1 This is ld.info, produced by makeinfo version 4.8 from ld.texinfo.
3 INFO-DIR-SECTION Software development
5 * Ld: (ld). The GNU linker.
8 This file documents the GNU linker LD (GNU Binutils) version 2.23.1.
10 Copyright (C) 1991, 1992, 1993, 1994, 1995, 1996, 1997, 1998, 1999,
11 2000, 2001, 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009 Free
12 Software Foundation, Inc.
14 Permission is granted to copy, distribute and/or modify this document
15 under the terms of the GNU Free Documentation License, Version 1.3 or
16 any later version published by the Free Software Foundation; with no
17 Invariant Sections, with no Front-Cover Texts, and with no Back-Cover
18 Texts. A copy of the license is included in the section entitled "GNU
19 Free Documentation License".
22 File: ld.info, Node: Top, Next: Overview, Up: (dir)
27 This file documents the GNU linker ld (GNU Binutils) version 2.23.1.
29 This document is distributed under the terms of the GNU Free
30 Documentation License version 1.3. A copy of the license is included
31 in the section entitled "GNU Free Documentation License".
36 * Invocation:: Invocation
37 * Scripts:: Linker Scripts
39 * Machine Dependent:: Machine Dependent Features
43 * Reporting Bugs:: Reporting Bugs
44 * MRI:: MRI Compatible Script Files
45 * GNU Free Documentation License:: GNU Free Documentation License
49 File: ld.info, Node: Overview, Next: Invocation, Prev: Top, Up: Top
54 `ld' combines a number of object and archive files, relocates their
55 data and ties up symbol references. Usually the last step in compiling
56 a program is to run `ld'.
58 `ld' accepts Linker Command Language files written in a superset of
59 AT&T's Link Editor Command Language syntax, to provide explicit and
60 total control over the linking process.
62 This version of `ld' uses the general purpose BFD libraries to
63 operate on object files. This allows `ld' to read, combine, and write
64 object files in many different formats--for example, COFF or `a.out'.
65 Different formats may be linked together to produce any available kind
66 of object file. *Note BFD::, for more information.
68 Aside from its flexibility, the GNU linker is more helpful than other
69 linkers in providing diagnostic information. Many linkers abandon
70 execution immediately upon encountering an error; whenever possible,
71 `ld' continues executing, allowing you to identify other errors (or, in
72 some cases, to get an output file in spite of the error).
75 File: ld.info, Node: Invocation, Next: Scripts, Prev: Overview, Up: Top
80 The GNU linker `ld' is meant to cover a broad range of situations, and
81 to be as compatible as possible with other linkers. As a result, you
82 have many choices to control its behavior.
86 * Options:: Command Line Options
87 * Environment:: Environment Variables
90 File: ld.info, Node: Options, Next: Environment, Up: Invocation
92 2.1 Command Line Options
93 ========================
95 The linker supports a plethora of command-line options, but in actual
96 practice few of them are used in any particular context. For instance,
97 a frequent use of `ld' is to link standard Unix object files on a
98 standard, supported Unix system. On such a system, to link a file
101 ld -o OUTPUT /lib/crt0.o hello.o -lc
103 This tells `ld' to produce a file called OUTPUT as the result of
104 linking the file `/lib/crt0.o' with `hello.o' and the library `libc.a',
105 which will come from the standard search directories. (See the
106 discussion of the `-l' option below.)
108 Some of the command-line options to `ld' may be specified at any
109 point in the command line. However, options which refer to files, such
110 as `-l' or `-T', cause the file to be read at the point at which the
111 option appears in the command line, relative to the object files and
112 other file options. Repeating non-file options with a different
113 argument will either have no further effect, or override prior
114 occurrences (those further to the left on the command line) of that
115 option. Options which may be meaningfully specified more than once are
116 noted in the descriptions below.
118 Non-option arguments are object files or archives which are to be
119 linked together. They may follow, precede, or be mixed in with
120 command-line options, except that an object file argument may not be
121 placed between an option and its argument.
123 Usually the linker is invoked with at least one object file, but you
124 can specify other forms of binary input files using `-l', `-R', and the
125 script command language. If _no_ binary input files at all are
126 specified, the linker does not produce any output, and issues the
127 message `No input files'.
129 If the linker cannot recognize the format of an object file, it will
130 assume that it is a linker script. A script specified in this way
131 augments the main linker script used for the link (either the default
132 linker script or the one specified by using `-T'). This feature
133 permits the linker to link against a file which appears to be an object
134 or an archive, but actually merely defines some symbol values, or uses
135 `INPUT' or `GROUP' to load other objects. Specifying a script in this
136 way merely augments the main linker script, with the extra commands
137 placed after the main script; use the `-T' option to replace the
138 default linker script entirely, but note the effect of the `INSERT'
139 command. *Note Scripts::.
141 For options whose names are a single letter, option arguments must
142 either follow the option letter without intervening whitespace, or be
143 given as separate arguments immediately following the option that
146 For options whose names are multiple letters, either one dash or two
147 can precede the option name; for example, `-trace-symbol' and
148 `--trace-symbol' are equivalent. Note--there is one exception to this
149 rule. Multiple letter options that start with a lower case 'o' can
150 only be preceded by two dashes. This is to reduce confusion with the
151 `-o' option. So for example `-omagic' sets the output file name to
152 `magic' whereas `--omagic' sets the NMAGIC flag on the output.
154 Arguments to multiple-letter options must either be separated from
155 the option name by an equals sign, or be given as separate arguments
156 immediately following the option that requires them. For example,
157 `--trace-symbol foo' and `--trace-symbol=foo' are equivalent. Unique
158 abbreviations of the names of multiple-letter options are accepted.
160 Note--if the linker is being invoked indirectly, via a compiler
161 driver (e.g. `gcc') then all the linker command line options should be
162 prefixed by `-Wl,' (or whatever is appropriate for the particular
163 compiler driver) like this:
165 gcc -Wl,--start-group foo.o bar.o -Wl,--end-group
167 This is important, because otherwise the compiler driver program may
168 silently drop the linker options, resulting in a bad link. Confusion
169 may also arise when passing options that require values through a
170 driver, as the use of a space between option and argument acts as a
171 separator, and causes the driver to pass only the option to the linker
172 and the argument to the compiler. In this case, it is simplest to use
173 the joined forms of both single- and multiple-letter options, such as:
175 gcc foo.o bar.o -Wl,-eENTRY -Wl,-Map=a.map
177 Here is a table of the generic command line switches accepted by the
181 Read command-line options from FILE. The options read are
182 inserted in place of the original @FILE option. If FILE does not
183 exist, or cannot be read, then the option will be treated
184 literally, and not removed.
186 Options in FILE are separated by whitespace. A whitespace
187 character may be included in an option by surrounding the entire
188 option in either single or double quotes. Any character
189 (including a backslash) may be included by prefixing the character
190 to be included with a backslash. The FILE may itself contain
191 additional @FILE options; any such options will be processed
195 This option is supported for HP/UX compatibility. The KEYWORD
196 argument must be one of the strings `archive', `shared', or
197 `default'. `-aarchive' is functionally equivalent to `-Bstatic',
198 and the other two keywords are functionally equivalent to
199 `-Bdynamic'. This option may be used any number of times.
202 Adds AUDITLIB to the `DT_AUDIT' entry of the dynamic section.
203 AUDITLIB is not checked for existence, nor will it use the
204 DT_SONAME specified in the library. If specified multiple times
205 `DT_AUDIT' will contain a colon separated list of audit interfaces
206 to use. If the linker finds an object with an audit entry while
207 searching for shared libraries, it will add a corresponding
208 `DT_DEPAUDIT' entry in the output file. This option is only
209 meaningful on ELF platforms supporting the rtld-audit interface.
212 `--architecture=ARCHITECTURE'
213 In the current release of `ld', this option is useful only for the
214 Intel 960 family of architectures. In that `ld' configuration, the
215 ARCHITECTURE argument identifies the particular architecture in
216 the 960 family, enabling some safeguards and modifying the
217 archive-library search path. *Note `ld' and the Intel 960 family:
220 Future releases of `ld' may support similar functionality for
221 other architecture families.
224 `--format=INPUT-FORMAT'
225 `ld' may be configured to support more than one kind of object
226 file. If your `ld' is configured this way, you can use the `-b'
227 option to specify the binary format for input object files that
228 follow this option on the command line. Even when `ld' is
229 configured to support alternative object formats, you don't
230 usually need to specify this, as `ld' should be configured to
231 expect as a default input format the most usual format on each
232 machine. INPUT-FORMAT is a text string, the name of a particular
233 format supported by the BFD libraries. (You can list the
234 available binary formats with `objdump -i'.) *Note BFD::.
236 You may want to use this option if you are linking files with an
237 unusual binary format. You can also use `-b' to switch formats
238 explicitly (when linking object files of different formats), by
239 including `-b INPUT-FORMAT' before each group of object files in a
242 The default format is taken from the environment variable
243 `GNUTARGET'. *Note Environment::. You can also define the input
244 format from a script, using the command `TARGET'; see *Note Format
248 `--mri-script=MRI-COMMANDFILE'
249 For compatibility with linkers produced by MRI, `ld' accepts script
250 files written in an alternate, restricted command language,
251 described in *Note MRI Compatible Script Files: MRI. Introduce
252 MRI script files with the option `-c'; use the `-T' option to run
253 linker scripts written in the general-purpose `ld' scripting
254 language. If MRI-CMDFILE does not exist, `ld' looks for it in the
255 directories specified by any `-L' options.
260 These three options are equivalent; multiple forms are supported
261 for compatibility with other linkers. They assign space to common
262 symbols even if a relocatable output file is specified (with
263 `-r'). The script command `FORCE_COMMON_ALLOCATION' has the same
264 effect. *Note Miscellaneous Commands::.
266 `--depaudit AUDITLIB'
268 Adds AUDITLIB to the `DT_DEPAUDIT' entry of the dynamic section.
269 AUDITLIB is not checked for existence, nor will it use the
270 DT_SONAME specified in the library. If specified multiple times
271 `DT_DEPAUDIT' will contain a colon separated list of audit
272 interfaces to use. This option is only meaningful on ELF
273 platforms supporting the rtld-audit interface. The -P option is
274 provided for Solaris compatibility.
278 Use ENTRY as the explicit symbol for beginning execution of your
279 program, rather than the default entry point. If there is no
280 symbol named ENTRY, the linker will try to parse ENTRY as a number,
281 and use that as the entry address (the number will be interpreted
282 in base 10; you may use a leading `0x' for base 16, or a leading
283 `0' for base 8). *Note Entry Point::, for a discussion of defaults
284 and other ways of specifying the entry point.
286 `--exclude-libs LIB,LIB,...'
287 Specifies a list of archive libraries from which symbols should
288 not be automatically exported. The library names may be delimited
289 by commas or colons. Specifying `--exclude-libs ALL' excludes
290 symbols in all archive libraries from automatic export. This
291 option is available only for the i386 PE targeted port of the
292 linker and for ELF targeted ports. For i386 PE, symbols
293 explicitly listed in a .def file are still exported, regardless of
294 this option. For ELF targeted ports, symbols affected by this
295 option will be treated as hidden.
297 `--exclude-modules-for-implib MODULE,MODULE,...'
298 Specifies a list of object files or archive members, from which
299 symbols should not be automatically exported, but which should be
300 copied wholesale into the import library being generated during
301 the link. The module names may be delimited by commas or colons,
302 and must match exactly the filenames used by `ld' to open the
303 files; for archive members, this is simply the member name, but
304 for object files the name listed must include and match precisely
305 any path used to specify the input file on the linker's
306 command-line. This option is available only for the i386 PE
307 targeted port of the linker. Symbols explicitly listed in a .def
308 file are still exported, regardless of this option.
312 `--no-export-dynamic'
313 When creating a dynamically linked executable, using the `-E'
314 option or the `--export-dynamic' option causes the linker to add
315 all symbols to the dynamic symbol table. The dynamic symbol table
316 is the set of symbols which are visible from dynamic objects at
319 If you do not use either of these options (or use the
320 `--no-export-dynamic' option to restore the default behavior), the
321 dynamic symbol table will normally contain only those symbols
322 which are referenced by some dynamic object mentioned in the link.
324 If you use `dlopen' to load a dynamic object which needs to refer
325 back to the symbols defined by the program, rather than some other
326 dynamic object, then you will probably need to use this option when
327 linking the program itself.
329 You can also use the dynamic list to control what symbols should
330 be added to the dynamic symbol table if the output format supports
331 it. See the description of `--dynamic-list'.
333 Note that this option is specific to ELF targeted ports. PE
334 targets support a similar function to export all symbols from a
335 DLL or EXE; see the description of `--export-all-symbols' below.
338 Link big-endian objects. This affects the default output format.
341 Link little-endian objects. This affects the default output
346 When creating an ELF shared object, set the internal DT_AUXILIARY
347 field to the specified name. This tells the dynamic linker that
348 the symbol table of the shared object should be used as an
349 auxiliary filter on the symbol table of the shared object NAME.
351 If you later link a program against this filter object, then, when
352 you run the program, the dynamic linker will see the DT_AUXILIARY
353 field. If the dynamic linker resolves any symbols from the filter
354 object, it will first check whether there is a definition in the
355 shared object NAME. If there is one, it will be used instead of
356 the definition in the filter object. The shared object NAME need
357 not exist. Thus the shared object NAME may be used to provide an
358 alternative implementation of certain functions, perhaps for
359 debugging or for machine specific performance.
361 This option may be specified more than once. The DT_AUXILIARY
362 entries will be created in the order in which they appear on the
367 When creating an ELF shared object, set the internal DT_FILTER
368 field to the specified name. This tells the dynamic linker that
369 the symbol table of the shared object which is being created
370 should be used as a filter on the symbol table of the shared
373 If you later link a program against this filter object, then, when
374 you run the program, the dynamic linker will see the DT_FILTER
375 field. The dynamic linker will resolve symbols according to the
376 symbol table of the filter object as usual, but it will actually
377 link to the definitions found in the shared object NAME. Thus the
378 filter object can be used to select a subset of the symbols
379 provided by the object NAME.
381 Some older linkers used the `-F' option throughout a compilation
382 toolchain for specifying object-file format for both input and
383 output object files. The GNU linker uses other mechanisms for
384 this purpose: the `-b', `--format', `--oformat' options, the
385 `TARGET' command in linker scripts, and the `GNUTARGET'
386 environment variable. The GNU linker will ignore the `-F' option
387 when not creating an ELF shared object.
390 When creating an ELF executable or shared object, call NAME when
391 the executable or shared object is unloaded, by setting DT_FINI to
392 the address of the function. By default, the linker uses `_fini'
393 as the function to call.
396 Ignored. Provided for compatibility with other tools.
400 Set the maximum size of objects to be optimized using the GP
401 register to SIZE. This is only meaningful for object file formats
402 such as MIPS ECOFF which supports putting large and small objects
403 into different sections. This is ignored for other object file
408 When creating an ELF shared object, set the internal DT_SONAME
409 field to the specified name. When an executable is linked with a
410 shared object which has a DT_SONAME field, then when the
411 executable is run the dynamic linker will attempt to load the
412 shared object specified by the DT_SONAME field rather than the
413 using the file name given to the linker.
416 Perform an incremental link (same as option `-r').
419 When creating an ELF executable or shared object, call NAME when
420 the executable or shared object is loaded, by setting DT_INIT to
421 the address of the function. By default, the linker uses `_init'
422 as the function to call.
426 Add the archive or object file specified by NAMESPEC to the list
427 of files to link. This option may be used any number of times.
428 If NAMESPEC is of the form `:FILENAME', `ld' will search the
429 library path for a file called FILENAME, otherwise it will search
430 the library path for a file called `libNAMESPEC.a'.
432 On systems which support shared libraries, `ld' may also search for
433 files other than `libNAMESPEC.a'. Specifically, on ELF and SunOS
434 systems, `ld' will search a directory for a library called
435 `libNAMESPEC.so' before searching for one called `libNAMESPEC.a'.
436 (By convention, a `.so' extension indicates a shared library.)
437 Note that this behavior does not apply to `:FILENAME', which
438 always specifies a file called FILENAME.
440 The linker will search an archive only once, at the location where
441 it is specified on the command line. If the archive defines a
442 symbol which was undefined in some object which appeared before
443 the archive on the command line, the linker will include the
444 appropriate file(s) from the archive. However, an undefined
445 symbol in an object appearing later on the command line will not
446 cause the linker to search the archive again.
448 See the `-(' option for a way to force the linker to search
449 archives multiple times.
451 You may list the same archive multiple times on the command line.
453 This type of archive searching is standard for Unix linkers.
454 However, if you are using `ld' on AIX, note that it is different
455 from the behaviour of the AIX linker.
458 `--library-path=SEARCHDIR'
459 Add path SEARCHDIR to the list of paths that `ld' will search for
460 archive libraries and `ld' control scripts. You may use this
461 option any number of times. The directories are searched in the
462 order in which they are specified on the command line.
463 Directories specified on the command line are searched before the
464 default directories. All `-L' options apply to all `-l' options,
465 regardless of the order in which the options appear. `-L' options
466 do not affect how `ld' searches for a linker script unless `-T'
469 If SEARCHDIR begins with `=', then the `=' will be replaced by the
470 "sysroot prefix", a path specified when the linker is configured.
472 The default set of paths searched (without being specified with
473 `-L') depends on which emulation mode `ld' is using, and in some
474 cases also on how it was configured. *Note Environment::.
476 The paths can also be specified in a link script with the
477 `SEARCH_DIR' command. Directories specified this way are searched
478 at the point in which the linker script appears in the command
482 Emulate the EMULATION linker. You can list the available
483 emulations with the `--verbose' or `-V' options.
485 If the `-m' option is not used, the emulation is taken from the
486 `LDEMULATION' environment variable, if that is defined.
488 Otherwise, the default emulation depends upon how the linker was
493 Print a link map to the standard output. A link map provides
494 information about the link, including the following:
496 * Where object files are mapped into memory.
498 * How common symbols are allocated.
500 * All archive members included in the link, with a mention of
501 the symbol which caused the archive member to be brought in.
503 * The values assigned to symbols.
505 Note - symbols whose values are computed by an expression
506 which involves a reference to a previous value of the same
507 symbol may not have correct result displayed in the link map.
508 This is because the linker discards intermediate results and
509 only retains the final value of an expression. Under such
510 circumstances the linker will display the final value
511 enclosed by square brackets. Thus for example a linker
518 will produce the following output in the link map if the `-M'
522 [0x0000000c] foo = (foo * 0x4)
523 [0x0000000c] foo = (foo + 0x8)
525 See *Note Expressions:: for more information about
526 expressions in linker scripts.
530 Turn off page alignment of sections, and disable linking against
531 shared libraries. If the output format supports Unix style magic
532 numbers, mark the output as `NMAGIC'.
536 Set the text and data sections to be readable and writable. Also,
537 do not page-align the data segment, and disable linking against
538 shared libraries. If the output format supports Unix style magic
539 numbers, mark the output as `OMAGIC'. Note: Although a writable
540 text section is allowed for PE-COFF targets, it does not conform
541 to the format specification published by Microsoft.
544 This option negates most of the effects of the `-N' option. It
545 sets the text section to be read-only, and forces the data segment
546 to be page-aligned. Note - this option does not enable linking
547 against shared libraries. Use `-Bdynamic' for this.
551 Use OUTPUT as the name for the program produced by `ld'; if this
552 option is not specified, the name `a.out' is used by default. The
553 script command `OUTPUT' can also specify the output file name.
556 If LEVEL is a numeric values greater than zero `ld' optimizes the
557 output. This might take significantly longer and therefore
558 probably should only be enabled for the final binary. At the
559 moment this option only affects ELF shared library generation.
560 Future releases of the linker may make more use of this option.
561 Also currently there is no difference in the linker's behaviour
562 for different non-zero values of this option. Again this may
563 change with future releases.
567 Leave relocation sections and contents in fully linked executables.
568 Post link analysis and optimization tools may need this
569 information in order to perform correct modifications of
570 executables. This results in larger executables.
572 This option is currently only supported on ELF platforms.
575 Force the output file to have dynamic sections. This option is
576 specific to VxWorks targets.
580 Generate relocatable output--i.e., generate an output file that
581 can in turn serve as input to `ld'. This is often called "partial
582 linking". As a side effect, in environments that support standard
583 Unix magic numbers, this option also sets the output file's magic
584 number to `OMAGIC'. If this option is not specified, an absolute
585 file is produced. When linking C++ programs, this option _will
586 not_ resolve references to constructors; to do that, use `-Ur'.
588 When an input file does not have the same format as the output
589 file, partial linking is only supported if that input file does
590 not contain any relocations. Different output formats can have
591 further restrictions; for example some `a.out'-based formats do
592 not support partial linking with input files in other formats at
595 This option does the same thing as `-i'.
598 `--just-symbols=FILENAME'
599 Read symbol names and their addresses from FILENAME, but do not
600 relocate it or include it in the output. This allows your output
601 file to refer symbolically to absolute locations of memory defined
602 in other programs. You may use this option more than once.
604 For compatibility with other ELF linkers, if the `-R' option is
605 followed by a directory name, rather than a file name, it is
606 treated as the `-rpath' option.
610 Omit all symbol information from the output file.
614 Omit debugger symbol information (but not all symbols) from the
619 Print the names of the input files as `ld' processes them.
622 `--script=SCRIPTFILE'
623 Use SCRIPTFILE as the linker script. This script replaces `ld''s
624 default linker script (rather than adding to it), so COMMANDFILE
625 must specify everything necessary to describe the output file.
626 *Note Scripts::. If SCRIPTFILE does not exist in the current
627 directory, `ld' looks for it in the directories specified by any
628 preceding `-L' options. Multiple `-T' options accumulate.
631 `--default-script=SCRIPTFILE'
632 Use SCRIPTFILE as the default linker script. *Note Scripts::.
634 This option is similar to the `--script' option except that
635 processing of the script is delayed until after the rest of the
636 command line has been processed. This allows options placed after
637 the `--default-script' option on the command line to affect the
638 behaviour of the linker script, which can be important when the
639 linker command line cannot be directly controlled by the user.
640 (eg because the command line is being constructed by another tool,
645 Force SYMBOL to be entered in the output file as an undefined
646 symbol. Doing this may, for example, trigger linking of additional
647 modules from standard libraries. `-u' may be repeated with
648 different option arguments to enter additional undefined symbols.
649 This option is equivalent to the `EXTERN' linker script command.
652 For anything other than C++ programs, this option is equivalent to
653 `-r': it generates relocatable output--i.e., an output file that
654 can in turn serve as input to `ld'. When linking C++ programs,
655 `-Ur' _does_ resolve references to constructors, unlike `-r'. It
656 does not work to use `-Ur' on files that were themselves linked
657 with `-Ur'; once the constructor table has been built, it cannot
658 be added to. Use `-Ur' only for the last partial link, and `-r'
662 Creates a separate output section for every input section matching
663 SECTION, or if the optional wildcard SECTION argument is missing,
664 for every orphan input section. An orphan section is one not
665 specifically mentioned in a linker script. You may use this option
666 multiple times on the command line; It prevents the normal
667 merging of input sections with the same name, overriding output
668 section assignments in a linker script.
673 Display the version number for `ld'. The `-V' option also lists
674 the supported emulations.
678 Delete all local symbols.
682 Delete all temporary local symbols. (These symbols start with
683 system-specific local label prefixes, typically `.L' for ELF
684 systems or `L' for traditional a.out systems.)
687 `--trace-symbol=SYMBOL'
688 Print the name of each linked file in which SYMBOL appears. This
689 option may be given any number of times. On many systems it is
690 necessary to prepend an underscore.
692 This option is useful when you have an undefined symbol in your
693 link but don't know where the reference is coming from.
696 Add PATH to the default library search path. This option exists
697 for Solaris compatibility.
700 The recognized keywords are:
702 Combines multiple reloc sections and sorts them to make
703 dynamic symbol lookup caching possible.
706 Disallows undefined symbols in object files. Undefined
707 symbols in shared libraries are still allowed.
710 Marks the object as requiring executable stack.
713 This option is only meaningful when building a shared object.
714 It marks the object so that its runtime initialization will
715 occur before the runtime initialization of any other objects
716 brought into the process at the same time. Similarly the
717 runtime finalization of the object will occur after the
718 runtime finalization of any other objects.
721 Marks the object that its symbol table interposes before all
722 symbols but the primary executable.
725 When generating an executable or shared library, mark it to
726 tell the dynamic linker to defer function call resolution to
727 the point when the function is called (lazy binding), rather
728 than at load time. Lazy binding is the default.
731 Marks the object that its filters be processed immediately at
735 Allows multiple definitions.
738 Disables multiple reloc sections combining.
741 Disables production of copy relocs.
744 Marks the object that the search for dependencies of this
745 object will ignore any default library search paths.
748 Marks the object shouldn't be unloaded at runtime.
751 Marks the object not available to `dlopen'.
754 Marks the object can not be dumped by `dldump'.
757 Marks the object as not requiring executable stack.
760 Don't create an ELF `PT_GNU_RELRO' segment header in the
764 When generating an executable or shared library, mark it to
765 tell the dynamic linker to resolve all symbols when the
766 program is started, or when the shared library is linked to
767 using dlopen, instead of deferring function call resolution
768 to the point when the function is first called.
771 Marks the object may contain $ORIGIN.
774 Create an ELF `PT_GNU_RELRO' segment header in the object.
776 `max-page-size=VALUE'
777 Set the emulation maximum page size to VALUE.
779 `common-page-size=VALUE'
780 Set the emulation common page size to VALUE.
783 Other keywords are ignored for Solaris compatibility.
786 `--start-group ARCHIVES --end-group'
787 The ARCHIVES should be a list of archive files. They may be
788 either explicit file names, or `-l' options.
