kindex --bank-window
2871 @item --bank-window @var{name}
2872 This option indicates to the linker the name of the memory region in
2873 the @samp{MEMORY} specification that describes the memory bank window.
2874 The definition of such region is then used by the linker to compute
2875 paging and addresses within the memory window.
2883 @subsection Options specific to Motorola 68K target
2885 @c man begin OPTIONS
2887 The following options are supported to control handling of GOT generation
2888 when linking for 68K targets.
2893 @item --got=@var{type}
2894 This option tells the linker which GOT generation scheme to use.
2895 @var{type} should be one of @samp{single}, @samp{negative},
2896 @samp{multigot} or @samp{target}. For more information refer to the
2897 Info entry for @file{ld}.
2905 @subsection Options specific to MIPS targets
2907 @c man begin OPTIONS
2909 The following options are supported to control microMIPS instruction
2910 generation when linking for MIPS targets.
2918 These options control the choice of microMIPS instructions used in code
2919 generated by the linker, such as that in the PLT or lazy binding stubs,
2920 or in relaxation. If @samp{--insn32} is used, then the linker only uses
2921 32-bit instruction encodings. By default or if @samp{--no-insn32} is
2922 used, all instruction encodings are used, including 16-bit ones where
2932 @section Environment Variables
2934 @c man begin ENVIRONMENT
2936 You can change the behaviour of @command{ld} with the environment variables
2937 @ifclear SingleFormat
2940 @code{LDEMULATION} and @code{COLLECT_NO_DEMANGLE}.
2942 @ifclear SingleFormat
2944 @cindex default input format
2945 @code{GNUTARGET} determines the input-file object format if you don't
2946 use @samp{-b} (or its synonym @samp{--format}). Its value should be one
2947 of the BFD names for an input format (@pxref{BFD}). If there is no
2948 @code{GNUTARGET} in the environment, @command{ld} uses the natural format
2949 of the target. If @code{GNUTARGET} is set to @code{default} then BFD
2950 attempts to discover the input format by examining binary input files;
2951 this method often succeeds, but there are potential ambiguities, since
2952 there is no method of ensuring that the magic number used to specify
2953 object-file formats is unique. However, the configuration procedure for
2954 BFD on each system places the conventional format for that system first
2955 in the search-list, so ambiguities are resolved in favor of convention.
2959 @cindex default emulation
2960 @cindex emulation, default
2961 @code{LDEMULATION} determines the default emulation if you don't use the
2962 @samp{-m} option. The emulation can affect various aspects of linker
2963 behaviour, particularly the default linker script. You can list the
2964 available emulations with the @samp{--verbose} or @samp{-V} options. If
2965 the @samp{-m} option is not used, and the @code{LDEMULATION} environment
2966 variable is not defined, the default emulation depends upon how the
2967 linker was configured.
2969 @kindex COLLECT_NO_DEMANGLE
2970 @cindex demangling, default
2971 Normally, the linker will default to demangling symbols. However, if
2972 @code{COLLECT_NO_DEMANGLE} is set in the environment, then it will
2973 default to not demangling symbols. This environment variable is used in
2974 a similar fashion by the @code{gcc} linker wrapper program. The default
2975 may be overridden by the @samp{--demangle} and @samp{--no-demangle}
2982 @chapter Linker Scripts
2985 @cindex linker scripts
2986 @cindex command files
2987 Every link is controlled by a @dfn{linker script}. This script is
2988 written in the linker command language.
2990 The main purpose of the linker script is to describe how the sections in
2991 the input files should be mapped into the output file, and to control
2992 the memory layout of the output file. Most linker scripts do nothing
2993 more than this. However, when necessary, the linker script can also
2994 direct the linker to perform many other operations, using the commands
2997 The linker always uses a linker script. If you do not supply one
2998 yourself, the linker will use a default script that is compiled into the
2999 linker executable. You can use the @samp{--verbose} command line option
3000 to display the default linker script. Certain command line options,
3001 such as @samp{-r} or @samp{-N}, will affect the default linker script.
3003 You may supply your own linker script by using the @samp{-T} command
3004 line option. When you do this, your linker script will replace the
3005 default linker script.
3007 You may also use linker scripts implicitly by naming them as input files
3008 to the linker, as though they were files to be linked. @xref{Implicit
3012 * Basic Script Concepts:: Basic Linker Script Concepts
3013 * Script Format:: Linker Script Format
3014 * Simple Example:: Simple Linker Script Example
3015 * Simple Commands:: Simple Linker Script Commands
3016 * Assignments:: Assigning Values to Symbols
3017 * SECTIONS:: SECTIONS Command
3018 * MEMORY:: MEMORY Command
3019 * PHDRS:: PHDRS Command
3020 * VERSION:: VERSION Command
3021 * Expressions:: Expressions in Linker Scripts
3022 * Implicit Linker Scripts:: Implicit Linker Scripts
3025 @node Basic Script Concepts
3026 @section Basic Linker Script Concepts
3027 @cindex linker script concepts
3028 We need to define some basic concepts and vocabulary in order to
3029 describe the linker script language.
3031 The linker combines input files into a single output file. The output
3032 file and each input file are in a special data format known as an
3033 @dfn{object file format}. Each file is called an @dfn{object file}.
3034 The output file is often called an @dfn{executable}, but for our
3035 purposes we will also call it an object file. Each object file has,
3036 among other things, a list of @dfn{sections}. We sometimes refer to a
3037 section in an input file as an @dfn{input section}; similarly, a section
3038 in the output file is an @dfn{output section}.
3040 Each section in an object file has a name and a size. Most sections
3041 also have an associated block of data, known as the @dfn{section
3042 contents}. A section may be marked as @dfn{loadable}, which means that
3043 the contents should be loaded into memory when the output file is run.
3044 A section with no contents may be @dfn{allocatable}, which means that an
3045 area in memory should be set aside, but nothing in particular should be
3046 loaded there (in some cases this memory must be zeroed out). A section
3047 which is neither loadable nor allocatable typically contains some sort
3048 of debugging information.
3050 Every loadable or allocatable output section has two addresses. The
3051 first is the @dfn{VMA}, or virtual memory address. This is the address
3052 the section will have when the output file is run. The second is the
3053 @dfn{LMA}, or load memory address. This is the address at which the
3054 section will be loaded. In most cases the two addresses will be the
3055 same. An example of when they might be different is when a data section
3056 is loaded into ROM, and then copied into RAM when the program starts up
3057 (this technique is often used to initialize global variables in a ROM
3058 based system). In this case the ROM address would be the LMA, and the
3059 RAM address would be the VMA.
3061 You can see the sections in an object file by using the @code{objdump}
3062 program with the @samp{-h} option.
3064 Every object file also has a list of @dfn{symbols}, known as the
3065 @dfn{symbol table}. A symbol may be defined or undefined. Each symbol
3066 has a name, and each defined symbol has an address, among other
3067 information. If you compile a C or C++ program into an object file, you
3068 will get a defined symbol for every defined function and global or
3069 static variable. Every undefined function or global variable which is
3070 referenced in the input file will become an undefined symbol.
3072 You can see the symbols in an object file by using the @code{nm}
3073 program, or by using the @code{objdump} program with the @samp{-t}
3077 @section Linker Script Format
3078 @cindex linker script format
3079 Linker scripts are text files.
3081 You write a linker script as a series of commands. Each command is
3082 either a keyword, possibly followed by arguments, or an assignment to a
3083 symbol. You may separate commands using semicolons. Whitespace is
3086 Strings such as file or format names can normally be entered directly.
3087 If the file name contains a character such as a comma which would
3088 otherwise serve to separate file names, you may put the file name in
3089 double quotes. There is no way to use a double quote character in a
3092 You may include comments in linker scripts just as in C, delimited by
3093 @samp{/*} and @samp{*/}. As in C, comments are syntactically equivalent
3096 @node Simple Example
3097 @section Simple Linker Script Example
3098 @cindex linker script example
3099 @cindex example of linker script
3100 Many linker scripts are fairly simple.
3102 The simplest possible linker script has just one command:
3103 @samp{SECTIONS}. You use the @samp{SECTIONS} command to describe the
3104 memory layout of the output file.
3106 The @samp{SECTIONS} command is a powerful command. Here we will
3107 describe a simple use of it. Let's assume your program consists only of
3108 code, initialized data, and uninitialized data. These will be in the
3109 @samp{.text}, @samp{.data}, and @samp{.bss} sections, respectively.
3110 Let's assume further that these are the only sections which appear in
3113 For this example, let's say that the code should be loaded at address
3114 0x10000, and that the data should start at address 0x8000000. Here is a
3115 linker script which will do that:
3120 .text : @{ *(.text) @}
3122 .data : @{ *(.data) @}
3123 .bss : @{ *(.bss) @}
3127 You write the @samp{SECTIONS} command as the keyword @samp{SECTIONS},
3128 followed by a series of symbol assignments and output section
3129 descriptions enclosed in curly braces.
3131 The first line inside the @samp{SECTIONS} command of the above example
3132 sets the value of the special symbol @samp{.}, which is the location
3133 counter. If you do not specify the address of an output section in some
3134 other way (other ways are described later), the address is set from the
3135 current value of the location counter. The location counter is then
3136 incremented by the size of the output section. At the start of the
3137 @samp{SECTIONS} command, the location counter has the value @samp{0}.
3139 The second line defines an output section, @samp{.text}. The colon is
3140 required syntax which may be ignored for now. Within the curly braces
3141 after the output section name, you list the names of the input sections
3142 which should be placed into this output section. The @samp{*} is a
3143 wildcard which matches any file name. The expression @samp{*(.text)}
3144 means all @samp{.text} input sections in all input files.
3146 Since the location counter is @samp{0x10000} when the output section
3147 @samp{.text} is defined, the linker will set the address of the
3148 @samp{.text} section in the output file to be @samp{0x10000}.
3150 The remaining lines define the @samp{.data} and @samp{.bss} sections in
3151 the output file. The linker will place the @samp{.data} output section
3152 at address @samp{0x8000000}. After the linker places the @samp{.data}
3153 output section, the value of the location counter will be
3154 @samp{0x8000000} plus the size of the @samp{.data} output section. The
3155 effect is that the linker will place the @samp{.bss} output section
3156 immediately after the @samp{.data} output section in memory.
3158 The linker will ensure that each output section has the required
3159 alignment, by increasing the location counter if necessary. In this
3160 example, the specified addresses for the @samp{.text} and @samp{.data}
3161 sections will probably satisfy any alignment constraints, but the linker
3162 may have to create a small gap between the @samp{.data} and @samp{.bss}
3165 That's it! That's a simple and complete linker script.
3167 @node Simple Commands
3168 @section Simple Linker Script Commands
3169 @cindex linker script simple commands
3170 In this section we describe the simple linker script commands.
3173 * Entry Point:: Setting the entry point
3174 * File Commands:: Commands dealing with files
3175 @ifclear SingleFormat
3176 * Format Commands:: Commands dealing with object file formats
3179 * REGION_ALIAS:: Assign alias names to memory regions
3180 * Miscellaneous Commands:: Other linker script commands
3184 @subsection Setting the Entry Point
3185 @kindex ENTRY(@var{symbol})
3186 @cindex start of execution
3187 @cindex first instruction
3189 The first instruction to execute in a program is called the @dfn{entry
3190 point}. You can use the @code{ENTRY} linker script command to set the
3191 entry point. The argument is a symbol name:
3196 There are several ways to set the entry point. The linker will set the
3197 entry point by trying each of the following methods in order, and
3198 stopping when one of them succeeds:
3201 the @samp{-e} @var{entry} command-line option;
3203 the @code{ENTRY(@var{symbol})} command in a linker script;
3205 the value of a target specific symbol, if it is defined; For many
3206 targets this is @code{start}, but PE and BeOS based systems for example
3207 check a list of possible entry symbols, matching the first one found.
3209 the address of the first byte of the @samp{.text} section, if present;
3211 The address @code{0}.
3215 @subsection Commands Dealing with Files
3216 @cindex linker script file commands
3217 Several linker script commands deal with files.
3220 @item INCLUDE @var{filename}
3221 @kindex INCLUDE @var{filename}
3222 @cindex including a linker script
3223 Include the linker script @var{filename} at this point. The file will
3224 be searched for in the current directory, and in any directory specified
3225 with the @option{-L} option. You can nest calls to @code{INCLUDE} up to
3228 You can place @code{INCLUDE} directives at the top level, in @code{MEMORY} or
3229 @code{SECTIONS} commands, or in output section descriptions.
3231 @item INPUT(@var{file}, @var{file}, @dots{})
3232 @itemx INPUT(@var{file} @var{file} @dots{})
3233 @kindex INPUT(@var{files})
3234 @cindex input files in linker scripts
3235 @cindex input object files in linker scripts
3236 @cindex linker script input object files
3237 The @code{INPUT} command directs the linker to include the named files
3238 in the link, as though they were named on the command line.
3240 For example, if you always want to include @file{subr.o} any time you do
3241 a link, but you can't be bothered to put it on every link command line,
3242 then you can put @samp{INPUT (subr.o)} in your linker script.
3244 In fact, if you like, you can list all of your input files in the linker
3245 script, and then invoke the linker with nothing but a @samp{-T} option.
3247 In case a @dfn{sysroot prefix} is configured, and the filename starts
3248 with the @samp{/} character, and the script being processed was
3249 located inside the @dfn{sysroot prefix}, the filename will be looked
3250 for in the @dfn{sysroot prefix}. Otherwise, the linker will try to
3251 open the file in the current directory. If it is not found, the
3252 linker will search through the archive library search path.
3253 The @dfn{sysroot prefix} can also be forced by specifying @code{=}
3254 as the first character in the filename path. See also the
3255 description of @samp{-L} in @ref{Options,,Command Line Options}.
3257 If you use @samp{INPUT (-l@var{file})}, @command{ld} will transform the
3258 name to @code{lib@var{file}.a}, as with the command line argument
3261 When you use the @code{INPUT} command in an implicit linker script, the
3262 files will be included in the link at the point at which the linker
3263 script file is included. This can affect archive searching.
3265 @item GROUP(@var{file}, @var{file}, @dots{})
3266 @itemx GROUP(@var{file} @var{file} @dots{})
3267 @kindex GROUP(@var{files})
3268 @cindex grouping input files
3269 The @code{GROUP} command is like @code{INPUT}, except that the named
3270 files should all be archives, and they are searched repeatedly until no
3271 new undefined references are created. See the description of @samp{-(}
3272 in @ref{Options,,Command Line Options}.
3274 @item AS_NEEDED(@var{file}, @var{file}, @dots{})
3275 @itemx AS_NEEDED(@var{file} @var{file} @dots{})
3276 @kindex AS_NEEDED(@var{files})
3277 This construct can appear only inside of the @code{INPUT} or @code{GROUP}
3278 commands, among other filenames. The files listed will be handled
3279 as if they appear directly in the @code{INPUT} or @code{GROUP} commands,
3280 with the exception of ELF shared libraries, that will be added only
3281 when they are actually needed. This construct essentially enables
3282 @option{--as-needed} option for all the files listed inside of it
3283 and restores previous @option{--as-needed} resp. @option{--no-as-needed}
3286 @item OUTPUT(@var{filename})
3287 @kindex OUTPUT(@var{filename})
3288 @cindex output file name in linker script
3289 The @code{OUTPUT} command names the output file. Using
3290 @code{OUTPUT(@var{filename})} in the linker script is exactly like using
3291 @samp{-o @var{filename}} on the command line (@pxref{Options,,Command
3292 Line Options}). If both are used, the command line option takes
3295 You can use the @code{OUTPUT} command to define a default name for the
3296 output file other than the usual default of @file{a.out}.
3298 @item SEARCH_DIR(@var{path})
3299 @kindex SEARCH_DIR(@var{path})
3300 @cindex library search path in linker script
3301 @cindex archive search path in linker script
3302 @cindex search path in linker script
3303 The @code{SEARCH_DIR} command adds @var{path} to the list of paths where
3304 @command{ld} looks for archive libraries. Using
3305 @code{SEARCH_DIR(@var{path})} is exactly like using @samp{-L @var{path}}
3306 on the command line (@pxref{Options,,Command Line Options}). If both
3307 are used, then the linker will search both paths. Paths specified using
3308 the command line option are searched first.
3310 @item STARTUP(@var{filename})
3311 @kindex STARTUP(@var{filename})
3312 @cindex first input file
3313 The @code{STARTUP} command is just like the @code{INPUT} command, except
3314 that @var{filename} will become the first input file to be linked, as
3315 though it were specified first on the command line. This may be useful
3316 when using a system in which the entry point is always the start of the
3320 @ifclear SingleFormat
3321 @node Format Commands
3322 @subsection Commands Dealing with Object File Formats
3323 A couple of linker script commands deal with object file formats.
3326 @item OUTPUT_FORMAT(@var{bfdname})
3327 @itemx OUTPUT_FORMAT(@var{default}, @var{big}, @var{little})
3328 @kindex OUTPUT_FORMAT(@var{bfdname})
3329 @cindex output file format in linker script
3330 The @code{OUTPUT_FORMAT} command names the BFD format to use for the
3331 output file (@pxref{BFD}). Using @code{OUTPUT_FORMAT(@var{bfdname})} is
3332 exactly like using @samp{--oformat @var{bfdname}} on the command line
3333 (@pxref{Options,,Command Line Options}). If both are used, the command
3334 line option takes precedence.
3336 You can use @code{OUTPUT_FORMAT} with three arguments to use different
3337 formats based on the @samp{-EB} and @samp{-EL} command line options.
3338 This permits the linker script to set the output format based on the
3341 If neither @samp{-EB} nor @samp{-EL} are used, then the output format
3342 will be the first argument, @var{default}. If @samp{-EB} is used, the
3343 output format will be the second argument, @var{big}. If @samp{-EL} is
3344 used, the output format will be the third argument, @var{little}.
3346 For example, the default linker script for the MIPS ELF target uses this
3349 OUTPUT_FORMAT(elf32-bigmips, elf32-bigmips, elf32-littlemips)
3351 This says that the default format for the output file is
3352 @samp{elf32-bigmips}, but if the user uses the @samp{-EL} command line
3353 option, the output file will be created in the @samp{elf32-littlemips}
3356 @item TARGET(@var{bfdname})
3357 @kindex TARGET(@var{bfdname})
3358 @cindex input file format in linker script
3359 The @code{TARGET} command names the BFD format to use when reading input
3360 files. It affects subsequent @code{INPUT} and @code{GROUP} commands.
3361 This command is like using @samp{-b @var{bfdname}} on the command line
3362 (@pxref{Options,,Command Line Options}). If the @code{TARGET} command
3363 is used but @code{OUTPUT_FORMAT} is not, then the last @code{TARGET}
3364 command is also used to set the format for the output file. @xref{BFD}.
3369 @subsection Assign alias names to memory regions
3370 @kindex REGION_ALIAS(@var{alias}, @var{region})
3371 @cindex region alias
3372 @cindex region names
3374 Alias names can be added to existing memory regions created with the
3375 @ref{MEMORY} command. Each name corresponds to at most one memory region.
3378 REGION_ALIAS(@var{alias}, @var{region})
3381 The @code{REGION_ALIAS} function creates an alias name @var{alias} for the
3382 memory region @var{region}. This allows a flexible mapping of output sections
3383 to memory regions. An example follows.
3385 Suppose we have an application for embedded systems which come with various
3386 memory storage devices. All have a general purpose, volatile memory @code{RAM}
3387 that allows code execution or data storage. Some may have a read-only,
3388 non-volatile memory @code{ROM} that allows code execution and read-only data
3389 access. The last variant is a read-only, non-volatile memory @code{ROM2} with
3390 read-only data access and no code execution capability. We have four output
3395 @code{.text} program code;
3397 @code{.rodata} read-only data;
3399 @code{.data} read-write initialized data;
3401 @code{.bss} read-write zero initialized data.
3404 The goal is to provide a linker command file that contains a system independent
3405 part defining the output sections and a system dependent part mapping the
3406 output sections to the memory regions available on the system. Our embedded
3407 systems come with three different memory setups @code{A}, @code{B} and
3409 @multitable @columnfractions .25 .25 .25 .25
3410 @item Section @tab Variant A @tab Variant B @tab Variant C
3411 @item .text @tab RAM @tab ROM @tab ROM
3412 @item .rodata @tab RAM @tab ROM @tab ROM2
3413 @item .data @tab RAM @tab RAM/ROM @tab RAM/ROM2
3414 @item .bss @tab RAM @tab RAM @tab RAM
3416 The notation @code{RAM/ROM} or @code{RAM/ROM2} means that this section is
3417 loaded into region @code{ROM} or @code{ROM2} respectively. Please note that
3418 the load address of the @code{.data} section starts in all three variants at
3419 the end of the @code{.rodata} section.
3421 The base linker script that deals with the output sections follows. It
3422 includes the system dependent @code{linkcmds.memory} file that describes the
3425 INCLUDE linkcmds.memory
3438 .data : AT (rodata_end)
3443 data_size = SIZEOF(.data);
3444 data_load_start = LOADADDR(.data);
3452 Now we need three different @code{linkcmds.memory} files to define memory
3453 regions and alias names. The content of @code{linkcmds.memory} for the three
3454 variants @code{A}, @code{B} and @code{C}:
3457 Here everything goes into the @code{RAM}.
3461 RAM : ORIGIN = 0, LENGTH = 4M
3464 REGION_ALIAS("REGION_TEXT", RAM);
3465 REGION_ALIAS("REGION_RODATA", RAM);
3466 REGION_ALIAS("REGION_DATA", RAM);
3467 REGION_ALIAS("REGION_BSS", RAM);
3470 Program code and read-only data go into the @code{ROM}. Read-write data goes
3471 into the @code{RAM}. An image of the initialized data is loaded into the
3472 @code{ROM} and will be copied during system start into the @code{RAM}.
3476 ROM : ORIGIN = 0, LENGTH = 3M
3477 RAM : ORIGIN = 0x10000000, LENGTH = 1M
3480 REGION_ALIAS("REGION_TEXT", ROM);
3481 REGION_ALIAS("REGION_RODATA", ROM);
3482 REGION_ALIAS("REGION_DATA", RAM);
3483 REGION_ALIAS("REGION_BSS", RAM);
3486 Program code goes into the @code{ROM}. Read-only data goes into the
3487 @code{ROM2}. Read-write data goes into the @code{RAM}. An image of the
3488 initialized data is loaded into the @code{ROM2} and will be copied during
3489 system start into the @code{RAM}.
3493 ROM : ORIGIN = 0, LENGTH = 2M
3494 ROM2 : ORIGIN = 0x10000000, LENGTH = 1M
3495 RAM : ORIGIN = 0x20000000, LENGTH = 1M
3498 REGION_ALIAS("REGION_TEXT", ROM);
3499 REGION_ALIAS("REGION_RODATA", ROM2);
3500 REGION_ALIAS("REGION_DATA", RAM);
3501 REGION_ALIAS("REGION_BSS", RAM);
3505 It is possible to write a common system initialization routine to copy the
3506 @code{.data} section from @code{ROM} or @code{ROM2} into the @code{RAM} if
3511 extern char data_start [];
3512 extern char data_size [];
3513 extern char data_load_start [];
3515 void copy_data(void)
3517 if (data_start != data_load_start)
3519 memcpy(data_start, data_load_start, (size_t) data_size);
3524 @node Miscellaneous Commands
3525 @subsection Other Linker Script Commands
3526 There are a few other linker scripts commands.
3529 @item ASSERT(@var{exp}, @var{message})
3531 @cindex assertion in linker script
3532 Ensure that @var{exp} is non-zero. If it is zero, then exit the linker
3533 with an error code, and print @var{message}.
