3 @c Copyright 1991, 1992, 1993, 1994, 1995, 1996, 1997, 1998, 1999, 2000,
4 @c 2001, 2002, 2003, 2004, 2005 Free Software Foundation, Inc.
6 @include configdoc.texi
7 @c (configdoc.texi is generated by the Makefile)
12 @macro gcctabopt{body}
18 @c Configure for the generation of man pages
55 * Ld: (ld). The GNU linker.
61 This file documents the @sc{gnu} linker LD version @value{VERSION}.
63 Copyright (C) 1991, 92, 93, 94, 95, 96, 97, 98, 99, 2000,
64 2001, 2002, 2003, 2004 Free Software Foundation, Inc.
68 Permission is granted to copy, distribute and/or modify this document
69 under the terms of the GNU Free Documentation License, Version 1.1
70 or any later version published by the Free Software Foundation;
71 with no Invariant Sections, with no Front-Cover Texts, and with no
72 Back-Cover Texts. A copy of the license is included in the
73 section entitled ``GNU Free Documentation License''.
75 Permission is granted to process this file through Tex and print the
76 results, provided the printed document carries copying permission
77 notice identical to this one except for the removal of this paragraph
78 (this paragraph not being relevant to the printed manual).
84 @setchapternewpage odd
85 @settitle Using LD, the GNU linker
88 @subtitle The GNU linker
90 @subtitle @code{ld} version 2
91 @subtitle Version @value{VERSION}
92 @author Steve Chamberlain
93 @author Ian Lance Taylor
98 \hfill Red Hat Inc\par
99 \hfill nickc\@credhat.com, doc\@redhat.com\par
100 \hfill {\it Using LD, the GNU linker}\par
101 \hfill Edited by Jeffrey Osier (jeffrey\@cygnus.com)\par
103 \global\parindent=0pt % Steve likes it this way.
106 @vskip 0pt plus 1filll
107 @c man begin COPYRIGHT
108 Copyright @copyright{} 1991, 92, 93, 94, 95, 96, 97, 98, 99, 2000, 2001,
109 2002, 2003, 2004 Free Software Foundation, Inc.
111 Permission is granted to copy, distribute and/or modify this document
112 under the terms of the GNU Free Documentation License, Version 1.1
113 or any later version published by the Free Software Foundation;
114 with no Invariant Sections, with no Front-Cover Texts, and with no
115 Back-Cover Texts. A copy of the license is included in the
116 section entitled ``GNU Free Documentation License''.
121 @c FIXME: Talk about importance of *order* of args, cmds to linker!
126 This file documents the @sc{gnu} linker ld version @value{VERSION}.
128 This document is distributed under the terms of the GNU Free
129 Documentation License. A copy of the license is included in the
130 section entitled ``GNU Free Documentation License''.
133 * Overview:: Overview
134 * Invocation:: Invocation
135 * Scripts:: Linker Scripts
137 * Machine Dependent:: Machine Dependent Features
141 * H8/300:: ld and the H8/300
144 * Renesas:: ld and other Renesas micros
147 * i960:: ld and the Intel 960 family
150 * ARM:: ld and the ARM family
153 * HPPA ELF32:: ld and HPPA 32-bit ELF
156 * M68HC11/68HC12:: ld and the Motorola 68HC11 and 68HC12 families
159 * TI COFF:: ld and the TI COFF
162 * Win32:: ld and WIN32 (cygwin/mingw)
165 * Xtensa:: ld and Xtensa Processors
168 @ifclear SingleFormat
171 @c Following blank line required for remaining bug in makeinfo conds/menus
173 * Reporting Bugs:: Reporting Bugs
174 * MRI:: MRI Compatible Script Files
175 * GNU Free Documentation License:: GNU Free Documentation License
183 @cindex @sc{gnu} linker
184 @cindex what is this?
187 @c man begin SYNOPSIS
188 ld [@b{options}] @var{objfile} @dots{}
192 ar(1), nm(1), objcopy(1), objdump(1), readelf(1) and
193 the Info entries for @file{binutils} and
198 @c man begin DESCRIPTION
200 @command{ld} combines a number of object and archive files, relocates
201 their data and ties up symbol references. Usually the last step in
202 compiling a program is to run @command{ld}.
204 @command{ld} accepts Linker Command Language files written in
205 a superset of AT&T's Link Editor Command Language syntax,
206 to provide explicit and total control over the linking process.
210 This man page does not describe the command language; see the
211 @command{ld} entry in @code{info}, or the manual
212 ld: the GNU linker, for full details on the command language and
213 on other aspects of the GNU linker.
216 @ifclear SingleFormat
217 This version of @command{ld} uses the general purpose BFD libraries
218 to operate on object files. This allows @command{ld} to read, combine, and
219 write object files in many different formats---for example, COFF or
220 @code{a.out}. Different formats may be linked together to produce any
221 available kind of object file. @xref{BFD}, for more information.
224 Aside from its flexibility, the @sc{gnu} linker is more helpful than other
225 linkers in providing diagnostic information. Many linkers abandon
226 execution immediately upon encountering an error; whenever possible,
227 @command{ld} continues executing, allowing you to identify other errors
228 (or, in some cases, to get an output file in spite of the error).
235 @c man begin DESCRIPTION
237 The @sc{gnu} linker @command{ld} is meant to cover a broad range of situations,
238 and to be as compatible as possible with other linkers. As a result,
239 you have many choices to control its behavior.
245 * Options:: Command Line Options
246 * Environment:: Environment Variables
250 @section Command Line Options
258 The linker supports a plethora of command-line options, but in actual
259 practice few of them are used in any particular context.
260 @cindex standard Unix system
261 For instance, a frequent use of @command{ld} is to link standard Unix
262 object files on a standard, supported Unix system. On such a system, to
263 link a file @code{hello.o}:
266 ld -o @var{output} /lib/crt0.o hello.o -lc
269 This tells @command{ld} to produce a file called @var{output} as the
270 result of linking the file @code{/lib/crt0.o} with @code{hello.o} and
271 the library @code{libc.a}, which will come from the standard search
272 directories. (See the discussion of the @samp{-l} option below.)
274 Some of the command-line options to @command{ld} may be specified at any
275 point in the command line. However, options which refer to files, such
276 as @samp{-l} or @samp{-T}, cause the file to be read at the point at
277 which the option appears in the command line, relative to the object
278 files and other file options. Repeating non-file options with a
279 different argument will either have no further effect, or override prior
280 occurrences (those further to the left on the command line) of that
281 option. Options which may be meaningfully specified more than once are
282 noted in the descriptions below.
285 Non-option arguments are object files or archives which are to be linked
286 together. They may follow, precede, or be mixed in with command-line
287 options, except that an object file argument may not be placed between
288 an option and its argument.
290 Usually the linker is invoked with at least one object file, but you can
291 specify other forms of binary input files using @samp{-l}, @samp{-R},
292 and the script command language. If @emph{no} binary input files at all
293 are specified, the linker does not produce any output, and issues the
294 message @samp{No input files}.
296 If the linker cannot recognize the format of an object file, it will
297 assume that it is a linker script. A script specified in this way
298 augments the main linker script used for the link (either the default
299 linker script or the one specified by using @samp{-T}). This feature
300 permits the linker to link against a file which appears to be an object
301 or an archive, but actually merely defines some symbol values, or uses
302 @code{INPUT} or @code{GROUP} to load other objects. Note that
303 specifying a script in this way merely augments the main linker script;
304 use the @samp{-T} option to replace the default linker script entirely.
307 For options whose names are a single letter,
308 option arguments must either follow the option letter without intervening
309 whitespace, or be given as separate arguments immediately following the
310 option that requires them.
312 For options whose names are multiple letters, either one dash or two can
313 precede the option name; for example, @samp{-trace-symbol} and
314 @samp{--trace-symbol} are equivalent. Note---there is one exception to
315 this rule. Multiple letter options that start with a lower case 'o' can
316 only be preceeded by two dashes. This is to reduce confusion with the
317 @samp{-o} option. So for example @samp{-omagic} sets the output file
318 name to @samp{magic} whereas @samp{--omagic} sets the NMAGIC flag on the
321 Arguments to multiple-letter options must either be separated from the
322 option name by an equals sign, or be given as separate arguments
323 immediately following the option that requires them. For example,
324 @samp{--trace-symbol foo} and @samp{--trace-symbol=foo} are equivalent.
325 Unique abbreviations of the names of multiple-letter options are
328 Note---if the linker is being invoked indirectly, via a compiler driver
329 (e.g. @samp{gcc}) then all the linker command line options should be
330 prefixed by @samp{-Wl,} (or whatever is appropriate for the particular
331 compiler driver) like this:
334 gcc -Wl,--startgroup foo.o bar.o -Wl,--endgroup
337 This is important, because otherwise the compiler driver program may
338 silently drop the linker options, resulting in a bad link.
340 Here is a table of the generic command line switches accepted by the GNU
344 @kindex -a@var{keyword}
345 @item -a@var{keyword}
346 This option is supported for HP/UX compatibility. The @var{keyword}
347 argument must be one of the strings @samp{archive}, @samp{shared}, or
348 @samp{default}. @samp{-aarchive} is functionally equivalent to
349 @samp{-Bstatic}, and the other two keywords are functionally equivalent
350 to @samp{-Bdynamic}. This option may be used any number of times.
353 @cindex architectures
355 @item -A@var{architecture}
356 @kindex --architecture=@var{arch}
357 @itemx --architecture=@var{architecture}
358 In the current release of @command{ld}, this option is useful only for the
359 Intel 960 family of architectures. In that @command{ld} configuration, the
360 @var{architecture} argument identifies the particular architecture in
361 the 960 family, enabling some safeguards and modifying the
362 archive-library search path. @xref{i960,,@command{ld} and the Intel 960
363 family}, for details.
365 Future releases of @command{ld} may support similar functionality for
366 other architecture families.
369 @ifclear SingleFormat
370 @cindex binary input format
371 @kindex -b @var{format}
372 @kindex --format=@var{format}
375 @item -b @var{input-format}
376 @itemx --format=@var{input-format}
377 @command{ld} may be configured to support more than one kind of object
378 file. If your @command{ld} is configured this way, you can use the
379 @samp{-b} option to specify the binary format for input object files
380 that follow this option on the command line. Even when @command{ld} is
381 configured to support alternative object formats, you don't usually need
382 to specify this, as @command{ld} should be configured to expect as a
383 default input format the most usual format on each machine.
384 @var{input-format} is a text string, the name of a particular format
385 supported by the BFD libraries. (You can list the available binary
386 formats with @samp{objdump -i}.)
389 You may want to use this option if you are linking files with an unusual
390 binary format. You can also use @samp{-b} to switch formats explicitly (when
391 linking object files of different formats), by including
392 @samp{-b @var{input-format}} before each group of object files in a
395 The default format is taken from the environment variable
400 You can also define the input format from a script, using the command
403 see @ref{Format Commands}.
407 @kindex -c @var{MRI-cmdfile}
408 @kindex --mri-script=@var{MRI-cmdfile}
409 @cindex compatibility, MRI
410 @item -c @var{MRI-commandfile}
411 @itemx --mri-script=@var{MRI-commandfile}
412 For compatibility with linkers produced by MRI, @command{ld} accepts script
413 files written in an alternate, restricted command language, described in
415 @ref{MRI,,MRI Compatible Script Files}.
418 the MRI Compatible Script Files section of GNU ld documentation.
420 Introduce MRI script files with
421 the option @samp{-c}; use the @samp{-T} option to run linker
422 scripts written in the general-purpose @command{ld} scripting language.
423 If @var{MRI-cmdfile} does not exist, @command{ld} looks for it in the directories
424 specified by any @samp{-L} options.
426 @cindex common allocation
433 These three options are equivalent; multiple forms are supported for
434 compatibility with other linkers. They assign space to common symbols
435 even if a relocatable output file is specified (with @samp{-r}). The
436 script command @code{FORCE_COMMON_ALLOCATION} has the same effect.
437 @xref{Miscellaneous Commands}.
439 @cindex entry point, from command line
440 @kindex -e @var{entry}
441 @kindex --entry=@var{entry}
443 @itemx --entry=@var{entry}
444 Use @var{entry} as the explicit symbol for beginning execution of your
445 program, rather than the default entry point. If there is no symbol
446 named @var{entry}, the linker will try to parse @var{entry} as a number,
447 and use that as the entry address (the number will be interpreted in
448 base 10; you may use a leading @samp{0x} for base 16, or a leading
449 @samp{0} for base 8). @xref{Entry Point}, for a discussion of defaults
450 and other ways of specifying the entry point.
452 @kindex --exclude-libs
453 @item --exclude-libs @var{lib},@var{lib},...
454 Specifies a list of archive libraries from which symbols should not be automatically
455 exported. The library names may be delimited by commas or colons. Specifying
456 @code{--exclude-libs ALL} excludes symbols in all archive libraries from
457 automatic export. This option is available only for the i386 PE targeted
458 port of the linker and for ELF targeted ports. For i386 PE, symbols
459 explicitly listed in a .def file are still exported, regardless of this
460 option. For ELF targeted ports, symbols affected by this option will
461 be treated as hidden.
463 @cindex dynamic symbol table
465 @kindex --export-dynamic
467 @itemx --export-dynamic
468 When creating a dynamically linked executable, add all symbols to the
469 dynamic symbol table. The dynamic symbol table is the set of symbols
470 which are visible from dynamic objects at run time.
472 If you do not use this option, the dynamic symbol table will normally
473 contain only those symbols which are referenced by some dynamic object
474 mentioned in the link.
476 If you use @code{dlopen} to load a dynamic object which needs to refer
477 back to the symbols defined by the program, rather than some other
478 dynamic object, then you will probably need to use this option when
479 linking the program itself.
481 You can also use the version script to control what symbols should
482 be added to the dynamic symbol table if the output format supports it.
483 See the description of @samp{--version-script} in @ref{VERSION}.
485 @ifclear SingleFormat
486 @cindex big-endian objects
490 Link big-endian objects. This affects the default output format.
492 @cindex little-endian objects
495 Link little-endian objects. This affects the default output format.
501 @itemx --auxiliary @var{name}
502 When creating an ELF shared object, set the internal DT_AUXILIARY field
503 to the specified name. This tells the dynamic linker that the symbol
504 table of the shared object should be used as an auxiliary filter on the
505 symbol table of the shared object @var{name}.
507 If you later link a program against this filter object, then, when you
508 run the program, the dynamic linker will see the DT_AUXILIARY field. If
509 the dynamic linker resolves any symbols from the filter object, it will
510 first check whether there is a definition in the shared object
511 @var{name}. If there is one, it will be used instead of the definition
512 in the filter object. The shared object @var{name} need not exist.
513 Thus the shared object @var{name} may be used to provide an alternative
514 implementation of certain functions, perhaps for debugging or for
515 machine specific performance.
517 This option may be specified more than once. The DT_AUXILIARY entries
518 will be created in the order in which they appear on the command line.
523 @itemx --filter @var{name}
524 When creating an ELF shared object, set the internal DT_FILTER field to
525 the specified name. This tells the dynamic linker that the symbol table
526 of the shared object which is being created should be used as a filter
527 on the symbol table of the shared object @var{name}.
529 If you later link a program against this filter object, then, when you
530 run the program, the dynamic linker will see the DT_FILTER field. The
531 dynamic linker will resolve symbols according to the symbol table of the
532 filter object as usual, but it will actually link to the definitions
533 found in the shared object @var{name}. Thus the filter object can be
534 used to select a subset of the symbols provided by the object
537 Some older linkers used the @option{-F} option throughout a compilation
538 toolchain for specifying object-file format for both input and output
540 @ifclear SingleFormat
541 The @sc{gnu} linker uses other mechanisms for this purpose: the
542 @option{-b}, @option{--format}, @option{--oformat} options, the
543 @code{TARGET} command in linker scripts, and the @code{GNUTARGET}
544 environment variable.
546 The @sc{gnu} linker will ignore the @option{-F} option when not
547 creating an ELF shared object.
549 @cindex finalization function
551 @item -fini @var{name}
552 When creating an ELF executable or shared object, call NAME when the
553 executable or shared object is unloaded, by setting DT_FINI to the
554 address of the function. By default, the linker uses @code{_fini} as
555 the function to call.
559 Ignored. Provided for compatibility with other tools.
565 @itemx --gpsize=@var{value}
566 Set the maximum size of objects to be optimized using the GP register to
567 @var{size}. This is only meaningful for object file formats such as
568 MIPS ECOFF which supports putting large and small objects into different
569 sections. This is ignored for other object file formats.
571 @cindex runtime library name
573 @kindex -soname=@var{name}
575 @itemx -soname=@var{name}
576 When creating an ELF shared object, set the internal DT_SONAME field to
577 the specified name. When an executable is linked with a shared object
578 which has a DT_SONAME field, then when the executable is run the dynamic
579 linker will attempt to load the shared object specified by the DT_SONAME
580 field rather than the using the file name given to the linker.
583 @cindex incremental link
585 Perform an incremental link (same as option @samp{-r}).
587 @cindex initialization function
589 @item -init @var{name}
590 When creating an ELF executable or shared object, call NAME when the
591 executable or shared object is loaded, by setting DT_INIT to the address
592 of the function. By default, the linker uses @code{_init} as the
595 @cindex archive files, from cmd line
596 @kindex -l@var{archive}
597 @kindex --library=@var{archive}
598 @item -l@var{archive}
599 @itemx --library=@var{archive}
600 Add archive file @var{archive} to the list of files to link. This
601 option may be used any number of times. @command{ld} will search its
602 path-list for occurrences of @code{lib@var{archive}.a} for every
603 @var{archive} specified.
605 On systems which support shared libraries, @command{ld} may also search for
606 libraries with extensions other than @code{.a}. Specifically, on ELF
607 and SunOS systems, @command{ld} will search a directory for a library with
608 an extension of @code{.so} before searching for one with an extension of
609 @code{.a}. By convention, a @code{.so} extension indicates a shared
612 The linker will search an archive only once, at the location where it is
613 specified on the command line. If the archive defines a symbol which
614 was undefined in some object which appeared before the archive on the
615 command line, the linker will include the appropriate file(s) from the
616 archive. However, an undefined symbol in an object appearing later on
617 the command line will not cause the linker to search the archive again.
619 See the @option{-(} option for a way to force the linker to search
620 archives multiple times.
622 You may list the same archive multiple times on the command line.
625 This type of archive searching is standard for Unix linkers. However,
626 if you are using @command{ld} on AIX, note that it is different from the
627 behaviour of the AIX linker.
630 @cindex search directory, from cmd line
632 @kindex --library-path=@var{dir}
633 @item -L@var{searchdir}
634 @itemx --library-path=@var{searchdir}
635 Add path @var{searchdir} to the list of paths that @command{ld} will search
636 for archive libraries and @command{ld} control scripts. You may use this
637 option any number of times. The directories are searched in the order
638 in which they are specified on the command line. Directories specified
639 on the command line are searched before the default directories. All
640 @option{-L} options apply to all @option{-l} options, regardless of the
641 order in which the options appear.
643 If @var{searchdir} begins with @code{=}, then the @code{=} will be replaced
644 by the @dfn{sysroot prefix}, a path specified when the linker is configured.
647 The default set of paths searched (without being specified with
648 @samp{-L}) depends on which emulation mode @command{ld} is using, and in
649 some cases also on how it was configured. @xref{Environment}.
652 The paths can also be specified in a link script with the
653 @code{SEARCH_DIR} command. Directories specified this way are searched
654 at the point in which the linker script appears in the command line.
657 @kindex -m @var{emulation}
658 @item -m@var{emulation}
659 Emulate the @var{emulation} linker. You can list the available
660 emulations with the @samp{--verbose} or @samp{-V} options.
662 If the @samp{-m} option is not used, the emulation is taken from the
663 @code{LDEMULATION} environment variable, if that is defined.
665 Otherwise, the default emulation depends upon how the linker was
673 Print a link map to the standard output. A link map provides
674 information about the link, including the following:
678 Where object files are mapped into memory.
680 How common symbols are allocated.
682 All archive members included in the link, with a mention of the symbol
683 which caused the archive member to be brought in.
685 The values assigned to symbols.
687 Note - symbols whose values are computed by an expression which
688 involves a reference to a previous value of the same symbol may not
689 have correct result displayed in the link map. This is because the
690 linker discards intermediate results and only retains the final value
691 of an expression. Under such circumstances the linker will display
692 the final value enclosed by square brackets. Thus for example a
693 linker script containing:
701 will produce the following output in the link map if the @option{-M}
706 [0x0000000c] foo = (foo * 0x4)
707 [0x0000000c] foo = (foo + 0x8)
710 See @ref{Expressions} for more information about expressions in linker
715 @cindex read-only text
720 Turn off page alignment of sections, and mark the output as
721 @code{NMAGIC} if possible.
725 @cindex read/write from cmd line
729 Set the text and data sections to be readable and writable. Also, do
730 not page-align the data segment, and disable linking against shared
731 libraries. If the output format supports Unix style magic numbers,
732 mark the output as @code{OMAGIC}. Note: Although a writable text section
733 is allowed for PE-COFF targets, it does not conform to the format
734 specification published by Microsoft.
739 This option negates most of the effects of the @option{-N} option. It
740 sets the text section to be read-only, and forces the data segment to
741 be page-aligned. Note - this option does not enable linking against
742 shared libraries. Use @option{-Bdynamic} for this.
744 @kindex -o @var{output}
745 @kindex --output=@var{output}
746 @cindex naming the output file
747 @item -o @var{output}
748 @itemx --output=@var{output}
749 Use @var{output} as the name for the program produced by @command{ld}; if this
750 option is not specified, the name @file{a.out} is used by default. The
751 script command @code{OUTPUT} can also specify the output file name.
753 @kindex -O @var{level}
754 @cindex generating optimized output
756 If @var{level} is a numeric values greater than zero @command{ld} optimizes
757 the output. This might take significantly longer and therefore probably
758 should only be enabled for the final binary.
761 @kindex --emit-relocs
762 @cindex retain relocations in final executable
765 Leave relocation sections and contents in fully linked exececutables.
766 Post link analysis and optimization tools may need this information in
767 order to perform correct modifications of executables. This results
768 in larger executables.
770 This option is currently only supported on ELF platforms.
773 @cindex relocatable output
775 @kindex --relocatable
778 Generate relocatable output---i.e., generate an output file that can in
779 turn serve as input to @command{ld}. This is often called @dfn{partial
780 linking}. As a side effect, in environments that support standard Unix
781 magic numbers, this option also sets the output file's magic number to
783 @c ; see @option{-N}.
784 If this option is not specified, an absolute file is produced. When
785 linking C++ programs, this option @emph{will not} resolve references to
786 constructors; to do that, use @samp{-Ur}.
788 When an input file does not have the same format as the output file,
789 partial linking is only supported if that input file does not contain any
790 relocations. Different output formats can have further restrictions; for
791 example some @code{a.out}-based formats do not support partial linking
792 with input files in other formats at all.
794 This option does the same thing as @samp{-i}.
796 @kindex -R @var{file}
797 @kindex --just-symbols=@var{file}
798 @cindex symbol-only input
799 @item -R @var{filename}
800 @itemx --just-symbols=@var{filename}
801 Read symbol names and their addresses from @var{filename}, but do not
802 relocate it or include it in the output. This allows your output file
803 to refer symbolically to absolute locations of memory defined in other
804 programs. You may use this option more than once.
806 For compatibility with other ELF linkers, if the @option{-R} option is
807 followed by a directory name, rather than a file name, it is treated as
808 the @option{-rpath} option.
812 @cindex strip all symbols
815 Omit all symbol information from the output file.
818 @kindex --strip-debug
819 @cindex strip debugger symbols
822 Omit debugger symbol information (but not all symbols) from the output file.
826 @cindex input files, displaying
829 Print the names of the input files as @command{ld} processes them.
831 @kindex -T @var{script}
832 @kindex --script=@var{script}
834 @item -T @var{scriptfile}
835 @itemx --script=@var{scriptfile}
836 Use @var{scriptfile} as the linker script. This script replaces
837 @command{ld}'s default linker script (rather than adding to it), so
838 @var{commandfile} must specify everything necessary to describe the
839 output file. @xref{Scripts}. If @var{scriptfile} does not exist in
840 the current directory, @code{ld} looks for it in the directories
841 specified by any preceding @samp{-L} options. Multiple @samp{-T}
844 @kindex -u @var{symbol}
845 @kindex --undefined=@var{symbol}
846 @cindex undefined symbol
847 @item -u @var{symbol}
848 @itemx --undefined=@var{symbol}
849 Force @var{symbol} to be entered in the output file as an undefined
850 symbol. Doing this may, for example, trigger linking of additional
851 modules from standard libraries. @samp{-u} may be repeated with
852 different option arguments to enter additional undefined symbols. This
853 option is equivalent to the @code{EXTERN} linker script command.
858 For anything other than C++ programs, this option is equivalent to
859 @samp{-r}: it generates relocatable output---i.e., an output file that can in
860 turn serve as input to @command{ld}. When linking C++ programs, @samp{-Ur}
861 @emph{does} resolve references to constructors, unlike @samp{-r}.
862 It does not work to use @samp{-Ur} on files that were themselves linked
863 with @samp{-Ur}; once the constructor table has been built, it cannot
864 be added to. Use @samp{-Ur} only for the last partial link, and
865 @samp{-r} for the others.
867 @kindex --unique[=@var{SECTION}]
868 @item --unique[=@var{SECTION}]
869 Creates a separate output section for every input section matching
870 @var{SECTION}, or if the optional wildcard @var{SECTION} argument is
871 missing, for every orphan input section. An orphan section is one not
872 specifically mentioned in a linker script. You may use this option
873 multiple times on the command line; It prevents the normal merging of
874 input sections with the same name, overriding output section assignments
884 Display the version number for @command{ld}. The @option{-V} option also
885 lists the supported emulations.
888 @kindex --discard-all
889 @cindex deleting local symbols
892 Delete all local symbols.
895 @kindex --discard-locals
896 @cindex local symbols, deleting
897 @cindex L, deleting symbols beginning
899 @itemx --discard-locals
900 Delete all temporary local symbols. For most targets, this is all local
901 symbols whose names begin with @samp{L}.
903 @kindex -y @var{symbol}
904 @kindex --trace-symbol=@var{symbol}
905 @cindex symbol tracing
906 @item -y @var{symbol}
907 @itemx --trace-symbol=@var{symbol}
908 Print the name of each linked file in which @var{symbol} appears. This
909 option may be given any number of times. On many systems it is necessary
910 to prepend an underscore.
