4 @include configdoc.texi
5 @c (configdoc.texi is generated by the Makefile)
12 * Ld: (ld). The GNU linker.
18 This file documents the @sc{gnu} linker LD.
20 Copyright (C) 1991, 92, 93, 94, 1995 Free Software Foundation, Inc.
22 Permission is granted to make and distribute verbatim copies of
23 this manual provided the copyright notice and this permission notice
24 are preserved on all copies.
26 Permission is granted to copy and distribute modified versions of this
27 manual under the conditions for verbatim copying, provided also that
28 the entire resulting derived work is distributed under the terms of a
29 permission notice identical to this one.
31 Permission is granted to copy and distribute translations of this manual
32 into another language, under the above conditions for modified versions.
35 Permission is granted to process this file through Tex and print the
36 results, provided the printed document carries copying permission
37 notice identical to this one except for the removal of this paragraph
38 (this paragraph not being relevant to the printed manual).
44 @setchapternewpage odd
45 @settitle Using LD, the GNU linker
48 @subtitle The GNU linker
50 @subtitle @code{ld} version 2
51 @subtitle January 1994
52 @author Steve Chamberlain
53 @author Cygnus Support
58 \hfill Cygnus Support\par
59 \hfill steve\@cygnus.com, doc\@cygnus.com\par
60 \hfill {\it Using LD, the GNU linker}\par
61 \hfill Edited by Jeffrey Osier (jeffrey\@cygnus.com)\par
63 \global\parindent=0pt % Steve likes it this way.
66 @vskip 0pt plus 1filll
67 Copyright @copyright{} 1991, 92, 93, 94, 1995 Free Software Foundation, Inc.
69 Permission is granted to make and distribute verbatim copies of
70 this manual provided the copyright notice and this permission notice
71 are preserved on all copies.
73 Permission is granted to copy and distribute modified versions of this
74 manual under the conditions for verbatim copying, provided also that
75 the entire resulting derived work is distributed under the terms of a
76 permission notice identical to this one.
78 Permission is granted to copy and distribute translations of this manual
79 into another language, under the above conditions for modified versions.
82 @c FIXME: Talk about importance of *order* of args, cmds to linker!
87 This file documents the @sc{gnu} linker ld.
91 * Invocation:: Invocation
92 * Commands:: Command Language
94 * Machine Dependent:: Machine Dependent Features
98 * H8/300:: ld and the H8/300
101 * Hitachi:: ld and other Hitachi micros
104 * i960:: ld and the Intel 960 family
107 @ifclear SingleFormat
110 @c Following blank line required for remaining bug in makeinfo conds/menus
112 * MRI:: MRI Compatible Script Files
120 @cindex @sc{gnu} linker
121 @cindex what is this?
122 @code{ld} combines a number of object and archive files, relocates
123 their data and ties up symbol references. Usually the last step in
124 compiling a program is to run @code{ld}.
126 @code{ld} accepts Linker Command Language files written in
127 a superset of AT&T's Link Editor Command Language syntax,
128 to provide explicit and total control over the linking process.
130 @ifclear SingleFormat
131 This version of @code{ld} uses the general purpose BFD libraries
132 to operate on object files. This allows @code{ld} to read, combine, and
133 write object files in many different formats---for example, COFF or
134 @code{a.out}. Different formats may be linked together to produce any
135 available kind of object file. @xref{BFD}, for more information.
138 Aside from its flexibility, the @sc{gnu} linker is more helpful than other
139 linkers in providing diagnostic information. Many linkers abandon
140 execution immediately upon encountering an error; whenever possible,
141 @code{ld} continues executing, allowing you to identify other errors
142 (or, in some cases, to get an output file in spite of the error).
147 The @sc{gnu} linker @code{ld} is meant to cover a broad range of situations,
148 and to be as compatible as possible with other linkers. As a result,
149 you have many choices to control its behavior.
153 * Options:: Command Line Options
154 * Environment:: Environment Variables
158 @section Command Line Options
163 Here is a summary of the options you can use on the @code{ld} command
166 @c FIXME! -relax only avail h8/300, i960. Conditionals screwed in examples.
168 ld [ -o @var{output} ] @var{objfile}@dots{}
169 [ -A@var{architecture} ] [ -b @var{input-format} ]
170 [ -Bstatic ] [ -Bdynamic ] [ -Bsymbolic ]
171 [ -c @var{MRI-commandfile} ] [ -d | -dc | -dp ]
172 [ -defsym @var{symbol}=@var{expression} ]
173 [ -dynamic-linker @var{file} ] [ -embedded-relocs ] [ -export-dynamic ]
174 [ -e @var{entry} ] [ -F ] [ -F @var{format} ]
175 [ -format @var{input-format} ] [ -g ] [ -G @var{size} ]
176 [ -help ] [ -i ] [ -l@var{archive} ] [ -L@var{searchdir} ]
177 [ -M ] [ -Map @var{mapfile} ] [ -m @var{emulation} ]
178 [ -N | -n ] [ -noinhibit-exec ] [ -no-keep-memory ]
179 [ -oformat @var{output-format} ] [ -R @var{filename} ]
180 [ -relax ] [ -retain-symbols-file @var{filename} ]
181 [ -r | -Ur ] [ -rpath @var{dir} ] [-rpath-link @var{dir} ]
182 [ -S ] [ -s ] [ -soname @var{name} ] [ -shared ]
183 [ -sort-common ] [ -stats ] [ -T @var{commandfile} ]
184 [ -Ttext @var{org} ] [ -Tdata @var{org} ]
185 [ -Tbss @var{org} ] [ -t ] [ -traditional-format ]
186 [ -u @var{symbol}] [-V] [-v] [ -verbose] [ -version ]
187 [ -warn-common ] [ -warn-constructors] [ -warn-once ]
188 [ -y @var{symbol} ] [ -X ] [-x ]
189 [ -( [ archives ] -) ]
190 [ --start-group [ archives ] --end-group ]
191 [ -split-by-reloc @var{count} ] [ -split-by-file ]
195 This plethora of command-line options may seem intimidating, but in
196 actual practice few of them are used in any particular context.
197 @cindex standard Unix system
198 For instance, a frequent use of @code{ld} is to link standard Unix
199 object files on a standard, supported Unix system. On such a system, to
200 link a file @code{hello.o}:
203 ld -o @var{output} /lib/crt0.o hello.o -lc
206 This tells @code{ld} to produce a file called @var{output} as the
207 result of linking the file @code{/lib/crt0.o} with @code{hello.o} and
208 the library @code{libc.a}, which will come from the standard search
209 directories. (See the discussion of the @samp{-l} option below.)
211 The command-line options to @code{ld} may be specified in any order, and
212 may be repeated at will. Repeating most options with a
213 different argument will either have no further effect, or override prior
214 occurrences (those further to the left on the command line) of that
217 @ifclear SingleFormat
218 The exceptions---which may meaningfully be used more than once---are
219 @samp{-A}, @samp{-b} (or its synonym @samp{-format}), @samp{-defsym},
220 @samp{-L}, @samp{-l}, @samp{-R}, @samp{-u}, and @samp{-(} (or its
221 synonym @samp{--start-group})..
224 The exceptions---which may meaningfully be used more than once---are
225 @samp{-A}, @samp{-defsym}, @samp{-L}, @samp{-l}, @samp{-R}, @samp{-u},
226 and @samp{-(} (or its synonym @samp{--start-group}).
230 The list of object files to be linked together, shown as @var{objfile}@dots{},
231 may follow, precede, or be mixed in with command-line options, except that
232 an @var{objfile} argument may not be placed between an option and
235 Usually the linker is invoked with at least one object file, but you can
236 specify other forms of binary input files using @samp{-l}, @samp{-R},
237 and the script command language. If @emph{no} binary input files at all
238 are specified, the linker does not produce any output, and issues the
239 message @samp{No input files}.
241 If the linker can not recognize the format of an object file, it will
242 assume that it is a linker script. A script specified in this way
243 augments the main linker script used for the link (either the default
244 linker script or the one specified by using @samp{-T}). This feature
245 permits the linker to link against a file which appears to be an object
246 or an archive, but actually merely defines some symbol values, or uses
247 @code{INPUT} or @code{GROUP} to load other objects. @xref{Commands}.
249 For options whose names are a single letter,
250 option arguments must either follow the option letter without intervening
251 whitespace, or be given as separate arguments immediately following the
252 option that requires them.
254 For options whose names are multiple letters, either one dash or two can
255 precede the option name; for example, @samp{--oformat} and
256 @samp{-oformat} are equivalent. Arguments to multiple-letter options
257 must either be separated from the option name by an equals sign, or be
258 given as separate arguments immediately following the option that
259 requires them. For example, @samp{--oformat srec} and
260 @samp{--oformat=srec} are equivalent. Unique abbreviations of the names
261 of multiple-letter options are accepted.
265 @cindex architectures
267 @item -A@var{architecture}
268 In the current release of @code{ld}, this option is useful only for the
269 Intel 960 family of architectures. In that @code{ld} configuration, the
270 @var{architecture} argument identifies the particular architecture in
271 the 960 family, enabling some safeguards and modifying the
272 archive-library search path. @xref{i960,,@code{ld} and the Intel 960
273 family}, for details.
275 Future releases of @code{ld} may support similar functionality for
276 other architecture families.
279 @ifclear SingleFormat
280 @cindex binary input format
281 @kindex -b @var{format}
284 @item -b @var{input-format}
285 @code{ld} may be configured to support more than one kind of object
286 file. If your @code{ld} is configured this way, you can use the
287 @samp{-b} option to specify the binary format for input object files
288 that follow this option on the command line. Even when @code{ld} is
289 configured to support alternative object formats, you don't usually need
290 to specify this, as @code{ld} should be configured to expect as a
291 default input format the most usual format on each machine.
292 @var{input-format} is a text string, the name of a particular format
293 supported by the BFD libraries. (You can list the available binary
294 formats with @samp{objdump -i}.) @w{@samp{-format @var{input-format}}}
295 has the same effect, as does the script command @code{TARGET}.
298 You may want to use this option if you are linking files with an unusual
299 binary format. You can also use @samp{-b} to switch formats explicitly (when
300 linking object files of different formats), by including
301 @samp{-b @var{input-format}} before each group of object files in a
304 The default format is taken from the environment variable
309 You can also define the input
310 format from a script, using the command @code{TARGET}; see @ref{Option
316 Do not link against shared libraries. This is only meaningful on
317 platforms for which shared libraries are supported.
321 Link against dynamic libraries. This is only meaningful on platforms
322 for which shared libraries are supported. This option is normally the
323 default on such platforms.
327 When creating a shared library, bind references to global symbols to the
328 definition within the shared library, if any. Normally, it is possible
329 for a program linked against a shared library to override the definition
330 within the shared library. This option is only meaningful on ELF
331 platforms which support shared libraries.
333 @kindex -c @var{MRI-cmdfile}
334 @cindex compatibility, MRI
335 @item -c @var{MRI-commandfile}
336 For compatibility with linkers produced by MRI, @code{ld} accepts script
337 files written in an alternate, restricted command language, described in
338 @ref{MRI,,MRI Compatible Script Files}. Introduce MRI script files with
339 the option @samp{-c}; use the @samp{-T} option to run linker
340 scripts written in the general-purpose @code{ld} scripting language.