3535 Note that assertions are checked before the final stages of linking
3536 take place. This means that expressions involving symbols PROVIDEd
3537 inside section definitions will fail if the user has not set values
3538 for those symbols. The only exception to this rule is PROVIDEd
3539 symbols that just reference dot. Thus an assertion like this:
3544 PROVIDE (__stack = .);
3545 PROVIDE (__stack_size = 0x100);
3546 ASSERT ((__stack > (_end + __stack_size)), "Error: No room left for the stack");
3550 will fail if @code{__stack_size} is not defined elsewhere. Symbols
3551 PROVIDEd outside of section definitions are evaluated earlier, so they
3552 can be used inside ASSERTions. Thus:
3555 PROVIDE (__stack_size = 0x100);
3558 PROVIDE (__stack = .);
3559 ASSERT ((__stack > (_end + __stack_size)), "Error: No room left for the stack");
3565 @item EXTERN(@var{symbol} @var{symbol} @dots{})
3567 @cindex undefined symbol in linker script
3568 Force @var{symbol} to be entered in the output file as an undefined
3569 symbol. Doing this may, for example, trigger linking of additional
3570 modules from standard libraries. You may list several @var{symbol}s for
3571 each @code{EXTERN}, and you may use @code{EXTERN} multiple times. This
3572 command has the same effect as the @samp{-u} command-line option.
3574 @item FORCE_COMMON_ALLOCATION
3575 @kindex FORCE_COMMON_ALLOCATION
3576 @cindex common allocation in linker script
3577 This command has the same effect as the @samp{-d} command-line option:
3578 to make @command{ld} assign space to common symbols even if a relocatable
3579 output file is specified (@samp{-r}).
3581 @item INHIBIT_COMMON_ALLOCATION
3582 @kindex INHIBIT_COMMON_ALLOCATION
3583 @cindex common allocation in linker script
3584 This command has the same effect as the @samp{--no-define-common}
3585 command-line option: to make @code{ld} omit the assignment of addresses
3586 to common symbols even for a non-relocatable output file.
3588 @item INSERT [ AFTER | BEFORE ] @var{output_section}
3590 @cindex insert user script into default script
3591 This command is typically used in a script specified by @samp{-T} to
3592 augment the default @code{SECTIONS} with, for example, overlays. It
3593 inserts all prior linker script statements after (or before)
3594 @var{output_section}, and also causes @samp{-T} to not override the
3595 default linker script. The exact insertion point is as for orphan
3596 sections. @xref{Location Counter}. The insertion happens after the
3597 linker has mapped input sections to output sections. Prior to the
3598 insertion, since @samp{-T} scripts are parsed before the default
3599 linker script, statements in the @samp{-T} script occur before the
3600 default linker script statements in the internal linker representation
3601 of the script. In particular, input section assignments will be made
3602 to @samp{-T} output sections before those in the default script. Here
3603 is an example of how a @samp{-T} script using @code{INSERT} might look:
3610 .ov1 @{ ov1*(.text) @}
3611 .ov2 @{ ov2*(.text) @}
3617 @item NOCROSSREFS(@var{section} @var{section} @dots{})
3618 @kindex NOCROSSREFS(@var{sections})
3619 @cindex cross references
3620 This command may be used to tell @command{ld} to issue an error about any
3621 references among certain output sections.
3623 In certain types of programs, particularly on embedded systems when
3624 using overlays, when one section is loaded into memory, another section
3625 will not be. Any direct references between the two sections would be
3626 errors. For example, it would be an error if code in one section called
3627 a function defined in the other section.
3629 The @code{NOCROSSREFS} command takes a list of output section names. If
3630 @command{ld} detects any cross references between the sections, it reports
3631 an error and returns a non-zero exit status. Note that the
3632 @code{NOCROSSREFS} command uses output section names, not input section
3635 @ifclear SingleFormat
3636 @item OUTPUT_ARCH(@var{bfdarch})
3637 @kindex OUTPUT_ARCH(@var{bfdarch})
3638 @cindex machine architecture
3639 @cindex architecture
3640 Specify a particular output machine architecture. The argument is one
3641 of the names used by the BFD library (@pxref{BFD}). You can see the
3642 architecture of an object file by using the @code{objdump} program with
3643 the @samp{-f} option.
3646 @item LD_FEATURE(@var{string})
3647 @kindex LD_FEATURE(@var{string})
3648 This command may be used to modify @command{ld} behavior. If
3649 @var{string} is @code{"SANE_EXPR"} then absolute symbols and numbers
3650 in a script are simply treated as numbers everywhere.
3651 @xref{Expression Section}.
3655 @section Assigning Values to Symbols
3656 @cindex assignment in scripts
3657 @cindex symbol definition, scripts
3658 @cindex variables, defining
3659 You may assign a value to a symbol in a linker script. This will define
3660 the symbol and place it into the symbol table with a global scope.
3663 * Simple Assignments:: Simple Assignments
3666 * PROVIDE_HIDDEN:: PROVIDE_HIDDEN
3667 * Source Code Reference:: How to use a linker script defined symbol in source code
3670 @node Simple Assignments
3671 @subsection Simple Assignments
3673 You may assign to a symbol using any of the C assignment operators:
3676 @item @var{symbol} = @var{expression} ;
3677 @itemx @var{symbol} += @var{expression} ;
3678 @itemx @var{symbol} -= @var{expression} ;
3679 @itemx @var{symbol} *= @var{expression} ;
3680 @itemx @var{symbol} /= @var{expression} ;
3681 @itemx @var{symbol} <<= @var{expression} ;
3682 @itemx @var{symbol} >>= @var{expression} ;
3683 @itemx @var{symbol} &= @var{expression} ;
3684 @itemx @var{symbol} |= @var{expression} ;
3687 The first case will define @var{symbol} to the value of
3688 @var{expression}. In the other cases, @var{symbol} must already be
3689 defined, and the value will be adjusted accordingly.
3691 The special symbol name @samp{.} indicates the location counter. You
3692 may only use this within a @code{SECTIONS} command. @xref{Location Counter}.
3694 The semicolon after @var{expression} is required.
3696 Expressions are defined below; see @ref{Expressions}.
3698 You may write symbol assignments as commands in their own right, or as
3699 statements within a @code{SECTIONS} command, or as part of an output
3700 section description in a @code{SECTIONS} command.
3702 The section of the symbol will be set from the section of the
3703 expression; for more information, see @ref{Expression Section}.
3705 Here is an example showing the three different places that symbol
3706 assignments may be used:
3717 _bdata = (. + 3) & ~ 3;
3718 .data : @{ *(.data) @}
3722 In this example, the symbol @samp{floating_point} will be defined as
3723 zero. The symbol @samp{_etext} will be defined as the address following
3724 the last @samp{.text} input section. The symbol @samp{_bdata} will be
3725 defined as the address following the @samp{.text} output section aligned
3726 upward to a 4 byte boundary.
3731 For ELF targeted ports, define a symbol that will be hidden and won't be
3732 exported. The syntax is @code{HIDDEN(@var{symbol} = @var{expression})}.
3734 Here is the example from @ref{Simple Assignments}, rewritten to use
3738 HIDDEN(floating_point = 0);
3746 HIDDEN(_bdata = (. + 3) & ~ 3);
3747 .data : @{ *(.data) @}
3751 In this case none of the three symbols will be visible outside this module.
3756 In some cases, it is desirable for a linker script to define a symbol
3757 only if it is referenced and is not defined by any object included in
3758 the link. For example, traditional linkers defined the symbol
3759 @samp{etext}. However, ANSI C requires that the user be able to use
3760 @samp{etext} as a function name without encountering an error. The
3761 @code{PROVIDE} keyword may be used to define a symbol, such as
3762 @samp{etext}, only if it is referenced but not defined. The syntax is
3763 @code{PROVIDE(@var{symbol} = @var{expression})}.
3765 Here is an example of using @code{PROVIDE} to define @samp{etext}:
3778 In this example, if the program defines @samp{_etext} (with a leading
3779 underscore), the linker will give a multiple definition error. If, on
3780 the other hand, the program defines @samp{etext} (with no leading
3781 underscore), the linker will silently use the definition in the program.
3782 If the program references @samp{etext} but does not define it, the
3783 linker will use the definition in the linker script.
3785 @node PROVIDE_HIDDEN
3786 @subsection PROVIDE_HIDDEN
3787 @cindex PROVIDE_HIDDEN
3788 Similar to @code{PROVIDE}. For ELF targeted ports, the symbol will be
3789 hidden and won't be exported.
3791 @node Source Code Reference
3792 @subsection Source Code Reference
3794 Accessing a linker script defined variable from source code is not
3795 intuitive. In particular a linker script symbol is not equivalent to
3796 a variable declaration in a high level language, it is instead a
3797 symbol that does not have a value.
3799 Before going further, it is important to note that compilers often
3800 transform names in the source code into different names when they are
3801 stored in the symbol table. For example, Fortran compilers commonly
3802 prepend or append an underscore, and C++ performs extensive @samp{name
3803 mangling}. Therefore there might be a discrepancy between the name
3804 of a variable as it is used in source code and the name of the same
3805 variable as it is defined in a linker script. For example in C a
3806 linker script variable might be referred to as:
3812 But in the linker script it might be defined as:
3818 In the remaining examples however it is assumed that no name
3819 transformation has taken place.
3821 When a symbol is declared in a high level language such as C, two
3822 things happen. The first is that the compiler reserves enough space
3823 in the program's memory to hold the @emph{value} of the symbol. The
3824 second is that the compiler creates an entry in the program's symbol
3825 table which holds the symbol's @emph{address}. ie the symbol table
3826 contains the address of the block of memory holding the symbol's
3827 value. So for example the following C declaration, at file scope:
3833 creates an entry called @samp{foo} in the symbol table. This entry
3834 holds the address of an @samp{int} sized block of memory where the
3835 number 1000 is initially stored.
3837 When a program references a symbol the compiler generates code that
3838 first accesses the symbol table to find the address of the symbol's
3839 memory block and then code to read the value from that memory block.
3846 looks up the symbol @samp{foo} in the symbol table, gets the address
3847 associated with this symbol and then writes the value 1 into that
3854 looks up the symbol @samp{foo} in the symbol table, gets its address
3855 and then copies this address into the block of memory associated with
3856 the variable @samp{a}.
3858 Linker scripts symbol declarations, by contrast, create an entry in
3859 the symbol table but do not assign any memory to them. Thus they are
3860 an address without a value. So for example the linker script definition:
3866 creates an entry in the symbol table called @samp{foo} which holds
3867 the address of memory location 1000, but nothing special is stored at
3868 address 1000. This means that you cannot access the @emph{value} of a
3869 linker script defined symbol - it has no value - all you can do is
3870 access the @emph{address} of a linker script defined symbol.
3872 Hence when you are using a linker script defined symbol in source code
3873 you should always take the address of the symbol, and never attempt to
3874 use its value. For example suppose you want to copy the contents of a
3875 section of memory called .ROM into a section called .FLASH and the
3876 linker script contains these declarations:
3880 start_of_ROM = .ROM;
3881 end_of_ROM = .ROM + sizeof (.ROM);
3882 start_of_FLASH = .FLASH;
3886 Then the C source code to perform the copy would be:
3890 extern char start_of_ROM, end_of_ROM, start_of_FLASH;
3892 memcpy (& start_of_FLASH, & start_of_ROM, & end_of_ROM - & start_of_ROM);
3896 Note the use of the @samp{&} operators. These are correct.
3897 Alternatively the symbols can be treated as the names of vectors or
3898 arrays and then the code will again work as expected:
3902 extern char start_of_ROM[], end_of_ROM[], start_of_FLASH[];
3904 memcpy (start_of_FLASH, start_of_ROM, end_of_ROM - start_of_ROM);
3908 Note how using this method does not require the use of @samp{&}
3912 @section SECTIONS Command
3914 The @code{SECTIONS} command tells the linker how to map input sections
3915 into output sections, and how to place the output sections in memory.
3917 The format of the @code{SECTIONS} command is:
3921 @var{sections-command}
3922 @var{sections-command}
3927 Each @var{sections-command} may of be one of the following:
3931 an @code{ENTRY} command (@pxref{Entry Point,,Entry command})
3933 a symbol assignment (@pxref{Assignments})
3935 an output section description
3937 an overlay description
3940 The @code{ENTRY} command and symbol assignments are permitted inside the
3941 @code{SECTIONS} command for convenience in using the location counter in
3942 those commands. This can also make the linker script easier to
3943 understand because you can use those commands at meaningful points in
3944 the layout of the output file.
3946 Output section descriptions and overlay descriptions are described
3949 If you do not use a @code{SECTIONS} command in your linker script, the
3950 linker will place each input section into an identically named output
3951 section in the order that the sections are first encountered in the
3952 input files. If all input sections are present in the first file, for
3953 example, the order of sections in the output file will match the order
3954 in the first input file. The first section will be at address zero.
3957 * Output Section Description:: Output section description
3958 * Output Section Name:: Output section name
3959 * Output Section Address:: Output section address
3960 * Input Section:: Input section description
3961 * Output Section Data:: Output section data
3962 * Output Section Keywords:: Output section keywords
3963 * Output Section Discarding:: Output section discarding
3964 * Output Section Attributes:: Output section attributes
3965 * Overlay Description:: Overlay description
3968 @node Output Section Description
3969 @subsection Output Section Description
3970 The full description of an output section looks like this:
3973 @var{section} [@var{address}] [(@var{type})] :
3975 [ALIGN(@var{section_align}) | ALIGN_WITH_INPUT]
3976 [SUBALIGN(@var{subsection_align})]
3979 @var{output-section-command}
3980 @var{output-section-command}
3982 @} [>@var{region}] [AT>@var{lma_region}] [:@var{phdr} :@var{phdr} @dots{}] [=@var{fillexp}] [,]
3986 Most output sections do not use most of the optional section attributes.
3988 The whitespace around @var{section} is required, so that the section
3989 name is unambiguous. The colon and the curly braces are also required.
3990 The comma at the end may be required if a @var{fillexp} is used and
3991 the next @var{sections-command} looks like a continuation of the expression.
3992 The line breaks and other white space are optional.
3994 Each @var{output-section-command} may be one of the following:
3998 a symbol assignment (@pxref{Assignments})
4000 an input section description (@pxref{Input Section})
4002 data values to include directly (@pxref{Output Section Data})
4004 a special output section keyword (@pxref{Output Section Keywords})
4007 @node Output Section Name
4008 @subsection Output Section Name
4009 @cindex name, section
4010 @cindex section name
4011 The name of the output section is @var{section}. @var{section} must
4012 meet the constraints of your output format. In formats which only
4013 support a limited number of sections, such as @code{a.out}, the name
4014 must be one of the names supported by the format (@code{a.out}, for
4015 example, allows only @samp{.text}, @samp{.data} or @samp{.bss}). If the
4016 output format supports any number of sections, but with numbers and not
4017 names (as is the case for Oasys), the name should be supplied as a
4018 quoted numeric string. A section name may consist of any sequence of
4019 characters, but a name which contains any unusual characters such as
4020 commas must be quoted.
4022 The output section name @samp{/DISCARD/} is special; @ref{Output Section
4025 @node Output Section Address
4026 @subsection Output Section Address
4027 @cindex address, section
4028 @cindex section address
4029 The @var{address} is an expression for the VMA (the virtual memory
4030 address) of the output section. This address is optional, but if it
4031 is provided then the output address will be set exactly as specified.
4033 If the output address is not specified then one will be chosen for the
4034 section, based on the heuristic below. This address will be adjusted
4035 to fit the alignment requirement of the output section. The
4036 alignment requirement is the strictest alignment of any input section
4037 contained within the output section.
4039 The output section address heuristic is as follows:
4043 If an output memory @var{region} is set for the section then it
4044 is added to this region and its address will be the next free address
4048 If the MEMORY command has been used to create a list of memory
4049 regions then the first region which has attributes compatible with the
4050 section is selected to contain it. The section's output address will
4051 be the next free address in that region; @ref{MEMORY}.
4054 If no memory regions were specified, or none match the section then
4055 the output address will be based on the current value of the location
4063 .text . : @{ *(.text) @}
4070 .text : @{ *(.text) @}
4074 are subtly different. The first will set the address of the
4075 @samp{.text} output section to the current value of the location
4076 counter. The second will set it to the current value of the location
4077 counter aligned to the strictest alignment of any of the @samp{.text}
4080 The @var{address} may be an arbitrary expression; @ref{Expressions}.
4081 For example, if you want to align the section on a 0x10 byte boundary,
4082 so that the lowest four bits of the section address are zero, you could
4083 do something like this:
4085 .text ALIGN(0x10) : @{ *(.text) @}
4088 This works because @code{ALIGN} returns the current location counter
4089 aligned upward to the specified value.
4091 Specifying @var{address} for a section will change the value of the
4092 location counter, provided that the section is non-empty. (Empty
4093 sections are ignored).
4096 @subsection Input Section Description
4097 @cindex input sections
4098 @cindex mapping input sections to output sections
4099 The most common output section command is an input section description.
4101 The input section description is the most basic linker script operation.
4102 You use output sections to tell the linker how to lay out your program
4103 in memory. You use input section descriptions to tell the linker how to
4104 map the input files into your memory layout.
4107 * Input Section Basics:: Input section basics
4108 * Input Section Wildcards:: Input section wildcard patterns
4109 * Input Section Common:: Input section for common symbols
4110 * Input Section Keep:: Input section and garbage collection
4111 * Input Section Example:: Input section example
4114 @node Input Section Basics
4115 @subsubsection Input Section Basics
4116 @cindex input section basics
4117 An input section description consists of a file name optionally followed
4118 by a list of section names in parentheses.
4120 The file name and the section name may be wildcard patterns, which we
4121 describe further below (@pxref{Input Section Wildcards}).
4123 The most common input section description is to include all input
4124 sections with a particular name in the output section. For example, to
4125 include all input @samp{.text} sections, you would write:
4130 Here the @samp{*} is a wildcard which matches any file name. To exclude a list
4131 of files from matching the file name wildcard, EXCLUDE_FILE may be used to
4132 match all files except the ones specified in the EXCLUDE_FILE list. For
4135 *(EXCLUDE_FILE (*crtend.o *otherfile.o) .ctors)
4137 will cause all .ctors sections from all files except @file{crtend.o} and
4138 @file{otherfile.o} to be included.
4140 There are two ways to include more than one section:
4146 The difference between these is the order in which the @samp{.text} and
4147 @samp{.rdata} input sections will appear in the output section. In the
4148 first example, they will be intermingled, appearing in the same order as
4149 they are found in the linker input. In the second example, all
4150 @samp{.text} input sections will appear first, followed by all
4151 @samp{.rdata} input sections.
4153 You can specify a file name to include sections from a particular file.
4154 You would do this if one or more of your files contain special data that
4155 needs to be at a particular location in memory. For example:
4160 To refine the sections that are included based on the section flags
4161 of an input section, INPUT_SECTION_FLAGS may be used.
4163 Here is a simple example for using Section header flags for ELF sections:
4168 .text : @{ INPUT_SECTION_FLAGS (SHF_MERGE & SHF_STRINGS) *(.text) @}
4169 .text2 : @{ INPUT_SECTION_FLAGS (!SHF_WRITE) *(.text) @}
4174 In this example, the output section @samp{.text} will be comprised of any
4175 input section matching the name *(.text) whose section header flags
4176 @code{SHF_MERGE} and @code{SHF_STRINGS} are set. The output section
4177 @samp{.text2} will be comprised of any input section matching the name *(.text)
4178 whose section header flag @code{SHF_WRITE} is clear.
4180 You can also specify files within archives by writing a pattern
4181 matching the archive, a colon, then the pattern matching the file,
4182 with no whitespace around the colon.
4186 matches file within archive
4188 matches the whole archive
4190 matches file but not one in an archive
4193 Either one or both of @samp{archive} and @samp{file} can contain shell
4194 wildcards. On DOS based file systems, the linker will assume that a
4195 single letter followed by a colon is a drive specifier, so
4196 @samp{c:myfile.o} is a simple file specification, not @samp{myfile.o}
4197 within an archive called @samp{c}. @samp{archive:file} filespecs may
4198 also be used within an @code{EXCLUDE_FILE} list, but may not appear in
4199 other linker script contexts. For instance, you cannot extract a file
4200 from an archive by using @samp{archive:file} in an @code{INPUT}
4203 If you use a file name without a list of sections, then all sections in
4204 the input file will be included in the output section. This is not
4205 commonly done, but it may by useful on occasion. For example:
4210 When you use a file name which is not an @samp{archive:file} specifier
4211 and does not contain any wild card
4212 characters, the linker will first see if you also specified the file
4213 name on the linker command line or in an @code{INPUT} command. If you
4214 did not, the linker will attempt to open the file as an input file, as
4215 though it appeared on the command line. Note that this differs from an
4216 @code{INPUT} command, because the linker will not search for the file in
4217 the archive search path.
4219 @node Input Section Wildcards
4220 @subsubsection Input Section Wildcard Patterns
4221 @cindex input section wildcards
4222 @cindex wildcard file name patterns
4223 @cindex file name wildcard patterns
4224 @cindex section name wildcard patterns
4225 In an input section description, either the file name or the section
4226 name or both may be wildcard patterns.
4228 The file name of @samp{*} seen in many examples is a simple wildcard
4229 pattern for the file name.
4231 The wildcard patterns are like those used by the Unix shell.
4235 matches any number of characters
4237 matches any single character
4239 matches a single instance of any of the @var{chars}; the @samp{-}
4240 character may be used to specify a range of characters, as in
4241 @samp{[a-z]} to match any lower case letter
4243 quotes the following character
4246 When a file name is matched with a wildcard, the wildcard characters
4247 will not match a @samp{/} character (used to separate directory names on
4248 Unix). A pattern consisting of a single @samp{*} character is an
4249 exception; it will always match any file name, whether it contains a
4250 @samp{/} or not. In a section name, the wildcard characters will match
4251 a @samp{/} character.
4253 File name wildcard patterns only match files which are explicitly
4254 specified on the command line or in an @code{INPUT} command. The linker
4255 does not search directories to expand wildcards.
4257 If a file name matches more than one wildcard pattern, or if a file name
4258 appears explicitly and is also matched by a wildcard pattern, the linker
4259 will use the first match in the linker script. For example, this
4260 sequence of input section descriptions is probably in error, because the
4261 @file{data.o} rule will not be used:
4263 .data : @{ *(.data) @}
4264 .data1 : @{ data.o(.data) @}
4267 @cindex SORT_BY_NAME
4268 Normally, the linker will place files and sections matched by wildcards
4269 in the order in which they are seen during the link. You can change
4270 this by using the @code{SORT_BY_NAME} keyword, which appears before a wildcard
4271 pattern in parentheses (e.g., @code{SORT_BY_NAME(.text*)}). When the
4272 @code{SORT_BY_NAME} keyword is used, the linker will sort the files or sections
4273 into ascending order by name before placing them in the output file.