912 This option is useful when you have an undefined symbol in your link but
913 don't know where the reference is coming from.
915 @kindex -Y @var{path}
917 Add @var{path} to the default library search path. This option exists
918 for Solaris compatibility.
920 @kindex -z @var{keyword}
921 @item -z @var{keyword}
922 The recognized keywords are:
926 Combines multiple reloc sections and sorts them to make dynamic symbol
927 lookup caching possible.
930 Disallows undefined symbols in object files. Undefined symbols in
931 shared libraries are still allowed.
934 This option is only meaningful when building a shared object.
935 It marks the object so that its runtime initialization will occur
936 before the runtime initialization of any other objects brought into
937 the process at the same time. Similarly the runtime finalization of
938 the object will occur after the runtime finalization of any other
942 Marks the object that its symbol table interposes before all symbols
943 but the primary executable.
946 Marks the object that its filters be processed immediately at
950 Allows multiple definitions.
953 Disables multiple reloc sections combining.
956 Disables production of copy relocs.
959 Marks the object that the search for dependencies of this object will
960 ignore any default library search paths.
963 Marks the object shouldn't be unloaded at runtime.
966 Marks the object not available to @code{dlopen}.
969 Marks the object can not be dumped by @code{dldump}.
972 When generating an executable or shared library, mark it to tell the
973 dynamic linker to resolve all symbols when the program is started, or
974 when the shared library is linked to using dlopen, instead of
975 deferring function call resolution to the point when the function is
979 Marks the object may contain $ORIGIN.
983 Other keywords are ignored for Solaris compatibility.
986 @cindex groups of archives
987 @item -( @var{archives} -)
988 @itemx --start-group @var{archives} --end-group
989 The @var{archives} should be a list of archive files. They may be
990 either explicit file names, or @samp{-l} options.
992 The specified archives are searched repeatedly until no new undefined
993 references are created. Normally, an archive is searched only once in
994 the order that it is specified on the command line. If a symbol in that
995 archive is needed to resolve an undefined symbol referred to by an
996 object in an archive that appears later on the command line, the linker
997 would not be able to resolve that reference. By grouping the archives,
998 they all be searched repeatedly until all possible references are
1001 Using this option has a significant performance cost. It is best to use
1002 it only when there are unavoidable circular references between two or
1005 @kindex --accept-unknown-input-arch
1006 @kindex --no-accept-unknown-input-arch
1007 @item --accept-unknown-input-arch
1008 @itemx --no-accept-unknown-input-arch
1009 Tells the linker to accept input files whose architecture cannot be
1010 recognised. The assumption is that the user knows what they are doing
1011 and deliberately wants to link in these unknown input files. This was
1012 the default behaviour of the linker, before release 2.14. The default
1013 behaviour from release 2.14 onwards is to reject such input files, and
1014 so the @samp{--accept-unknown-input-arch} option has been added to
1015 restore the old behaviour.
1018 @kindex --no-as-needed
1020 @itemx --no-as-needed
1021 This option affects ELF DT_NEEDED tags for dynamic libraries mentioned
1022 on the command line after the @option{--as-needed} option. Normally,
1023 the linker will add a DT_NEEDED tag for each dynamic library mentioned
1024 on the command line, regardless of whether the library is actually
1025 needed. @option{--as-needed} causes DT_NEEDED tags to only be emitted
1026 for libraries that satisfy some symbol reference from regular objects
1027 which is undefined at the point that the library was linked.
1028 @option{--no-as-needed} restores the default behaviour.
1030 @kindex --add-needed
1031 @kindex --no-add-needed
1033 @itemx --no-add-needed
1034 This option affects the treatment of dynamic libraries from ELF
1035 DT_NEEDED tags in dynamic libraries mentioned on the command line after
1036 the @option{--no-add-needed} option. Normally, the linker will add
1037 a DT_NEEDED tag for each dynamic library from DT_NEEDED tags.
1038 @option{--no-add-needed} causes DT_NEEDED tags will never be emitted
1039 for those libraries from DT_NEEDED tags. @option{--add-needed} restores
1040 the default behaviour.
1042 @kindex -assert @var{keyword}
1043 @item -assert @var{keyword}
1044 This option is ignored for SunOS compatibility.
1048 @kindex -call_shared
1052 Link against dynamic libraries. This is only meaningful on platforms
1053 for which shared libraries are supported. This option is normally the
1054 default on such platforms. The different variants of this option are
1055 for compatibility with various systems. You may use this option
1056 multiple times on the command line: it affects library searching for
1057 @option{-l} options which follow it.
1061 Set the @code{DF_1_GROUP} flag in the @code{DT_FLAGS_1} entry in the dynamic
1062 section. This causes the runtime linker to handle lookups in this
1063 object and its dependencies to be performed only inside the group.
1064 @option{--unresolved-symbols=report-all} is implied. This option is
1065 only meaningful on ELF platforms which support shared libraries.
1075 Do not link against shared libraries. This is only meaningful on
1076 platforms for which shared libraries are supported. The different
1077 variants of this option are for compatibility with various systems. You
1078 may use this option multiple times on the command line: it affects
1079 library searching for @option{-l} options which follow it. This
1080 option also implies @option{--unresolved-symbols=report-all}. This
1081 option can be used with @option{-shared}. Doing so means that a
1082 shared library is being created but that all of the library's external
1083 references must be resolved by pulling in entries from static
1088 When creating a shared library, bind references to global symbols to the
1089 definition within the shared library, if any. Normally, it is possible
1090 for a program linked against a shared library to override the definition
1091 within the shared library. This option is only meaningful on ELF
1092 platforms which support shared libraries.
1094 @kindex --check-sections
1095 @kindex --no-check-sections
1096 @item --check-sections
1097 @itemx --no-check-sections
1098 Asks the linker @emph{not} to check section addresses after they have
1099 been assigned to see if there any overlaps. Normally the linker will
1100 perform this check, and if it finds any overlaps it will produce
1101 suitable error messages. The linker does know about, and does make
1102 allowances for sections in overlays. The default behaviour can be
1103 restored by using the command line switch @option{--check-sections}.
1105 @cindex cross reference table
1108 Output a cross reference table. If a linker map file is being
1109 generated, the cross reference table is printed to the map file.
1110 Otherwise, it is printed on the standard output.
1112 The format of the table is intentionally simple, so that it may be
1113 easily processed by a script if necessary. The symbols are printed out,
1114 sorted by name. For each symbol, a list of file names is given. If the
1115 symbol is defined, the first file listed is the location of the
1116 definition. The remaining files contain references to the symbol.
1118 @cindex common allocation
1119 @kindex --no-define-common
1120 @item --no-define-common
1121 This option inhibits the assignment of addresses to common symbols.
1122 The script command @code{INHIBIT_COMMON_ALLOCATION} has the same effect.
1123 @xref{Miscellaneous Commands}.
1125 The @samp{--no-define-common} option allows decoupling
1126 the decision to assign addresses to Common symbols from the choice
1127 of the output file type; otherwise a non-Relocatable output type
1128 forces assigning addresses to Common symbols.
1129 Using @samp{--no-define-common} allows Common symbols that are referenced
1130 from a shared library to be assigned addresses only in the main program.
1131 This eliminates the unused duplicate space in the shared library,
1132 and also prevents any possible confusion over resolving to the wrong
1133 duplicate when there are many dynamic modules with specialized search
1134 paths for runtime symbol resolution.
1136 @cindex symbols, from command line
1137 @kindex --defsym @var{symbol}=@var{exp}
1138 @item --defsym @var{symbol}=@var{expression}
1139 Create a global symbol in the output file, containing the absolute
1140 address given by @var{expression}. You may use this option as many
1141 times as necessary to define multiple symbols in the command line. A
1142 limited form of arithmetic is supported for the @var{expression} in this
1143 context: you may give a hexadecimal constant or the name of an existing
1144 symbol, or use @code{+} and @code{-} to add or subtract hexadecimal
1145 constants or symbols. If you need more elaborate expressions, consider
1146 using the linker command language from a script (@pxref{Assignments,,
1147 Assignment: Symbol Definitions}). @emph{Note:} there should be no white
1148 space between @var{symbol}, the equals sign (``@key{=}''), and
1151 @cindex demangling, from command line
1152 @kindex --demangle[=@var{style}]
1153 @kindex --no-demangle
1154 @item --demangle[=@var{style}]
1155 @itemx --no-demangle
1156 These options control whether to demangle symbol names in error messages
1157 and other output. When the linker is told to demangle, it tries to
1158 present symbol names in a readable fashion: it strips leading
1159 underscores if they are used by the object file format, and converts C++
1160 mangled symbol names into user readable names. Different compilers have
1161 different mangling styles. The optional demangling style argument can be used
1162 to choose an appropriate demangling style for your compiler. The linker will
1163 demangle by default unless the environment variable @samp{COLLECT_NO_DEMANGLE}
1164 is set. These options may be used to override the default.
1166 @cindex dynamic linker, from command line
1167 @kindex -I@var{file}
1168 @kindex --dynamic-linker @var{file}
1169 @item --dynamic-linker @var{file}
1170 Set the name of the dynamic linker. This is only meaningful when
1171 generating dynamically linked ELF executables. The default dynamic
1172 linker is normally correct; don't use this unless you know what you are
1176 @kindex --fatal-warnings
1177 @item --fatal-warnings
1178 Treat all warnings as errors.
1180 @kindex --force-exe-suffix
1181 @item --force-exe-suffix
1182 Make sure that an output file has a .exe suffix.
1184 If a successfully built fully linked output file does not have a
1185 @code{.exe} or @code{.dll} suffix, this option forces the linker to copy
1186 the output file to one of the same name with a @code{.exe} suffix. This
1187 option is useful when using unmodified Unix makefiles on a Microsoft
1188 Windows host, since some versions of Windows won't run an image unless
1189 it ends in a @code{.exe} suffix.
1191 @kindex --gc-sections
1192 @kindex --no-gc-sections
1193 @cindex garbage collection
1194 @item --no-gc-sections
1195 @itemx --gc-sections
1196 Enable garbage collection of unused input sections. It is ignored on
1197 targets that do not support this option. This option is not compatible
1198 with @samp{-r}. The default behaviour (of not performing this garbage
1199 collection) can be restored by specifying @samp{--no-gc-sections} on
1206 Print a summary of the command-line options on the standard output and exit.
1208 @kindex --target-help
1210 Print a summary of all target specific options on the standard output and exit.
1213 @item -Map @var{mapfile}
1214 Print a link map to the file @var{mapfile}. See the description of the
1215 @option{-M} option, above.
1217 @cindex memory usage
1218 @kindex --no-keep-memory
1219 @item --no-keep-memory
1220 @command{ld} normally optimizes for speed over memory usage by caching the
1221 symbol tables of input files in memory. This option tells @command{ld} to
1222 instead optimize for memory usage, by rereading the symbol tables as
1223 necessary. This may be required if @command{ld} runs out of memory space
1224 while linking a large executable.
1226 @kindex --no-undefined
1228 @item --no-undefined
1230 Report unresolved symbol references from regular object files. This
1231 is done even if the linker is creating a non-symbolic shared library.
1232 The switch @option{--[no-]allow-shlib-undefined} controls the
1233 behaviour for reporting unresolved references found in shared
1234 libraries being linked in.
1236 @kindex --allow-multiple-definition
1238 @item --allow-multiple-definition
1240 Normally when a symbol is defined multiple times, the linker will
1241 report a fatal error. These options allow multiple definitions and the
1242 first definition will be used.
1244 @kindex --allow-shlib-undefined
1245 @kindex --no-allow-shlib-undefined
1246 @item --allow-shlib-undefined
1247 @itemx --no-allow-shlib-undefined
1248 Allows (the default) or disallows undefined symbols in shared libraries.
1249 This switch is similar to @option{--no-undefined} except that it
1250 determines the behaviour when the undefined symbols are in a
1251 shared library rather than a regular object file. It does not affect
1252 how undefined symbols in regular object files are handled.
1254 The reason that @option{--allow-shlib-undefined} is the default is that
1255 the shared library being specified at link time may not be the same as
1256 the one that is available at load time, so the symbols might actually be
1257 resolvable at load time. Plus there are some systems, (eg BeOS) where
1258 undefined symbols in shared libraries is normal. (The kernel patches
1259 them at load time to select which function is most appropriate
1260 for the current architecture. This is used for example to dynamically
1261 select an appropriate memset function). Apparently it is also normal
1262 for HPPA shared libraries to have undefined symbols.
1264 @kindex --no-undefined-version
1265 @item --no-undefined-version
1266 Normally when a symbol has an undefined version, the linker will ignore
1267 it. This option disallows symbols with undefined version and a fatal error
1268 will be issued instead.
1270 @kindex --default-symver
1271 @item --default-symver
1272 Create and use a default symbol version (the soname) for unversioned
1275 @kindex --default-imported-symver
1276 @item --default-imported-symver
1277 Create and use a default symbol version (the soname) for unversioned
1280 @kindex --no-warn-mismatch
1281 @item --no-warn-mismatch
1282 Normally @command{ld} will give an error if you try to link together input
1283 files that are mismatched for some reason, perhaps because they have
1284 been compiled for different processors or for different endiannesses.
1285 This option tells @command{ld} that it should silently permit such possible
1286 errors. This option should only be used with care, in cases when you
1287 have taken some special action that ensures that the linker errors are
1290 @kindex --no-whole-archive
1291 @item --no-whole-archive
1292 Turn off the effect of the @option{--whole-archive} option for subsequent
1295 @cindex output file after errors
1296 @kindex --noinhibit-exec
1297 @item --noinhibit-exec
1298 Retain the executable output file whenever it is still usable.
1299 Normally, the linker will not produce an output file if it encounters
1300 errors during the link process; it exits without writing an output file
1301 when it issues any error whatsoever.
1305 Only search library directories explicitly specified on the
1306 command line. Library directories specified in linker scripts
1307 (including linker scripts specified on the command line) are ignored.
1309 @ifclear SingleFormat
1311 @item --oformat @var{output-format}
1312 @command{ld} may be configured to support more than one kind of object
1313 file. If your @command{ld} is configured this way, you can use the
1314 @samp{--oformat} option to specify the binary format for the output
1315 object file. Even when @command{ld} is configured to support alternative
1316 object formats, you don't usually need to specify this, as @command{ld}
1317 should be configured to produce as a default output format the most
1318 usual format on each machine. @var{output-format} is a text string, the
1319 name of a particular format supported by the BFD libraries. (You can
1320 list the available binary formats with @samp{objdump -i}.) The script
1321 command @code{OUTPUT_FORMAT} can also specify the output format, but
1322 this option overrides it. @xref{BFD}.
1326 @kindex --pic-executable
1328 @itemx --pic-executable
1329 @cindex position independent executables
1330 Create a position independent executable. This is currently only supported on
1331 ELF platforms. Position independent executables are similar to shared
1332 libraries in that they are relocated by the dynamic linker to the virtual
1333 address the OS chooses for them (which can vary between invocations). Like
1334 normal dynamically linked executables they can be executed and symbols
1335 defined in the executable cannot be overridden by shared libraries.
1339 This option is ignored for Linux compatibility.
1343 This option is ignored for SVR4 compatibility.
1346 @cindex synthesizing linker
1347 @cindex relaxing addressing modes
1349 An option with machine dependent effects.
1351 This option is only supported on a few targets.
1354 @xref{H8/300,,@command{ld} and the H8/300}.
1357 @xref{i960,, @command{ld} and the Intel 960 family}.
1360 @xref{Xtensa,, @command{ld} and Xtensa Processors}.
1363 @xref{M68HC11/68HC12,,@command{ld} and the 68HC11 and 68HC12}.
1366 On some platforms, the @samp{--relax} option performs global
1367 optimizations that become possible when the linker resolves addressing
1368 in the program, such as relaxing address modes and synthesizing new
1369 instructions in the output object file.
1371 On some platforms these link time global optimizations may make symbolic
1372 debugging of the resulting executable impossible.
1375 the case for the Matsushita MN10200 and MN10300 family of processors.
1379 On platforms where this is not supported, @samp{--relax} is accepted,
1383 @cindex retaining specified symbols
1384 @cindex stripping all but some symbols
1385 @cindex symbols, retaining selectively
1386 @item --retain-symbols-file @var{filename}
1387 Retain @emph{only} the symbols listed in the file @var{filename},
1388 discarding all others. @var{filename} is simply a flat file, with one
1389 symbol name per line. This option is especially useful in environments
1393 where a large global symbol table is accumulated gradually, to conserve
1396 @samp{--retain-symbols-file} does @emph{not} discard undefined symbols,
1397 or symbols needed for relocations.
1399 You may only specify @samp{--retain-symbols-file} once in the command
1400 line. It overrides @samp{-s} and @samp{-S}.
1403 @item -rpath @var{dir}
1404 @cindex runtime library search path
1406 Add a directory to the runtime library search path. This is used when
1407 linking an ELF executable with shared objects. All @option{-rpath}
1408 arguments are concatenated and passed to the runtime linker, which uses
1409 them to locate shared objects at runtime. The @option{-rpath} option is
1410 also used when locating shared objects which are needed by shared
1411 objects explicitly included in the link; see the description of the
1412 @option{-rpath-link} option. If @option{-rpath} is not used when linking an
1413 ELF executable, the contents of the environment variable
1414 @code{LD_RUN_PATH} will be used if it is defined.
1416 The @option{-rpath} option may also be used on SunOS. By default, on
1417 SunOS, the linker will form a runtime search patch out of all the
1418 @option{-L} options it is given. If a @option{-rpath} option is used, the
1419 runtime search path will be formed exclusively using the @option{-rpath}
1420 options, ignoring the @option{-L} options. This can be useful when using
1421 gcc, which adds many @option{-L} options which may be on NFS mounted
1424 For compatibility with other ELF linkers, if the @option{-R} option is
1425 followed by a directory name, rather than a file name, it is treated as
1426 the @option{-rpath} option.
1430 @cindex link-time runtime library search path
1432 @item -rpath-link @var{DIR}
1433 When using ELF or SunOS, one shared library may require another. This
1434 happens when an @code{ld -shared} link includes a shared library as one
1437 When the linker encounters such a dependency when doing a non-shared,
1438 non-relocatable link, it will automatically try to locate the required
1439 shared library and include it in the link, if it is not included
1440 explicitly. In such a case, the @option{-rpath-link} option
1441 specifies the first set of directories to search. The
1442 @option{-rpath-link} option may specify a sequence of directory names
1443 either by specifying a list of names separated by colons, or by
1444 appearing multiple times.
1446 This option should be used with caution as it overrides the search path
1447 that may have been hard compiled into a shared library. In such a case it
1448 is possible to use unintentionally a different search path than the
1449 runtime linker would do.
1451 The linker uses the following search paths to locate required shared
1455 Any directories specified by @option{-rpath-link} options.
1457 Any directories specified by @option{-rpath} options. The difference
1458 between @option{-rpath} and @option{-rpath-link} is that directories
1459 specified by @option{-rpath} options are included in the executable and
1460 used at runtime, whereas the @option{-rpath-link} option is only effective
1461 at link time. It is for the native linker only.
1463 On an ELF system, if the @option{-rpath} and @code{rpath-link} options
1464 were not used, search the contents of the environment variable
1465 @code{LD_RUN_PATH}. It is for the native linker only.
1467 On SunOS, if the @option{-rpath} option was not used, search any
1468 directories specified using @option{-L} options.
1470 For a native linker, the contents of the environment variable
1471 @code{LD_LIBRARY_PATH}.
1473 For a native ELF linker, the directories in @code{DT_RUNPATH} or
1474 @code{DT_RPATH} of a shared library are searched for shared
1475 libraries needed by it. The @code{DT_RPATH} entries are ignored if
1476 @code{DT_RUNPATH} entries exist.
1478 The default directories, normally @file{/lib} and @file{/usr/lib}.
1480 For a native linker on an ELF system, if the file @file{/etc/ld.so.conf}
1481 exists, the list of directories found in that file.
1484 If the required shared library is not found, the linker will issue a
1485 warning and continue with the link.
1492 @cindex shared libraries
1493 Create a shared library. This is currently only supported on ELF, XCOFF
1494 and SunOS platforms. On SunOS, the linker will automatically create a
1495 shared library if the @option{-e} option is not used and there are
1496 undefined symbols in the link.
1499 @kindex --sort-common
1500 This option tells @command{ld} to sort the common symbols by size when it
1501 places them in the appropriate output sections. First come all the one
1502 byte symbols, then all the two byte, then all the four byte, and then
1503 everything else. This is to prevent gaps between symbols due to
1504 alignment constraints.
1506 @kindex --sort-section name
1507 @item --sort-section name
1508 This option will apply @code{SORT_BY_NAME} to all wildcard section
1509 patterns in the linker script.
1511 @kindex --sort-section alignment
1512 @item --sort-section alignment
1513 This option will apply @code{SORT_BY_ALIGNMENT} to all wildcard section
1514 patterns in the linker script.
1516 @kindex --split-by-file
1517 @item --split-by-file [@var{size}]
1518 Similar to @option{--split-by-reloc} but creates a new output section for
1519 each input file when @var{size} is reached. @var{size} defaults to a
1520 size of 1 if not given.
1522 @kindex --split-by-reloc
1523 @item --split-by-reloc [@var{count}]
1524 Tries to creates extra sections in the output file so that no single
1525 output section in the file contains more than @var{count} relocations.
1526 This is useful when generating huge relocatable files for downloading into
1527 certain real time kernels with the COFF object file format; since COFF
1528 cannot represent more than 65535 relocations in a single section. Note
1529 that this will fail to work with object file formats which do not
1530 support arbitrary sections. The linker will not split up individual
1531 input sections for redistribution, so if a single input section contains
1532 more than @var{count} relocations one output section will contain that
1533 many relocations. @var{count} defaults to a value of 32768.
1537 Compute and display statistics about the operation of the linker, such
1538 as execution time and memory usage.
1541 @item --sysroot=@var{directory}
1542 Use @var{directory} as the location of the sysroot, overriding the
1543 configure-time default. This option is only supported by linkers
1544 that were configured using @option{--with-sysroot}.
1546 @kindex --traditional-format
1547 @cindex traditional format
1548 @item --traditional-format
1549 For some targets, the output of @command{ld} is different in some ways from
1550 the output of some existing linker. This switch requests @command{ld} to
1551 use the traditional format instead.
1554 For example, on SunOS, @command{ld} combines duplicate entries in the
1555 symbol string table. This can reduce the size of an output file with
1556 full debugging information by over 30 percent. Unfortunately, the SunOS
1557 @code{dbx} program can not read the resulting program (@code{gdb} has no
1558 trouble). The @samp{--traditional-format} switch tells @command{ld} to not
1559 combine duplicate entries.
1561 @kindex --section-start @var{sectionname}=@var{org}
1562 @item --section-start @var{sectionname}=@var{org}
1563 Locate a section in the output file at the absolute
1564 address given by @var{org}. You may use this option as many
1565 times as necessary to locate multiple sections in the command
1567 @var{org} must be a single hexadecimal integer;
1568 for compatibility with other linkers, you may omit the leading
1569 @samp{0x} usually associated with hexadecimal values. @emph{Note:} there
1570 should be no white space between @var{sectionname}, the equals
1571 sign (``@key{=}''), and @var{org}.
1573 @kindex -Tbss @var{org}
1574 @kindex -Tdata @var{org}
1575 @kindex -Ttext @var{org}
1576 @cindex segment origins, cmd line
1577 @item -Tbss @var{org}
1578 @itemx -Tdata @var{org}
1579 @itemx -Ttext @var{org}
1580 Same as --section-start, with @code{.bss}, @code{.data} or
1581 @code{.text} as the @var{sectionname}.
1583 @kindex --unresolved-symbols
1584 @item --unresolved-symbols=@var{method}
1585 Determine how to handle unresolved symbols. There are four possible
1586 values for @samp{method}:
1590 Do not report any unresolved symbols.
1593 Report all unresolved symbols. This is the default.
1595 @item ignore-in-object-files
1596 Report unresolved symbols that are contained in shared libraries, but
1597 ignore them if they come from regular object files.
1599 @item ignore-in-shared-libs
1600 Report unresolved symbols that come from regular object files, but
1601 ignore them if they come from shared libraries. This can be useful
1602 when creating a dynamic binary and it is known that all the shared
1603 libraries that it should be referencing are included on the linker's
1607 The behaviour for shared libraries on their own can also be controlled
1608 by the @option{--[no-]allow-shlib-undefined} option.
1610 Normally the linker will generate an error message for each reported
1611 unresolved symbol but the option @option{--warn-unresolved-symbols}
1612 can change this to a warning.
1618 Display the version number for @command{ld} and list the linker emulations
1619 supported. Display which input files can and cannot be opened. Display
1620 the linker script being used by the linker.
1622 @kindex --version-script=@var{version-scriptfile}
1623 @cindex version script, symbol versions
1624 @itemx --version-script=@var{version-scriptfile}
1625 Specify the name of a version script to the linker. This is typically
1626 used when creating shared libraries to specify additional information
1627 about the version hierarchy for the library being created. This option
1628 is only meaningful on ELF platforms which support shared libraries.
1631 @kindex --warn-common
1632 @cindex warnings, on combining symbols
1633 @cindex combining symbols, warnings on
1635 Warn when a common symbol is combined with another common symbol or with
1636 a symbol definition. Unix linkers allow this somewhat sloppy practise,
1637 but linkers on some other operating systems do not. This option allows
1638 you to find potential problems from combining global symbols.
1639 Unfortunately, some C libraries use this practise, so you may get some
1640 warnings about symbols in the libraries as well as in your programs.
1642 There are three kinds of global symbols, illustrated here by C examples:
1646 A definition, which goes in the initialized data section of the output
1650 An undefined reference, which does not allocate space.
1651 There must be either a definition or a common symbol for the
1655 A common symbol. If there are only (one or more) common symbols for a
1656 variable, it goes in the uninitialized data area of the output file.
1657 The linker merges multiple common symbols for the same variable into a
1658 single symbol. If they are of different sizes, it picks the largest
1659 size. The linker turns a common symbol into a declaration, if there is
1660 a definition of the same variable.
1663 The @samp{--warn-common} option can produce five kinds of warnings.
1664 Each warning consists of a pair of lines: the first describes the symbol
1665 just encountered, and the second describes the previous symbol
1666 encountered with the same name. One or both of the two symbols will be
1671 Turning a common symbol into a reference, because there is already a
1672 definition for the symbol.