341 If @var{MRI-cmdfile} does not exist, @code{ld} looks for it in the directories
342 specified by any @samp{-L} options.
344 @cindex common allocation
351 These three options are equivalent; multiple forms are supported for
352 compatibility with other linkers. They
353 assign space to common symbols even if a relocatable output file is
354 specified (with @samp{-r}). The script command
355 @code{FORCE_COMMON_ALLOCATION} has the same effect. @xref{Option
358 @cindex symbols, from command line
359 @kindex -defsym @var{symbol}=@var{exp}
360 @item -defsym @var{symbol}=@var{expression}
361 Create a global symbol in the output file, containing the absolute
362 address given by @var{expression}. You may use this option as many
363 times as necessary to define multiple symbols in the command line. A
364 limited form of arithmetic is supported for the @var{expression} in this
365 context: you may give a hexadecimal constant or the name of an existing
366 symbol, or use @code{+} and @code{-} to add or subtract hexadecimal
367 constants or symbols. If you need more elaborate expressions, consider
368 using the linker command language from a script (@pxref{Assignment, ,
369 Assignment: Symbol Definitions}). @emph{Note:} there should be no
370 white space between @var{symbol}, the equals sign (``@key{=}''), and
374 @cindex dynamic linker, from command line
375 @kindex -dynamic-linker @var{file}
376 @item -dynamic-linker @var{file}
377 Set the name of the dynamic linker. This is only meaningful when
378 generating dynamically linked ELF executables. The default dynamic
379 linker is normally correct; don't use this unless you know what you are
383 @cindex MIPS embedded PIC code
384 @kindex -embedded-relocs
385 @item -embedded-relocs
386 This option is only meaningful when linking MIPS embedded PIC code,
387 generated by the -membedded-pic option to the @sc{gnu} compiler and
388 assembler. It causes the linker to create a table which may be used at
389 runtime to relocate any data which was statically initialized to pointer
390 values. See the code in testsuite/ld-empic for details.
392 @cindex entry point, from command line
393 @kindex -e @var{entry}
395 Use @var{entry} as the explicit symbol for beginning execution of your
396 program, rather than the default entry point. @xref{Entry Point}, for a
397 discussion of defaults and other ways of specifying the
400 @cindex dynamic symbol table
401 @kindex -export-dynamic
402 @item -export-dynamic
403 When creating an ELF file, add all symbols to the dynamic symbol table.
404 Normally, the dynamic symbol table contains only symbols which are used
405 by a dynamic object. This option is needed for some uses of
408 @ifclear SingleFormat
411 @itemx -F@var{format}
412 Ignored. Some older linkers used this option throughout a compilation
413 toolchain for specifying object-file format for both input and output
414 object files. The mechanisms @code{ld} uses for this purpose (the
415 @samp{-b} or @samp{-format} options for input files, @samp{-oformat}
416 option or the @code{TARGET} command in linker scripts for output files,
417 the @code{GNUTARGET} environment variable) are more flexible, but
418 @code{ld} accepts the @samp{-F} option for compatibility with scripts
419 written to call the old linker.
422 @item -format @var{input-format}
423 Synonym for @samp{-b @var{input-format}}.
428 Ignored. Provided for compatibility with other tools.
433 @itemx -G @var{value}
434 Set the maximum size of objects to be optimized using the GP register to
435 @var{size} under MIPS ECOFF. Ignored for other object file formats.
441 Print a summary of the command-line options on the standard output and exit.
444 @cindex incremental link
446 Perform an incremental link (same as option @samp{-r}).
448 @cindex archive files, from cmd line
449 @kindex -l@var{archive}
451 Add archive file @var{archive} to the list of files to link. This
452 option may be used any number of times. @code{ld} will search its
453 path-list for occurrences of @code{lib@var{ar}.a} for every @var{archive}
456 @cindex search directory, from cmd line
458 @item -L@var{searchdir}
459 @itemx -L @var{searchdir}
460 Add path @var{searchdir} to the list of paths that @code{ld} will search
461 for archive libraries and @code{ld} control scripts. You may use this
462 option any number of times. The directories are searched in the order
463 in which they are specified on the command line. Directories specified
464 on the command line are searched before the default directories. All
465 @code{-L} options apply to all @code{-l} options, regardless of the
466 order in which the options appear.
469 The default set of paths searched (without being specified with
470 @samp{-L}) depends on which emulation mode @code{ld} is using, and in
471 some cases also on how it was configured. @xref{Environment}.
474 The paths can also be specified in a link script with the
475 @code{SEARCH_DIR} command. Directories specified this way are searched
476 at the point in which the linker script appears in the command line.
481 Print (to the standard output) a link map---diagnostic information about
482 where symbols are mapped by @code{ld}, and information on global common
487 @item -Map @var{mapfile}
488 Print to the file @var{mapfile} a link map---diagnostic information
489 about where symbols are mapped by @code{ld}, and information on global
490 common storage allocation.
493 @kindex -m @var{emulation}
494 @item -m@var{emulation}
495 @itemx -m @var{emulation}
496 Emulate the @var{emulation} linker. You can list the available
497 emulations with the @samp{--verbose} or @samp{-V} options. The default
498 depends on how your @code{ld} was configured.
501 @cindex read/write from cmd line
504 Set the text and data sections to be readable and writable. Also, do
505 not page-align the data segment. If the output format supports Unix
506 style magic numbers, mark the output as @code{OMAGIC}.
509 @cindex read-only text
512 Set the text segment to be read only, and mark the output as
513 @code{NMAGIC} if possible.
515 @cindex output file after errors
516 @kindex -noinhibit-exec
517 @item -noinhibit-exec
518 Retain the executable output file whenever it is still usable.
519 Normally, the linker will not produce an output file if it encounters
520 errors during the link process; it exits without writing an output file
521 when it issues any error whatsoever.
524 @kindex -no-keep-memory
525 @item -no-keep-memory
526 @code{ld} normally optimizes for speed over memory usage by caching the
527 symbol tables of input files in memory. This option tells @code{ld} to
528 instead optimize for memory usage, by rereading the symbol tables as
529 necessary. This may be required if @code{ld} runs out of memory space
530 while linking a large executable.
532 @kindex -o @var{output}
533 @cindex naming the output file
534 @item -o @var{output}
535 Use @var{output} as the name for the program produced by @code{ld}; if this
536 option is not specified, the name @file{a.out} is used by default. The
537 script command @code{OUTPUT} can also specify the output file name.
539 @ifclear SingleFormat
541 @item -oformat @var{output-format}
542 @code{ld} may be configured to support more than one kind of object
543 file. If your @code{ld} is configured this way, you can use the
544 @samp{-oformat} option to specify the binary format for the output
545 object file. Even when @code{ld} is configured to support alternative
546 object formats, you don't usually need to specify this, as @code{ld}
547 should be configured to produce as a default output format the most
548 usual format on each machine. @var{output-format} is a text string, the
549 name of a particular format supported by the BFD libraries. (You can
550 list the available binary formats with @samp{objdump -i}.) The script
551 command @code{OUTPUT_FORMAT} can also specify the output format, but
552 this option overrides it. @xref{BFD}.
555 @kindex -R @var{file}
556 @cindex symbol-only input
557 @item -R @var{filename}
558 Read symbol names and their addresses from @var{filename}, but do not
559 relocate it or include it in the output. This allows your output file
560 to refer symbolically to absolute locations of memory defined in other
563 For compatibility with other ELF linkers, if the @code{-R} option is
564 followed by a directory name, rather than a file name, it is treated as
565 the @code{-rpath} option.
568 @cindex synthesizing linker
569 @cindex relaxing addressing modes
571 An option with machine dependent effects.
573 Currently this option is only supported on the H8/300 and the Intel 960.
576 @xref{H8/300,,@code{ld} and the H8/300}.
579 @xref{i960,, @code{ld} and the Intel 960 family}.
582 On some platforms, the @samp{-relax} option performs global optimizations that
583 become possible when the linker resolves addressing in the program, such
584 as relaxing address modes and synthesizing new instructions in the
588 On platforms where this is not supported, @samp{-relax} is accepted, but
592 @cindex retaining specified symbols
593 @cindex stripping all but some symbols
594 @cindex symbols, retaining selectively
595 @item -retain-symbols-file @var{filename}
596 Retain @emph{only} the symbols listed in the file @var{filename},
597 discarding all others. @var{filename} is simply a flat file, with one
598 symbol name per line. This option is especially useful in environments
602 where a large global symbol table is accumulated gradually, to conserve
605 @samp{-retain-symbols-file} does @emph{not} discard undefined symbols,
606 or symbols needed for relocations.
608 You may only specify @samp{-retain-symbols-file} once in the command
609 line. It overrides @samp{-s} and @samp{-S}.
612 @item -rpath @var{dir}
613 @cindex runtime library search path
615 Add a directory to the runtime library search path. This is used when
616 linking an ELF executable with shared objects. All @code{-rpath}
617 arguments are concatenated and passed to the runtime linker, which uses
618 them to locate shared objects at runtime. The @code{-rpath} option is
619 also used when locating shared objects which are needed by shared
620 objects explicitly included in the link; see the description of the
621 @code{-rpath-link} option. If @code{-rpath} is not used when linking an
622 ELF executable, the contents of the environment variable
623 @code{LD_RUN_PATH} will be used if it is defined.
625 The @code{-rpath} option may also be used on SunOS. By default, on
626 SunOS, the linker will form a runtime search patch out of all the
627 @code{-L} options it is given. If a @code{-rpath} option is used, the
628 runtime search path will be formed exclusively using the @code{-rpath}
629 options, ignoring the @code{-L} options. This can be useful when using
630 gcc, which adds many @code{-L} options which may be on NFS mounted
633 For compatibility with other ELF linkers, if the @code{-R} option is
634 followed by a directory name, rather than a file name, it is treated as
635 the @code{-rpath} option.
639 @cindex link-time runtime library search path
641 @item -rpath-link @var{DIR}
642 When using ELF or SunOS, one shared library may require another. This
643 happens when an @code{ld -shared} link includes a shared library as one
646 When the linker encounters such a dependency when doing a non-shared,
647 non-relocateable link, it will automatically try to locate the required
648 shared library and include it in the link, if it is not included
649 explicitly. In such a case, the @code{-rpath-link} option
650 specifies the first set of directories to search. The
651 @code{-rpath-link} option may specify a sequence of directory names
652 either by specifying a list of names separated by colons, or by
653 appearing multiple times.
655 The linker uses the following search paths to locate required shared
659 Any directories specified by @code{-rpath-link} options.
661 Any directories specified by @code{-rpath} options. The difference
662 between @code{-rpath} and @code{-rpath-link} is that directories
663 specified by @code{-rpath} options are included in the executable and
664 used at runtime, whereas the @code{-rpath-link} option is only effective
667 On an ELF system, if the @code{-rpath} and @code{rpath-link} options
668 were not used, search the contents of the environment variable
671 On SunOS, if the @code{-rpath} option was not used, search any
672 directories specified using @code{-L} options.