4275 @cindex SORT_BY_ALIGNMENT
4276 @code{SORT_BY_ALIGNMENT} is very similar to @code{SORT_BY_NAME}. The
4277 difference is @code{SORT_BY_ALIGNMENT} will sort sections into
4278 descending order by alignment before placing them in the output file.
4279 Larger alignments are placed before smaller alignments in order to
4280 reduce the amount of padding necessary.
4282 @cindex SORT_BY_INIT_PRIORITY
4283 @code{SORT_BY_INIT_PRIORITY} is very similar to @code{SORT_BY_NAME}. The
4284 difference is @code{SORT_BY_INIT_PRIORITY} will sort sections into
4285 ascending order by numerical value of the GCC init_priority attribute
4286 encoded in the section name before placing them in the output file.
4289 @code{SORT} is an alias for @code{SORT_BY_NAME}.
4291 When there are nested section sorting commands in linker script, there
4292 can be at most 1 level of nesting for section sorting commands.
4296 @code{SORT_BY_NAME} (@code{SORT_BY_ALIGNMENT} (wildcard section pattern)).
4297 It will sort the input sections by name first, then by alignment if two
4298 sections have the same name.
4300 @code{SORT_BY_ALIGNMENT} (@code{SORT_BY_NAME} (wildcard section pattern)).
4301 It will sort the input sections by alignment first, then by name if two
4302 sections have the same alignment.
4304 @code{SORT_BY_NAME} (@code{SORT_BY_NAME} (wildcard section pattern)) is
4305 treated the same as @code{SORT_BY_NAME} (wildcard section pattern).
4307 @code{SORT_BY_ALIGNMENT} (@code{SORT_BY_ALIGNMENT} (wildcard section pattern))
4308 is treated the same as @code{SORT_BY_ALIGNMENT} (wildcard section pattern).
4310 All other nested section sorting commands are invalid.
4313 When both command line section sorting option and linker script
4314 section sorting command are used, section sorting command always
4315 takes precedence over the command line option.
4317 If the section sorting command in linker script isn't nested, the
4318 command line option will make the section sorting command to be
4319 treated as nested sorting command.
4323 @code{SORT_BY_NAME} (wildcard section pattern ) with
4324 @option{--sort-sections alignment} is equivalent to
4325 @code{SORT_BY_NAME} (@code{SORT_BY_ALIGNMENT} (wildcard section pattern)).
4327 @code{SORT_BY_ALIGNMENT} (wildcard section pattern) with
4328 @option{--sort-section name} is equivalent to
4329 @code{SORT_BY_ALIGNMENT} (@code{SORT_BY_NAME} (wildcard section pattern)).
4332 If the section sorting command in linker script is nested, the
4333 command line option will be ignored.
4336 @code{SORT_NONE} disables section sorting by ignoring the command line
4337 section sorting option.
4339 If you ever get confused about where input sections are going, use the
4340 @samp{-M} linker option to generate a map file. The map file shows
4341 precisely how input sections are mapped to output sections.
4343 This example shows how wildcard patterns might be used to partition
4344 files. This linker script directs the linker to place all @samp{.text}
4345 sections in @samp{.text} and all @samp{.bss} sections in @samp{.bss}.
4346 The linker will place the @samp{.data} section from all files beginning
4347 with an upper case character in @samp{.DATA}; for all other files, the
4348 linker will place the @samp{.data} section in @samp{.data}.
4352 .text : @{ *(.text) @}
4353 .DATA : @{ [A-Z]*(.data) @}
4354 .data : @{ *(.data) @}
4355 .bss : @{ *(.bss) @}
4360 @node Input Section Common
4361 @subsubsection Input Section for Common Symbols
4362 @cindex common symbol placement
4363 @cindex uninitialized data placement
4364 A special notation is needed for common symbols, because in many object
4365 file formats common symbols do not have a particular input section. The
4366 linker treats common symbols as though they are in an input section
4367 named @samp{COMMON}.
4369 You may use file names with the @samp{COMMON} section just as with any
4370 other input sections. You can use this to place common symbols from a
4371 particular input file in one section while common symbols from other
4372 input files are placed in another section.
4374 In most cases, common symbols in input files will be placed in the
4375 @samp{.bss} section in the output file. For example:
4377 .bss @{ *(.bss) *(COMMON) @}
4380 @cindex scommon section
4381 @cindex small common symbols
4382 Some object file formats have more than one type of common symbol. For
4383 example, the MIPS ELF object file format distinguishes standard common
4384 symbols and small common symbols. In this case, the linker will use a
4385 different special section name for other types of common symbols. In
4386 the case of MIPS ELF, the linker uses @samp{COMMON} for standard common
4387 symbols and @samp{.scommon} for small common symbols. This permits you
4388 to map the different types of common symbols into memory at different
4392 You will sometimes see @samp{[COMMON]} in old linker scripts. This
4393 notation is now considered obsolete. It is equivalent to
4396 @node Input Section Keep
4397 @subsubsection Input Section and Garbage Collection
4399 @cindex garbage collection
4400 When link-time garbage collection is in use (@samp{--gc-sections}),
4401 it is often useful to mark sections that should not be eliminated.
4402 This is accomplished by surrounding an input section's wildcard entry
4403 with @code{KEEP()}, as in @code{KEEP(*(.init))} or
4404 @code{KEEP(SORT_BY_NAME(*)(.ctors))}.
4406 @node Input Section Example
4407 @subsubsection Input Section Example
4408 The following example is a complete linker script. It tells the linker
4409 to read all of the sections from file @file{all.o} and place them at the
4410 start of output section @samp{outputa} which starts at location
4411 @samp{0x10000}. All of section @samp{.input1} from file @file{foo.o}
4412 follows immediately, in the same output section. All of section
4413 @samp{.input2} from @file{foo.o} goes into output section
4414 @samp{outputb}, followed by section @samp{.input1} from @file{foo1.o}.
4415 All of the remaining @samp{.input1} and @samp{.input2} sections from any
4416 files are written to output section @samp{outputc}.
4444 @node Output Section Data
4445 @subsection Output Section Data
4447 @cindex section data
4448 @cindex output section data
4449 @kindex BYTE(@var{expression})
4450 @kindex SHORT(@var{expression})
4451 @kindex LONG(@var{expression})
4452 @kindex QUAD(@var{expression})
4453 @kindex SQUAD(@var{expression})
4454 You can include explicit bytes of data in an output section by using
4455 @code{BYTE}, @code{SHORT}, @code{LONG}, @code{QUAD}, or @code{SQUAD} as
4456 an output section command. Each keyword is followed by an expression in
4457 parentheses providing the value to store (@pxref{Expressions}). The
4458 value of the expression is stored at the current value of the location
4461 The @code{BYTE}, @code{SHORT}, @code{LONG}, and @code{QUAD} commands
4462 store one, two, four, and eight bytes (respectively). After storing the
4463 bytes, the location counter is incremented by the number of bytes
4466 For example, this will store the byte 1 followed by the four byte value
4467 of the symbol @samp{addr}:
4473 When using a 64 bit host or target, @code{QUAD} and @code{SQUAD} are the
4474 same; they both store an 8 byte, or 64 bit, value. When both host and
4475 target are 32 bits, an expression is computed as 32 bits. In this case
4476 @code{QUAD} stores a 32 bit value zero extended to 64 bits, and
4477 @code{SQUAD} stores a 32 bit value sign extended to 64 bits.
4479 If the object file format of the output file has an explicit endianness,
4480 which is the normal case, the value will be stored in that endianness.
4481 When the object file format does not have an explicit endianness, as is
4482 true of, for example, S-records, the value will be stored in the
4483 endianness of the first input object file.
4485 Note---these commands only work inside a section description and not
4486 between them, so the following will produce an error from the linker:
4488 SECTIONS @{@ .text : @{@ *(.text) @}@ LONG(1) .data : @{@ *(.data) @}@ @}@
4490 whereas this will work:
4492 SECTIONS @{@ .text : @{@ *(.text) ; LONG(1) @}@ .data : @{@ *(.data) @}@ @}@
4495 @kindex FILL(@var{expression})
4496 @cindex holes, filling
4497 @cindex unspecified memory
4498 You may use the @code{FILL} command to set the fill pattern for the
4499 current section. It is followed by an expression in parentheses. Any
4500 otherwise unspecified regions of memory within the section (for example,
4501 gaps left due to the required alignment of input sections) are filled
4502 with the value of the expression, repeated as
4503 necessary. A @code{FILL} statement covers memory locations after the
4504 point at which it occurs in the section definition; by including more
4505 than one @code{FILL} statement, you can have different fill patterns in
4506 different parts of an output section.
4508 This example shows how to fill unspecified regions of memory with the
4514 The @code{FILL} command is similar to the @samp{=@var{fillexp}} output
4515 section attribute, but it only affects the
4516 part of the section following the @code{FILL} command, rather than the
4517 entire section. If both are used, the @code{FILL} command takes
4518 precedence. @xref{Output Section Fill}, for details on the fill
4521 @node Output Section Keywords
4522 @subsection Output Section Keywords
4523 There are a couple of keywords which can appear as output section
4527 @kindex CREATE_OBJECT_SYMBOLS
4528 @cindex input filename symbols
4529 @cindex filename symbols
4530 @item CREATE_OBJECT_SYMBOLS
4531 The command tells the linker to create a symbol for each input file.
4532 The name of each symbol will be the name of the corresponding input
4533 file. The section of each symbol will be the output section in which
4534 the @code{CREATE_OBJECT_SYMBOLS} command appears.
4536 This is conventional for the a.out object file format. It is not
4537 normally used for any other object file format.
4539 @kindex CONSTRUCTORS
4540 @cindex C++ constructors, arranging in link
4541 @cindex constructors, arranging in link
4543 When linking using the a.out object file format, the linker uses an
4544 unusual set construct to support C++ global constructors and
4545 destructors. When linking object file formats which do not support
4546 arbitrary sections, such as ECOFF and XCOFF, the linker will
4547 automatically recognize C++ global constructors and destructors by name.
4548 For these object file formats, the @code{CONSTRUCTORS} command tells the
4549 linker to place constructor information in the output section where the
4550 @code{CONSTRUCTORS} command appears. The @code{CONSTRUCTORS} command is
4551 ignored for other object file formats.
4553 The symbol @w{@code{__CTOR_LIST__}} marks the start of the global
4554 constructors, and the symbol @w{@code{__CTOR_END__}} marks the end.
4555 Similarly, @w{@code{__DTOR_LIST__}} and @w{@code{__DTOR_END__}} mark
4556 the start and end of the global destructors. The
4557 first word in the list is the number of entries, followed by the address
4558 of each constructor or destructor, followed by a zero word. The
4559 compiler must arrange to actually run the code. For these object file
4560 formats @sc{gnu} C++ normally calls constructors from a subroutine
4561 @code{__main}; a call to @code{__main} is automatically inserted into
4562 the startup code for @code{main}. @sc{gnu} C++ normally runs
4563 destructors either by using @code{atexit}, or directly from the function
4566 For object file formats such as @code{COFF} or @code{ELF} which support
4567 arbitrary section names, @sc{gnu} C++ will normally arrange to put the
4568 addresses of global constructors and destructors into the @code{.ctors}
4569 and @code{.dtors} sections. Placing the following sequence into your
4570 linker script will build the sort of table which the @sc{gnu} C++
4571 runtime code expects to see.
4575 LONG((__CTOR_END__ - __CTOR_LIST__) / 4 - 2)
4580 LONG((__DTOR_END__ - __DTOR_LIST__) / 4 - 2)
4586 If you are using the @sc{gnu} C++ support for initialization priority,
4587 which provides some control over the order in which global constructors
4588 are run, you must sort the constructors at link time to ensure that they
4589 are executed in the correct order. When using the @code{CONSTRUCTORS}
4590 command, use @samp{SORT_BY_NAME(CONSTRUCTORS)} instead. When using the
4591 @code{.ctors} and @code{.dtors} sections, use @samp{*(SORT_BY_NAME(.ctors))} and
4592 @samp{*(SORT_BY_NAME(.dtors))} instead of just @samp{*(.ctors)} and
4595 Normally the compiler and linker will handle these issues automatically,
4596 and you will not need to concern yourself with them. However, you may
4597 need to consider this if you are using C++ and writing your own linker
4602 @node Output Section Discarding
4603 @subsection Output Section Discarding
4604 @cindex discarding sections
4605 @cindex sections, discarding
4606 @cindex removing sections
4607 The linker will not normally create output sections with no contents.
4608 This is for convenience when referring to input sections that may or
4609 may not be present in any of the input files. For example:
4611 .foo : @{ *(.foo) @}
4614 will only create a @samp{.foo} section in the output file if there is a
4615 @samp{.foo} section in at least one input file, and if the input
4616 sections are not all empty. Other link script directives that allocate
4617 space in an output section will also create the output section. So
4618 too will assignments to dot even if the assignment does not create
4619 space, except for @samp{. = 0}, @samp{. = . + 0}, @samp{. = sym},
4620 @samp{. = . + sym} and @samp{. = ALIGN (. != 0, expr, 1)} when
4621 @samp{sym} is an absolute symbol of value 0 defined in the script.
4622 This allows you to force output of an empty section with @samp{. = .}.
4624 The linker will ignore address assignments (@pxref{Output Section Address})
4625 on discarded output sections, except when the linker script defines
4626 symbols in the output section. In that case the linker will obey
4627 the address assignments, possibly advancing dot even though the
4628 section is discarded.
4631 The special output section name @samp{/DISCARD/} may be used to discard
4632 input sections. Any input sections which are assigned to an output
4633 section named @samp{/DISCARD/} are not included in the output file.
4635 @node Output Section Attributes
4636 @subsection Output Section Attributes
4637 @cindex output section attributes
4638 We showed above that the full description of an output section looked
4643 @var{section} [@var{address}] [(@var{type})] :
4645 [ALIGN(@var{section_align})]
4646 [SUBALIGN(@var{subsection_align})]
4649 @var{output-section-command}
4650 @var{output-section-command}
4652 @} [>@var{region}] [AT>@var{lma_region}] [:@var{phdr} :@var{phdr} @dots{}] [=@var{fillexp}]
4656 We've already described @var{section}, @var{address}, and
4657 @var{output-section-command}. In this section we will describe the
4658 remaining section attributes.
4661 * Output Section Type:: Output section type
4662 * Output Section LMA:: Output section LMA
4663 * Forced Output Alignment:: Forced Output Alignment
4664 * Forced Input Alignment:: Forced Input Alignment
4665 * Output Section Constraint:: Output section constraint
4666 * Output Section Region:: Output section region
4667 * Output Section Phdr:: Output section phdr
4668 * Output Section Fill:: Output section fill
4671 @node Output Section Type
4672 @subsubsection Output Section Type
4673 Each output section may have a type. The type is a keyword in
4674 parentheses. The following types are defined:
4678 The section should be marked as not loadable, so that it will not be
4679 loaded into memory when the program is run.
4684 These type names are supported for backward compatibility, and are
4685 rarely used. They all have the same effect: the section should be
4686 marked as not allocatable, so that no memory is allocated for the
4687 section when the program is run.
4691 @cindex prevent unnecessary loading
4692 @cindex loading, preventing
4693 The linker normally sets the attributes of an output section based on
4694 the input sections which map into it. You can override this by using
4695 the section type. For example, in the script sample below, the
4696 @samp{ROM} section is addressed at memory location @samp{0} and does not
4697 need to be loaded when the program is run.
4701 ROM 0 (NOLOAD) : @{ @dots{} @}
4707 @node Output Section LMA
4708 @subsubsection Output Section LMA
4709 @kindex AT>@var{lma_region}
4710 @kindex AT(@var{lma})
4711 @cindex load address
4712 @cindex section load address
4713 Every section has a virtual address (VMA) and a load address (LMA); see
4714 @ref{Basic Script Concepts}. The virtual address is specified by the
4715 @pxref{Output Section Address} described earlier. The load address is
4716 specified by the @code{AT} or @code{AT>} keywords. Specifying a load
4717 address is optional.
4719 The @code{AT} keyword takes an expression as an argument. This
4720 specifies the exact load address of the section. The @code{AT>} keyword
4721 takes the name of a memory region as an argument. @xref{MEMORY}. The
4722 load address of the section is set to the next free address in the
4723 region, aligned to the section's alignment requirements.
4725 If neither @code{AT} nor @code{AT>} is specified for an allocatable
4726 section, the linker will use the following heuristic to determine the
4731 If the section has a specific VMA address, then this is used as
4732 the LMA address as well.
4735 If the section is not allocatable then its LMA is set to its VMA.
4738 Otherwise if a memory region can be found that is compatible
4739 with the current section, and this region contains at least one
4740 section, then the LMA is set so the difference between the
4741 VMA and LMA is the same as the difference between the VMA and LMA of
4742 the last section in the located region.
4745 If no memory regions have been declared then a default region
4746 that covers the entire address space is used in the previous step.
4749 If no suitable region could be found, or there was no previous
4750 section then the LMA is set equal to the VMA.
4753 @cindex ROM initialized data
4754 @cindex initialized data in ROM
4755 This feature is designed to make it easy to build a ROM image. For
4756 example, the following linker script creates three output sections: one
4757 called @samp{.text}, which starts at @code{0x1000}, one called
4758 @samp{.mdata}, which is loaded at the end of the @samp{.text} section
4759 even though its VMA is @code{0x2000}, and one called @samp{.bss} to hold
4760 uninitialized data at address @code{0x3000}. The symbol @code{_data} is
4761 defined with the value @code{0x2000}, which shows that the location
4762 counter holds the VMA value, not the LMA value.
4768 .text 0x1000 : @{ *(.text) _etext = . ; @}
4770 AT ( ADDR (.text) + SIZEOF (.text) )
4771 @{ _data = . ; *(.data); _edata = . ; @}
4773 @{ _bstart = . ; *(.bss) *(COMMON) ; _bend = . ;@}
4778 The run-time initialization code for use with a program generated with
4779 this linker script would include something like the following, to copy
4780 the initialized data from the ROM image to its runtime address. Notice
4781 how this code takes advantage of the symbols defined by the linker
4786 extern char _etext, _data, _edata, _bstart, _bend;
4787 char *src = &_etext;
4790 /* ROM has data at end of text; copy it. */
4791 while (dst < &_edata)
4795 for (dst = &_bstart; dst< &_bend; dst++)
4800 @node Forced Output Alignment
4801 @subsubsection Forced Output Alignment
4802 @kindex ALIGN(@var{section_align})
4803 @cindex forcing output section alignment
4804 @cindex output section alignment
4805 You can increase an output section's alignment by using ALIGN. As an
4806 alternative you can enforce that the difference between the VMA and LMA remains
4807 intact throughout this output section with the ALIGN_WITH_INPUT attribute.
4809 @node Forced Input Alignment
4810 @subsubsection Forced Input Alignment
4811 @kindex SUBALIGN(@var{subsection_align})
4812 @cindex forcing input section alignment
4813 @cindex input section alignment
4814 You can force input section alignment within an output section by using
4815 SUBALIGN. The value specified overrides any alignment given by input
4816 sections, whether larger or smaller.
4818 @node Output Section Constraint
4819 @subsubsection Output Section Constraint
4822 @cindex constraints on output sections
4823 You can specify that an output section should only be created if all
4824 of its input sections are read-only or all of its input sections are
4825 read-write by using the keyword @code{ONLY_IF_RO} and
4826 @code{ONLY_IF_RW} respectively.
4828 @node Output Section Region
4829 @subsubsection Output Section Region
4830 @kindex >@var{region}
4831 @cindex section, assigning to memory region
4832 @cindex memory regions and sections
4833 You can assign a section to a previously defined region of memory by
4834 using @samp{>@var{region}}. @xref{MEMORY}.
4836 Here is a simple example:
4839 MEMORY @{ rom : ORIGIN = 0x1000, LENGTH = 0x1000 @}
4840 SECTIONS @{ ROM : @{ *(.text) @} >rom @}
4844 @node Output Section Phdr
4845 @subsubsection Output Section Phdr
4847 @cindex section, assigning to program header
4848 @cindex program headers and sections
4849 You can assign a section to a previously defined program segment by
4850 using @samp{:@var{phdr}}. @xref{PHDRS}. If a section is assigned to
4851 one or more segments, then all subsequent allocated sections will be
4852 assigned to those segments as well, unless they use an explicitly
4853 @code{:@var{phdr}} modifier. You can use @code{:NONE} to tell the
4854 linker to not put the section in any segment at all.
4856 Here is a simple example:
4859 PHDRS @{ text PT_LOAD ; @}
4860 SECTIONS @{ .text : @{ *(.text) @} :text @}
4864 @node Output Section Fill
4865 @subsubsection Output Section Fill
4866 @kindex =@var{fillexp}
4867 @cindex section fill pattern
4868 @cindex fill pattern, entire section
4869 You can set the fill pattern for an entire section by using
4870 @samp{=@var{fillexp}}. @var{fillexp} is an expression
4871 (@pxref{Expressions}). Any otherwise unspecified regions of memory
4872 within the output section (for example, gaps left due to the required
4873 alignment of input sections) will be filled with the value, repeated as
4874 necessary. If the fill expression is a simple hex number, ie. a string
4875 of hex digit starting with @samp{0x} and without a trailing @samp{k} or @samp{M}, then
4876 an arbitrarily long sequence of hex digits can be used to specify the
4877 fill pattern; Leading zeros become part of the pattern too. For all
4878 other cases, including extra parentheses or a unary @code{+}, the fill
4879 pattern is the four least significant bytes of the value of the
4880 expression. In all cases, the number is big-endian.
4882 You can also change the fill value with a @code{FILL} command in the
4883 output section commands; (@pxref{Output Section Data}).
4885 Here is a simple example:
4888 SECTIONS @{ .text : @{ *(.text) @} =0x90909090 @}
4892 @node Overlay Description
4893 @subsection Overlay Description
4896 An overlay description provides an easy way to describe sections which
4897 are to be loaded as part of a single memory image but are to be run at
4898 the same memory address. At run time, some sort of overlay manager will
4899 copy the overlaid sections in and out of the runtime memory address as
4900 required, perhaps by simply manipulating addressing bits. This approach
4901 can be useful, for example, when a certain region of memory is faster
4904 Overlays are described using the @code{OVERLAY} command. The
4905 @code{OVERLAY} command is used within a @code{SECTIONS} command, like an
4906 output section description. The full syntax of the @code{OVERLAY}
4907 command is as follows:
4910 OVERLAY [@var{start}] : [NOCROSSREFS] [AT ( @var{ldaddr} )]
4914 @var{output-section-command}
4915 @var{output-section-command}
4917 @} [:@var{phdr}@dots{}] [=@var{fill}]
4920 @var{output-section-command}
4921 @var{output-section-command}
4923 @} [:@var{phdr}@dots{}] [=@var{fill}]
4925 @} [>@var{region}] [:@var{phdr}@dots{}] [=@var{fill}] [,]
4929 Everything is optional except @code{OVERLAY} (a keyword), and each
4930 section must have a name (@var{secname1} and @var{secname2} above). The
4931 section definitions within the @code{OVERLAY} construct are identical to
4932 those within the general @code{SECTIONS} construct (@pxref{SECTIONS}),
4933 except that no addresses and no memory regions may be defined for
4934 sections within an @code{OVERLAY}.
4936 The comma at the end may be required if a @var{fill} is used and
4937 the next @var{sections-command} looks like a continuation of the expression.