1674 @var{file}(@var{section}): warning: common of `@var{symbol}'
1675 overridden by definition
1676 @var{file}(@var{section}): warning: defined here
1680 Turning a common symbol into a reference, because a later definition for
1681 the symbol is encountered. This is the same as the previous case,
1682 except that the symbols are encountered in a different order.
1684 @var{file}(@var{section}): warning: definition of `@var{symbol}'
1686 @var{file}(@var{section}): warning: common is here
1690 Merging a common symbol with a previous same-sized common symbol.
1692 @var{file}(@var{section}): warning: multiple common
1694 @var{file}(@var{section}): warning: previous common is here
1698 Merging a common symbol with a previous larger common symbol.
1700 @var{file}(@var{section}): warning: common of `@var{symbol}'
1701 overridden by larger common
1702 @var{file}(@var{section}): warning: larger common is here
1706 Merging a common symbol with a previous smaller common symbol. This is
1707 the same as the previous case, except that the symbols are
1708 encountered in a different order.
1710 @var{file}(@var{section}): warning: common of `@var{symbol}'
1711 overriding smaller common
1712 @var{file}(@var{section}): warning: smaller common is here
1716 @kindex --warn-constructors
1717 @item --warn-constructors
1718 Warn if any global constructors are used. This is only useful for a few
1719 object file formats. For formats like COFF or ELF, the linker can not
1720 detect the use of global constructors.
1722 @kindex --warn-multiple-gp
1723 @item --warn-multiple-gp
1724 Warn if multiple global pointer values are required in the output file.
1725 This is only meaningful for certain processors, such as the Alpha.
1726 Specifically, some processors put large-valued constants in a special
1727 section. A special register (the global pointer) points into the middle
1728 of this section, so that constants can be loaded efficiently via a
1729 base-register relative addressing mode. Since the offset in
1730 base-register relative mode is fixed and relatively small (e.g., 16
1731 bits), this limits the maximum size of the constant pool. Thus, in
1732 large programs, it is often necessary to use multiple global pointer
1733 values in order to be able to address all possible constants. This
1734 option causes a warning to be issued whenever this case occurs.
1737 @cindex warnings, on undefined symbols
1738 @cindex undefined symbols, warnings on
1740 Only warn once for each undefined symbol, rather than once per module
1743 @kindex --warn-section-align
1744 @cindex warnings, on section alignment
1745 @cindex section alignment, warnings on
1746 @item --warn-section-align
1747 Warn if the address of an output section is changed because of
1748 alignment. Typically, the alignment will be set by an input section.
1749 The address will only be changed if it not explicitly specified; that
1750 is, if the @code{SECTIONS} command does not specify a start address for
1751 the section (@pxref{SECTIONS}).
1753 @kindex --warn-shared-textrel
1754 @item --warn-shared-textrel
1755 Warn if the linker adds a DT_TEXTREL to a shared object.
1757 @kindex --warn-unresolved-symbols
1758 @item --warn-unresolved-symbols
1759 If the linker is going to report an unresolved symbol (see the option
1760 @option{--unresolved-symbols}) it will normally generate an error.
1761 This option makes it generate a warning instead.
1763 @kindex --error-unresolved-symbols
1764 @item --error-unresolved-symbols
1765 This restores the linker's default behaviour of generating errors when
1766 it is reporting unresolved symbols.
1768 @kindex --whole-archive
1769 @cindex including an entire archive
1770 @item --whole-archive
1771 For each archive mentioned on the command line after the
1772 @option{--whole-archive} option, include every object file in the archive
1773 in the link, rather than searching the archive for the required object
1774 files. This is normally used to turn an archive file into a shared
1775 library, forcing every object to be included in the resulting shared
1776 library. This option may be used more than once.
1778 Two notes when using this option from gcc: First, gcc doesn't know
1779 about this option, so you have to use @option{-Wl,-whole-archive}.
1780 Second, don't forget to use @option{-Wl,-no-whole-archive} after your
1781 list of archives, because gcc will add its own list of archives to
1782 your link and you may not want this flag to affect those as well.
1785 @item --wrap @var{symbol}
1786 Use a wrapper function for @var{symbol}. Any undefined reference to
1787 @var{symbol} will be resolved to @code{__wrap_@var{symbol}}. Any
1788 undefined reference to @code{__real_@var{symbol}} will be resolved to
1791 This can be used to provide a wrapper for a system function. The
1792 wrapper function should be called @code{__wrap_@var{symbol}}. If it
1793 wishes to call the system function, it should call
1794 @code{__real_@var{symbol}}.
1796 Here is a trivial example:
1800 __wrap_malloc (size_t c)
1802 printf ("malloc called with %zu\n", c);
1803 return __real_malloc (c);
1807 If you link other code with this file using @option{--wrap malloc}, then
1808 all calls to @code{malloc} will call the function @code{__wrap_malloc}
1809 instead. The call to @code{__real_malloc} in @code{__wrap_malloc} will
1810 call the real @code{malloc} function.
1812 You may wish to provide a @code{__real_malloc} function as well, so that
1813 links without the @option{--wrap} option will succeed. If you do this,
1814 you should not put the definition of @code{__real_malloc} in the same
1815 file as @code{__wrap_malloc}; if you do, the assembler may resolve the
1816 call before the linker has a chance to wrap it to @code{malloc}.
1818 @kindex --enable-new-dtags
1819 @kindex --disable-new-dtags
1820 @item --enable-new-dtags
1821 @itemx --disable-new-dtags
1822 This linker can create the new dynamic tags in ELF. But the older ELF
1823 systems may not understand them. If you specify
1824 @option{--enable-new-dtags}, the dynamic tags will be created as needed.
1825 If you specify @option{--disable-new-dtags}, no new dynamic tags will be
1826 created. By default, the new dynamic tags are not created. Note that
1827 those options are only available for ELF systems.
1829 @kindex --hash-size=@var{number}
1830 @item --hash-size=@var{number}
1831 Set the default size of the linker's hash tables to a prime number
1832 close to @var{number}. Increasing this value can reduce the length of
1833 time it takes the linker to perform its tasks, at the expense of
1834 increasing the linker's memory requirements. Similarly reducing this
1835 value can reduce the memory requirements at the expense of speed.
1837 @kindex --reduce-memory-overheads
1838 @item --reduce-memory-overheads
1839 This option reduces memory requirements at ld runtime, at the expense of
1840 linking speed. This was introduced to to select the old O(n^2) algorithm
1841 for link map file generation, rather than the new O(n) algorithm which uses
1842 about 40% more memory for symbol storage.
1844 Another affect of the switch is to set the default hash table size to
1845 1021, which again saves memory at the cost of lengthening the linker's
1846 run time. This is not done however if the @option{--hash-size} switch
1849 The @option{--reduce-memory-overheads} switch may be also be used to
1850 enable other tradeoffs in future versions of the linker.
1856 @subsection Options Specific to i386 PE Targets
1858 @c man begin OPTIONS
1860 The i386 PE linker supports the @option{-shared} option, which causes
1861 the output to be a dynamically linked library (DLL) instead of a
1862 normal executable. You should name the output @code{*.dll} when you
1863 use this option. In addition, the linker fully supports the standard
1864 @code{*.def} files, which may be specified on the linker command line
1865 like an object file (in fact, it should precede archives it exports
1866 symbols from, to ensure that they get linked in, just like a normal
1869 In addition to the options common to all targets, the i386 PE linker
1870 support additional command line options that are specific to the i386
1871 PE target. Options that take values may be separated from their
1872 values by either a space or an equals sign.
1876 @kindex --add-stdcall-alias
1877 @item --add-stdcall-alias
1878 If given, symbols with a stdcall suffix (@@@var{nn}) will be exported
1879 as-is and also with the suffix stripped.
1880 [This option is specific to the i386 PE targeted port of the linker]
1883 @item --base-file @var{file}
1884 Use @var{file} as the name of a file in which to save the base
1885 addresses of all the relocations needed for generating DLLs with
1887 [This is an i386 PE specific option]
1891 Create a DLL instead of a regular executable. You may also use
1892 @option{-shared} or specify a @code{LIBRARY} in a given @code{.def}
1894 [This option is specific to the i386 PE targeted port of the linker]
1896 @kindex --enable-stdcall-fixup
1897 @kindex --disable-stdcall-fixup
1898 @item --enable-stdcall-fixup
1899 @itemx --disable-stdcall-fixup
1900 If the link finds a symbol that it cannot resolve, it will attempt to
1901 do ``fuzzy linking'' by looking for another defined symbol that differs
1902 only in the format of the symbol name (cdecl vs stdcall) and will
1903 resolve that symbol by linking to the match. For example, the
1904 undefined symbol @code{_foo} might be linked to the function
1905 @code{_foo@@12}, or the undefined symbol @code{_bar@@16} might be linked
1906 to the function @code{_bar}. When the linker does this, it prints a
1907 warning, since it normally should have failed to link, but sometimes
1908 import libraries generated from third-party dlls may need this feature
1909 to be usable. If you specify @option{--enable-stdcall-fixup}, this
1910 feature is fully enabled and warnings are not printed. If you specify
1911 @option{--disable-stdcall-fixup}, this feature is disabled and such
1912 mismatches are considered to be errors.
1913 [This option is specific to the i386 PE targeted port of the linker]
1915 @cindex DLLs, creating
1916 @kindex --export-all-symbols
1917 @item --export-all-symbols
1918 If given, all global symbols in the objects used to build a DLL will
1919 be exported by the DLL. Note that this is the default if there
1920 otherwise wouldn't be any exported symbols. When symbols are
1921 explicitly exported via DEF files or implicitly exported via function
1922 attributes, the default is to not export anything else unless this
1923 option is given. Note that the symbols @code{DllMain@@12},
1924 @code{DllEntryPoint@@0}, @code{DllMainCRTStartup@@12}, and
1925 @code{impure_ptr} will not be automatically
1926 exported. Also, symbols imported from other DLLs will not be
1927 re-exported, nor will symbols specifying the DLL's internal layout
1928 such as those beginning with @code{_head_} or ending with
1929 @code{_iname}. In addition, no symbols from @code{libgcc},
1930 @code{libstd++}, @code{libmingw32}, or @code{crtX.o} will be exported.
1931 Symbols whose names begin with @code{__rtti_} or @code{__builtin_} will
1932 not be exported, to help with C++ DLLs. Finally, there is an
1933 extensive list of cygwin-private symbols that are not exported
1934 (obviously, this applies on when building DLLs for cygwin targets).
1935 These cygwin-excludes are: @code{_cygwin_dll_entry@@12},
1936 @code{_cygwin_crt0_common@@8}, @code{_cygwin_noncygwin_dll_entry@@12},
1937 @code{_fmode}, @code{_impure_ptr}, @code{cygwin_attach_dll},
1938 @code{cygwin_premain0}, @code{cygwin_premain1}, @code{cygwin_premain2},
1939 @code{cygwin_premain3}, and @code{environ}.
1940 [This option is specific to the i386 PE targeted port of the linker]
1942 @kindex --exclude-symbols
1943 @item --exclude-symbols @var{symbol},@var{symbol},...
1944 Specifies a list of symbols which should not be automatically
1945 exported. The symbol names may be delimited by commas or colons.
1946 [This option is specific to the i386 PE targeted port of the linker]
1948 @kindex --file-alignment
1949 @item --file-alignment
1950 Specify the file alignment. Sections in the file will always begin at
1951 file offsets which are multiples of this number. This defaults to
1953 [This option is specific to the i386 PE targeted port of the linker]
1957 @item --heap @var{reserve}
1958 @itemx --heap @var{reserve},@var{commit}
1959 Specify the amount of memory to reserve (and optionally commit) to be
1960 used as heap for this program. The default is 1Mb reserved, 4K
1962 [This option is specific to the i386 PE targeted port of the linker]
1965 @kindex --image-base
1966 @item --image-base @var{value}
1967 Use @var{value} as the base address of your program or dll. This is
1968 the lowest memory location that will be used when your program or dll
1969 is loaded. To reduce the need to relocate and improve performance of
1970 your dlls, each should have a unique base address and not overlap any
1971 other dlls. The default is 0x400000 for executables, and 0x10000000
1973 [This option is specific to the i386 PE targeted port of the linker]
1977 If given, the stdcall suffixes (@@@var{nn}) will be stripped from
1978 symbols before they are exported.
1979 [This option is specific to the i386 PE targeted port of the linker]
1981 @kindex --large-address-aware
1982 @item --large-address-aware
1983 If given, the appropriate bit in the ``Charateristics'' field of the COFF
1984 header is set to indicate that this executable supports virtual addresses
1985 greater than 2 gigabytes. This should be used in conjuction with the /3GB
1986 or /USERVA=@var{value} megabytes switch in the ``[operating systems]''
1987 section of the BOOT.INI. Otherwise, this bit has no effect.
1988 [This option is specific to PE targeted ports of the linker]
1990 @kindex --major-image-version
1991 @item --major-image-version @var{value}
1992 Sets the major number of the ``image version''. Defaults to 1.
1993 [This option is specific to the i386 PE targeted port of the linker]
1995 @kindex --major-os-version
1996 @item --major-os-version @var{value}
1997 Sets the major number of the ``os version''. Defaults to 4.
1998 [This option is specific to the i386 PE targeted port of the linker]
2000 @kindex --major-subsystem-version
2001 @item --major-subsystem-version @var{value}
2002 Sets the major number of the ``subsystem version''. Defaults to 4.
2003 [This option is specific to the i386 PE targeted port of the linker]
2005 @kindex --minor-image-version
2006 @item --minor-image-version @var{value}
2007 Sets the minor number of the ``image version''. Defaults to 0.
2008 [This option is specific to the i386 PE targeted port of the linker]
2010 @kindex --minor-os-version
2011 @item --minor-os-version @var{value}
2012 Sets the minor number of the ``os version''. Defaults to 0.
2013 [This option is specific to the i386 PE targeted port of the linker]
2015 @kindex --minor-subsystem-version
2016 @item --minor-subsystem-version @var{value}
2017 Sets the minor number of the ``subsystem version''. Defaults to 0.
2018 [This option is specific to the i386 PE targeted port of the linker]
2020 @cindex DEF files, creating
2021 @cindex DLLs, creating
2022 @kindex --output-def
2023 @item --output-def @var{file}
2024 The linker will create the file @var{file} which will contain a DEF
2025 file corresponding to the DLL the linker is generating. This DEF file
2026 (which should be called @code{*.def}) may be used to create an import
2027 library with @code{dlltool} or may be used as a reference to
2028 automatically or implicitly exported symbols.
2029 [This option is specific to the i386 PE targeted port of the linker]
2031 @cindex DLLs, creating
2032 @kindex --out-implib
2033 @item --out-implib @var{file}
2034 The linker will create the file @var{file} which will contain an
2035 import lib corresponding to the DLL the linker is generating. This
2036 import lib (which should be called @code{*.dll.a} or @code{*.a}
2037 may be used to link clients against the generated DLL; this behaviour
2038 makes it possible to skip a separate @code{dlltool} import library
2040 [This option is specific to the i386 PE targeted port of the linker]
2042 @kindex --enable-auto-image-base
2043 @item --enable-auto-image-base
2044 Automatically choose the image base for DLLs, unless one is specified
2045 using the @code{--image-base} argument. By using a hash generated
2046 from the dllname to create unique image bases for each DLL, in-memory
2047 collisions and relocations which can delay program execution are
2049 [This option is specific to the i386 PE targeted port of the linker]
2051 @kindex --disable-auto-image-base
2052 @item --disable-auto-image-base
2053 Do not automatically generate a unique image base. If there is no
2054 user-specified image base (@code{--image-base}) then use the platform
2056 [This option is specific to the i386 PE targeted port of the linker]
2058 @cindex DLLs, linking to
2059 @kindex --dll-search-prefix
2060 @item --dll-search-prefix @var{string}
2061 When linking dynamically to a dll without an import library,
2062 search for @code{<string><basename>.dll} in preference to
2063 @code{lib<basename>.dll}. This behaviour allows easy distinction
2064 between DLLs built for the various "subplatforms": native, cygwin,
2065 uwin, pw, etc. For instance, cygwin DLLs typically use
2066 @code{--dll-search-prefix=cyg}.
2067 [This option is specific to the i386 PE targeted port of the linker]
2069 @kindex --enable-auto-import
2070 @item --enable-auto-import
2071 Do sophisticated linking of @code{_symbol} to @code{__imp__symbol} for
2072 DATA imports from DLLs, and create the necessary thunking symbols when
2073 building the import libraries with those DATA exports. Note: Use of the
2074 'auto-import' extension will cause the text section of the image file
2075 to be made writable. This does not conform to the PE-COFF format
2076 specification published by Microsoft.
2078 Using 'auto-import' generally will 'just work' -- but sometimes you may
2081 "variable '<var>' can't be auto-imported. Please read the
2082 documentation for ld's @code{--enable-auto-import} for details."
2084 This message occurs when some (sub)expression accesses an address
2085 ultimately given by the sum of two constants (Win32 import tables only
2086 allow one). Instances where this may occur include accesses to member
2087 fields of struct variables imported from a DLL, as well as using a
2088 constant index into an array variable imported from a DLL. Any
2089 multiword variable (arrays, structs, long long, etc) may trigger
2090 this error condition. However, regardless of the exact data type
2091 of the offending exported variable, ld will always detect it, issue
2092 the warning, and exit.
2094 There are several ways to address this difficulty, regardless of the
2095 data type of the exported variable:
2097 One way is to use --enable-runtime-pseudo-reloc switch. This leaves the task
2098 of adjusting references in your client code for runtime environment, so
2099 this method works only when runtime environment supports this feature.
2101 A second solution is to force one of the 'constants' to be a variable --
2102 that is, unknown and un-optimizable at compile time. For arrays,
2103 there are two possibilities: a) make the indexee (the array's address)
2104 a variable, or b) make the 'constant' index a variable. Thus:
2107 extern type extern_array[];
2109 @{ volatile type *t=extern_array; t[1] @}
2115 extern type extern_array[];
2117 @{ volatile int t=1; extern_array[t] @}
2120 For structs (and most other multiword data types) the only option
2121 is to make the struct itself (or the long long, or the ...) variable:
2124 extern struct s extern_struct;
2125 extern_struct.field -->
2126 @{ volatile struct s *t=&extern_struct; t->field @}
2132 extern long long extern_ll;
2134 @{ volatile long long * local_ll=&extern_ll; *local_ll @}
2137 A third method of dealing with this difficulty is to abandon
2138 'auto-import' for the offending symbol and mark it with
2139 @code{__declspec(dllimport)}. However, in practise that
2140 requires using compile-time #defines to indicate whether you are
2141 building a DLL, building client code that will link to the DLL, or
2142 merely building/linking to a static library. In making the choice
2143 between the various methods of resolving the 'direct address with
2144 constant offset' problem, you should consider typical real-world usage:
2152 void main(int argc, char **argv)@{
2153 printf("%d\n",arr[1]);
2163 void main(int argc, char **argv)@{
2164 /* This workaround is for win32 and cygwin; do not "optimize" */
2165 volatile int *parr = arr;
2166 printf("%d\n",parr[1]);
2173 /* Note: auto-export is assumed (no __declspec(dllexport)) */
2174 #if (defined(_WIN32) || defined(__CYGWIN__)) && \
2175 !(defined(FOO_BUILD_DLL) || defined(FOO_STATIC))
2176 #define FOO_IMPORT __declspec(dllimport)
2180 extern FOO_IMPORT int arr[];
2183 void main(int argc, char **argv)@{
2184 printf("%d\n",arr[1]);
2188 A fourth way to avoid this problem is to re-code your
2189 library to use a functional interface rather than a data interface
2190 for the offending variables (e.g. set_foo() and get_foo() accessor
2192 [This option is specific to the i386 PE targeted port of the linker]
2194 @kindex --disable-auto-import
2195 @item --disable-auto-import
2196 Do not attempt to do sophisticated linking of @code{_symbol} to
2197 @code{__imp__symbol} for DATA imports from DLLs.
2198 [This option is specific to the i386 PE targeted port of the linker]
2200 @kindex --enable-runtime-pseudo-reloc
2201 @item --enable-runtime-pseudo-reloc
2202 If your code contains expressions described in --enable-auto-import section,
2203 that is, DATA imports from DLL with non-zero offset, this switch will create
2204 a vector of 'runtime pseudo relocations' which can be used by runtime
2205 environment to adjust references to such data in your client code.
2206 [This option is specific to the i386 PE targeted port of the linker]
2208 @kindex --disable-runtime-pseudo-reloc
2209 @item --disable-runtime-pseudo-reloc
2210 Do not create pseudo relocations for non-zero offset DATA imports from
2211 DLLs. This is the default.
2212 [This option is specific to the i386 PE targeted port of the linker]
2214 @kindex --enable-extra-pe-debug
2215 @item --enable-extra-pe-debug
2216 Show additional debug info related to auto-import symbol thunking.
2217 [This option is specific to the i386 PE targeted port of the linker]
2219 @kindex --section-alignment
2220 @item --section-alignment
2221 Sets the section alignment. Sections in memory will always begin at
2222 addresses which are a multiple of this number. Defaults to 0x1000.
2223 [This option is specific to the i386 PE targeted port of the linker]
2227 @item --stack @var{reserve}
2228 @itemx --stack @var{reserve},@var{commit}
2229 Specify the amount of memory to reserve (and optionally commit) to be
2230 used as stack for this program. The default is 2Mb reserved, 4K
2232 [This option is specific to the i386 PE targeted port of the linker]
2235 @item --subsystem @var{which}
2236 @itemx --subsystem @var{which}:@var{major}
2237 @itemx --subsystem @var{which}:@var{major}.@var{minor}
2238 Specifies the subsystem under which your program will execute. The
2239 legal values for @var{which} are @code{native}, @code{windows},
2240 @code{console}, @code{posix}, and @code{xbox}. You may optionally set
2241 the subsystem version also. Numeric values are also accepted for
2243 [This option is specific to the i386 PE targeted port of the linker]
2250 @subsection Options specific to Motorola 68HC11 and 68HC12 targets
2252 @c man begin OPTIONS
2254 The 68HC11 and 68HC12 linkers support specific options to control the
2255 memory bank switching mapping and trampoline code generation.
2259 @kindex --no-trampoline
2260 @item --no-trampoline
2261 This option disables the generation of trampoline. By default a trampoline
2262 is generated for each far function which is called using a @code{jsr}
2263 instruction (this happens when a pointer to a far function is taken).
2265 @kindex --bank-window
2266 @item --bank-window @var{name}
2267 This option indicates to the linker the name of the memory region in
2268 the @samp{MEMORY} specification that describes the memory bank window.
2269 The definition of such region is then used by the linker to compute
2270 paging and addresses within the memory window.
2279 @section Environment Variables
2281 @c man begin ENVIRONMENT
2283 You can change the behaviour of @command{ld} with the environment variables
2284 @ifclear SingleFormat
2287 @code{LDEMULATION} and @code{COLLECT_NO_DEMANGLE}.
2289 @ifclear SingleFormat
2291 @cindex default input format
2292 @code{GNUTARGET} determines the input-file object format if you don't
2293 use @samp{-b} (or its synonym @samp{--format}). Its value should be one
2294 of the BFD names for an input format (@pxref{BFD}). If there is no
2295 @code{GNUTARGET} in the environment, @command{ld} uses the natural format
2296 of the target. If @code{GNUTARGET} is set to @code{default} then BFD
2297 attempts to discover the input format by examining binary input files;
2298 this method often succeeds, but there are potential ambiguities, since
2299 there is no method of ensuring that the magic number used to specify
2300 object-file formats is unique. However, the configuration procedure for
2301 BFD on each system places the conventional format for that system first
2302 in the search-list, so ambiguities are resolved in favor of convention.
2306 @cindex default emulation
2307 @cindex emulation, default
2308 @code{LDEMULATION} determines the default emulation if you don't use the
2309 @samp{-m} option. The emulation can affect various aspects of linker
2310 behaviour, particularly the default linker script. You can list the
2311 available emulations with the @samp{--verbose} or @samp{-V} options. If
2312 the @samp{-m} option is not used, and the @code{LDEMULATION} environment
2313 variable is not defined, the default emulation depends upon how the
2314 linker was configured.
2316 @kindex COLLECT_NO_DEMANGLE
2317 @cindex demangling, default
2318 Normally, the linker will default to demangling symbols. However, if
2319 @code{COLLECT_NO_DEMANGLE} is set in the environment, then it will
2320 default to not demangling symbols. This environment variable is used in
2321 a similar fashion by the @code{gcc} linker wrapper program. The default
2322 may be overridden by the @samp{--demangle} and @samp{--no-demangle}
2329 @chapter Linker Scripts
2332 @cindex linker scripts
2333 @cindex command files
2334 Every link is controlled by a @dfn{linker script}. This script is
2335 written in the linker command language.
2337 The main purpose of the linker script is to describe how the sections in
2338 the input files should be mapped into the output file, and to control
2339 the memory layout of the output file. Most linker scripts do nothing
2340 more than this. However, when necessary, the linker script can also
2341 direct the linker to perform many other operations, using the commands
2344 The linker always uses a linker script. If you do not supply one
2345 yourself, the linker will use a default script that is compiled into the
2346 linker executable. You can use the @samp{--verbose} command line option
2347 to display the default linker script. Certain command line options,
2348 such as @samp{-r} or @samp{-N}, will affect the default linker script.
2350 You may supply your own linker script by using the @samp{-T} command
2351 line option. When you do this, your linker script will replace the
2352 default linker script.
2354 You may also use linker scripts implicitly by naming them as input files
2355 to the linker, as though they were files to be linked. @xref{Implicit
2359 * Basic Script Concepts:: Basic Linker Script Concepts
2360 * Script Format:: Linker Script Format
2361 * Simple Example:: Simple Linker Script Example
2362 * Simple Commands:: Simple Linker Script Commands
2363 * Assignments:: Assigning Values to Symbols
2364 * SECTIONS:: SECTIONS Command
2365 * MEMORY:: MEMORY Command
2366 * PHDRS:: PHDRS Command
2367 * VERSION:: VERSION Command
2368 * Expressions:: Expressions in Linker Scripts
2369 * Implicit Linker Scripts:: Implicit Linker Scripts
2372 @node Basic Script Concepts
2373 @section Basic Linker Script Concepts
2374 @cindex linker script concepts
2375 We need to define some basic concepts and vocabulary in order to
2376 describe the linker script language.
2378 The linker combines input files into a single output file. The output
2379 file and each input file are in a special data format known as an
2380 @dfn{object file format}. Each file is called an @dfn{object file}.