674 For a native linker, the contents of the environment variable
675 @code{LD_LIBRARY_PATH}.
677 The default directories, normally @file{/lib} and @file{/usr/lib}.
680 If the required shared library is not found, the linker will issue a
681 warning and continue with the link.
685 @cindex relocatable output
688 Generate relocatable output---i.e., generate an output file that can in
689 turn serve as input to @code{ld}. This is often called @dfn{partial
690 linking}. As a side effect, in environments that support standard Unix
691 magic numbers, this option also sets the output file's magic number to
694 If this option is not specified, an absolute file is produced. When
695 linking C++ programs, this option @emph{will not} resolve references to
696 constructors; to do that, use @samp{-Ur}.
698 This option does the same thing as @samp{-i}.
701 @cindex strip debugger symbols
703 Omit debugger symbol information (but not all symbols) from the output file.
706 @cindex strip all symbols
708 Omit all symbol information from the output file.
711 @cindex runtime library name
713 @item -soname @var{name}
714 When creating an ELF shared object, set the internal DT_SONAME field to
715 the specified name. When an executable is linked with a shared object
716 which has a DT_SONAME field, then when the executable is run the dynamic
717 linker will attempt to load the shared object specified by the DT_SONAME
718 field rather than the using the file name given to the linker.
722 @cindex shared libraries
724 Create a shared library. This is currently only supported on ELF and
725 SunOS platforms. On SunOS, the linker will automatically create a
726 shared library if the @code{-e} option is not used and there are
727 undefined symbols in the link.
731 Normally, when @code{ld} places the global common symbols in the
732 appropriate output sections, it sorts them by size. First come all the
733 one byte symbols, then all the two bytes, then all the four bytes, and
734 then everything else. This is to prevent gaps between symbols due to
735 alignment constraints. This option disables that sorting.
738 @item -split-by-reloc @var{count}
739 Trys to creates extra sections in the output file so that no single output section
740 in the file contains more than @var{count} relocations. This
741 is useful when generating huge relocatable for downloading into
742 certain real time kernels with the COFF object file format; since
743 COFF cannot represent more than 65535 relocations in a single section.
744 Note that this will fail to work with object file formats which do not
745 support arbitrary sections. The linker will not split up individual input
746 sections for redistribution, so if a single input section contains
747 more than @var{count} relocations one output section will contain that
752 Similar to -split-by-reloc but creates a new output section for each
756 Compute and display statistics about the operation of the linker,
757 such as execution time and memory usage.
759 @kindex -Tbss @var{org}
760 @kindex -Tdata @var{org}
761 @kindex -Ttext @var{org}
762 @cindex segment origins, cmd line
763 @item -Tbss @var{org}
764 @itemx -Tdata @var{org}
765 @itemx -Ttext @var{org}
766 Use @var{org} as the starting address for---respectively---the
767 @code{bss}, @code{data}, or the @code{text} segment of the output file.
768 @var{org} must be a single hexadecimal integer;
769 for compatibility with other linkers, you may omit the leading
770 @samp{0x} usually associated with hexadecimal values.
772 @kindex -T @var{script}
774 @item -T @var{commandfile}
775 @itemx -T@var{commandfile}
776 Read link commands from the file @var{commandfile}. These commands
777 replace @code{ld}'s default link script (rather than adding
778 to it), so @var{commandfile} must specify everything necessary to describe
779 the target format. @xref{Commands}. If @var{commandfile} does not
780 exist, @code{ld} looks for it in the directories specified by any
781 preceding @samp{-L} options. Multiple @samp{-T} options accumulate.
785 @cindex input files, displaying
787 Print the names of the input files as @code{ld} processes them.
789 @kindex -traditional-format
790 @cindex traditional format
791 @item -traditional-format
792 For some targets, the output of @code{ld} is different in some ways from
793 the output of some existing linker. This switch requests @code{ld} to
794 use the traditional format instead.
797 For example, on SunOS, @code{ld} combines duplicate entries in the
798 symbol string table. This can reduce the size of an output file with
799 full debugging information by over 30 percent. Unfortunately, the SunOS
800 @code{dbx} program can not read the resulting program (@code{gdb} has no
801 trouble). The @samp{-traditional-format} switch tells @code{ld} to not
802 combine duplicate entries.
804 @kindex -u @var{symbol}
805 @cindex undefined symbol
806 @item -u @var{symbol}
807 Force @var{symbol} to be entered in the output file as an undefined symbol.
808 Doing this may, for example, trigger linking of additional modules from
809 standard libraries. @samp{-u} may be repeated with different option
810 arguments to enter additional undefined symbols.
811 @c Nice idea, but no such command: This option is equivalent
812 @c to the @code{EXTERN} linker command.
817 For anything other than C++ programs, this option is equivalent to
818 @samp{-r}: it generates relocatable output---i.e., an output file that can in
819 turn serve as input to @code{ld}. When linking C++ programs, @samp{-Ur}
820 @emph{does} resolve references to constructors, unlike @samp{-r}.
821 It does not work to use @samp{-Ur} on files that were themselves linked
822 with @samp{-Ur}; once the constructor table has been built, it cannot
823 be added to. Use @samp{-Ur} only for the last partial link, and
824 @samp{-r} for the others.
829 Display the version number for @code{ld} and list the linker emulations
830 supported. Display which input files can and cannot be opened.
837 Display the version number for @code{ld}. The @code{-V} option also
838 lists the supported emulations.
842 Display the version number for @code{ld} and exit.
845 @cindex warnings, on combining symbols
846 @cindex combining symbols, warnings on
848 Warn when a common symbol is combined with another common symbol or with
849 a symbol definition. Unix linkers allow this somewhat sloppy practice,
850 but linkers on some other operating systems do not. This option allows
851 you to find potential problems from combining global symbols.
852 Unfortunately, some C libraries use this practice, so you may get some
853 warnings about symbols in the libraries as well as in your programs.
855 There are three kinds of global symbols, illustrated here by C examples:
859 A definition, which goes in the initialized data section of the output
863 An undefined reference, which does not allocate space.
864 There must be either a definition or a common symbol for the
868 A common symbol. If there are only (one or more) common symbols for a
869 variable, it goes in the uninitialized data area of the output file.
870 The linker merges multiple common symbols for the same variable into a
871 single symbol. If they are of different sizes, it picks the largest
872 size. The linker turns a common symbol into a declaration, if there is
873 a definition of the same variable.
876 The @samp{-warn-common} option can produce five kinds of warnings. Each
877 warning consists of a pair of lines: the first describes the symbol just
878 encountered, and the second describes the previous symbol encountered
879 with the same name. One or both of the two symbols will be a common
884 Turning a common symbol into a reference, because there is already a
885 definition for the symbol.
887 @var{file}(@var{section}): warning: common of `@var{symbol}'
888 overridden by definition
889 @var{file}(@var{section}): warning: defined here
893 Turning a common symbol into a reference, because a later definition for
894 the symbol is encountered. This is the same as the previous case,
895 except that the symbols are encountered in a different order.
897 @var{file}(@var{section}): warning: definition of `@var{symbol}'
899 @var{file}(@var{section}): warning: common is here
903 Merging a common symbol with a previous same-sized common symbol.
905 @var{file}(@var{section}): warning: multiple common
907 @var{file}(@var{section}): warning: previous common is here
911 Merging a common symbol with a previous larger common symbol.
913 @var{file}(@var{section}): warning: common of `@var{symbol}'
914 overridden by larger common
915 @var{file}(@var{section}): warning: larger common is here
919 Merging a common symbol with a previous smaller common symbol. This is
920 the same as the previous case, except that the symbols are
921 encountered in a different order.
923 @var{file}(@var{section}): warning: common of `@var{symbol}'
924 overriding smaller common
925 @var{file}(@var{section}): warning: smaller common is here
929 @kindex -warn-constructors
930 @item -warn-constructors
931 Warn if any global constructors are used. This is only useful for a few
932 object file formats. For formats like COFF or ELF, the linker can not
933 detect the use of global constructors.
936 @cindex warnings, on undefined symbols
937 @cindex undefined symbols, warnings on
939 Only warn once for each undefined symbol, rather than once per module
942 @kindex --whole-archive
943 @cindex including an entire archive
944 For each archive mentioned on the command line, include every object
945 file in the archive in the link, rather than searching the archive for
946 the required object files. This is normally used to turn an archive
947 file into a shared library, forcing every object to be included in the
948 resulting shared library.
951 @cindex local symbols, deleting
952 @cindex L, deleting symbols beginning
954 Delete all temporary local symbols. For most targets, this is all local
955 symbols whose names begin with @samp{L}.
958 @cindex deleting local symbols
960 Delete all local symbols.
962 @kindex -y @var{symbol}
963 @cindex symbol tracing
964 @item -y @var{symbol}
965 Print the name of each linked file in which @var{symbol} appears. This
966 option may be given any number of times. On many systems it is necessary
967 to prepend an underscore.
969 This option is useful when you have an undefined symbol in your link but
970 don't know where the reference is coming from.
973 @cindex groups of archives
974 @item -( @var{archives} -)
975 @itemx --start-group @var{archives} --end-group
976 The @var{archives} should be a list of archive files. They may be
977 either explicit file names, or @samp{-l} options.
979 The specified archives are searched repeatedly until no new undefined
980 references are created. Normally, an archive is searched only once in
981 the order that it is specified on the command line. If a symbol in that
982 archive is needed to resolve an undefined symbol referred to by an
983 object in an archive that appears later on the command line, the linker
984 would not be able to resolve that reference. By grouping the archives,
985 they all be searched repeatedly until all possible references are
988 Using this option has a significant performance cost. It is best to use
989 it only when there are unavoidable circular references between two or
995 @section Environment Variables
997 You can change the behavior of @code{ld} with the environment
998 variable @code{GNUTARGET}.
1001 @cindex default input format
1002 @code{GNUTARGET} determines the input-file object format if you don't
1003 use @samp{-b} (or its synonym @samp{-format}). Its value should be one
1004 of the BFD names for an input format (@pxref{BFD}). If there is no
1005 @code{GNUTARGET} in the environment, @code{ld} uses the natural format
1006 of the target. If @code{GNUTARGET} is set to @code{default} then BFD attempts to discover the
1007 input format by examining binary input files; this method often
1008 succeeds, but there are potential ambiguities, since there is no method
1009 of ensuring that the magic number used to specify object-file formats is
1010 unique. However, the configuration procedure for BFD on each system
1011 places the conventional format for that system first in the search-list,
1012 so ambiguities are resolved in favor of convention.
1016 @chapter Command Language
1018 @cindex command files
1019 The command language provides explicit control over the link process,
1020 allowing complete specification of the mapping between the linker's
1021 input files and its output. It controls:
1030 addresses of sections
1032 placement of common blocks
1035 You may supply a command file (also known as a link script) to the
1036 linker either explicitly through the @samp{-T} option, or implicitly as
1037 an ordinary file. If the linker opens a file which it cannot recognize
1038 as a supported object or archive format, it reports an error.