790 The specified archives are searched repeatedly until no new
791 undefined references are created. Normally, an archive is
792 searched only once in the order that it is specified on the
793 command line. If a symbol in that archive is needed to resolve an
794 undefined symbol referred to by an object in an archive that
795 appears later on the command line, the linker would not be able to
796 resolve that reference. By grouping the archives, they all be
797 searched repeatedly until all possible references are resolved.
799 Using this option has a significant performance cost. It is best
800 to use it only when there are unavoidable circular references
801 between two or more archives.
803 `--accept-unknown-input-arch'
804 `--no-accept-unknown-input-arch'
805 Tells the linker to accept input files whose architecture cannot be
806 recognised. The assumption is that the user knows what they are
807 doing and deliberately wants to link in these unknown input files.
808 This was the default behaviour of the linker, before release
809 2.14. The default behaviour from release 2.14 onwards is to
810 reject such input files, and so the `--accept-unknown-input-arch'
811 option has been added to restore the old behaviour.
815 This option affects ELF DT_NEEDED tags for dynamic libraries
816 mentioned on the command line after the `--as-needed' option.
817 Normally the linker will add a DT_NEEDED tag for each dynamic
818 library mentioned on the command line, regardless of whether the
819 library is actually needed or not. `--as-needed' causes a
820 DT_NEEDED tag to only be emitted for a library that satisfies an
821 undefined symbol reference from a regular object file or, if the
822 library is not found in the DT_NEEDED lists of other libraries
823 linked up to that point, an undefined symbol reference from
824 another dynamic library. `--no-as-needed' restores the default
829 These two options have been deprecated because of the similarity of
830 their names to the `--as-needed' and `--no-as-needed' options.
831 They have been replaced by `--copy-dt-needed-entries' and
832 `--no-copy-dt-needed-entries'.
835 This option is ignored for SunOS compatibility.
840 Link against dynamic libraries. This is only meaningful on
841 platforms for which shared libraries are supported. This option
842 is normally the default on such platforms. The different variants
843 of this option are for compatibility with various systems. You
844 may use this option multiple times on the command line: it affects
845 library searching for `-l' options which follow it.
848 Set the `DF_1_GROUP' flag in the `DT_FLAGS_1' entry in the dynamic
849 section. This causes the runtime linker to handle lookups in this
850 object and its dependencies to be performed only inside the group.
851 `--unresolved-symbols=report-all' is implied. This option is only
852 meaningful on ELF platforms which support shared libraries.
858 Do not link against shared libraries. This is only meaningful on
859 platforms for which shared libraries are supported. The different
860 variants of this option are for compatibility with various
861 systems. You may use this option multiple times on the command
862 line: it affects library searching for `-l' options which follow
863 it. This option also implies `--unresolved-symbols=report-all'.
864 This option can be used with `-shared'. Doing so means that a
865 shared library is being created but that all of the library's
866 external references must be resolved by pulling in entries from
870 When creating a shared library, bind references to global symbols
871 to the definition within the shared library, if any. Normally, it
872 is possible for a program linked against a shared library to
873 override the definition within the shared library. This option is
874 only meaningful on ELF platforms which support shared libraries.
876 `-Bsymbolic-functions'
877 When creating a shared library, bind references to global function
878 symbols to the definition within the shared library, if any. This
879 option is only meaningful on ELF platforms which support shared
882 `--dynamic-list=DYNAMIC-LIST-FILE'
883 Specify the name of a dynamic list file to the linker. This is
884 typically used when creating shared libraries to specify a list of
885 global symbols whose references shouldn't be bound to the
886 definition within the shared library, or creating dynamically
887 linked executables to specify a list of symbols which should be
888 added to the symbol table in the executable. This option is only
889 meaningful on ELF platforms which support shared libraries.
891 The format of the dynamic list is the same as the version node
892 without scope and node name. See *Note VERSION:: for more
895 `--dynamic-list-data'
896 Include all global data symbols to the dynamic list.
898 `--dynamic-list-cpp-new'
899 Provide the builtin dynamic list for C++ operator new and delete.
900 It is mainly useful for building shared libstdc++.
902 `--dynamic-list-cpp-typeinfo'
903 Provide the builtin dynamic list for C++ runtime type
907 `--no-check-sections'
908 Asks the linker _not_ to check section addresses after they have
909 been assigned to see if there are any overlaps. Normally the
910 linker will perform this check, and if it finds any overlaps it
911 will produce suitable error messages. The linker does know about,
912 and does make allowances for sections in overlays. The default
913 behaviour can be restored by using the command line switch
914 `--check-sections'. Section overlap is not usually checked for
915 relocatable links. You can force checking in that case by using
916 the `--check-sections' option.
918 `--copy-dt-needed-entries'
919 `--no-copy-dt-needed-entries'
920 This option affects the treatment of dynamic libraries referred to
921 by DT_NEEDED tags _inside_ ELF dynamic libraries mentioned on the
922 command line. Normally the linker won't add a DT_NEEDED tag to the
923 output binary for each library mentioned in a DT_NEEDED tag in an
924 input dynamic library. With `--copy-dt-needed-entries' specified
925 on the command line however any dynamic libraries that follow it
926 will have their DT_NEEDED entries added. The default behaviour
927 can be restored with `--no-copy-dt-needed-entries'.
929 This option also has an effect on the resolution of symbols in
930 dynamic libraries. With `--copy-dt-needed-entries' dynamic
931 libraries mentioned on the command line will be recursively
932 searched, following their DT_NEEDED tags to other libraries, in
933 order to resolve symbols required by the output binary. With the
934 default setting however the searching of dynamic libraries that
935 follow it will stop with the dynamic library itself. No DT_NEEDED
936 links will be traversed to resolve symbols.
939 Output a cross reference table. If a linker map file is being
940 generated, the cross reference table is printed to the map file.
941 Otherwise, it is printed on the standard output.
943 The format of the table is intentionally simple, so that it may be
944 easily processed by a script if necessary. The symbols are
945 printed out, sorted by name. For each symbol, a list of file
946 names is given. If the symbol is defined, the first file listed
947 is the location of the definition. The remaining files contain
948 references to the symbol.
951 This option inhibits the assignment of addresses to common symbols.
952 The script command `INHIBIT_COMMON_ALLOCATION' has the same effect.
953 *Note Miscellaneous Commands::.
955 The `--no-define-common' option allows decoupling the decision to
956 assign addresses to Common symbols from the choice of the output
957 file type; otherwise a non-Relocatable output type forces
958 assigning addresses to Common symbols. Using `--no-define-common'
959 allows Common symbols that are referenced from a shared library to
960 be assigned addresses only in the main program. This eliminates
961 the unused duplicate space in the shared library, and also
962 prevents any possible confusion over resolving to the wrong
963 duplicate when there are many dynamic modules with specialized
964 search paths for runtime symbol resolution.
966 `--defsym=SYMBOL=EXPRESSION'
967 Create a global symbol in the output file, containing the absolute
968 address given by EXPRESSION. You may use this option as many
969 times as necessary to define multiple symbols in the command line.
970 A limited form of arithmetic is supported for the EXPRESSION in
971 this context: you may give a hexadecimal constant or the name of
972 an existing symbol, or use `+' and `-' to add or subtract
973 hexadecimal constants or symbols. If you need more elaborate
974 expressions, consider using the linker command language from a
975 script (*note Assignment: Symbol Definitions: Assignments.).
976 _Note:_ there should be no white space between SYMBOL, the equals
977 sign ("<=>"), and EXPRESSION.
981 These options control whether to demangle symbol names in error
982 messages and other output. When the linker is told to demangle,
983 it tries to present symbol names in a readable fashion: it strips
984 leading underscores if they are used by the object file format,
985 and converts C++ mangled symbol names into user readable names.
986 Different compilers have different mangling styles. The optional
987 demangling style argument can be used to choose an appropriate
988 demangling style for your compiler. The linker will demangle by
989 default unless the environment variable `COLLECT_NO_DEMANGLE' is
990 set. These options may be used to override the default.
993 `--dynamic-linker=FILE'
994 Set the name of the dynamic linker. This is only meaningful when
995 generating dynamically linked ELF executables. The default dynamic
996 linker is normally correct; don't use this unless you know what
1000 `--no-fatal-warnings'
1001 Treat all warnings as errors. The default behaviour can be
1002 restored with the option `--no-fatal-warnings'.
1004 `--force-exe-suffix'
1005 Make sure that an output file has a .exe suffix.
1007 If a successfully built fully linked output file does not have a
1008 `.exe' or `.dll' suffix, this option forces the linker to copy the
1009 output file to one of the same name with a `.exe' suffix. This
1010 option is useful when using unmodified Unix makefiles on a
1011 Microsoft Windows host, since some versions of Windows won't run
1012 an image unless it ends in a `.exe' suffix.
1016 Enable garbage collection of unused input sections. It is ignored
1017 on targets that do not support this option. The default behaviour
1018 (of not performing this garbage collection) can be restored by
1019 specifying `--no-gc-sections' on the command line.
1021 `--gc-sections' decides which input sections are used by examining
1022 symbols and relocations. The section containing the entry symbol
1023 and all sections containing symbols undefined on the command-line
1024 will be kept, as will sections containing symbols referenced by
1025 dynamic objects. Note that when building shared libraries, the
1026 linker must assume that any visible symbol is referenced. Once
1027 this initial set of sections has been determined, the linker
1028 recursively marks as used any section referenced by their
1029 relocations. See `--entry' and `--undefined'.
1031 This option can be set when doing a partial link (enabled with
1032 option `-r'). In this case the root of symbols kept must be
1033 explicitly specified either by an `--entry' or `--undefined'
1034 option or by a `ENTRY' command in the linker script.
1036 `--print-gc-sections'
1037 `--no-print-gc-sections'
1038 List all sections removed by garbage collection. The listing is
1039 printed on stderr. This option is only effective if garbage
1040 collection has been enabled via the `--gc-sections') option. The
1041 default behaviour (of not listing the sections that are removed)
1042 can be restored by specifying `--no-print-gc-sections' on the
1045 `--print-output-format'
1046 Print the name of the default output format (perhaps influenced by
1047 other command-line options). This is the string that would appear
1048 in an `OUTPUT_FORMAT' linker script command (*note File
1052 Print a summary of the command-line options on the standard output
1056 Print a summary of all target specific options on the standard
1060 Print a link map to the file MAPFILE. See the description of the
1064 `ld' normally optimizes for speed over memory usage by caching the
1065 symbol tables of input files in memory. This option tells `ld' to
1066 instead optimize for memory usage, by rereading the symbol tables
1067 as necessary. This may be required if `ld' runs out of memory
1068 space while linking a large executable.
1072 Report unresolved symbol references from regular object files.
1073 This is done even if the linker is creating a non-symbolic shared
1074 library. The switch `--[no-]allow-shlib-undefined' controls the
1075 behaviour for reporting unresolved references found in shared
1076 libraries being linked in.
1078 `--allow-multiple-definition'
1080 Normally when a symbol is defined multiple times, the linker will
1081 report a fatal error. These options allow multiple definitions and
1082 the first definition will be used.
1084 `--allow-shlib-undefined'
1085 `--no-allow-shlib-undefined'
1086 Allows or disallows undefined symbols in shared libraries. This
1087 switch is similar to `--no-undefined' except that it determines
1088 the behaviour when the undefined symbols are in a shared library
1089 rather than a regular object file. It does not affect how
1090 undefined symbols in regular object files are handled.
1092 The default behaviour is to report errors for any undefined symbols
1093 referenced in shared libraries if the linker is being used to
1094 create an executable, but to allow them if the linker is being
1095 used to create a shared library.
1097 The reasons for allowing undefined symbol references in shared
1098 libraries specified at link time are that:
1100 * A shared library specified at link time may not be the same
1101 as the one that is available at load time, so the symbol
1102 might actually be resolvable at load time.
1104 * There are some operating systems, eg BeOS and HPPA, where
1105 undefined symbols in shared libraries are normal.
1107 The BeOS kernel for example patches shared libraries at load
1108 time to select whichever function is most appropriate for the
1109 current architecture. This is used, for example, to
1110 dynamically select an appropriate memset function.
1112 `--no-undefined-version'
1113 Normally when a symbol has an undefined version, the linker will
1114 ignore it. This option disallows symbols with undefined version
1115 and a fatal error will be issued instead.
1118 Create and use a default symbol version (the soname) for
1119 unversioned exported symbols.
1121 `--default-imported-symver'
1122 Create and use a default symbol version (the soname) for
1123 unversioned imported symbols.
1125 `--no-warn-mismatch'
1126 Normally `ld' will give an error if you try to link together input
1127 files that are mismatched for some reason, perhaps because they
1128 have been compiled for different processors or for different
1129 endiannesses. This option tells `ld' that it should silently
1130 permit such possible errors. This option should only be used with
1131 care, in cases when you have taken some special action that
1132 ensures that the linker errors are inappropriate.
1134 `--no-warn-search-mismatch'
1135 Normally `ld' will give a warning if it finds an incompatible
1136 library during a library search. This option silences the warning.
1138 `--no-whole-archive'
1139 Turn off the effect of the `--whole-archive' option for subsequent
1143 Retain the executable output file whenever it is still usable.
1144 Normally, the linker will not produce an output file if it
1145 encounters errors during the link process; it exits without
1146 writing an output file when it issues any error whatsoever.
1149 Only search library directories explicitly specified on the
1150 command line. Library directories specified in linker scripts
1151 (including linker scripts specified on the command line) are
1154 `--oformat=OUTPUT-FORMAT'
1155 `ld' may be configured to support more than one kind of object
1156 file. If your `ld' is configured this way, you can use the
1157 `--oformat' option to specify the binary format for the output
1158 object file. Even when `ld' is configured to support alternative
1159 object formats, you don't usually need to specify this, as `ld'
1160 should be configured to produce as a default output format the most
1161 usual format on each machine. OUTPUT-FORMAT is a text string, the
1162 name of a particular format supported by the BFD libraries. (You
1163 can list the available binary formats with `objdump -i'.) The
1164 script command `OUTPUT_FORMAT' can also specify the output format,
1165 but this option overrides it. *Note BFD::.
1169 Create a position independent executable. This is currently only
1170 supported on ELF platforms. Position independent executables are
1171 similar to shared libraries in that they are relocated by the
1172 dynamic linker to the virtual address the OS chooses for them
1173 (which can vary between invocations). Like normal dynamically
1174 linked executables they can be executed and symbols defined in the
1175 executable cannot be overridden by shared libraries.
1178 This option is ignored for Linux compatibility.
1181 This option is ignored for SVR4 compatibility.
1185 An option with machine dependent effects. This option is only
1186 supported on a few targets. *Note `ld' and the H8/300: H8/300.
1187 *Note `ld' and the Intel 960 family: i960. *Note `ld' and Xtensa
1188 Processors: Xtensa. *Note `ld' and the 68HC11 and 68HC12:
1189 M68HC11/68HC12. *Note `ld' and PowerPC 32-bit ELF Support:
1192 On some platforms the `--relax' option performs target specific,
1193 global optimizations that become possible when the linker resolves
1194 addressing in the program, such as relaxing address modes,
1195 synthesizing new instructions, selecting shorter version of current
1196 instructions, and combinig constant values.
1198 On some platforms these link time global optimizations may make
1199 symbolic debugging of the resulting executable impossible. This
1200 is known to be the case for the Matsushita MN10200 and MN10300
1201 family of processors.
1203 On platforms where this is not supported, `--relax' is accepted,
1206 On platforms where `--relax' is accepted the option `--no-relax'
1207 can be used to disable the feature.
1209 `--retain-symbols-file=FILENAME'
1210 Retain _only_ the symbols listed in the file FILENAME, discarding
1211 all others. FILENAME is simply a flat file, with one symbol name
1212 per line. This option is especially useful in environments (such
1213 as VxWorks) where a large global symbol table is accumulated
1214 gradually, to conserve run-time memory.
1216 `--retain-symbols-file' does _not_ discard undefined symbols, or
1217 symbols needed for relocations.
1219 You may only specify `--retain-symbols-file' once in the command
1220 line. It overrides `-s' and `-S'.
1223 Add a directory to the runtime library search path. This is used
1224 when linking an ELF executable with shared objects. All `-rpath'
1225 arguments are concatenated and passed to the runtime linker, which
1226 uses them to locate shared objects at runtime. The `-rpath'
1227 option is also used when locating shared objects which are needed
1228 by shared objects explicitly included in the link; see the
1229 description of the `-rpath-link' option. If `-rpath' is not used
1230 when linking an ELF executable, the contents of the environment
1231 variable `LD_RUN_PATH' will be used if it is defined.
1233 The `-rpath' option may also be used on SunOS. By default, on
1234 SunOS, the linker will form a runtime search patch out of all the
1235 `-L' options it is given. If a `-rpath' option is used, the
1236 runtime search path will be formed exclusively using the `-rpath'
1237 options, ignoring the `-L' options. This can be useful when using
1238 gcc, which adds many `-L' options which may be on NFS mounted file
1241 For compatibility with other ELF linkers, if the `-R' option is
1242 followed by a directory name, rather than a file name, it is
1243 treated as the `-rpath' option.
1246 When using ELF or SunOS, one shared library may require another.
1247 This happens when an `ld -shared' link includes a shared library
1248 as one of the input files.
1250 When the linker encounters such a dependency when doing a
1251 non-shared, non-relocatable link, it will automatically try to
1252 locate the required shared library and include it in the link, if
1253 it is not included explicitly. In such a case, the `-rpath-link'
1254 option specifies the first set of directories to search. The
1255 `-rpath-link' option may specify a sequence of directory names
1256 either by specifying a list of names separated by colons, or by
1257 appearing multiple times.
1259 This option should be used with caution as it overrides the search
1260 path that may have been hard compiled into a shared library. In
1261 such a case it is possible to use unintentionally a different
1262 search path than the runtime linker would do.
1264 The linker uses the following search paths to locate required
1266 1. Any directories specified by `-rpath-link' options.
1268 2. Any directories specified by `-rpath' options. The difference
1269 between `-rpath' and `-rpath-link' is that directories
1270 specified by `-rpath' options are included in the executable
1271 and used at runtime, whereas the `-rpath-link' option is only
1272 effective at link time. Searching `-rpath' in this way is
1273 only supported by native linkers and cross linkers which have
1274 been configured with the `--with-sysroot' option.
1276 3. On an ELF system, for native linkers, if the `-rpath' and
1277 `-rpath-link' options were not used, search the contents of
1278 the environment variable `LD_RUN_PATH'.
1280 4. On SunOS, if the `-rpath' option was not used, search any
1281 directories specified using `-L' options.
1283 5. For a native linker, the search the contents of the
1284 environment variable `LD_LIBRARY_PATH'.
1286 6. For a native ELF linker, the directories in `DT_RUNPATH' or
1287 `DT_RPATH' of a shared library are searched for shared
1288 libraries needed by it. The `DT_RPATH' entries are ignored if
1289 `DT_RUNPATH' entries exist.
1291 7. The default directories, normally `/lib' and `/usr/lib'.
1293 8. For a native linker on an ELF system, if the file
1294 `/etc/ld.so.conf' exists, the list of directories found in
1297 If the required shared library is not found, the linker will issue
1298 a warning and continue with the link.
1302 Create a shared library. This is currently only supported on ELF,
1303 XCOFF and SunOS platforms. On SunOS, the linker will
1304 automatically create a shared library if the `-e' option is not
1305 used and there are undefined symbols in the link.
1308 `--sort-common=ascending'
1309 `--sort-common=descending'
1310 This option tells `ld' to sort the common symbols by alignment in
1311 ascending or descending order when it places them in the
1312 appropriate output sections. The symbol alignments considered are
1313 sixteen-byte or larger, eight-byte, four-byte, two-byte, and
1314 one-byte. This is to prevent gaps between symbols due to alignment
1315 constraints. If no sorting order is specified, then descending
1318 `--sort-section=name'
1319 This option will apply `SORT_BY_NAME' to all wildcard section
1320 patterns in the linker script.
1322 `--sort-section=alignment'
1323 This option will apply `SORT_BY_ALIGNMENT' to all wildcard section
1324 patterns in the linker script.
1326 `--split-by-file[=SIZE]'
1327 Similar to `--split-by-reloc' but creates a new output section for
1328 each input file when SIZE is reached. SIZE defaults to a size of
1331 `--split-by-reloc[=COUNT]'
1332 Tries to creates extra sections in the output file so that no
1333 single output section in the file contains more than COUNT
1334 relocations. This is useful when generating huge relocatable
1335 files for downloading into certain real time kernels with the COFF
1336 object file format; since COFF cannot represent more than 65535
1337 relocations in a single section. Note that this will fail to work
1338 with object file formats which do not support arbitrary sections.
1339 The linker will not split up individual input sections for
1340 redistribution, so if a single input section contains more than
1341 COUNT relocations one output section will contain that many
1342 relocations. COUNT defaults to a value of 32768.
1345 Compute and display statistics about the operation of the linker,
1346 such as execution time and memory usage.
1348 `--sysroot=DIRECTORY'
1349 Use DIRECTORY as the location of the sysroot, overriding the
1350 configure-time default. This option is only supported by linkers
1351 that were configured using `--with-sysroot'.
1353 `--traditional-format'
1354 For some targets, the output of `ld' is different in some ways from
1355 the output of some existing linker. This switch requests `ld' to
1356 use the traditional format instead.
1358 For example, on SunOS, `ld' combines duplicate entries in the
1359 symbol string table. This can reduce the size of an output file
1360 with full debugging information by over 30 percent.
1361 Unfortunately, the SunOS `dbx' program can not read the resulting
1362 program (`gdb' has no trouble). The `--traditional-format' switch
1363 tells `ld' to not combine duplicate entries.
1365 `--section-start=SECTIONNAME=ORG'
1366 Locate a section in the output file at the absolute address given
1367 by ORG. You may use this option as many times as necessary to
1368 locate multiple sections in the command line. ORG must be a
1369 single hexadecimal integer; for compatibility with other linkers,
1370 you may omit the leading `0x' usually associated with hexadecimal
1371 values. _Note:_ there should be no white space between
1372 SECTIONNAME, the equals sign ("<=>"), and ORG.
1377 Same as `--section-start', with `.bss', `.data' or `.text' as the
1380 `-Ttext-segment=ORG'
1381 When creating an ELF executable or shared object, it will set the
1382 address of the first byte of the text segment.
1384 `--unresolved-symbols=METHOD'
1385 Determine how to handle unresolved symbols. There are four
1386 possible values for `method':
1389 Do not report any unresolved symbols.
1392 Report all unresolved symbols. This is the default.
1394 `ignore-in-object-files'
1395 Report unresolved symbols that are contained in shared
1396 libraries, but ignore them if they come from regular object
1399 `ignore-in-shared-libs'
1400 Report unresolved symbols that come from regular object
1401 files, but ignore them if they come from shared libraries.
1402 This can be useful when creating a dynamic binary and it is
1403 known that all the shared libraries that it should be
1404 referencing are included on the linker's command line.
1406 The behaviour for shared libraries on their own can also be
1407 controlled by the `--[no-]allow-shlib-undefined' option.
1409 Normally the linker will generate an error message for each
1410 reported unresolved symbol but the option
1411 `--warn-unresolved-symbols' can change this to a warning.
1414 `--verbose[=NUMBER]'
1415 Display the version number for `ld' and list the linker emulations
1416 supported. Display which input files can and cannot be opened.
1417 Display the linker script being used by the linker. If the
1418 optional NUMBER argument > 1, plugin symbol status will also be
1421 `--version-script=VERSION-SCRIPTFILE'
1422 Specify the name of a version script to the linker. This is
1423 typically used when creating shared libraries to specify
1424 additional information about the version hierarchy for the library
1425 being created. This option is only fully supported on ELF
1426 platforms which support shared libraries; see *Note VERSION::. It
1427 is partially supported on PE platforms, which can use version
1428 scripts to filter symbol visibility in auto-export mode: any
1429 symbols marked `local' in the version script will not be exported.
1433 Warn when a common symbol is combined with another common symbol
1434 or with a symbol definition. Unix linkers allow this somewhat
1435 sloppy practise, but linkers on some other operating systems do
1436 not. This option allows you to find potential problems from
1437 combining global symbols. Unfortunately, some C libraries use
1438 this practise, so you may get some warnings about symbols in the
1439 libraries as well as in your programs.
1441 There are three kinds of global symbols, illustrated here by C
1445 A definition, which goes in the initialized data section of
1449 An undefined reference, which does not allocate space. There
1450 must be either a definition or a common symbol for the
1454 A common symbol. If there are only (one or more) common
1455 symbols for a variable, it goes in the uninitialized data
1456 area of the output file. The linker merges multiple common
1457 symbols for the same variable into a single symbol. If they
1458 are of different sizes, it picks the largest size. The
1459 linker turns a common symbol into a declaration, if there is
1460 a definition of the same variable.
1462 The `--warn-common' option can produce five kinds of warnings.
1463 Each warning consists of a pair of lines: the first describes the
1464 symbol just encountered, and the second describes the previous
1465 symbol encountered with the same name. One or both of the two
1466 symbols will be a common symbol.
1468 1. Turning a common symbol into a reference, because there is
1469 already a definition for the symbol.
1470 FILE(SECTION): warning: common of `SYMBOL'
1471 overridden by definition
1472 FILE(SECTION): warning: defined here
1474 2. Turning a common symbol into a reference, because a later
1475 definition for the symbol is encountered. This is the same
1476 as the previous case, except that the symbols are encountered
1477 in a different order.
1478 FILE(SECTION): warning: definition of `SYMBOL'
1480 FILE(SECTION): warning: common is here
1482 3. Merging a common symbol with a previous same-sized common
1484 FILE(SECTION): warning: multiple common
1486 FILE(SECTION): warning: previous common is here
1488 4. Merging a common symbol with a previous larger common symbol.
1489 FILE(SECTION): warning: common of `SYMBOL'
1490 overridden by larger common
1491 FILE(SECTION): warning: larger common is here
1493 5. Merging a common symbol with a previous smaller common
1494 symbol. This is the same as the previous case, except that
1495 the symbols are encountered in a different order.