2381 The output file is often called an @dfn{executable}, but for our
2382 purposes we will also call it an object file. Each object file has,
2383 among other things, a list of @dfn{sections}. We sometimes refer to a
2384 section in an input file as an @dfn{input section}; similarly, a section
2385 in the output file is an @dfn{output section}.
2387 Each section in an object file has a name and a size. Most sections
2388 also have an associated block of data, known as the @dfn{section
2389 contents}. A section may be marked as @dfn{loadable}, which mean that
2390 the contents should be loaded into memory when the output file is run.
2391 A section with no contents may be @dfn{allocatable}, which means that an
2392 area in memory should be set aside, but nothing in particular should be
2393 loaded there (in some cases this memory must be zeroed out). A section
2394 which is neither loadable nor allocatable typically contains some sort
2395 of debugging information.
2397 Every loadable or allocatable output section has two addresses. The
2398 first is the @dfn{VMA}, or virtual memory address. This is the address
2399 the section will have when the output file is run. The second is the
2400 @dfn{LMA}, or load memory address. This is the address at which the
2401 section will be loaded. In most cases the two addresses will be the
2402 same. An example of when they might be different is when a data section
2403 is loaded into ROM, and then copied into RAM when the program starts up
2404 (this technique is often used to initialize global variables in a ROM
2405 based system). In this case the ROM address would be the LMA, and the
2406 RAM address would be the VMA.
2408 You can see the sections in an object file by using the @code{objdump}
2409 program with the @samp{-h} option.
2411 Every object file also has a list of @dfn{symbols}, known as the
2412 @dfn{symbol table}. A symbol may be defined or undefined. Each symbol
2413 has a name, and each defined symbol has an address, among other
2414 information. If you compile a C or C++ program into an object file, you
2415 will get a defined symbol for every defined function and global or
2416 static variable. Every undefined function or global variable which is
2417 referenced in the input file will become an undefined symbol.
2419 You can see the symbols in an object file by using the @code{nm}
2420 program, or by using the @code{objdump} program with the @samp{-t}
2424 @section Linker Script Format
2425 @cindex linker script format
2426 Linker scripts are text files.
2428 You write a linker script as a series of commands. Each command is
2429 either a keyword, possibly followed by arguments, or an assignment to a
2430 symbol. You may separate commands using semicolons. Whitespace is
2433 Strings such as file or format names can normally be entered directly.
2434 If the file name contains a character such as a comma which would
2435 otherwise serve to separate file names, you may put the file name in
2436 double quotes. There is no way to use a double quote character in a
2439 You may include comments in linker scripts just as in C, delimited by
2440 @samp{/*} and @samp{*/}. As in C, comments are syntactically equivalent
2443 @node Simple Example
2444 @section Simple Linker Script Example
2445 @cindex linker script example
2446 @cindex example of linker script
2447 Many linker scripts are fairly simple.
2449 The simplest possible linker script has just one command:
2450 @samp{SECTIONS}. You use the @samp{SECTIONS} command to describe the
2451 memory layout of the output file.
2453 The @samp{SECTIONS} command is a powerful command. Here we will
2454 describe a simple use of it. Let's assume your program consists only of
2455 code, initialized data, and uninitialized data. These will be in the
2456 @samp{.text}, @samp{.data}, and @samp{.bss} sections, respectively.
2457 Let's assume further that these are the only sections which appear in
2460 For this example, let's say that the code should be loaded at address
2461 0x10000, and that the data should start at address 0x8000000. Here is a
2462 linker script which will do that:
2467 .text : @{ *(.text) @}
2469 .data : @{ *(.data) @}
2470 .bss : @{ *(.bss) @}
2474 You write the @samp{SECTIONS} command as the keyword @samp{SECTIONS},
2475 followed by a series of symbol assignments and output section
2476 descriptions enclosed in curly braces.
2478 The first line inside the @samp{SECTIONS} command of the above example
2479 sets the value of the special symbol @samp{.}, which is the location
2480 counter. If you do not specify the address of an output section in some
2481 other way (other ways are described later), the address is set from the
2482 current value of the location counter. The location counter is then
2483 incremented by the size of the output section. At the start of the
2484 @samp{SECTIONS} command, the location counter has the value @samp{0}.
2486 The second line defines an output section, @samp{.text}. The colon is
2487 required syntax which may be ignored for now. Within the curly braces
2488 after the output section name, you list the names of the input sections
2489 which should be placed into this output section. The @samp{*} is a
2490 wildcard which matches any file name. The expression @samp{*(.text)}
2491 means all @samp{.text} input sections in all input files.
2493 Since the location counter is @samp{0x10000} when the output section
2494 @samp{.text} is defined, the linker will set the address of the
2495 @samp{.text} section in the output file to be @samp{0x10000}.
2497 The remaining lines define the @samp{.data} and @samp{.bss} sections in
2498 the output file. The linker will place the @samp{.data} output section
2499 at address @samp{0x8000000}. After the linker places the @samp{.data}
2500 output section, the value of the location counter will be
2501 @samp{0x8000000} plus the size of the @samp{.data} output section. The
2502 effect is that the linker will place the @samp{.bss} output section
2503 immediately after the @samp{.data} output section in memory.
2505 The linker will ensure that each output section has the required
2506 alignment, by increasing the location counter if necessary. In this
2507 example, the specified addresses for the @samp{.text} and @samp{.data}
2508 sections will probably satisfy any alignment constraints, but the linker
2509 may have to create a small gap between the @samp{.data} and @samp{.bss}
2512 That's it! That's a simple and complete linker script.
2514 @node Simple Commands
2515 @section Simple Linker Script Commands
2516 @cindex linker script simple commands
2517 In this section we describe the simple linker script commands.
2520 * Entry Point:: Setting the entry point
2521 * File Commands:: Commands dealing with files
2522 @ifclear SingleFormat
2523 * Format Commands:: Commands dealing with object file formats
2526 * Miscellaneous Commands:: Other linker script commands
2530 @subsection Setting the Entry Point
2531 @kindex ENTRY(@var{symbol})
2532 @cindex start of execution
2533 @cindex first instruction
2535 The first instruction to execute in a program is called the @dfn{entry
2536 point}. You can use the @code{ENTRY} linker script command to set the
2537 entry point. The argument is a symbol name:
2542 There are several ways to set the entry point. The linker will set the
2543 entry point by trying each of the following methods in order, and
2544 stopping when one of them succeeds:
2547 the @samp{-e} @var{entry} command-line option;
2549 the @code{ENTRY(@var{symbol})} command in a linker script;
2551 the value of the symbol @code{start}, if defined;
2553 the address of the first byte of the @samp{.text} section, if present;
2555 The address @code{0}.
2559 @subsection Commands Dealing with Files
2560 @cindex linker script file commands
2561 Several linker script commands deal with files.
2564 @item INCLUDE @var{filename}
2565 @kindex INCLUDE @var{filename}
2566 @cindex including a linker script
2567 Include the linker script @var{filename} at this point. The file will
2568 be searched for in the current directory, and in any directory specified
2569 with the @option{-L} option. You can nest calls to @code{INCLUDE} up to
2572 @item INPUT(@var{file}, @var{file}, @dots{})
2573 @itemx INPUT(@var{file} @var{file} @dots{})
2574 @kindex INPUT(@var{files})
2575 @cindex input files in linker scripts
2576 @cindex input object files in linker scripts
2577 @cindex linker script input object files
2578 The @code{INPUT} command directs the linker to include the named files
2579 in the link, as though they were named on the command line.
2581 For example, if you always want to include @file{subr.o} any time you do
2582 a link, but you can't be bothered to put it on every link command line,
2583 then you can put @samp{INPUT (subr.o)} in your linker script.
2585 In fact, if you like, you can list all of your input files in the linker
2586 script, and then invoke the linker with nothing but a @samp{-T} option.
2588 In case a @dfn{sysroot prefix} is configured, and the filename starts
2589 with the @samp{/} character, and the script being processed was
2590 located inside the @dfn{sysroot prefix}, the filename will be looked
2591 for in the @dfn{sysroot prefix}. Otherwise, the linker will try to
2592 open the file in the current directory. If it is not found, the
2593 linker will search through the archive library search path. See the
2594 description of @samp{-L} in @ref{Options,,Command Line Options}.
2596 If you use @samp{INPUT (-l@var{file})}, @command{ld} will transform the
2597 name to @code{lib@var{file}.a}, as with the command line argument
2600 When you use the @code{INPUT} command in an implicit linker script, the
2601 files will be included in the link at the point at which the linker
2602 script file is included. This can affect archive searching.
2604 @item GROUP(@var{file}, @var{file}, @dots{})
2605 @itemx GROUP(@var{file} @var{file} @dots{})
2606 @kindex GROUP(@var{files})
2607 @cindex grouping input files
2608 The @code{GROUP} command is like @code{INPUT}, except that the named
2609 files should all be archives, and they are searched repeatedly until no
2610 new undefined references are created. See the description of @samp{-(}
2611 in @ref{Options,,Command Line Options}.
2613 @item AS_NEEDED(@var{file}, @var{file}, @dots{})
2614 @itemx AS_NEEDED(@var{file} @var{file} @dots{})
2615 @kindex AS_NEEDED(@var{files})
2616 This construct can appear only inside of the @code{INPUT} or @code{GROUP}
2617 commands, among other filenames. The files listed will be handled
2618 as if they appear directly in the @code{INPUT} or @code{GROUP} commands,
2619 with the exception of ELF shared libraries, that will be added only
2620 when they are actually needed. This construct essentially enables
2621 @option{--as-needed} option for all the files listed inside of it
2622 and restores previous @option{--as-needed} resp. @option{--no-as-needed}
2625 @item OUTPUT(@var{filename})
2626 @kindex OUTPUT(@var{filename})
2627 @cindex output file name in linker scripot
2628 The @code{OUTPUT} command names the output file. Using
2629 @code{OUTPUT(@var{filename})} in the linker script is exactly like using
2630 @samp{-o @var{filename}} on the command line (@pxref{Options,,Command
2631 Line Options}). If both are used, the command line option takes
2634 You can use the @code{OUTPUT} command to define a default name for the
2635 output file other than the usual default of @file{a.out}.
2637 @item SEARCH_DIR(@var{path})
2638 @kindex SEARCH_DIR(@var{path})
2639 @cindex library search path in linker script
2640 @cindex archive search path in linker script
2641 @cindex search path in linker script
2642 The @code{SEARCH_DIR} command adds @var{path} to the list of paths where
2643 @command{ld} looks for archive libraries. Using
2644 @code{SEARCH_DIR(@var{path})} is exactly like using @samp{-L @var{path}}
2645 on the command line (@pxref{Options,,Command Line Options}). If both
2646 are used, then the linker will search both paths. Paths specified using
2647 the command line option are searched first.
2649 @item STARTUP(@var{filename})
2650 @kindex STARTUP(@var{filename})
2651 @cindex first input file
2652 The @code{STARTUP} command is just like the @code{INPUT} command, except
2653 that @var{filename} will become the first input file to be linked, as
2654 though it were specified first on the command line. This may be useful
2655 when using a system in which the entry point is always the start of the
2659 @ifclear SingleFormat
2660 @node Format Commands
2661 @subsection Commands Dealing with Object File Formats
2662 A couple of linker script commands deal with object file formats.
2665 @item OUTPUT_FORMAT(@var{bfdname})
2666 @itemx OUTPUT_FORMAT(@var{default}, @var{big}, @var{little})
2667 @kindex OUTPUT_FORMAT(@var{bfdname})
2668 @cindex output file format in linker script
2669 The @code{OUTPUT_FORMAT} command names the BFD format to use for the
2670 output file (@pxref{BFD}). Using @code{OUTPUT_FORMAT(@var{bfdname})} is
2671 exactly like using @samp{--oformat @var{bfdname}} on the command line
2672 (@pxref{Options,,Command Line Options}). If both are used, the command
2673 line option takes precedence.
2675 You can use @code{OUTPUT_FORMAT} with three arguments to use different
2676 formats based on the @samp{-EB} and @samp{-EL} command line options.
2677 This permits the linker script to set the output format based on the
2680 If neither @samp{-EB} nor @samp{-EL} are used, then the output format
2681 will be the first argument, @var{default}. If @samp{-EB} is used, the
2682 output format will be the second argument, @var{big}. If @samp{-EL} is
2683 used, the output format will be the third argument, @var{little}.
2685 For example, the default linker script for the MIPS ELF target uses this
2688 OUTPUT_FORMAT(elf32-bigmips, elf32-bigmips, elf32-littlemips)
2690 This says that the default format for the output file is
2691 @samp{elf32-bigmips}, but if the user uses the @samp{-EL} command line
2692 option, the output file will be created in the @samp{elf32-littlemips}
2695 @item TARGET(@var{bfdname})
2696 @kindex TARGET(@var{bfdname})
2697 @cindex input file format in linker script
2698 The @code{TARGET} command names the BFD format to use when reading input
2699 files. It affects subsequent @code{INPUT} and @code{GROUP} commands.
2700 This command is like using @samp{-b @var{bfdname}} on the command line
2701 (@pxref{Options,,Command Line Options}). If the @code{TARGET} command
2702 is used but @code{OUTPUT_FORMAT} is not, then the last @code{TARGET}
2703 command is also used to set the format for the output file. @xref{BFD}.
2707 @node Miscellaneous Commands
2708 @subsection Other Linker Script Commands
2709 There are a few other linker scripts commands.
2712 @item ASSERT(@var{exp}, @var{message})
2714 @cindex assertion in linker script
2715 Ensure that @var{exp} is non-zero. If it is zero, then exit the linker
2716 with an error code, and print @var{message}.
2718 @item EXTERN(@var{symbol} @var{symbol} @dots{})
2720 @cindex undefined symbol in linker script
2721 Force @var{symbol} to be entered in the output file as an undefined
2722 symbol. Doing this may, for example, trigger linking of additional
2723 modules from standard libraries. You may list several @var{symbol}s for
2724 each @code{EXTERN}, and you may use @code{EXTERN} multiple times. This
2725 command has the same effect as the @samp{-u} command-line option.
2727 @item FORCE_COMMON_ALLOCATION
2728 @kindex FORCE_COMMON_ALLOCATION
2729 @cindex common allocation in linker script
2730 This command has the same effect as the @samp{-d} command-line option:
2731 to make @command{ld} assign space to common symbols even if a relocatable
2732 output file is specified (@samp{-r}).
2734 @item INHIBIT_COMMON_ALLOCATION
2735 @kindex INHIBIT_COMMON_ALLOCATION
2736 @cindex common allocation in linker script
2737 This command has the same effect as the @samp{--no-define-common}
2738 command-line option: to make @code{ld} omit the assignment of addresses
2739 to common symbols even for a non-relocatable output file.
2741 @item NOCROSSREFS(@var{section} @var{section} @dots{})
2742 @kindex NOCROSSREFS(@var{sections})
2743 @cindex cross references
2744 This command may be used to tell @command{ld} to issue an error about any
2745 references among certain output sections.
2747 In certain types of programs, particularly on embedded systems when
2748 using overlays, when one section is loaded into memory, another section
2749 will not be. Any direct references between the two sections would be
2750 errors. For example, it would be an error if code in one section called
2751 a function defined in the other section.
2753 The @code{NOCROSSREFS} command takes a list of output section names. If
2754 @command{ld} detects any cross references between the sections, it reports
2755 an error and returns a non-zero exit status. Note that the
2756 @code{NOCROSSREFS} command uses output section names, not input section
2759 @ifclear SingleFormat
2760 @item OUTPUT_ARCH(@var{bfdarch})
2761 @kindex OUTPUT_ARCH(@var{bfdarch})
2762 @cindex machine architecture
2763 @cindex architecture
2764 Specify a particular output machine architecture. The argument is one
2765 of the names used by the BFD library (@pxref{BFD}). You can see the
2766 architecture of an object file by using the @code{objdump} program with
2767 the @samp{-f} option.
2772 @section Assigning Values to Symbols
2773 @cindex assignment in scripts
2774 @cindex symbol definition, scripts
2775 @cindex variables, defining
2776 You may assign a value to a symbol in a linker script. This will define
2777 the symbol and place it into the symbol table with a global scope.
2780 * Simple Assignments:: Simple Assignments
2782 * PROVIDE_HIDDEN:: PROVIDE_HIDDEN
2783 * Source Code Reference:: How to use a linker script defined symbol in source code
2786 @node Simple Assignments
2787 @subsection Simple Assignments
2789 You may assign to a symbol using any of the C assignment operators:
2792 @item @var{symbol} = @var{expression} ;
2793 @itemx @var{symbol} += @var{expression} ;
2794 @itemx @var{symbol} -= @var{expression} ;
2795 @itemx @var{symbol} *= @var{expression} ;
2796 @itemx @var{symbol} /= @var{expression} ;
2797 @itemx @var{symbol} <<= @var{expression} ;
2798 @itemx @var{symbol} >>= @var{expression} ;
2799 @itemx @var{symbol} &= @var{expression} ;
2800 @itemx @var{symbol} |= @var{expression} ;
2803 The first case will define @var{symbol} to the value of
2804 @var{expression}. In the other cases, @var{symbol} must already be
2805 defined, and the value will be adjusted accordingly.
2807 The special symbol name @samp{.} indicates the location counter. You
2808 may only use this within a @code{SECTIONS} command. @xref{Location Counter}.
2810 The semicolon after @var{expression} is required.
2812 Expressions are defined below; see @ref{Expressions}.
2814 You may write symbol assignments as commands in their own right, or as
2815 statements within a @code{SECTIONS} command, or as part of an output
2816 section description in a @code{SECTIONS} command.
2818 The section of the symbol will be set from the section of the
2819 expression; for more information, see @ref{Expression Section}.
2821 Here is an example showing the three different places that symbol
2822 assignments may be used:
2833 _bdata = (. + 3) & ~ 3;
2834 .data : @{ *(.data) @}
2838 In this example, the symbol @samp{floating_point} will be defined as
2839 zero. The symbol @samp{_etext} will be defined as the address following
2840 the last @samp{.text} input section. The symbol @samp{_bdata} will be
2841 defined as the address following the @samp{.text} output section aligned
2842 upward to a 4 byte boundary.
2847 In some cases, it is desirable for a linker script to define a symbol
2848 only if it is referenced and is not defined by any object included in
2849 the link. For example, traditional linkers defined the symbol
2850 @samp{etext}. However, ANSI C requires that the user be able to use
2851 @samp{etext} as a function name without encountering an error. The
2852 @code{PROVIDE} keyword may be used to define a symbol, such as
2853 @samp{etext}, only if it is referenced but not defined. The syntax is
2854 @code{PROVIDE(@var{symbol} = @var{expression})}.
2856 Here is an example of using @code{PROVIDE} to define @samp{etext}:
2869 In this example, if the program defines @samp{_etext} (with a leading
2870 underscore), the linker will give a multiple definition error. If, on
2871 the other hand, the program defines @samp{etext} (with no leading
2872 underscore), the linker will silently use the definition in the program.
2873 If the program references @samp{etext} but does not define it, the
2874 linker will use the definition in the linker script.
2876 @node PROVIDE_HIDDEN
2877 @subsection PROVIDE_HIDDEN
2878 @cindex PROVIDE_HIDDEN
2879 Similar to @code{PROVIDE}. For ELF targeted ports, the symbol will be
2880 hidden and won't be exported.
2882 @node Source Code Reference
2883 @subsection Source Code Reference
2885 Accessing a linker script defined variable from source code is not
2886 intuitive. In particular a linker script symbol is not equivalent to
2887 a variable declaration in a high level language, it is instead a
2888 symbol that does not have a value.
2890 Before going further, it is important to note that compilers often
2891 transform names in the source code into different names when they are
2892 stored in the symbol table. For example, Fortran compilers commonly
2893 prepend or append an underscore, and C++ performs extensive @samp{name
2894 mangling}. Therefore there might be a discrepancy between the name
2895 of a variable as it is used in source code and the name of the same
2896 variable as it is defined in a linker script. For example in C a
2897 linker script variable might be referred to as:
2903 But in the linker script it might be defined as:
2909 In the remaining examples however it is assumed that no name
2910 transformation has taken place.
2912 When a symbol is declared in a high level language such as C, two
2913 things happen. The first is that the compiler reserves enough space
2914 in the program's memory to hold the @emph{value} of the symbol. The
2915 second is that the compiler creates an entry in the program's symbol
2916 table which holds the symbol's @emph{address}. ie the symbol table
2917 contains the address of the block of memory holding the symbol's
2918 value. So for example the following C declaration, at file scope:
2924 creates a entry called @samp{foo} in the symbol table. This entry
2925 holds the address of an @samp{int} sized block of memory where the
2926 number 1000 is initially stored.
2928 When a program references a symbol the compiler generates code that
2929 first accesses the symbol table to find the address of the symbol's
2930 memory block and then code to read the value from that memory block.
2937 looks up the symbol @samp{foo} in the symbol table, gets the address
2938 associated with this symbol and then writes the value 1 into that
2945 looks up the symbol @samp{foo} in the symbol table, gets it address
2946 and then copies this address into the block of memory associated with
2947 the variable @samp{a}.
2949 Linker scripts symbol declarations, by contrast, create an entry in
2950 the symbol table but do not assign any memory to them. Thus they are
2951 an address without a value. So for example the linker script definition:
2957 creates an entry in the symbol table called @samp{foo} which holds
2958 the address of memory location 1000, but nothing special is stored at
2959 address 1000. This means that you cannot access the @emph{value} of a
2960 linker script defined symbol - it has no value - all you can do is
2961 access the @emph{address} of a linker script defined symbol.
2963 Hence when you are using a linker script defined symbol in source code
2964 you should always take the address of the symbol, and never attempt to
2965 use its value. For example suppose you want to copy the contents of a
2966 section of memory called .ROM into a section called .FLASH and the
2967 linker script contains these declarations:
2971 start_of_ROM = .ROM;
2972 end_of_ROM = .ROM + sizeof (.ROM) - 1;
2973 start_of_FLASH = .FLASH;
2977 Then the C source code to perform the copy would be:
2981 extern char start_of_ROM, end_of_ROM, start_of_FLASH;
2983 memcpy (& start_of_FLASH, & start_of_ROM, & end_of_ROM - & start_of_ROM);
2987 Note the use of the @samp{&} operators. These are correct.
2990 @section SECTIONS Command
2992 The @code{SECTIONS} command tells the linker how to map input sections
2993 into output sections, and how to place the output sections in memory.
2995 The format of the @code{SECTIONS} command is:
2999 @var{sections-command}
3000 @var{sections-command}
3005 Each @var{sections-command} may of be one of the following:
3009 an @code{ENTRY} command (@pxref{Entry Point,,Entry command})
3011 a symbol assignment (@pxref{Assignments})
3013 an output section description
3015 an overlay description
3018 The @code{ENTRY} command and symbol assignments are permitted inside the
3019 @code{SECTIONS} command for convenience in using the location counter in
3020 those commands. This can also make the linker script easier to
3021 understand because you can use those commands at meaningful points in
3022 the layout of the output file.
3024 Output section descriptions and overlay descriptions are described
3027 If you do not use a @code{SECTIONS} command in your linker script, the
3028 linker will place each input section into an identically named output
3029 section in the order that the sections are first encountered in the
3030 input files. If all input sections are present in the first file, for
3031 example, the order of sections in the output file will match the order
3032 in the first input file. The first section will be at address zero.
3035 * Output Section Description:: Output section description
3036 * Output Section Name:: Output section name
3037 * Output Section Address:: Output section address
3038 * Input Section:: Input section description
3039 * Output Section Data:: Output section data
3040 * Output Section Keywords:: Output section keywords
3041 * Output Section Discarding:: Output section discarding
3042 * Output Section Attributes:: Output section attributes
3043 * Overlay Description:: Overlay description
3046 @node Output Section Description
3047 @subsection Output Section Description
3048 The full description of an output section looks like this:
3051 @var{section} [@var{address}] [(@var{type})] :
3052 [AT(@var{lma})] [SUBALIGN(@var{subsection_align})]
3054 @var{output-section-command}
3055 @var{output-section-command}
3057 @} [>@var{region}] [AT>@var{lma_region}] [:@var{phdr} :@var{phdr} @dots{}] [=@var{fillexp}]
3061 Most output sections do not use most of the optional section attributes.
3063 The whitespace around @var{section} is required, so that the section
3064 name is unambiguous. The colon and the curly braces are also required.
3065 The line breaks and other white space are optional.
3067 Each @var{output-section-command} may be one of the following:
3071 a symbol assignment (@pxref{Assignments})
3073 an input section description (@pxref{Input Section})
3075 data values to include directly (@pxref{Output Section Data})
3077 a special output section keyword (@pxref{Output Section Keywords})
3080 @node Output Section Name
3081 @subsection Output Section Name
3082 @cindex name, section
3083 @cindex section name
3084 The name of the output section is @var{section}. @var{section} must
3085 meet the constraints of your output format. In formats which only
3086 support a limited number of sections, such as @code{a.out}, the name
3087 must be one of the names supported by the format (@code{a.out}, for
3088 example, allows only @samp{.text}, @samp{.data} or @samp{.bss}). If the
3089 output format supports any number of sections, but with numbers and not
3090 names (as is the case for Oasys), the name should be supplied as a
3091 quoted numeric string. A section name may consist of any sequence of
3092 characters, but a name which contains any unusual characters such as
3093 commas must be quoted.
3095 The output section name @samp{/DISCARD/} is special; @ref{Output Section
3098 @node Output Section Address
3099 @subsection Output Section Address
3100 @cindex address, section
3101 @cindex section address
3102 The @var{address} is an expression for the VMA (the virtual memory
3103 address) of the output section. If you do not provide @var{address},
3104 the linker will set it based on @var{region} if present, or otherwise
3105 based on the current value of the location counter.
3107 If you provide @var{address}, the address of the output section will be
3108 set to precisely that. If you provide neither @var{address} nor
3109 @var{region}, then the address of the output section will be set to the
3110 current value of the location counter aligned to the alignment
3111 requirements of the output section. The alignment requirement of the
3112 output section is the strictest alignment of any input section contained
3113 within the output section.
3117 .text . : @{ *(.text) @}
3122 .text : @{ *(.text) @}
3125 are subtly different. The first will set the address of the
3126 @samp{.text} output section to the current value of the location
3127 counter. The second will set it to the current value of the location
3128 counter aligned to the strictest alignment of a @samp{.text} input
3131 The @var{address} may be an arbitrary expression; @ref{Expressions}.
3132 For example, if you want to align the section on a 0x10 byte boundary,
3133 so that the lowest four bits of the section address are zero, you could
3134 do something like this:
3136 .text ALIGN(0x10) : @{ *(.text) @}
3139 This works because @code{ALIGN} returns the current location counter
3140 aligned upward to the specified value.