1041 * Scripts:: Linker Scripts
1042 * Expressions:: Expressions
1043 * MEMORY:: MEMORY Command
1044 * SECTIONS:: SECTIONS Command
1045 * PHDRS:: PHDRS Command
1046 * Entry Point:: The Entry Point
1047 * Option Commands:: Option Commands
1051 @section Linker Scripts
1052 The @code{ld} command language is a collection of statements; some are
1053 simple keywords setting a particular option, some are used to select and
1054 group input files or name output files; and two statement
1055 types have a fundamental and pervasive impact on the linking process.
1057 @cindex fundamental script commands
1058 @cindex commands, fundamental
1059 @cindex output file layout
1060 @cindex layout of output file
1061 The most fundamental command of the @code{ld} command language is the
1062 @code{SECTIONS} command (@pxref{SECTIONS}). Every meaningful command
1063 script must have a @code{SECTIONS} command: it specifies a
1064 ``picture'' of the output file's layout, in varying degrees of detail.
1065 No other command is required in all cases.
1067 The @code{MEMORY} command complements @code{SECTIONS} by describing the
1068 available memory in the target architecture. This command is optional;
1069 if you don't use a @code{MEMORY} command, @code{ld} assumes sufficient
1070 memory is available in a contiguous block for all output.
1074 You may include comments in linker scripts just as in C: delimited
1075 by @samp{/*} and @samp{*/}. As in C, comments are syntactically
1076 equivalent to whitespace.
1079 @section Expressions
1080 @cindex expression syntax
1082 Many useful commands involve arithmetic expressions. The syntax for
1083 expressions in the command language is identical to that of C
1084 expressions, with the following features:
1087 All expressions evaluated as integers and
1088 are of ``long'' or ``unsigned long'' type.
1090 All constants are integers.
1092 All of the C arithmetic operators are provided.
1094 You may reference, define, and create global variables.
1096 You may call special purpose built-in functions.
1100 * Integers:: Integers
1101 * Symbols:: Symbol Names
1102 * Location Counter:: The Location Counter
1103 * Operators:: Operators
1104 * Evaluation:: Evaluation
1105 * Assignment:: Assignment: Defining Symbols
1106 * Arithmetic Functions:: Built-In Functions
1110 @subsection Integers
1111 @cindex integer notation
1112 @cindex octal integers
1113 An octal integer is @samp{0} followed by zero or more of the octal
1114 digits (@samp{01234567}).
1116 _as_octal = 0157255;
1119 @cindex decimal integers
1120 A decimal integer starts with a non-zero digit followed by zero or
1121 more digits (@samp{0123456789}).
1123 _as_decimal = 57005;
1126 @cindex hexadecimal integers
1128 A hexadecimal integer is @samp{0x} or @samp{0X} followed by one or
1129 more hexadecimal digits chosen from @samp{0123456789abcdefABCDEF}.
1134 @cindex negative integers
1135 To write a negative integer, use
1136 the prefix operator @samp{-}; @pxref{Operators}.
1141 @cindex scaled integers
1142 @cindex K and M integer suffixes
1143 @cindex M and K integer suffixes
1144 @cindex suffixes for integers
1145 @cindex integer suffixes
1146 Additionally the suffixes @code{K} and @code{M} may be used to scale a
1150 @c END TEXI2ROFF-KILL
1151 @code{1024} or @code{1024*1024}
1155 ${\rm 1024}$ or ${\rm 1024}^2$
1157 @c END TEXI2ROFF-KILL
1158 respectively. For example, the following all refer to the same quantity:
1167 @subsection Symbol Names
1168 @cindex symbol names
1170 @cindex quoted symbol names
1172 Unless quoted, symbol names start with a letter, underscore, or point
1173 and may include any letters, underscores, digits, points,
1174 and hyphens. Unquoted symbol names must not conflict with any
1175 keywords. You can specify a symbol which contains odd characters or has
1176 the same name as a keyword, by surrounding the symbol name in double quotes:
1179 "with a space" = "also with a space" + 10;
1182 Since symbols can contain many non-alphabetic characters, it is safest
1183 to delimit symbols with spaces. For example, @samp{A-B} is one symbol,
1184 whereas @samp{A - B} is an expression involving subtraction.
1186 @node Location Counter
1187 @subsection The Location Counter
1190 @cindex location counter
1191 @cindex current output location
1192 The special linker variable @dfn{dot} @samp{.} always contains the
1193 current output location counter. Since the @code{.} always refers to
1194 a location in an output section, it must always appear in an
1195 expression within a @code{SECTIONS} command. The @code{.} symbol
1196 may appear anywhere that an ordinary symbol is allowed in an
1197 expression, but its assignments have a side effect. Assigning a value
1198 to the @code{.} symbol will cause the location counter to be moved.
1200 This may be used to create holes in the output section. The location
1201 counter may never be moved backwards.
1216 In the previous example, @code{file1} is located at the beginning of the
1217 output section, then there is a 1000 byte gap. Then @code{file2}
1218 appears, also with a 1000 byte gap following before @code{file3} is
1219 loaded. The notation @samp{= 0x1234} specifies what data to write in
1220 the gaps (@pxref{Section Options}).
1228 @subsection Operators
1229 @cindex Operators for arithmetic
1230 @cindex arithmetic operators
1231 @cindex precedence in expressions
1232 The linker recognizes the standard C set of arithmetic operators, with
1233 the standard bindings and precedence levels:
1236 @c END TEXI2ROFF-KILL
1238 precedence associativity Operators Notes
1244 5 left == != > < <= >=
1250 11 right &= += -= *= /= (2)
1254 (1) Prefix operators
1255 (2) @xref{Assignment}
1259 \vskip \baselineskip
1260 %"lispnarrowing" is the extra indent used generally for @smallexample
1261 \hskip\lispnarrowing\vbox{\offinterlineskip
1264 {\vrule#&\strut\hfil\ #\ \hfil&\vrule#&\strut\hfil\ #\ \hfil&\vrule#&\strut\hfil\ {\tt #}\ \hfil&\vrule#\cr
1265 height2pt&\omit&&\omit&&\omit&\cr
1266 &Precedence&& Associativity &&{\rm Operators}&\cr
1267 height2pt&\omit&&\omit&&\omit&\cr
1269 height2pt&\omit&&\omit&&\omit&\cr
1271 % '176 is tilde, '~' in tt font
1272 &1&&left&&\qquad- \char'176\ !\qquad\dag&\cr
1273 &2&&left&&* / \%&\cr
1276 &5&&left&&== != > < <= >=&\cr
1279 &8&&left&&{\&\&}&\cr
1282 &11&&right&&\qquad\&= += -= *= /=\qquad\ddag&\cr
1284 height2pt&\omit&&\omit&&\omit&\cr}
1289 @obeylines@parskip=0pt@parindent=0pt
1290 @dag@quad Prefix operators.
1291 @ddag@quad @xref{Assignment}.
1294 @c END TEXI2ROFF-KILL
1297 @subsection Evaluation
1299 @cindex lazy evaluation
1300 @cindex expression evaluation order
1301 The linker uses ``lazy evaluation'' for expressions; it only calculates
1302 an expression when absolutely necessary. The linker needs the value of
1303 the start address, and the lengths of memory regions, in order to do any
1304 linking at all; these values are computed as soon as possible when the
1305 linker reads in the command file. However, other values (such as symbol
1306 values) are not known or needed until after storage allocation. Such
1307 values are evaluated later, when other information (such as the sizes of
1308 output sections) is available for use in the symbol assignment
1312 @subsection Assignment: Defining Symbols
1313 @cindex assignment in scripts
1314 @cindex symbol definition, scripts
1315 @cindex variables, defining
1316 You may create global symbols, and assign values (addresses) to global
1317 symbols, using any of the C assignment operators:
1320 @item @var{symbol} = @var{expression} ;
1321 @itemx @var{symbol} &= @var{expression} ;
1322 @itemx @var{symbol} += @var{expression} ;
1323 @itemx @var{symbol} -= @var{expression} ;
1324 @itemx @var{symbol} *= @var{expression} ;
1325 @itemx @var{symbol} /= @var{expression} ;
1328 Two things distinguish assignment from other operators in @code{ld}
1332 Assignment may only be used at the root of an expression;
1333 @samp{a=b+3;} is allowed, but @samp{a+b=3;} is an error.
1338 You must place a trailing semicolon (``@key{;}'') at the end of an
1339 assignment statement.
1342 Assignment statements may appear:
1345 as commands in their own right in an @code{ld} script; or
1347 as independent statements within a @code{SECTIONS} command; or
1349 as part of the contents of a section definition in a
1350 @code{SECTIONS} command.
1353 The first two cases are equivalent in effect---both define a symbol with
1354 an absolute address. The last case defines a symbol whose address is
1355 relative to a particular section (@pxref{SECTIONS}).
1357 @cindex absolute and relocatable symbols
1358 @cindex relocatable and absolute symbols
1359 @cindex symbols, relocatable and absolute
1360 When a linker expression is evaluated and assigned to a variable, it is
1361 given either an absolute or a relocatable type. An absolute expression
1362 type is one in which the symbol contains the value that it will have in
1363 the output file; a relocatable expression type is one in which the
1364 value is expressed as a fixed offset from the base of a section.
1366 The type of the expression is controlled by its position in the script
1367 file. A symbol assigned within a section definition is created relative
1368 to the base of the section; a symbol assigned in any other place is
1369 created as an absolute symbol. Since a symbol created within a
1370 section definition is relative to the base of the section, it
1371 will remain relocatable if relocatable output is requested. A symbol
1372 may be created with an absolute value even when assigned to within a
1373 section definition by using the absolute assignment function
1374 @code{ABSOLUTE}. For example, to create an absolute symbol whose address
1375 is the last byte of an output section named @code{.data}:
1381 _edata = ABSOLUTE(.) ;
1386 The linker tries to put off the evaluation of an assignment until all
1387 the terms in the source expression are known (@pxref{Evaluation}). For
1388 instance, the sizes of sections cannot be known until after allocation,
1389 so assignments dependent upon these are not performed until after
1390 allocation. Some expressions, such as those depending upon the location
1391 counter @dfn{dot}, @samp{.} must be evaluated during allocation. If the
1392 result of an expression is required, but the value is not available,
1393 then an error results. For example, a script like the following
1396 text 9+this_isnt_constant :
1401 @kindex Non constant expression
1403 will cause the error message ``@code{Non constant expression for initial
1407 In some cases, it is desirable for a linker script to define a symbol
1408 only if it is referenced, and only if it is not defined by any object
1409 included in the link. For example, traditional linkers defined the
1410 symbol @samp{etext}. However, ANSI C requires that the user be able to
1411 use @samp{etext} as a function name without encountering an error.
1412 The @code{PROVIDE} keyword may be used to define a symbol, such as
1413 @samp{etext}, only if it is referenced but not defined. The syntax is
1414 @code{PROVIDE(@var{symbol} = @var{expression})}.
1416 @node Arithmetic Functions
1417 @subsection Arithmetic Functions
1418 @cindex functions in expression language
1419 The command language includes a number of built-in
1420 functions for use in link script expressions.