4939 The sections are all defined with the same starting address. The load
4940 addresses of the sections are arranged such that they are consecutive in
4941 memory starting at the load address used for the @code{OVERLAY} as a
4942 whole (as with normal section definitions, the load address is optional,
4943 and defaults to the start address; the start address is also optional,
4944 and defaults to the current value of the location counter).
4946 If the @code{NOCROSSREFS} keyword is used, and there are any
4947 references among the sections, the linker will report an error. Since
4948 the sections all run at the same address, it normally does not make
4949 sense for one section to refer directly to another.
4950 @xref{Miscellaneous Commands, NOCROSSREFS}.
4952 For each section within the @code{OVERLAY}, the linker automatically
4953 provides two symbols. The symbol @code{__load_start_@var{secname}} is
4954 defined as the starting load address of the section. The symbol
4955 @code{__load_stop_@var{secname}} is defined as the final load address of
4956 the section. Any characters within @var{secname} which are not legal
4957 within C identifiers are removed. C (or assembler) code may use these
4958 symbols to move the overlaid sections around as necessary.
4960 At the end of the overlay, the value of the location counter is set to
4961 the start address of the overlay plus the size of the largest section.
4963 Here is an example. Remember that this would appear inside a
4964 @code{SECTIONS} construct.
4967 OVERLAY 0x1000 : AT (0x4000)
4969 .text0 @{ o1/*.o(.text) @}
4970 .text1 @{ o2/*.o(.text) @}
4975 This will define both @samp{.text0} and @samp{.text1} to start at
4976 address 0x1000. @samp{.text0} will be loaded at address 0x4000, and
4977 @samp{.text1} will be loaded immediately after @samp{.text0}. The
4978 following symbols will be defined if referenced: @code{__load_start_text0},
4979 @code{__load_stop_text0}, @code{__load_start_text1},
4980 @code{__load_stop_text1}.
4982 C code to copy overlay @code{.text1} into the overlay area might look
4987 extern char __load_start_text1, __load_stop_text1;
4988 memcpy ((char *) 0x1000, &__load_start_text1,
4989 &__load_stop_text1 - &__load_start_text1);
4993 Note that the @code{OVERLAY} command is just syntactic sugar, since
4994 everything it does can be done using the more basic commands. The above
4995 example could have been written identically as follows.
4999 .text0 0x1000 : AT (0x4000) @{ o1/*.o(.text) @}
5000 PROVIDE (__load_start_text0 = LOADADDR (.text0));
5001 PROVIDE (__load_stop_text0 = LOADADDR (.text0) + SIZEOF (.text0));
5002 .text1 0x1000 : AT (0x4000 + SIZEOF (.text0)) @{ o2/*.o(.text) @}
5003 PROVIDE (__load_start_text1 = LOADADDR (.text1));
5004 PROVIDE (__load_stop_text1 = LOADADDR (.text1) + SIZEOF (.text1));
5005 . = 0x1000 + MAX (SIZEOF (.text0), SIZEOF (.text1));
5010 @section MEMORY Command
5012 @cindex memory regions
5013 @cindex regions of memory
5014 @cindex allocating memory
5015 @cindex discontinuous memory
5016 The linker's default configuration permits allocation of all available
5017 memory. You can override this by using the @code{MEMORY} command.
5019 The @code{MEMORY} command describes the location and size of blocks of
5020 memory in the target. You can use it to describe which memory regions
5021 may be used by the linker, and which memory regions it must avoid. You
5022 can then assign sections to particular memory regions. The linker will
5023 set section addresses based on the memory regions, and will warn about
5024 regions that become too full. The linker will not shuffle sections
5025 around to fit into the available regions.
5027 A linker script may contain many uses of the @code{MEMORY} command,
5028 however, all memory blocks defined are treated as if they were
5029 specified inside a single @code{MEMORY} command. The syntax for
5035 @var{name} [(@var{attr})] : ORIGIN = @var{origin}, LENGTH = @var{len}
5041 The @var{name} is a name used in the linker script to refer to the
5042 region. The region name has no meaning outside of the linker script.
5043 Region names are stored in a separate name space, and will not conflict
5044 with symbol names, file names, or section names. Each memory region
5045 must have a distinct name within the @code{MEMORY} command. However you can
5046 add later alias names to existing memory regions with the @ref{REGION_ALIAS}
5049 @cindex memory region attributes
5050 The @var{attr} string is an optional list of attributes that specify
5051 whether to use a particular memory region for an input section which is
5052 not explicitly mapped in the linker script. As described in
5053 @ref{SECTIONS}, if you do not specify an output section for some input
5054 section, the linker will create an output section with the same name as
5055 the input section. If you define region attributes, the linker will use
5056 them to select the memory region for the output section that it creates.
5058 The @var{attr} string must consist only of the following characters:
5073 Invert the sense of any of the attributes that follow
5076 If a unmapped section matches any of the listed attributes other than
5077 @samp{!}, it will be placed in the memory region. The @samp{!}
5078 attribute reverses this test, so that an unmapped section will be placed
5079 in the memory region only if it does not match any of the listed
5085 The @var{origin} is an numerical expression for the start address of
5086 the memory region. The expression must evaluate to a constant and it
5087 cannot involve any symbols. The keyword @code{ORIGIN} may be
5088 abbreviated to @code{org} or @code{o} (but not, for example,
5094 The @var{len} is an expression for the size in bytes of the memory
5095 region. As with the @var{origin} expression, the expression must
5096 be numerical only and must evaluate to a constant. The keyword
5097 @code{LENGTH} may be abbreviated to @code{len} or @code{l}.
5099 In the following example, we specify that there are two memory regions
5100 available for allocation: one starting at @samp{0} for 256 kilobytes,
5101 and the other starting at @samp{0x40000000} for four megabytes. The
5102 linker will place into the @samp{rom} memory region every section which
5103 is not explicitly mapped into a memory region, and is either read-only
5104 or executable. The linker will place other sections which are not
5105 explicitly mapped into a memory region into the @samp{ram} memory
5112 rom (rx) : ORIGIN = 0, LENGTH = 256K
5113 ram (!rx) : org = 0x40000000, l = 4M
5118 Once you define a memory region, you can direct the linker to place
5119 specific output sections into that memory region by using the
5120 @samp{>@var{region}} output section attribute. For example, if you have
5121 a memory region named @samp{mem}, you would use @samp{>mem} in the
5122 output section definition. @xref{Output Section Region}. If no address
5123 was specified for the output section, the linker will set the address to
5124 the next available address within the memory region. If the combined
5125 output sections directed to a memory region are too large for the
5126 region, the linker will issue an error message.
5128 It is possible to access the origin and length of a memory in an
5129 expression via the @code{ORIGIN(@var{memory})} and
5130 @code{LENGTH(@var{memory})} functions:
5134 _fstack = ORIGIN(ram) + LENGTH(ram) - 4;
5139 @section PHDRS Command
5141 @cindex program headers
5142 @cindex ELF program headers
5143 @cindex program segments
5144 @cindex segments, ELF
5145 The ELF object file format uses @dfn{program headers}, also knows as
5146 @dfn{segments}. The program headers describe how the program should be
5147 loaded into memory. You can print them out by using the @code{objdump}
5148 program with the @samp{-p} option.
5150 When you run an ELF program on a native ELF system, the system loader
5151 reads the program headers in order to figure out how to load the
5152 program. This will only work if the program headers are set correctly.
5153 This manual does not describe the details of how the system loader
5154 interprets program headers; for more information, see the ELF ABI.
5156 The linker will create reasonable program headers by default. However,
5157 in some cases, you may need to specify the program headers more
5158 precisely. You may use the @code{PHDRS} command for this purpose. When
5159 the linker sees the @code{PHDRS} command in the linker script, it will
5160 not create any program headers other than the ones specified.
5162 The linker only pays attention to the @code{PHDRS} command when
5163 generating an ELF output file. In other cases, the linker will simply
5164 ignore @code{PHDRS}.
5166 This is the syntax of the @code{PHDRS} command. The words @code{PHDRS},
5167 @code{FILEHDR}, @code{AT}, and @code{FLAGS} are keywords.
5173 @var{name} @var{type} [ FILEHDR ] [ PHDRS ] [ AT ( @var{address} ) ]
5174 [ FLAGS ( @var{flags} ) ] ;
5179 The @var{name} is used only for reference in the @code{SECTIONS} command
5180 of the linker script. It is not put into the output file. Program
5181 header names are stored in a separate name space, and will not conflict
5182 with symbol names, file names, or section names. Each program header
5183 must have a distinct name. The headers are processed in order and it
5184 is usual for them to map to sections in ascending load address order.
5186 Certain program header types describe segments of memory which the
5187 system loader will load from the file. In the linker script, you
5188 specify the contents of these segments by placing allocatable output
5189 sections in the segments. You use the @samp{:@var{phdr}} output section
5190 attribute to place a section in a particular segment. @xref{Output
5193 It is normal to put certain sections in more than one segment. This
5194 merely implies that one segment of memory contains another. You may
5195 repeat @samp{:@var{phdr}}, using it once for each segment which should
5196 contain the section.
5198 If you place a section in one or more segments using @samp{:@var{phdr}},
5199 then the linker will place all subsequent allocatable sections which do
5200 not specify @samp{:@var{phdr}} in the same segments. This is for
5201 convenience, since generally a whole set of contiguous sections will be
5202 placed in a single segment. You can use @code{:NONE} to override the
5203 default segment and tell the linker to not put the section in any
5208 You may use the @code{FILEHDR} and @code{PHDRS} keywords after
5209 the program header type to further describe the contents of the segment.
5210 The @code{FILEHDR} keyword means that the segment should include the ELF
5211 file header. The @code{PHDRS} keyword means that the segment should
5212 include the ELF program headers themselves. If applied to a loadable
5213 segment (@code{PT_LOAD}), all prior loadable segments must have one of
5216 The @var{type} may be one of the following. The numbers indicate the
5217 value of the keyword.
5220 @item @code{PT_NULL} (0)
5221 Indicates an unused program header.
5223 @item @code{PT_LOAD} (1)
5224 Indicates that this program header describes a segment to be loaded from
5227 @item @code{PT_DYNAMIC} (2)
5228 Indicates a segment where dynamic linking information can be found.
5230 @item @code{PT_INTERP} (3)
5231 Indicates a segment where the name of the program interpreter may be
5234 @item @code{PT_NOTE} (4)
5235 Indicates a segment holding note information.
5237 @item @code{PT_SHLIB} (5)
5238 A reserved program header type, defined but not specified by the ELF
5241 @item @code{PT_PHDR} (6)
5242 Indicates a segment where the program headers may be found.
5244 @item @var{expression}
5245 An expression giving the numeric type of the program header. This may
5246 be used for types not defined above.
5249 You can specify that a segment should be loaded at a particular address
5250 in memory by using an @code{AT} expression. This is identical to the
5251 @code{AT} command used as an output section attribute (@pxref{Output
5252 Section LMA}). The @code{AT} command for a program header overrides the
5253 output section attribute.
5255 The linker will normally set the segment flags based on the sections
5256 which comprise the segment. You may use the @code{FLAGS} keyword to
5257 explicitly specify the segment flags. The value of @var{flags} must be
5258 an integer. It is used to set the @code{p_flags} field of the program
5261 Here is an example of @code{PHDRS}. This shows a typical set of program
5262 headers used on a native ELF system.
5268 headers PT_PHDR PHDRS ;
5270 text PT_LOAD FILEHDR PHDRS ;
5272 dynamic PT_DYNAMIC ;
5278 .interp : @{ *(.interp) @} :text :interp
5279 .text : @{ *(.text) @} :text
5280 .rodata : @{ *(.rodata) @} /* defaults to :text */
5282 . = . + 0x1000; /* move to a new page in memory */
5283 .data : @{ *(.data) @} :data
5284 .dynamic : @{ *(.dynamic) @} :data :dynamic
5291 @section VERSION Command
5292 @kindex VERSION @{script text@}
5293 @cindex symbol versions
5294 @cindex version script
5295 @cindex versions of symbols
5296 The linker supports symbol versions when using ELF. Symbol versions are
5297 only useful when using shared libraries. The dynamic linker can use
5298 symbol versions to select a specific version of a function when it runs
5299 a program that may have been linked against an earlier version of the
5302 You can include a version script directly in the main linker script, or
5303 you can supply the version script as an implicit linker script. You can
5304 also use the @samp{--version-script} linker option.
5306 The syntax of the @code{VERSION} command is simply
5308 VERSION @{ version-script-commands @}
5311 The format of the version script commands is identical to that used by
5312 Sun's linker in Solaris 2.5. The version script defines a tree of
5313 version nodes. You specify the node names and interdependencies in the
5314 version script. You can specify which symbols are bound to which
5315 version nodes, and you can reduce a specified set of symbols to local
5316 scope so that they are not globally visible outside of the shared
5319 The easiest way to demonstrate the version script language is with a few
5345 This example version script defines three version nodes. The first
5346 version node defined is @samp{VERS_1.1}; it has no other dependencies.
5347 The script binds the symbol @samp{foo1} to @samp{VERS_1.1}. It reduces
5348 a number of symbols to local scope so that they are not visible outside
5349 of the shared library; this is done using wildcard patterns, so that any
5350 symbol whose name begins with @samp{old}, @samp{original}, or @samp{new}
5351 is matched. The wildcard patterns available are the same as those used
5352 in the shell when matching filenames (also known as ``globbing'').
5353 However, if you specify the symbol name inside double quotes, then the
5354 name is treated as literal, rather than as a glob pattern.
5356 Next, the version script defines node @samp{VERS_1.2}. This node
5357 depends upon @samp{VERS_1.1}. The script binds the symbol @samp{foo2}
5358 to the version node @samp{VERS_1.2}.
5360 Finally, the version script defines node @samp{VERS_2.0}. This node
5361 depends upon @samp{VERS_1.2}. The scripts binds the symbols @samp{bar1}
5362 and @samp{bar2} are bound to the version node @samp{VERS_2.0}.
5364 When the linker finds a symbol defined in a library which is not
5365 specifically bound to a version node, it will effectively bind it to an
5366 unspecified base version of the library. You can bind all otherwise
5367 unspecified symbols to a given version node by using @samp{global: *;}
5368 somewhere in the version script. Note that it's slightly crazy to use
5369 wildcards in a global spec except on the last version node. Global
5370 wildcards elsewhere run the risk of accidentally adding symbols to the
5371 set exported for an old version. That's wrong since older versions
5372 ought to have a fixed set of symbols.
5374 The names of the version nodes have no specific meaning other than what
5375 they might suggest to the person reading them. The @samp{2.0} version
5376 could just as well have appeared in between @samp{1.1} and @samp{1.2}.
5377 However, this would be a confusing way to write a version script.
5379 Node name can be omitted, provided it is the only version node
5380 in the version script. Such version script doesn't assign any versions to
5381 symbols, only selects which symbols will be globally visible out and which
5385 @{ global: foo; bar; local: *; @};
5388 When you link an application against a shared library that has versioned
5389 symbols, the application itself knows which version of each symbol it
5390 requires, and it also knows which version nodes it needs from each
5391 shared library it is linked against. Thus at runtime, the dynamic
5392 loader can make a quick check to make sure that the libraries you have
5393 linked against do in fact supply all of the version nodes that the
5394 application will need to resolve all of the dynamic symbols. In this
5395 way it is possible for the dynamic linker to know with certainty that
5396 all external symbols that it needs will be resolvable without having to
5397 search for each symbol reference.
5399 The symbol versioning is in effect a much more sophisticated way of
5400 doing minor version checking that SunOS does. The fundamental problem
5401 that is being addressed here is that typically references to external
5402 functions are bound on an as-needed basis, and are not all bound when
5403 the application starts up. If a shared library is out of date, a
5404 required interface may be missing; when the application tries to use
5405 that interface, it may suddenly and unexpectedly fail. With symbol
5406 versioning, the user will get a warning when they start their program if
5407 the libraries being used with the application are too old.
5409 There are several GNU extensions to Sun's versioning approach. The
5410 first of these is the ability to bind a symbol to a version node in the
5411 source file where the symbol is defined instead of in the versioning
5412 script. This was done mainly to reduce the burden on the library
5413 maintainer. You can do this by putting something like:
5415 __asm__(".symver original_foo,foo@@VERS_1.1");
5418 in the C source file. This renames the function @samp{original_foo} to
5419 be an alias for @samp{foo} bound to the version node @samp{VERS_1.1}.
5420 The @samp{local:} directive can be used to prevent the symbol
5421 @samp{original_foo} from being exported. A @samp{.symver} directive
5422 takes precedence over a version script.
5424 The second GNU extension is to allow multiple versions of the same
5425 function to appear in a given shared library. In this way you can make
5426 an incompatible change to an interface without increasing the major
5427 version number of the shared library, while still allowing applications
5428 linked against the old interface to continue to function.
5430 To do this, you must use multiple @samp{.symver} directives in the
5431 source file. Here is an example:
5434 __asm__(".symver original_foo,foo@@");
5435 __asm__(".symver old_foo,foo@@VERS_1.1");
5436 __asm__(".symver old_foo1,foo@@VERS_1.2");
5437 __asm__(".symver new_foo,foo@@@@VERS_2.0");
5440 In this example, @samp{foo@@} represents the symbol @samp{foo} bound to the
5441 unspecified base version of the symbol. The source file that contains this
5442 example would define 4 C functions: @samp{original_foo}, @samp{old_foo},
5443 @samp{old_foo1}, and @samp{new_foo}.
5445 When you have multiple definitions of a given symbol, there needs to be
5446 some way to specify a default version to which external references to
5447 this symbol will be bound. You can do this with the
5448 @samp{foo@@@@VERS_2.0} type of @samp{.symver} directive. You can only
5449 declare one version of a symbol as the default in this manner; otherwise
5450 you would effectively have multiple definitions of the same symbol.
5452 If you wish to bind a reference to a specific version of the symbol
5453 within the shared library, you can use the aliases of convenience
5454 (i.e., @samp{old_foo}), or you can use the @samp{.symver} directive to
5455 specifically bind to an external version of the function in question.
5457 You can also specify the language in the version script:
5460 VERSION extern "lang" @{ version-script-commands @}
5463 The supported @samp{lang}s are @samp{C}, @samp{C++}, and @samp{Java}.
5464 The linker will iterate over the list of symbols at the link time and
5465 demangle them according to @samp{lang} before matching them to the
5466 patterns specified in @samp{version-script-commands}. The default
5467 @samp{lang} is @samp{C}.
5469 Demangled names may contains spaces and other special characters. As
5470 described above, you can use a glob pattern to match demangled names,
5471 or you can use a double-quoted string to match the string exactly. In
5472 the latter case, be aware that minor differences (such as differing
5473 whitespace) between the version script and the demangler output will
5474 cause a mismatch. As the exact string generated by the demangler
5475 might change in the future, even if the mangled name does not, you
5476 should check that all of your version directives are behaving as you
5477 expect when you upgrade.
5480 @section Expressions in Linker Scripts
5483 The syntax for expressions in the linker script language is identical to
5484 that of C expressions. All expressions are evaluated as integers. All
5485 expressions are evaluated in the same size, which is 32 bits if both the
5486 host and target are 32 bits, and is otherwise 64 bits.
5488 You can use and set symbol values in expressions.
5490 The linker defines several special purpose builtin functions for use in
5494 * Constants:: Constants
5495 * Symbolic Constants:: Symbolic constants
5496 * Symbols:: Symbol Names
5497 * Orphan Sections:: Orphan Sections
5498 * Location Counter:: The Location Counter
5499 * Operators:: Operators
5500 * Evaluation:: Evaluation
5501 * Expression Section:: The Section of an Expression
5502 * Builtin Functions:: Builtin Functions
5506 @subsection Constants
5507 @cindex integer notation
5508 @cindex constants in linker scripts
5509 All constants are integers.
5511 As in C, the linker considers an integer beginning with @samp{0} to be
5512 octal, and an integer beginning with @samp{0x} or @samp{0X} to be
5513 hexadecimal. Alternatively the linker accepts suffixes of @samp{h} or
5514 @samp{H} for hexadecimal, @samp{o} or @samp{O} for octal, @samp{b} or
5515 @samp{B} for binary and @samp{d} or @samp{D} for decimal. Any integer
5516 value without a prefix or a suffix is considered to be decimal.
5518 @cindex scaled integers
5519 @cindex K and M integer suffixes
5520 @cindex M and K integer suffixes
5521 @cindex suffixes for integers
5522 @cindex integer suffixes
5523 In addition, you can use the suffixes @code{K} and @code{M} to scale a
5527 @c END TEXI2ROFF-KILL
5528 @code{1024} or @code{1024*1024}
5532 ${\rm 1024}$ or ${\rm 1024}^2$
5534 @c END TEXI2ROFF-KILL
5535 respectively. For example, the following
5536 all refer to the same quantity:
5545 Note - the @code{K} and @code{M} suffixes cannot be used in
5546 conjunction with the base suffixes mentioned above.
5548 @node Symbolic Constants
5549 @subsection Symbolic Constants
5550 @cindex symbolic constants
5552 It is possible to refer to target specific constants via the use of
5553 the @code{CONSTANT(@var{name})} operator, where @var{name} is one of:
5558 The target's maximum page size.
5560 @item COMMONPAGESIZE
5561 @kindex COMMONPAGESIZE
5562 The target's default page size.
5568 .text ALIGN (CONSTANT (MAXPAGESIZE)) : @{ *(.text) @}
5571 will create a text section aligned to the largest page boundary
5572 supported by the target.
5575 @subsection Symbol Names
5576 @cindex symbol names
5578 @cindex quoted symbol names
5580 Unless quoted, symbol names start with a letter, underscore, or period
5581 and may include letters, digits, underscores, periods, and hyphens.
5582 Unquoted symbol names must not conflict with any keywords. You can
5583 specify a symbol which contains odd characters or has the same name as a
5584 keyword by surrounding the symbol name in double quotes:
5587 "with a space" = "also with a space" + 10;
5590 Since symbols can contain many non-alphabetic characters, it is safest
5591 to delimit symbols with spaces. For example, @samp{A-B} is one symbol,
5592 whereas @samp{A - B} is an expression involving subtraction.
5594 @node Orphan Sections
5595 @subsection Orphan Sections
5597 Orphan sections are sections present in the input files which
5598 are not explicitly placed into the output file by the linker
5599 script. The linker will still copy these sections into the
5600 output file, but it has to guess as to where they should be
5601 placed. The linker uses a simple heuristic to do this. It
5602 attempts to place orphan sections after non-orphan sections of the
5603 same attribute, such as code vs data, loadable vs non-loadable, etc.
5604 If there is not enough room to do this then it places
5605 at the end of the file.
5607 For ELF targets, the attribute of the section includes section type as
5608 well as section flag.
5610 The command line options @samp{--orphan-handling} and @samp{--unique}
5611 (@pxref{Options,,Command Line Options}) can be used to control which
5612 output sections an orphan is placed in.
5614 If an orphaned section's name is representable as a C identifier then
5615 the linker will automatically @pxref{PROVIDE} two symbols:
5616 __start_SECNAME and __stop_SECNAME, where SECNAME is the name of the
5617 section. These indicate the start address and end address of the
5618 orphaned section respectively. Note: most section names are not
5619 representable as C identifiers because they contain a @samp{.}
5622 @node Location Counter
5623 @subsection The Location Counter
5626 @cindex location counter
5627 @cindex current output location
5628 The special linker variable @dfn{dot} @samp{.} always contains the
5629 current output location counter. Since the @code{.} always refers to a
5630 location in an output section, it may only appear in an expression
5631 within a @code{SECTIONS} command. The @code{.} symbol may appear
5632 anywhere that an ordinary symbol is allowed in an expression.