3142 Specifying @var{address} for a section will change the value of the
3146 @subsection Input Section Description
3147 @cindex input sections
3148 @cindex mapping input sections to output sections
3149 The most common output section command is an input section description.
3151 The input section description is the most basic linker script operation.
3152 You use output sections to tell the linker how to lay out your program
3153 in memory. You use input section descriptions to tell the linker how to
3154 map the input files into your memory layout.
3157 * Input Section Basics:: Input section basics
3158 * Input Section Wildcards:: Input section wildcard patterns
3159 * Input Section Common:: Input section for common symbols
3160 * Input Section Keep:: Input section and garbage collection
3161 * Input Section Example:: Input section example
3164 @node Input Section Basics
3165 @subsubsection Input Section Basics
3166 @cindex input section basics
3167 An input section description consists of a file name optionally followed
3168 by a list of section names in parentheses.
3170 The file name and the section name may be wildcard patterns, which we
3171 describe further below (@pxref{Input Section Wildcards}).
3173 The most common input section description is to include all input
3174 sections with a particular name in the output section. For example, to
3175 include all input @samp{.text} sections, you would write:
3180 Here the @samp{*} is a wildcard which matches any file name. To exclude a list
3181 of files from matching the file name wildcard, EXCLUDE_FILE may be used to
3182 match all files except the ones specified in the EXCLUDE_FILE list. For
3185 (*(EXCLUDE_FILE (*crtend.o *otherfile.o) .ctors))
3187 will cause all .ctors sections from all files except @file{crtend.o} and
3188 @file{otherfile.o} to be included.
3190 There are two ways to include more than one section:
3196 The difference between these is the order in which the @samp{.text} and
3197 @samp{.rdata} input sections will appear in the output section. In the
3198 first example, they will be intermingled, appearing in the same order as
3199 they are found in the linker input. In the second example, all
3200 @samp{.text} input sections will appear first, followed by all
3201 @samp{.rdata} input sections.
3203 You can specify a file name to include sections from a particular file.
3204 You would do this if one or more of your files contain special data that
3205 needs to be at a particular location in memory. For example:
3210 If you use a file name without a list of sections, then all sections in
3211 the input file will be included in the output section. This is not
3212 commonly done, but it may by useful on occasion. For example:
3217 When you use a file name which does not contain any wild card
3218 characters, the linker will first see if you also specified the file
3219 name on the linker command line or in an @code{INPUT} command. If you
3220 did not, the linker will attempt to open the file as an input file, as
3221 though it appeared on the command line. Note that this differs from an
3222 @code{INPUT} command, because the linker will not search for the file in
3223 the archive search path.
3225 @node Input Section Wildcards
3226 @subsubsection Input Section Wildcard Patterns
3227 @cindex input section wildcards
3228 @cindex wildcard file name patterns
3229 @cindex file name wildcard patterns
3230 @cindex section name wildcard patterns
3231 In an input section description, either the file name or the section
3232 name or both may be wildcard patterns.
3234 The file name of @samp{*} seen in many examples is a simple wildcard
3235 pattern for the file name.
3237 The wildcard patterns are like those used by the Unix shell.
3241 matches any number of characters
3243 matches any single character
3245 matches a single instance of any of the @var{chars}; the @samp{-}
3246 character may be used to specify a range of characters, as in
3247 @samp{[a-z]} to match any lower case letter
3249 quotes the following character
3252 When a file name is matched with a wildcard, the wildcard characters
3253 will not match a @samp{/} character (used to separate directory names on
3254 Unix). A pattern consisting of a single @samp{*} character is an
3255 exception; it will always match any file name, whether it contains a
3256 @samp{/} or not. In a section name, the wildcard characters will match
3257 a @samp{/} character.
3259 File name wildcard patterns only match files which are explicitly
3260 specified on the command line or in an @code{INPUT} command. The linker
3261 does not search directories to expand wildcards.
3263 If a file name matches more than one wildcard pattern, or if a file name
3264 appears explicitly and is also matched by a wildcard pattern, the linker
3265 will use the first match in the linker script. For example, this
3266 sequence of input section descriptions is probably in error, because the
3267 @file{data.o} rule will not be used:
3269 .data : @{ *(.data) @}
3270 .data1 : @{ data.o(.data) @}
3273 @cindex SORT_BY_NAME
3274 Normally, the linker will place files and sections matched by wildcards
3275 in the order in which they are seen during the link. You can change
3276 this by using the @code{SORT_BY_NAME} keyword, which appears before a wildcard
3277 pattern in parentheses (e.g., @code{SORT_BY_NAME(.text*)}). When the
3278 @code{SORT_BY_NAME} keyword is used, the linker will sort the files or sections
3279 into ascending order by name before placing them in the output file.
3281 @cindex SORT_BY_ALIGNMENT
3282 @code{SORT_BY_ALIGNMENT} is very similar to @code{SORT_BY_NAME}. The
3283 difference is @code{SORT_BY_ALIGNMENT} will sort sections into
3284 ascending order by alignment before placing them in the output file.
3287 @code{SORT} is an alias for @code{SORT_BY_NAME}.
3289 When there are nested section sorting commands in linker script, there
3290 can be at most 1 level of nesting for section sorting commands.
3294 @code{SORT_BY_NAME} (@code{SORT_BY_ALIGNMENT} (wildcard section pattern)).
3295 It will sort the input sections by name first, then by alignment if 2
3296 sections have the same name.
3298 @code{SORT_BY_ALIGNMENT} (@code{SORT_BY_NAME} (wildcard section pattern)).
3299 It will sort the input sections by alignment first, then by name if 2
3300 sections have the same alignment.
3302 @code{SORT_BY_NAME} (@code{SORT_BY_NAME} (wildcard section pattern)) is
3303 treated the same as @code{SORT_BY_NAME} (wildcard section pattern).
3305 @code{SORT_BY_ALIGNMENT} (@code{SORT_BY_ALIGNMENT} (wildcard section pattern))
3306 is treated the same as @code{SORT_BY_ALIGNMENT} (wildcard section pattern).
3308 All other nested section sorting commands are invalid.
3311 When both command line section sorting option and linker script
3312 section sorting command are used, section sorting command always
3313 takes precedence over the command line option.
3315 If the section sorting command in linker script isn't nested, the
3316 command line option will make the section sorting command to be
3317 treated as nested sorting command.
3321 @code{SORT_BY_NAME} (wildcard section pattern ) with
3322 @option{--sort-sections alignment} is equivalent to
3323 @code{SORT_BY_NAME} (@code{SORT_BY_ALIGNMENT} (wildcard section pattern)).
3325 @code{SORT_BY_ALIGNMENT} (wildcard section pattern) with
3326 @option{--sort-section name} is equivalent to
3327 @code{SORT_BY_ALIGNMENT} (@code{SORT_BY_NAME} (wildcard section pattern)).
3330 If the section sorting command in linker script is nested, the
3331 command line option will be ignored.
3333 If you ever get confused about where input sections are going, use the
3334 @samp{-M} linker option to generate a map file. The map file shows
3335 precisely how input sections are mapped to output sections.
3337 This example shows how wildcard patterns might be used to partition
3338 files. This linker script directs the linker to place all @samp{.text}
3339 sections in @samp{.text} and all @samp{.bss} sections in @samp{.bss}.
3340 The linker will place the @samp{.data} section from all files beginning
3341 with an upper case character in @samp{.DATA}; for all other files, the
3342 linker will place the @samp{.data} section in @samp{.data}.
3346 .text : @{ *(.text) @}
3347 .DATA : @{ [A-Z]*(.data) @}
3348 .data : @{ *(.data) @}
3349 .bss : @{ *(.bss) @}
3354 @node Input Section Common
3355 @subsubsection Input Section for Common Symbols
3356 @cindex common symbol placement
3357 @cindex uninitialized data placement
3358 A special notation is needed for common symbols, because in many object
3359 file formats common symbols do not have a particular input section. The
3360 linker treats common symbols as though they are in an input section
3361 named @samp{COMMON}.
3363 You may use file names with the @samp{COMMON} section just as with any
3364 other input sections. You can use this to place common symbols from a
3365 particular input file in one section while common symbols from other
3366 input files are placed in another section.
3368 In most cases, common symbols in input files will be placed in the
3369 @samp{.bss} section in the output file. For example:
3371 .bss @{ *(.bss) *(COMMON) @}
3374 @cindex scommon section
3375 @cindex small common symbols
3376 Some object file formats have more than one type of common symbol. For
3377 example, the MIPS ELF object file format distinguishes standard common
3378 symbols and small common symbols. In this case, the linker will use a
3379 different special section name for other types of common symbols. In
3380 the case of MIPS ELF, the linker uses @samp{COMMON} for standard common
3381 symbols and @samp{.scommon} for small common symbols. This permits you
3382 to map the different types of common symbols into memory at different
3386 You will sometimes see @samp{[COMMON]} in old linker scripts. This
3387 notation is now considered obsolete. It is equivalent to
3390 @node Input Section Keep
3391 @subsubsection Input Section and Garbage Collection
3393 @cindex garbage collection
3394 When link-time garbage collection is in use (@samp{--gc-sections}),
3395 it is often useful to mark sections that should not be eliminated.
3396 This is accomplished by surrounding an input section's wildcard entry
3397 with @code{KEEP()}, as in @code{KEEP(*(.init))} or
3398 @code{KEEP(SORT_BY_NAME(*)(.ctors))}.
3400 @node Input Section Example
3401 @subsubsection Input Section Example
3402 The following example is a complete linker script. It tells the linker
3403 to read all of the sections from file @file{all.o} and place them at the
3404 start of output section @samp{outputa} which starts at location
3405 @samp{0x10000}. All of section @samp{.input1} from file @file{foo.o}
3406 follows immediately, in the same output section. All of section
3407 @samp{.input2} from @file{foo.o} goes into output section
3408 @samp{outputb}, followed by section @samp{.input1} from @file{foo1.o}.
3409 All of the remaining @samp{.input1} and @samp{.input2} sections from any
3410 files are written to output section @samp{outputc}.
3438 @node Output Section Data
3439 @subsection Output Section Data
3441 @cindex section data
3442 @cindex output section data
3443 @kindex BYTE(@var{expression})
3444 @kindex SHORT(@var{expression})
3445 @kindex LONG(@var{expression})
3446 @kindex QUAD(@var{expression})
3447 @kindex SQUAD(@var{expression})
3448 You can include explicit bytes of data in an output section by using
3449 @code{BYTE}, @code{SHORT}, @code{LONG}, @code{QUAD}, or @code{SQUAD} as
3450 an output section command. Each keyword is followed by an expression in
3451 parentheses providing the value to store (@pxref{Expressions}). The
3452 value of the expression is stored at the current value of the location
3455 The @code{BYTE}, @code{SHORT}, @code{LONG}, and @code{QUAD} commands
3456 store one, two, four, and eight bytes (respectively). After storing the
3457 bytes, the location counter is incremented by the number of bytes
3460 For example, this will store the byte 1 followed by the four byte value
3461 of the symbol @samp{addr}:
3467 When using a 64 bit host or target, @code{QUAD} and @code{SQUAD} are the
3468 same; they both store an 8 byte, or 64 bit, value. When both host and
3469 target are 32 bits, an expression is computed as 32 bits. In this case
3470 @code{QUAD} stores a 32 bit value zero extended to 64 bits, and
3471 @code{SQUAD} stores a 32 bit value sign extended to 64 bits.
3473 If the object file format of the output file has an explicit endianness,
3474 which is the normal case, the value will be stored in that endianness.
3475 When the object file format does not have an explicit endianness, as is
3476 true of, for example, S-records, the value will be stored in the
3477 endianness of the first input object file.
3479 Note---these commands only work inside a section description and not
3480 between them, so the following will produce an error from the linker:
3482 SECTIONS @{@ .text : @{@ *(.text) @}@ LONG(1) .data : @{@ *(.data) @}@ @}@
3484 whereas this will work:
3486 SECTIONS @{@ .text : @{@ *(.text) ; LONG(1) @}@ .data : @{@ *(.data) @}@ @}@
3489 @kindex FILL(@var{expression})
3490 @cindex holes, filling
3491 @cindex unspecified memory
3492 You may use the @code{FILL} command to set the fill pattern for the
3493 current section. It is followed by an expression in parentheses. Any
3494 otherwise unspecified regions of memory within the section (for example,
3495 gaps left due to the required alignment of input sections) are filled
3496 with the value of the expression, repeated as
3497 necessary. A @code{FILL} statement covers memory locations after the
3498 point at which it occurs in the section definition; by including more
3499 than one @code{FILL} statement, you can have different fill patterns in
3500 different parts of an output section.
3502 This example shows how to fill unspecified regions of memory with the
3508 The @code{FILL} command is similar to the @samp{=@var{fillexp}} output
3509 section attribute, but it only affects the
3510 part of the section following the @code{FILL} command, rather than the
3511 entire section. If both are used, the @code{FILL} command takes
3512 precedence. @xref{Output Section Fill}, for details on the fill
3515 @node Output Section Keywords
3516 @subsection Output Section Keywords
3517 There are a couple of keywords which can appear as output section
3521 @kindex CREATE_OBJECT_SYMBOLS
3522 @cindex input filename symbols
3523 @cindex filename symbols
3524 @item CREATE_OBJECT_SYMBOLS
3525 The command tells the linker to create a symbol for each input file.
3526 The name of each symbol will be the name of the corresponding input
3527 file. The section of each symbol will be the output section in which
3528 the @code{CREATE_OBJECT_SYMBOLS} command appears.
3530 This is conventional for the a.out object file format. It is not
3531 normally used for any other object file format.
3533 @kindex CONSTRUCTORS
3534 @cindex C++ constructors, arranging in link
3535 @cindex constructors, arranging in link
3537 When linking using the a.out object file format, the linker uses an
3538 unusual set construct to support C++ global constructors and
3539 destructors. When linking object file formats which do not support
3540 arbitrary sections, such as ECOFF and XCOFF, the linker will
3541 automatically recognize C++ global constructors and destructors by name.
3542 For these object file formats, the @code{CONSTRUCTORS} command tells the
3543 linker to place constructor information in the output section where the
3544 @code{CONSTRUCTORS} command appears. The @code{CONSTRUCTORS} command is
3545 ignored for other object file formats.
3547 The symbol @w{@code{__CTOR_LIST__}} marks the start of the global
3548 constructors, and the symbol @w{@code{__CTOR_END__}} marks the end.
3549 Similarly, @w{@code{__DTOR_LIST__}} and @w{@code{__DTOR_END__}} mark
3550 the start and end of the global destructors. The
3551 first word in the list is the number of entries, followed by the address
3552 of each constructor or destructor, followed by a zero word. The
3553 compiler must arrange to actually run the code. For these object file
3554 formats @sc{gnu} C++ normally calls constructors from a subroutine
3555 @code{__main}; a call to @code{__main} is automatically inserted into
3556 the startup code for @code{main}. @sc{gnu} C++ normally runs
3557 destructors either by using @code{atexit}, or directly from the function
3560 For object file formats such as @code{COFF} or @code{ELF} which support
3561 arbitrary section names, @sc{gnu} C++ will normally arrange to put the
3562 addresses of global constructors and destructors into the @code{.ctors}
3563 and @code{.dtors} sections. Placing the following sequence into your
3564 linker script will build the sort of table which the @sc{gnu} C++
3565 runtime code expects to see.
3569 LONG((__CTOR_END__ - __CTOR_LIST__) / 4 - 2)
3574 LONG((__DTOR_END__ - __DTOR_LIST__) / 4 - 2)
3580 If you are using the @sc{gnu} C++ support for initialization priority,
3581 which provides some control over the order in which global constructors
3582 are run, you must sort the constructors at link time to ensure that they
3583 are executed in the correct order. When using the @code{CONSTRUCTORS}
3584 command, use @samp{SORT_BY_NAME(CONSTRUCTORS)} instead. When using the
3585 @code{.ctors} and @code{.dtors} sections, use @samp{*(SORT_BY_NAME(.ctors))} and
3586 @samp{*(SORT_BY_NAME(.dtors))} instead of just @samp{*(.ctors)} and
3589 Normally the compiler and linker will handle these issues automatically,
3590 and you will not need to concern yourself with them. However, you may
3591 need to consider this if you are using C++ and writing your own linker
3596 @node Output Section Discarding
3597 @subsection Output Section Discarding
3598 @cindex discarding sections
3599 @cindex sections, discarding
3600 @cindex removing sections
3601 The linker will not create output section which do not have any
3602 contents. This is for convenience when referring to input sections that
3603 may or may not be present in any of the input files. For example:
3608 will only create a @samp{.foo} section in the output file if there is a
3609 @samp{.foo} section in at least one input file.
3611 If you use anything other than an input section description as an output
3612 section command, such as a symbol assignment, then the output section
3613 will always be created, even if there are no matching input sections.
3616 The special output section name @samp{/DISCARD/} may be used to discard
3617 input sections. Any input sections which are assigned to an output
3618 section named @samp{/DISCARD/} are not included in the output file.
3620 @node Output Section Attributes
3621 @subsection Output Section Attributes
3622 @cindex output section attributes
3623 We showed above that the full description of an output section looked
3627 @var{section} [@var{address}] [(@var{type})] :
3628 [AT(@var{lma})] [SUBALIGN(@var{subsection_align})]
3630 @var{output-section-command}
3631 @var{output-section-command}
3633 @} [>@var{region}] [AT>@var{lma_region}] [:@var{phdr} :@var{phdr} @dots{}] [=@var{fillexp}]
3636 We've already described @var{section}, @var{address}, and
3637 @var{output-section-command}. In this section we will describe the
3638 remaining section attributes.
3641 * Output Section Type:: Output section type
3642 * Output Section LMA:: Output section LMA
3643 * Forced Input Alignment:: Forced Input Alignment
3644 * Output Section Region:: Output section region
3645 * Output Section Phdr:: Output section phdr
3646 * Output Section Fill:: Output section fill
3649 @node Output Section Type
3650 @subsubsection Output Section Type
3651 Each output section may have a type. The type is a keyword in
3652 parentheses. The following types are defined:
3656 The section should be marked as not loadable, so that it will not be
3657 loaded into memory when the program is run.
3662 These type names are supported for backward compatibility, and are
3663 rarely used. They all have the same effect: the section should be
3664 marked as not allocatable, so that no memory is allocated for the
3665 section when the program is run.
3669 @cindex prevent unnecessary loading
3670 @cindex loading, preventing
3671 The linker normally sets the attributes of an output section based on
3672 the input sections which map into it. You can override this by using
3673 the section type. For example, in the script sample below, the
3674 @samp{ROM} section is addressed at memory location @samp{0} and does not
3675 need to be loaded when the program is run. The contents of the
3676 @samp{ROM} section will appear in the linker output file as usual.
3680 ROM 0 (NOLOAD) : @{ @dots{} @}
3686 @node Output Section LMA
3687 @subsubsection Output Section LMA
3688 @kindex AT>@var{lma_region}
3689 @kindex AT(@var{lma})
3690 @cindex load address
3691 @cindex section load address
3692 Every section has a virtual address (VMA) and a load address (LMA); see
3693 @ref{Basic Script Concepts}. The address expression which may appear in
3694 an output section description sets the VMA (@pxref{Output Section
3697 The linker will normally set the LMA equal to the VMA. You can change
3698 that by using the @code{AT} keyword. The expression @var{lma} that
3699 follows the @code{AT} keyword specifies the load address of the
3702 Alternatively, with @samp{AT>@var{lma_region}} expression, you may
3703 specify a memory region for the section's load address. @xref{MEMORY}.
3704 Note that if the section has not had a VMA assigned to it then the
3705 linker will use the @var{lma_region} as the VMA region as well.
3706 @xref{Output Section Region}.
3708 @cindex ROM initialized data
3709 @cindex initialized data in ROM
3710 This feature is designed to make it easy to build a ROM image. For
3711 example, the following linker script creates three output sections: one
3712 called @samp{.text}, which starts at @code{0x1000}, one called
3713 @samp{.mdata}, which is loaded at the end of the @samp{.text} section
3714 even though its VMA is @code{0x2000}, and one called @samp{.bss} to hold
3715 uninitialized data at address @code{0x3000}. The symbol @code{_data} is
3716 defined with the value @code{0x2000}, which shows that the location
3717 counter holds the VMA value, not the LMA value.
3723 .text 0x1000 : @{ *(.text) _etext = . ; @}
3725 AT ( ADDR (.text) + SIZEOF (.text) )
3726 @{ _data = . ; *(.data); _edata = . ; @}
3728 @{ _bstart = . ; *(.bss) *(COMMON) ; _bend = . ;@}
3733 The run-time initialization code for use with a program generated with
3734 this linker script would include something like the following, to copy
3735 the initialized data from the ROM image to its runtime address. Notice
3736 how this code takes advantage of the symbols defined by the linker
3741 extern char _etext, _data, _edata, _bstart, _bend;
3742 char *src = &_etext;
3745 /* ROM has data at end of text; copy it. */
3746 while (dst < &_edata) @{
3751 for (dst = &_bstart; dst< &_bend; dst++)
3756 @node Forced Input Alignment
3757 @subsubsection Forced Input Alignment
3758 @kindex SUBALIGN(@var{subsection_align})
3759 @cindex forcing input section alignment
3760 @cindex input section alignment
3761 You can force input section alignment within an output section by using
3762 SUBALIGN. The value specified overrides any alignment given by input
3763 sections, whether larger or smaller.
3765 @node Output Section Region
3766 @subsubsection Output Section Region
3767 @kindex >@var{region}
3768 @cindex section, assigning to memory region
3769 @cindex memory regions and sections
3770 You can assign a section to a previously defined region of memory by
3771 using @samp{>@var{region}}. @xref{MEMORY}.
3773 Here is a simple example:
3776 MEMORY @{ rom : ORIGIN = 0x1000, LENGTH = 0x1000 @}
3777 SECTIONS @{ ROM : @{ *(.text) @} >rom @}
3781 @node Output Section Phdr
3782 @subsubsection Output Section Phdr
3784 @cindex section, assigning to program header
3785 @cindex program headers and sections
3786 You can assign a section to a previously defined program segment by
3787 using @samp{:@var{phdr}}. @xref{PHDRS}. If a section is assigned to
3788 one or more segments, then all subsequent allocated sections will be
3789 assigned to those segments as well, unless they use an explicitly
3790 @code{:@var{phdr}} modifier. You can use @code{:NONE} to tell the
3791 linker to not put the section in any segment at all.
3793 Here is a simple example:
3796 PHDRS @{ text PT_LOAD ; @}
3797 SECTIONS @{ .text : @{ *(.text) @} :text @}
3801 @node Output Section Fill
3802 @subsubsection Output Section Fill
3803 @kindex =@var{fillexp}
3804 @cindex section fill pattern
3805 @cindex fill pattern, entire section
3806 You can set the fill pattern for an entire section by using
3807 @samp{=@var{fillexp}}. @var{fillexp} is an expression
3808 (@pxref{Expressions}). Any otherwise unspecified regions of memory
3809 within the output section (for example, gaps left due to the required
3810 alignment of input sections) will be filled with the value, repeated as
3811 necessary. If the fill expression is a simple hex number, ie. a string
3812 of hex digit starting with @samp{0x} and without a trailing @samp{k} or @samp{M}, then
3813 an arbitrarily long sequence of hex digits can be used to specify the
3814 fill pattern; Leading zeros become part of the pattern too. For all
3815 other cases, including extra parentheses or a unary @code{+}, the fill
3816 pattern is the four least significant bytes of the value of the
3817 expression. In all cases, the number is big-endian.
3819 You can also change the fill value with a @code{FILL} command in the
3820 output section commands; (@pxref{Output Section Data}).
3822 Here is a simple example:
3825 SECTIONS @{ .text : @{ *(.text) @} =0x90909090 @}
3829 @node Overlay Description
3830 @subsection Overlay Description
3833 An overlay description provides an easy way to describe sections which
3834 are to be loaded as part of a single memory image but are to be run at
3835 the same memory address. At run time, some sort of overlay manager will
3836 copy the overlaid sections in and out of the runtime memory address as
3837 required, perhaps by simply manipulating addressing bits. This approach
3838 can be useful, for example, when a certain region of memory is faster
3841 Overlays are described using the @code{OVERLAY} command. The
3842 @code{OVERLAY} command is used within a @code{SECTIONS} command, like an
3843 output section description. The full syntax of the @code{OVERLAY}
3844 command is as follows:
3847 OVERLAY [@var{start}] : [NOCROSSREFS] [AT ( @var{ldaddr} )]
3851 @var{output-section-command}
3852 @var{output-section-command}
3854 @} [:@var{phdr}@dots{}] [=@var{fill}]
3857 @var{output-section-command}
3858 @var{output-section-command}
3860 @} [:@var{phdr}@dots{}] [=@var{fill}]
3862 @} [>@var{region}] [:@var{phdr}@dots{}] [=@var{fill}]
3866 Everything is optional except @code{OVERLAY} (a keyword), and each
3867 section must have a name (@var{secname1} and @var{secname2} above). The
3868 section definitions within the @code{OVERLAY} construct are identical to
3869 those within the general @code{SECTIONS} contruct (@pxref{SECTIONS}),
3870 except that no addresses and no memory regions may be defined for
3871 sections within an @code{OVERLAY}.
3873 The sections are all defined with the same starting address. The load
3874 addresses of the sections are arranged such that they are consecutive in
3875 memory starting at the load address used for the @code{OVERLAY} as a
3876 whole (as with normal section definitions, the load address is optional,
3877 and defaults to the start address; the start address is also optional,
3878 and defaults to the current value of the location counter).
3880 If the @code{NOCROSSREFS} keyword is used, and there any references
3881 among the sections, the linker will report an error. Since the sections
3882 all run at the same address, it normally does not make sense for one
3883 section to refer directly to another. @xref{Miscellaneous Commands,
3886 For each section within the @code{OVERLAY}, the linker automatically
3887 defines two symbols. The symbol @code{__load_start_@var{secname}} is
3888 defined as the starting load address of the section. The symbol
3889 @code{__load_stop_@var{secname}} is defined as the final load address of
3890 the section. Any characters within @var{secname} which are not legal
3891 within C identifiers are removed. C (or assembler) code may use these
3892 symbols to move the overlaid sections around as necessary.