1422 @kindex ABSOLUTE(@var{exp})
1423 @cindex expression, absolute
1424 @item ABSOLUTE(@var{exp})
1425 Return the absolute (non-relocatable, as opposed to non-negative) value
1426 of the expression @var{exp}. Primarily useful to assign an absolute
1427 value to a symbol within a section definition, where symbol values are
1428 normally section-relative.
1430 @kindex ADDR(@var{section})
1431 @cindex section address
1432 @item ADDR(@var{section})
1433 Return the absolute address of the named @var{section}. Your script must
1434 previously have defined the location of that section. In the following
1435 example, @code{symbol_1} and @code{symbol_2} are assigned identical
1442 start_of_output_1 = ABSOLUTE(.);
1447 symbol_1 = ADDR(.output1);
1448 symbol_2 = start_of_output_1;
1454 @kindex ALIGN(@var{exp})
1455 @cindex rounding up location counter
1456 @item ALIGN(@var{exp})
1457 Return the result of the current location counter (@code{.}) aligned to
1458 the next @var{exp} boundary. @var{exp} must be an expression whose
1459 value is a power of two. This is equivalent to
1461 (. + @var{exp} - 1) & ~(@var{exp} - 1)
1464 @code{ALIGN} doesn't change the value of the location counter---it just
1465 does arithmetic on it. As an example, to align the output @code{.data}
1466 section to the next @code{0x2000} byte boundary after the preceding
1467 section and to set a variable within the section to the next
1468 @code{0x8000} boundary after the input sections:
1472 .data ALIGN(0x2000): @{
1474 variable = ALIGN(0x8000);
1480 The first use of @code{ALIGN} in this example specifies the location of
1481 a section because it is used as the optional @var{start} attribute of a
1482 section definition (@pxref{Section Options}). The second use simply
1483 defines the value of a variable.
1485 The built-in @code{NEXT} is closely related to @code{ALIGN}.
1487 @kindex DEFINED(@var{symbol})
1488 @cindex symbol defaults
1489 @item DEFINED(@var{symbol})
1490 Return 1 if @var{symbol} is in the linker global symbol table and is
1491 defined, otherwise return 0. You can use this function to provide default
1492 values for symbols. For example, the following command-file fragment shows how
1493 to set a global symbol @code{begin} to the first location in the
1494 @code{.text} section---but if a symbol called @code{begin} already
1495 existed, its value is preserved:
1501 begin = DEFINED(begin) ? begin : . ;
1508 @kindex NEXT(@var{exp})
1509 @cindex unallocated address, next
1510 @item NEXT(@var{exp})
1511 Return the next unallocated address that is a multiple of @var{exp}.
1512 This function is closely related to @code{ALIGN(@var{exp})}; unless you
1513 use the @code{MEMORY} command to define discontinuous memory for the
1514 output file, the two functions are equivalent.
1516 @kindex SIZEOF(@var{section})
1517 @cindex section size
1518 @item SIZEOF(@var{section})
1519 Return the size in bytes of the named @var{section}, if that section has
1520 been allocated. In the following example, @code{symbol_1} and
1521 @code{symbol_2} are assigned identical values:
1522 @c What does it return if the section hasn't been allocated? 0?
1531 symbol_1 = .end - .start ;
1532 symbol_2 = SIZEOF(.output);
1537 @kindex SIZEOF_HEADERS
1539 @kindex sizeof_headers
1540 @item SIZEOF_HEADERS
1541 @itemx sizeof_headers
1542 Return the size in bytes of the output file's headers. You can use this number
1543 as the start address of the first section, if you choose, to facilitate
1549 @section Memory Layout
1551 @cindex regions of memory
1552 @cindex discontinuous memory
1553 @cindex allocating memory
1554 The linker's default configuration permits allocation of all available memory.
1555 You can override this configuration by using the @code{MEMORY} command. The
1556 @code{MEMORY} command describes the location and size of blocks of
1557 memory in the target. By using it carefully, you can describe which
1558 memory regions may be used by the linker, and which memory regions it
1559 must avoid. The linker does not shuffle sections to fit into the
1560 available regions, but does move the requested sections into the correct
1561 regions and issue errors when the regions become too full.
1563 A command file may contain at most one use of the @code{MEMORY}
1564 command; however, you can define as many blocks of memory within it as
1565 you wish. The syntax is:
1571 @var{name} (@var{attr}) : ORIGIN = @var{origin}, LENGTH = @var{len}
1577 @cindex naming memory regions
1579 is a name used internally by the linker to refer to the region. Any
1580 symbol name may be used. The region names are stored in a separate
1581 name space, and will not conflict with symbols, file names or section
1582 names. Use distinct names to specify multiple regions.
1584 @cindex memory region attributes
1586 is an optional list of attributes, permitted for compatibility with the
1587 AT&T linker but not used by @code{ld} beyond checking that the
1588 attribute list is valid. Valid attribute lists must be made up of the
1589 characters ``@code{LIRWX}''. If you omit the attribute list, you may
1590 omit the parentheses around it as well.
1596 is the start address of the region in physical memory. It is
1597 an expression that must evaluate to a constant before
1598 memory allocation is performed. The keyword @code{ORIGIN} may be
1599 abbreviated to @code{org} or @code{o} (but not, for example, @samp{ORG}).
1605 is the size in bytes of the region (an expression).
1606 The keyword @code{LENGTH} may be abbreviated to @code{len} or @code{l}.
1609 For example, to specify that memory has two regions available for
1610 allocation---one starting at 0 for 256 kilobytes, and the other
1611 starting at @code{0x40000000} for four megabytes:
1617 rom : ORIGIN = 0, LENGTH = 256K
1618 ram : org = 0x40000000, l = 4M
1623 Once you have defined a region of memory named @var{mem}, you can direct
1624 specific output sections there by using a command ending in
1625 @samp{>@var{mem}} within the @code{SECTIONS} command (@pxref{Section
1626 Options}). If the combined output sections directed to a region are too
1627 big for the region, the linker will issue an error message.
1630 @section Specifying Output Sections
1633 The @code{SECTIONS} command controls exactly where input sections are
1634 placed into output sections, their order in the output file, and to
1635 which output sections they are allocated.
1637 You may use at most one @code{SECTIONS} command in a script file,
1638 but you can have as many statements within it as you wish. Statements
1639 within the @code{SECTIONS} command can do one of three things:
1643 define the entry point;
1646 assign a value to a symbol;
1649 describe the placement of a named output section, and which input
1650 sections go into it.
1653 You can also use the first two operations---defining the entry point and
1654 defining symbols---outside the @code{SECTIONS} command: @pxref{Entry
1655 Point}, and @pxref{Assignment}. They are permitted here as well for
1656 your convenience in reading the script, so that symbols and the entry
1657 point can be defined at meaningful points in your output-file layout.
1659 If you do not use a @code{SECTIONS} command, the linker places each input
1660 section into an identically named output section in the order that the
1661 sections are first encountered in the input files. If all input sections
1662 are present in the first file, for example, the order of sections in the
1663 output file will match the order in the first input file.
1666 * Section Definition:: Section Definitions
1667 * Section Placement:: Section Placement
1668 * Section Data Expressions:: Section Data Expressions
1669 * Section Options:: Optional Section Attributes
1672 @node Section Definition
1673 @subsection Section Definitions
1674 @cindex section definition
1675 The most frequently used statement in the @code{SECTIONS} command is
1676 the @dfn{section definition}, which specifies the
1677 properties of an output section: its location, alignment, contents,
1678 fill pattern, and target memory region. Most of
1679 these specifications are optional; the simplest form of a section
1688 @cindex naming output sections
1690 @var{secname} is the name of the output section, and @var{contents} a
1691 specification of what goes there---for example, a list of input files or
1692 sections of input files (@pxref{Section Placement}). As you might
1693 assume, the whitespace shown is optional. You do need the colon
1694 @samp{:} and the braces @samp{@{@}}, however.
1696 @var{secname} must meet the constraints of your output format. In
1697 formats which only support a limited number of sections, such as
1698 @code{a.out}, the name must be one of the names supported by the format
1699 (@code{a.out}, for example, allows only @code{.text}, @code{.data} or
1700 @code{.bss}). If the output format supports any number of sections, but
1701 with numbers and not names (as is the case for Oasys), the name should be
1702 supplied as a quoted numeric string. A section name may consist of any
1703 sequence of characters, but any name which does not conform to the standard
1704 @code{ld} symbol name syntax must be quoted.
1705 @xref{Symbols, , Symbol Names}.
1707 The linker will not create output sections which do not have any
1708 contents. This is for convenience when referring to input sections that
1709 may or may not exist. For example,
1713 will only create a @samp{.foo} section in the output file if there is a
1714 @samp{.foo} section in at least one input file.
1716 @node Section Placement
1717 @subsection Section Placement
1719 @cindex contents of a section
1720 In a section definition, you can specify the contents of an output
1721 section by listing particular input files, by listing particular
1722 input-file sections, or by a combination of the two. You can also place
1723 arbitrary data in the section, and define symbols relative to the
1724 beginning of the section.
1726 The @var{contents} of a section definition may include any of the
1727 following kinds of statement. You can include as many of these as you
1728 like in a single section definition, separated from one another by
1732 @kindex @var{filename}
1733 @cindex input files, section defn
1734 @cindex files, including in output sections
1735 @item @var{filename}
1736 You may simply name a particular input file to be placed in the current
1737 output section; @emph{all} sections from that file are placed in the
1738 current section definition. If the file name has already been mentioned
1739 in another section definition, with an explicit section name list, then
1740 only those sections which have not yet been allocated are used.
1742 To specify a list of particular files by name:
1744 .data : @{ afile.o bfile.o cfile.o @}
1747 The example also illustrates that multiple statements can be included in
1748 the contents of a section definition, since each file name is a separate
1751 @kindex @var{filename}(@var{section})
1752 @cindex files and sections, section defn
1753 @item @var{filename}( @var{section} )
1754 @itemx @var{filename}( @var{section}, @var{section}, @dots{} )
1755 @itemx @var{filename}( @var{section} @var{section} @dots{} )
1756 You can name one or more sections from your input files, for
1757 insertion in the current output section. If you wish to specify a list
1758 of input-file sections inside the parentheses, you may separate the
1759 section names by either commas or whitespace.
1761 @cindex input sections to output section
1762 @kindex *(@var{section})
1763 @item * (@var{section})
1764 @itemx * (@var{section}, @var{section}, @dots{})
1765 @itemx * (@var{section} @var{section} @dots{})
1766 Instead of explicitly naming particular input files in a link control
1767 script, you can refer to @emph{all} files from the @code{ld} command
1768 line: use @samp{*} instead of a particular file name before the
1769 parenthesized input-file section list.
1771 If you have already explicitly included some files by name, @samp{*}
1772 refers to all @emph{remaining} files---those whose places in the output
1773 file have not yet been defined.
1775 For example, to copy sections @code{1} through @code{4} from an Oasys file
1776 into the @code{.text} section of an @code{a.out} file, and sections @code{13}
1777 and @code{14} into the @code{.data} section:
1792 @cindex @code{[@var{section}@dots{}]}, not supported
1793 @samp{[ @var{section} @dots{} ]} used to be accepted as an alternate way
1794 to specify named sections from all unallocated input files. Because
1795 some operating systems (VMS) allow brackets in file names, that notation
1796 is no longer supported.