1496 FILE(SECTION): warning: common of `SYMBOL'
1497 overriding smaller common
1498 FILE(SECTION): warning: smaller common is here
1500 `--warn-constructors'
1501 Warn if any global constructors are used. This is only useful for
1502 a few object file formats. For formats like COFF or ELF, the
1503 linker can not detect the use of global constructors.
1505 `--warn-multiple-gp'
1506 Warn if multiple global pointer values are required in the output
1507 file. This is only meaningful for certain processors, such as the
1508 Alpha. Specifically, some processors put large-valued constants
1509 in a special section. A special register (the global pointer)
1510 points into the middle of this section, so that constants can be
1511 loaded efficiently via a base-register relative addressing mode.
1512 Since the offset in base-register relative mode is fixed and
1513 relatively small (e.g., 16 bits), this limits the maximum size of
1514 the constant pool. Thus, in large programs, it is often necessary
1515 to use multiple global pointer values in order to be able to
1516 address all possible constants. This option causes a warning to
1517 be issued whenever this case occurs.
1520 Only warn once for each undefined symbol, rather than once per
1521 module which refers to it.
1523 `--warn-section-align'
1524 Warn if the address of an output section is changed because of
1525 alignment. Typically, the alignment will be set by an input
1526 section. The address will only be changed if it not explicitly
1527 specified; that is, if the `SECTIONS' command does not specify a
1528 start address for the section (*note SECTIONS::).
1530 `--warn-shared-textrel'
1531 Warn if the linker adds a DT_TEXTREL to a shared object.
1533 `--warn-alternate-em'
1534 Warn if an object has alternate ELF machine code.
1536 `--warn-unresolved-symbols'
1537 If the linker is going to report an unresolved symbol (see the
1538 option `--unresolved-symbols') it will normally generate an error.
1539 This option makes it generate a warning instead.
1541 `--error-unresolved-symbols'
1542 This restores the linker's default behaviour of generating errors
1543 when it is reporting unresolved symbols.
1546 For each archive mentioned on the command line after the
1547 `--whole-archive' option, include every object file in the archive
1548 in the link, rather than searching the archive for the required
1549 object files. This is normally used to turn an archive file into
1550 a shared library, forcing every object to be included in the
1551 resulting shared library. This option may be used more than once.
1553 Two notes when using this option from gcc: First, gcc doesn't know
1554 about this option, so you have to use `-Wl,-whole-archive'.
1555 Second, don't forget to use `-Wl,-no-whole-archive' after your
1556 list of archives, because gcc will add its own list of archives to
1557 your link and you may not want this flag to affect those as well.
1560 Use a wrapper function for SYMBOL. Any undefined reference to
1561 SYMBOL will be resolved to `__wrap_SYMBOL'. Any undefined
1562 reference to `__real_SYMBOL' will be resolved to SYMBOL.
1564 This can be used to provide a wrapper for a system function. The
1565 wrapper function should be called `__wrap_SYMBOL'. If it wishes
1566 to call the system function, it should call `__real_SYMBOL'.
1568 Here is a trivial example:
1571 __wrap_malloc (size_t c)
1573 printf ("malloc called with %zu\n", c);
1574 return __real_malloc (c);
1577 If you link other code with this file using `--wrap malloc', then
1578 all calls to `malloc' will call the function `__wrap_malloc'
1579 instead. The call to `__real_malloc' in `__wrap_malloc' will call
1580 the real `malloc' function.
1582 You may wish to provide a `__real_malloc' function as well, so that
1583 links without the `--wrap' option will succeed. If you do this,
1584 you should not put the definition of `__real_malloc' in the same
1585 file as `__wrap_malloc'; if you do, the assembler may resolve the
1586 call before the linker has a chance to wrap it to `malloc'.
1589 Request creation of `.eh_frame_hdr' section and ELF
1590 `PT_GNU_EH_FRAME' segment header.
1592 `--no-ld-generated-unwind-info'
1593 Request creation of `.eh_frame' unwind info for linker generated
1594 code sections like PLT. This option is on by default if linker
1595 generated unwind info is supported.
1597 `--enable-new-dtags'
1598 `--disable-new-dtags'
1599 This linker can create the new dynamic tags in ELF. But the older
1600 ELF systems may not understand them. If you specify
1601 `--enable-new-dtags', the dynamic tags will be created as needed.
1602 If you specify `--disable-new-dtags', no new dynamic tags will be
1603 created. By default, the new dynamic tags are not created. Note
1604 that those options are only available for ELF systems.
1606 `--hash-size=NUMBER'
1607 Set the default size of the linker's hash tables to a prime number
1608 close to NUMBER. Increasing this value can reduce the length of
1609 time it takes the linker to perform its tasks, at the expense of
1610 increasing the linker's memory requirements. Similarly reducing
1611 this value can reduce the memory requirements at the expense of
1614 `--hash-style=STYLE'
1615 Set the type of linker's hash table(s). STYLE can be either
1616 `sysv' for classic ELF `.hash' section, `gnu' for new style GNU
1617 `.gnu.hash' section or `both' for both the classic ELF `.hash' and
1618 new style GNU `.gnu.hash' hash tables. The default is `sysv'.
1620 `--reduce-memory-overheads'
1621 This option reduces memory requirements at ld runtime, at the
1622 expense of linking speed. This was introduced to select the old
1623 O(n^2) algorithm for link map file generation, rather than the new
1624 O(n) algorithm which uses about 40% more memory for symbol storage.
1626 Another effect of the switch is to set the default hash table size
1627 to 1021, which again saves memory at the cost of lengthening the
1628 linker's run time. This is not done however if the `--hash-size'
1629 switch has been used.
1631 The `--reduce-memory-overheads' switch may be also be used to
1632 enable other tradeoffs in future versions of the linker.
1636 Request creation of `.note.gnu.build-id' ELF note section. The
1637 contents of the note are unique bits identifying this linked file.
1638 STYLE can be `uuid' to use 128 random bits, `sha1' to use a
1639 160-bit SHA1 hash on the normative parts of the output contents,
1640 `md5' to use a 128-bit MD5 hash on the normative parts of the
1641 output contents, or `0xHEXSTRING' to use a chosen bit string
1642 specified as an even number of hexadecimal digits (`-' and `:'
1643 characters between digit pairs are ignored). If STYLE is omitted,
1646 The `md5' and `sha1' styles produces an identifier that is always
1647 the same in an identical output file, but will be unique among all
1648 nonidentical output files. It is not intended to be compared as a
1649 checksum for the file's contents. A linked file may be changed
1650 later by other tools, but the build ID bit string identifying the
1651 original linked file does not change.
1653 Passing `none' for STYLE disables the setting from any
1654 `--build-id' options earlier on the command line.
1656 2.1.1 Options Specific to i386 PE Targets
1657 -----------------------------------------
1659 The i386 PE linker supports the `-shared' option, which causes the
1660 output to be a dynamically linked library (DLL) instead of a normal
1661 executable. You should name the output `*.dll' when you use this
1662 option. In addition, the linker fully supports the standard `*.def'
1663 files, which may be specified on the linker command line like an object
1664 file (in fact, it should precede archives it exports symbols from, to
1665 ensure that they get linked in, just like a normal object file).
1667 In addition to the options common to all targets, the i386 PE linker
1668 support additional command line options that are specific to the i386
1669 PE target. Options that take values may be separated from their values
1670 by either a space or an equals sign.
1672 `--add-stdcall-alias'
1673 If given, symbols with a stdcall suffix (@NN) will be exported
1674 as-is and also with the suffix stripped. [This option is specific
1675 to the i386 PE targeted port of the linker]
1678 Use FILE as the name of a file in which to save the base addresses
1679 of all the relocations needed for generating DLLs with `dlltool'.
1680 [This is an i386 PE specific option]
1683 Create a DLL instead of a regular executable. You may also use
1684 `-shared' or specify a `LIBRARY' in a given `.def' file. [This
1685 option is specific to the i386 PE targeted port of the linker]
1687 `--enable-long-section-names'
1688 `--disable-long-section-names'
1689 The PE variants of the Coff object format add an extension that
1690 permits the use of section names longer than eight characters, the
1691 normal limit for Coff. By default, these names are only allowed
1692 in object files, as fully-linked executable images do not carry
1693 the Coff string table required to support the longer names. As a
1694 GNU extension, it is possible to allow their use in executable
1695 images as well, or to (probably pointlessly!) disallow it in
1696 object files, by using these two options. Executable images
1697 generated with these long section names are slightly non-standard,
1698 carrying as they do a string table, and may generate confusing
1699 output when examined with non-GNU PE-aware tools, such as file
1700 viewers and dumpers. However, GDB relies on the use of PE long
1701 section names to find Dwarf-2 debug information sections in an
1702 executable image at runtime, and so if neither option is specified
1703 on the command-line, `ld' will enable long section names,
1704 overriding the default and technically correct behaviour, when it
1705 finds the presence of debug information while linking an executable
1706 image and not stripping symbols. [This option is valid for all PE
1707 targeted ports of the linker]
1709 `--enable-stdcall-fixup'
1710 `--disable-stdcall-fixup'
1711 If the link finds a symbol that it cannot resolve, it will attempt
1712 to do "fuzzy linking" by looking for another defined symbol that
1713 differs only in the format of the symbol name (cdecl vs stdcall)
1714 and will resolve that symbol by linking to the match. For
1715 example, the undefined symbol `_foo' might be linked to the
1716 function `_foo@12', or the undefined symbol `_bar@16' might be
1717 linked to the function `_bar'. When the linker does this, it
1718 prints a warning, since it normally should have failed to link,
1719 but sometimes import libraries generated from third-party dlls may
1720 need this feature to be usable. If you specify
1721 `--enable-stdcall-fixup', this feature is fully enabled and
1722 warnings are not printed. If you specify
1723 `--disable-stdcall-fixup', this feature is disabled and such
1724 mismatches are considered to be errors. [This option is specific
1725 to the i386 PE targeted port of the linker]
1727 `--leading-underscore'
1728 `--no-leading-underscore'
1729 For most targets default symbol-prefix is an underscore and is
1730 defined in target's description. By this option it is possible to
1731 disable/enable the default underscore symbol-prefix.
1733 `--export-all-symbols'
1734 If given, all global symbols in the objects used to build a DLL
1735 will be exported by the DLL. Note that this is the default if
1736 there otherwise wouldn't be any exported symbols. When symbols are
1737 explicitly exported via DEF files or implicitly exported via
1738 function attributes, the default is to not export anything else
1739 unless this option is given. Note that the symbols `DllMain@12',
1740 `DllEntryPoint@0', `DllMainCRTStartup@12', and `impure_ptr' will
1741 not be automatically exported. Also, symbols imported from other
1742 DLLs will not be re-exported, nor will symbols specifying the
1743 DLL's internal layout such as those beginning with `_head_' or
1744 ending with `_iname'. In addition, no symbols from `libgcc',
1745 `libstd++', `libmingw32', or `crtX.o' will be exported. Symbols
1746 whose names begin with `__rtti_' or `__builtin_' will not be
1747 exported, to help with C++ DLLs. Finally, there is an extensive
1748 list of cygwin-private symbols that are not exported (obviously,
1749 this applies on when building DLLs for cygwin targets). These
1750 cygwin-excludes are: `_cygwin_dll_entry@12',
1751 `_cygwin_crt0_common@8', `_cygwin_noncygwin_dll_entry@12',
1752 `_fmode', `_impure_ptr', `cygwin_attach_dll', `cygwin_premain0',
1753 `cygwin_premain1', `cygwin_premain2', `cygwin_premain3', and
1754 `environ'. [This option is specific to the i386 PE targeted port
1757 `--exclude-symbols SYMBOL,SYMBOL,...'
1758 Specifies a list of symbols which should not be automatically
1759 exported. The symbol names may be delimited by commas or colons.
1760 [This option is specific to the i386 PE targeted port of the
1763 `--exclude-all-symbols'
1764 Specifies no symbols should be automatically exported. [This
1765 option is specific to the i386 PE targeted port of the linker]
1768 Specify the file alignment. Sections in the file will always
1769 begin at file offsets which are multiples of this number. This
1770 defaults to 512. [This option is specific to the i386 PE targeted
1774 `--heap RESERVE,COMMIT'
1775 Specify the number of bytes of memory to reserve (and optionally
1776 commit) to be used as heap for this program. The default is 1Mb
1777 reserved, 4K committed. [This option is specific to the i386 PE
1778 targeted port of the linker]
1780 `--image-base VALUE'
1781 Use VALUE as the base address of your program or dll. This is the
1782 lowest memory location that will be used when your program or dll
1783 is loaded. To reduce the need to relocate and improve performance
1784 of your dlls, each should have a unique base address and not
1785 overlap any other dlls. The default is 0x400000 for executables,
1786 and 0x10000000 for dlls. [This option is specific to the i386 PE
1787 targeted port of the linker]
1790 If given, the stdcall suffixes (@NN) will be stripped from symbols
1791 before they are exported. [This option is specific to the i386 PE
1792 targeted port of the linker]
1794 `--large-address-aware'
1795 If given, the appropriate bit in the "Characteristics" field of
1796 the COFF header is set to indicate that this executable supports
1797 virtual addresses greater than 2 gigabytes. This should be used
1798 in conjunction with the /3GB or /USERVA=VALUE megabytes switch in
1799 the "[operating systems]" section of the BOOT.INI. Otherwise,
1800 this bit has no effect. [This option is specific to PE targeted
1801 ports of the linker]
1803 `--major-image-version VALUE'
1804 Sets the major number of the "image version". Defaults to 1.
1805 [This option is specific to the i386 PE targeted port of the
1808 `--major-os-version VALUE'
1809 Sets the major number of the "os version". Defaults to 4. [This
1810 option is specific to the i386 PE targeted port of the linker]
1812 `--major-subsystem-version VALUE'
1813 Sets the major number of the "subsystem version". Defaults to 4.
1814 [This option is specific to the i386 PE targeted port of the
1817 `--minor-image-version VALUE'
1818 Sets the minor number of the "image version". Defaults to 0.
1819 [This option is specific to the i386 PE targeted port of the
1822 `--minor-os-version VALUE'
1823 Sets the minor number of the "os version". Defaults to 0. [This
1824 option is specific to the i386 PE targeted port of the linker]
1826 `--minor-subsystem-version VALUE'
1827 Sets the minor number of the "subsystem version". Defaults to 0.
1828 [This option is specific to the i386 PE targeted port of the
1832 The linker will create the file FILE which will contain a DEF file
1833 corresponding to the DLL the linker is generating. This DEF file
1834 (which should be called `*.def') may be used to create an import
1835 library with `dlltool' or may be used as a reference to
1836 automatically or implicitly exported symbols. [This option is
1837 specific to the i386 PE targeted port of the linker]
1840 The linker will create the file FILE which will contain an import
1841 lib corresponding to the DLL the linker is generating. This import
1842 lib (which should be called `*.dll.a' or `*.a' may be used to link
1843 clients against the generated DLL; this behaviour makes it
1844 possible to skip a separate `dlltool' import library creation step.
1845 [This option is specific to the i386 PE targeted port of the
1848 `--enable-auto-image-base'
1849 Automatically choose the image base for DLLs, unless one is
1850 specified using the `--image-base' argument. By using a hash
1851 generated from the dllname to create unique image bases for each
1852 DLL, in-memory collisions and relocations which can delay program
1853 execution are avoided. [This option is specific to the i386 PE
1854 targeted port of the linker]
1856 `--disable-auto-image-base'
1857 Do not automatically generate a unique image base. If there is no
1858 user-specified image base (`--image-base') then use the platform
1859 default. [This option is specific to the i386 PE targeted port of
1862 `--dll-search-prefix STRING'
1863 When linking dynamically to a dll without an import library,
1864 search for `<string><basename>.dll' in preference to
1865 `lib<basename>.dll'. This behaviour allows easy distinction
1866 between DLLs built for the various "subplatforms": native, cygwin,
1867 uwin, pw, etc. For instance, cygwin DLLs typically use
1868 `--dll-search-prefix=cyg'. [This option is specific to the i386
1869 PE targeted port of the linker]
1871 `--enable-auto-import'
1872 Do sophisticated linking of `_symbol' to `__imp__symbol' for DATA
1873 imports from DLLs, and create the necessary thunking symbols when
1874 building the import libraries with those DATA exports. Note: Use
1875 of the 'auto-import' extension will cause the text section of the
1876 image file to be made writable. This does not conform to the
1877 PE-COFF format specification published by Microsoft.
1879 Note - use of the 'auto-import' extension will also cause read only
1880 data which would normally be placed into the .rdata section to be
1881 placed into the .data section instead. This is in order to work
1882 around a problem with consts that is described here:
1883 http://www.cygwin.com/ml/cygwin/2004-09/msg01101.html
1885 Using 'auto-import' generally will 'just work' - but sometimes you
1886 may see this message:
1888 "variable '<var>' can't be auto-imported. Please read the
1889 documentation for ld's `--enable-auto-import' for details."
1891 This message occurs when some (sub)expression accesses an address
1892 ultimately given by the sum of two constants (Win32 import tables
1893 only allow one). Instances where this may occur include accesses
1894 to member fields of struct variables imported from a DLL, as well
1895 as using a constant index into an array variable imported from a
1896 DLL. Any multiword variable (arrays, structs, long long, etc) may
1897 trigger this error condition. However, regardless of the exact
1898 data type of the offending exported variable, ld will always
1899 detect it, issue the warning, and exit.
1901 There are several ways to address this difficulty, regardless of
1902 the data type of the exported variable:
1904 One way is to use -enable-runtime-pseudo-reloc switch. This leaves
1905 the task of adjusting references in your client code for runtime
1906 environment, so this method works only when runtime environment
1907 supports this feature.
1909 A second solution is to force one of the 'constants' to be a
1910 variable - that is, unknown and un-optimizable at compile time.
1911 For arrays, there are two possibilities: a) make the indexee (the
1912 array's address) a variable, or b) make the 'constant' index a
1915 extern type extern_array[];
1917 { volatile type *t=extern_array; t[1] }
1921 extern type extern_array[];
1923 { volatile int t=1; extern_array[t] }
1925 For structs (and most other multiword data types) the only option
1926 is to make the struct itself (or the long long, or the ...)
1929 extern struct s extern_struct;
1930 extern_struct.field -->
1931 { volatile struct s *t=&extern_struct; t->field }
1935 extern long long extern_ll;
1937 { volatile long long * local_ll=&extern_ll; *local_ll }
1939 A third method of dealing with this difficulty is to abandon
1940 'auto-import' for the offending symbol and mark it with
1941 `__declspec(dllimport)'. However, in practise that requires using
1942 compile-time #defines to indicate whether you are building a DLL,
1943 building client code that will link to the DLL, or merely
1944 building/linking to a static library. In making the choice
1945 between the various methods of resolving the 'direct address with
1946 constant offset' problem, you should consider typical real-world
1954 void main(int argc, char **argv){
1955 printf("%d\n",arr[1]);
1963 void main(int argc, char **argv){
1964 /* This workaround is for win32 and cygwin; do not "optimize" */
1965 volatile int *parr = arr;
1966 printf("%d\n",parr[1]);
1971 /* Note: auto-export is assumed (no __declspec(dllexport)) */
1972 #if (defined(_WIN32) || defined(__CYGWIN__)) && \
1973 !(defined(FOO_BUILD_DLL) || defined(FOO_STATIC))
1974 #define FOO_IMPORT __declspec(dllimport)
1978 extern FOO_IMPORT int arr[];
1981 void main(int argc, char **argv){
1982 printf("%d\n",arr[1]);
1985 A fourth way to avoid this problem is to re-code your library to
1986 use a functional interface rather than a data interface for the
1987 offending variables (e.g. set_foo() and get_foo() accessor
1988 functions). [This option is specific to the i386 PE targeted port
1991 `--disable-auto-import'
1992 Do not attempt to do sophisticated linking of `_symbol' to
1993 `__imp__symbol' for DATA imports from DLLs. [This option is
1994 specific to the i386 PE targeted port of the linker]
1996 `--enable-runtime-pseudo-reloc'
1997 If your code contains expressions described in -enable-auto-import
1998 section, that is, DATA imports from DLL with non-zero offset, this
1999 switch will create a vector of 'runtime pseudo relocations' which
2000 can be used by runtime environment to adjust references to such
2001 data in your client code. [This option is specific to the i386 PE
2002 targeted port of the linker]
2004 `--disable-runtime-pseudo-reloc'
2005 Do not create pseudo relocations for non-zero offset DATA imports
2006 from DLLs. This is the default. [This option is specific to the
2007 i386 PE targeted port of the linker]
2009 `--enable-extra-pe-debug'
2010 Show additional debug info related to auto-import symbol thunking.
2011 [This option is specific to the i386 PE targeted port of the
2014 `--section-alignment'
2015 Sets the section alignment. Sections in memory will always begin
2016 at addresses which are a multiple of this number. Defaults to
2017 0x1000. [This option is specific to the i386 PE targeted port of
2021 `--stack RESERVE,COMMIT'
2022 Specify the number of bytes of memory to reserve (and optionally
2023 commit) to be used as stack for this program. The default is 2Mb
2024 reserved, 4K committed. [This option is specific to the i386 PE
2025 targeted port of the linker]
2028 `--subsystem WHICH:MAJOR'
2029 `--subsystem WHICH:MAJOR.MINOR'
2030 Specifies the subsystem under which your program will execute. The
2031 legal values for WHICH are `native', `windows', `console',
2032 `posix', and `xbox'. You may optionally set the subsystem version
2033 also. Numeric values are also accepted for WHICH. [This option
2034 is specific to the i386 PE targeted port of the linker]
2036 The following options set flags in the `DllCharacteristics' field
2037 of the PE file header: [These options are specific to PE targeted
2038 ports of the linker]
2041 The image base address may be relocated using address space layout
2042 randomization (ASLR). This feature was introduced with MS Windows
2043 Vista for i386 PE targets.
2046 Code integrity checks are enforced.
2049 The image is compatible with the Data Execution Prevention. This
2050 feature was introduced with MS Windows XP SP2 for i386 PE targets.
2053 Although the image understands isolation, do not isolate the image.
2056 The image does not use SEH. No SE handler may be called from this
2060 Do not bind this image.
2063 The driver uses the MS Windows Driver Model.
2066 The image is Terminal Server aware.
2069 2.1.2 Options specific to C6X uClinux targets
2070 ---------------------------------------------
2072 The C6X uClinux target uses a binary format called DSBT to support
2073 shared libraries. Each shared library in the system needs to have a
2074 unique index; all executables use an index of 0.
2077 This option sets the number of entires in the DSBT of the current
2078 executable or shared library to SIZE. The default is to create a
2079 table with 64 entries.
2081 `--dsbt-index INDEX'
2082 This option sets the DSBT index of the current executable or
2083 shared library to INDEX. The default is 0, which is appropriate
2084 for generating executables. If a shared library is generated with
2085 a DSBT index of 0, the `R_C6000_DSBT_INDEX' relocs are copied into
2088 The `--no-merge-exidx-entries' switch disables the merging of
2089 adjacent exidx entries in frame unwind info.
2092 2.1.3 Options specific to Motorola 68HC11 and 68HC12 targets
2093 ------------------------------------------------------------
2095 The 68HC11 and 68HC12 linkers support specific options to control the
2096 memory bank switching mapping and trampoline code generation.
2099 This option disables the generation of trampoline. By default a
2100 trampoline is generated for each far function which is called
2101 using a `jsr' instruction (this happens when a pointer to a far
2104 `--bank-window NAME'
2105 This option indicates to the linker the name of the memory region
2106 in the `MEMORY' specification that describes the memory bank
2107 window. The definition of such region is then used by the linker
2108 to compute paging and addresses within the memory window.
2111 2.1.4 Options specific to Motorola 68K target
2112 ---------------------------------------------
2114 The following options are supported to control handling of GOT
2115 generation when linking for 68K targets.
2118 This option tells the linker which GOT generation scheme to use.
2119 TYPE should be one of `single', `negative', `multigot' or
2120 `target'. For more information refer to the Info entry for `ld'.
2124 File: ld.info, Node: Environment, Prev: Options, Up: Invocation
2126 2.2 Environment Variables
2127 =========================
2129 You can change the behaviour of `ld' with the environment variables
2130 `GNUTARGET', `LDEMULATION' and `COLLECT_NO_DEMANGLE'.
2132 `GNUTARGET' determines the input-file object format if you don't use
2133 `-b' (or its synonym `--format'). Its value should be one of the BFD
2134 names for an input format (*note BFD::). If there is no `GNUTARGET' in
2135 the environment, `ld' uses the natural format of the target. If
2136 `GNUTARGET' is set to `default' then BFD attempts to discover the input
2137 format by examining binary input files; this method often succeeds, but
2138 there are potential ambiguities, since there is no method of ensuring
2139 that the magic number used to specify object-file formats is unique.
2140 However, the configuration procedure for BFD on each system places the
2141 conventional format for that system first in the search-list, so
2142 ambiguities are resolved in favor of convention.
2144 `LDEMULATION' determines the default emulation if you don't use the
2145 `-m' option. The emulation can affect various aspects of linker
2146 behaviour, particularly the default linker script. You can list the
2147 available emulations with the `--verbose' or `-V' options. If the `-m'
2148 option is not used, and the `LDEMULATION' environment variable is not
2149 defined, the default emulation depends upon how the linker was
2152 Normally, the linker will default to demangling symbols. However, if
2153 `COLLECT_NO_DEMANGLE' is set in the environment, then it will default
2154 to not demangling symbols. This environment variable is used in a
2155 similar fashion by the `gcc' linker wrapper program. The default may
2156 be overridden by the `--demangle' and `--no-demangle' options.
2159 File: ld.info, Node: Scripts, Next: Machine Dependent, Prev: Invocation, Up: Top
2164 Every link is controlled by a "linker script". This script is written
2165 in the linker command language.