5635 Assigning a value to @code{.} will cause the location counter to be
5636 moved. This may be used to create holes in the output section. The
5637 location counter may not be moved backwards inside an output section,
5638 and may not be moved backwards outside of an output section if so
5639 doing creates areas with overlapping LMAs.
5655 In the previous example, the @samp{.text} section from @file{file1} is
5656 located at the beginning of the output section @samp{output}. It is
5657 followed by a 1000 byte gap. Then the @samp{.text} section from
5658 @file{file2} appears, also with a 1000 byte gap following before the
5659 @samp{.text} section from @file{file3}. The notation @samp{= 0x12345678}
5660 specifies what data to write in the gaps (@pxref{Output Section Fill}).
5662 @cindex dot inside sections
5663 Note: @code{.} actually refers to the byte offset from the start of the
5664 current containing object. Normally this is the @code{SECTIONS}
5665 statement, whose start address is 0, hence @code{.} can be used as an
5666 absolute address. If @code{.} is used inside a section description
5667 however, it refers to the byte offset from the start of that section,
5668 not an absolute address. Thus in a script like this:
5686 The @samp{.text} section will be assigned a starting address of 0x100
5687 and a size of exactly 0x200 bytes, even if there is not enough data in
5688 the @samp{.text} input sections to fill this area. (If there is too
5689 much data, an error will be produced because this would be an attempt to
5690 move @code{.} backwards). The @samp{.data} section will start at 0x500
5691 and it will have an extra 0x600 bytes worth of space after the end of
5692 the values from the @samp{.data} input sections and before the end of
5693 the @samp{.data} output section itself.
5695 @cindex dot outside sections
5696 Setting symbols to the value of the location counter outside of an
5697 output section statement can result in unexpected values if the linker
5698 needs to place orphan sections. For example, given the following:
5704 .text: @{ *(.text) @}
5708 .data: @{ *(.data) @}
5713 If the linker needs to place some input section, e.g. @code{.rodata},
5714 not mentioned in the script, it might choose to place that section
5715 between @code{.text} and @code{.data}. You might think the linker
5716 should place @code{.rodata} on the blank line in the above script, but
5717 blank lines are of no particular significance to the linker. As well,
5718 the linker doesn't associate the above symbol names with their
5719 sections. Instead, it assumes that all assignments or other
5720 statements belong to the previous output section, except for the
5721 special case of an assignment to @code{.}. I.e., the linker will
5722 place the orphan @code{.rodata} section as if the script was written
5729 .text: @{ *(.text) @}
5733 .rodata: @{ *(.rodata) @}
5734 .data: @{ *(.data) @}
5739 This may or may not be the script author's intention for the value of
5740 @code{start_of_data}. One way to influence the orphan section
5741 placement is to assign the location counter to itself, as the linker
5742 assumes that an assignment to @code{.} is setting the start address of
5743 a following output section and thus should be grouped with that
5744 section. So you could write:
5750 .text: @{ *(.text) @}
5755 .data: @{ *(.data) @}
5760 Now, the orphan @code{.rodata} section will be placed between
5761 @code{end_of_text} and @code{start_of_data}.
5765 @subsection Operators
5766 @cindex operators for arithmetic
5767 @cindex arithmetic operators
5768 @cindex precedence in expressions
5769 The linker recognizes the standard C set of arithmetic operators, with
5770 the standard bindings and precedence levels:
5773 @c END TEXI2ROFF-KILL
5775 precedence associativity Operators Notes
5781 5 left == != > < <= >=
5787 11 right &= += -= *= /= (2)
5791 (1) Prefix operators
5792 (2) @xref{Assignments}.
5796 \vskip \baselineskip
5797 %"lispnarrowing" is the extra indent used generally for smallexample
5798 \hskip\lispnarrowing\vbox{\offinterlineskip
5801 {\vrule#&\strut\hfil\ #\ \hfil&\vrule#&\strut\hfil\ #\ \hfil&\vrule#&\strut\hfil\ {\tt #}\ \hfil&\vrule#\cr
5802 height2pt&\omit&&\omit&&\omit&\cr
5803 &Precedence&& Associativity &&{\rm Operators}&\cr
5804 height2pt&\omit&&\omit&&\omit&\cr
5806 height2pt&\omit&&\omit&&\omit&\cr
5808 % '176 is tilde, '~' in tt font
5809 &1&&left&&\qquad- \char'176\ !\qquad\dag&\cr
5810 &2&&left&&* / \%&\cr
5813 &5&&left&&== != > < <= >=&\cr
5816 &8&&left&&{\&\&}&\cr
5819 &11&&right&&\qquad\&= += -= *= /=\qquad\ddag&\cr
5821 height2pt&\omit&&\omit&&\omit&\cr}
5826 @obeylines@parskip=0pt@parindent=0pt
5827 @dag@quad Prefix operators.
5828 @ddag@quad @xref{Assignments}.
5831 @c END TEXI2ROFF-KILL
5834 @subsection Evaluation
5835 @cindex lazy evaluation
5836 @cindex expression evaluation order
5837 The linker evaluates expressions lazily. It only computes the value of
5838 an expression when absolutely necessary.
5840 The linker needs some information, such as the value of the start
5841 address of the first section, and the origins and lengths of memory
5842 regions, in order to do any linking at all. These values are computed
5843 as soon as possible when the linker reads in the linker script.
5845 However, other values (such as symbol values) are not known or needed
5846 until after storage allocation. Such values are evaluated later, when
5847 other information (such as the sizes of output sections) is available
5848 for use in the symbol assignment expression.
5850 The sizes of sections cannot be known until after allocation, so
5851 assignments dependent upon these are not performed until after
5854 Some expressions, such as those depending upon the location counter
5855 @samp{.}, must be evaluated during section allocation.
5857 If the result of an expression is required, but the value is not
5858 available, then an error results. For example, a script like the
5864 .text 9+this_isnt_constant :
5870 will cause the error message @samp{non constant expression for initial
5873 @node Expression Section
5874 @subsection The Section of an Expression
5875 @cindex expression sections
5876 @cindex absolute expressions
5877 @cindex relative expressions
5878 @cindex absolute and relocatable symbols
5879 @cindex relocatable and absolute symbols
5880 @cindex symbols, relocatable and absolute
5881 Addresses and symbols may be section relative, or absolute. A section
5882 relative symbol is relocatable. If you request relocatable output
5883 using the @samp{-r} option, a further link operation may change the
5884 value of a section relative symbol. On the other hand, an absolute
5885 symbol will retain the same value throughout any further link
5888 Some terms in linker expressions are addresses. This is true of
5889 section relative symbols and for builtin functions that return an
5890 address, such as @code{ADDR}, @code{LOADADDR}, @code{ORIGIN} and
5891 @code{SEGMENT_START}. Other terms are simply numbers, or are builtin
5892 functions that return a non-address value, such as @code{LENGTH}.
5893 One complication is that unless you set @code{LD_FEATURE ("SANE_EXPR")}
5894 (@pxref{Miscellaneous Commands}), numbers and absolute symbols are treated
5895 differently depending on their location, for compatibility with older
5896 versions of @code{ld}. Expressions appearing outside an output
5897 section definition treat all numbers as absolute addresses.
5898 Expressions appearing inside an output section definition treat
5899 absolute symbols as numbers. If @code{LD_FEATURE ("SANE_EXPR")} is
5900 given, then absolute symbols and numbers are simply treated as numbers
5903 In the following simple example,
5910 __executable_start = 0x100;
5914 __data_start = 0x10;
5922 both @code{.} and @code{__executable_start} are set to the absolute
5923 address 0x100 in the first two assignments, then both @code{.} and
5924 @code{__data_start} are set to 0x10 relative to the @code{.data}
5925 section in the second two assignments.
5927 For expressions involving numbers, relative addresses and absolute
5928 addresses, ld follows these rules to evaluate terms:
5932 Unary operations on an absolute address or number, and binary
5933 operations on two absolute addresses or two numbers, or between one
5934 absolute address and a number, apply the operator to the value(s).
5936 Unary operations on a relative address, and binary operations on two
5937 relative addresses in the same section or between one relative address
5938 and a number, apply the operator to the offset part of the address(es).
5940 Other binary operations, that is, between two relative addresses not
5941 in the same section, or between a relative address and an absolute
5942 address, first convert any non-absolute term to an absolute address
5943 before applying the operator.
5946 The result section of each sub-expression is as follows:
5950 An operation involving only numbers results in a number.
5952 The result of comparisons, @samp{&&} and @samp{||} is also a number.
5954 The result of other binary arithmetic and logical operations on two
5955 relative addresses in the same section or two absolute addresses
5956 (after above conversions) is also a number.
5958 The result of other operations on relative addresses or one
5959 relative address and a number, is a relative address in the same
5960 section as the relative operand(s).
5962 The result of other operations on absolute addresses (after above
5963 conversions) is an absolute address.
5966 You can use the builtin function @code{ABSOLUTE} to force an expression
5967 to be absolute when it would otherwise be relative. For example, to
5968 create an absolute symbol set to the address of the end of the output
5969 section @samp{.data}:
5973 .data : @{ *(.data) _edata = ABSOLUTE(.); @}
5977 If @samp{ABSOLUTE} were not used, @samp{_edata} would be relative to the
5978 @samp{.data} section.
5980 Using @code{LOADADDR} also forces an expression absolute, since this
5981 particular builtin function returns an absolute address.
5983 @node Builtin Functions
5984 @subsection Builtin Functions
5985 @cindex functions in expressions
5986 The linker script language includes a number of builtin functions for
5987 use in linker script expressions.
5990 @item ABSOLUTE(@var{exp})
5991 @kindex ABSOLUTE(@var{exp})
5992 @cindex expression, absolute
5993 Return the absolute (non-relocatable, as opposed to non-negative) value
5994 of the expression @var{exp}. Primarily useful to assign an absolute
5995 value to a symbol within a section definition, where symbol values are
5996 normally section relative. @xref{Expression Section}.
5998 @item ADDR(@var{section})
5999 @kindex ADDR(@var{section})
6000 @cindex section address in expression
6001 Return the address (VMA) of the named @var{section}. Your
6002 script must previously have defined the location of that section. In
6003 the following example, @code{start_of_output_1}, @code{symbol_1} and
6004 @code{symbol_2} are assigned equivalent values, except that
6005 @code{symbol_1} will be relative to the @code{.output1} section while
6006 the other two will be absolute:
6012 start_of_output_1 = ABSOLUTE(.);
6017 symbol_1 = ADDR(.output1);
6018 symbol_2 = start_of_output_1;
6024 @item ALIGN(@var{align})
6025 @itemx ALIGN(@var{exp},@var{align})
6026 @kindex ALIGN(@var{align})
6027 @kindex ALIGN(@var{exp},@var{align})
6028 @cindex round up location counter
6029 @cindex align location counter
6030 @cindex round up expression
6031 @cindex align expression
6032 Return the location counter (@code{.}) or arbitrary expression aligned
6033 to the next @var{align} boundary. The single operand @code{ALIGN}
6034 doesn't change the value of the location counter---it just does
6035 arithmetic on it. The two operand @code{ALIGN} allows an arbitrary
6036 expression to be aligned upwards (@code{ALIGN(@var{align})} is
6037 equivalent to @code{ALIGN(ABSOLUTE(.), @var{align})}).
6039 Here is an example which aligns the output @code{.data} section to the
6040 next @code{0x2000} byte boundary after the preceding section and sets a
6041 variable within the section to the next @code{0x8000} boundary after the
6046 .data ALIGN(0x2000): @{
6048 variable = ALIGN(0x8000);
6054 The first use of @code{ALIGN} in this example specifies the location of
6055 a section because it is used as the optional @var{address} attribute of
6056 a section definition (@pxref{Output Section Address}). The second use
6057 of @code{ALIGN} is used to defines the value of a symbol.
6059 The builtin function @code{NEXT} is closely related to @code{ALIGN}.
6061 @item ALIGNOF(@var{section})
6062 @kindex ALIGNOF(@var{section})
6063 @cindex section alignment
6064 Return the alignment in bytes of the named @var{section}, if that section has
6065 been allocated. If the section has not been allocated when this is
6066 evaluated, the linker will report an error. In the following example,
6067 the alignment of the @code{.output} section is stored as the first
6068 value in that section.
6073 LONG (ALIGNOF (.output))
6080 @item BLOCK(@var{exp})
6081 @kindex BLOCK(@var{exp})
6082 This is a synonym for @code{ALIGN}, for compatibility with older linker
6083 scripts. It is most often seen when setting the address of an output
6086 @item DATA_SEGMENT_ALIGN(@var{maxpagesize}, @var{commonpagesize})
6087 @kindex DATA_SEGMENT_ALIGN(@var{maxpagesize}, @var{commonpagesize})
6088 This is equivalent to either
6090 (ALIGN(@var{maxpagesize}) + (. & (@var{maxpagesize} - 1)))
6094 (ALIGN(@var{maxpagesize}) + (. & (@var{maxpagesize} - @var{commonpagesize})))
6097 depending on whether the latter uses fewer @var{commonpagesize} sized pages
6098 for the data segment (area between the result of this expression and
6099 @code{DATA_SEGMENT_END}) than the former or not.
6100 If the latter form is used, it means @var{commonpagesize} bytes of runtime
6101 memory will be saved at the expense of up to @var{commonpagesize} wasted
6102 bytes in the on-disk file.
6104 This expression can only be used directly in @code{SECTIONS} commands, not in
6105 any output section descriptions and only once in the linker script.
6106 @var{commonpagesize} should be less or equal to @var{maxpagesize} and should
6107 be the system page size the object wants to be optimized for (while still
6108 working on system page sizes up to @var{maxpagesize}).
6113 . = DATA_SEGMENT_ALIGN(0x10000, 0x2000);
6116 @item DATA_SEGMENT_END(@var{exp})
6117 @kindex DATA_SEGMENT_END(@var{exp})
6118 This defines the end of data segment for @code{DATA_SEGMENT_ALIGN}
6119 evaluation purposes.
6122 . = DATA_SEGMENT_END(.);
6125 @item DATA_SEGMENT_RELRO_END(@var{offset}, @var{exp})
6126 @kindex DATA_SEGMENT_RELRO_END(@var{offset}, @var{exp})
6127 This defines the end of the @code{PT_GNU_RELRO} segment when
6128 @samp{-z relro} option is used.
6129 When @samp{-z relro} option is not present, @code{DATA_SEGMENT_RELRO_END}
6130 does nothing, otherwise @code{DATA_SEGMENT_ALIGN} is padded so that
6131 @var{exp} + @var{offset} is aligned to the most commonly used page
6132 boundary for particular target. If present in the linker script,
6133 it must always come in between @code{DATA_SEGMENT_ALIGN} and
6134 @code{DATA_SEGMENT_END}. Evaluates to the second argument plus any
6135 padding needed at the end of the @code{PT_GNU_RELRO} segment due to
6139 . = DATA_SEGMENT_RELRO_END(24, .);
6142 @item DEFINED(@var{symbol})
6143 @kindex DEFINED(@var{symbol})
6144 @cindex symbol defaults
6145 Return 1 if @var{symbol} is in the linker global symbol table and is
6146 defined before the statement using DEFINED in the script, otherwise
6147 return 0. You can use this function to provide
6148 default values for symbols. For example, the following script fragment
6149 shows how to set a global symbol @samp{begin} to the first location in
6150 the @samp{.text} section---but if a symbol called @samp{begin} already
6151 existed, its value is preserved:
6157 begin = DEFINED(begin) ? begin : . ;
6165 @item LENGTH(@var{memory})
6166 @kindex LENGTH(@var{memory})
6167 Return the length of the memory region named @var{memory}.
6169 @item LOADADDR(@var{section})
6170 @kindex LOADADDR(@var{section})
6171 @cindex section load address in expression
6172 Return the absolute LMA of the named @var{section}. (@pxref{Output
6175 @item LOG2CEIL(@var{exp})
6176 @kindex LOG2CEIL(@var{exp})
6177 Return the binary logarithm of @var{exp} rounded towards infinity.
6178 @code{LOG2CEIL(0)} returns 0.
6181 @item MAX(@var{exp1}, @var{exp2})
6182 Returns the maximum of @var{exp1} and @var{exp2}.
6185 @item MIN(@var{exp1}, @var{exp2})
6186 Returns the minimum of @var{exp1} and @var{exp2}.
6188 @item NEXT(@var{exp})
6189 @kindex NEXT(@var{exp})
6190 @cindex unallocated address, next
6191 Return the next unallocated address that is a multiple of @var{exp}.
6192 This function is closely related to @code{ALIGN(@var{exp})}; unless you
6193 use the @code{MEMORY} command to define discontinuous memory for the
6194 output file, the two functions are equivalent.
6196 @item ORIGIN(@var{memory})
6197 @kindex ORIGIN(@var{memory})
6198 Return the origin of the memory region named @var{memory}.
6200 @item SEGMENT_START(@var{segment}, @var{default})
6201 @kindex SEGMENT_START(@var{segment}, @var{default})
6202 Return the base address of the named @var{segment}. If an explicit
6203 value has already been given for this segment (with a command-line
6204 @samp{-T} option) then that value will be returned otherwise the value
6205 will be @var{default}. At present, the @samp{-T} command-line option
6206 can only be used to set the base address for the ``text'', ``data'', and
6207 ``bss'' sections, but you can use @code{SEGMENT_START} with any segment
6210 @item SIZEOF(@var{section})
6211 @kindex SIZEOF(@var{section})
6212 @cindex section size
6213 Return the size in bytes of the named @var{section}, if that section has
6214 been allocated. If the section has not been allocated when this is
6215 evaluated, the linker will report an error. In the following example,
6216 @code{symbol_1} and @code{symbol_2} are assigned identical values:
6225 symbol_1 = .end - .start ;
6226 symbol_2 = SIZEOF(.output);
6231 @item SIZEOF_HEADERS
6232 @itemx sizeof_headers
6233 @kindex SIZEOF_HEADERS
6235 Return the size in bytes of the output file's headers. This is
6236 information which appears at the start of the output file. You can use
6237 this number when setting the start address of the first section, if you
6238 choose, to facilitate paging.
6240 @cindex not enough room for program headers
6241 @cindex program headers, not enough room
6242 When producing an ELF output file, if the linker script uses the
6243 @code{SIZEOF_HEADERS} builtin function, the linker must compute the
6244 number of program headers before it has determined all the section
6245 addresses and sizes. If the linker later discovers that it needs
6246 additional program headers, it will report an error @samp{not enough
6247 room for program headers}. To avoid this error, you must avoid using
6248 the @code{SIZEOF_HEADERS} function, or you must rework your linker
6249 script to avoid forcing the linker to use additional program headers, or
6250 you must define the program headers yourself using the @code{PHDRS}
6251 command (@pxref{PHDRS}).
6254 @node Implicit Linker Scripts
6255 @section Implicit Linker Scripts
6256 @cindex implicit linker scripts
6257 If you specify a linker input file which the linker can not recognize as
6258 an object file or an archive file, it will try to read the file as a
6259 linker script. If the file can not be parsed as a linker script, the
6260 linker will report an error.
6262 An implicit linker script will not replace the default linker script.
6264 Typically an implicit linker script would contain only symbol
6265 assignments, or the @code{INPUT}, @code{GROUP}, or @code{VERSION}
6268 Any input files read because of an implicit linker script will be read
6269 at the position in the command line where the implicit linker script was
6270 read. This can affect archive searching.
6273 @node Machine Dependent
6274 @chapter Machine Dependent Features
6276 @cindex machine dependencies
6277 @command{ld} has additional features on some platforms; the following
6278 sections describe them. Machines where @command{ld} has no additional
6279 functionality are not listed.
6283 * H8/300:: @command{ld} and the H8/300
6286 * i960:: @command{ld} and the Intel 960 family
6289 * M68HC11/68HC12:: @code{ld} and the Motorola 68HC11 and 68HC12 families
6292 * ARM:: @command{ld} and the ARM family
6295 * HPPA ELF32:: @command{ld} and HPPA 32-bit ELF
6298 * M68K:: @command{ld} and the Motorola 68K family
6301 * MIPS:: @command{ld} and the MIPS family
6304 * MMIX:: @command{ld} and MMIX
6307 * MSP430:: @command{ld} and MSP430
6310 * NDS32:: @command{ld} and NDS32
6313 * Nios II:: @command{ld} and the Altera Nios II
6316 * PowerPC ELF32:: @command{ld} and PowerPC 32-bit ELF Support
6319 * PowerPC64 ELF64:: @command{ld} and PowerPC64 64-bit ELF Support
6322 * SPU ELF:: @command{ld} and SPU ELF Support
6325 * TI COFF:: @command{ld} and TI COFF
6328 * WIN32:: @command{ld} and WIN32 (cygwin/mingw)
6331 * Xtensa:: @command{ld} and Xtensa Processors
6342 @section @command{ld} and the H8/300
6344 @cindex H8/300 support
6345 For the H8/300, @command{ld} can perform these global optimizations when
6346 you specify the @samp{--relax} command-line option.
6349 @cindex relaxing on H8/300
6350 @item relaxing address modes
6351 @command{ld} finds all @code{jsr} and @code{jmp} instructions whose
6352 targets are within eight bits, and turns them into eight-bit
6353 program-counter relative @code{bsr} and @code{bra} instructions,
6356 @cindex synthesizing on H8/300
6357 @item synthesizing instructions
6358 @c FIXME: specifically mov.b, or any mov instructions really? -> mov.b only, at least on H8, H8H, H8S
6359 @command{ld} finds all @code{mov.b} instructions which use the
6360 sixteen-bit absolute address form, but refer to the top
6361 page of memory, and changes them to use the eight-bit address form.
6362 (That is: the linker turns @samp{mov.b @code{@@}@var{aa}:16} into
6363 @samp{mov.b @code{@@}@var{aa}:8} whenever the address @var{aa} is in the
6364 top page of memory).
6366 @command{ld} finds all @code{mov} instructions which use the register
6367 indirect with 32-bit displacement addressing mode, but use a small
6368 displacement inside 16-bit displacement range, and changes them to use
6369 the 16-bit displacement form. (That is: the linker turns @samp{mov.b
6370 @code{@@}@var{d}:32,ERx} into @samp{mov.b @code{@@}@var{d}:16,ERx}
6371 whenever the displacement @var{d} is in the 16 bit signed integer
6372 range. Only implemented in ELF-format ld).
6374 @item bit manipulation instructions
6375 @command{ld} finds all bit manipulation instructions like @code{band, bclr,
6376 biand, bild, bior, bist, bixor, bld, bnot, bor, bset, bst, btst, bxor}
6377 which use 32 bit and 16 bit absolute address form, but refer to the top
6378 page of memory, and changes them to use the 8 bit address form.
6379 (That is: the linker turns @samp{bset #xx:3,@code{@@}@var{aa}:32} into
6380 @samp{bset #xx:3,@code{@@}@var{aa}:8} whenever the address @var{aa} is in
6381 the top page of memory).
6383 @item system control instructions
6384 @command{ld} finds all @code{ldc.w, stc.w} instructions which use the
6385 32 bit absolute address form, but refer to the top page of memory, and
6386 changes them to use 16 bit address form.