1798 @cindex uninitialized data
1799 @cindex commons in output
1801 @item @var{filename}@code{( COMMON )}
1803 Specify where in your output file to place uninitialized data
1804 with this notation. @code{*(COMMON)} by itself refers to all
1805 uninitialized data from all input files (so far as it is not yet
1806 allocated); @var{filename}@code{(COMMON)} refers to uninitialized data
1807 from a particular file. Both are special cases of the general
1808 mechanisms for specifying where to place input-file sections:
1809 @code{ld} permits you to refer to uninitialized data as if it
1810 were in an input-file section named @code{COMMON}, regardless of the
1811 input file's format.
1814 For example, the following command script arranges the output file into
1815 three consecutive sections, named @code{.text}, @code{.data}, and
1816 @code{.bss}, taking the input for each from the correspondingly named
1817 sections of all the input files:
1822 .text : @{ *(.text) @}
1823 .data : @{ *(.data) @}
1824 .bss : @{ *(.bss) *(COMMON) @}
1829 The following example reads all of the sections from file @code{all.o}
1830 and places them at the start of output section @code{outputa} which
1831 starts at location @code{0x10000}. All of section @code{.input1} from
1832 file @code{foo.o} follows immediately, in the same output section. All
1833 of section @code{.input2} from @code{foo.o} goes into output section
1834 @code{outputb}, followed by section @code{.input1} from @code{foo1.o}.
1835 All of the remaining @code{.input1} and @code{.input2} sections from any
1836 files are written to output section @code{outputc}.
1860 @node Section Data Expressions
1861 @subsection Section Data Expressions
1863 @cindex expressions in a section
1864 The foregoing statements arrange, in your output file, data originating
1865 from your input files. You can also place data directly in an output
1866 section from the link command script. Most of these additional
1867 statements involve expressions; @pxref{Expressions}. Although these
1868 statements are shown separately here for ease of presentation, no such
1869 segregation is needed within a section definition in the @code{SECTIONS}
1870 command; you can intermix them freely with any of the statements we've
1874 @cindex input filename symbols
1875 @cindex filename symbols
1876 @kindex CREATE_OBJECT_SYMBOLS
1877 @item CREATE_OBJECT_SYMBOLS
1878 Create a symbol for each input file
1879 in the current section, set to the address of the first byte of
1880 data written from that input file. For instance, with @code{a.out}
1881 files it is conventional to have a symbol for each input file. You can
1882 accomplish this by defining the output @code{.text} section as follows:
1888 CREATE_OBJECT_SYMBOLS
1890 _etext = ALIGN(0x2000);
1897 If @code{sample.ld} is a file containing this script, and @code{a.o},
1898 @code{b.o}, @code{c.o}, and @code{d.o} are four input files with
1899 contents like the following---
1911 @samp{ld -M -T sample.ld a.o b.o c.o d.o} would create a map like this,
1912 containing symbols matching the object file names:
1914 00000000 A __DYNAMIC
1917 00002020 T _afunction
1920 00002038 T _bfunction
1923 00002050 T _cfunction
1926 00002068 T _dfunction
1936 @kindex @var{symbol} = @var{expression} ;
1937 @kindex @var{symbol} @var{f}= @var{expression} ;
1938 @item @var{symbol} = @var{expression} ;
1939 @itemx @var{symbol} @var{f}= @var{expression} ;
1940 @var{symbol} is any symbol name (@pxref{Symbols}). ``@var{f}=''
1941 refers to any of the operators @code{&= += -= *= /=} which combine
1942 arithmetic and assignment.
1944 @cindex assignment, in section defn
1945 When you assign a value to a symbol within a particular section
1946 definition, the value is relative to the beginning of the section
1947 (@pxref{Assignment}). If you write
1954 .data : @{ @dots{} rel = 14 ; @dots{} @}
1955 abs2 = 14 + ADDR(.data);
1961 @c FIXME: Try above example!
1963 @code{abs} and @code{rel} do not have the same value; @code{rel} has the
1964 same value as @code{abs2}.
1966 @kindex BYTE(@var{expression})
1967 @kindex SHORT(@var{expression})
1968 @kindex LONG(@var{expression})
1969 @kindex QUAD(@var{expression})
1970 @cindex direct output
1971 @item BYTE(@var{expression})
1972 @itemx SHORT(@var{expression})
1973 @itemx LONG(@var{expression})
1974 @itemx QUAD(@var{expression})
1975 By including one of these four statements in a section definition, you
1976 can explicitly place one, two, four, or eight bytes (respectively) at
1977 the current address of that section. @code{QUAD} is only supported when
1978 using a 64 bit host or target.
1980 @ifclear SingleFormat
1981 Multiple-byte quantities are represented in whatever byte order is
1982 appropriate for the output file format (@pxref{BFD}).
1985 @kindex FILL(@var{expression})
1986 @cindex holes, filling
1987 @cindex unspecified memory
1988 @item FILL(@var{expression})
1989 Specify the ``fill pattern'' for the current section. Any otherwise
1990 unspecified regions of memory within the section (for example, regions
1991 you skip over by assigning a new value to the location counter @samp{.})
1992 are filled with the two least significant bytes from the
1993 @var{expression} argument. A @code{FILL} statement covers memory
1994 locations @emph{after} the point it occurs in the section definition; by
1995 including more than one @code{FILL} statement, you can have different
1996 fill patterns in different parts of an output section.
1999 @node Section Options
2000 @subsection Optional Section Attributes
2001 @cindex section defn, full syntax
2002 Here is the full syntax of a section definition, including all the
2009 @var{secname} @var{start} BLOCK(@var{align}) (NOLOAD) : AT ( @var{ldadr} )
2010 @{ @var{contents} @} >@var{region} :@var{phdr} =@var{fill}
2016 @var{secname} and @var{contents} are required. @xref{Section
2017 Definition}, and @pxref{Section Placement} for details on
2018 @var{contents}. The remaining elements---@var{start},
2019 @code{BLOCK(@var{align)}}, @code{(NOLOAD)}, @code{AT ( @var{ldadr} )},
2020 @code{>@var{region}}, @code{:@var{phdr}}, and @code{=@var{fill}}---are
2024 @cindex start address, section
2025 @cindex section start
2026 @cindex section address
2028 You can force the output section to be loaded at a specified address by
2029 specifying @var{start} immediately following the section name.
2030 @var{start} can be represented as any expression. The following
2031 example generates section @var{output} at location
2038 output 0x40000000: @{
2046 @kindex BLOCK(@var{align})
2047 @cindex section alignment
2048 @cindex aligning sections
2049 @item BLOCK(@var{align})
2050 You can include @code{BLOCK()} specification to advance
2051 the location counter @code{.} prior to the beginning of the section, so
2052 that the section will begin at the specified alignment. @var{align} is
2056 @cindex prevent unnecessary loading
2057 @cindex loading, preventing
2059 Use @samp{(NOLOAD)} to prevent a section from being loaded into memory
2060 each time it is accessed. For example, in the script sample below, the
2061 @code{ROM} segment is addressed at memory location @samp{0} and does not
2062 need to be loaded into each object file:
2067 ROM 0 (NOLOAD) : @{ @dots{} @}
2073 @kindex AT ( @var{ldadr} )
2074 @cindex specify load address
2075 @cindex load address, specifying
2076 @item AT ( @var{ldadr} )
2077 The expression @var{ldadr} that follows the @code{AT} keyword specifies
2078 the load address of the section. The default (if you do not use the
2079 @code{AT} keyword) is to make the load address the same as the
2080 relocation address. This feature is designed to make it easy to build a
2081 ROM image. For example, this @code{SECTIONS} definition creates two
2082 output sections: one called @samp{.text}, which starts at @code{0x1000},
2083 and one called @samp{.mdata}, which is loaded at the end of the
2084 @samp{.text} section even though its relocation address is
2085 @code{0x2000}. The symbol @code{_data} is defined with the value
2092 .text 0x1000 : @{ *(.text) _etext = . ; @}
2094 AT ( ADDR(.text) + SIZEOF ( .text ) )
2095 @{ _data = . ; *(.data); _edata = . ; @}
2097 @{ _bstart = . ; *(.bss) *(COMMON) ; _bend = . ;@}
2102 The run-time initialization code (for C programs, usually @code{crt0})
2103 for use with a ROM generated this way has to include something like
2104 the following, to copy the initialized data from the ROM image to its runtime
2112 /* ROM has data at end of text; copy it. */
2113 while (dst < _edata) @{
2118 for (dst = _bstart; dst< _bend; dst++)
2123 @kindex >@var{region}
2124 @cindex section, assigning to memory region
2125 @cindex memory regions and sections
2127 Assign this section to a previously defined region of memory.
2131 @cindex section, assigning to program header
2132 @cindex program headers and sections
2134 Assign this section to a segment described by a program header.
2135 @xref{PHDRS}. If a section is assigned to one or more segments, than
2136 all subsequent allocated sections will be assigned to those segments as
2137 well, unless they use an explicitly @code{:@var{phdr}} modifier. To
2138 prevent a section from being assigned to a segment when it would
2139 normally default to one, use @code{:NONE}.
2142 @cindex section fill pattern
2143 @cindex fill pattern, entire section
2145 Including @code{=@var{fill}} in a section definition specifies the
2146 initial fill value for that section. You may use any expression to
2147 specify @var{fill}. Any unallocated holes in the current output section
2148 when written to the output file will be filled with the two least
2149 significant bytes of the value, repeated as necessary. You can also
2150 change the fill value with a @code{FILL} statement in the @var{contents}
2151 of a section definition.
2156 @section ELF Program Headers
2158 @kindex program headers
2159 @kindex ELF program headers
2161 The ELF object file format uses @dfn{program headers}, which are read by
2162 the system loader and describe how the program should be loaded into
2163 memory. These program headers must be set correctly in order to run the
2164 program on a native ELF system. The linker will create reasonable
2165 program headers by default. However, in some cases, it is desirable to
2166 specify the program headers more precisely; the @code{PHDRS} command may
2167 be used for this purpose. When the @code{PHDRS} command is used, the
2168 linker will not generate any program headers itself.
2170 The @code{PHDRS} command is only meaningful when generating an ELF
2171 output file. It is ignored in other cases. This manual does not
2172 describe the details of how the system loader interprets program
2173 headers; for more information, see the ELF ABI. The program headers of
2174 an ELF file may be displayed using the @samp{-p} option of the
2175 @code{objdump} command.
2177 This is the syntax of the @code{PHDRS} command. The words @code{PHDRS},
2178 @code{FILEHDR}, @code{AT}, and @code{FLAGS} are keywords.
2184 @var{name} @var{type} [ FILEHDR ] [ PHDRS ] [ AT ( @var{address} ) ]
2185 [ FLAGS ( @var{flags} ) ] ;
2190 The @var{name} is used only for reference in the @code{SECTIONS} command
2191 of the linker script. It does not get put into the output file.