2167 The main purpose of the linker script is to describe how the
2168 sections in the input files should be mapped into the output file, and
2169 to control the memory layout of the output file. Most linker scripts
2170 do nothing more than this. However, when necessary, the linker script
2171 can also direct the linker to perform many other operations, using the
2172 commands described below.
2174 The linker always uses a linker script. If you do not supply one
2175 yourself, the linker will use a default script that is compiled into the
2176 linker executable. You can use the `--verbose' command line option to
2177 display the default linker script. Certain command line options, such
2178 as `-r' or `-N', will affect the default linker script.
2180 You may supply your own linker script by using the `-T' command line
2181 option. When you do this, your linker script will replace the default
2184 You may also use linker scripts implicitly by naming them as input
2185 files to the linker, as though they were files to be linked. *Note
2186 Implicit Linker Scripts::.
2190 * Basic Script Concepts:: Basic Linker Script Concepts
2191 * Script Format:: Linker Script Format
2192 * Simple Example:: Simple Linker Script Example
2193 * Simple Commands:: Simple Linker Script Commands
2194 * Assignments:: Assigning Values to Symbols
2195 * SECTIONS:: SECTIONS Command
2196 * MEMORY:: MEMORY Command
2197 * PHDRS:: PHDRS Command
2198 * VERSION:: VERSION Command
2199 * Expressions:: Expressions in Linker Scripts
2200 * Implicit Linker Scripts:: Implicit Linker Scripts
2203 File: ld.info, Node: Basic Script Concepts, Next: Script Format, Up: Scripts
2205 3.1 Basic Linker Script Concepts
2206 ================================
2208 We need to define some basic concepts and vocabulary in order to
2209 describe the linker script language.
2211 The linker combines input files into a single output file. The
2212 output file and each input file are in a special data format known as an
2213 "object file format". Each file is called an "object file". The
2214 output file is often called an "executable", but for our purposes we
2215 will also call it an object file. Each object file has, among other
2216 things, a list of "sections". We sometimes refer to a section in an
2217 input file as an "input section"; similarly, a section in the output
2218 file is an "output section".
2220 Each section in an object file has a name and a size. Most sections
2221 also have an associated block of data, known as the "section contents".
2222 A section may be marked as "loadable", which mean that the contents
2223 should be loaded into memory when the output file is run. A section
2224 with no contents may be "allocatable", which means that an area in
2225 memory should be set aside, but nothing in particular should be loaded
2226 there (in some cases this memory must be zeroed out). A section which
2227 is neither loadable nor allocatable typically contains some sort of
2228 debugging information.
2230 Every loadable or allocatable output section has two addresses. The
2231 first is the "VMA", or virtual memory address. This is the address the
2232 section will have when the output file is run. The second is the
2233 "LMA", or load memory address. This is the address at which the
2234 section will be loaded. In most cases the two addresses will be the
2235 same. An example of when they might be different is when a data section
2236 is loaded into ROM, and then copied into RAM when the program starts up
2237 (this technique is often used to initialize global variables in a ROM
2238 based system). In this case the ROM address would be the LMA, and the
2239 RAM address would be the VMA.
2241 You can see the sections in an object file by using the `objdump'
2242 program with the `-h' option.
2244 Every object file also has a list of "symbols", known as the "symbol
2245 table". A symbol may be defined or undefined. Each symbol has a name,
2246 and each defined symbol has an address, among other information. If
2247 you compile a C or C++ program into an object file, you will get a
2248 defined symbol for every defined function and global or static
2249 variable. Every undefined function or global variable which is
2250 referenced in the input file will become an undefined symbol.
2252 You can see the symbols in an object file by using the `nm' program,
2253 or by using the `objdump' program with the `-t' option.
2256 File: ld.info, Node: Script Format, Next: Simple Example, Prev: Basic Script Concepts, Up: Scripts
2258 3.2 Linker Script Format
2259 ========================
2261 Linker scripts are text files.
2263 You write a linker script as a series of commands. Each command is
2264 either a keyword, possibly followed by arguments, or an assignment to a
2265 symbol. You may separate commands using semicolons. Whitespace is
2268 Strings such as file or format names can normally be entered
2269 directly. If the file name contains a character such as a comma which
2270 would otherwise serve to separate file names, you may put the file name
2271 in double quotes. There is no way to use a double quote character in a
2274 You may include comments in linker scripts just as in C, delimited by
2275 `/*' and `*/'. As in C, comments are syntactically equivalent to
2279 File: ld.info, Node: Simple Example, Next: Simple Commands, Prev: Script Format, Up: Scripts
2281 3.3 Simple Linker Script Example
2282 ================================
2284 Many linker scripts are fairly simple.
2286 The simplest possible linker script has just one command:
2287 `SECTIONS'. You use the `SECTIONS' command to describe the memory
2288 layout of the output file.
2290 The `SECTIONS' command is a powerful command. Here we will describe
2291 a simple use of it. Let's assume your program consists only of code,
2292 initialized data, and uninitialized data. These will be in the
2293 `.text', `.data', and `.bss' sections, respectively. Let's assume
2294 further that these are the only sections which appear in your input
2297 For this example, let's say that the code should be loaded at address
2298 0x10000, and that the data should start at address 0x8000000. Here is a
2299 linker script which will do that:
2303 .text : { *(.text) }
2305 .data : { *(.data) }
2309 You write the `SECTIONS' command as the keyword `SECTIONS', followed
2310 by a series of symbol assignments and output section descriptions
2311 enclosed in curly braces.
2313 The first line inside the `SECTIONS' command of the above example
2314 sets the value of the special symbol `.', which is the location
2315 counter. If you do not specify the address of an output section in some
2316 other way (other ways are described later), the address is set from the
2317 current value of the location counter. The location counter is then
2318 incremented by the size of the output section. At the start of the
2319 `SECTIONS' command, the location counter has the value `0'.
2321 The second line defines an output section, `.text'. The colon is
2322 required syntax which may be ignored for now. Within the curly braces
2323 after the output section name, you list the names of the input sections
2324 which should be placed into this output section. The `*' is a wildcard
2325 which matches any file name. The expression `*(.text)' means all
2326 `.text' input sections in all input files.
2328 Since the location counter is `0x10000' when the output section
2329 `.text' is defined, the linker will set the address of the `.text'
2330 section in the output file to be `0x10000'.
2332 The remaining lines define the `.data' and `.bss' sections in the
2333 output file. The linker will place the `.data' output section at
2334 address `0x8000000'. After the linker places the `.data' output
2335 section, the value of the location counter will be `0x8000000' plus the
2336 size of the `.data' output section. The effect is that the linker will
2337 place the `.bss' output section immediately after the `.data' output
2340 The linker will ensure that each output section has the required
2341 alignment, by increasing the location counter if necessary. In this
2342 example, the specified addresses for the `.text' and `.data' sections
2343 will probably satisfy any alignment constraints, but the linker may
2344 have to create a small gap between the `.data' and `.bss' sections.
2346 That's it! That's a simple and complete linker script.
2349 File: ld.info, Node: Simple Commands, Next: Assignments, Prev: Simple Example, Up: Scripts
2351 3.4 Simple Linker Script Commands
2352 =================================
2354 In this section we describe the simple linker script commands.
2358 * Entry Point:: Setting the entry point
2359 * File Commands:: Commands dealing with files
2361 * Format Commands:: Commands dealing with object file formats
2363 * REGION_ALIAS:: Assign alias names to memory regions
2364 * Miscellaneous Commands:: Other linker script commands
2367 File: ld.info, Node: Entry Point, Next: File Commands, Up: Simple Commands
2369 3.4.1 Setting the Entry Point
2370 -----------------------------
2372 The first instruction to execute in a program is called the "entry
2373 point". You can use the `ENTRY' linker script command to set the entry
2374 point. The argument is a symbol name:
2377 There are several ways to set the entry point. The linker will set
2378 the entry point by trying each of the following methods in order, and
2379 stopping when one of them succeeds:
2380 * the `-e' ENTRY command-line option;
2382 * the `ENTRY(SYMBOL)' command in a linker script;
2384 * the value of a target specific symbol, if it is defined; For many
2385 targets this is `start', but PE and BeOS based systems for example
2386 check a list of possible entry symbols, matching the first one
2389 * the address of the first byte of the `.text' section, if present;
2394 File: ld.info, Node: File Commands, Next: Format Commands, Prev: Entry Point, Up: Simple Commands
2396 3.4.2 Commands Dealing with Files
2397 ---------------------------------
2399 Several linker script commands deal with files.
2402 Include the linker script FILENAME at this point. The file will
2403 be searched for in the current directory, and in any directory
2404 specified with the `-L' option. You can nest calls to `INCLUDE'
2405 up to 10 levels deep.
2407 You can place `INCLUDE' directives at the top level, in `MEMORY' or
2408 `SECTIONS' commands, or in output section descriptions.
2410 `INPUT(FILE, FILE, ...)'
2411 `INPUT(FILE FILE ...)'
2412 The `INPUT' command directs the linker to include the named files
2413 in the link, as though they were named on the command line.
2415 For example, if you always want to include `subr.o' any time you do
2416 a link, but you can't be bothered to put it on every link command
2417 line, then you can put `INPUT (subr.o)' in your linker script.
2419 In fact, if you like, you can list all of your input files in the
2420 linker script, and then invoke the linker with nothing but a `-T'
2423 In case a "sysroot prefix" is configured, and the filename starts
2424 with the `/' character, and the script being processed was located
2425 inside the "sysroot prefix", the filename will be looked for in
2426 the "sysroot prefix". Otherwise, the linker will try to open the
2427 file in the current directory. If it is not found, the linker
2428 will search through the archive library search path. See the
2429 description of `-L' in *Note Command Line Options: Options.
2431 If you use `INPUT (-lFILE)', `ld' will transform the name to
2432 `libFILE.a', as with the command line argument `-l'.
2434 When you use the `INPUT' command in an implicit linker script, the
2435 files will be included in the link at the point at which the linker
2436 script file is included. This can affect archive searching.
2438 `GROUP(FILE, FILE, ...)'
2439 `GROUP(FILE FILE ...)'
2440 The `GROUP' command is like `INPUT', except that the named files
2441 should all be archives, and they are searched repeatedly until no
2442 new undefined references are created. See the description of `-('
2443 in *Note Command Line Options: Options.
2445 `AS_NEEDED(FILE, FILE, ...)'
2446 `AS_NEEDED(FILE FILE ...)'
2447 This construct can appear only inside of the `INPUT' or `GROUP'
2448 commands, among other filenames. The files listed will be handled
2449 as if they appear directly in the `INPUT' or `GROUP' commands,
2450 with the exception of ELF shared libraries, that will be added only
2451 when they are actually needed. This construct essentially enables
2452 `--as-needed' option for all the files listed inside of it and
2453 restores previous `--as-needed' resp. `--no-as-needed' setting
2457 The `OUTPUT' command names the output file. Using
2458 `OUTPUT(FILENAME)' in the linker script is exactly like using `-o
2459 FILENAME' on the command line (*note Command Line Options:
2460 Options.). If both are used, the command line option takes
2463 You can use the `OUTPUT' command to define a default name for the
2464 output file other than the usual default of `a.out'.
2467 The `SEARCH_DIR' command adds PATH to the list of paths where `ld'
2468 looks for archive libraries. Using `SEARCH_DIR(PATH)' is exactly
2469 like using `-L PATH' on the command line (*note Command Line
2470 Options: Options.). If both are used, then the linker will search
2471 both paths. Paths specified using the command line option are
2475 The `STARTUP' command is just like the `INPUT' command, except
2476 that FILENAME will become the first input file to be linked, as
2477 though it were specified first on the command line. This may be
2478 useful when using a system in which the entry point is always the
2479 start of the first file.
2482 File: ld.info, Node: Format Commands, Next: REGION_ALIAS, Prev: File Commands, Up: Simple Commands
2484 3.4.3 Commands Dealing with Object File Formats
2485 -----------------------------------------------
2487 A couple of linker script commands deal with object file formats.
2489 `OUTPUT_FORMAT(BFDNAME)'
2490 `OUTPUT_FORMAT(DEFAULT, BIG, LITTLE)'
2491 The `OUTPUT_FORMAT' command names the BFD format to use for the
2492 output file (*note BFD::). Using `OUTPUT_FORMAT(BFDNAME)' is
2493 exactly like using `--oformat BFDNAME' on the command line (*note
2494 Command Line Options: Options.). If both are used, the command
2495 line option takes precedence.
2497 You can use `OUTPUT_FORMAT' with three arguments to use different
2498 formats based on the `-EB' and `-EL' command line options. This
2499 permits the linker script to set the output format based on the
2502 If neither `-EB' nor `-EL' are used, then the output format will
2503 be the first argument, DEFAULT. If `-EB' is used, the output
2504 format will be the second argument, BIG. If `-EL' is used, the
2505 output format will be the third argument, LITTLE.
2507 For example, the default linker script for the MIPS ELF target
2509 OUTPUT_FORMAT(elf32-bigmips, elf32-bigmips, elf32-littlemips)
2510 This says that the default format for the output file is
2511 `elf32-bigmips', but if the user uses the `-EL' command line
2512 option, the output file will be created in the `elf32-littlemips'
2516 The `TARGET' command names the BFD format to use when reading input
2517 files. It affects subsequent `INPUT' and `GROUP' commands. This
2518 command is like using `-b BFDNAME' on the command line (*note
2519 Command Line Options: Options.). If the `TARGET' command is used
2520 but `OUTPUT_FORMAT' is not, then the last `TARGET' command is also
2521 used to set the format for the output file. *Note BFD::.
2524 File: ld.info, Node: REGION_ALIAS, Next: Miscellaneous Commands, Prev: Format Commands, Up: Simple Commands
2526 3.4.4 Assign alias names to memory regions
2527 ------------------------------------------
2529 Alias names can be added to existing memory regions created with the
2530 *Note MEMORY:: command. Each name corresponds to at most one memory
2533 REGION_ALIAS(ALIAS, REGION)
2535 The `REGION_ALIAS' function creates an alias name ALIAS for the
2536 memory region REGION. This allows a flexible mapping of output sections
2537 to memory regions. An example follows.
2539 Suppose we have an application for embedded systems which come with
2540 various memory storage devices. All have a general purpose, volatile
2541 memory `RAM' that allows code execution or data storage. Some may have
2542 a read-only, non-volatile memory `ROM' that allows code execution and
2543 read-only data access. The last variant is a read-only, non-volatile
2544 memory `ROM2' with read-only data access and no code execution
2545 capability. We have four output sections:
2547 * `.text' program code;
2549 * `.rodata' read-only data;
2551 * `.data' read-write initialized data;
2553 * `.bss' read-write zero initialized data.
2555 The goal is to provide a linker command file that contains a system
2556 independent part defining the output sections and a system dependent
2557 part mapping the output sections to the memory regions available on the
2558 system. Our embedded systems come with three different memory setups
2560 Section Variant A Variant B Variant C
2562 .rodata RAM ROM ROM2
2563 .data RAM RAM/ROM RAM/ROM2
2565 The notation `RAM/ROM' or `RAM/ROM2' means that this section is
2566 loaded into region `ROM' or `ROM2' respectively. Please note that the
2567 load address of the `.data' section starts in all three variants at the
2568 end of the `.rodata' section.
2570 The base linker script that deals with the output sections follows.
2571 It includes the system dependent `linkcmds.memory' file that describes
2573 INCLUDE linkcmds.memory
2586 .data : AT (rodata_end)
2591 data_size = SIZEOF(.data);
2592 data_load_start = LOADADDR(.data);
2599 Now we need three different `linkcmds.memory' files to define memory
2600 regions and alias names. The content of `linkcmds.memory' for the three
2601 variants `A', `B' and `C':
2603 Here everything goes into the `RAM'.
2606 RAM : ORIGIN = 0, LENGTH = 4M
2609 REGION_ALIAS("REGION_TEXT", RAM);
2610 REGION_ALIAS("REGION_RODATA", RAM);
2611 REGION_ALIAS("REGION_DATA", RAM);
2612 REGION_ALIAS("REGION_BSS", RAM);
2615 Program code and read-only data go into the `ROM'. Read-write
2616 data goes into the `RAM'. An image of the initialized data is
2617 loaded into the `ROM' and will be copied during system start into
2621 ROM : ORIGIN = 0, LENGTH = 3M
2622 RAM : ORIGIN = 0x10000000, LENGTH = 1M
2625 REGION_ALIAS("REGION_TEXT", ROM);
2626 REGION_ALIAS("REGION_RODATA", ROM);
2627 REGION_ALIAS("REGION_DATA", RAM);
2628 REGION_ALIAS("REGION_BSS", RAM);
2631 Program code goes into the `ROM'. Read-only data goes into the
2632 `ROM2'. Read-write data goes into the `RAM'. An image of the
2633 initialized data is loaded into the `ROM2' and will be copied
2634 during system start into the `RAM'.
2637 ROM : ORIGIN = 0, LENGTH = 2M
2638 ROM2 : ORIGIN = 0x10000000, LENGTH = 1M
2639 RAM : ORIGIN = 0x20000000, LENGTH = 1M
2642 REGION_ALIAS("REGION_TEXT", ROM);
2643 REGION_ALIAS("REGION_RODATA", ROM2);
2644 REGION_ALIAS("REGION_DATA", RAM);
2645 REGION_ALIAS("REGION_BSS", RAM);
2647 It is possible to write a common system initialization routine to
2648 copy the `.data' section from `ROM' or `ROM2' into the `RAM' if
2652 extern char data_start [];
2653 extern char data_size [];
2654 extern char data_load_start [];
2656 void copy_data(void)
2658 if (data_start != data_load_start)
2660 memcpy(data_start, data_load_start, (size_t) data_size);
2665 File: ld.info, Node: Miscellaneous Commands, Prev: REGION_ALIAS, Up: Simple Commands
2667 3.4.5 Other Linker Script Commands
2668 ----------------------------------
2670 There are a few other linker scripts commands.
2672 `ASSERT(EXP, MESSAGE)'
2673 Ensure that EXP is non-zero. If it is zero, then exit the linker
2674 with an error code, and print MESSAGE.
2676 `EXTERN(SYMBOL SYMBOL ...)'
2677 Force SYMBOL to be entered in the output file as an undefined
2678 symbol. Doing this may, for example, trigger linking of additional
2679 modules from standard libraries. You may list several SYMBOLs for
2680 each `EXTERN', and you may use `EXTERN' multiple times. This
2681 command has the same effect as the `-u' command-line option.
2683 `FORCE_COMMON_ALLOCATION'
2684 This command has the same effect as the `-d' command-line option:
2685 to make `ld' assign space to common symbols even if a relocatable
2686 output file is specified (`-r').
2688 `INHIBIT_COMMON_ALLOCATION'
2689 This command has the same effect as the `--no-define-common'
2690 command-line option: to make `ld' omit the assignment of addresses
2691 to common symbols even for a non-relocatable output file.
2693 `INSERT [ AFTER | BEFORE ] OUTPUT_SECTION'
2694 This command is typically used in a script specified by `-T' to
2695 augment the default `SECTIONS' with, for example, overlays. It
2696 inserts all prior linker script statements after (or before)
2697 OUTPUT_SECTION, and also causes `-T' to not override the default
2698 linker script. The exact insertion point is as for orphan
2699 sections. *Note Location Counter::. The insertion happens after
2700 the linker has mapped input sections to output sections. Prior to
2701 the insertion, since `-T' scripts are parsed before the default
2702 linker script, statements in the `-T' script occur before the
2703 default linker script statements in the internal linker
2704 representation of the script. In particular, input section
2705 assignments will be made to `-T' output sections before those in
2706 the default script. Here is an example of how a `-T' script using
2707 `INSERT' might look:
2713 .ov1 { ov1*(.text) }
2714 .ov2 { ov2*(.text) }
2719 `NOCROSSREFS(SECTION SECTION ...)'
2720 This command may be used to tell `ld' to issue an error about any
2721 references among certain output sections.
2723 In certain types of programs, particularly on embedded systems when
2724 using overlays, when one section is loaded into memory, another
2725 section will not be. Any direct references between the two
2726 sections would be errors. For example, it would be an error if
2727 code in one section called a function defined in the other section.
2729 The `NOCROSSREFS' command takes a list of output section names. If
2730 `ld' detects any cross references between the sections, it reports
2731 an error and returns a non-zero exit status. Note that the
2732 `NOCROSSREFS' command uses output section names, not input section
2735 `OUTPUT_ARCH(BFDARCH)'
2736 Specify a particular output machine architecture. The argument is
2737 one of the names used by the BFD library (*note BFD::). You can
2738 see the architecture of an object file by using the `objdump'
2739 program with the `-f' option.
2741 `LD_FEATURE(STRING)'
2742 This command may be used to modify `ld' behavior. If STRING is
2743 `"SANE_EXPR"' then absolute symbols and numbers in a script are
2744 simply treated as numbers everywhere. *Note Expression Section::.
2747 File: ld.info, Node: Assignments, Next: SECTIONS, Prev: Simple Commands, Up: Scripts
2749 3.5 Assigning Values to Symbols
2750 ===============================
2752 You may assign a value to a symbol in a linker script. This will define
2753 the symbol and place it into the symbol table with a global scope.
2757 * Simple Assignments:: Simple Assignments
2760 * PROVIDE_HIDDEN:: PROVIDE_HIDDEN
2761 * Source Code Reference:: How to use a linker script defined symbol in source code
2764 File: ld.info, Node: Simple Assignments, Next: HIDDEN, Up: Assignments
2766 3.5.1 Simple Assignments
2767 ------------------------
2769 You may assign to a symbol using any of the C assignment operators:
2771 `SYMBOL = EXPRESSION ;'
2772 `SYMBOL += EXPRESSION ;'
2773 `SYMBOL -= EXPRESSION ;'
2774 `SYMBOL *= EXPRESSION ;'
2775 `SYMBOL /= EXPRESSION ;'
2776 `SYMBOL <<= EXPRESSION ;'
2777 `SYMBOL >>= EXPRESSION ;'
2778 `SYMBOL &= EXPRESSION ;'
2779 `SYMBOL |= EXPRESSION ;'
2781 The first case will define SYMBOL to the value of EXPRESSION. In
2782 the other cases, SYMBOL must already be defined, and the value will be
2783 adjusted accordingly.
2785 The special symbol name `.' indicates the location counter. You may
2786 only use this within a `SECTIONS' command. *Note Location Counter::.
2788 The semicolon after EXPRESSION is required.
2790 Expressions are defined below; see *Note Expressions::.
2792 You may write symbol assignments as commands in their own right, or
2793 as statements within a `SECTIONS' command, or as part of an output
2794 section description in a `SECTIONS' command.
2796 The section of the symbol will be set from the section of the
2797 expression; for more information, see *Note Expression Section::.
2799 Here is an example showing the three different places that symbol
2800 assignments may be used:
2810 _bdata = (. + 3) & ~ 3;
2811 .data : { *(.data) }
2813 In this example, the symbol `floating_point' will be defined as
2814 zero. The symbol `_etext' will be defined as the address following the
2815 last `.text' input section. The symbol `_bdata' will be defined as the
2816 address following the `.text' output section aligned upward to a 4 byte
2820 File: ld.info, Node: HIDDEN, Next: PROVIDE, Prev: Simple Assignments, Up: Assignments
2825 For ELF targeted ports, define a symbol that will be hidden and won't be
2826 exported. The syntax is `HIDDEN(SYMBOL = EXPRESSION)'.
2828 Here is the example from *Note Simple Assignments::, rewritten to use
2831 HIDDEN(floating_point = 0);
2839 HIDDEN(_bdata = (. + 3) & ~ 3);
2840 .data : { *(.data) }
2842 In this case none of the three symbols will be visible outside this
2846 File: ld.info, Node: PROVIDE, Next: PROVIDE_HIDDEN, Prev: HIDDEN, Up: Assignments
2851 In some cases, it is desirable for a linker script to define a symbol
2852 only if it is referenced and is not defined by any object included in
2853 the link. For example, traditional linkers defined the symbol `etext'.
2854 However, ANSI C requires that the user be able to use `etext' as a
2855 function name without encountering an error. The `PROVIDE' keyword may
2856 be used to define a symbol, such as `etext', only if it is referenced
2857 but not defined. The syntax is `PROVIDE(SYMBOL = EXPRESSION)'.
2859 Here is an example of using `PROVIDE' to define `etext':
2870 In this example, if the program defines `_etext' (with a leading
2871 underscore), the linker will give a multiple definition error. If, on
2872 the other hand, the program defines `etext' (with no leading
2873 underscore), the linker will silently use the definition in the program.
2874 If the program references `etext' but does not define it, the linker
2875 will use the definition in the linker script.
2878 File: ld.info, Node: PROVIDE_HIDDEN, Next: Source Code Reference, Prev: PROVIDE, Up: Assignments
2880 3.5.4 PROVIDE_HIDDEN
2881 --------------------
2883 Similar to `PROVIDE'. For ELF targeted ports, the symbol will be
2884 hidden and won't be exported.
2887 File: ld.info, Node: Source Code Reference, Prev: PROVIDE_HIDDEN, Up: Assignments
2889 3.5.5 Source Code Reference
2890 ---------------------------
2892 Accessing a linker script defined variable from source code is not
2893 intuitive. In particular a linker script symbol is not equivalent to a
2894 variable declaration in a high level language, it is instead a symbol
2895 that does not have a value.
2897 Before going further, it is important to note that compilers often
2898 transform names in the source code into different names when they are
2899 stored in the symbol table. For example, Fortran compilers commonly
2900 prepend or append an underscore, and C++ performs extensive `name
2901 mangling'. Therefore there might be a discrepancy between the name of
2902 a variable as it is used in source code and the name of the same
2903 variable as it is defined in a linker script. For example in C a
2904 linker script variable might be referred to as:
2908 But in the linker script it might be defined as:
2912 In the remaining examples however it is assumed that no name
2913 transformation has taken place.