6387 (That is: the linker turns @samp{ldc.w @code{@@}@var{aa}:32,ccr} into
6388 @samp{ldc.w @code{@@}@var{aa}:16,ccr} whenever the address @var{aa} is in
6389 the top page of memory).
6399 @c This stuff is pointless to say unless you're especially concerned
6400 @c with Renesas chips; don't enable it for generic case, please.
6402 @chapter @command{ld} and Other Renesas Chips
6404 @command{ld} also supports the Renesas (formerly Hitachi) H8/300H,
6405 H8/500, and SH chips. No special features, commands, or command-line
6406 options are required for these chips.
6416 @section @command{ld} and the Intel 960 Family
6418 @cindex i960 support
6420 You can use the @samp{-A@var{architecture}} command line option to
6421 specify one of the two-letter names identifying members of the 960
6422 family; the option specifies the desired output target, and warns of any
6423 incompatible instructions in the input files. It also modifies the
6424 linker's search strategy for archive libraries, to support the use of
6425 libraries specific to each particular architecture, by including in the
6426 search loop names suffixed with the string identifying the architecture.
6428 For example, if your @command{ld} command line included @w{@samp{-ACA}} as
6429 well as @w{@samp{-ltry}}, the linker would look (in its built-in search
6430 paths, and in any paths you specify with @samp{-L}) for a library with
6443 The first two possibilities would be considered in any event; the last
6444 two are due to the use of @w{@samp{-ACA}}.
6446 You can meaningfully use @samp{-A} more than once on a command line, since
6447 the 960 architecture family allows combination of target architectures; each
6448 use will add another pair of name variants to search for when @w{@samp{-l}}
6449 specifies a library.
6451 @cindex @option{--relax} on i960
6452 @cindex relaxing on i960
6453 @command{ld} supports the @samp{--relax} option for the i960 family. If
6454 you specify @samp{--relax}, @command{ld} finds all @code{balx} and
6455 @code{calx} instructions whose targets are within 24 bits, and turns
6456 them into 24-bit program-counter relative @code{bal} and @code{cal}
6457 instructions, respectively. @command{ld} also turns @code{cal}
6458 instructions into @code{bal} instructions when it determines that the
6459 target subroutine is a leaf routine (that is, the target subroutine does
6460 not itself call any subroutines).
6477 @node M68HC11/68HC12
6478 @section @command{ld} and the Motorola 68HC11 and 68HC12 families
6480 @cindex M68HC11 and 68HC12 support
6482 @subsection Linker Relaxation
6484 For the Motorola 68HC11, @command{ld} can perform these global
6485 optimizations when you specify the @samp{--relax} command-line option.
6488 @cindex relaxing on M68HC11
6489 @item relaxing address modes
6490 @command{ld} finds all @code{jsr} and @code{jmp} instructions whose
6491 targets are within eight bits, and turns them into eight-bit
6492 program-counter relative @code{bsr} and @code{bra} instructions,
6495 @command{ld} also looks at all 16-bit extended addressing modes and
6496 transforms them in a direct addressing mode when the address is in
6497 page 0 (between 0 and 0x0ff).
6499 @item relaxing gcc instruction group
6500 When @command{gcc} is called with @option{-mrelax}, it can emit group
6501 of instructions that the linker can optimize to use a 68HC11 direct
6502 addressing mode. These instructions consists of @code{bclr} or
6503 @code{bset} instructions.
6507 @subsection Trampoline Generation
6509 @cindex trampoline generation on M68HC11
6510 @cindex trampoline generation on M68HC12
6511 For 68HC11 and 68HC12, @command{ld} can generate trampoline code to
6512 call a far function using a normal @code{jsr} instruction. The linker
6513 will also change the relocation to some far function to use the
6514 trampoline address instead of the function address. This is typically the
6515 case when a pointer to a function is taken. The pointer will in fact
6516 point to the function trampoline.
6524 @section @command{ld} and the ARM family
6526 @cindex ARM interworking support
6527 @kindex --support-old-code
6528 For the ARM, @command{ld} will generate code stubs to allow functions calls
6529 between ARM and Thumb code. These stubs only work with code that has
6530 been compiled and assembled with the @samp{-mthumb-interwork} command
6531 line option. If it is necessary to link with old ARM object files or
6532 libraries, which have not been compiled with the -mthumb-interwork
6533 option then the @samp{--support-old-code} command line switch should be
6534 given to the linker. This will make it generate larger stub functions
6535 which will work with non-interworking aware ARM code. Note, however,
6536 the linker does not support generating stubs for function calls to
6537 non-interworking aware Thumb code.
6539 @cindex thumb entry point
6540 @cindex entry point, thumb
6541 @kindex --thumb-entry=@var{entry}
6542 The @samp{--thumb-entry} switch is a duplicate of the generic
6543 @samp{--entry} switch, in that it sets the program's starting address.
6544 But it also sets the bottom bit of the address, so that it can be
6545 branched to using a BX instruction, and the program will start
6546 executing in Thumb mode straight away.
6548 @cindex PE import table prefixing
6549 @kindex --use-nul-prefixed-import-tables
6550 The @samp{--use-nul-prefixed-import-tables} switch is specifying, that
6551 the import tables idata4 and idata5 have to be generated with a zero
6552 element prefix for import libraries. This is the old style to generate
6553 import tables. By default this option is turned off.
6557 The @samp{--be8} switch instructs @command{ld} to generate BE8 format
6558 executables. This option is only valid when linking big-endian
6559 objects - ie ones which have been assembled with the @option{-EB}
6560 option. The resulting image will contain big-endian data and
6564 @kindex --target1-rel
6565 @kindex --target1-abs
6566 The @samp{R_ARM_TARGET1} relocation is typically used for entries in the
6567 @samp{.init_array} section. It is interpreted as either @samp{R_ARM_REL32}
6568 or @samp{R_ARM_ABS32}, depending on the target. The @samp{--target1-rel}
6569 and @samp{--target1-abs} switches override the default.
6572 @kindex --target2=@var{type}
6573 The @samp{--target2=type} switch overrides the default definition of the
6574 @samp{R_ARM_TARGET2} relocation. Valid values for @samp{type}, their
6575 meanings, and target defaults are as follows:
6578 @samp{R_ARM_REL32} (arm*-*-elf, arm*-*-eabi)
6580 @samp{R_ARM_ABS32} (arm*-*-symbianelf)
6582 @samp{R_ARM_GOT_PREL} (arm*-*-linux, arm*-*-*bsd)
6587 The @samp{R_ARM_V4BX} relocation (defined by the ARM AAELF
6588 specification) enables objects compiled for the ARMv4 architecture to be
6589 interworking-safe when linked with other objects compiled for ARMv4t, but
6590 also allows pure ARMv4 binaries to be built from the same ARMv4 objects.
6592 In the latter case, the switch @option{--fix-v4bx} must be passed to the
6593 linker, which causes v4t @code{BX rM} instructions to be rewritten as
6594 @code{MOV PC,rM}, since v4 processors do not have a @code{BX} instruction.
6596 In the former case, the switch should not be used, and @samp{R_ARM_V4BX}
6597 relocations are ignored.
6599 @cindex FIX_V4BX_INTERWORKING
6600 @kindex --fix-v4bx-interworking
6601 Replace @code{BX rM} instructions identified by @samp{R_ARM_V4BX}
6602 relocations with a branch to the following veneer:
6610 This allows generation of libraries/applications that work on ARMv4 cores
6611 and are still interworking safe. Note that the above veneer clobbers the
6612 condition flags, so may cause incorrect program behavior in rare cases.
6616 The @samp{--use-blx} switch enables the linker to use ARM/Thumb
6617 BLX instructions (available on ARMv5t and above) in various
6618 situations. Currently it is used to perform calls via the PLT from Thumb
6619 code using BLX rather than using BX and a mode-switching stub before
6620 each PLT entry. This should lead to such calls executing slightly faster.
6622 This option is enabled implicitly for SymbianOS, so there is no need to
6623 specify it if you are using that target.
6625 @cindex VFP11_DENORM_FIX
6626 @kindex --vfp11-denorm-fix
6627 The @samp{--vfp11-denorm-fix} switch enables a link-time workaround for a
6628 bug in certain VFP11 coprocessor hardware, which sometimes allows
6629 instructions with denorm operands (which must be handled by support code)
6630 to have those operands overwritten by subsequent instructions before
6631 the support code can read the intended values.
6633 The bug may be avoided in scalar mode if you allow at least one
6634 intervening instruction between a VFP11 instruction which uses a register
6635 and another instruction which writes to the same register, or at least two
6636 intervening instructions if vector mode is in use. The bug only affects
6637 full-compliance floating-point mode: you do not need this workaround if
6638 you are using "runfast" mode. Please contact ARM for further details.
6640 If you know you are using buggy VFP11 hardware, you can
6641 enable this workaround by specifying the linker option
6642 @samp{--vfp-denorm-fix=scalar} if you are using the VFP11 scalar
6643 mode only, or @samp{--vfp-denorm-fix=vector} if you are using
6644 vector mode (the latter also works for scalar code). The default is
6645 @samp{--vfp-denorm-fix=none}.
6647 If the workaround is enabled, instructions are scanned for
6648 potentially-troublesome sequences, and a veneer is created for each
6649 such sequence which may trigger the erratum. The veneer consists of the
6650 first instruction of the sequence and a branch back to the subsequent
6651 instruction. The original instruction is then replaced with a branch to
6652 the veneer. The extra cycles required to call and return from the veneer
6653 are sufficient to avoid the erratum in both the scalar and vector cases.
6655 @cindex ARM1176 erratum workaround
6656 @kindex --fix-arm1176
6657 @kindex --no-fix-arm1176
6658 The @samp{--fix-arm1176} switch enables a link-time workaround for an erratum
6659 in certain ARM1176 processors. The workaround is enabled by default if you
6660 are targeting ARM v6 (excluding ARM v6T2) or earlier. It can be disabled
6661 unconditionally by specifying @samp{--no-fix-arm1176}.
6663 Further information is available in the ``ARM1176JZ-S and ARM1176JZF-S
6664 Programmer Advice Notice'' available on the ARM documentation website at:
6665 http://infocenter.arm.com/.
6667 @cindex NO_ENUM_SIZE_WARNING
6668 @kindex --no-enum-size-warning
6669 The @option{--no-enum-size-warning} switch prevents the linker from
6670 warning when linking object files that specify incompatible EABI
6671 enumeration size attributes. For example, with this switch enabled,
6672 linking of an object file using 32-bit enumeration values with another
6673 using enumeration values fitted into the smallest possible space will
6676 @cindex NO_WCHAR_SIZE_WARNING
6677 @kindex --no-wchar-size-warning
6678 The @option{--no-wchar-size-warning} switch prevents the linker from
6679 warning when linking object files that specify incompatible EABI
6680 @code{wchar_t} size attributes. For example, with this switch enabled,
6681 linking of an object file using 32-bit @code{wchar_t} values with another
6682 using 16-bit @code{wchar_t} values will not be diagnosed.
6685 @kindex --pic-veneer
6686 The @samp{--pic-veneer} switch makes the linker use PIC sequences for
6687 ARM/Thumb interworking veneers, even if the rest of the binary
6688 is not PIC. This avoids problems on uClinux targets where
6689 @samp{--emit-relocs} is used to generate relocatable binaries.
6691 @cindex STUB_GROUP_SIZE
6692 @kindex --stub-group-size=@var{N}
6693 The linker will automatically generate and insert small sequences of
6694 code into a linked ARM ELF executable whenever an attempt is made to
6695 perform a function call to a symbol that is too far away. The
6696 placement of these sequences of instructions - called stubs - is
6697 controlled by the command line option @option{--stub-group-size=N}.
6698 The placement is important because a poor choice can create a need for
6699 duplicate stubs, increasing the code size. The linker will try to
6700 group stubs together in order to reduce interruptions to the flow of
6701 code, but it needs guidance as to how big these groups should be and
6702 where they should be placed.
6704 The value of @samp{N}, the parameter to the
6705 @option{--stub-group-size=} option controls where the stub groups are
6706 placed. If it is negative then all stubs are placed after the first
6707 branch that needs them. If it is positive then the stubs can be
6708 placed either before or after the branches that need them. If the
6709 value of @samp{N} is 1 (either +1 or -1) then the linker will choose
6710 exactly where to place groups of stubs, using its built in heuristics.
6711 A value of @samp{N} greater than 1 (or smaller than -1) tells the
6712 linker that a single group of stubs can service at most @samp{N} bytes
6713 from the input sections.
6715 The default, if @option{--stub-group-size=} is not specified, is
6718 Farcalls stubs insertion is fully supported for the ARM-EABI target
6719 only, because it relies on object files properties not present
6722 @cindex Cortex-A8 erratum workaround
6723 @kindex --fix-cortex-a8
6724 @kindex --no-fix-cortex-a8
6725 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}.
6727 The erratum only affects Thumb-2 code. Please contact ARM for further details.
6729 @cindex Cortex-A53 erratum 835769 workaround
6730 @kindex --fix-cortex-a53-835769
6731 @kindex --no-fix-cortex-a53-835769
6732 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}.
6734 Please contact ARM for further details.
6736 @kindex --merge-exidx-entries
6737 @kindex --no-merge-exidx-entries
6738 @cindex Merging exidx entries
6739 The @samp{--no-merge-exidx-entries} switch disables the merging of adjacent exidx entries in debuginfo.
6742 @cindex 32-bit PLT entries
6743 The @samp{--long-plt} option enables the use of 16 byte PLT entries
6744 which support up to 4Gb of code. The default is to use 12 byte PLT
6745 entries which only support 512Mb of code.
6758 @section @command{ld} and HPPA 32-bit ELF Support
6759 @cindex HPPA multiple sub-space stubs
6760 @kindex --multi-subspace
6761 When generating a shared library, @command{ld} will by default generate
6762 import stubs suitable for use with a single sub-space application.
6763 The @samp{--multi-subspace} switch causes @command{ld} to generate export
6764 stubs, and different (larger) import stubs suitable for use with
6765 multiple sub-spaces.
6767 @cindex HPPA stub grouping
6768 @kindex --stub-group-size=@var{N}
6769 Long branch stubs and import/export stubs are placed by @command{ld} in
6770 stub sections located between groups of input sections.
6771 @samp{--stub-group-size} specifies the maximum size of a group of input
6772 sections handled by one stub section. Since branch offsets are signed,
6773 a stub section may serve two groups of input sections, one group before
6774 the stub section, and one group after it. However, when using
6775 conditional branches that require stubs, it may be better (for branch
6776 prediction) that stub sections only serve one group of input sections.
6777 A negative value for @samp{N} chooses this scheme, ensuring that
6778 branches to stubs always use a negative offset. Two special values of
6779 @samp{N} are recognized, @samp{1} and @samp{-1}. These both instruct
6780 @command{ld} to automatically size input section groups for the branch types
6781 detected, with the same behaviour regarding stub placement as other
6782 positive or negative values of @samp{N} respectively.
6784 Note that @samp{--stub-group-size} does not split input sections. A
6785 single input section larger than the group size specified will of course
6786 create a larger group (of one section). If input sections are too
6787 large, it may not be possible for a branch to reach its stub.
6800 @section @command{ld} and the Motorola 68K family
6802 @cindex Motorola 68K GOT generation
6803 @kindex --got=@var{type}
6804 The @samp{--got=@var{type}} option lets you choose the GOT generation scheme.
6805 The choices are @samp{single}, @samp{negative}, @samp{multigot} and
6806 @samp{target}. When @samp{target} is selected the linker chooses
6807 the default GOT generation scheme for the current target.
6808 @samp{single} tells the linker to generate a single GOT with
6809 entries only at non-negative offsets.
6810 @samp{negative} instructs the linker to generate a single GOT with
6811 entries at both negative and positive offsets. Not all environments
6813 @samp{multigot} allows the linker to generate several GOTs in the
6814 output file. All GOT references from a single input object
6815 file access the same GOT, but references from different input object
6816 files might access different GOTs. Not all environments support such GOTs.
6829 @section @command{ld} and the MIPS family
6831 @cindex MIPS microMIPS instruction choice selection
6834 The @samp{--insn32} and @samp{--no-insn32} options control the choice of
6835 microMIPS instructions used in code generated by the linker, such as that
6836 in the PLT or lazy binding stubs, or in relaxation. If @samp{--insn32} is
6837 used, then the linker only uses 32-bit instruction encodings. By default
6838 or if @samp{--no-insn32} is used, all instruction encodings are used,
6839 including 16-bit ones where possible.
6852 @section @code{ld} and MMIX
6853 For MMIX, there is a choice of generating @code{ELF} object files or
6854 @code{mmo} object files when linking. The simulator @code{mmix}
6855 understands the @code{mmo} format. The binutils @code{objcopy} utility
6856 can translate between the two formats.
6858 There is one special section, the @samp{.MMIX.reg_contents} section.
6859 Contents in this section is assumed to correspond to that of global
6860 registers, and symbols referring to it are translated to special symbols,
6861 equal to registers. In a final link, the start address of the
6862 @samp{.MMIX.reg_contents} section corresponds to the first allocated
6863 global register multiplied by 8. Register @code{$255} is not included in
6864 this section; it is always set to the program entry, which is at the
6865 symbol @code{Main} for @code{mmo} files.
6867 Global symbols with the prefix @code{__.MMIX.start.}, for example
6868 @code{__.MMIX.start..text} and @code{__.MMIX.start..data} are special.
6869 The default linker script uses these to set the default start address
6872 Initial and trailing multiples of zero-valued 32-bit words in a section,
6873 are left out from an mmo file.
6886 @section @code{ld} and MSP430
6887 For the MSP430 it is possible to select the MPU architecture. The flag @samp{-m [mpu type]}
6888 will select an appropriate linker script for selected MPU type. (To get a list of known MPUs
6889 just pass @samp{-m help} option to the linker).
6891 @cindex MSP430 extra sections
6892 The linker will recognize some extra sections which are MSP430 specific:
6895 @item @samp{.vectors}
6896 Defines a portion of ROM where interrupt vectors located.
6898 @item @samp{.bootloader}
6899 Defines the bootloader portion of the ROM (if applicable). Any code
6900 in this section will be uploaded to the MPU.
6902 @item @samp{.infomem}
6903 Defines an information memory section (if applicable). Any code in
6904 this section will be uploaded to the MPU.
6906 @item @samp{.infomemnobits}
6907 This is the same as the @samp{.infomem} section except that any code
6908 in this section will not be uploaded to the MPU.
6910 @item @samp{.noinit}
6911 Denotes a portion of RAM located above @samp{.bss} section.
6913 The last two sections are used by gcc.
6927 @section @code{ld} and NDS32
6928 @kindex relaxing on NDS32
6929 For NDS32, there are some options to select relaxation behavior. The linker
6930 relaxes objects according to these options.
6933 @item @samp{--m[no-]fp-as-gp}
6934 Disable/enable fp-as-gp relaxation.
6936 @item @samp{--mexport-symbols=FILE}
6937 Exporting symbols and their address into FILE as linker script.
6939 @item @samp{--m[no-]ex9}
6940 Disable/enable link-time EX9 relaxation.
6942 @item @samp{--mexport-ex9=FILE}
6943 Export the EX9 table after linking.
6945 @item @samp{--mimport-ex9=FILE}
6946 Import the Ex9 table for EX9 relaxation.
6948 @item @samp{--mupdate-ex9}
6949 Update the existing EX9 table.
6951 @item @samp{--mex9-limit=NUM}
6952 Maximum number of entries in the ex9 table.
6954 @item @samp{--mex9-loop-aware}
6955 Avoid generating the EX9 instruction inside the loop.
6957 @item @samp{--m[no-]ifc}
6958 Disable/enable the link-time IFC optimization.
6960 @item @samp{--mifc-loop-aware}
6961 Avoid generating the IFC instruction inside the loop.
6975 @section @command{ld} and the Altera Nios II
6976 @cindex Nios II call relaxation
6977 @kindex --relax on Nios II
6979 Call and immediate jump instructions on Nios II processors are limited to
6980 transferring control to addresses in the same 256MB memory segment,
6981 which may result in @command{ld} giving
6982 @samp{relocation truncated to fit} errors with very large programs.
6983 The command-line option @option{--relax} enables the generation of
6984 trampolines that can access the entire 32-bit address space for calls
6985 outside the normal @code{call} and @code{jmpi} address range. These
6986 trampolines are inserted at section boundaries, so may not themselves
6987 be reachable if an input section and its associated call trampolines are
6990 The @option{--relax} option is enabled by default unless @option{-r}
6991 is also specified. You can disable trampoline generation by using the
6992 @option{--no-relax} linker option. You can also disable this optimization
6993 locally by using the @samp{set .noat} directive in assembly-language
6994 source files, as the linker-inserted trampolines use the @code{at}
6995 register as a temporary.
6997 Note that the linker @option{--relax} option is independent of assembler
6998 relaxation options, and that using the GNU assembler's @option{-relax-all}
6999 option interferes with the linker's more selective call instruction relaxation.
7012 @section @command{ld} and PowerPC 32-bit ELF Support
7013 @cindex PowerPC long branches
7014 @kindex --relax on PowerPC
7015 Branches on PowerPC processors are limited to a signed 26-bit
7016 displacement, which may result in @command{ld} giving
7017 @samp{relocation truncated to fit} errors with very large programs.
7018 @samp{--relax} enables the generation of trampolines that can access
7019 the entire 32-bit address space. These trampolines are inserted at
7020 section boundaries, so may not themselves be reachable if an input
7021 section exceeds 33M in size. You may combine @samp{-r} and
7022 @samp{--relax} to add trampolines in a partial link. In that case
7023 both branches to undefined symbols and inter-section branches are also
7024 considered potentially out of range, and trampolines inserted.
7026 @cindex PowerPC ELF32 options
7031 Current PowerPC GCC accepts a @samp{-msecure-plt} option that
7032 generates code capable of using a newer PLT and GOT layout that has
7033 the security advantage of no executable section ever needing to be
7034 writable and no writable section ever being executable. PowerPC
7035 @command{ld} will generate this layout, including stubs to access the
7036 PLT, if all input files (including startup and static libraries) were
7037 compiled with @samp{-msecure-plt}. @samp{--bss-plt} forces the old
7038 BSS PLT (and GOT layout) which can give slightly better performance.
7040 @kindex --secure-plt
7042 @command{ld} will use the new PLT and GOT layout if it is linking new
7043 @samp{-fpic} or @samp{-fPIC} code, but does not do so automatically
7044 when linking non-PIC code. This option requests the new PLT and GOT
7045 layout. A warning will be given if some object file requires the old
7051 The new secure PLT and GOT are placed differently relative to other
7052 sections compared to older BSS PLT and GOT placement. The location of
7053 @code{.plt} must change because the new secure PLT is an initialized
7054 section while the old PLT is uninitialized. The reason for the
7055 @code{.got} change is more subtle: The new placement allows
7056 @code{.got} to be read-only in applications linked with
7057 @samp{-z relro -z now}. However, this placement means that
7058 @code{.sdata} cannot always be used in shared libraries, because the
7059 PowerPC ABI accesses @code{.sdata} in shared libraries from the GOT
7060 pointer. @samp{--sdata-got} forces the old GOT placement. PowerPC
7061 GCC doesn't use @code{.sdata} in shared libraries, so this option is
7062 really only useful for other compilers that may do so.