2193 Certain program header types describe segments of memory which are
2194 loaded from the file by the system loader. In the linker script, the
2195 contents of these segments are specified by directing allocated output
2196 sections to be placed in the segment. To do this, the command
2197 describing the output section in the @code{SECTIONS} command should use
2198 @samp{:@var{name}}, where @var{name} is the name of the program header
2199 as it appears in the @code{PHDRS} command. @xref{Section Options}.
2201 It is normal for certain sections to appear in more than one segment.
2202 This merely implies that one segment of memory contains another. This
2203 is specified by repeating @samp{:@var{name}}, using it once for each
2204 program header in which the section is to appear.
2206 If a section is placed in one or more segments using @samp{:@var{name}},
2207 then all subsequent allocated sections which do not specify
2208 @samp{:@var{name}} are placed in the same segments. This is for
2209 convenience, since generally a whole set of contiguous sections will be
2210 placed in a single segment. To prevent a section from being assigned to
2211 a segment when it would normally default to one, use @code{:NONE}.
2213 The @code{FILEHDR} and @code{PHDRS} keywords which may appear after the
2214 program header type also indicate contents of the segment of memory.
2215 The @code{FILEHDR} keyword means that the segment should include the ELF
2216 file header. The @code{PHDRS} keyword means that the segment should
2217 include the ELF program headers themselves.
2219 The @var{type} may be one of the following. The numbers indicate the
2220 value of the keyword.
2223 @item @code{PT_NULL} (0)
2224 Indicates an unused program header.
2226 @item @code{PT_LOAD} (1)
2227 Indicates that this program header describes a segment to be loaded from
2230 @item @code{PT_DYNAMIC} (2)
2231 Indicates a segment where dynamic linking information can be found.
2233 @item @code{PT_INTERP} (3)
2234 Indicates a segment where the name of the program interpreter may be
2237 @item @code{PT_NOTE} (4)
2238 Indicates a segment holding note information.
2240 @item @code{PT_SHLIB} (5)
2241 A reserved program header type, defined but not specified by the ELF
2244 @item @code{PT_PHDR} (6)
2245 Indicates a segment where the program headers may be found.
2247 @item @var{expression}
2248 An expression giving the numeric type of the program header. This may
2249 be used for types not defined above.
2252 It is possible to specify that a segment should be loaded at a
2253 particular address in memory. This is done using an @code{AT}
2254 expression. This is identical to the @code{AT} command used in the
2255 @code{SECTIONS} command (@pxref{Section Options}). Using the @code{AT}
2256 command for a program header overrides any information in the
2257 @code{SECTIONS} command.
2259 Normally the segment flags are set based on the sections. The
2260 @code{FLAGS} keyword may be used to explicitly specify the segment
2261 flags. The value of @var{flags} must be an integer. It is used to
2262 set the @code{p_flags} field of the program header.
2264 Here is an example of the use of @code{PHDRS}. This shows a typical set
2265 of program headers used on a native ELF system.
2271 headers PT_PHDR PHDRS ;
2273 text PT_LOAD FILEHDR PHDRS ;
2275 dynamic PT_DYNAMIC ;
2281 .interp : @{ *(.interp) @} :text :interp
2282 .text : @{ *(.text) @} :text
2283 .rodata : @{ *(.rodata) @} /* defaults to :text */
2285 . = . + 0x1000; /* move to a new page in memory */
2286 .data : @{ *(.data) @} :data
2287 .dynamic : @{ *(.dynamic) @} :data :dynamic
2294 @section The Entry Point
2295 @kindex ENTRY(@var{symbol})
2296 @cindex start of execution
2297 @cindex first instruction
2298 The linker command language includes a command specifically for
2299 defining the first executable instruction in an output file (its
2300 @dfn{entry point}). Its argument is a symbol name:
2305 Like symbol assignments, the @code{ENTRY} command may be placed either
2306 as an independent command in the command file, or among the section
2307 definitions within the @code{SECTIONS} command---whatever makes the most
2308 sense for your layout.
2310 @cindex entry point, defaults
2311 @code{ENTRY} is only one of several ways of choosing the entry point.
2312 You may indicate it in any of the following ways (shown in descending
2313 order of priority: methods higher in the list override methods lower down).
2316 the @samp{-e} @var{entry} command-line option;
2318 the @code{ENTRY(@var{symbol})} command in a linker control script;
2320 the value of the symbol @code{start}, if present;
2322 the address of the first byte of the @code{.text} section, if present;
2324 The address @code{0}.
2327 For example, you can use these rules to generate an entry point with an
2328 assignment statement: if no symbol @code{start} is defined within your
2329 input files, you can simply define it, assigning it an appropriate
2337 The example shows an absolute address, but you can use any expression.
2338 For example, if your input object files use some other symbol-name
2339 convention for the entry point, you can just assign the value of
2340 whatever symbol contains the start address to @code{start}:
2343 start = other_symbol ;
2346 @node Option Commands
2347 @section Option Commands
2348 The command language includes a number of other commands that you can
2349 use for specialized purposes. They are similar in purpose to
2350 command-line options.
2353 @kindex CONSTRUCTORS
2354 @cindex C++ constructors, arranging in link
2355 @cindex constructors, arranging in link
2357 This command ties up C++ style constructor and destructor records. The
2358 details of the constructor representation vary from one object format to
2359 another, but usually lists of constructors and destructors appear as
2360 special sections. The @code{CONSTRUCTORS} command specifies where the
2361 linker is to place the data from these sections, relative to the rest of
2362 the linked output. Constructor data is marked by the symbol
2363 @w{@code{__CTOR_LIST__}} at the start, and @w{@code{__CTOR_LIST_END}} at
2364 the end; destructor data is bracketed similarly, between
2365 @w{@code{__DTOR_LIST__}} and @w{@code{__DTOR_LIST_END}}. (The compiler
2366 must arrange to actually run this code; @sc{gnu} C++ calls constructors from
2367 a subroutine @code{__main}, which it inserts automatically into the
2368 startup code for @code{main}, and destructors from @code{_exit}.)
2375 These keywords were used in some older linkers to request a particular
2376 math subroutine library. @code{ld} doesn't use the keywords, assuming
2377 instead that any necessary subroutines are in libraries specified using
2378 the general mechanisms for linking to archives; but to permit the use of
2379 scripts that were written for the older linkers, the keywords
2380 @code{FLOAT} and @code{NOFLOAT} are accepted and ignored.
2382 @kindex FORCE_COMMON_ALLOCATION
2383 @cindex common allocation
2384 @item FORCE_COMMON_ALLOCATION
2385 This command has the same effect as the @samp{-d} command-line option:
2386 to make @code{ld} assign space to common symbols even if a relocatable
2387 output file is specified (@samp{-r}).
2389 @kindex INPUT ( @var{files} )
2390 @cindex binary input files
2391 @item INPUT ( @var{file}, @var{file}, @dots{} )
2392 @itemx INPUT ( @var{file} @var{file} @dots{} )
2393 Use this command to include binary input files in the link, without
2394 including them in a particular section definition.
2395 Specify the full name for each @var{file}, including @samp{.a} if
2398 @code{ld} searches for each @var{file} through the archive-library
2399 search path, just as for files you specify on the command line.
2400 See the description of @samp{-L} in @ref{Options,,Command Line
2403 If you use @samp{-l@var{file}}, @code{ld} will transform the name to
2404 @code{lib@var{file}.a} as with the command line argument @samp{-l}.
2406 @kindex GROUP ( @var{files} )
2407 @cindex grouping input files
2408 @item GROUP ( @var{file}, @var{file}, @dots{} )
2409 @itemx GROUP ( @var{file} @var{file} @dots{} )
2410 This command is like @code{INPUT}, except that the named files should
2411 all be archives, and they are searched repeatedly until no new undefined
2412 references are created. See the description of @samp{-(} in
2413 @ref{Options,,Command Line Options}.
2416 @kindex MAP ( @var{name} )
2417 @item MAP ( @var{name} )
2418 @c MAP(...) appears to look for an F in the arg, ignoring all other
2419 @c chars; if it finds one, it sets "map_option_f" to true. But nothing
2420 @c checks map_option_f. Apparently a stub for the future...
2423 @kindex OUTPUT ( @var{filename} )
2424 @cindex naming the output file
2425 @item OUTPUT ( @var{filename} )
2426 Use this command to name the link output file @var{filename}. The
2427 effect of @code{OUTPUT(@var{filename})} is identical to the effect of
2428 @w{@samp{-o @var{filename}}}, which overrides it. You can use this
2429 command to supply a default output-file name other than @code{a.out}.
2431 @ifclear SingleFormat
2432 @kindex OUTPUT_ARCH ( @var{bfdname} )
2433 @cindex machine architecture, output
2434 @item OUTPUT_ARCH ( @var{bfdname} )
2435 Specify a particular output machine architecture, with one of the names
2436 used by the BFD back-end routines (@pxref{BFD}). This command is often
2437 unnecessary; the architecture is most often set implicitly by either the
2438 system BFD configuration or as a side effect of the @code{OUTPUT_FORMAT}
2441 @kindex OUTPUT_FORMAT ( @var{bfdname} )
2442 @cindex format, output file
2443 @item OUTPUT_FORMAT ( @var{bfdname} )
2444 When @code{ld} is configured to support multiple object code formats,
2445 you can use this command to specify a particular output format.
2446 @var{bfdname} is one of the names used by the BFD back-end routines
2447 (@pxref{BFD}). The effect is identical to the effect of the
2448 @samp{-oformat} command-line option. This selection affects only
2449 the output file; the related command @code{TARGET} affects primarily
2453 @kindex SEARCH_DIR ( @var{path} )
2454 @cindex path for libraries
2455 @cindex search path, libraries
2456 @item SEARCH_DIR ( @var{path} )
2457 Add @var{path} to the list of paths where @code{ld} looks for
2458 archive libraries. @code{SEARCH_DIR(@var{path})} has the same
2459 effect as @samp{-L@var{path}} on the command line.
2461 @kindex STARTUP ( @var{filename} )
2462 @cindex first input file
2463 @item STARTUP ( @var{filename} )
2464 Ensure that @var{filename} is the first input file used in the link
2467 @ifclear SingleFormat
2468 @cindex input file format
2469 @kindex TARGET ( @var{format} )
2470 @item TARGET ( @var{format} )
2471 When @code{ld} is configured to support multiple object code formats,
2472 you can use this command to change the input-file object code format
2473 (like the command-line option @samp{-b} or its synonym @samp{-format}).
2474 The argument @var{format} is one of the strings used by BFD to name
2475 binary formats. If @code{TARGET} is specified but @code{OUTPUT_FORMAT}
2476 is not, the last @code{TARGET} argument is also used as the default
2477 format for the @code{ld} output file. @xref{BFD}.
2480 If you don't use the @code{TARGET} command, @code{ld} uses the value of
2481 the environment variable @code{GNUTARGET}, if available, to select the
2482 output file format. If that variable is also absent, @code{ld} uses
2483 the default format configured for your machine in the BFD libraries.