3894 At the end of the overlay, the value of the location counter is set to
3895 the start address of the overlay plus the size of the largest section.
3897 Here is an example. Remember that this would appear inside a
3898 @code{SECTIONS} construct.
3901 OVERLAY 0x1000 : AT (0x4000)
3903 .text0 @{ o1/*.o(.text) @}
3904 .text1 @{ o2/*.o(.text) @}
3909 This will define both @samp{.text0} and @samp{.text1} to start at
3910 address 0x1000. @samp{.text0} will be loaded at address 0x4000, and
3911 @samp{.text1} will be loaded immediately after @samp{.text0}. The
3912 following symbols will be defined: @code{__load_start_text0},
3913 @code{__load_stop_text0}, @code{__load_start_text1},
3914 @code{__load_stop_text1}.
3916 C code to copy overlay @code{.text1} into the overlay area might look
3921 extern char __load_start_text1, __load_stop_text1;
3922 memcpy ((char *) 0x1000, &__load_start_text1,
3923 &__load_stop_text1 - &__load_start_text1);
3927 Note that the @code{OVERLAY} command is just syntactic sugar, since
3928 everything it does can be done using the more basic commands. The above
3929 example could have been written identically as follows.
3933 .text0 0x1000 : AT (0x4000) @{ o1/*.o(.text) @}
3934 __load_start_text0 = LOADADDR (.text0);
3935 __load_stop_text0 = LOADADDR (.text0) + SIZEOF (.text0);
3936 .text1 0x1000 : AT (0x4000 + SIZEOF (.text0)) @{ o2/*.o(.text) @}
3937 __load_start_text1 = LOADADDR (.text1);
3938 __load_stop_text1 = LOADADDR (.text1) + SIZEOF (.text1);
3939 . = 0x1000 + MAX (SIZEOF (.text0), SIZEOF (.text1));
3944 @section MEMORY Command
3946 @cindex memory regions
3947 @cindex regions of memory
3948 @cindex allocating memory
3949 @cindex discontinuous memory
3950 The linker's default configuration permits allocation of all available
3951 memory. You can override this by using the @code{MEMORY} command.
3953 The @code{MEMORY} command describes the location and size of blocks of
3954 memory in the target. You can use it to describe which memory regions
3955 may be used by the linker, and which memory regions it must avoid. You
3956 can then assign sections to particular memory regions. The linker will
3957 set section addresses based on the memory regions, and will warn about
3958 regions that become too full. The linker will not shuffle sections
3959 around to fit into the available regions.
3961 A linker script may contain at most one use of the @code{MEMORY}
3962 command. However, you can define as many blocks of memory within it as
3963 you wish. The syntax is:
3968 @var{name} [(@var{attr})] : ORIGIN = @var{origin}, LENGTH = @var{len}
3974 The @var{name} is a name used in the linker script to refer to the
3975 region. The region name has no meaning outside of the linker script.
3976 Region names are stored in a separate name space, and will not conflict
3977 with symbol names, file names, or section names. Each memory region
3978 must have a distinct name.
3980 @cindex memory region attributes
3981 The @var{attr} string is an optional list of attributes that specify
3982 whether to use a particular memory region for an input section which is
3983 not explicitly mapped in the linker script. As described in
3984 @ref{SECTIONS}, if you do not specify an output section for some input
3985 section, the linker will create an output section with the same name as
3986 the input section. If you define region attributes, the linker will use
3987 them to select the memory region for the output section that it creates.
3989 The @var{attr} string must consist only of the following characters:
4004 Invert the sense of any of the preceding attributes
4007 If a unmapped section matches any of the listed attributes other than
4008 @samp{!}, it will be placed in the memory region. The @samp{!}
4009 attribute reverses this test, so that an unmapped section will be placed
4010 in the memory region only if it does not match any of the listed
4016 The @var{origin} is an numerical expression for the start address of
4017 the memory region. The expression must evaluate to a constant and it
4018 cannot involve any symbols. The keyword @code{ORIGIN} may be
4019 abbreviated to @code{org} or @code{o} (but not, for example,
4025 The @var{len} is an expression for the size in bytes of the memory
4026 region. As with the @var{origin} expression, the expression must
4027 be numerical only and must evaluate to a constant. The keyword
4028 @code{LENGTH} may be abbreviated to @code{len} or @code{l}.
4030 In the following example, we specify that there are two memory regions
4031 available for allocation: one starting at @samp{0} for 256 kilobytes,
4032 and the other starting at @samp{0x40000000} for four megabytes. The
4033 linker will place into the @samp{rom} memory region every section which
4034 is not explicitly mapped into a memory region, and is either read-only
4035 or executable. The linker will place other sections which are not
4036 explicitly mapped into a memory region into the @samp{ram} memory
4043 rom (rx) : ORIGIN = 0, LENGTH = 256K
4044 ram (!rx) : org = 0x40000000, l = 4M
4049 Once you define a memory region, you can direct the linker to place
4050 specific output sections into that memory region by using the
4051 @samp{>@var{region}} output section attribute. For example, if you have
4052 a memory region named @samp{mem}, you would use @samp{>mem} in the
4053 output section definition. @xref{Output Section Region}. If no address
4054 was specified for the output section, the linker will set the address to
4055 the next available address within the memory region. If the combined
4056 output sections directed to a memory region are too large for the
4057 region, the linker will issue an error message.
4059 It is possible to access the origin and length of a memory in an
4060 expression via the @code{ORIGIN(@var{memory})} and
4061 @code{LENGTH(@var{memory})} functions:
4065 _fstack = ORIGIN(ram) + LENGTH(ram) - 4;
4070 @section PHDRS Command
4072 @cindex program headers
4073 @cindex ELF program headers
4074 @cindex program segments
4075 @cindex segments, ELF
4076 The ELF object file format uses @dfn{program headers}, also knows as
4077 @dfn{segments}. The program headers describe how the program should be
4078 loaded into memory. You can print them out by using the @code{objdump}
4079 program with the @samp{-p} option.
4081 When you run an ELF program on a native ELF system, the system loader
4082 reads the program headers in order to figure out how to load the
4083 program. This will only work if the program headers are set correctly.
4084 This manual does not describe the details of how the system loader
4085 interprets program headers; for more information, see the ELF ABI.
4087 The linker will create reasonable program headers by default. However,
4088 in some cases, you may need to specify the program headers more
4089 precisely. You may use the @code{PHDRS} command for this purpose. When
4090 the linker sees the @code{PHDRS} command in the linker script, it will
4091 not create any program headers other than the ones specified.
4093 The linker only pays attention to the @code{PHDRS} command when
4094 generating an ELF output file. In other cases, the linker will simply
4095 ignore @code{PHDRS}.
4097 This is the syntax of the @code{PHDRS} command. The words @code{PHDRS},
4098 @code{FILEHDR}, @code{AT}, and @code{FLAGS} are keywords.
4104 @var{name} @var{type} [ FILEHDR ] [ PHDRS ] [ AT ( @var{address} ) ]
4105 [ FLAGS ( @var{flags} ) ] ;
4110 The @var{name} is used only for reference in the @code{SECTIONS} command
4111 of the linker script. It is not put into the output file. Program
4112 header names are stored in a separate name space, and will not conflict
4113 with symbol names, file names, or section names. Each program header
4114 must have a distinct name.
4116 Certain program header types describe segments of memory which the
4117 system loader will load from the file. In the linker script, you
4118 specify the contents of these segments by placing allocatable output
4119 sections in the segments. You use the @samp{:@var{phdr}} output section
4120 attribute to place a section in a particular segment. @xref{Output
4123 It is normal to put certain sections in more than one segment. This
4124 merely implies that one segment of memory contains another. You may
4125 repeat @samp{:@var{phdr}}, using it once for each segment which should
4126 contain the section.
4128 If you place a section in one or more segments using @samp{:@var{phdr}},
4129 then the linker will place all subsequent allocatable sections which do
4130 not specify @samp{:@var{phdr}} in the same segments. This is for
4131 convenience, since generally a whole set of contiguous sections will be
4132 placed in a single segment. You can use @code{:NONE} to override the
4133 default segment and tell the linker to not put the section in any
4138 You may use the @code{FILEHDR} and @code{PHDRS} keywords appear after
4139 the program header type to further describe the contents of the segment.
4140 The @code{FILEHDR} keyword means that the segment should include the ELF
4141 file header. The @code{PHDRS} keyword means that the segment should
4142 include the ELF program headers themselves.
4144 The @var{type} may be one of the following. The numbers indicate the
4145 value of the keyword.
4148 @item @code{PT_NULL} (0)
4149 Indicates an unused program header.
4151 @item @code{PT_LOAD} (1)
4152 Indicates that this program header describes a segment to be loaded from
4155 @item @code{PT_DYNAMIC} (2)
4156 Indicates a segment where dynamic linking information can be found.
4158 @item @code{PT_INTERP} (3)
4159 Indicates a segment where the name of the program interpreter may be
4162 @item @code{PT_NOTE} (4)
4163 Indicates a segment holding note information.
4165 @item @code{PT_SHLIB} (5)
4166 A reserved program header type, defined but not specified by the ELF
4169 @item @code{PT_PHDR} (6)
4170 Indicates a segment where the program headers may be found.
4172 @item @var{expression}
4173 An expression giving the numeric type of the program header. This may
4174 be used for types not defined above.
4177 You can specify that a segment should be loaded at a particular address
4178 in memory by using an @code{AT} expression. This is identical to the
4179 @code{AT} command used as an output section attribute (@pxref{Output
4180 Section LMA}). The @code{AT} command for a program header overrides the
4181 output section attribute.
4183 The linker will normally set the segment flags based on the sections
4184 which comprise the segment. You may use the @code{FLAGS} keyword to
4185 explicitly specify the segment flags. The value of @var{flags} must be
4186 an integer. It is used to set the @code{p_flags} field of the program
4189 Here is an example of @code{PHDRS}. This shows a typical set of program
4190 headers used on a native ELF system.
4196 headers PT_PHDR PHDRS ;
4198 text PT_LOAD FILEHDR PHDRS ;
4200 dynamic PT_DYNAMIC ;
4206 .interp : @{ *(.interp) @} :text :interp
4207 .text : @{ *(.text) @} :text
4208 .rodata : @{ *(.rodata) @} /* defaults to :text */
4210 . = . + 0x1000; /* move to a new page in memory */
4211 .data : @{ *(.data) @} :data
4212 .dynamic : @{ *(.dynamic) @} :data :dynamic
4219 @section VERSION Command
4220 @kindex VERSION @{script text@}
4221 @cindex symbol versions
4222 @cindex version script
4223 @cindex versions of symbols
4224 The linker supports symbol versions when using ELF. Symbol versions are
4225 only useful when using shared libraries. The dynamic linker can use
4226 symbol versions to select a specific version of a function when it runs
4227 a program that may have been linked against an earlier version of the
4230 You can include a version script directly in the main linker script, or
4231 you can supply the version script as an implicit linker script. You can
4232 also use the @samp{--version-script} linker option.
4234 The syntax of the @code{VERSION} command is simply
4236 VERSION @{ version-script-commands @}
4239 The format of the version script commands is identical to that used by
4240 Sun's linker in Solaris 2.5. The version script defines a tree of
4241 version nodes. You specify the node names and interdependencies in the
4242 version script. You can specify which symbols are bound to which
4243 version nodes, and you can reduce a specified set of symbols to local
4244 scope so that they are not globally visible outside of the shared
4247 The easiest way to demonstrate the version script language is with a few
4269 This example version script defines three version nodes. The first
4270 version node defined is @samp{VERS_1.1}; it has no other dependencies.
4271 The script binds the symbol @samp{foo1} to @samp{VERS_1.1}. It reduces
4272 a number of symbols to local scope so that they are not visible outside
4273 of the shared library; this is done using wildcard patterns, so that any
4274 symbol whose name begins with @samp{old}, @samp{original}, or @samp{new}
4275 is matched. The wildcard patterns available are the same as those used
4276 in the shell when matching filenames (also known as ``globbing'').
4278 Next, the version script defines node @samp{VERS_1.2}. This node
4279 depends upon @samp{VERS_1.1}. The script binds the symbol @samp{foo2}
4280 to the version node @samp{VERS_1.2}.
4282 Finally, the version script defines node @samp{VERS_2.0}. This node
4283 depends upon @samp{VERS_1.2}. The scripts binds the symbols @samp{bar1}
4284 and @samp{bar2} are bound to the version node @samp{VERS_2.0}.
4286 When the linker finds a symbol defined in a library which is not
4287 specifically bound to a version node, it will effectively bind it to an
4288 unspecified base version of the library. You can bind all otherwise
4289 unspecified symbols to a given version node by using @samp{global: *;}
4290 somewhere in the version script.
4292 The names of the version nodes have no specific meaning other than what
4293 they might suggest to the person reading them. The @samp{2.0} version
4294 could just as well have appeared in between @samp{1.1} and @samp{1.2}.
4295 However, this would be a confusing way to write a version script.
4297 Node name can be omited, provided it is the only version node
4298 in the version script. Such version script doesn't assign any versions to
4299 symbols, only selects which symbols will be globally visible out and which
4303 @{ global: foo; bar; local: *; @};
4306 When you link an application against a shared library that has versioned
4307 symbols, the application itself knows which version of each symbol it
4308 requires, and it also knows which version nodes it needs from each
4309 shared library it is linked against. Thus at runtime, the dynamic
4310 loader can make a quick check to make sure that the libraries you have
4311 linked against do in fact supply all of the version nodes that the
4312 application will need to resolve all of the dynamic symbols. In this
4313 way it is possible for the dynamic linker to know with certainty that
4314 all external symbols that it needs will be resolvable without having to
4315 search for each symbol reference.
4317 The symbol versioning is in effect a much more sophisticated way of
4318 doing minor version checking that SunOS does. The fundamental problem
4319 that is being addressed here is that typically references to external
4320 functions are bound on an as-needed basis, and are not all bound when
4321 the application starts up. If a shared library is out of date, a
4322 required interface may be missing; when the application tries to use
4323 that interface, it may suddenly and unexpectedly fail. With symbol
4324 versioning, the user will get a warning when they start their program if
4325 the libraries being used with the application are too old.
4327 There are several GNU extensions to Sun's versioning approach. The
4328 first of these is the ability to bind a symbol to a version node in the
4329 source file where the symbol is defined instead of in the versioning
4330 script. This was done mainly to reduce the burden on the library
4331 maintainer. You can do this by putting something like:
4333 __asm__(".symver original_foo,foo@@VERS_1.1");
4336 in the C source file. This renames the function @samp{original_foo} to
4337 be an alias for @samp{foo} bound to the version node @samp{VERS_1.1}.
4338 The @samp{local:} directive can be used to prevent the symbol
4339 @samp{original_foo} from being exported. A @samp{.symver} directive
4340 takes precedence over a version script.
4342 The second GNU extension is to allow multiple versions of the same
4343 function to appear in a given shared library. In this way you can make
4344 an incompatible change to an interface without increasing the major
4345 version number of the shared library, while still allowing applications
4346 linked against the old interface to continue to function.
4348 To do this, you must use multiple @samp{.symver} directives in the
4349 source file. Here is an example:
4352 __asm__(".symver original_foo,foo@@");
4353 __asm__(".symver old_foo,foo@@VERS_1.1");
4354 __asm__(".symver old_foo1,foo@@VERS_1.2");
4355 __asm__(".symver new_foo,foo@@@@VERS_2.0");
4358 In this example, @samp{foo@@} represents the symbol @samp{foo} bound to the
4359 unspecified base version of the symbol. The source file that contains this
4360 example would define 4 C functions: @samp{original_foo}, @samp{old_foo},
4361 @samp{old_foo1}, and @samp{new_foo}.
4363 When you have multiple definitions of a given symbol, there needs to be
4364 some way to specify a default version to which external references to
4365 this symbol will be bound. You can do this with the
4366 @samp{foo@@@@VERS_2.0} type of @samp{.symver} directive. You can only
4367 declare one version of a symbol as the default in this manner; otherwise
4368 you would effectively have multiple definitions of the same symbol.
4370 If you wish to bind a reference to a specific version of the symbol
4371 within the shared library, you can use the aliases of convenience
4372 (i.e., @samp{old_foo}), or you can use the @samp{.symver} directive to
4373 specifically bind to an external version of the function in question.
4375 You can also specify the language in the version script:
4378 VERSION extern "lang" @{ version-script-commands @}
4381 The supported @samp{lang}s are @samp{C}, @samp{C++}, and @samp{Java}.
4382 The linker will iterate over the list of symbols at the link time and
4383 demangle them according to @samp{lang} before matching them to the
4384 patterns specified in @samp{version-script-commands}.
4387 @section Expressions in Linker Scripts
4390 The syntax for expressions in the linker script language is identical to
4391 that of C expressions. All expressions are evaluated as integers. All
4392 expressions are evaluated in the same size, which is 32 bits if both the
4393 host and target are 32 bits, and is otherwise 64 bits.
4395 You can use and set symbol values in expressions.
4397 The linker defines several special purpose builtin functions for use in
4401 * Constants:: Constants
4402 * Symbols:: Symbol Names
4403 * Location Counter:: The Location Counter
4404 * Operators:: Operators
4405 * Evaluation:: Evaluation
4406 * Expression Section:: The Section of an Expression
4407 * Builtin Functions:: Builtin Functions
4411 @subsection Constants
4412 @cindex integer notation
4413 @cindex constants in linker scripts
4414 All constants are integers.
4416 As in C, the linker considers an integer beginning with @samp{0} to be
4417 octal, and an integer beginning with @samp{0x} or @samp{0X} to be
4418 hexadecimal. The linker considers other integers to be decimal.
4420 @cindex scaled integers
4421 @cindex K and M integer suffixes
4422 @cindex M and K integer suffixes
4423 @cindex suffixes for integers
4424 @cindex integer suffixes
4425 In addition, you can use the suffixes @code{K} and @code{M} to scale a
4429 @c END TEXI2ROFF-KILL
4430 @code{1024} or @code{1024*1024}
4434 ${\rm 1024}$ or ${\rm 1024}^2$
4436 @c END TEXI2ROFF-KILL
4437 respectively. For example, the following all refer to the same quantity:
4445 @subsection Symbol Names
4446 @cindex symbol names
4448 @cindex quoted symbol names
4450 Unless quoted, symbol names start with a letter, underscore, or period
4451 and may include letters, digits, underscores, periods, and hyphens.
4452 Unquoted symbol names must not conflict with any keywords. You can
4453 specify a symbol which contains odd characters or has the same name as a
4454 keyword by surrounding the symbol name in double quotes:
4457 "with a space" = "also with a space" + 10;
4460 Since symbols can contain many non-alphabetic characters, it is safest
4461 to delimit symbols with spaces. For example, @samp{A-B} is one symbol,
4462 whereas @samp{A - B} is an expression involving subtraction.
4464 @node Location Counter
4465 @subsection The Location Counter
4468 @cindex location counter
4469 @cindex current output location
4470 The special linker variable @dfn{dot} @samp{.} always contains the
4471 current output location counter. Since the @code{.} always refers to a
4472 location in an output section, it may only appear in an expression
4473 within a @code{SECTIONS} command. The @code{.} symbol may appear
4474 anywhere that an ordinary symbol is allowed in an expression.
4477 Assigning a value to @code{.} will cause the location counter to be
4478 moved. This may be used to create holes in the output section. The
4479 location counter may never be moved backwards.
4495 In the previous example, the @samp{.text} section from @file{file1} is
4496 located at the beginning of the output section @samp{output}. It is
4497 followed by a 1000 byte gap. Then the @samp{.text} section from
4498 @file{file2} appears, also with a 1000 byte gap following before the
4499 @samp{.text} section from @file{file3}. The notation @samp{= 0x12345678}
4500 specifies what data to write in the gaps (@pxref{Output Section Fill}).
4502 @cindex dot inside sections
4503 Note: @code{.} actually refers to the byte offset from the start of the
4504 current containing object. Normally this is the @code{SECTIONS}
4505 statement, whose start address is 0, hence @code{.} can be used as an
4506 absolute address. If @code{.} is used inside a section description
4507 however, it refers to the byte offset from the start of that section,
4508 not an absolute address. Thus in a script like this:
4526 The @samp{.text} section will be assigned a starting address of 0x100
4527 and a size of exactly 0x200 bytes, even if there is not enough data in
4528 the @samp{.text} input sections to fill this area. (If there is too
4529 much data, an error will be produced because this would be an attempt to
4530 move @code{.} backwards). The @samp{.data} section will start at 0x500
4531 and it will have an extra 0x600 bytes worth of space after the end of
4532 the values from the @samp{.data} input sections and before the end of
4533 the @samp{.data} output section itself.
4535 @cindex dot outside sections
4536 Setting symbols to the value of the location counter outside of an
4537 output section statement can result in unexpected values if the linker
4538 needs to place orphan sections. For example, given the following:
4544 .text: @{ *(.text) @}
4548 .data: @{ *(.data) @}
4553 If the linker needs to place some input section, e.g. @code{.rodata},
4554 not mentioned in the script, it might choose to place that section
4555 between @code{.text} and @code{.data}. You might think the linker
4556 should place @code{.rodata} on the blank line in the above script, but
4557 blank lines are of no particular significance to the linker. As well,
4558 the linker doesn't associate the above symbol names with their
4559 sections. Instead, it assumes that all assignments or other
4560 statements belong to the previous output section, except for the
4561 special case of an assignment to @code{.}. I.e., the linker will
4562 place the orphan @code{.rodata} section as if the script was written
4569 .text: @{ *(.text) @}
4573 .rodata: @{ *(.rodata) @}
4574 .data: @{ *(.data) @}
4579 This may or may not be the script author's intention for the value of
4580 @code{start_of_data}. One way to influence the orphan section
4581 placement is to assign the location counter to itself, as the linker
4582 assumes that an assignment to @code{.} is setting the start address of
4583 a following output section and thus should be grouped with that
4584 section. So you could write:
4590 .text: @{ *(.text) @}
4595 .data: @{ *(.data) @}
4600 Now, the orphan @code{.rodata} section will be placed between
4601 @code{end_of_text} and @code{start_of_data}.
4605 @subsection Operators
4606 @cindex operators for arithmetic
4607 @cindex arithmetic operators
4608 @cindex precedence in expressions
4609 The linker recognizes the standard C set of arithmetic operators, with
4610 the standard bindings and precedence levels:
4613 @c END TEXI2ROFF-KILL
4615 precedence associativity Operators Notes
4621 5 left == != > < <= >=
4627 11 right &= += -= *= /= (2)
4631 (1) Prefix operators
4632 (2) @xref{Assignments}.
4636 \vskip \baselineskip
4637 %"lispnarrowing" is the extra indent used generally for smallexample
4638 \hskip\lispnarrowing\vbox{\offinterlineskip
4641 {\vrule#&\strut\hfil\ #\ \hfil&\vrule#&\strut\hfil\ #\ \hfil&\vrule#&\strut\hfil\ {\tt #}\ \hfil&\vrule#\cr
4642 height2pt&\omit&&\omit&&\omit&\cr
4643 &Precedence&& Associativity &&{\rm Operators}&\cr
4644 height2pt&\omit&&\omit&&\omit&\cr
4646 height2pt&\omit&&\omit&&\omit&\cr
4648 % '176 is tilde, '~' in tt font
4649 &1&&left&&\qquad- \char'176\ !\qquad\dag&\cr
4650 &2&&left&&* / \%&\cr
4653 &5&&left&&== != > < <= >=&\cr
4656 &8&&left&&{\&\&}&\cr
4659 &11&&right&&\qquad\&= += -= *= /=\qquad\ddag&\cr
4661 height2pt&\omit&&\omit&&\omit&\cr}
4666 @obeylines@parskip=0pt@parindent=0pt
4667 @dag@quad Prefix operators.
4668 @ddag@quad @xref{Assignments}.
4671 @c END TEXI2ROFF-KILL
4674 @subsection Evaluation
4675 @cindex lazy evaluation
4676 @cindex expression evaluation order
4677 The linker evaluates expressions lazily. It only computes the value of
4678 an expression when absolutely necessary.
4680 The linker needs some information, such as the value of the start
4681 address of the first section, and the origins and lengths of memory
4682 regions, in order to do any linking at all. These values are computed
4683 as soon as possible when the linker reads in the linker script.
4685 However, other values (such as symbol values) are not known or needed
4686 until after storage allocation. Such values are evaluated later, when
4687 other information (such as the sizes of output sections) is available
4688 for use in the symbol assignment expression.
4690 The sizes of sections cannot be known until after allocation, so
4691 assignments dependent upon these are not performed until after
4694 Some expressions, such as those depending upon the location counter
4695 @samp{.}, must be evaluated during section allocation.
4697 If the result of an expression is required, but the value is not
4698 available, then an error results. For example, a script like the
4704 .text 9+this_isnt_constant :
4710 will cause the error message @samp{non constant expression for initial
4713 @node Expression Section
4714 @subsection The Section of an Expression
4715 @cindex expression sections
4716 @cindex absolute expressions
4717 @cindex relative expressions
4718 @cindex absolute and relocatable symbols
4719 @cindex relocatable and absolute symbols
4720 @cindex symbols, relocatable and absolute
4721 When the linker evaluates an expression, the result is either absolute
4722 or relative to some section. A relative expression is expressed as a
4723 fixed offset from the base of a section.
4725 The position of the expression within the linker script determines
4726 whether it is absolute or relative. An expression which appears within
4727 an output section definition is relative to the base of the output
4728 section. An expression which appears elsewhere will be absolute.
4730 A symbol set to a relative expression will be relocatable if you request
4731 relocatable output using the @samp{-r} option. That means that a
4732 further link operation may change the value of the symbol. The symbol's
4733 section will be the section of the relative expression.
4735 A symbol set to an absolute expression will retain the same value
4736 through any further link operation. The symbol will be absolute, and
4737 will not have any particular associated section.
4739 You can use the builtin function @code{ABSOLUTE} to force an expression
4740 to be absolute when it would otherwise be relative. For example, to
4741 create an absolute symbol set to the address of the end of the output
4742 section @samp{.data}:
4746 .data : @{ *(.data) _edata = ABSOLUTE(.); @}
4750 If @samp{ABSOLUTE} were not used, @samp{_edata} would be relative to the
4751 @samp{.data} section.
4753 @node Builtin Functions
4754 @subsection Builtin Functions
4755 @cindex functions in expressions
4756 The linker script language includes a number of builtin functions for
4757 use in linker script expressions.