2915 When a symbol is declared in a high level language such as C, two
2916 things happen. The first is that the compiler reserves enough space in
2917 the program's memory to hold the _value_ of the symbol. The second is
2918 that the compiler creates an entry in the program's symbol table which
2919 holds the symbol's _address_. ie the symbol table contains the address
2920 of the block of memory holding the symbol's value. So for example the
2921 following C declaration, at file scope:
2925 creates a entry called `foo' in the symbol table. This entry holds
2926 the address of an `int' sized block of memory where the number 1000 is
2929 When a program references a symbol the compiler generates code that
2930 first accesses the symbol table to find the address of the symbol's
2931 memory block and then code to read the value from that memory block.
2936 looks up the symbol `foo' in the symbol table, gets the address
2937 associated with this symbol and then writes the value 1 into that
2942 looks up the symbol `foo' in the symbol table, gets it address and
2943 then copies this address into the block of memory associated with the
2946 Linker scripts symbol declarations, by contrast, create an entry in
2947 the symbol table but do not assign any memory to them. Thus they are
2948 an address without a value. So for example the linker script
2953 creates an entry in the symbol table called `foo' which holds the
2954 address of memory location 1000, but nothing special is stored at
2955 address 1000. This means that you cannot access the _value_ of a
2956 linker script defined symbol - it has no value - all you can do is
2957 access the _address_ of a linker script defined symbol.
2959 Hence when you are using a linker script defined symbol in source
2960 code you should always take the address of the symbol, and never
2961 attempt to use its value. For example suppose you want to copy the
2962 contents of a section of memory called .ROM into a section called
2963 .FLASH and the linker script contains these declarations:
2965 start_of_ROM = .ROM;
2966 end_of_ROM = .ROM + sizeof (.ROM) - 1;
2967 start_of_FLASH = .FLASH;
2969 Then the C source code to perform the copy would be:
2971 extern char start_of_ROM, end_of_ROM, start_of_FLASH;
2973 memcpy (& start_of_FLASH, & start_of_ROM, & end_of_ROM - & start_of_ROM);
2975 Note the use of the `&' operators. These are correct.
2978 File: ld.info, Node: SECTIONS, Next: MEMORY, Prev: Assignments, Up: Scripts
2980 3.6 SECTIONS Command
2981 ====================
2983 The `SECTIONS' command tells the linker how to map input sections into
2984 output sections, and how to place the output sections in memory.
2986 The format of the `SECTIONS' command is:
2994 Each SECTIONS-COMMAND may of be one of the following:
2996 * an `ENTRY' command (*note Entry command: Entry Point.)
2998 * a symbol assignment (*note Assignments::)
3000 * an output section description
3002 * an overlay description
3004 The `ENTRY' command and symbol assignments are permitted inside the
3005 `SECTIONS' command for convenience in using the location counter in
3006 those commands. This can also make the linker script easier to
3007 understand because you can use those commands at meaningful points in
3008 the layout of the output file.
3010 Output section descriptions and overlay descriptions are described
3013 If you do not use a `SECTIONS' command in your linker script, the
3014 linker will place each input section into an identically named output
3015 section in the order that the sections are first encountered in the
3016 input files. If all input sections are present in the first file, for
3017 example, the order of sections in the output file will match the order
3018 in the first input file. The first section will be at address zero.
3022 * Output Section Description:: Output section description
3023 * Output Section Name:: Output section name
3024 * Output Section Address:: Output section address
3025 * Input Section:: Input section description
3026 * Output Section Data:: Output section data
3027 * Output Section Keywords:: Output section keywords
3028 * Output Section Discarding:: Output section discarding
3029 * Output Section Attributes:: Output section attributes
3030 * Overlay Description:: Overlay description
3033 File: ld.info, Node: Output Section Description, Next: Output Section Name, Up: SECTIONS
3035 3.6.1 Output Section Description
3036 --------------------------------
3038 The full description of an output section looks like this:
3039 SECTION [ADDRESS] [(TYPE)] :
3041 [ALIGN(SECTION_ALIGN)]
3042 [SUBALIGN(SUBSECTION_ALIGN)]
3045 OUTPUT-SECTION-COMMAND
3046 OUTPUT-SECTION-COMMAND
3048 } [>REGION] [AT>LMA_REGION] [:PHDR :PHDR ...] [=FILLEXP]
3050 Most output sections do not use most of the optional section
3053 The whitespace around SECTION is required, so that the section name
3054 is unambiguous. The colon and the curly braces are also required. The
3055 line breaks and other white space are optional.
3057 Each OUTPUT-SECTION-COMMAND may be one of the following:
3059 * a symbol assignment (*note Assignments::)
3061 * an input section description (*note Input Section::)
3063 * data values to include directly (*note Output Section Data::)
3065 * a special output section keyword (*note Output Section Keywords::)
3068 File: ld.info, Node: Output Section Name, Next: Output Section Address, Prev: Output Section Description, Up: SECTIONS
3070 3.6.2 Output Section Name
3071 -------------------------
3073 The name of the output section is SECTION. SECTION must meet the
3074 constraints of your output format. In formats which only support a
3075 limited number of sections, such as `a.out', the name must be one of
3076 the names supported by the format (`a.out', for example, allows only
3077 `.text', `.data' or `.bss'). If the output format supports any number
3078 of sections, but with numbers and not names (as is the case for Oasys),
3079 the name should be supplied as a quoted numeric string. A section name
3080 may consist of any sequence of characters, but a name which contains
3081 any unusual characters such as commas must be quoted.
3083 The output section name `/DISCARD/' is special; *Note Output Section
3087 File: ld.info, Node: Output Section Address, Next: Input Section, Prev: Output Section Name, Up: SECTIONS
3089 3.6.3 Output Section Address
3090 ----------------------------
3092 The ADDRESS is an expression for the VMA (the virtual memory address)
3093 of the output section. This address is optional, but if it is provided
3094 then the output address will be set exactly as specified.
3096 If the output address is not specified then one will be chosen for
3097 the section, based on the heuristic below. This address will be
3098 adjusted to fit the alignment requirement of the output section. The
3099 alignment requirement is the strictest alignment of any input section
3100 contained within the output section.
3102 The output section address heuristic is as follows:
3104 * If an output memory REGION is set for the section then it is added
3105 to this region and its address will be the next free address in
3108 * If the MEMORY command has been used to create a list of memory
3109 regions then the first region which has attributes compatible with
3110 the section is selected to contain it. The section's output
3111 address will be the next free address in that region; *Note
3114 * If no memory regions were specified, or none match the section then
3115 the output address will be based on the current value of the
3120 .text . : { *(.text) }
3124 .text : { *(.text) }
3126 are subtly different. The first will set the address of the `.text'
3127 output section to the current value of the location counter. The
3128 second will set it to the current value of the location counter aligned
3129 to the strictest alignment of any of the `.text' input sections.
3131 The ADDRESS may be an arbitrary expression; *Note Expressions::.
3132 For example, if you want to align the section on a 0x10 byte boundary,
3133 so that the lowest four bits of the section address are zero, you could
3134 do something like this:
3135 .text ALIGN(0x10) : { *(.text) }
3136 This works because `ALIGN' returns the current location counter
3137 aligned upward to the specified value.
3139 Specifying ADDRESS for a section will change the value of the
3140 location counter, provided that the section is non-empty. (Empty
3141 sections are ignored).
3144 File: ld.info, Node: Input Section, Next: Output Section Data, Prev: Output Section Address, Up: SECTIONS
3146 3.6.4 Input Section Description
3147 -------------------------------
3149 The most common output section command is an input section description.
3151 The input section description is the most basic linker script
3152 operation. You use output sections to tell the linker how to lay out
3153 your program in memory. You use input section descriptions to tell the
3154 linker how to map the input files into your memory layout.
3158 * Input Section Basics:: Input section basics
3159 * Input Section Wildcards:: Input section wildcard patterns
3160 * Input Section Common:: Input section for common symbols
3161 * Input Section Keep:: Input section and garbage collection
3162 * Input Section Example:: Input section example
3165 File: ld.info, Node: Input Section Basics, Next: Input Section Wildcards, Up: Input Section
3167 3.6.4.1 Input Section Basics
3168 ............................
3170 An input section description consists of a file name optionally followed
3171 by a list of section names in parentheses.
3173 The file name and the section name may be wildcard patterns, which we
3174 describe further below (*note Input Section Wildcards::).
3176 The most common input section description is to include all input
3177 sections with a particular name in the output section. For example, to
3178 include all input `.text' sections, you would write:
3180 Here the `*' is a wildcard which matches any file name. To exclude
3181 a list of files from matching the file name wildcard, EXCLUDE_FILE may
3182 be used to match all files except the ones specified in the
3183 EXCLUDE_FILE list. For example:
3184 *(EXCLUDE_FILE (*crtend.o *otherfile.o) .ctors)
3185 will cause all .ctors sections from all files except `crtend.o' and
3186 `otherfile.o' to be included.
3188 There are two ways to include more than one section:
3191 The difference between these is the order in which the `.text' and
3192 `.rdata' input sections will appear in the output section. In the
3193 first example, they will be intermingled, appearing in the same order as
3194 they are found in the linker input. In the second example, all `.text'
3195 input sections will appear first, followed by all `.rdata' input
3198 You can specify a file name to include sections from a particular
3199 file. You would do this if one or more of your files contain special
3200 data that needs to be at a particular location in memory. For example:
3203 To refine the sections that are included based on the section flags
3204 of an input section, INPUT_SECTION_FLAGS may be used.
3206 Here is a simple example for using Section header flags for ELF
3210 .text : { INPUT_SECTION_FLAGS (SHF_MERGE & SHF_STRINGS) *(.text) }
3211 .text2 : { INPUT_SECTION_FLAGS (!SHF_WRITE) *(.text) }
3214 In this example, the output section `.text' will be comprised of any
3215 input section matching the name *(.text) whose section header flags
3216 `SHF_MERGE' and `SHF_STRINGS' are set. The output section `.text2'
3217 will be comprised of any input section matching the name *(.text) whose
3218 section header flag `SHF_WRITE' is clear.
3220 You can also specify files within archives by writing a pattern
3221 matching the archive, a colon, then the pattern matching the file, with
3222 no whitespace around the colon.
3225 matches file within archive
3228 matches the whole archive
3231 matches file but not one in an archive
3233 Either one or both of `archive' and `file' can contain shell
3234 wildcards. On DOS based file systems, the linker will assume that a
3235 single letter followed by a colon is a drive specifier, so `c:myfile.o'
3236 is a simple file specification, not `myfile.o' within an archive called
3237 `c'. `archive:file' filespecs may also be used within an
3238 `EXCLUDE_FILE' list, but may not appear in other linker script
3239 contexts. For instance, you cannot extract a file from an archive by
3240 using `archive:file' in an `INPUT' command.
3242 If you use a file name without a list of sections, then all sections
3243 in the input file will be included in the output section. This is not
3244 commonly done, but it may by useful on occasion. For example:
3247 When you use a file name which is not an `archive:file' specifier
3248 and does not contain any wild card characters, the linker will first
3249 see if you also specified the file name on the linker command line or
3250 in an `INPUT' command. If you did not, the linker will attempt to open
3251 the file as an input file, as though it appeared on the command line.
3252 Note that this differs from an `INPUT' command, because the linker will
3253 not search for the file in the archive search path.
3256 File: ld.info, Node: Input Section Wildcards, Next: Input Section Common, Prev: Input Section Basics, Up: Input Section
3258 3.6.4.2 Input Section Wildcard Patterns
3259 .......................................
3261 In an input section description, either the file name or the section
3262 name or both may be wildcard patterns.
3264 The file name of `*' seen in many examples is a simple wildcard
3265 pattern for the file name.
3267 The wildcard patterns are like those used by the Unix shell.
3270 matches any number of characters
3273 matches any single character
3276 matches a single instance of any of the CHARS; the `-' character
3277 may be used to specify a range of characters, as in `[a-z]' to
3278 match any lower case letter
3281 quotes the following character
3283 When a file name is matched with a wildcard, the wildcard characters
3284 will not match a `/' character (used to separate directory names on
3285 Unix). A pattern consisting of a single `*' character is an exception;
3286 it will always match any file name, whether it contains a `/' or not.
3287 In a section name, the wildcard characters will match a `/' character.
3289 File name wildcard patterns only match files which are explicitly
3290 specified on the command line or in an `INPUT' command. The linker
3291 does not search directories to expand wildcards.
3293 If a file name matches more than one wildcard pattern, or if a file
3294 name appears explicitly and is also matched by a wildcard pattern, the
3295 linker will use the first match in the linker script. For example, this
3296 sequence of input section descriptions is probably in error, because the
3297 `data.o' rule will not be used:
3298 .data : { *(.data) }
3299 .data1 : { data.o(.data) }
3301 Normally, the linker will place files and sections matched by
3302 wildcards in the order in which they are seen during the link. You can
3303 change this by using the `SORT_BY_NAME' keyword, which appears before a
3304 wildcard pattern in parentheses (e.g., `SORT_BY_NAME(.text*)'). When
3305 the `SORT_BY_NAME' keyword is used, the linker will sort the files or
3306 sections into ascending order by name before placing them in the output
3309 `SORT_BY_ALIGNMENT' is very similar to `SORT_BY_NAME'. The
3310 difference is `SORT_BY_ALIGNMENT' will sort sections into ascending
3311 order by alignment before placing them in the output file.
3313 `SORT_BY_INIT_PRIORITY' is very similar to `SORT_BY_NAME'. The
3314 difference is `SORT_BY_INIT_PRIORITY' will sort sections into ascending
3315 order by numerical value of the GCC init_priority attribute encoded in
3316 the section name before placing them in the output file.
3318 `SORT' is an alias for `SORT_BY_NAME'.
3320 When there are nested section sorting commands in linker script,
3321 there can be at most 1 level of nesting for section sorting commands.
3323 1. `SORT_BY_NAME' (`SORT_BY_ALIGNMENT' (wildcard section pattern)).
3324 It will sort the input sections by name first, then by alignment
3325 if 2 sections have the same name.
3327 2. `SORT_BY_ALIGNMENT' (`SORT_BY_NAME' (wildcard section pattern)).
3328 It will sort the input sections by alignment first, then by name
3329 if 2 sections have the same alignment.
3331 3. `SORT_BY_NAME' (`SORT_BY_NAME' (wildcard section pattern)) is
3332 treated the same as `SORT_BY_NAME' (wildcard section pattern).
3334 4. `SORT_BY_ALIGNMENT' (`SORT_BY_ALIGNMENT' (wildcard section
3335 pattern)) is treated the same as `SORT_BY_ALIGNMENT' (wildcard
3338 5. All other nested section sorting commands are invalid.
3340 When both command line section sorting option and linker script
3341 section sorting command are used, section sorting command always takes
3342 precedence over the command line option.
3344 If the section sorting command in linker script isn't nested, the
3345 command line option will make the section sorting command to be treated
3346 as nested sorting command.
3348 1. `SORT_BY_NAME' (wildcard section pattern ) with `--sort-sections
3349 alignment' is equivalent to `SORT_BY_NAME' (`SORT_BY_ALIGNMENT'
3350 (wildcard section pattern)).
3352 2. `SORT_BY_ALIGNMENT' (wildcard section pattern) with
3353 `--sort-section name' is equivalent to `SORT_BY_ALIGNMENT'
3354 (`SORT_BY_NAME' (wildcard section pattern)).
3356 If the section sorting command in linker script is nested, the
3357 command line option will be ignored.
3359 `SORT_NONE' disables section sorting by ignoring the command line
3360 section sorting option.
3362 If you ever get confused about where input sections are going, use
3363 the `-M' linker option to generate a map file. The map file shows
3364 precisely how input sections are mapped to output sections.
3366 This example shows how wildcard patterns might be used to partition
3367 files. This linker script directs the linker to place all `.text'
3368 sections in `.text' and all `.bss' sections in `.bss'. The linker will
3369 place the `.data' section from all files beginning with an upper case
3370 character in `.DATA'; for all other files, the linker will place the
3371 `.data' section in `.data'.
3373 .text : { *(.text) }
3374 .DATA : { [A-Z]*(.data) }
3375 .data : { *(.data) }
3380 File: ld.info, Node: Input Section Common, Next: Input Section Keep, Prev: Input Section Wildcards, Up: Input Section
3382 3.6.4.3 Input Section for Common Symbols
3383 ........................................
3385 A special notation is needed for common symbols, because in many object
3386 file formats common symbols do not have a particular input section. The
3387 linker treats common symbols as though they are in an input section
3390 You may use file names with the `COMMON' section just as with any
3391 other input sections. You can use this to place common symbols from a
3392 particular input file in one section while common symbols from other
3393 input files are placed in another section.
3395 In most cases, common symbols in input files will be placed in the
3396 `.bss' section in the output file. For example:
3397 .bss { *(.bss) *(COMMON) }
3399 Some object file formats have more than one type of common symbol.
3400 For example, the MIPS ELF object file format distinguishes standard
3401 common symbols and small common symbols. In this case, the linker will
3402 use a different special section name for other types of common symbols.
3403 In the case of MIPS ELF, the linker uses `COMMON' for standard common
3404 symbols and `.scommon' for small common symbols. This permits you to
3405 map the different types of common symbols into memory at different
3408 You will sometimes see `[COMMON]' in old linker scripts. This
3409 notation is now considered obsolete. It is equivalent to `*(COMMON)'.
3412 File: ld.info, Node: Input Section Keep, Next: Input Section Example, Prev: Input Section Common, Up: Input Section
3414 3.6.4.4 Input Section and Garbage Collection
3415 ............................................
3417 When link-time garbage collection is in use (`--gc-sections'), it is
3418 often useful to mark sections that should not be eliminated. This is
3419 accomplished by surrounding an input section's wildcard entry with
3420 `KEEP()', as in `KEEP(*(.init))' or `KEEP(SORT_BY_NAME(*)(.ctors))'.
3423 File: ld.info, Node: Input Section Example, Prev: Input Section Keep, Up: Input Section
3425 3.6.4.5 Input Section Example
3426 .............................
3428 The following example is a complete linker script. It tells the linker
3429 to read all of the sections from file `all.o' and place them at the
3430 start of output section `outputa' which starts at location `0x10000'.
3431 All of section `.input1' from file `foo.o' follows immediately, in the
3432 same output section. All of section `.input2' from `foo.o' goes into
3433 output section `outputb', followed by section `.input1' from `foo1.o'.
3434 All of the remaining `.input1' and `.input2' sections from any files
3435 are written to output section `outputc'.
3456 File: ld.info, Node: Output Section Data, Next: Output Section Keywords, Prev: Input Section, Up: SECTIONS
3458 3.6.5 Output Section Data
3459 -------------------------
3461 You can include explicit bytes of data in an output section by using
3462 `BYTE', `SHORT', `LONG', `QUAD', or `SQUAD' as an output section
3463 command. Each keyword is followed by an expression in parentheses
3464 providing the value to store (*note Expressions::). The value of the
3465 expression is stored at the current value of the location counter.
3467 The `BYTE', `SHORT', `LONG', and `QUAD' commands store one, two,
3468 four, and eight bytes (respectively). After storing the bytes, the
3469 location counter is incremented by the number of bytes stored.
3471 For example, this will store the byte 1 followed by the four byte
3472 value of the symbol `addr':
3476 When using a 64 bit host or target, `QUAD' and `SQUAD' are the same;
3477 they both store an 8 byte, or 64 bit, value. When both host and target
3478 are 32 bits, an expression is computed as 32 bits. In this case `QUAD'
3479 stores a 32 bit value zero extended to 64 bits, and `SQUAD' stores a 32
3480 bit value sign extended to 64 bits.
3482 If the object file format of the output file has an explicit
3483 endianness, which is the normal case, the value will be stored in that
3484 endianness. When the object file format does not have an explicit
3485 endianness, as is true of, for example, S-records, the value will be
3486 stored in the endianness of the first input object file.
3488 Note--these commands only work inside a section description and not
3489 between them, so the following will produce an error from the linker:
3490 SECTIONS { .text : { *(.text) } LONG(1) .data : { *(.data) } }
3491 whereas this will work:
3492 SECTIONS { .text : { *(.text) ; LONG(1) } .data : { *(.data) } }
3494 You may use the `FILL' command to set the fill pattern for the
3495 current section. It is followed by an expression in parentheses. Any
3496 otherwise unspecified regions of memory within the section (for example,
3497 gaps left due to the required alignment of input sections) are filled
3498 with the value of the expression, repeated as necessary. A `FILL'
3499 statement covers memory locations after the point at which it occurs in
3500 the section definition; by including more than one `FILL' statement,
3501 you can have different fill patterns in different parts of an output
3504 This example shows how to fill unspecified regions of memory with the
3508 The `FILL' command is similar to the `=FILLEXP' output section
3509 attribute, but it only affects the part of the section following the
3510 `FILL' command, rather than the entire section. If both are used, the
3511 `FILL' command takes precedence. *Note Output Section Fill::, for
3512 details on the fill expression.
3515 File: ld.info, Node: Output Section Keywords, Next: Output Section Discarding, Prev: Output Section Data, Up: SECTIONS
3517 3.6.6 Output Section Keywords
3518 -----------------------------
3520 There are a couple of keywords which can appear as output section
3523 `CREATE_OBJECT_SYMBOLS'
3524 The command tells the linker to create a symbol for each input
3525 file. The name of each symbol will be the name of the
3526 corresponding input file. The section of each symbol will be the
3527 output section in which the `CREATE_OBJECT_SYMBOLS' command
3530 This is conventional for the a.out object file format. It is not
3531 normally used for any other object file format.
3534 When linking using the a.out object file format, the linker uses an
3535 unusual set construct to support C++ global constructors and
3536 destructors. When linking object file formats which do not support
3537 arbitrary sections, such as ECOFF and XCOFF, the linker will
3538 automatically recognize C++ global constructors and destructors by
3539 name. For these object file formats, the `CONSTRUCTORS' command
3540 tells the linker to place constructor information in the output
3541 section where the `CONSTRUCTORS' command appears. The
3542 `CONSTRUCTORS' command is ignored for other object file formats.
3544 The symbol `__CTOR_LIST__' marks the start of the global
3545 constructors, and the symbol `__CTOR_END__' marks the end.
3546 Similarly, `__DTOR_LIST__' and `__DTOR_END__' mark the start and
3547 end of the global destructors. The first word in the list is the
3548 number of entries, followed by the address of each constructor or
3549 destructor, followed by a zero word. The compiler must arrange to
3550 actually run the code. For these object file formats GNU C++
3551 normally calls constructors from a subroutine `__main'; a call to
3552 `__main' is automatically inserted into the startup code for
3553 `main'. GNU C++ normally runs destructors either by using
3554 `atexit', or directly from the function `exit'.
3556 For object file formats such as `COFF' or `ELF' which support
3557 arbitrary section names, GNU C++ will normally arrange to put the
3558 addresses of global constructors and destructors into the `.ctors'
3559 and `.dtors' sections. Placing the following sequence into your
3560 linker script will build the sort of table which the GNU C++
3561 runtime code expects to see.
3564 LONG((__CTOR_END__ - __CTOR_LIST__) / 4 - 2)
3569 LONG((__DTOR_END__ - __DTOR_LIST__) / 4 - 2)
3574 If you are using the GNU C++ support for initialization priority,
3575 which provides some control over the order in which global
3576 constructors are run, you must sort the constructors at link time
3577 to ensure that they are executed in the correct order. When using
3578 the `CONSTRUCTORS' command, use `SORT_BY_NAME(CONSTRUCTORS)'
3579 instead. When using the `.ctors' and `.dtors' sections, use
3580 `*(SORT_BY_NAME(.ctors))' and `*(SORT_BY_NAME(.dtors))' instead of
3581 just `*(.ctors)' and `*(.dtors)'.
3583 Normally the compiler and linker will handle these issues
3584 automatically, and you will not need to concern yourself with
3585 them. However, you may need to consider this if you are using C++
3586 and writing your own linker scripts.
3590 File: ld.info, Node: Output Section Discarding, Next: Output Section Attributes, Prev: Output Section Keywords, Up: SECTIONS
3592 3.6.7 Output Section Discarding
3593 -------------------------------
3595 The linker will not create output sections with no contents. This is
3596 for convenience when referring to input sections that may or may not be
3597 present in any of the input files. For example:
3599 will only create a `.foo' section in the output file if there is a
3600 `.foo' section in at least one input file, and if the input sections
3601 are not all empty. Other link script directives that allocate space in
3602 an output section will also create the output section.
3604 The linker will ignore address assignments (*note Output Section
3605 Address::) on discarded output sections, except when the linker script
3606 defines symbols in the output section. In that case the linker will
3607 obey the address assignments, possibly advancing dot even though the
3608 section is discarded.
3610 The special output section name `/DISCARD/' may be used to discard
3611 input sections. Any input sections which are assigned to an output
3612 section named `/DISCARD/' are not included in the output file.
3615 File: ld.info, Node: Output Section Attributes, Next: Overlay Description, Prev: Output Section Discarding, Up: SECTIONS
3617 3.6.8 Output Section Attributes
3618 -------------------------------
3620 We showed above that the full description of an output section looked
3623 SECTION [ADDRESS] [(TYPE)] :
3625 [ALIGN(SECTION_ALIGN)]
3626 [SUBALIGN(SUBSECTION_ALIGN)]
3629 OUTPUT-SECTION-COMMAND
3630 OUTPUT-SECTION-COMMAND
3632 } [>REGION] [AT>LMA_REGION] [:PHDR :PHDR ...] [=FILLEXP]
3634 We've already described SECTION, ADDRESS, and
3635 OUTPUT-SECTION-COMMAND. In this section we will describe the remaining
3640 * Output Section Type:: Output section type
3641 * Output Section LMA:: Output section LMA
3642 * Forced Output Alignment:: Forced Output Alignment
3643 * Forced Input Alignment:: Forced Input Alignment
3644 * Output Section Constraint:: Output section constraint
3645 * Output Section Region:: Output section region
3646 * Output Section Phdr:: Output section phdr
3647 * Output Section Fill:: Output section fill
3650 File: ld.info, Node: Output Section Type, Next: Output Section LMA, Up: Output Section Attributes
3652 3.6.8.1 Output Section Type
3653 ...........................