7064 @cindex PowerPC stub symbols
7065 @kindex --emit-stub-syms
7066 @item --emit-stub-syms
7067 This option causes @command{ld} to label linker stubs with a local
7068 symbol that encodes the stub type and destination.
7070 @cindex PowerPC TLS optimization
7071 @kindex --no-tls-optimize
7072 @item --no-tls-optimize
7073 PowerPC @command{ld} normally performs some optimization of code
7074 sequences used to access Thread-Local Storage. Use this option to
7075 disable the optimization.
7088 @node PowerPC64 ELF64
7089 @section @command{ld} and PowerPC64 64-bit ELF Support
7091 @cindex PowerPC64 ELF64 options
7093 @cindex PowerPC64 stub grouping
7094 @kindex --stub-group-size
7095 @item --stub-group-size
7096 Long branch stubs, PLT call stubs and TOC adjusting stubs are placed
7097 by @command{ld} in stub sections located between groups of input sections.
7098 @samp{--stub-group-size} specifies the maximum size of a group of input
7099 sections handled by one stub section. Since branch offsets are signed,
7100 a stub section may serve two groups of input sections, one group before
7101 the stub section, and one group after it. However, when using
7102 conditional branches that require stubs, it may be better (for branch
7103 prediction) that stub sections only serve one group of input sections.
7104 A negative value for @samp{N} chooses this scheme, ensuring that
7105 branches to stubs always use a negative offset. Two special values of
7106 @samp{N} are recognized, @samp{1} and @samp{-1}. These both instruct
7107 @command{ld} to automatically size input section groups for the branch types
7108 detected, with the same behaviour regarding stub placement as other
7109 positive or negative values of @samp{N} respectively.
7111 Note that @samp{--stub-group-size} does not split input sections. A
7112 single input section larger than the group size specified will of course
7113 create a larger group (of one section). If input sections are too
7114 large, it may not be possible for a branch to reach its stub.
7116 @cindex PowerPC64 stub symbols
7117 @kindex --emit-stub-syms
7118 @item --emit-stub-syms
7119 This option causes @command{ld} to label linker stubs with a local
7120 symbol that encodes the stub type and destination.
7122 @cindex PowerPC64 dot symbols
7124 @kindex --no-dotsyms
7125 @item --dotsyms, --no-dotsyms
7126 These two options control how @command{ld} interprets version patterns
7127 in a version script. Older PowerPC64 compilers emitted both a
7128 function descriptor symbol with the same name as the function, and a
7129 code entry symbol with the name prefixed by a dot (@samp{.}). To
7130 properly version a function @samp{foo}, the version script thus needs
7131 to control both @samp{foo} and @samp{.foo}. The option
7132 @samp{--dotsyms}, on by default, automatically adds the required
7133 dot-prefixed patterns. Use @samp{--no-dotsyms} to disable this
7136 @cindex PowerPC64 register save/restore functions
7137 @kindex --save-restore-funcs
7138 @kindex --no-save-restore-funcs
7139 @item --save-restore-funcs, --no-save-restore-funcs
7140 These two options control whether PowerPC64 @command{ld} automatically
7141 provides out-of-line register save and restore functions used by
7142 @samp{-Os} code. The default is to provide any such referenced
7143 function for a normal final link, and to not do so for a relocatable
7146 @cindex PowerPC64 TLS optimization
7147 @kindex --no-tls-optimize
7148 @item --no-tls-optimize
7149 PowerPC64 @command{ld} normally performs some optimization of code
7150 sequences used to access Thread-Local Storage. Use this option to
7151 disable the optimization.
7153 @cindex PowerPC64 __tls_get_addr optimization
7154 @kindex --tls-get-addr-optimize
7155 @kindex --no-tls-get-addr-optimize
7156 @item --tls-get-addr-optimize, --no-tls-get-addr-optimize
7157 These options control whether PowerPC64 @command{ld} uses a special
7158 stub to call __tls_get_addr. PowerPC64 glibc 2.22 and later support
7159 an optimization that allows the second and subsequent calls to
7160 @code{__tls_get_addr} for a given symbol to be resolved by the special
7161 stub without calling in to glibc. By default the linker enables this
7162 option when glibc advertises the availability of __tls_get_addr_opt.
7163 Forcing this option on when using an older glibc won't do much besides
7164 slow down your applications, but may be useful if linking an
7165 application against an older glibc with the expectation that it will
7166 normally be used on systems having a newer glibc.
7168 @cindex PowerPC64 OPD optimization
7169 @kindex --no-opd-optimize
7170 @item --no-opd-optimize
7171 PowerPC64 @command{ld} normally removes @code{.opd} section entries
7172 corresponding to deleted link-once functions, or functions removed by
7173 the action of @samp{--gc-sections} or linker script @code{/DISCARD/}.
7174 Use this option to disable @code{.opd} optimization.
7176 @cindex PowerPC64 OPD spacing
7177 @kindex --non-overlapping-opd
7178 @item --non-overlapping-opd
7179 Some PowerPC64 compilers have an option to generate compressed
7180 @code{.opd} entries spaced 16 bytes apart, overlapping the third word,
7181 the static chain pointer (unused in C) with the first word of the next
7182 entry. This option expands such entries to the full 24 bytes.
7184 @cindex PowerPC64 TOC optimization
7185 @kindex --no-toc-optimize
7186 @item --no-toc-optimize
7187 PowerPC64 @command{ld} normally removes unused @code{.toc} section
7188 entries. Such entries are detected by examining relocations that
7189 reference the TOC in code sections. A reloc in a deleted code section
7190 marks a TOC word as unneeded, while a reloc in a kept code section
7191 marks a TOC word as needed. Since the TOC may reference itself, TOC
7192 relocs are also examined. TOC words marked as both needed and
7193 unneeded will of course be kept. TOC words without any referencing
7194 reloc are assumed to be part of a multi-word entry, and are kept or
7195 discarded as per the nearest marked preceding word. This works
7196 reliably for compiler generated code, but may be incorrect if assembly
7197 code is used to insert TOC entries. Use this option to disable the
7200 @cindex PowerPC64 multi-TOC
7201 @kindex --no-multi-toc
7202 @item --no-multi-toc
7203 If given any toc option besides @code{-mcmodel=medium} or
7204 @code{-mcmodel=large}, PowerPC64 GCC generates code for a TOC model
7206 entries are accessed with a 16-bit offset from r2. This limits the
7207 total TOC size to 64K. PowerPC64 @command{ld} extends this limit by
7208 grouping code sections such that each group uses less than 64K for its
7209 TOC entries, then inserts r2 adjusting stubs between inter-group
7210 calls. @command{ld} does not split apart input sections, so cannot
7211 help if a single input file has a @code{.toc} section that exceeds
7212 64K, most likely from linking multiple files with @command{ld -r}.
7213 Use this option to turn off this feature.
7215 @cindex PowerPC64 TOC sorting
7216 @kindex --no-toc-sort
7218 By default, @command{ld} sorts TOC sections so that those whose file
7219 happens to have a section called @code{.init} or @code{.fini} are
7220 placed first, followed by TOC sections referenced by code generated
7221 with PowerPC64 gcc's @code{-mcmodel=small}, and lastly TOC sections
7222 referenced only by code generated with PowerPC64 gcc's
7223 @code{-mcmodel=medium} or @code{-mcmodel=large} options. Doing this
7224 results in better TOC grouping for multi-TOC. Use this option to turn
7227 @cindex PowerPC64 PLT stub alignment
7229 @kindex --no-plt-align
7231 @itemx --no-plt-align
7232 Use these options to control whether individual PLT call stubs are
7233 padded so that they don't cross a 32-byte boundary, or to the
7234 specified power of two boundary when using @code{--plt-align=}. Note
7235 that this isn't alignment in the usual sense. By default PLT call
7236 stubs are packed tightly.
7238 @cindex PowerPC64 PLT call stub static chain
7239 @kindex --plt-static-chain
7240 @kindex --no-plt-static-chain
7241 @item --plt-static-chain
7242 @itemx --no-plt-static-chain
7243 Use these options to control whether PLT call stubs load the static
7244 chain pointer (r11). @code{ld} defaults to not loading the static
7245 chain since there is never any need to do so on a PLT call.
7247 @cindex PowerPC64 PLT call stub thread safety
7248 @kindex --plt-thread-safe
7249 @kindex --no-plt-thread-safe
7250 @item --plt-thread-safe
7251 @itemx --no-thread-safe
7252 With power7's weakly ordered memory model, it is possible when using
7253 lazy binding for ld.so to update a plt entry in one thread and have
7254 another thread see the individual plt entry words update in the wrong
7255 order, despite ld.so carefully writing in the correct order and using
7256 memory write barriers. To avoid this we need some sort of read
7257 barrier in the call stub, or use LD_BIND_NOW=1. By default, @code{ld}
7258 looks for calls to commonly used functions that create threads, and if
7259 seen, adds the necessary barriers. Use these options to change the
7274 @section @command{ld} and SPU ELF Support
7276 @cindex SPU ELF options
7282 This option marks an executable as a PIC plugin module.
7284 @cindex SPU overlays
7285 @kindex --no-overlays
7287 Normally, @command{ld} recognizes calls to functions within overlay
7288 regions, and redirects such calls to an overlay manager via a stub.
7289 @command{ld} also provides a built-in overlay manager. This option
7290 turns off all this special overlay handling.
7292 @cindex SPU overlay stub symbols
7293 @kindex --emit-stub-syms
7294 @item --emit-stub-syms
7295 This option causes @command{ld} to label overlay stubs with a local
7296 symbol that encodes the stub type and destination.
7298 @cindex SPU extra overlay stubs
7299 @kindex --extra-overlay-stubs
7300 @item --extra-overlay-stubs
7301 This option causes @command{ld} to add overlay call stubs on all
7302 function calls out of overlay regions. Normally stubs are not added
7303 on calls to non-overlay regions.
7305 @cindex SPU local store size
7306 @kindex --local-store=lo:hi
7307 @item --local-store=lo:hi
7308 @command{ld} usually checks that a final executable for SPU fits in
7309 the address range 0 to 256k. This option may be used to change the
7310 range. Disable the check entirely with @option{--local-store=0:0}.
7313 @kindex --stack-analysis
7314 @item --stack-analysis
7315 SPU local store space is limited. Over-allocation of stack space
7316 unnecessarily limits space available for code and data, while
7317 under-allocation results in runtime failures. If given this option,
7318 @command{ld} will provide an estimate of maximum stack usage.
7319 @command{ld} does this by examining symbols in code sections to
7320 determine the extents of functions, and looking at function prologues
7321 for stack adjusting instructions. A call-graph is created by looking
7322 for relocations on branch instructions. The graph is then searched
7323 for the maximum stack usage path. Note that this analysis does not
7324 find calls made via function pointers, and does not handle recursion
7325 and other cycles in the call graph. Stack usage may be
7326 under-estimated if your code makes such calls. Also, stack usage for
7327 dynamic allocation, e.g. alloca, will not be detected. If a link map
7328 is requested, detailed information about each function's stack usage
7329 and calls will be given.
7332 @kindex --emit-stack-syms
7333 @item --emit-stack-syms
7334 This option, if given along with @option{--stack-analysis} will result
7335 in @command{ld} emitting stack sizing symbols for each function.
7336 These take the form @code{__stack_<function_name>} for global
7337 functions, and @code{__stack_<number>_<function_name>} for static
7338 functions. @code{<number>} is the section id in hex. The value of
7339 such symbols is the stack requirement for the corresponding function.
7340 The symbol size will be zero, type @code{STT_NOTYPE}, binding
7341 @code{STB_LOCAL}, and section @code{SHN_ABS}.
7355 @section @command{ld}'s Support for Various TI COFF Versions
7356 @cindex TI COFF versions
7357 @kindex --format=@var{version}
7358 The @samp{--format} switch allows selection of one of the various
7359 TI COFF versions. The latest of this writing is 2; versions 0 and 1 are
7360 also supported. The TI COFF versions also vary in header byte-order
7361 format; @command{ld} will read any version or byte order, but the output
7362 header format depends on the default specified by the specific target.
7375 @section @command{ld} and WIN32 (cygwin/mingw)
7377 This section describes some of the win32 specific @command{ld} issues.
7378 See @ref{Options,,Command Line Options} for detailed description of the
7379 command line options mentioned here.
7382 @cindex import libraries
7383 @item import libraries
7384 The standard Windows linker creates and uses so-called import
7385 libraries, which contains information for linking to dll's. They are
7386 regular static archives and are handled as any other static
7387 archive. The cygwin and mingw ports of @command{ld} have specific
7388 support for creating such libraries provided with the
7389 @samp{--out-implib} command line option.
7391 @item exporting DLL symbols
7392 @cindex exporting DLL symbols
7393 The cygwin/mingw @command{ld} has several ways to export symbols for dll's.
7396 @item using auto-export functionality
7397 @cindex using auto-export functionality
7398 By default @command{ld} exports symbols with the auto-export functionality,
7399 which is controlled by the following command line options:
7402 @item --export-all-symbols [This is the default]
7403 @item --exclude-symbols
7404 @item --exclude-libs
7405 @item --exclude-modules-for-implib
7406 @item --version-script
7409 When auto-export is in operation, @command{ld} will export all the non-local
7410 (global and common) symbols it finds in a DLL, with the exception of a few
7411 symbols known to belong to the system's runtime and libraries. As it will
7412 often not be desirable to export all of a DLL's symbols, which may include
7413 private functions that are not part of any public interface, the command-line
7414 options listed above may be used to filter symbols out from the list for
7415 exporting. The @samp{--output-def} option can be used in order to see the
7416 final list of exported symbols with all exclusions taken into effect.
7418 If @samp{--export-all-symbols} is not given explicitly on the
7419 command line, then the default auto-export behavior will be @emph{disabled}
7420 if either of the following are true:
7423 @item A DEF file is used.
7424 @item Any symbol in any object file was marked with the __declspec(dllexport) attribute.
7427 @item using a DEF file
7428 @cindex using a DEF file
7429 Another way of exporting symbols is using a DEF file. A DEF file is
7430 an ASCII file containing definitions of symbols which should be
7431 exported when a dll is created. Usually it is named @samp{<dll
7432 name>.def} and is added as any other object file to the linker's
7433 command line. The file's name must end in @samp{.def} or @samp{.DEF}.
7436 gcc -o <output> <objectfiles> <dll name>.def
7439 Using a DEF file turns off the normal auto-export behavior, unless the
7440 @samp{--export-all-symbols} option is also used.
7442 Here is an example of a DEF file for a shared library called @samp{xyz.dll}:
7445 LIBRARY "xyz.dll" BASE=0x20000000
7451 another_foo = abc.dll.afoo
7457 This example defines a DLL with a non-default base address and seven
7458 symbols in the export table. The third exported symbol @code{_bar} is an
7459 alias for the second. The fourth symbol, @code{another_foo} is resolved
7460 by "forwarding" to another module and treating it as an alias for
7461 @code{afoo} exported from the DLL @samp{abc.dll}. The final symbol
7462 @code{var1} is declared to be a data object. The @samp{doo} symbol in
7463 export library is an alias of @samp{foo}, which gets the string name
7464 in export table @samp{foo2}. The @samp{eoo} symbol is an data export
7465 symbol, which gets in export table the name @samp{var1}.
7467 The optional @code{LIBRARY <name>} command indicates the @emph{internal}
7468 name of the output DLL. If @samp{<name>} does not include a suffix,
7469 the default library suffix, @samp{.DLL} is appended.
7471 When the .DEF file is used to build an application, rather than a
7472 library, the @code{NAME <name>} command should be used instead of
7473 @code{LIBRARY}. If @samp{<name>} does not include a suffix, the default
7474 executable suffix, @samp{.EXE} is appended.
7476 With either @code{LIBRARY <name>} or @code{NAME <name>} the optional
7477 specification @code{BASE = <number>} may be used to specify a
7478 non-default base address for the image.
7480 If neither @code{LIBRARY <name>} nor @code{NAME <name>} is specified,
7481 or they specify an empty string, the internal name is the same as the
7482 filename specified on the command line.
7484 The complete specification of an export symbol is:
7488 ( ( ( <name1> [ = <name2> ] )
7489 | ( <name1> = <module-name> . <external-name>))
7490 [ @@ <integer> ] [NONAME] [DATA] [CONSTANT] [PRIVATE] [== <name3>] ) *
7493 Declares @samp{<name1>} as an exported symbol from the DLL, or declares
7494 @samp{<name1>} as an exported alias for @samp{<name2>}; or declares
7495 @samp{<name1>} as a "forward" alias for the symbol
7496 @samp{<external-name>} in the DLL @samp{<module-name>}.
7497 Optionally, the symbol may be exported by the specified ordinal
7498 @samp{<integer>} alias. The optional @samp{<name3>} is the to be used
7499 string in import/export table for the symbol.
7501 The optional keywords that follow the declaration indicate:
7503 @code{NONAME}: Do not put the symbol name in the DLL's export table. It
7504 will still be exported by its ordinal alias (either the value specified
7505 by the .def specification or, otherwise, the value assigned by the
7506 linker). The symbol name, however, does remain visible in the import
7507 library (if any), unless @code{PRIVATE} is also specified.
7509 @code{DATA}: The symbol is a variable or object, rather than a function.
7510 The import lib will export only an indirect reference to @code{foo} as
7511 the symbol @code{_imp__foo} (ie, @code{foo} must be resolved as
7514 @code{CONSTANT}: Like @code{DATA}, but put the undecorated @code{foo} as
7515 well as @code{_imp__foo} into the import library. Both refer to the
7516 read-only import address table's pointer to the variable, not to the
7517 variable itself. This can be dangerous. If the user code fails to add
7518 the @code{dllimport} attribute and also fails to explicitly add the
7519 extra indirection that the use of the attribute enforces, the
7520 application will behave unexpectedly.
7522 @code{PRIVATE}: Put the symbol in the DLL's export table, but do not put
7523 it into the static import library used to resolve imports at link time. The
7524 symbol can still be imported using the @code{LoadLibrary/GetProcAddress}
7525 API at runtime or by by using the GNU ld extension of linking directly to
7526 the DLL without an import library.
7528 See ld/deffilep.y in the binutils sources for the full specification of
7529 other DEF file statements
7531 @cindex creating a DEF file
7532 While linking a shared dll, @command{ld} is able to create a DEF file
7533 with the @samp{--output-def <file>} command line option.
7535 @item Using decorations
7536 @cindex Using decorations
7537 Another way of marking symbols for export is to modify the source code
7538 itself, so that when building the DLL each symbol to be exported is
7542 __declspec(dllexport) int a_variable
7543 __declspec(dllexport) void a_function(int with_args)
7546 All such symbols will be exported from the DLL. If, however,
7547 any of the object files in the DLL contain symbols decorated in
7548 this way, then the normal auto-export behavior is disabled, unless
7549 the @samp{--export-all-symbols} option is also used.
7551 Note that object files that wish to access these symbols must @emph{not}
7552 decorate them with dllexport. Instead, they should use dllimport,
7556 __declspec(dllimport) int a_variable
7557 __declspec(dllimport) void a_function(int with_args)
7560 This complicates the structure of library header files, because
7561 when included by the library itself the header must declare the
7562 variables and functions as dllexport, but when included by client
7563 code the header must declare them as dllimport. There are a number
7564 of idioms that are typically used to do this; often client code can
7565 omit the __declspec() declaration completely. See
7566 @samp{--enable-auto-import} and @samp{automatic data imports} for more
7570 @cindex automatic data imports
7571 @item automatic data imports
7572 The standard Windows dll format supports data imports from dlls only
7573 by adding special decorations (dllimport/dllexport), which let the
7574 compiler produce specific assembler instructions to deal with this
7575 issue. This increases the effort necessary to port existing Un*x
7576 code to these platforms, especially for large
7577 c++ libraries and applications. The auto-import feature, which was
7578 initially provided by Paul Sokolovsky, allows one to omit the
7579 decorations to achieve a behavior that conforms to that on POSIX/Un*x
7580 platforms. This feature is enabled with the @samp{--enable-auto-import}
7581 command-line option, although it is enabled by default on cygwin/mingw.
7582 The @samp{--enable-auto-import} option itself now serves mainly to
7583 suppress any warnings that are ordinarily emitted when linked objects
7584 trigger the feature's use.
7586 auto-import of variables does not always work flawlessly without
7587 additional assistance. Sometimes, you will see this message
7589 "variable '<var>' can't be auto-imported. Please read the
7590 documentation for ld's @code{--enable-auto-import} for details."
7592 The @samp{--enable-auto-import} documentation explains why this error
7593 occurs, and several methods that can be used to overcome this difficulty.
7594 One of these methods is the @emph{runtime pseudo-relocs} feature, described
7597 @cindex runtime pseudo-relocation
7598 For complex variables imported from DLLs (such as structs or classes),
7599 object files typically contain a base address for the variable and an
7600 offset (@emph{addend}) within the variable--to specify a particular
7601 field or public member, for instance. Unfortunately, the runtime loader used
7602 in win32 environments is incapable of fixing these references at runtime
7603 without the additional information supplied by dllimport/dllexport decorations.
7604 The standard auto-import feature described above is unable to resolve these
7607 The @samp{--enable-runtime-pseudo-relocs} switch allows these references to
7608 be resolved without error, while leaving the task of adjusting the references
7609 themselves (with their non-zero addends) to specialized code provided by the
7610 runtime environment. Recent versions of the cygwin and mingw environments and
7611 compilers provide this runtime support; older versions do not. However, the
7612 support is only necessary on the developer's platform; the compiled result will
7613 run without error on an older system.
7615 @samp{--enable-runtime-pseudo-relocs} is not the default; it must be explicitly
7618 @cindex direct linking to a dll
7619 @item direct linking to a dll
7620 The cygwin/mingw ports of @command{ld} support the direct linking,
7621 including data symbols, to a dll without the usage of any import
7622 libraries. This is much faster and uses much less memory than does the
7623 traditional import library method, especially when linking large
7624 libraries or applications. When @command{ld} creates an import lib, each
7625 function or variable exported from the dll is stored in its own bfd, even
7626 though a single bfd could contain many exports. The overhead involved in
7627 storing, loading, and processing so many bfd's is quite large, and explains the
7628 tremendous time, memory, and storage needed to link against particularly
7629 large or complex libraries when using import libs.
7631 Linking directly to a dll uses no extra command-line switches other than
7632 @samp{-L} and @samp{-l}, because @command{ld} already searches for a number
7633 of names to match each library. All that is needed from the developer's
7634 perspective is an understanding of this search, in order to force ld to
7635 select the dll instead of an import library.
7638 For instance, when ld is called with the argument @samp{-lxxx} it will attempt
7639 to find, in the first directory of its search path,
7651 before moving on to the next directory in the search path.
7653 (*) Actually, this is not @samp{cygxxx.dll} but in fact is @samp{<prefix>xxx.dll},
7654 where @samp{<prefix>} is set by the @command{ld} option
7655 @samp{--dll-search-prefix=<prefix>}. In the case of cygwin, the standard gcc spec
7656 file includes @samp{--dll-search-prefix=cyg}, so in effect we actually search for
7659 Other win32-based unix environments, such as mingw or pw32, may use other
7660 @samp{<prefix>}es, although at present only cygwin makes use of this feature. It
7661 was originally intended to help avoid name conflicts among dll's built for the
7662 various win32/un*x environments, so that (for example) two versions of a zlib dll
7663 could coexist on the same machine.