2488 @node Machine Dependent
2489 @chapter Machine Dependent Features
2491 @cindex machine dependencies
2492 @code{ld} has additional features on some platforms; the following
2493 sections describe them. Machines where @code{ld} has no additional
2494 functionality are not listed.
2497 * H8/300:: @code{ld} and the H8/300
2498 * i960:: @code{ld} and the Intel 960 family
2502 @c FIXME! This could use @raisesections/@lowersections, but there seems to be a conflict
2503 @c between those and node-defaulting.
2509 @section @code{ld} and the H8/300
2511 @cindex H8/300 support
2512 For the H8/300, @code{ld} can perform these global optimizations when
2513 you specify the @samp{-relax} command-line option.
2516 @cindex relaxing on H8/300
2517 @item relaxing address modes
2518 @code{ld} finds all @code{jsr} and @code{jmp} instructions whose
2519 targets are within eight bits, and turns them into eight-bit
2520 program-counter relative @code{bsr} and @code{bra} instructions,
2523 @cindex synthesizing on H8/300
2524 @item synthesizing instructions
2525 @c FIXME: specifically mov.b, or any mov instructions really?
2526 @code{ld} finds all @code{mov.b} instructions which use the
2527 sixteen-bit absolute address form, but refer to the top
2528 page of memory, and changes them to use the eight-bit address form.
2529 (That is: the linker turns @samp{mov.b @code{@@}@var{aa}:16} into
2530 @samp{mov.b @code{@@}@var{aa}:8} whenever the address @var{aa} is in the
2531 top page of memory).
2540 @c This stuff is pointless to say unless you're especially concerned
2541 @c with Hitachi chips; don't enable it for generic case, please.
2543 @chapter @code{ld} and other Hitachi chips
2545 @code{ld} also supports the H8/300H, the H8/500, and the Hitachi SH. No
2546 special features, commands, or command-line options are required for
2556 @section @code{ld} and the Intel 960 family
2558 @cindex i960 support
2560 You can use the @samp{-A@var{architecture}} command line option to
2561 specify one of the two-letter names identifying members of the 960
2562 family; the option specifies the desired output target, and warns of any
2563 incompatible instructions in the input files. It also modifies the
2564 linker's search strategy for archive libraries, to support the use of
2565 libraries specific to each particular architecture, by including in the
2566 search loop names suffixed with the string identifying the architecture.
2568 For example, if your @code{ld} command line included @w{@samp{-ACA}} as
2569 well as @w{@samp{-ltry}}, the linker would look (in its built-in search
2570 paths, and in any paths you specify with @samp{-L}) for a library with
2583 The first two possibilities would be considered in any event; the last
2584 two are due to the use of @w{@samp{-ACA}}.
2586 You can meaningfully use @samp{-A} more than once on a command line, since
2587 the 960 architecture family allows combination of target architectures; each
2588 use will add another pair of name variants to search for when @w{@samp{-l}}
2589 specifies a library.
2591 @cindex @code{-relax} on i960
2592 @cindex relaxing on i960
2593 @code{ld} supports the @samp{-relax} option for the i960 family. If you
2594 specify @samp{-relax}, @code{ld} finds all @code{balx} and @code{calx}
2595 instructions whose targets are within 24 bits, and turns them into
2596 24-bit program-counter relative @code{bal} and @code{cal}
2597 instructions, respectively. @code{ld} also turns @code{cal}
2598 instructions into @code{bal} instructions when it determines that the
2599 target subroutine is a leaf routine (that is, the target subroutine does
2600 not itself call any subroutines).
2607 @ifclear SingleFormat
2612 @cindex object file management
2613 @cindex object formats available
2615 The linker accesses object and archive files using the BFD libraries.
2616 These libraries allow the linker to use the same routines to operate on
2617 object files whatever the object file format. A different object file
2618 format can be supported simply by creating a new BFD back end and adding
2619 it to the library. To conserve runtime memory, however, the linker and
2620 associated tools are usually configured to support only a subset of the
2621 object file formats available. You can use @code{objdump -i}
2622 (@pxref{objdump,,objdump,binutils.info,The GNU Binary Utilities}) to
2623 list all the formats available for your configuration.
2625 @cindex BFD requirements
2626 @cindex requirements for BFD
2627 As with most implementations, BFD is a compromise between
2628 several conflicting requirements. The major factor influencing
2629 BFD design was efficiency: any time used converting between
2630 formats is time which would not have been spent had BFD not
2631 been involved. This is partly offset by abstraction payback; since
2632 BFD simplifies applications and back ends, more time and care
2633 may be spent optimizing algorithms for a greater speed.
2635 One minor artifact of the BFD solution which you should bear in
2636 mind is the potential for information loss. There are two places where
2637 useful information can be lost using the BFD mechanism: during
2638 conversion and during output. @xref{BFD information loss}.
2641 * BFD outline:: How it works: an outline of BFD
2645 @section How it works: an outline of BFD
2646 @cindex opening object files
2647 @include bfdsumm.texi
2651 @appendix MRI Compatible Script Files
2652 @cindex MRI compatibility
2653 To aid users making the transition to @sc{gnu} @code{ld} from the MRI
2654 linker, @code{ld} can use MRI compatible linker scripts as an
2655 alternative to the more general-purpose linker scripting language
2656 described in @ref{Commands,,Command Language}. MRI compatible linker
2657 scripts have a much simpler command set than the scripting language
2658 otherwise used with @code{ld}. @sc{gnu} @code{ld} supports the most
2659 commonly used MRI linker commands; these commands are described here.
2661 In general, MRI scripts aren't of much use with the @code{a.out} object
2662 file format, since it only has three sections and MRI scripts lack some
2663 features to make use of them.
2665 You can specify a file containing an MRI-compatible script using the
2666 @samp{-c} command-line option.
2668 Each command in an MRI-compatible script occupies its own line; each
2669 command line starts with the keyword that identifies the command (though
2670 blank lines are also allowed for punctuation). If a line of an
2671 MRI-compatible script begins with an unrecognized keyword, @code{ld}
2672 issues a warning message, but continues processing the script.
2674 Lines beginning with @samp{*} are comments.
2676 You can write these commands using all upper-case letters, or all
2677 lower case; for example, @samp{chip} is the same as @samp{CHIP}.
2678 The following list shows only the upper-case form of each command.
2681 @cindex @code{ABSOLUTE} (MRI)
2682 @item ABSOLUTE @var{secname}
2683 @itemx ABSOLUTE @var{secname}, @var{secname}, @dots{} @var{secname}
2684 Normally, @code{ld} includes in the output file all sections from all
2685 the input files. However, in an MRI-compatible script, you can use the
2686 @code{ABSOLUTE} command to restrict the sections that will be present in
2687 your output program. If the @code{ABSOLUTE} command is used at all in a
2688 script, then only the sections named explicitly in @code{ABSOLUTE}
2689 commands will appear in the linker output. You can still use other
2690 input sections (whatever you select on the command line, or using
2691 @code{LOAD}) to resolve addresses in the output file.
2693 @cindex @code{ALIAS} (MRI)
2694 @item ALIAS @var{out-secname}, @var{in-secname}
2695 Use this command to place the data from input section @var{in-secname}
2696 in a section called @var{out-secname} in the linker output file.
2698 @var{in-secname} may be an integer.
2700 @cindex @code{ALIGN} (MRI)
2701 @item ALIGN @var{secname} = @var{expression}
2702 Align the section called @var{secname} to @var{expression}. The
2703 @var{expression} should be a power of two.
2705 @cindex @code{BASE} (MRI)
2706 @item BASE @var{expression}
2707 Use the value of @var{expression} as the lowest address (other than
2708 absolute addresses) in the output file.
2710 @cindex @code{CHIP} (MRI)
2711 @item CHIP @var{expression}
2712 @itemx CHIP @var{expression}, @var{expression}
2713 This command does nothing; it is accepted only for compatibility.
2715 @cindex @code{END} (MRI)
2717 This command does nothing whatever; it's only accepted for compatibility.
2719 @cindex @code{FORMAT} (MRI)
2720 @item FORMAT @var{output-format}
2721 Similar to the @code{OUTPUT_FORMAT} command in the more general linker
2722 language, but restricted to one of these output formats:
2726 S-records, if @var{output-format} is @samp{S}
2729 IEEE, if @var{output-format} is @samp{IEEE}
2732 COFF (the @samp{coff-m68k} variant in BFD), if @var{output-format} is
2736 @cindex @code{LIST} (MRI)
2737 @item LIST @var{anything}@dots{}
2738 Print (to the standard output file) a link map, as produced by the
2739 @code{ld} command-line option @samp{-M}.
2741 The keyword @code{LIST} may be followed by anything on the
2742 same line, with no change in its effect.
2744 @cindex @code{LOAD} (MRI)
2745 @item LOAD @var{filename}
2746 @itemx LOAD @var{filename}, @var{filename}, @dots{} @var{filename}
2747 Include one or more object file @var{filename} in the link; this has the
2748 same effect as specifying @var{filename} directly on the @code{ld}
2751 @cindex @code{NAME} (MRI)
2752 @item NAME @var{output-name}
2753 @var{output-name} is the name for the program produced by @code{ld}; the
2754 MRI-compatible command @code{NAME} is equivalent to the command-line
2755 option @samp{-o} or the general script language command @code{OUTPUT}.
2757 @cindex @code{ORDER} (MRI)
2758 @item ORDER @var{secname}, @var{secname}, @dots{} @var{secname}
2759 @itemx ORDER @var{secname} @var{secname} @var{secname}
2760 Normally, @code{ld} orders the sections in its output file in the
2761 order in which they first appear in the input files. In an MRI-compatible
2762 script, you can override this ordering with the @code{ORDER} command. The
2763 sections you list with @code{ORDER} will appear first in your output
2764 file, in the order specified.
2766 @cindex @code{PUBLIC} (MRI)
2767 @item PUBLIC @var{name}=@var{expression}
2768 @itemx PUBLIC @var{name},@var{expression}
2769 @itemx PUBLIC @var{name} @var{expression}
2770 Supply a value (@var{expression}) for external symbol
2771 @var{name} used in the linker input files.
2773 @cindex @code{SECT} (MRI)
2774 @item SECT @var{secname}, @var{expression}
2775 @itemx SECT @var{secname}=@var{expression}
2776 @itemx SECT @var{secname} @var{expression}
2777 You can use any of these three forms of the @code{SECT} command to
2778 specify the start address (@var{expression}) for section @var{secname}.
2779 If you have more than one @code{SECT} statement for the same
2780 @var{secname}, only the @emph{first} sets the start address.
2790 % I think something like @colophon should be in texinfo. In the
2792 \long\def\colophon{\hbox to0pt{}\vfill
2793 \centerline{The body of this manual is set in}
2794 \centerline{\fontname\tenrm,}
2795 \centerline{with headings in {\bf\fontname\tenbf}}
2796 \centerline{and examples in {\tt\fontname\tentt}.}
2797 \centerline{{\it\fontname\tenit\/} and}
2798 \centerline{{\sl\fontname\tensl\/}}
2799 \centerline{are used for emphasis.}\vfill}
2801 % Blame: doc@cygnus.com, 28mar91.