3655 Each output section may have a type. The type is a keyword in
3656 parentheses. The following types are defined:
3659 The section should be marked as not loadable, so that it will not
3660 be loaded into memory when the program is run.
3666 These type names are supported for backward compatibility, and are
3667 rarely used. They all have the same effect: the section should be
3668 marked as not allocatable, so that no memory is allocated for the
3669 section when the program is run.
3671 The linker normally sets the attributes of an output section based on
3672 the input sections which map into it. You can override this by using
3673 the section type. For example, in the script sample below, the `ROM'
3674 section is addressed at memory location `0' and does not need to be
3675 loaded when the program is run.
3677 ROM 0 (NOLOAD) : { ... }
3682 File: ld.info, Node: Output Section LMA, Next: Forced Output Alignment, Prev: Output Section Type, Up: Output Section Attributes
3684 3.6.8.2 Output Section LMA
3685 ..........................
3687 Every section has a virtual address (VMA) and a load address (LMA); see
3688 *Note Basic Script Concepts::. The virtual address is specified by the
3689 *note Output Section Address:: described earlier. The load address is
3690 specified by the `AT' or `AT>' keywords. Specifying a load address is
3693 The `AT' keyword takes an expression as an argument. This specifies
3694 the exact load address of the section. The `AT>' keyword takes the
3695 name of a memory region as an argument. *Note MEMORY::. The load
3696 address of the section is set to the next free address in the region,
3697 aligned to the section's alignment requirements.
3699 If neither `AT' nor `AT>' is specified for an allocatable section,
3700 the linker will use the following heuristic to determine the load
3703 * If the section has a specific VMA address, then this is used as
3704 the LMA address as well.
3706 * If the section is not allocatable then its LMA is set to its VMA.
3708 * Otherwise if a memory region can be found that is compatible with
3709 the current section, and this region contains at least one
3710 section, then the LMA is set so the difference between the VMA and
3711 LMA is the same as the difference between the VMA and LMA of the
3712 last section in the located region.
3714 * If no memory regions have been declared then a default region that
3715 covers the entire address space is used in the previous step.
3717 * If no suitable region could be found, or there was no previous
3718 section then the LMA is set equal to the VMA.
3720 This feature is designed to make it easy to build a ROM image. For
3721 example, the following linker script creates three output sections: one
3722 called `.text', which starts at `0x1000', one called `.mdata', which is
3723 loaded at the end of the `.text' section even though its VMA is
3724 `0x2000', and one called `.bss' to hold uninitialized data at address
3725 `0x3000'. The symbol `_data' is defined with the value `0x2000', which
3726 shows that the location counter holds the VMA value, not the LMA value.
3730 .text 0x1000 : { *(.text) _etext = . ; }
3732 AT ( ADDR (.text) + SIZEOF (.text) )
3733 { _data = . ; *(.data); _edata = . ; }
3735 { _bstart = . ; *(.bss) *(COMMON) ; _bend = . ;}
3738 The run-time initialization code for use with a program generated
3739 with this linker script would include something like the following, to
3740 copy the initialized data from the ROM image to its runtime address.
3741 Notice how this code takes advantage of the symbols defined by the
3744 extern char _etext, _data, _edata, _bstart, _bend;
3745 char *src = &_etext;
3748 /* ROM has data at end of text; copy it. */
3749 while (dst < &_edata)
3753 for (dst = &_bstart; dst< &_bend; dst++)
3757 File: ld.info, Node: Forced Output Alignment, Next: Forced Input Alignment, Prev: Output Section LMA, Up: Output Section Attributes
3759 3.6.8.3 Forced Output Alignment
3760 ...............................
3762 You can increase an output section's alignment by using ALIGN.
3765 File: ld.info, Node: Forced Input Alignment, Next: Output Section Constraint, Prev: Forced Output Alignment, Up: Output Section Attributes
3767 3.6.8.4 Forced Input Alignment
3768 ..............................
3770 You can force input section alignment within an output section by using
3771 SUBALIGN. The value specified overrides any alignment given by input
3772 sections, whether larger or smaller.
3775 File: ld.info, Node: Output Section Constraint, Next: Output Section Region, Prev: Forced Input Alignment, Up: Output Section Attributes
3777 3.6.8.5 Output Section Constraint
3778 .................................
3780 You can specify that an output section should only be created if all of
3781 its input sections are read-only or all of its input sections are
3782 read-write by using the keyword `ONLY_IF_RO' and `ONLY_IF_RW'
3786 File: ld.info, Node: Output Section Region, Next: Output Section Phdr, Prev: Output Section Constraint, Up: Output Section Attributes
3788 3.6.8.6 Output Section Region
3789 .............................
3791 You can assign a section to a previously defined region of memory by
3792 using `>REGION'. *Note MEMORY::.
3794 Here is a simple example:
3795 MEMORY { rom : ORIGIN = 0x1000, LENGTH = 0x1000 }
3796 SECTIONS { ROM : { *(.text) } >rom }
3799 File: ld.info, Node: Output Section Phdr, Next: Output Section Fill, Prev: Output Section Region, Up: Output Section Attributes
3801 3.6.8.7 Output Section Phdr
3802 ...........................
3804 You can assign a section to a previously defined program segment by
3805 using `:PHDR'. *Note PHDRS::. If a section is assigned to one or more
3806 segments, then all subsequent allocated sections will be assigned to
3807 those segments as well, unless they use an explicitly `:PHDR' modifier.
3808 You can use `:NONE' to tell the linker to not put the section in any
3811 Here is a simple example:
3812 PHDRS { text PT_LOAD ; }
3813 SECTIONS { .text : { *(.text) } :text }
3816 File: ld.info, Node: Output Section Fill, Prev: Output Section Phdr, Up: Output Section Attributes
3818 3.6.8.8 Output Section Fill
3819 ...........................
3821 You can set the fill pattern for an entire section by using `=FILLEXP'.
3822 FILLEXP is an expression (*note Expressions::). Any otherwise
3823 unspecified regions of memory within the output section (for example,
3824 gaps left due to the required alignment of input sections) will be
3825 filled with the value, repeated as necessary. If the fill expression
3826 is a simple hex number, ie. a string of hex digit starting with `0x'
3827 and without a trailing `k' or `M', then an arbitrarily long sequence of
3828 hex digits can be used to specify the fill pattern; Leading zeros
3829 become part of the pattern too. For all other cases, including extra
3830 parentheses or a unary `+', the fill pattern is the four least
3831 significant bytes of the value of the expression. In all cases, the
3832 number is big-endian.
3834 You can also change the fill value with a `FILL' command in the
3835 output section commands; (*note Output Section Data::).
3837 Here is a simple example:
3838 SECTIONS { .text : { *(.text) } =0x90909090 }
3841 File: ld.info, Node: Overlay Description, Prev: Output Section Attributes, Up: SECTIONS
3843 3.6.9 Overlay Description
3844 -------------------------
3846 An overlay description provides an easy way to describe sections which
3847 are to be loaded as part of a single memory image but are to be run at
3848 the same memory address. At run time, some sort of overlay manager will
3849 copy the overlaid sections in and out of the runtime memory address as
3850 required, perhaps by simply manipulating addressing bits. This approach
3851 can be useful, for example, when a certain region of memory is faster
3854 Overlays are described using the `OVERLAY' command. The `OVERLAY'
3855 command is used within a `SECTIONS' command, like an output section
3856 description. The full syntax of the `OVERLAY' command is as follows:
3857 OVERLAY [START] : [NOCROSSREFS] [AT ( LDADDR )]
3861 OUTPUT-SECTION-COMMAND
3862 OUTPUT-SECTION-COMMAND
3864 } [:PHDR...] [=FILL]
3867 OUTPUT-SECTION-COMMAND
3868 OUTPUT-SECTION-COMMAND
3870 } [:PHDR...] [=FILL]
3872 } [>REGION] [:PHDR...] [=FILL]
3874 Everything is optional except `OVERLAY' (a keyword), and each
3875 section must have a name (SECNAME1 and SECNAME2 above). The section
3876 definitions within the `OVERLAY' construct are identical to those
3877 within the general `SECTIONS' contruct (*note SECTIONS::), except that
3878 no addresses and no memory regions may be defined for sections within
3881 The sections are all defined with the same starting address. The
3882 load addresses of the sections are arranged such that they are
3883 consecutive in memory starting at the load address used for the
3884 `OVERLAY' as a whole (as with normal section definitions, the load
3885 address is optional, and defaults to the start address; the start
3886 address is also optional, and defaults to the current value of the
3889 If the `NOCROSSREFS' keyword is used, and there any references among
3890 the sections, the linker will report an error. Since the sections all
3891 run at the same address, it normally does not make sense for one
3892 section to refer directly to another. *Note NOCROSSREFS: Miscellaneous
3895 For each section within the `OVERLAY', the linker automatically
3896 provides two symbols. The symbol `__load_start_SECNAME' is defined as
3897 the starting load address of the section. The symbol
3898 `__load_stop_SECNAME' is defined as the final load address of the
3899 section. Any characters within SECNAME which are not legal within C
3900 identifiers are removed. C (or assembler) code may use these symbols
3901 to move the overlaid sections around as necessary.
3903 At the end of the overlay, the value of the location counter is set
3904 to the start address of the overlay plus the size of the largest
3907 Here is an example. Remember that this would appear inside a
3908 `SECTIONS' construct.
3909 OVERLAY 0x1000 : AT (0x4000)
3911 .text0 { o1/*.o(.text) }
3912 .text1 { o2/*.o(.text) }
3914 This will define both `.text0' and `.text1' to start at address
3915 0x1000. `.text0' will be loaded at address 0x4000, and `.text1' will
3916 be loaded immediately after `.text0'. The following symbols will be
3917 defined if referenced: `__load_start_text0', `__load_stop_text0',
3918 `__load_start_text1', `__load_stop_text1'.
3920 C code to copy overlay `.text1' into the overlay area might look
3923 extern char __load_start_text1, __load_stop_text1;
3924 memcpy ((char *) 0x1000, &__load_start_text1,
3925 &__load_stop_text1 - &__load_start_text1);
3927 Note that the `OVERLAY' command is just syntactic sugar, since
3928 everything it does can be done using the more basic commands. The above
3929 example could have been written identically as follows.
3931 .text0 0x1000 : AT (0x4000) { o1/*.o(.text) }
3932 PROVIDE (__load_start_text0 = LOADADDR (.text0));
3933 PROVIDE (__load_stop_text0 = LOADADDR (.text0) + SIZEOF (.text0));
3934 .text1 0x1000 : AT (0x4000 + SIZEOF (.text0)) { o2/*.o(.text) }
3935 PROVIDE (__load_start_text1 = LOADADDR (.text1));
3936 PROVIDE (__load_stop_text1 = LOADADDR (.text1) + SIZEOF (.text1));
3937 . = 0x1000 + MAX (SIZEOF (.text0), SIZEOF (.text1));
3940 File: ld.info, Node: MEMORY, Next: PHDRS, Prev: SECTIONS, Up: Scripts
3945 The linker's default configuration permits allocation of all available
3946 memory. You can override this by using the `MEMORY' command.
3948 The `MEMORY' command describes the location and size of blocks of
3949 memory in the target. You can use it to describe which memory regions
3950 may be used by the linker, and which memory regions it must avoid. You
3951 can then assign sections to particular memory regions. The linker will
3952 set section addresses based on the memory regions, and will warn about
3953 regions that become too full. The linker will not shuffle sections
3954 around to fit into the available regions.
3956 A linker script may contain at most one use of the `MEMORY' command.
3957 However, you can define as many blocks of memory within it as you
3958 wish. The syntax is:
3961 NAME [(ATTR)] : ORIGIN = ORIGIN, LENGTH = LEN
3965 The NAME is a name used in the linker script to refer to the region.
3966 The region name has no meaning outside of the linker script. Region
3967 names are stored in a separate name space, and will not conflict with
3968 symbol names, file names, or section names. Each memory region must
3969 have a distinct name within the `MEMORY' command. However you can add
3970 later alias names to existing memory regions with the *Note
3971 REGION_ALIAS:: command.
3973 The ATTR string is an optional list of attributes that specify
3974 whether to use a particular memory region for an input section which is
3975 not explicitly mapped in the linker script. As described in *Note
3976 SECTIONS::, if you do not specify an output section for some input
3977 section, the linker will create an output section with the same name as
3978 the input section. If you define region attributes, the linker will use
3979 them to select the memory region for the output section that it creates.
3981 The ATTR string must consist only of the following characters:
4001 Invert the sense of any of the attributes that follow
4003 If a unmapped section matches any of the listed attributes other than
4004 `!', it will be placed in the memory region. The `!' attribute
4005 reverses this test, so that an unmapped section will be placed in the
4006 memory region only if it does not match any of the listed attributes.
4008 The ORIGIN is an numerical expression for the start address of the
4009 memory region. The expression must evaluate to a constant and it
4010 cannot involve any symbols. The keyword `ORIGIN' may be abbreviated to
4011 `org' or `o' (but not, for example, `ORG').
4013 The LEN is an expression for the size in bytes of the memory region.
4014 As with the ORIGIN expression, the expression must be numerical only
4015 and must evaluate to a constant. The keyword `LENGTH' may be
4016 abbreviated to `len' or `l'.
4018 In the following example, we specify that there are two memory
4019 regions available for allocation: one starting at `0' for 256 kilobytes,
4020 and the other starting at `0x40000000' for four megabytes. The linker
4021 will place into the `rom' memory region every section which is not
4022 explicitly mapped into a memory region, and is either read-only or
4023 executable. The linker will place other sections which are not
4024 explicitly mapped into a memory region into the `ram' memory region.
4028 rom (rx) : ORIGIN = 0, LENGTH = 256K
4029 ram (!rx) : org = 0x40000000, l = 4M
4032 Once you define a memory region, you can direct the linker to place
4033 specific output sections into that memory region by using the `>REGION'
4034 output section attribute. For example, if you have a memory region
4035 named `mem', you would use `>mem' in the output section definition.
4036 *Note Output Section Region::. If no address was specified for the
4037 output section, the linker will set the address to the next available
4038 address within the memory region. If the combined output sections
4039 directed to a memory region are too large for the region, the linker
4040 will issue an error message.
4042 It is possible to access the origin and length of a memory in an
4043 expression via the `ORIGIN(MEMORY)' and `LENGTH(MEMORY)' functions:
4045 _fstack = ORIGIN(ram) + LENGTH(ram) - 4;
4048 File: ld.info, Node: PHDRS, Next: VERSION, Prev: MEMORY, Up: Scripts
4053 The ELF object file format uses "program headers", also knows as
4054 "segments". The program headers describe how the program should be
4055 loaded into memory. You can print them out by using the `objdump'
4056 program with the `-p' option.
4058 When you run an ELF program on a native ELF system, the system loader
4059 reads the program headers in order to figure out how to load the
4060 program. This will only work if the program headers are set correctly.
4061 This manual does not describe the details of how the system loader
4062 interprets program headers; for more information, see the ELF ABI.
4064 The linker will create reasonable program headers by default.
4065 However, in some cases, you may need to specify the program headers more
4066 precisely. You may use the `PHDRS' command for this purpose. When the
4067 linker sees the `PHDRS' command in the linker script, it will not
4068 create any program headers other than the ones specified.
4070 The linker only pays attention to the `PHDRS' command when
4071 generating an ELF output file. In other cases, the linker will simply
4074 This is the syntax of the `PHDRS' command. The words `PHDRS',
4075 `FILEHDR', `AT', and `FLAGS' are keywords.
4079 NAME TYPE [ FILEHDR ] [ PHDRS ] [ AT ( ADDRESS ) ]
4080 [ FLAGS ( FLAGS ) ] ;
4083 The NAME is used only for reference in the `SECTIONS' command of the
4084 linker script. It is not put into the output file. Program header
4085 names are stored in a separate name space, and will not conflict with
4086 symbol names, file names, or section names. Each program header must
4087 have a distinct name. The headers are processed in order and it is
4088 usual for them to map to sections in ascending load address order.
4090 Certain program header types describe segments of memory which the
4091 system loader will load from the file. In the linker script, you
4092 specify the contents of these segments by placing allocatable output
4093 sections in the segments. You use the `:PHDR' output section attribute
4094 to place a section in a particular segment. *Note Output Section
4097 It is normal to put certain sections in more than one segment. This
4098 merely implies that one segment of memory contains another. You may
4099 repeat `:PHDR', using it once for each segment which should contain the
4102 If you place a section in one or more segments using `:PHDR', then
4103 the linker will place all subsequent allocatable sections which do not
4104 specify `:PHDR' in the same segments. This is for convenience, since
4105 generally a whole set of contiguous sections will be placed in a single
4106 segment. You can use `:NONE' to override the default segment and tell
4107 the linker to not put the section in any segment at all.
4109 You may use the `FILEHDR' and `PHDRS' keywords after the program
4110 header type to further describe the contents of the segment. The
4111 `FILEHDR' keyword means that the segment should include the ELF file
4112 header. The `PHDRS' keyword means that the segment should include the
4113 ELF program headers themselves. If applied to a loadable segment
4114 (`PT_LOAD'), all prior loadable segments must have one of these
4117 The TYPE may be one of the following. The numbers indicate the
4118 value of the keyword.
4121 Indicates an unused program header.
4124 Indicates that this program header describes a segment to be
4125 loaded from the file.
4128 Indicates a segment where dynamic linking information can be found.
4131 Indicates a segment where the name of the program interpreter may
4135 Indicates a segment holding note information.
4138 A reserved program header type, defined but not specified by the
4142 Indicates a segment where the program headers may be found.
4145 An expression giving the numeric type of the program header. This
4146 may be used for types not defined above.
4148 You can specify that a segment should be loaded at a particular
4149 address in memory by using an `AT' expression. This is identical to the
4150 `AT' command used as an output section attribute (*note Output Section
4151 LMA::). The `AT' command for a program header overrides the output
4154 The linker will normally set the segment flags based on the sections
4155 which comprise the segment. You may use the `FLAGS' keyword to
4156 explicitly specify the segment flags. The value of FLAGS must be an
4157 integer. It is used to set the `p_flags' field of the program header.
4159 Here is an example of `PHDRS'. This shows a typical set of program
4160 headers used on a native ELF system.
4164 headers PT_PHDR PHDRS ;
4166 text PT_LOAD FILEHDR PHDRS ;
4168 dynamic PT_DYNAMIC ;
4174 .interp : { *(.interp) } :text :interp
4175 .text : { *(.text) } :text
4176 .rodata : { *(.rodata) } /* defaults to :text */
4178 . = . + 0x1000; /* move to a new page in memory */
4179 .data : { *(.data) } :data
4180 .dynamic : { *(.dynamic) } :data :dynamic
4185 File: ld.info, Node: VERSION, Next: Expressions, Prev: PHDRS, Up: Scripts
4190 The linker supports symbol versions when using ELF. Symbol versions are
4191 only useful when using shared libraries. The dynamic linker can use
4192 symbol versions to select a specific version of a function when it runs
4193 a program that may have been linked against an earlier version of the
4196 You can include a version script directly in the main linker script,
4197 or you can supply the version script as an implicit linker script. You
4198 can also use the `--version-script' linker option.
4200 The syntax of the `VERSION' command is simply
4201 VERSION { version-script-commands }
4203 The format of the version script commands is identical to that used
4204 by Sun's linker in Solaris 2.5. The version script defines a tree of
4205 version nodes. You specify the node names and interdependencies in the
4206 version script. You can specify which symbols are bound to which
4207 version nodes, and you can reduce a specified set of symbols to local
4208 scope so that they are not globally visible outside of the shared
4211 The easiest way to demonstrate the version script language is with a
4235 This example version script defines three version nodes. The first
4236 version node defined is `VERS_1.1'; it has no other dependencies. The
4237 script binds the symbol `foo1' to `VERS_1.1'. It reduces a number of
4238 symbols to local scope so that they are not visible outside of the
4239 shared library; this is done using wildcard patterns, so that any
4240 symbol whose name begins with `old', `original', or `new' is matched.
4241 The wildcard patterns available are the same as those used in the shell
4242 when matching filenames (also known as "globbing"). However, if you
4243 specify the symbol name inside double quotes, then the name is treated
4244 as literal, rather than as a glob pattern.
4246 Next, the version script defines node `VERS_1.2'. This node depends
4247 upon `VERS_1.1'. The script binds the symbol `foo2' to the version
4250 Finally, the version script defines node `VERS_2.0'. This node
4251 depends upon `VERS_1.2'. The scripts binds the symbols `bar1' and
4252 `bar2' are bound to the version node `VERS_2.0'.
4254 When the linker finds a symbol defined in a library which is not
4255 specifically bound to a version node, it will effectively bind it to an
4256 unspecified base version of the library. You can bind all otherwise
4257 unspecified symbols to a given version node by using `global: *;'
4258 somewhere in the version script. Note that it's slightly crazy to use
4259 wildcards in a global spec except on the last version node. Global
4260 wildcards elsewhere run the risk of accidentally adding symbols to the
4261 set exported for an old version. That's wrong since older versions
4262 ought to have a fixed set of symbols.
4264 The names of the version nodes have no specific meaning other than
4265 what they might suggest to the person reading them. The `2.0' version
4266 could just as well have appeared in between `1.1' and `1.2'. However,
4267 this would be a confusing way to write a version script.
4269 Node name can be omitted, provided it is the only version node in
4270 the version script. Such version script doesn't assign any versions to
4271 symbols, only selects which symbols will be globally visible out and
4274 { global: foo; bar; local: *; };
4276 When you link an application against a shared library that has
4277 versioned symbols, the application itself knows which version of each
4278 symbol it requires, and it also knows which version nodes it needs from
4279 each shared library it is linked against. Thus at runtime, the dynamic
4280 loader can make a quick check to make sure that the libraries you have
4281 linked against do in fact supply all of the version nodes that the
4282 application will need to resolve all of the dynamic symbols. In this
4283 way it is possible for the dynamic linker to know with certainty that
4284 all external symbols that it needs will be resolvable without having to
4285 search for each symbol reference.
4287 The symbol versioning is in effect a much more sophisticated way of
4288 doing minor version checking that SunOS does. The fundamental problem
4289 that is being addressed here is that typically references to external
4290 functions are bound on an as-needed basis, and are not all bound when
4291 the application starts up. If a shared library is out of date, a
4292 required interface may be missing; when the application tries to use
4293 that interface, it may suddenly and unexpectedly fail. With symbol
4294 versioning, the user will get a warning when they start their program if
4295 the libraries being used with the application are too old.
4297 There are several GNU extensions to Sun's versioning approach. The
4298 first of these is the ability to bind a symbol to a version node in the
4299 source file where the symbol is defined instead of in the versioning
4300 script. This was done mainly to reduce the burden on the library
4301 maintainer. You can do this by putting something like:
4302 __asm__(".symver original_foo,foo@VERS_1.1");
4303 in the C source file. This renames the function `original_foo' to
4304 be an alias for `foo' bound to the version node `VERS_1.1'. The
4305 `local:' directive can be used to prevent the symbol `original_foo'
4306 from being exported. A `.symver' directive takes precedence over a
4309 The second GNU extension is to allow multiple versions of the same
4310 function to appear in a given shared library. In this way you can make
4311 an incompatible change to an interface without increasing the major
4312 version number of the shared library, while still allowing applications
4313 linked against the old interface to continue to function.
4315 To do this, you must use multiple `.symver' directives in the source
4316 file. Here is an example:
4318 __asm__(".symver original_foo,foo@");
4319 __asm__(".symver old_foo,foo@VERS_1.1");
4320 __asm__(".symver old_foo1,foo@VERS_1.2");
4321 __asm__(".symver new_foo,foo@@VERS_2.0");
4323 In this example, `foo@' represents the symbol `foo' bound to the
4324 unspecified base version of the symbol. The source file that contains
4325 this example would define 4 C functions: `original_foo', `old_foo',
4326 `old_foo1', and `new_foo'.
4328 When you have multiple definitions of a given symbol, there needs to
4329 be some way to specify a default version to which external references to
4330 this symbol will be bound. You can do this with the `foo@@VERS_2.0'
4331 type of `.symver' directive. You can only declare one version of a
4332 symbol as the default in this manner; otherwise you would effectively
4333 have multiple definitions of the same symbol.
4335 If you wish to bind a reference to a specific version of the symbol
4336 within the shared library, you can use the aliases of convenience
4337 (i.e., `old_foo'), or you can use the `.symver' directive to
4338 specifically bind to an external version of the function in question.
4340 You can also specify the language in the version script:
4342 VERSION extern "lang" { version-script-commands }
4344 The supported `lang's are `C', `C++', and `Java'. The linker will
4345 iterate over the list of symbols at the link time and demangle them
4346 according to `lang' before matching them to the patterns specified in
4347 `version-script-commands'. The default `lang' is `C'.
4349 Demangled names may contains spaces and other special characters. As
4350 described above, you can use a glob pattern to match demangled names,
4351 or you can use a double-quoted string to match the string exactly. In
4352 the latter case, be aware that minor differences (such as differing
4353 whitespace) between the version script and the demangler output will
4354 cause a mismatch. As the exact string generated by the demangler might
4355 change in the future, even if the mangled name does not, you should
4356 check that all of your version directives are behaving as you expect
4360 File: ld.info, Node: Expressions, Next: Implicit Linker Scripts, Prev: VERSION, Up: Scripts
4362 3.10 Expressions in Linker Scripts
4363 ==================================
4365 The syntax for expressions in the linker script language is identical to
4366 that of C expressions. All expressions are evaluated as integers. All
4367 expressions are evaluated in the same size, which is 32 bits if both the
4368 host and target are 32 bits, and is otherwise 64 bits.