4760 @item ABSOLUTE(@var{exp})
4761 @kindex ABSOLUTE(@var{exp})
4762 @cindex expression, absolute
4763 Return the absolute (non-relocatable, as opposed to non-negative) value
4764 of the expression @var{exp}. Primarily useful to assign an absolute
4765 value to a symbol within a section definition, where symbol values are
4766 normally section relative. @xref{Expression Section}.
4768 @item ADDR(@var{section})
4769 @kindex ADDR(@var{section})
4770 @cindex section address in expression
4771 Return the absolute address (the VMA) of the named @var{section}. Your
4772 script must previously have defined the location of that section. In
4773 the following example, @code{symbol_1} and @code{symbol_2} are assigned
4780 start_of_output_1 = ABSOLUTE(.);
4785 symbol_1 = ADDR(.output1);
4786 symbol_2 = start_of_output_1;
4792 @item ALIGN(@var{align})
4793 @itemx ALIGN(@var{exp},@var{align})
4794 @kindex ALIGN(@var{align})
4795 @kindex ALIGN(@var{exp},@var{align})
4796 @cindex round up location counter
4797 @cindex align location counter
4798 @cindex round up expression
4799 @cindex align expression
4800 Return the location counter (@code{.}) or arbitrary expression aligned
4801 to the next @var{align} boundary. The single operand @code{ALIGN}
4802 doesn't change the value of the location counter---it just does
4803 arithmetic on it. The two operand @code{ALIGN} allows an arbitrary
4804 expression to be aligned upwards (@code{ALIGN(@var{align})} is
4805 equivalent to @code{ALIGN(., @var{align})}).
4807 Here is an example which aligns the output @code{.data} section to the
4808 next @code{0x2000} byte boundary after the preceding section and sets a
4809 variable within the section to the next @code{0x8000} boundary after the
4814 .data ALIGN(0x2000): @{
4816 variable = ALIGN(0x8000);
4822 The first use of @code{ALIGN} in this example specifies the location of
4823 a section because it is used as the optional @var{address} attribute of
4824 a section definition (@pxref{Output Section Address}). The second use
4825 of @code{ALIGN} is used to defines the value of a symbol.
4827 The builtin function @code{NEXT} is closely related to @code{ALIGN}.
4829 @item BLOCK(@var{exp})
4830 @kindex BLOCK(@var{exp})
4831 This is a synonym for @code{ALIGN}, for compatibility with older linker
4832 scripts. It is most often seen when setting the address of an output
4835 @item DATA_SEGMENT_ALIGN(@var{maxpagesize}, @var{commonpagesize})
4836 @kindex DATA_SEGMENT_ALIGN(@var{maxpagesize}, @var{commonpagesize})
4837 This is equivalent to either
4839 (ALIGN(@var{maxpagesize}) + (. & (@var{maxpagesize} - 1)))
4843 (ALIGN(@var{maxpagesize}) + (. & (@var{maxpagesize} - @var{commonpagesize})))
4846 depending on whether the latter uses fewer @var{commonpagesize} sized pages
4847 for the data segment (area between the result of this expression and
4848 @code{DATA_SEGMENT_END}) than the former or not.
4849 If the latter form is used, it means @var{commonpagesize} bytes of runtime
4850 memory will be saved at the expense of up to @var{commonpagesize} wasted
4851 bytes in the on-disk file.
4853 This expression can only be used directly in @code{SECTIONS} commands, not in
4854 any output section descriptions and only once in the linker script.
4855 @var{commonpagesize} should be less or equal to @var{maxpagesize} and should
4856 be the system page size the object wants to be optimized for (while still
4857 working on system page sizes up to @var{maxpagesize}).
4862 . = DATA_SEGMENT_ALIGN(0x10000, 0x2000);
4865 @item DATA_SEGMENT_END(@var{exp})
4866 @kindex DATA_SEGMENT_END(@var{exp})
4867 This defines the end of data segment for @code{DATA_SEGMENT_ALIGN}
4868 evaluation purposes.
4871 . = DATA_SEGMENT_END(.);
4874 @item DATA_SEGMENT_RELRO_END(@var{offset}, @var{exp})
4875 @kindex DATA_SEGMENT_RELRO_END(@var{offset}, @var{exp})
4876 This defines the end of the @code{PT_GNU_RELRO} segment when
4877 @samp{-z relro} option is used. Second argument is returned.
4878 When @samp{-z relro} option is not present, @code{DATA_SEGMENT_RELRO_END}
4879 does nothing, otherwise @code{DATA_SEGMENT_ALIGN} is padded so that
4880 @var{exp} + @var{offset} is aligned to the most commonly used page
4881 boundary for particular target. If present in the linker script,
4882 it must always come in between @code{DATA_SEGMENT_ALIGN} and
4883 @code{DATA_SEGMENT_END}.
4886 . = DATA_SEGMENT_RELRO_END(24, .);
4889 @item DEFINED(@var{symbol})
4890 @kindex DEFINED(@var{symbol})
4891 @cindex symbol defaults
4892 Return 1 if @var{symbol} is in the linker global symbol table and is
4893 defined before the statement using DEFINED in the script, otherwise
4894 return 0. You can use this function to provide
4895 default values for symbols. For example, the following script fragment
4896 shows how to set a global symbol @samp{begin} to the first location in
4897 the @samp{.text} section---but if a symbol called @samp{begin} already
4898 existed, its value is preserved:
4904 begin = DEFINED(begin) ? begin : . ;
4912 @item LENGTH(@var{memory})
4913 @kindex LENGTH(@var{memory})
4914 Return the length of the memory region named @var{memory}.
4916 @item LOADADDR(@var{section})
4917 @kindex LOADADDR(@var{section})
4918 @cindex section load address in expression
4919 Return the absolute LMA of the named @var{section}. This is normally
4920 the same as @code{ADDR}, but it may be different if the @code{AT}
4921 attribute is used in the output section definition (@pxref{Output
4925 @item MAX(@var{exp1}, @var{exp2})
4926 Returns the maximum of @var{exp1} and @var{exp2}.
4929 @item MIN(@var{exp1}, @var{exp2})
4930 Returns the minimum of @var{exp1} and @var{exp2}.
4932 @item NEXT(@var{exp})
4933 @kindex NEXT(@var{exp})
4934 @cindex unallocated address, next
4935 Return the next unallocated address that is a multiple of @var{exp}.
4936 This function is closely related to @code{ALIGN(@var{exp})}; unless you
4937 use the @code{MEMORY} command to define discontinuous memory for the
4938 output file, the two functions are equivalent.
4940 @item ORIGIN(@var{memory})
4941 @kindex ORIGIN(@var{memory})
4942 Return the origin of the memory region named @var{memory}.
4944 @item SEGMENT_START(@var{segment}, @var{default})
4945 @kindex SEGMENT_START(@var{segment}, @var{default})
4946 Return the base address of the named @var{segment}. If an explicit
4947 value has been given for this segment (with a command-line @samp{-T}
4948 option) that value will be returned; otherwise the value will be
4949 @var{default}. At present, the @samp{-T} command-line option can only
4950 be used to set the base address for the ``text'', ``data'', and
4951 ``bss'' sections, but you use @code{SEGMENT_START} with any segment
4954 @item SIZEOF(@var{section})
4955 @kindex SIZEOF(@var{section})
4956 @cindex section size
4957 Return the size in bytes of the named @var{section}, if that section has
4958 been allocated. If the section has not been allocated when this is
4959 evaluated, the linker will report an error. In the following example,
4960 @code{symbol_1} and @code{symbol_2} are assigned identical values:
4969 symbol_1 = .end - .start ;
4970 symbol_2 = SIZEOF(.output);
4975 @item SIZEOF_HEADERS
4976 @itemx sizeof_headers
4977 @kindex SIZEOF_HEADERS
4979 Return the size in bytes of the output file's headers. This is
4980 information which appears at the start of the output file. You can use
4981 this number when setting the start address of the first section, if you
4982 choose, to facilitate paging.
4984 @cindex not enough room for program headers
4985 @cindex program headers, not enough room
4986 When producing an ELF output file, if the linker script uses the
4987 @code{SIZEOF_HEADERS} builtin function, the linker must compute the
4988 number of program headers before it has determined all the section
4989 addresses and sizes. If the linker later discovers that it needs
4990 additional program headers, it will report an error @samp{not enough
4991 room for program headers}. To avoid this error, you must avoid using
4992 the @code{SIZEOF_HEADERS} function, or you must rework your linker
4993 script to avoid forcing the linker to use additional program headers, or
4994 you must define the program headers yourself using the @code{PHDRS}
4995 command (@pxref{PHDRS}).
4998 @node Implicit Linker Scripts
4999 @section Implicit Linker Scripts
5000 @cindex implicit linker scripts
5001 If you specify a linker input file which the linker can not recognize as
5002 an object file or an archive file, it will try to read the file as a
5003 linker script. If the file can not be parsed as a linker script, the
5004 linker will report an error.
5006 An implicit linker script will not replace the default linker script.
5008 Typically an implicit linker script would contain only symbol
5009 assignments, or the @code{INPUT}, @code{GROUP}, or @code{VERSION}
5012 Any input files read because of an implicit linker script will be read
5013 at the position in the command line where the implicit linker script was
5014 read. This can affect archive searching.
5017 @node Machine Dependent
5018 @chapter Machine Dependent Features
5020 @cindex machine dependencies
5021 @command{ld} has additional features on some platforms; the following
5022 sections describe them. Machines where @command{ld} has no additional
5023 functionality are not listed.
5027 * H8/300:: @command{ld} and the H8/300
5030 * i960:: @command{ld} and the Intel 960 family
5033 * ARM:: @command{ld} and the ARM family
5036 * HPPA ELF32:: @command{ld} and HPPA 32-bit ELF
5039 * MMIX:: @command{ld} and MMIX
5042 * MSP430:: @command{ld} and MSP430
5045 * M68HC11/68HC12:: @code{ld} and the Motorola 68HC11 and 68HC12 families
5048 * TI COFF:: @command{ld} and TI COFF
5051 * WIN32:: @command{ld} and WIN32 (cygwin/mingw)
5054 * Xtensa:: @command{ld} and Xtensa Processors
5065 @section @command{ld} and the H8/300
5067 @cindex H8/300 support
5068 For the H8/300, @command{ld} can perform these global optimizations when
5069 you specify the @samp{--relax} command-line option.
5072 @cindex relaxing on H8/300
5073 @item relaxing address modes
5074 @command{ld} finds all @code{jsr} and @code{jmp} instructions whose
5075 targets are within eight bits, and turns them into eight-bit
5076 program-counter relative @code{bsr} and @code{bra} instructions,
5079 @cindex synthesizing on H8/300
5080 @item synthesizing instructions
5081 @c FIXME: specifically mov.b, or any mov instructions really?
5082 @command{ld} finds all @code{mov.b} instructions which use the
5083 sixteen-bit absolute address form, but refer to the top
5084 page of memory, and changes them to use the eight-bit address form.
5085 (That is: the linker turns @samp{mov.b @code{@@}@var{aa}:16} into
5086 @samp{mov.b @code{@@}@var{aa}:8} whenever the address @var{aa} is in the
5087 top page of memory).
5089 @item bit manipulation instructions
5090 @command{ld} finds all bit manipulation instructions like @code{band, bclr,
5091 biand, bild, bior, bist, bixor, bld, bnot, bor, bset, bst, btst, bxor}
5092 which use 32 bit and 16 bit absolute address form, but refer to the top
5093 page of memory, and changes them to use the 8 bit address form.
5094 (That is: the linker turns @samp{bset #xx:3,@code{@@}@var{aa}:32} into
5095 @samp{bset #xx:3,@code{@@}@var{aa}:8} whenever the address @var{aa} is in
5096 the top page of memory).
5098 @item system control instructions
5099 @command{ld} finds all @code{ldc.w, stc.w} instrcutions which use the
5100 32 bit absolute address form, but refer to the top page of memory, and
5101 changes them to use 16 bit address form.
5102 (That is: the linker turns @samp{ldc.w @code{@@}@var{aa}:32,ccr} into
5103 @samp{ldc.w @code{@@}@var{aa}:16,ccr} whenever the address @var{aa} is in
5104 the top page of memory).
5114 @c This stuff is pointless to say unless you're especially concerned
5115 @c with Renesas chips; don't enable it for generic case, please.
5117 @chapter @command{ld} and Other Renesas Chips
5119 @command{ld} also supports the Renesas (formerly Hitachi) H8/300H,
5120 H8/500, and SH chips. No special features, commands, or command-line
5121 options are required for these chips.
5131 @section @command{ld} and the Intel 960 Family
5133 @cindex i960 support
5135 You can use the @samp{-A@var{architecture}} command line option to
5136 specify one of the two-letter names identifying members of the 960
5137 family; the option specifies the desired output target, and warns of any
5138 incompatible instructions in the input files. It also modifies the
5139 linker's search strategy for archive libraries, to support the use of
5140 libraries specific to each particular architecture, by including in the
5141 search loop names suffixed with the string identifying the architecture.
5143 For example, if your @command{ld} command line included @w{@samp{-ACA}} as
5144 well as @w{@samp{-ltry}}, the linker would look (in its built-in search
5145 paths, and in any paths you specify with @samp{-L}) for a library with
5158 The first two possibilities would be considered in any event; the last
5159 two are due to the use of @w{@samp{-ACA}}.
5161 You can meaningfully use @samp{-A} more than once on a command line, since
5162 the 960 architecture family allows combination of target architectures; each
5163 use will add another pair of name variants to search for when @w{@samp{-l}}
5164 specifies a library.
5166 @cindex @option{--relax} on i960
5167 @cindex relaxing on i960
5168 @command{ld} supports the @samp{--relax} option for the i960 family. If
5169 you specify @samp{--relax}, @command{ld} finds all @code{balx} and
5170 @code{calx} instructions whose targets are within 24 bits, and turns
5171 them into 24-bit program-counter relative @code{bal} and @code{cal}
5172 instructions, respectively. @command{ld} also turns @code{cal}
5173 instructions into @code{bal} instructions when it determines that the
5174 target subroutine is a leaf routine (that is, the target subroutine does
5175 not itself call any subroutines).
5192 @node M68HC11/68HC12
5193 @section @command{ld} and the Motorola 68HC11 and 68HC12 families
5195 @cindex M68HC11 and 68HC12 support
5197 @subsection Linker Relaxation
5199 For the Motorola 68HC11, @command{ld} can perform these global
5200 optimizations when you specify the @samp{--relax} command-line option.
5203 @cindex relaxing on M68HC11
5204 @item relaxing address modes
5205 @command{ld} finds all @code{jsr} and @code{jmp} instructions whose
5206 targets are within eight bits, and turns them into eight-bit
5207 program-counter relative @code{bsr} and @code{bra} instructions,
5210 @command{ld} also looks at all 16-bit extended addressing modes and
5211 transforms them in a direct addressing mode when the address is in
5212 page 0 (between 0 and 0x0ff).
5214 @item relaxing gcc instruction group
5215 When @command{gcc} is called with @option{-mrelax}, it can emit group
5216 of instructions that the linker can optimize to use a 68HC11 direct
5217 addressing mode. These instructions consists of @code{bclr} or
5218 @code{bset} instructions.
5222 @subsection Trampoline Generation
5224 @cindex trampoline generation on M68HC11
5225 @cindex trampoline generation on M68HC12
5226 For 68HC11 and 68HC12, @command{ld} can generate trampoline code to
5227 call a far function using a normal @code{jsr} instruction. The linker
5228 will also change the relocation to some far function to use the
5229 trampoline address instead of the function address. This is typically the
5230 case when a pointer to a function is taken. The pointer will in fact
5231 point to the function trampoline.
5239 @section @command{ld} and the ARM family
5241 @cindex ARM interworking support
5242 @kindex --support-old-code
5243 For the ARM, @command{ld} will generate code stubs to allow functions calls
5244 betweem ARM and Thumb code. These stubs only work with code that has
5245 been compiled and assembled with the @samp{-mthumb-interwork} command
5246 line option. If it is necessary to link with old ARM object files or
5247 libraries, which have not been compiled with the -mthumb-interwork
5248 option then the @samp{--support-old-code} command line switch should be
5249 given to the linker. This will make it generate larger stub functions
5250 which will work with non-interworking aware ARM code. Note, however,
5251 the linker does not support generating stubs for function calls to
5252 non-interworking aware Thumb code.
5254 @cindex thumb entry point
5255 @cindex entry point, thumb
5256 @kindex --thumb-entry=@var{entry}
5257 The @samp{--thumb-entry} switch is a duplicate of the generic
5258 @samp{--entry} switch, in that it sets the program's starting address.
5259 But it also sets the bottom bit of the address, so that it can be
5260 branched to using a BX instruction, and the program will start
5261 executing in Thumb mode straight away.
5265 The @samp{--be8} switch instructs @command{ld} to generate BE8 format
5266 executables. This option is only valid when linking big-endian objects.
5267 The resulting image will contain big-endian data and little-endian code.
5270 @kindex --target1-rel
5271 @kindex --target1-abs
5272 The @samp{R_ARM_TARGET1} relocation is typically used for entries in the
5273 @samp{.init_array} section. It is interpreted as either @samp{R_ARM_REL32}
5274 or @samp{R_ARM_ABS32}, depending on the target. The @samp{--target1-rel}
5275 and @samp{--target1-abs} switches override the default.
5278 @kindex --target2=@var{type}
5279 The @samp{--target2=type} switch overrides the default definition of the
5280 @samp{R_ARM_TARGET2} relocation. Valid values for @samp{type}, their
5281 meanings, and target defaults are as follows:
5284 @samp{R_ARM_REL32} (arm*-*-elf, arm*-*-eabi)
5286 @samp{R_ARM_ABS32} (arm*-*-symbianelf)
5288 @samp{R_ARM_GOT_PREL} (arm*-*-linux, arm*-*-*bsd)
5293 The @samp{R_ARM_V4BX} relocation (defined by the ARM AAELF
5294 specification) enables objects compiled for the ARMv4 architecture to be
5295 interworking-safe when linked with other objects compiled for ARMv4t, but
5296 also allows pure ARMv4 binaries to be built from the same ARMv4 objects.
5298 In the latter case, the switch @option{--fix-v4bx} must be passed to the
5299 linker, which causes v4t @code{BX rM} instructions to be rewritten as
5300 @code{MOV PC,rM}, since v4 processors do not have a @code{BX} instruction.
5302 In the former case, the switch should not be used, and @samp{R_ARM_V4BX}
5303 relocations are ignored.
5307 The @samp{--use-blx} switch enables the linker to use ARM/Thumb
5308 BLX instructions (available on ARMv5t and above) in various
5309 situations. Currently it is used to perform calls via the PLT from Thumb
5310 code using BLX rather than using BX and a mode-switching stub before
5311 each PLT entry. This should lead to such calls executing slightly faster.
5313 This option is enabled implicitly for SymbianOS, so there is no need to
5314 specify it if you are using that target.
5327 @section @command{ld} and HPPA 32-bit ELF Support
5328 @cindex HPPA multiple sub-space stubs
5329 @kindex --multi-subspace
5330 When generating a shared library, @command{ld} will by default generate
5331 import stubs suitable for use with a single sub-space application.
5332 The @samp{--multi-subspace} switch causes @command{ld} to generate export
5333 stubs, and different (larger) import stubs suitable for use with
5334 multiple sub-spaces.
5336 @cindex HPPA stub grouping
5337 @kindex --stub-group-size=@var{N}
5338 Long branch stubs and import/export stubs are placed by @command{ld} in
5339 stub sections located between groups of input sections.
5340 @samp{--stub-group-size} specifies the maximum size of a group of input
5341 sections handled by one stub section. Since branch offsets are signed,
5342 a stub section may serve two groups of input sections, one group before
5343 the stub section, and one group after it. However, when using
5344 conditional branches that require stubs, it may be better (for branch
5345 prediction) that stub sections only serve one group of input sections.
5346 A negative value for @samp{N} chooses this scheme, ensuring that
5347 branches to stubs always use a negative offset. Two special values of
5348 @samp{N} are recognized, @samp{1} and @samp{-1}. These both instruct
5349 @command{ld} to automatically size input section groups for the branch types
5350 detected, with the same behaviour regarding stub placement as other
5351 positive or negative values of @samp{N} respectively.
5353 Note that @samp{--stub-group-size} does not split input sections. A
5354 single input section larger than the group size specified will of course
5355 create a larger group (of one section). If input sections are too
5356 large, it may not be possible for a branch to reach its stub.
5369 @section @code{ld} and MMIX
5370 For MMIX, there is a choice of generating @code{ELF} object files or
5371 @code{mmo} object files when linking. The simulator @code{mmix}
5372 understands the @code{mmo} format. The binutils @code{objcopy} utility
5373 can translate between the two formats.
5375 There is one special section, the @samp{.MMIX.reg_contents} section.
5376 Contents in this section is assumed to correspond to that of global
5377 registers, and symbols referring to it are translated to special symbols,
5378 equal to registers. In a final link, the start address of the
5379 @samp{.MMIX.reg_contents} section corresponds to the first allocated
5380 global register multiplied by 8. Register @code{$255} is not included in
5381 this section; it is always set to the program entry, which is at the
5382 symbol @code{Main} for @code{mmo} files.
5384 Symbols with the prefix @code{__.MMIX.start.}, for example
5385 @code{__.MMIX.start..text} and @code{__.MMIX.start..data} are special;
5386 there must be only one each, even if they are local. The default linker
5387 script uses these to set the default start address of a section.
5389 Initial and trailing multiples of zero-valued 32-bit words in a section,
5390 are left out from an mmo file.
5403 @section @code{ld} and MSP430
5404 For the MSP430 it is possible to select the MPU architecture. The flag @samp{-m [mpu type]}
5405 will select an appropriate linker script for selected MPU type. (To get a list of known MPUs
5406 just pass @samp{-m help} option to the linker).
5408 @cindex MSP430 extra sections
5409 The linker will recognize some extra sections which are MSP430 specific:
5412 @item @samp{.vectors}
5413 Defines a portion of ROM where interrupt vectors located.
5415 @item @samp{.bootloader}
5416 Defines the bootloader portion of the ROM (if applicable). Any code
5417 in this section will be uploaded to the MPU.
5419 @item @samp{.infomem}
5420 Defines an information memory section (if applicable). Any code in
5421 this section will be uploaded to the MPU.
5423 @item @samp{.infomemnobits}
5424 This is the same as the @samp{.infomem} section except that any code
5425 in this section will not be uploaded to the MPU.
5427 @item @samp{.noinit}
5428 Denotes a portion of RAM located above @samp{.bss} section.
5430 The last two sections are used by gcc.
5444 @section @command{ld}'s Support for Various TI COFF Versions
5445 @cindex TI COFF versions
5446 @kindex --format=@var{version}
5447 The @samp{--format} switch allows selection of one of the various
5448 TI COFF versions. The latest of this writing is 2; versions 0 and 1 are
5449 also supported. The TI COFF versions also vary in header byte-order
5450 format; @command{ld} will read any version or byte order, but the output
5451 header format depends on the default specified by the specific target.
5464 @section @command{ld} and WIN32 (cygwin/mingw)
5466 This section describes some of the win32 specific @command{ld} issues.
5467 See @ref{Options,,Command Line Options} for detailed decription of the
5468 command line options mentioned here.
5471 @cindex import libraries
5472 @item import libraries
5473 The standard Windows linker creates and uses so-called import
5474 libraries, which contains information for linking to dll's. They are
5475 regular static archives and are handled as any other static
5476 archive. The cygwin and mingw ports of @command{ld} have specific
5477 support for creating such libraries provided with the
5478 @samp{--out-implib} command line option.
5480 @item exporting DLL symbols
5481 @cindex exporting DLL symbols
5482 The cygwin/mingw @command{ld} has several ways to export symbols for dll's.
5485 @item using auto-export functionality
5486 @cindex using auto-export functionality
5487 By default @command{ld} exports symbols with the auto-export functionality,
5488 which is controlled by the following command line options:
5491 @item --export-all-symbols [This is the default]
5492 @item --exclude-symbols
5493 @item --exclude-libs
5496 If, however, @samp{--export-all-symbols} is not given explicitly on the
5497 command line, then the default auto-export behavior will be @emph{disabled}
5498 if either of the following are true:
5501 @item A DEF file is used.
5502 @item Any symbol in any object file was marked with the __declspec(dllexport) attribute.
5505 @item using a DEF file
5506 @cindex using a DEF file
5507 Another way of exporting symbols is using a DEF file. A DEF file is
5508 an ASCII file containing definitions of symbols which should be
5509 exported when a dll is created. Usually it is named @samp{<dll
5510 name>.def} and is added as any other object file to the linker's
5511 command line. The file's name must end in @samp{.def} or @samp{.DEF}.
5514 gcc -o <output> <objectfiles> <dll name>.def
5517 Using a DEF file turns off the normal auto-export behavior, unless the
5518 @samp{--export-all-symbols} option is also used.
5520 Here is an example of a DEF file for a shared library called @samp{xyz.dll}:
5523 LIBRARY "xyz.dll" BASE=0x10000000
5531 This example defines a base address and three symbols. The third
5532 symbol is an alias for the second. For the complete format
5533 specification see ld/deffilep.y in the binutils sources.
5535 @cindex creating a DEF file
5536 While linking a shared dll, @command{ld} is able to create a DEF file
5537 with the @samp{--output-def <file>} command line option.
5539 @item Using decorations
5540 @cindex Using decorations
5541 Another way of marking symbols for export is to modify the source code
5542 itself, so that when building the DLL each symbol to be exported is
5546 __declspec(dllexport) int a_variable
5547 __declspec(dllexport) void a_function(int with_args)
5550 All such symbols will be exported from the DLL. If, however,
5551 any of the object files in the DLL contain symbols decorated in
5552 this way, then the normal auto-export behavior is disabled, unless
5553 the @samp{--export-all-symbols} option is also used.