7665 The generic cygwin/mingw path layout uses a @samp{bin} directory for
7666 applications and dll's and a @samp{lib} directory for the import
7667 libraries (using cygwin nomenclature):
7673 libxxx.dll.a (in case of dll's)
7674 libxxx.a (in case of static archive)
7677 Linking directly to a dll without using the import library can be
7680 1. Use the dll directly by adding the @samp{bin} path to the link line
7682 gcc -Wl,-verbose -o a.exe -L../bin/ -lxxx
7685 However, as the dll's often have version numbers appended to their names
7686 (@samp{cygncurses-5.dll}) this will often fail, unless one specifies
7687 @samp{-L../bin -lncurses-5} to include the version. Import libs are generally
7688 not versioned, and do not have this difficulty.
7690 2. Create a symbolic link from the dll to a file in the @samp{lib}
7691 directory according to the above mentioned search pattern. This
7692 should be used to avoid unwanted changes in the tools needed for
7696 ln -s bin/cygxxx.dll lib/[cyg|lib|]xxx.dll[.a]
7699 Then you can link without any make environment changes.
7702 gcc -Wl,-verbose -o a.exe -L../lib/ -lxxx
7705 This technique also avoids the version number problems, because the following is
7712 libxxx.dll.a -> ../bin/cygxxx-5.dll
7715 Linking directly to a dll without using an import lib will work
7716 even when auto-import features are exercised, and even when
7717 @samp{--enable-runtime-pseudo-relocs} is used.
7719 Given the improvements in speed and memory usage, one might justifiably
7720 wonder why import libraries are used at all. There are three reasons:
7722 1. Until recently, the link-directly-to-dll functionality did @emph{not}
7723 work with auto-imported data.
7725 2. Sometimes it is necessary to include pure static objects within the
7726 import library (which otherwise contains only bfd's for indirection
7727 symbols that point to the exports of a dll). Again, the import lib
7728 for the cygwin kernel makes use of this ability, and it is not
7729 possible to do this without an import lib.
7731 3. Symbol aliases can only be resolved using an import lib. This is
7732 critical when linking against OS-supplied dll's (eg, the win32 API)
7733 in which symbols are usually exported as undecorated aliases of their
7734 stdcall-decorated assembly names.
7736 So, import libs are not going away. But the ability to replace
7737 true import libs with a simple symbolic link to (or a copy of)
7738 a dll, in many cases, is a useful addition to the suite of tools
7739 binutils makes available to the win32 developer. Given the
7740 massive improvements in memory requirements during linking, storage
7741 requirements, and linking speed, we expect that many developers
7742 will soon begin to use this feature whenever possible.
7744 @item symbol aliasing
7746 @item adding additional names
7747 Sometimes, it is useful to export symbols with additional names.
7748 A symbol @samp{foo} will be exported as @samp{foo}, but it can also be
7749 exported as @samp{_foo} by using special directives in the DEF file
7750 when creating the dll. This will affect also the optional created
7751 import library. Consider the following DEF file:
7754 LIBRARY "xyz.dll" BASE=0x61000000
7761 The line @samp{_foo = foo} maps the symbol @samp{foo} to @samp{_foo}.
7763 Another method for creating a symbol alias is to create it in the
7764 source code using the "weak" attribute:
7767 void foo () @{ /* Do something. */; @}
7768 void _foo () __attribute__ ((weak, alias ("foo")));
7771 See the gcc manual for more information about attributes and weak
7774 @item renaming symbols
7775 Sometimes it is useful to rename exports. For instance, the cygwin
7776 kernel does this regularly. A symbol @samp{_foo} can be exported as
7777 @samp{foo} but not as @samp{_foo} by using special directives in the
7778 DEF file. (This will also affect the import library, if it is
7779 created). In the following example:
7782 LIBRARY "xyz.dll" BASE=0x61000000
7788 The line @samp{_foo = foo} maps the exported symbol @samp{foo} to
7792 Note: using a DEF file disables the default auto-export behavior,
7793 unless the @samp{--export-all-symbols} command line option is used.
7794 If, however, you are trying to rename symbols, then you should list
7795 @emph{all} desired exports in the DEF file, including the symbols
7796 that are not being renamed, and do @emph{not} use the
7797 @samp{--export-all-symbols} option. If you list only the
7798 renamed symbols in the DEF file, and use @samp{--export-all-symbols}
7799 to handle the other symbols, then the both the new names @emph{and}
7800 the original names for the renamed symbols will be exported.
7801 In effect, you'd be aliasing those symbols, not renaming them,
7802 which is probably not what you wanted.
7804 @cindex weak externals
7805 @item weak externals
7806 The Windows object format, PE, specifies a form of weak symbols called
7807 weak externals. When a weak symbol is linked and the symbol is not
7808 defined, the weak symbol becomes an alias for some other symbol. There
7809 are three variants of weak externals:
7811 @item Definition is searched for in objects and libraries, historically
7812 called lazy externals.
7813 @item Definition is searched for only in other objects, not in libraries.
7814 This form is not presently implemented.
7815 @item No search; the symbol is an alias. This form is not presently
7818 As a GNU extension, weak symbols that do not specify an alternate symbol
7819 are supported. If the symbol is undefined when linking, the symbol
7820 uses a default value.
7822 @cindex aligned common symbols
7823 @item aligned common symbols
7824 As a GNU extension to the PE file format, it is possible to specify the
7825 desired alignment for a common symbol. This information is conveyed from
7826 the assembler or compiler to the linker by means of GNU-specific commands
7827 carried in the object file's @samp{.drectve} section, which are recognized
7828 by @command{ld} and respected when laying out the common symbols. Native
7829 tools will be able to process object files employing this GNU extension,
7830 but will fail to respect the alignment instructions, and may issue noisy
7831 warnings about unknown linker directives.
7846 @section @code{ld} and Xtensa Processors
7848 @cindex Xtensa processors
7849 The default @command{ld} behavior for Xtensa processors is to interpret
7850 @code{SECTIONS} commands so that lists of explicitly named sections in a
7851 specification with a wildcard file will be interleaved when necessary to
7852 keep literal pools within the range of PC-relative load offsets. For
7853 example, with the command:
7865 @command{ld} may interleave some of the @code{.literal}
7866 and @code{.text} sections from different object files to ensure that the
7867 literal pools are within the range of PC-relative load offsets. A valid
7868 interleaving might place the @code{.literal} sections from an initial
7869 group of files followed by the @code{.text} sections of that group of
7870 files. Then, the @code{.literal} sections from the rest of the files
7871 and the @code{.text} sections from the rest of the files would follow.
7873 @cindex @option{--relax} on Xtensa
7874 @cindex relaxing on Xtensa
7875 Relaxation is enabled by default for the Xtensa version of @command{ld} and
7876 provides two important link-time optimizations. The first optimization
7877 is to combine identical literal values to reduce code size. A redundant
7878 literal will be removed and all the @code{L32R} instructions that use it
7879 will be changed to reference an identical literal, as long as the
7880 location of the replacement literal is within the offset range of all
7881 the @code{L32R} instructions. The second optimization is to remove
7882 unnecessary overhead from assembler-generated ``longcall'' sequences of
7883 @code{L32R}/@code{CALLX@var{n}} when the target functions are within
7884 range of direct @code{CALL@var{n}} instructions.
7886 For each of these cases where an indirect call sequence can be optimized
7887 to a direct call, the linker will change the @code{CALLX@var{n}}
7888 instruction to a @code{CALL@var{n}} instruction, remove the @code{L32R}
7889 instruction, and remove the literal referenced by the @code{L32R}
7890 instruction if it is not used for anything else. Removing the
7891 @code{L32R} instruction always reduces code size but can potentially
7892 hurt performance by changing the alignment of subsequent branch targets.
7893 By default, the linker will always preserve alignments, either by
7894 switching some instructions between 24-bit encodings and the equivalent
7895 density instructions or by inserting a no-op in place of the @code{L32R}
7896 instruction that was removed. If code size is more important than
7897 performance, the @option{--size-opt} option can be used to prevent the
7898 linker from widening density instructions or inserting no-ops, except in
7899 a few cases where no-ops are required for correctness.
7901 The following Xtensa-specific command-line options can be used to
7904 @cindex Xtensa options
7907 When optimizing indirect calls to direct calls, optimize for code size
7908 more than performance. With this option, the linker will not insert
7909 no-ops or widen density instructions to preserve branch target
7910 alignment. There may still be some cases where no-ops are required to
7911 preserve the correctness of the code.
7919 @ifclear SingleFormat
7924 @cindex object file management
7925 @cindex object formats available
7927 The linker accesses object and archive files using the BFD libraries.
7928 These libraries allow the linker to use the same routines to operate on
7929 object files whatever the object file format. A different object file
7930 format can be supported simply by creating a new BFD back end and adding
7931 it to the library. To conserve runtime memory, however, the linker and
7932 associated tools are usually configured to support only a subset of the
7933 object file formats available. You can use @code{objdump -i}
7934 (@pxref{objdump,,objdump,binutils.info,The GNU Binary Utilities}) to
7935 list all the formats available for your configuration.
7937 @cindex BFD requirements
7938 @cindex requirements for BFD
7939 As with most implementations, BFD is a compromise between
7940 several conflicting requirements. The major factor influencing
7941 BFD design was efficiency: any time used converting between
7942 formats is time which would not have been spent had BFD not
7943 been involved. This is partly offset by abstraction payback; since
7944 BFD simplifies applications and back ends, more time and care
7945 may be spent optimizing algorithms for a greater speed.
7947 One minor artifact of the BFD solution which you should bear in
7948 mind is the potential for information loss. There are two places where
7949 useful information can be lost using the BFD mechanism: during
7950 conversion and during output. @xref{BFD information loss}.
7953 * BFD outline:: How it works: an outline of BFD
7957 @section How It Works: An Outline of BFD
7958 @cindex opening object files
7959 @include bfdsumm.texi
7962 @node Reporting Bugs
7963 @chapter Reporting Bugs
7964 @cindex bugs in @command{ld}
7965 @cindex reporting bugs in @command{ld}
7967 Your bug reports play an essential role in making @command{ld} reliable.
7969 Reporting a bug may help you by bringing a solution to your problem, or
7970 it may not. But in any case the principal function of a bug report is
7971 to help the entire community by making the next version of @command{ld}
7972 work better. Bug reports are your contribution to the maintenance of
7975 In order for a bug report to serve its purpose, you must include the
7976 information that enables us to fix the bug.
7979 * Bug Criteria:: Have you found a bug?
7980 * Bug Reporting:: How to report bugs
7984 @section Have You Found a Bug?
7985 @cindex bug criteria
7987 If you are not sure whether you have found a bug, here are some guidelines:
7990 @cindex fatal signal
7991 @cindex linker crash
7992 @cindex crash of linker
7994 If the linker gets a fatal signal, for any input whatever, that is a
7995 @command{ld} bug. Reliable linkers never crash.
7997 @cindex error on valid input
7999 If @command{ld} produces an error message for valid input, that is a bug.
8001 @cindex invalid input
8003 If @command{ld} does not produce an error message for invalid input, that
8004 may be a bug. In the general case, the linker can not verify that
8005 object files are correct.
8008 If you are an experienced user of linkers, your suggestions for
8009 improvement of @command{ld} are welcome in any case.
8013 @section How to Report Bugs
8015 @cindex @command{ld} bugs, reporting
8017 A number of companies and individuals offer support for @sc{gnu}
8018 products. If you obtained @command{ld} from a support organization, we
8019 recommend you contact that organization first.
8021 You can find contact information for many support companies and
8022 individuals in the file @file{etc/SERVICE} in the @sc{gnu} Emacs
8026 Otherwise, send bug reports for @command{ld} to
8030 The fundamental principle of reporting bugs usefully is this:
8031 @strong{report all the facts}. If you are not sure whether to state a
8032 fact or leave it out, state it!
8034 Often people omit facts because they think they know what causes the
8035 problem and assume that some details do not matter. Thus, you might
8036 assume that the name of a symbol you use in an example does not
8037 matter. Well, probably it does not, but one cannot be sure. Perhaps
8038 the bug is a stray memory reference which happens to fetch from the
8039 location where that name is stored in memory; perhaps, if the name
8040 were different, the contents of that location would fool the linker
8041 into doing the right thing despite the bug. Play it safe and give a
8042 specific, complete example. That is the easiest thing for you to do,
8043 and the most helpful.
8045 Keep in mind that the purpose of a bug report is to enable us to fix
8046 the bug if it is new to us. Therefore, always write your bug reports
8047 on the assumption that the bug has not been reported previously.
8049 Sometimes people give a few sketchy facts and ask, ``Does this ring a
8050 bell?'' This cannot help us fix a bug, so it is basically useless. We
8051 respond by asking for enough details to enable us to investigate.
8052 You might as well expedite matters by sending them to begin with.
8054 To enable us to fix the bug, you should include all these things:
8058 The version of @command{ld}. @command{ld} announces it if you start it with
8059 the @samp{--version} argument.
8061 Without this, we will not know whether there is any point in looking for
8062 the bug in the current version of @command{ld}.
8065 Any patches you may have applied to the @command{ld} source, including any
8066 patches made to the @code{BFD} library.
8069 The type of machine you are using, and the operating system name and
8073 What compiler (and its version) was used to compile @command{ld}---e.g.
8077 The command arguments you gave the linker to link your example and
8078 observe the bug. To guarantee you will not omit something important,
8079 list them all. A copy of the Makefile (or the output from make) is
8082 If we were to try to guess the arguments, we would probably guess wrong
8083 and then we might not encounter the bug.
8086 A complete input file, or set of input files, that will reproduce the
8087 bug. It is generally most helpful to send the actual object files
8088 provided that they are reasonably small. Say no more than 10K. For
8089 bigger files you can either make them available by FTP or HTTP or else
8090 state that you are willing to send the object file(s) to whomever
8091 requests them. (Note - your email will be going to a mailing list, so
8092 we do not want to clog it up with large attachments). But small
8093 attachments are best.
8095 If the source files were assembled using @code{gas} or compiled using
8096 @code{gcc}, then it may be OK to send the source files rather than the
8097 object files. In this case, be sure to say exactly what version of
8098 @code{gas} or @code{gcc} was used to produce the object files. Also say
8099 how @code{gas} or @code{gcc} were configured.
8102 A description of what behavior you observe that you believe is
8103 incorrect. For example, ``It gets a fatal signal.''
8105 Of course, if the bug is that @command{ld} gets a fatal signal, then we
8106 will certainly notice it. But if the bug is incorrect output, we might
8107 not notice unless it is glaringly wrong. You might as well not give us
8108 a chance to make a mistake.
8110 Even if the problem you experience is a fatal signal, you should still
8111 say so explicitly. Suppose something strange is going on, such as, your
8112 copy of @command{ld} is out of sync, or you have encountered a bug in the
8113 C library on your system. (This has happened!) Your copy might crash
8114 and ours would not. If you told us to expect a crash, then when ours
8115 fails to crash, we would know that the bug was not happening for us. If
8116 you had not told us to expect a crash, then we would not be able to draw
8117 any conclusion from our observations.
8120 If you wish to suggest changes to the @command{ld} source, send us context
8121 diffs, as generated by @code{diff} with the @samp{-u}, @samp{-c}, or
8122 @samp{-p} option. Always send diffs from the old file to the new file.
8123 If you even discuss something in the @command{ld} source, refer to it by
8124 context, not by line number.
8126 The line numbers in our development sources will not match those in your
8127 sources. Your line numbers would convey no useful information to us.
8130 Here are some things that are not necessary:
8134 A description of the envelope of the bug.
8136 Often people who encounter a bug spend a lot of time investigating
8137 which changes to the input file will make the bug go away and which
8138 changes will not affect it.
8140 This is often time consuming and not very useful, because the way we
8141 will find the bug is by running a single example under the debugger
8142 with breakpoints, not by pure deduction from a series of examples.
8143 We recommend that you save your time for something else.
8145 Of course, if you can find a simpler example to report @emph{instead}
8146 of the original one, that is a convenience for us. Errors in the
8147 output will be easier to spot, running under the debugger will take
8148 less time, and so on.
8150 However, simplification is not vital; if you do not want to do this,
8151 report the bug anyway and send us the entire test case you used.
8154 A patch for the bug.
8156 A patch for the bug does help us if it is a good one. But do not omit
8157 the necessary information, such as the test case, on the assumption that
8158 a patch is all we need. We might see problems with your patch and decide
8159 to fix the problem another way, or we might not understand it at all.
8161 Sometimes with a program as complicated as @command{ld} it is very hard to
8162 construct an example that will make the program follow a certain path
8163 through the code. If you do not send us the example, we will not be
8164 able to construct one, so we will not be able to verify that the bug is
8167 And if we cannot understand what bug you are trying to fix, or why your
8168 patch should be an improvement, we will not install it. A test case will
8169 help us to understand.
8172 A guess about what the bug is or what it depends on.
8174 Such guesses are usually wrong. Even we cannot guess right about such
8175 things without first using the debugger to find the facts.
8179 @appendix MRI Compatible Script Files
8180 @cindex MRI compatibility
8181 To aid users making the transition to @sc{gnu} @command{ld} from the MRI
8182 linker, @command{ld} can use MRI compatible linker scripts as an
8183 alternative to the more general-purpose linker scripting language
8184 described in @ref{Scripts}. MRI compatible linker scripts have a much
8185 simpler command set than the scripting language otherwise used with
8186 @command{ld}. @sc{gnu} @command{ld} supports the most commonly used MRI
8187 linker commands; these commands are described here.
8189 In general, MRI scripts aren't of much use with the @code{a.out} object
8190 file format, since it only has three sections and MRI scripts lack some
8191 features to make use of them.
8193 You can specify a file containing an MRI-compatible script using the
8194 @samp{-c} command-line option.
8196 Each command in an MRI-compatible script occupies its own line; each
8197 command line starts with the keyword that identifies the command (though
8198 blank lines are also allowed for punctuation). If a line of an
8199 MRI-compatible script begins with an unrecognized keyword, @command{ld}
8200 issues a warning message, but continues processing the script.
8202 Lines beginning with @samp{*} are comments.
8204 You can write these commands using all upper-case letters, or all
8205 lower case; for example, @samp{chip} is the same as @samp{CHIP}.
8206 The following list shows only the upper-case form of each command.
8209 @cindex @code{ABSOLUTE} (MRI)
8210 @item ABSOLUTE @var{secname}
8211 @itemx ABSOLUTE @var{secname}, @var{secname}, @dots{} @var{secname}
8212 Normally, @command{ld} includes in the output file all sections from all
8213 the input files. However, in an MRI-compatible script, you can use the
8214 @code{ABSOLUTE} command to restrict the sections that will be present in
8215 your output program. If the @code{ABSOLUTE} command is used at all in a
8216 script, then only the sections named explicitly in @code{ABSOLUTE}
8217 commands will appear in the linker output. You can still use other
8218 input sections (whatever you select on the command line, or using
8219 @code{LOAD}) to resolve addresses in the output file.
8221 @cindex @code{ALIAS} (MRI)
8222 @item ALIAS @var{out-secname}, @var{in-secname}
8223 Use this command to place the data from input section @var{in-secname}
8224 in a section called @var{out-secname} in the linker output file.
8226 @var{in-secname} may be an integer.
8228 @cindex @code{ALIGN} (MRI)
8229 @item ALIGN @var{secname} = @var{expression}
8230 Align the section called @var{secname} to @var{expression}. The
8231 @var{expression} should be a power of two.
8233 @cindex @code{BASE} (MRI)
8234 @item BASE @var{expression}
8235 Use the value of @var{expression} as the lowest address (other than
8236 absolute addresses) in the output file.
8238 @cindex @code{CHIP} (MRI)
8239 @item CHIP @var{expression}
8240 @itemx CHIP @var{expression}, @var{expression}
8241 This command does nothing; it is accepted only for compatibility.
8243 @cindex @code{END} (MRI)
8245 This command does nothing whatever; it's only accepted for compatibility.
8247 @cindex @code{FORMAT} (MRI)
8248 @item FORMAT @var{output-format}
8249 Similar to the @code{OUTPUT_FORMAT} command in the more general linker
8250 language, but restricted to one of these output formats:
8254 S-records, if @var{output-format} is @samp{S}
8257 IEEE, if @var{output-format} is @samp{IEEE}
8260 COFF (the @samp{coff-m68k} variant in BFD), if @var{output-format} is
8264 @cindex @code{LIST} (MRI)
8265 @item LIST @var{anything}@dots{}
8266 Print (to the standard output file) a link map, as produced by the
8267 @command{ld} command-line option @samp{-M}.
8269 The keyword @code{LIST} may be followed by anything on the
8270 same line, with no change in its effect.
8272 @cindex @code{LOAD} (MRI)
8273 @item LOAD @var{filename}
8274 @itemx LOAD @var{filename}, @var{filename}, @dots{} @var{filename}
8275 Include one or more object file @var{filename} in the link; this has the
8276 same effect as specifying @var{filename} directly on the @command{ld}
8279 @cindex @code{NAME} (MRI)
8280 @item NAME @var{output-name}
8281 @var{output-name} is the name for the program produced by @command{ld}; the
8282 MRI-compatible command @code{NAME} is equivalent to the command-line
8283 option @samp{-o} or the general script language command @code{OUTPUT}.
8285 @cindex @code{ORDER} (MRI)
8286 @item ORDER @var{secname}, @var{secname}, @dots{} @var{secname}
8287 @itemx ORDER @var{secname} @var{secname} @var{secname}
8288 Normally, @command{ld} orders the sections in its output file in the
8289 order in which they first appear in the input files. In an MRI-compatible
8290 script, you can override this ordering with the @code{ORDER} command. The
8291 sections you list with @code{ORDER} will appear first in your output
8292 file, in the order specified.
8294 @cindex @code{PUBLIC} (MRI)
8295 @item PUBLIC @var{name}=@var{expression}
8296 @itemx PUBLIC @var{name},@var{expression}
8297 @itemx PUBLIC @var{name} @var{expression}
8298 Supply a value (@var{expression}) for external symbol
8299 @var{name} used in the linker input files.
8301 @cindex @code{SECT} (MRI)
8302 @item SECT @var{secname}, @var{expression}
8303 @itemx SECT @var{secname}=@var{expression}
8304 @itemx SECT @var{secname} @var{expression}
8305 You can use any of these three forms of the @code{SECT} command to
8306 specify the start address (@var{expression}) for section @var{secname}.
8307 If you have more than one @code{SECT} statement for the same
8308 @var{secname}, only the @emph{first} sets the start address.
8311 @node GNU Free Documentation License
8312 @appendix GNU Free Documentation License
8316 @unnumbered LD Index
8321 % I think something like @@colophon should be in texinfo. In the
8323 \long\def\colophon{\hbox to0pt{}\vfill
8324 \centerline{The body of this manual is set in}
8325 \centerline{\fontname\tenrm,}
8326 \centerline{with headings in {\bf\fontname\tenbf}}
8327 \centerline{and examples in {\tt\fontname\tentt}.}
8328 \centerline{{\it\fontname\tenit\/} and}
8329 \centerline{{\sl\fontname\tensl\/}}
8330 \centerline{are used for emphasis.}\vfill}
8332 % Blame: doc@@cygnus.com, 28mar91.