4370 You can use and set symbol values in expressions.
4372 The linker defines several special purpose builtin functions for use
4377 * Constants:: Constants
4378 * Symbolic Constants:: Symbolic constants
4379 * Symbols:: Symbol Names
4380 * Orphan Sections:: Orphan Sections
4381 * Location Counter:: The Location Counter
4382 * Operators:: Operators
4383 * Evaluation:: Evaluation
4384 * Expression Section:: The Section of an Expression
4385 * Builtin Functions:: Builtin Functions
4388 File: ld.info, Node: Constants, Next: Symbolic Constants, Up: Expressions
4393 All constants are integers.
4395 As in C, the linker considers an integer beginning with `0' to be
4396 octal, and an integer beginning with `0x' or `0X' to be hexadecimal.
4397 Alternatively the linker accepts suffixes of `h' or `H' for
4398 hexadeciaml, `o' or `O' for octal, `b' or `B' for binary and `d' or `D'
4399 for decimal. Any integer value without a prefix or a suffix is
4400 considered to be decimal.
4402 In addition, you can use the suffixes `K' and `M' to scale a
4403 constant by `1024' or `1024*1024' respectively. For example, the
4404 following all refer to the same quantity:
4411 Note - the `K' and `M' suffixes cannot be used in conjunction with
4412 the base suffixes mentioned above.
4415 File: ld.info, Node: Symbolic Constants, Next: Symbols, Prev: Constants, Up: Expressions
4417 3.10.2 Symbolic Constants
4418 -------------------------
4420 It is possible to refer to target specific constants via the use of the
4421 `CONSTANT(NAME)' operator, where NAME is one of:
4424 The target's maximum page size.
4427 The target's default page size.
4431 .text ALIGN (CONSTANT (MAXPAGESIZE)) : { *(.text) }
4433 will create a text section aligned to the largest page boundary
4434 supported by the target.
4437 File: ld.info, Node: Symbols, Next: Orphan Sections, Prev: Symbolic Constants, Up: Expressions
4442 Unless quoted, symbol names start with a letter, underscore, or period
4443 and may include letters, digits, underscores, periods, and hyphens.
4444 Unquoted symbol names must not conflict with any keywords. You can
4445 specify a symbol which contains odd characters or has the same name as a
4446 keyword by surrounding the symbol name in double quotes:
4448 "with a space" = "also with a space" + 10;
4450 Since symbols can contain many non-alphabetic characters, it is
4451 safest to delimit symbols with spaces. For example, `A-B' is one
4452 symbol, whereas `A - B' is an expression involving subtraction.
4455 File: ld.info, Node: Orphan Sections, Next: Location Counter, Prev: Symbols, Up: Expressions
4457 3.10.4 Orphan Sections
4458 ----------------------
4460 Orphan sections are sections present in the input files which are not
4461 explicitly placed into the output file by the linker script. The
4462 linker will still copy these sections into the output file, but it has
4463 to guess as to where they should be placed. The linker uses a simple
4464 heuristic to do this. It attempts to place orphan sections after
4465 non-orphan sections of the same attribute, such as code vs data,
4466 loadable vs non-loadable, etc. If there is not enough room to do this
4467 then it places at the end of the file.
4469 For ELF targets, the attribute of the section includes section type
4470 as well as section flag.
4472 If an orphaned section's name is representable as a C identifier then
4473 the linker will automatically *note PROVIDE:: two symbols:
4474 __start_SECNAME and __stop_SECNAME, where SECNAME is the name of the
4475 section. These indicate the start address and end address of the
4476 orphaned section respectively. Note: most section names are not
4477 representable as C identifiers because they contain a `.' character.
4480 File: ld.info, Node: Location Counter, Next: Operators, Prev: Orphan Sections, Up: Expressions
4482 3.10.5 The Location Counter
4483 ---------------------------
4485 The special linker variable "dot" `.' always contains the current
4486 output location counter. Since the `.' always refers to a location in
4487 an output section, it may only appear in an expression within a
4488 `SECTIONS' command. The `.' symbol may appear anywhere that an
4489 ordinary symbol is allowed in an expression.
4491 Assigning a value to `.' will cause the location counter to be
4492 moved. This may be used to create holes in the output section. The
4493 location counter may not be moved backwards inside an output section,
4494 and may not be moved backwards outside of an output section if so doing
4495 creates areas with overlapping LMAs.
4508 In the previous example, the `.text' section from `file1' is located
4509 at the beginning of the output section `output'. It is followed by a
4510 1000 byte gap. Then the `.text' section from `file2' appears, also
4511 with a 1000 byte gap following before the `.text' section from `file3'.
4512 The notation `= 0x12345678' specifies what data to write in the gaps
4513 (*note Output Section Fill::).
4515 Note: `.' actually refers to the byte offset from the start of the
4516 current containing object. Normally this is the `SECTIONS' statement,
4517 whose start address is 0, hence `.' can be used as an absolute address.
4518 If `.' is used inside a section description however, it refers to the
4519 byte offset from the start of that section, not an absolute address.
4520 Thus in a script like this:
4536 The `.text' section will be assigned a starting address of 0x100 and
4537 a size of exactly 0x200 bytes, even if there is not enough data in the
4538 `.text' input sections to fill this area. (If there is too much data,
4539 an error will be produced because this would be an attempt to move `.'
4540 backwards). The `.data' section will start at 0x500 and it will have
4541 an extra 0x600 bytes worth of space after the end of the values from
4542 the `.data' input sections and before the end of the `.data' output
4545 Setting symbols to the value of the location counter outside of an
4546 output section statement can result in unexpected values if the linker
4547 needs to place orphan sections. For example, given the following:
4560 If the linker needs to place some input section, e.g. `.rodata', not
4561 mentioned in the script, it might choose to place that section between
4562 `.text' and `.data'. You might think the linker should place `.rodata'
4563 on the blank line in the above script, but blank lines are of no
4564 particular significance to the linker. As well, the linker doesn't
4565 associate the above symbol names with their sections. Instead, it
4566 assumes that all assignments or other statements belong to the previous
4567 output section, except for the special case of an assignment to `.'.
4568 I.e., the linker will place the orphan `.rodata' section as if the
4569 script was written as follows:
4578 .rodata: { *(.rodata) }
4583 This may or may not be the script author's intention for the value of
4584 `start_of_data'. One way to influence the orphan section placement is
4585 to assign the location counter to itself, as the linker assumes that an
4586 assignment to `.' is setting the start address of a following output
4587 section and thus should be grouped with that section. So you could
4602 Now, the orphan `.rodata' section will be placed between
4603 `end_of_text' and `start_of_data'.
4606 File: ld.info, Node: Operators, Next: Evaluation, Prev: Location Counter, Up: Expressions
4611 The linker recognizes the standard C set of arithmetic operators, with
4612 the standard bindings and precedence levels:
4613 precedence associativity Operators Notes
4619 5 left == != > < <= >=
4625 11 right &= += -= *= /= (2)
4627 Notes: (1) Prefix operators (2) *Note Assignments::.
4630 File: ld.info, Node: Evaluation, Next: Expression Section, Prev: Operators, Up: Expressions
4635 The linker evaluates expressions lazily. It only computes the value of
4636 an expression when absolutely necessary.
4638 The linker needs some information, such as the value of the start
4639 address of the first section, and the origins and lengths of memory
4640 regions, in order to do any linking at all. These values are computed
4641 as soon as possible when the linker reads in the linker script.
4643 However, other values (such as symbol values) are not known or needed
4644 until after storage allocation. Such values are evaluated later, when
4645 other information (such as the sizes of output sections) is available
4646 for use in the symbol assignment expression.
4648 The sizes of sections cannot be known until after allocation, so
4649 assignments dependent upon these are not performed until after
4652 Some expressions, such as those depending upon the location counter
4653 `.', must be evaluated during section allocation.
4655 If the result of an expression is required, but the value is not
4656 available, then an error results. For example, a script like the
4660 .text 9+this_isnt_constant :
4663 will cause the error message `non constant expression for initial
4667 File: ld.info, Node: Expression Section, Next: Builtin Functions, Prev: Evaluation, Up: Expressions
4669 3.10.8 The Section of an Expression
4670 -----------------------------------
4672 Addresses and symbols may be section relative, or absolute. A section
4673 relative symbol is relocatable. If you request relocatable output
4674 using the `-r' option, a further link operation may change the value of
4675 a section relative symbol. On the other hand, an absolute symbol will
4676 retain the same value throughout any further link operations.
4678 Some terms in linker expressions are addresses. This is true of
4679 section relative symbols and for builtin functions that return an
4680 address, such as `ADDR', `LOADADDR', `ORIGIN' and `SEGMENT_START'.
4681 Other terms are simply numbers, or are builtin functions that return a
4682 non-address value, such as `LENGTH'. One complication is that unless
4683 you set `LD_FEATURE ("SANE_EXPR")' (*note Miscellaneous Commands::),
4684 numbers and absolute symbols are treated differently depending on their
4685 location, for compatibility with older versions of `ld'. Expressions
4686 appearing outside an output section definition treat all numbers as
4687 absolute addresses. Expressions appearing inside an output section
4688 definition treat absolute symbols as numbers. If `LD_FEATURE
4689 ("SANE_EXPR")' is given, then absolute symbols and numbers are simply
4690 treated as numbers everywhere.
4692 In the following simple example,
4697 __executable_start = 0x100;
4701 __data_start = 0x10;
4707 both `.' and `__executable_start' are set to the absolute address
4708 0x100 in the first two assignments, then both `.' and `__data_start'
4709 are set to 0x10 relative to the `.data' section in the second two
4712 For expressions involving numbers, relative addresses and absolute
4713 addresses, ld follows these rules to evaluate terms:
4715 * Unary operations on a relative address, and binary operations on
4716 two relative addresses in the same section or between one relative
4717 address and a number, apply the operator to the offset part of the
4720 * Unary operations on an absolute address, and binary operations on
4721 one or more absolute addresses or on two relative addresses not in
4722 the same section, first convert any non-absolute term to an
4723 absolute address before applying the operator.
4725 The result section of each sub-expression is as follows:
4727 * An operation involving only numbers results in a number.
4729 * The result of comparisons, `&&' and `||' is also a number.
4731 * The result of other binary arithmetic and logical operations on two
4732 relative addresses in the same section or two absolute addresess
4733 (after above conversions) is also a number.
4735 * The result of other operations on relative addresses or one
4736 relative address and a number, is a relative address in the same
4737 section as the relative operand(s).
4739 * The result of other operations on absolute addresses (after above
4740 conversions) is an absolute address.
4742 You can use the builtin function `ABSOLUTE' to force an expression
4743 to be absolute when it would otherwise be relative. For example, to
4744 create an absolute symbol set to the address of the end of the output
4748 .data : { *(.data) _edata = ABSOLUTE(.); }
4750 If `ABSOLUTE' were not used, `_edata' would be relative to the
4753 Using `LOADADDR' also forces an expression absolute, since this
4754 particular builtin function returns an absolute address.
4757 File: ld.info, Node: Builtin Functions, Prev: Expression Section, Up: Expressions
4759 3.10.9 Builtin Functions
4760 ------------------------
4762 The linker script language includes a number of builtin functions for
4763 use in linker script expressions.
4766 Return the absolute (non-relocatable, as opposed to non-negative)
4767 value of the expression EXP. Primarily useful to assign an
4768 absolute value to a symbol within a section definition, where
4769 symbol values are normally section relative. *Note Expression
4773 Return the address (VMA) of the named SECTION. Your script must
4774 previously have defined the location of that section. In the
4775 following example, `start_of_output_1', `symbol_1' and `symbol_2'
4776 are assigned equivalent values, except that `symbol_1' will be
4777 relative to the `.output1' section while the other two will be
4782 start_of_output_1 = ABSOLUTE(.);
4787 symbol_1 = ADDR(.output1);
4788 symbol_2 = start_of_output_1;
4794 Return the location counter (`.') or arbitrary expression aligned
4795 to the next ALIGN boundary. The single operand `ALIGN' doesn't
4796 change the value of the location counter--it just does arithmetic
4797 on it. The two operand `ALIGN' allows an arbitrary expression to
4798 be aligned upwards (`ALIGN(ALIGN)' is equivalent to `ALIGN(.,
4801 Here is an example which aligns the output `.data' section to the
4802 next `0x2000' byte boundary after the preceding section and sets a
4803 variable within the section to the next `0x8000' boundary after the
4806 .data ALIGN(0x2000): {
4808 variable = ALIGN(0x8000);
4811 The first use of `ALIGN' in this example specifies the
4812 location of a section because it is used as the optional ADDRESS
4813 attribute of a section definition (*note Output Section
4814 Address::). The second use of `ALIGN' is used to defines the
4817 The builtin function `NEXT' is closely related to `ALIGN'.
4820 Return the alignment in bytes of the named SECTION, if that
4821 section has been allocated. If the section has not been allocated
4822 when this is evaluated, the linker will report an error. In the
4823 following example, the alignment of the `.output' section is
4824 stored as the first value in that section.
4827 LONG (ALIGNOF (.output))
4833 This is a synonym for `ALIGN', for compatibility with older linker
4834 scripts. It is most often seen when setting the address of an
4837 `DATA_SEGMENT_ALIGN(MAXPAGESIZE, COMMONPAGESIZE)'
4838 This is equivalent to either
4839 (ALIGN(MAXPAGESIZE) + (. & (MAXPAGESIZE - 1)))
4841 (ALIGN(MAXPAGESIZE) + (. & (MAXPAGESIZE - COMMONPAGESIZE)))
4842 depending on whether the latter uses fewer COMMONPAGESIZE sized
4843 pages for the data segment (area between the result of this
4844 expression and `DATA_SEGMENT_END') than the former or not. If the
4845 latter form is used, it means COMMONPAGESIZE bytes of runtime
4846 memory will be saved at the expense of up to COMMONPAGESIZE wasted
4847 bytes in the on-disk file.
4849 This expression can only be used directly in `SECTIONS' commands,
4850 not in any output section descriptions and only once in the linker
4851 script. COMMONPAGESIZE should be less or equal to MAXPAGESIZE and
4852 should be the system page size the object wants to be optimized
4853 for (while still working on system page sizes up to MAXPAGESIZE).
4856 . = DATA_SEGMENT_ALIGN(0x10000, 0x2000);
4858 `DATA_SEGMENT_END(EXP)'
4859 This defines the end of data segment for `DATA_SEGMENT_ALIGN'
4860 evaluation purposes.
4862 . = DATA_SEGMENT_END(.);
4864 `DATA_SEGMENT_RELRO_END(OFFSET, EXP)'
4865 This defines the end of the `PT_GNU_RELRO' segment when `-z relro'
4866 option is used. Second argument is returned. When `-z relro'
4867 option is not present, `DATA_SEGMENT_RELRO_END' does nothing,
4868 otherwise `DATA_SEGMENT_ALIGN' is padded so that EXP + OFFSET is
4869 aligned to the most commonly used page boundary for particular
4870 target. If present in the linker script, it must always come in
4871 between `DATA_SEGMENT_ALIGN' and `DATA_SEGMENT_END'.
4873 . = DATA_SEGMENT_RELRO_END(24, .);
4876 Return 1 if SYMBOL is in the linker global symbol table and is
4877 defined before the statement using DEFINED in the script, otherwise
4878 return 0. You can use this function to provide default values for
4879 symbols. For example, the following script fragment shows how to
4880 set a global symbol `begin' to the first location in the `.text'
4881 section--but if a symbol called `begin' already existed, its value
4886 begin = DEFINED(begin) ? begin : . ;
4893 Return the length of the memory region named MEMORY.
4896 Return the absolute LMA of the named SECTION. (*note Output
4900 Returns the maximum of EXP1 and EXP2.
4903 Returns the minimum of EXP1 and EXP2.
4906 Return the next unallocated address that is a multiple of EXP.
4907 This function is closely related to `ALIGN(EXP)'; unless you use
4908 the `MEMORY' command to define discontinuous memory for the output
4909 file, the two functions are equivalent.
4912 Return the origin of the memory region named MEMORY.
4914 `SEGMENT_START(SEGMENT, DEFAULT)'
4915 Return the base address of the named SEGMENT. If an explicit
4916 value has been given for this segment (with a command-line `-T'
4917 option) that value will be returned; otherwise the value will be
4918 DEFAULT. At present, the `-T' command-line option can only be
4919 used to set the base address for the "text", "data", and "bss"
4920 sections, but you can use `SEGMENT_START' with any segment name.
4923 Return the size in bytes of the named SECTION, if that section has
4924 been allocated. If the section has not been allocated when this is
4925 evaluated, the linker will report an error. In the following
4926 example, `symbol_1' and `symbol_2' are assigned identical values:
4933 symbol_1 = .end - .start ;
4934 symbol_2 = SIZEOF(.output);
4939 Return the size in bytes of the output file's headers. This is
4940 information which appears at the start of the output file. You
4941 can use this number when setting the start address of the first
4942 section, if you choose, to facilitate paging.
4944 When producing an ELF output file, if the linker script uses the
4945 `SIZEOF_HEADERS' builtin function, the linker must compute the
4946 number of program headers before it has determined all the section
4947 addresses and sizes. If the linker later discovers that it needs
4948 additional program headers, it will report an error `not enough
4949 room for program headers'. To avoid this error, you must avoid
4950 using the `SIZEOF_HEADERS' function, or you must rework your linker
4951 script to avoid forcing the linker to use additional program
4952 headers, or you must define the program headers yourself using the
4953 `PHDRS' command (*note PHDRS::).
4956 File: ld.info, Node: Implicit Linker Scripts, Prev: Expressions, Up: Scripts
4958 3.11 Implicit Linker Scripts
4959 ============================
4961 If you specify a linker input file which the linker can not recognize as
4962 an object file or an archive file, it will try to read the file as a
4963 linker script. If the file can not be parsed as a linker script, the
4964 linker will report an error.
4966 An implicit linker script will not replace the default linker script.
4968 Typically an implicit linker script would contain only symbol
4969 assignments, or the `INPUT', `GROUP', or `VERSION' commands.
4971 Any input files read because of an implicit linker script will be
4972 read at the position in the command line where the implicit linker
4973 script was read. This can affect archive searching.
4976 File: ld.info, Node: Machine Dependent, Next: BFD, Prev: Scripts, Up: Top
4978 4 Machine Dependent Features
4979 ****************************
4981 `ld' has additional features on some platforms; the following sections
4982 describe them. Machines where `ld' has no additional functionality are
4988 * H8/300:: `ld' and the H8/300
4990 * i960:: `ld' and the Intel 960 family
4992 * ARM:: `ld' and the ARM family
4994 * HPPA ELF32:: `ld' and HPPA 32-bit ELF
4996 * M68K:: `ld' and the Motorola 68K family
4998 * MMIX:: `ld' and MMIX
5000 * MSP430:: `ld' and MSP430
5002 * M68HC11/68HC12:: `ld' and the Motorola 68HC11 and 68HC12 families
5004 * PowerPC ELF32:: `ld' and PowerPC 32-bit ELF Support
5006 * PowerPC64 ELF64:: `ld' and PowerPC64 64-bit ELF Support
5008 * SPU ELF:: `ld' and SPU ELF Support
5010 * TI COFF:: `ld' and TI COFF
5012 * WIN32:: `ld' and WIN32 (cygwin/mingw)
5014 * Xtensa:: `ld' and Xtensa Processors
5017 File: ld.info, Node: H8/300, Next: i960, Up: Machine Dependent
5019 4.1 `ld' and the H8/300
5020 =======================
5022 For the H8/300, `ld' can perform these global optimizations when you
5023 specify the `--relax' command-line option.
5025 _relaxing address modes_
5026 `ld' finds all `jsr' and `jmp' instructions whose targets are
5027 within eight bits, and turns them into eight-bit program-counter
5028 relative `bsr' and `bra' instructions, respectively.
5030 _synthesizing instructions_
5031 `ld' finds all `mov.b' instructions which use the sixteen-bit
5032 absolute address form, but refer to the top page of memory, and
5033 changes them to use the eight-bit address form. (That is: the
5034 linker turns `mov.b `@'AA:16' into `mov.b `@'AA:8' whenever the
5035 address AA is in the top page of memory).
5037 _bit manipulation instructions_
5038 `ld' finds all bit manipulation instructions like `band, bclr,
5039 biand, bild, bior, bist, bixor, bld, bnot, bor, bset, bst, btst,
5040 bxor' which use 32 bit and 16 bit absolute address form, but refer
5041 to the top page of memory, and changes them to use the 8 bit
5042 address form. (That is: the linker turns `bset #xx:3,`@'AA:32'
5043 into `bset #xx:3,`@'AA:8' whenever the address AA is in the top
5046 _system control instructions_
5047 `ld' finds all `ldc.w, stc.w' instructions which use the 32 bit
5048 absolute address form, but refer to the top page of memory, and
5049 changes them to use 16 bit address form. (That is: the linker
5050 turns `ldc.w `@'AA:32,ccr' into `ldc.w `@'AA:16,ccr' whenever the
5051 address AA is in the top page of memory).
5054 File: ld.info, Node: i960, Next: ARM, Prev: H8/300, Up: Machine Dependent
5056 4.2 `ld' and the Intel 960 Family
5057 =================================
5059 You can use the `-AARCHITECTURE' command line option to specify one of
5060 the two-letter names identifying members of the 960 family; the option
5061 specifies the desired output target, and warns of any incompatible
5062 instructions in the input files. It also modifies the linker's search
5063 strategy for archive libraries, to support the use of libraries
5064 specific to each particular architecture, by including in the search
5065 loop names suffixed with the string identifying the architecture.
5067 For example, if your `ld' command line included `-ACA' as well as
5068 `-ltry', the linker would look (in its built-in search paths, and in
5069 any paths you specify with `-L') for a library with the names
5076 The first two possibilities would be considered in any event; the last
5077 two are due to the use of `-ACA'.
5079 You can meaningfully use `-A' more than once on a command line, since
5080 the 960 architecture family allows combination of target architectures;
5081 each use will add another pair of name variants to search for when `-l'
5082 specifies a library.
5084 `ld' supports the `--relax' option for the i960 family. If you
5085 specify `--relax', `ld' finds all `balx' and `calx' instructions whose
5086 targets are within 24 bits, and turns them into 24-bit program-counter
5087 relative `bal' and `cal' instructions, respectively. `ld' also turns
5088 `cal' instructions into `bal' instructions when it determines that the
5089 target subroutine is a leaf routine (that is, the target subroutine does
5090 not itself call any subroutines).
5092 The `--fix-cortex-a8' switch enables a link-time workaround for an
5093 erratum in certain Cortex-A8 processors. The workaround is enabled by
5094 default if you are targeting the ARM v7-A architecture profile. It can
5095 be enabled otherwise by specifying `--fix-cortex-a8', or disabled
5096 unconditionally by specifying `--no-fix-cortex-a8'.
5098 The erratum only affects Thumb-2 code. Please contact ARM for
5101 The `--no-merge-exidx-entries' switch disables the merging of
5102 adjacent exidx entries in debuginfo.
5105 File: ld.info, Node: M68HC11/68HC12, Next: PowerPC ELF32, Prev: MSP430, Up: Machine Dependent
5107 4.3 `ld' and the Motorola 68HC11 and 68HC12 families
5108 ====================================================
5110 4.3.1 Linker Relaxation
5111 -----------------------
5113 For the Motorola 68HC11, `ld' can perform these global optimizations
5114 when you specify the `--relax' command-line option.
5116 _relaxing address modes_
5117 `ld' finds all `jsr' and `jmp' instructions whose targets are
5118 within eight bits, and turns them into eight-bit program-counter
5119 relative `bsr' and `bra' instructions, respectively.
5121 `ld' also looks at all 16-bit extended addressing modes and
5122 transforms them in a direct addressing mode when the address is in
5123 page 0 (between 0 and 0x0ff).
5125 _relaxing gcc instruction group_
5126 When `gcc' is called with `-mrelax', it can emit group of
5127 instructions that the linker can optimize to use a 68HC11 direct
5128 addressing mode. These instructions consists of `bclr' or `bset'
5132 4.3.2 Trampoline Generation
5133 ---------------------------
5135 For 68HC11 and 68HC12, `ld' can generate trampoline code to call a far
5136 function using a normal `jsr' instruction. The linker will also change
5137 the relocation to some far function to use the trampoline address
5138 instead of the function address. This is typically the case when a
5139 pointer to a function is taken. The pointer will in fact point to the
5140 function trampoline.
5143 File: ld.info, Node: ARM, Next: HPPA ELF32, Prev: i960, Up: Machine Dependent
5145 4.4 `ld' and the ARM family
5146 ===========================
5148 For the ARM, `ld' will generate code stubs to allow functions calls
5149 between ARM and Thumb code. These stubs only work with code that has
5150 been compiled and assembled with the `-mthumb-interwork' command line
5151 option. If it is necessary to link with old ARM object files or
5152 libraries, which have not been compiled with the -mthumb-interwork
5153 option then the `--support-old-code' command line switch should be
5154 given to the linker. This will make it generate larger stub functions
5155 which will work with non-interworking aware ARM code. Note, however,
5156 the linker does not support generating stubs for function calls to
5157 non-interworking aware Thumb code.
5159 The `--thumb-entry' switch is a duplicate of the generic `--entry'
5160 switch, in that it sets the program's starting address. But it also
5161 sets the bottom bit of the address, so that it can be branched to using
5162 a BX instruction, and the program will start executing in Thumb mode
5165 The `--use-nul-prefixed-import-tables' switch is specifying, that
5166 the import tables idata4 and idata5 have to be generated with a zero
5167 elememt prefix for import libraries. This is the old style to generate
5168 import tables. By default this option is turned off.
5170 The `--be8' switch instructs `ld' to generate BE8 format
5171 executables. This option is only valid when linking big-endian objects.