5555 Note that object files that wish to access these symbols must @emph{not}
5556 decorate them with dllexport. Instead, they should use dllimport,
5560 __declspec(dllimport) int a_variable
5561 __declspec(dllimport) void a_function(int with_args)
5564 This complicates the structure of library header files, because
5565 when included by the library itself the header must declare the
5566 variables and functions as dllexport, but when included by client
5567 code the header must declare them as dllimport. There are a number
5568 of idioms that are typically used to do this; often client code can
5569 omit the __declspec() declaration completely. See
5570 @samp{--enable-auto-import} and @samp{automatic data imports} for more
5574 @cindex automatic data imports
5575 @item automatic data imports
5576 The standard Windows dll format supports data imports from dlls only
5577 by adding special decorations (dllimport/dllexport), which let the
5578 compiler produce specific assembler instructions to deal with this
5579 issue. This increases the effort necessary to port existing Un*x
5580 code to these platforms, especially for large
5581 c++ libraries and applications. The auto-import feature, which was
5582 initially provided by Paul Sokolovsky, allows one to omit the
5583 decorations to archieve a behavior that conforms to that on POSIX/Un*x
5584 platforms. This feature is enabled with the @samp{--enable-auto-import}
5585 command-line option, although it is enabled by default on cygwin/mingw.
5586 The @samp{--enable-auto-import} option itself now serves mainly to
5587 suppress any warnings that are ordinarily emitted when linked objects
5588 trigger the feature's use.
5590 auto-import of variables does not always work flawlessly without
5591 additional assistance. Sometimes, you will see this message
5593 "variable '<var>' can't be auto-imported. Please read the
5594 documentation for ld's @code{--enable-auto-import} for details."
5596 The @samp{--enable-auto-import} documentation explains why this error
5597 occurs, and several methods that can be used to overcome this difficulty.
5598 One of these methods is the @emph{runtime pseudo-relocs} feature, described
5601 @cindex runtime pseudo-relocation
5602 For complex variables imported from DLLs (such as structs or classes),
5603 object files typically contain a base address for the variable and an
5604 offset (@emph{addend}) within the variable--to specify a particular
5605 field or public member, for instance. Unfortunately, the runtime loader used
5606 in win32 environments is incapable of fixing these references at runtime
5607 without the additional information supplied by dllimport/dllexport decorations.
5608 The standard auto-import feature described above is unable to resolve these
5611 The @samp{--enable-runtime-pseudo-relocs} switch allows these references to
5612 be resolved without error, while leaving the task of adjusting the references
5613 themselves (with their non-zero addends) to specialized code provided by the
5614 runtime environment. Recent versions of the cygwin and mingw environments and
5615 compilers provide this runtime support; older versions do not. However, the
5616 support is only necessary on the developer's platform; the compiled result will
5617 run without error on an older system.
5619 @samp{--enable-runtime-pseudo-relocs} is not the default; it must be explicitly
5622 @cindex direct linking to a dll
5623 @item direct linking to a dll
5624 The cygwin/mingw ports of @command{ld} support the direct linking,
5625 including data symbols, to a dll without the usage of any import
5626 libraries. This is much faster and uses much less memory than does the
5627 traditional import library method, expecially when linking large
5628 libraries or applications. When @command{ld} creates an import lib, each
5629 function or variable exported from the dll is stored in its own bfd, even
5630 though a single bfd could contain many exports. The overhead involved in
5631 storing, loading, and processing so many bfd's is quite large, and explains the
5632 tremendous time, memory, and storage needed to link against particularly
5633 large or complex libraries when using import libs.
5635 Linking directly to a dll uses no extra command-line switches other than
5636 @samp{-L} and @samp{-l}, because @command{ld} already searches for a number
5637 of names to match each library. All that is needed from the developer's
5638 perspective is an understanding of this search, in order to force ld to
5639 select the dll instead of an import library.
5642 For instance, when ld is called with the argument @samp{-lxxx} it will attempt
5643 to find, in the first directory of its search path,
5654 before moving on to the next directory in the search path.
5656 (*) Actually, this is not @samp{cygxxx.dll} but in fact is @samp{<prefix>xxx.dll},
5657 where @samp{<prefix>} is set by the @command{ld} option
5658 @samp{--dll-search-prefix=<prefix>}. In the case of cygwin, the standard gcc spec
5659 file includes @samp{--dll-search-prefix=cyg}, so in effect we actually search for
5662 Other win32-based unix environments, such as mingw or pw32, may use other
5663 @samp{<prefix>}es, although at present only cygwin makes use of this feature. It
5664 was originally intended to help avoid name conflicts among dll's built for the
5665 various win32/un*x environments, so that (for example) two versions of a zlib dll
5666 could coexist on the same machine.
5668 The generic cygwin/mingw path layout uses a @samp{bin} directory for
5669 applications and dll's and a @samp{lib} directory for the import
5670 libraries (using cygwin nomenclature):
5676 libxxx.dll.a (in case of dll's)
5677 libxxx.a (in case of static archive)
5680 Linking directly to a dll without using the import library can be
5683 1. Use the dll directly by adding the @samp{bin} path to the link line
5685 gcc -Wl,-verbose -o a.exe -L../bin/ -lxxx
5688 However, as the dll's often have version numbers appended to their names
5689 (@samp{cygncurses-5.dll}) this will often fail, unless one specifies
5690 @samp{-L../bin -lncurses-5} to include the version. Import libs are generally
5691 not versioned, and do not have this difficulty.
5693 2. Create a symbolic link from the dll to a file in the @samp{lib}
5694 directory according to the above mentioned search pattern. This
5695 should be used to avoid unwanted changes in the tools needed for
5699 ln -s bin/cygxxx.dll lib/[cyg|lib|]xxx.dll[.a]
5702 Then you can link without any make environment changes.
5705 gcc -Wl,-verbose -o a.exe -L../lib/ -lxxx
5708 This technique also avoids the version number problems, because the following is
5715 libxxx.dll.a -> ../bin/cygxxx-5.dll
5718 Linking directly to a dll without using an import lib will work
5719 even when auto-import features are exercised, and even when
5720 @samp{--enable-runtime-pseudo-relocs} is used.
5722 Given the improvements in speed and memory usage, one might justifiably
5723 wonder why import libraries are used at all. There are two reasons:
5725 1. Until recently, the link-directly-to-dll functionality did @emph{not}
5726 work with auto-imported data.
5728 2. Sometimes it is necessary to include pure static objects within the
5729 import library (which otherwise contains only bfd's for indirection
5730 symbols that point to the exports of a dll). Again, the import lib
5731 for the cygwin kernel makes use of this ability, and it is not
5732 possible to do this without an import lib.
5734 So, import libs are not going away. But the ability to replace
5735 true import libs with a simple symbolic link to (or a copy of)
5736 a dll, in most cases, is a useful addition to the suite of tools
5737 binutils makes available to the win32 developer. Given the
5738 massive improvements in memory requirements during linking, storage
5739 requirements, and linking speed, we expect that many developers
5740 will soon begin to use this feature whenever possible.
5742 @item symbol aliasing
5744 @item adding additional names
5745 Sometimes, it is useful to export symbols with additional names.
5746 A symbol @samp{foo} will be exported as @samp{foo}, but it can also be
5747 exported as @samp{_foo} by using special directives in the DEF file
5748 when creating the dll. This will affect also the optional created
5749 import library. Consider the following DEF file:
5752 LIBRARY "xyz.dll" BASE=0x61000000
5759 The line @samp{_foo = foo} maps the symbol @samp{foo} to @samp{_foo}.
5761 Another method for creating a symbol alias is to create it in the
5762 source code using the "weak" attribute:
5765 void foo () @{ /* Do something. */; @}
5766 void _foo () __attribute__ ((weak, alias ("foo")));
5769 See the gcc manual for more information about attributes and weak
5772 @item renaming symbols
5773 Sometimes it is useful to rename exports. For instance, the cygwin
5774 kernel does this regularly. A symbol @samp{_foo} can be exported as
5775 @samp{foo} but not as @samp{_foo} by using special directives in the
5776 DEF file. (This will also affect the import library, if it is
5777 created). In the following example:
5780 LIBRARY "xyz.dll" BASE=0x61000000
5786 The line @samp{_foo = foo} maps the exported symbol @samp{foo} to
5790 Note: using a DEF file disables the default auto-export behavior,
5791 unless the @samp{--export-all-symbols} command line option is used.
5792 If, however, you are trying to rename symbols, then you should list
5793 @emph{all} desired exports in the DEF file, including the symbols
5794 that are not being renamed, and do @emph{not} use the
5795 @samp{--export-all-symbols} option. If you list only the
5796 renamed symbols in the DEF file, and use @samp{--export-all-symbols}
5797 to handle the other symbols, then the both the new names @emph{and}
5798 the original names for the renamed symbols will be exported.
5799 In effect, you'd be aliasing those symbols, not renaming them,
5800 which is probably not what you wanted.
5802 @cindex weak externals
5803 @item weak externals
5804 The Windows object format, PE, specifies a form of weak symbols called
5805 weak externals. When a weak symbol is linked and the symbol is not
5806 defined, the weak symbol becomes an alias for some other symbol. There
5807 are three variants of weak externals:
5809 @item Definition is searched for in objects and libraries, historically
5810 called lazy externals.
5811 @item Definition is searched for only in other objects, not in libraries.
5812 This form is not presently implemented.
5813 @item No search; the symbol is an alias. This form is not presently
5816 As a GNU extension, weak symbols that do not specify an alternate symbol
5817 are supported. If the symbol is undefined when linking, the symbol
5818 uses a default value.
5832 @section @code{ld} and Xtensa Processors
5834 @cindex Xtensa processors
5835 The default @command{ld} behavior for Xtensa processors is to interpret
5836 @code{SECTIONS} commands so that lists of explicitly named sections in a
5837 specification with a wildcard file will be interleaved when necessary to
5838 keep literal pools within the range of PC-relative load offsets. For
5839 example, with the command:
5851 @command{ld} may interleave some of the @code{.literal}
5852 and @code{.text} sections from different object files to ensure that the
5853 literal pools are within the range of PC-relative load offsets. A valid
5854 interleaving might place the @code{.literal} sections from an initial
5855 group of files followed by the @code{.text} sections of that group of
5856 files. Then, the @code{.literal} sections from the rest of the files
5857 and the @code{.text} sections from the rest of the files would follow.
5859 @cindex @option{--relax} on Xtensa
5860 @cindex relaxing on Xtensa
5861 Relaxation is enabled by default for the Xtensa version of @command{ld} and
5862 provides two important link-time optimizations. The first optimization
5863 is to combine identical literal values to reduce code size. A redundant
5864 literal will be removed and all the @code{L32R} instructions that use it
5865 will be changed to reference an identical literal, as long as the
5866 location of the replacement literal is within the offset range of all
5867 the @code{L32R} instructions. The second optimization is to remove
5868 unnecessary overhead from assembler-generated ``longcall'' sequences of
5869 @code{L32R}/@code{CALLX@var{n}} when the target functions are within
5870 range of direct @code{CALL@var{n}} instructions.
5872 For each of these cases where an indirect call sequence can be optimized
5873 to a direct call, the linker will change the @code{CALLX@var{n}}
5874 instruction to a @code{CALL@var{n}} instruction, remove the @code{L32R}
5875 instruction, and remove the literal referenced by the @code{L32R}
5876 instruction if it is not used for anything else. Removing the
5877 @code{L32R} instruction always reduces code size but can potentially
5878 hurt performance by changing the alignment of subsequent branch targets.
5879 By default, the linker will always preserve alignments, either by
5880 switching some instructions between 24-bit encodings and the equivalent
5881 density instructions or by inserting a no-op in place of the @code{L32R}
5882 instruction that was removed. If code size is more important than
5883 performance, the @option{--size-opt} option can be used to prevent the
5884 linker from widening density instructions or inserting no-ops, except in
5885 a few cases where no-ops are required for correctness.
5887 The following Xtensa-specific command-line options can be used to
5890 @cindex Xtensa options
5894 Since the Xtensa version of @code{ld} enables the @option{--relax} option
5895 by default, the @option{--no-relax} option is provided to disable
5899 When optimizing indirect calls to direct calls, optimize for code size
5900 more than performance. With this option, the linker will not insert
5901 no-ops or widen density instructions to preserve branch target
5902 alignment. There may still be some cases where no-ops are required to
5903 preserve the correctness of the code.
5911 @ifclear SingleFormat
5916 @cindex object file management
5917 @cindex object formats available
5919 The linker accesses object and archive files using the BFD libraries.
5920 These libraries allow the linker to use the same routines to operate on
5921 object files whatever the object file format. A different object file
5922 format can be supported simply by creating a new BFD back end and adding
5923 it to the library. To conserve runtime memory, however, the linker and
5924 associated tools are usually configured to support only a subset of the
5925 object file formats available. You can use @code{objdump -i}
5926 (@pxref{objdump,,objdump,binutils.info,The GNU Binary Utilities}) to
5927 list all the formats available for your configuration.
5929 @cindex BFD requirements
5930 @cindex requirements for BFD
5931 As with most implementations, BFD is a compromise between
5932 several conflicting requirements. The major factor influencing
5933 BFD design was efficiency: any time used converting between
5934 formats is time which would not have been spent had BFD not
5935 been involved. This is partly offset by abstraction payback; since
5936 BFD simplifies applications and back ends, more time and care
5937 may be spent optimizing algorithms for a greater speed.
5939 One minor artifact of the BFD solution which you should bear in
5940 mind is the potential for information loss. There are two places where
5941 useful information can be lost using the BFD mechanism: during
5942 conversion and during output. @xref{BFD information loss}.
5945 * BFD outline:: How it works: an outline of BFD
5949 @section How It Works: An Outline of BFD
5950 @cindex opening object files
5951 @include bfdsumm.texi
5954 @node Reporting Bugs
5955 @chapter Reporting Bugs
5956 @cindex bugs in @command{ld}
5957 @cindex reporting bugs in @command{ld}
5959 Your bug reports play an essential role in making @command{ld} reliable.
5961 Reporting a bug may help you by bringing a solution to your problem, or
5962 it may not. But in any case the principal function of a bug report is
5963 to help the entire community by making the next version of @command{ld}
5964 work better. Bug reports are your contribution to the maintenance of
5967 In order for a bug report to serve its purpose, you must include the
5968 information that enables us to fix the bug.
5971 * Bug Criteria:: Have you found a bug?
5972 * Bug Reporting:: How to report bugs
5976 @section Have You Found a Bug?
5977 @cindex bug criteria
5979 If you are not sure whether you have found a bug, here are some guidelines:
5982 @cindex fatal signal
5983 @cindex linker crash
5984 @cindex crash of linker
5986 If the linker gets a fatal signal, for any input whatever, that is a
5987 @command{ld} bug. Reliable linkers never crash.
5989 @cindex error on valid input
5991 If @command{ld} produces an error message for valid input, that is a bug.
5993 @cindex invalid input
5995 If @command{ld} does not produce an error message for invalid input, that
5996 may be a bug. In the general case, the linker can not verify that
5997 object files are correct.
6000 If you are an experienced user of linkers, your suggestions for
6001 improvement of @command{ld} are welcome in any case.
6005 @section How to Report Bugs
6007 @cindex @command{ld} bugs, reporting
6009 A number of companies and individuals offer support for @sc{gnu}
6010 products. If you obtained @command{ld} from a support organization, we
6011 recommend you contact that organization first.
6013 You can find contact information for many support companies and
6014 individuals in the file @file{etc/SERVICE} in the @sc{gnu} Emacs
6017 Otherwise, send bug reports for @command{ld} to
6018 @samp{bug-binutils@@gnu.org}.
6020 The fundamental principle of reporting bugs usefully is this:
6021 @strong{report all the facts}. If you are not sure whether to state a
6022 fact or leave it out, state it!
6024 Often people omit facts because they think they know what causes the
6025 problem and assume that some details do not matter. Thus, you might
6026 assume that the name of a symbol you use in an example does not
6027 matter. Well, probably it does not, but one cannot be sure. Perhaps
6028 the bug is a stray memory reference which happens to fetch from the
6029 location where that name is stored in memory; perhaps, if the name
6030 were different, the contents of that location would fool the linker
6031 into doing the right thing despite the bug. Play it safe and give a
6032 specific, complete example. That is the easiest thing for you to do,
6033 and the most helpful.
6035 Keep in mind that the purpose of a bug report is to enable us to fix
6036 the bug if it is new to us. Therefore, always write your bug reports
6037 on the assumption that the bug has not been reported previously.
6039 Sometimes people give a few sketchy facts and ask, ``Does this ring a
6040 bell?'' This cannot help us fix a bug, so it is basically useless. We
6041 respond by asking for enough details to enable us to investigate.
6042 You might as well expedite matters by sending them to begin with.
6044 To enable us to fix the bug, you should include all these things:
6048 The version of @command{ld}. @command{ld} announces it if you start it with
6049 the @samp{--version} argument.
6051 Without this, we will not know whether there is any point in looking for
6052 the bug in the current version of @command{ld}.
6055 Any patches you may have applied to the @command{ld} source, including any
6056 patches made to the @code{BFD} library.
6059 The type of machine you are using, and the operating system name and
6063 What compiler (and its version) was used to compile @command{ld}---e.g.
6067 The command arguments you gave the linker to link your example and
6068 observe the bug. To guarantee you will not omit something important,
6069 list them all. A copy of the Makefile (or the output from make) is
6072 If we were to try to guess the arguments, we would probably guess wrong
6073 and then we might not encounter the bug.
6076 A complete input file, or set of input files, that will reproduce the
6077 bug. It is generally most helpful to send the actual object files
6078 provided that they are reasonably small. Say no more than 10K. For
6079 bigger files you can either make them available by FTP or HTTP or else
6080 state that you are willing to send the object file(s) to whomever
6081 requests them. (Note - your email will be going to a mailing list, so
6082 we do not want to clog it up with large attachments). But small
6083 attachments are best.
6085 If the source files were assembled using @code{gas} or compiled using
6086 @code{gcc}, then it may be OK to send the source files rather than the
6087 object files. In this case, be sure to say exactly what version of
6088 @code{gas} or @code{gcc} was used to produce the object files. Also say
6089 how @code{gas} or @code{gcc} were configured.
6092 A description of what behavior you observe that you believe is
6093 incorrect. For example, ``It gets a fatal signal.''
6095 Of course, if the bug is that @command{ld} gets a fatal signal, then we
6096 will certainly notice it. But if the bug is incorrect output, we might
6097 not notice unless it is glaringly wrong. You might as well not give us
6098 a chance to make a mistake.
6100 Even if the problem you experience is a fatal signal, you should still
6101 say so explicitly. Suppose something strange is going on, such as, your
6102 copy of @command{ld} is out of synch, or you have encountered a bug in the
6103 C library on your system. (This has happened!) Your copy might crash
6104 and ours would not. If you told us to expect a crash, then when ours
6105 fails to crash, we would know that the bug was not happening for us. If
6106 you had not told us to expect a crash, then we would not be able to draw
6107 any conclusion from our observations.
6110 If you wish to suggest changes to the @command{ld} source, send us context
6111 diffs, as generated by @code{diff} with the @samp{-u}, @samp{-c}, or
6112 @samp{-p} option. Always send diffs from the old file to the new file.
6113 If you even discuss something in the @command{ld} source, refer to it by
6114 context, not by line number.
6116 The line numbers in our development sources will not match those in your
6117 sources. Your line numbers would convey no useful information to us.
6120 Here are some things that are not necessary:
6124 A description of the envelope of the bug.
6126 Often people who encounter a bug spend a lot of time investigating
6127 which changes to the input file will make the bug go away and which
6128 changes will not affect it.
6130 This is often time consuming and not very useful, because the way we
6131 will find the bug is by running a single example under the debugger
6132 with breakpoints, not by pure deduction from a series of examples.
6133 We recommend that you save your time for something else.
6135 Of course, if you can find a simpler example to report @emph{instead}
6136 of the original one, that is a convenience for us. Errors in the
6137 output will be easier to spot, running under the debugger will take
6138 less time, and so on.
6140 However, simplification is not vital; if you do not want to do this,
6141 report the bug anyway and send us the entire test case you used.
6144 A patch for the bug.
6146 A patch for the bug does help us if it is a good one. But do not omit
6147 the necessary information, such as the test case, on the assumption that
6148 a patch is all we need. We might see problems with your patch and decide
6149 to fix the problem another way, or we might not understand it at all.
6151 Sometimes with a program as complicated as @command{ld} it is very hard to
6152 construct an example that will make the program follow a certain path
6153 through the code. If you do not send us the example, we will not be
6154 able to construct one, so we will not be able to verify that the bug is
6157 And if we cannot understand what bug you are trying to fix, or why your
6158 patch should be an improvement, we will not install it. A test case will
6159 help us to understand.
6162 A guess about what the bug is or what it depends on.
6164 Such guesses are usually wrong. Even we cannot guess right about such
6165 things without first using the debugger to find the facts.
6169 @appendix MRI Compatible Script Files
6170 @cindex MRI compatibility
6171 To aid users making the transition to @sc{gnu} @command{ld} from the MRI
6172 linker, @command{ld} can use MRI compatible linker scripts as an
6173 alternative to the more general-purpose linker scripting language
6174 described in @ref{Scripts}. MRI compatible linker scripts have a much
6175 simpler command set than the scripting language otherwise used with
6176 @command{ld}. @sc{gnu} @command{ld} supports the most commonly used MRI
6177 linker commands; these commands are described here.
6179 In general, MRI scripts aren't of much use with the @code{a.out} object
6180 file format, since it only has three sections and MRI scripts lack some
6181 features to make use of them.
6183 You can specify a file containing an MRI-compatible script using the
6184 @samp{-c} command-line option.
6186 Each command in an MRI-compatible script occupies its own line; each
6187 command line starts with the keyword that identifies the command (though
6188 blank lines are also allowed for punctuation). If a line of an
6189 MRI-compatible script begins with an unrecognized keyword, @command{ld}
6190 issues a warning message, but continues processing the script.
6192 Lines beginning with @samp{*} are comments.
6194 You can write these commands using all upper-case letters, or all
6195 lower case; for example, @samp{chip} is the same as @samp{CHIP}.
6196 The following list shows only the upper-case form of each command.
6199 @cindex @code{ABSOLUTE} (MRI)
6200 @item ABSOLUTE @var{secname}
6201 @itemx ABSOLUTE @var{secname}, @var{secname}, @dots{} @var{secname}
6202 Normally, @command{ld} includes in the output file all sections from all
6203 the input files. However, in an MRI-compatible script, you can use the
6204 @code{ABSOLUTE} command to restrict the sections that will be present in
6205 your output program. If the @code{ABSOLUTE} command is used at all in a
6206 script, then only the sections named explicitly in @code{ABSOLUTE}
6207 commands will appear in the linker output. You can still use other
6208 input sections (whatever you select on the command line, or using
6209 @code{LOAD}) to resolve addresses in the output file.
6211 @cindex @code{ALIAS} (MRI)
6212 @item ALIAS @var{out-secname}, @var{in-secname}
6213 Use this command to place the data from input section @var{in-secname}
6214 in a section called @var{out-secname} in the linker output file.
6216 @var{in-secname} may be an integer.
6218 @cindex @code{ALIGN} (MRI)
6219 @item ALIGN @var{secname} = @var{expression}
6220 Align the section called @var{secname} to @var{expression}. The
6221 @var{expression} should be a power of two.
6223 @cindex @code{BASE} (MRI)
6224 @item BASE @var{expression}
6225 Use the value of @var{expression} as the lowest address (other than
6226 absolute addresses) in the output file.
6228 @cindex @code{CHIP} (MRI)
6229 @item CHIP @var{expression}
6230 @itemx CHIP @var{expression}, @var{expression}
6231 This command does nothing; it is accepted only for compatibility.
6233 @cindex @code{END} (MRI)
6235 This command does nothing whatever; it's only accepted for compatibility.
6237 @cindex @code{FORMAT} (MRI)
6238 @item FORMAT @var{output-format}
6239 Similar to the @code{OUTPUT_FORMAT} command in the more general linker
6240 language, but restricted to one of these output formats:
6244 S-records, if @var{output-format} is @samp{S}
6247 IEEE, if @var{output-format} is @samp{IEEE}
6250 COFF (the @samp{coff-m68k} variant in BFD), if @var{output-format} is
6254 @cindex @code{LIST} (MRI)
6255 @item LIST @var{anything}@dots{}
6256 Print (to the standard output file) a link map, as produced by the
6257 @command{ld} command-line option @samp{-M}.
6259 The keyword @code{LIST} may be followed by anything on the
6260 same line, with no change in its effect.
6262 @cindex @code{LOAD} (MRI)
6263 @item LOAD @var{filename}
6264 @itemx LOAD @var{filename}, @var{filename}, @dots{} @var{filename}
6265 Include one or more object file @var{filename} in the link; this has the
6266 same effect as specifying @var{filename} directly on the @command{ld}
6269 @cindex @code{NAME} (MRI)
6270 @item NAME @var{output-name}
6271 @var{output-name} is the name for the program produced by @command{ld}; the
6272 MRI-compatible command @code{NAME} is equivalent to the command-line
6273 option @samp{-o} or the general script language command @code{OUTPUT}.
6275 @cindex @code{ORDER} (MRI)
6276 @item ORDER @var{secname}, @var{secname}, @dots{} @var{secname}
6277 @itemx ORDER @var{secname} @var{secname} @var{secname}
6278 Normally, @command{ld} orders the sections in its output file in the
6279 order in which they first appear in the input files. In an MRI-compatible
6280 script, you can override this ordering with the @code{ORDER} command. The
6281 sections you list with @code{ORDER} will appear first in your output
6282 file, in the order specified.
6284 @cindex @code{PUBLIC} (MRI)
6285 @item PUBLIC @var{name}=@var{expression}
6286 @itemx PUBLIC @var{name},@var{expression}
6287 @itemx PUBLIC @var{name} @var{expression}
6288 Supply a value (@var{expression}) for external symbol
6289 @var{name} used in the linker input files.
6291 @cindex @code{SECT} (MRI)
6292 @item SECT @var{secname}, @var{expression}
6293 @itemx SECT @var{secname}=@var{expression}
6294 @itemx SECT @var{secname} @var{expression}
6295 You can use any of these three forms of the @code{SECT} command to
6296 specify the start address (@var{expression}) for section @var{secname}.
6297 If you have more than one @code{SECT} statement for the same
6298 @var{secname}, only the @emph{first} sets the start address.
6309 % I think something like @colophon should be in texinfo. In the
6311 \long\def\colophon{\hbox to0pt{}\vfill
6312 \centerline{The body of this manual is set in}
6313 \centerline{\fontname\tenrm,}
6314 \centerline{with headings in {\bf\fontname\tenbf}}
6315 \centerline{and examples in {\tt\fontname\tentt}.}
6316 \centerline{{\it\fontname\tenit\/} and}
6317 \centerline{{\sl\fontname\tensl\/}}
6318 \centerline{are used for emphasis.}\vfill}
6320 % Blame: doc@cygnus.com, 28mar91.