5172 The resulting image will contain big-endian data and little-endian code.
5174 The `R_ARM_TARGET1' relocation is typically used for entries in the
5175 `.init_array' section. It is interpreted as either `R_ARM_REL32' or
5176 `R_ARM_ABS32', depending on the target. The `--target1-rel' and
5177 `--target1-abs' switches override the default.
5179 The `--target2=type' switch overrides the default definition of the
5180 `R_ARM_TARGET2' relocation. Valid values for `type', their meanings,
5181 and target defaults are as follows:
5183 `R_ARM_REL32' (arm*-*-elf, arm*-*-eabi)
5186 `R_ARM_ABS32' (arm*-*-symbianelf)
5189 `R_ARM_GOT_PREL' (arm*-*-linux, arm*-*-*bsd)
5191 The `R_ARM_V4BX' relocation (defined by the ARM AAELF specification)
5192 enables objects compiled for the ARMv4 architecture to be
5193 interworking-safe when linked with other objects compiled for ARMv4t,
5194 but also allows pure ARMv4 binaries to be built from the same ARMv4
5197 In the latter case, the switch `--fix-v4bx' must be passed to the
5198 linker, which causes v4t `BX rM' instructions to be rewritten as `MOV
5199 PC,rM', since v4 processors do not have a `BX' instruction.
5201 In the former case, the switch should not be used, and `R_ARM_V4BX'
5202 relocations are ignored.
5204 Replace `BX rM' instructions identified by `R_ARM_V4BX' relocations
5205 with a branch to the following veneer:
5211 This allows generation of libraries/applications that work on ARMv4
5212 cores and are still interworking safe. Note that the above veneer
5213 clobbers the condition flags, so may cause incorrect progrm behavior in
5216 The `--use-blx' switch enables the linker to use ARM/Thumb BLX
5217 instructions (available on ARMv5t and above) in various situations.
5218 Currently it is used to perform calls via the PLT from Thumb code using
5219 BLX rather than using BX and a mode-switching stub before each PLT
5220 entry. This should lead to such calls executing slightly faster.
5222 This option is enabled implicitly for SymbianOS, so there is no need
5223 to specify it if you are using that target.
5225 The `--vfp11-denorm-fix' switch enables a link-time workaround for a
5226 bug in certain VFP11 coprocessor hardware, which sometimes allows
5227 instructions with denorm operands (which must be handled by support
5228 code) to have those operands overwritten by subsequent instructions
5229 before the support code can read the intended values.
5231 The bug may be avoided in scalar mode if you allow at least one
5232 intervening instruction between a VFP11 instruction which uses a
5233 register and another instruction which writes to the same register, or
5234 at least two intervening instructions if vector mode is in use. The bug
5235 only affects full-compliance floating-point mode: you do not need this
5236 workaround if you are using "runfast" mode. Please contact ARM for
5239 If you know you are using buggy VFP11 hardware, you can enable this
5240 workaround by specifying the linker option `--vfp-denorm-fix=scalar' if
5241 you are using the VFP11 scalar mode only, or `--vfp-denorm-fix=vector'
5242 if you are using vector mode (the latter also works for scalar code).
5243 The default is `--vfp-denorm-fix=none'.
5245 If the workaround is enabled, instructions are scanned for
5246 potentially-troublesome sequences, and a veneer is created for each
5247 such sequence which may trigger the erratum. The veneer consists of the
5248 first instruction of the sequence and a branch back to the subsequent
5249 instruction. The original instruction is then replaced with a branch to
5250 the veneer. The extra cycles required to call and return from the veneer
5251 are sufficient to avoid the erratum in both the scalar and vector cases.
5253 The `--fix-arm1176' switch enables a link-time workaround for an
5254 erratum in certain ARM1176 processors. The workaround is enabled by
5255 default if you are targetting ARM v6 (excluding ARM v6T2) or earlier.
5256 It can be disabled unconditionally by specifying `--no-fix-arm1176'.
5258 Further information is available in the "ARM1176JZ-S and ARM1176JZF-S
5259 Programmer Advice Notice" available on the ARM documentaion website at:
5260 http://infocenter.arm.com/.
5262 The `--no-enum-size-warning' switch prevents the linker from warning
5263 when linking object files that specify incompatible EABI enumeration
5264 size attributes. For example, with this switch enabled, linking of an
5265 object file using 32-bit enumeration values with another using
5266 enumeration values fitted into the smallest possible space will not be
5269 The `--no-wchar-size-warning' switch prevents the linker from
5270 warning when linking object files that specify incompatible EABI
5271 `wchar_t' size attributes. For example, with this switch enabled,
5272 linking of an object file using 32-bit `wchar_t' values with another
5273 using 16-bit `wchar_t' values will not be diagnosed.
5275 The `--pic-veneer' switch makes the linker use PIC sequences for
5276 ARM/Thumb interworking veneers, even if the rest of the binary is not
5277 PIC. This avoids problems on uClinux targets where `--emit-relocs' is
5278 used to generate relocatable binaries.
5280 The linker will automatically generate and insert small sequences of
5281 code into a linked ARM ELF executable whenever an attempt is made to
5282 perform a function call to a symbol that is too far away. The
5283 placement of these sequences of instructions - called stubs - is
5284 controlled by the command line option `--stub-group-size=N'. The
5285 placement is important because a poor choice can create a need for
5286 duplicate stubs, increasing the code sizw. The linker will try to
5287 group stubs together in order to reduce interruptions to the flow of
5288 code, but it needs guidance as to how big these groups should be and
5289 where they should be placed.
5291 The value of `N', the parameter to the `--stub-group-size=' option
5292 controls where the stub groups are placed. If it is negative then all
5293 stubs are placed after the first branch that needs them. If it is
5294 positive then the stubs can be placed either before or after the
5295 branches that need them. If the value of `N' is 1 (either +1 or -1)
5296 then the linker will choose exactly where to place groups of stubs,
5297 using its built in heuristics. A value of `N' greater than 1 (or
5298 smaller than -1) tells the linker that a single group of stubs can
5299 service at most `N' bytes from the input sections.
5301 The default, if `--stub-group-size=' is not specified, is `N = +1'.
5303 Farcalls stubs insertion is fully supported for the ARM-EABI target
5304 only, because it relies on object files properties not present
5308 File: ld.info, Node: HPPA ELF32, Next: M68K, Prev: ARM, Up: Machine Dependent
5310 4.5 `ld' and HPPA 32-bit ELF Support
5311 ====================================
5313 When generating a shared library, `ld' will by default generate import
5314 stubs suitable for use with a single sub-space application. The
5315 `--multi-subspace' switch causes `ld' to generate export stubs, and
5316 different (larger) import stubs suitable for use with multiple
5319 Long branch stubs and import/export stubs are placed by `ld' in stub
5320 sections located between groups of input sections. `--stub-group-size'
5321 specifies the maximum size of a group of input sections handled by one
5322 stub section. Since branch offsets are signed, a stub section may
5323 serve two groups of input sections, one group before the stub section,
5324 and one group after it. However, when using conditional branches that
5325 require stubs, it may be better (for branch prediction) that stub
5326 sections only serve one group of input sections. A negative value for
5327 `N' chooses this scheme, ensuring that branches to stubs always use a
5328 negative offset. Two special values of `N' are recognized, `1' and
5329 `-1'. These both instruct `ld' to automatically size input section
5330 groups for the branch types detected, with the same behaviour regarding
5331 stub placement as other positive or negative values of `N' respectively.
5333 Note that `--stub-group-size' does not split input sections. A
5334 single input section larger than the group size specified will of course
5335 create a larger group (of one section). If input sections are too
5336 large, it may not be possible for a branch to reach its stub.
5339 File: ld.info, Node: M68K, Next: MMIX, Prev: HPPA ELF32, Up: Machine Dependent
5341 4.6 `ld' and the Motorola 68K family
5342 ====================================
5344 The `--got=TYPE' option lets you choose the GOT generation scheme. The
5345 choices are `single', `negative', `multigot' and `target'. When
5346 `target' is selected the linker chooses the default GOT generation
5347 scheme for the current target. `single' tells the linker to generate a
5348 single GOT with entries only at non-negative offsets. `negative'
5349 instructs the linker to generate a single GOT with entries at both
5350 negative and positive offsets. Not all environments support such GOTs.
5351 `multigot' allows the linker to generate several GOTs in the output
5352 file. All GOT references from a single input object file access the
5353 same GOT, but references from different input object files might access
5354 different GOTs. Not all environments support such GOTs.
5357 File: ld.info, Node: MMIX, Next: MSP430, Prev: M68K, Up: Machine Dependent
5362 For MMIX, there is a choice of generating `ELF' object files or `mmo'
5363 object files when linking. The simulator `mmix' understands the `mmo'
5364 format. The binutils `objcopy' utility can translate between the two
5367 There is one special section, the `.MMIX.reg_contents' section.
5368 Contents in this section is assumed to correspond to that of global
5369 registers, and symbols referring to it are translated to special
5370 symbols, equal to registers. In a final link, the start address of the
5371 `.MMIX.reg_contents' section corresponds to the first allocated global
5372 register multiplied by 8. Register `$255' is not included in this
5373 section; it is always set to the program entry, which is at the symbol
5374 `Main' for `mmo' files.
5376 Global symbols with the prefix `__.MMIX.start.', for example
5377 `__.MMIX.start..text' and `__.MMIX.start..data' are special. The
5378 default linker script uses these to set the default start address of a
5381 Initial and trailing multiples of zero-valued 32-bit words in a
5382 section, are left out from an mmo file.
5385 File: ld.info, Node: MSP430, Next: M68HC11/68HC12, Prev: MMIX, Up: Machine Dependent
5390 For the MSP430 it is possible to select the MPU architecture. The flag
5391 `-m [mpu type]' will select an appropriate linker script for selected
5392 MPU type. (To get a list of known MPUs just pass `-m help' option to
5395 The linker will recognize some extra sections which are MSP430
5399 Defines a portion of ROM where interrupt vectors located.
5402 Defines the bootloader portion of the ROM (if applicable). Any
5403 code in this section will be uploaded to the MPU.
5406 Defines an information memory section (if applicable). Any code in
5407 this section will be uploaded to the MPU.
5410 This is the same as the `.infomem' section except that any code in
5411 this section will not be uploaded to the MPU.
5414 Denotes a portion of RAM located above `.bss' section.
5416 The last two sections are used by gcc.
5419 File: ld.info, Node: PowerPC ELF32, Next: PowerPC64 ELF64, Prev: M68HC11/68HC12, Up: Machine Dependent
5421 4.9 `ld' and PowerPC 32-bit ELF Support
5422 =======================================
5424 Branches on PowerPC processors are limited to a signed 26-bit
5425 displacement, which may result in `ld' giving `relocation truncated to
5426 fit' errors with very large programs. `--relax' enables the generation
5427 of trampolines that can access the entire 32-bit address space. These
5428 trampolines are inserted at section boundaries, so may not themselves
5429 be reachable if an input section exceeds 33M in size. You may combine
5430 `-r' and `--relax' to add trampolines in a partial link. In that case
5431 both branches to undefined symbols and inter-section branches are also
5432 considered potentially out of range, and trampolines inserted.
5435 Current PowerPC GCC accepts a `-msecure-plt' option that generates
5436 code capable of using a newer PLT and GOT layout that has the
5437 security advantage of no executable section ever needing to be
5438 writable and no writable section ever being executable. PowerPC
5439 `ld' will generate this layout, including stubs to access the PLT,
5440 if all input files (including startup and static libraries) were
5441 compiled with `-msecure-plt'. `--bss-plt' forces the old BSS PLT
5442 (and GOT layout) which can give slightly better performance.
5445 `ld' will use the new PLT and GOT layout if it is linking new
5446 `-fpic' or `-fPIC' code, but does not do so automatically when
5447 linking non-PIC code. This option requests the new PLT and GOT
5448 layout. A warning will be given if some object file requires the
5452 The new secure PLT and GOT are placed differently relative to other
5453 sections compared to older BSS PLT and GOT placement. The
5454 location of `.plt' must change because the new secure PLT is an
5455 initialized section while the old PLT is uninitialized. The
5456 reason for the `.got' change is more subtle: The new placement
5457 allows `.got' to be read-only in applications linked with `-z
5458 relro -z now'. However, this placement means that `.sdata' cannot
5459 always be used in shared libraries, because the PowerPC ABI
5460 accesses `.sdata' in shared libraries from the GOT pointer.
5461 `--sdata-got' forces the old GOT placement. PowerPC GCC doesn't
5462 use `.sdata' in shared libraries, so this option is really only
5463 useful for other compilers that may do so.
5466 This option causes `ld' to label linker stubs with a local symbol
5467 that encodes the stub type and destination.
5470 PowerPC `ld' normally performs some optimization of code sequences
5471 used to access Thread-Local Storage. Use this option to disable
5475 File: ld.info, Node: PowerPC64 ELF64, Next: SPU ELF, Prev: PowerPC ELF32, Up: Machine Dependent
5477 4.10 `ld' and PowerPC64 64-bit ELF Support
5478 ==========================================
5481 Long branch stubs, PLT call stubs and TOC adjusting stubs are
5482 placed by `ld' in stub sections located between groups of input
5483 sections. `--stub-group-size' specifies the maximum size of a
5484 group of input sections handled by one stub section. Since branch
5485 offsets are signed, a stub section may serve two groups of input
5486 sections, one group before the stub section, and one group after
5487 it. However, when using conditional branches that require stubs,
5488 it may be better (for branch prediction) that stub sections only
5489 serve one group of input sections. A negative value for `N'
5490 chooses this scheme, ensuring that branches to stubs always use a
5491 negative offset. Two special values of `N' are recognized, `1'
5492 and `-1'. These both instruct `ld' to automatically size input
5493 section groups for the branch types detected, with the same
5494 behaviour regarding stub placement as other positive or negative
5495 values of `N' respectively.
5497 Note that `--stub-group-size' does not split input sections. A
5498 single input section larger than the group size specified will of
5499 course create a larger group (of one section). If input sections
5500 are too large, it may not be possible for a branch to reach its
5504 This option causes `ld' to label linker stubs with a local symbol
5505 that encodes the stub type and destination.
5507 `--dotsyms, --no-dotsyms'
5508 These two options control how `ld' interprets version patterns in
5509 a version script. Older PowerPC64 compilers emitted both a
5510 function descriptor symbol with the same name as the function, and
5511 a code entry symbol with the name prefixed by a dot (`.'). To
5512 properly version a function `foo', the version script thus needs
5513 to control both `foo' and `.foo'. The option `--dotsyms', on by
5514 default, automatically adds the required dot-prefixed patterns.
5515 Use `--no-dotsyms' to disable this feature.
5518 PowerPC64 `ld' normally performs some optimization of code
5519 sequences used to access Thread-Local Storage. Use this option to
5520 disable the optimization.
5523 PowerPC64 `ld' normally removes `.opd' section entries
5524 corresponding to deleted link-once functions, or functions removed
5525 by the action of `--gc-sections' or linker script `/DISCARD/'.
5526 Use this option to disable `.opd' optimization.
5528 `--non-overlapping-opd'
5529 Some PowerPC64 compilers have an option to generate compressed
5530 `.opd' entries spaced 16 bytes apart, overlapping the third word,
5531 the static chain pointer (unused in C) with the first word of the
5532 next entry. This option expands such entries to the full 24 bytes.
5535 PowerPC64 `ld' normally removes unused `.toc' section entries.
5536 Such entries are detected by examining relocations that reference
5537 the TOC in code sections. A reloc in a deleted code section marks
5538 a TOC word as unneeded, while a reloc in a kept code section marks
5539 a TOC word as needed. Since the TOC may reference itself, TOC
5540 relocs are also examined. TOC words marked as both needed and
5541 unneeded will of course be kept. TOC words without any referencing
5542 reloc are assumed to be part of a multi-word entry, and are kept or
5543 discarded as per the nearest marked preceding word. This works
5544 reliably for compiler generated code, but may be incorrect if
5545 assembly code is used to insert TOC entries. Use this option to
5546 disable the optimization.
5549 If given any toc option besides `-mcmodel=medium' or
5550 `-mcmodel=large', PowerPC64 GCC generates code for a TOC model
5551 where TOC entries are accessed with a 16-bit offset from r2. This
5552 limits the total TOC size to 64K. PowerPC64 `ld' extends this
5553 limit by grouping code sections such that each group uses less
5554 than 64K for its TOC entries, then inserts r2 adjusting stubs
5555 between inter-group calls. `ld' does not split apart input
5556 sections, so cannot help if a single input file has a `.toc'
5557 section that exceeds 64K, most likely from linking multiple files
5558 with `ld -r'. Use this option to turn off this feature.
5561 By default, `ld' sorts TOC sections so that those whose file
5562 happens to have a section called `.init' or `.fini' are placed
5563 first, followed by TOC sections referenced by code generated with
5564 PowerPC64 gcc's `-mcmodel=small', and lastly TOC sections
5565 referenced only by code generated with PowerPC64 gcc's
5566 `-mcmodel=medium' or `-mcmodel=large' options. Doing this results
5567 in better TOC grouping for multi-TOC. Use this option to turn off
5572 Use these options to control whether individual PLT call stubs are
5573 aligned to a 32-byte boundary, or to the specified power of two
5574 boundary when using `--plt-align='. By default PLT call stubs are
5577 `--plt-static-chain'
5578 `--no-plt-static-chain'
5579 Use these options to control whether PLT call stubs load the static
5580 chain pointer (r11). `ld' defaults to not loading the static
5581 chain since there is never any need to do so on a PLT call.
5585 With power7's weakly ordered memory model, it is possible when
5586 using lazy binding for ld.so to update a plt entry in one thread
5587 and have another thread see the individual plt entry words update
5588 in the wrong order, despite ld.so carefully writing in the correct
5589 order and using memory write barriers. To avoid this we need some
5590 sort of read barrier in the call stub, or use LD_BIND_NOW=1. By
5591 default, `ld' looks for calls to commonly used functions that
5592 create threads, and if seen, adds the necessary barriers. Use
5593 these options to change the default behaviour.
5596 File: ld.info, Node: SPU ELF, Next: TI COFF, Prev: PowerPC64 ELF64, Up: Machine Dependent
5598 4.11 `ld' and SPU ELF Support
5599 =============================
5602 This option marks an executable as a PIC plugin module.
5605 Normally, `ld' recognizes calls to functions within overlay
5606 regions, and redirects such calls to an overlay manager via a stub.
5607 `ld' also provides a built-in overlay manager. This option turns
5608 off all this special overlay handling.
5611 This option causes `ld' to label overlay stubs with a local symbol
5612 that encodes the stub type and destination.
5614 `--extra-overlay-stubs'
5615 This option causes `ld' to add overlay call stubs on all function
5616 calls out of overlay regions. Normally stubs are not added on
5617 calls to non-overlay regions.
5619 `--local-store=lo:hi'
5620 `ld' usually checks that a final executable for SPU fits in the
5621 address range 0 to 256k. This option may be used to change the
5622 range. Disable the check entirely with `--local-store=0:0'.
5625 SPU local store space is limited. Over-allocation of stack space
5626 unnecessarily limits space available for code and data, while
5627 under-allocation results in runtime failures. If given this
5628 option, `ld' will provide an estimate of maximum stack usage.
5629 `ld' does this by examining symbols in code sections to determine
5630 the extents of functions, and looking at function prologues for
5631 stack adjusting instructions. A call-graph is created by looking
5632 for relocations on branch instructions. The graph is then searched
5633 for the maximum stack usage path. Note that this analysis does not
5634 find calls made via function pointers, and does not handle
5635 recursion and other cycles in the call graph. Stack usage may be
5636 under-estimated if your code makes such calls. Also, stack usage
5637 for dynamic allocation, e.g. alloca, will not be detected. If a
5638 link map is requested, detailed information about each function's
5639 stack usage and calls will be given.
5642 This option, if given along with `--stack-analysis' will result in
5643 `ld' emitting stack sizing symbols for each function. These take
5644 the form `__stack_<function_name>' for global functions, and
5645 `__stack_<number>_<function_name>' for static functions.
5646 `<number>' is the section id in hex. The value of such symbols is
5647 the stack requirement for the corresponding function. The symbol
5648 size will be zero, type `STT_NOTYPE', binding `STB_LOCAL', and
5652 File: ld.info, Node: TI COFF, Next: WIN32, Prev: SPU ELF, Up: Machine Dependent
5654 4.12 `ld''s Support for Various TI COFF Versions
5655 ================================================
5657 The `--format' switch allows selection of one of the various TI COFF
5658 versions. The latest of this writing is 2; versions 0 and 1 are also
5659 supported. The TI COFF versions also vary in header byte-order format;
5660 `ld' will read any version or byte order, but the output header format
5661 depends on the default specified by the specific target.
5664 File: ld.info, Node: WIN32, Next: Xtensa, Prev: TI COFF, Up: Machine Dependent
5666 4.13 `ld' and WIN32 (cygwin/mingw)
5667 ==================================
5669 This section describes some of the win32 specific `ld' issues. See
5670 *Note Command Line Options: Options. for detailed description of the
5671 command line options mentioned here.
5674 The standard Windows linker creates and uses so-called import
5675 libraries, which contains information for linking to dll's. They
5676 are regular static archives and are handled as any other static
5677 archive. The cygwin and mingw ports of `ld' have specific support
5678 for creating such libraries provided with the `--out-implib'
5679 command line option.
5681 _exporting DLL symbols_
5682 The cygwin/mingw `ld' has several ways to export symbols for dll's.
5684 _using auto-export functionality_
5685 By default `ld' exports symbols with the auto-export
5686 functionality, which is controlled by the following command
5689 * -export-all-symbols [This is the default]
5695 * -exclude-modules-for-implib
5699 When auto-export is in operation, `ld' will export all the
5700 non-local (global and common) symbols it finds in a DLL, with
5701 the exception of a few symbols known to belong to the
5702 system's runtime and libraries. As it will often not be
5703 desirable to export all of a DLL's symbols, which may include
5704 private functions that are not part of any public interface,
5705 the command-line options listed above may be used to filter
5706 symbols out from the list for exporting. The `--output-def'
5707 option can be used in order to see the final list of exported
5708 symbols with all exclusions taken into effect.
5710 If `--export-all-symbols' is not given explicitly on the
5711 command line, then the default auto-export behavior will be
5712 _disabled_ if either of the following are true:
5714 * A DEF file is used.
5716 * Any symbol in any object file was marked with the
5717 __declspec(dllexport) attribute.
5720 Another way of exporting symbols is using a DEF file. A DEF
5721 file is an ASCII file containing definitions of symbols which
5722 should be exported when a dll is created. Usually it is
5723 named `<dll name>.def' and is added as any other object file
5724 to the linker's command line. The file's name must end in
5727 gcc -o <output> <objectfiles> <dll name>.def
5729 Using a DEF file turns off the normal auto-export behavior,
5730 unless the `--export-all-symbols' option is also used.
5732 Here is an example of a DEF file for a shared library called
5735 LIBRARY "xyz.dll" BASE=0x20000000
5741 another_foo = abc.dll.afoo
5746 This example defines a DLL with a non-default base address
5747 and seven symbols in the export table. The third exported
5748 symbol `_bar' is an alias for the second. The fourth symbol,
5749 `another_foo' is resolved by "forwarding" to another module
5750 and treating it as an alias for `afoo' exported from the DLL
5751 `abc.dll'. The final symbol `var1' is declared to be a data
5752 object. The `doo' symbol in export library is an alias of
5753 `foo', which gets the string name in export table `foo2'. The
5754 `eoo' symbol is an data export symbol, which gets in export
5755 table the name `var1'.
5757 The optional `LIBRARY <name>' command indicates the _internal_
5758 name of the output DLL. If `<name>' does not include a suffix,
5759 the default library suffix, `.DLL' is appended.
5761 When the .DEF file is used to build an application, rather
5762 than a library, the `NAME <name>' command should be used
5763 instead of `LIBRARY'. If `<name>' does not include a suffix,
5764 the default executable suffix, `.EXE' is appended.
5766 With either `LIBRARY <name>' or `NAME <name>' the optional
5767 specification `BASE = <number>' may be used to specify a
5768 non-default base address for the image.
5770 If neither `LIBRARY <name>' nor `NAME <name>' is specified,
5771 or they specify an empty string, the internal name is the
5772 same as the filename specified on the command line.
5774 The complete specification of an export symbol is:
5777 ( ( ( <name1> [ = <name2> ] )
5778 | ( <name1> = <module-name> . <external-name>))
5779 [ @ <integer> ] [NONAME] [DATA] [CONSTANT] [PRIVATE] [== <name3>] ) *
5781 Declares `<name1>' as an exported symbol from the DLL, or
5782 declares `<name1>' as an exported alias for `<name2>'; or
5783 declares `<name1>' as a "forward" alias for the symbol
5784 `<external-name>' in the DLL `<module-name>'. Optionally,
5785 the symbol may be exported by the specified ordinal
5786 `<integer>' alias. The optional `<name3>' is the to be used
5787 string in import/export table for the symbol.
5789 The optional keywords that follow the declaration indicate:
5791 `NONAME': Do not put the symbol name in the DLL's export
5792 table. It will still be exported by its ordinal alias
5793 (either the value specified by the .def specification or,
5794 otherwise, the value assigned by the linker). The symbol
5795 name, however, does remain visible in the import library (if
5796 any), unless `PRIVATE' is also specified.
5798 `DATA': The symbol is a variable or object, rather than a
5799 function. The import lib will export only an indirect
5800 reference to `foo' as the symbol `_imp__foo' (ie, `foo' must
5801 be resolved as `*_imp__foo').
5803 `CONSTANT': Like `DATA', but put the undecorated `foo' as
5804 well as `_imp__foo' into the import library. Both refer to the
5805 read-only import address table's pointer to the variable, not
5806 to the variable itself. This can be dangerous. If the user
5807 code fails to add the `dllimport' attribute and also fails to
5808 explicitly add the extra indirection that the use of the
5809 attribute enforces, the application will behave unexpectedly.
5811 `PRIVATE': Put the symbol in the DLL's export table, but do
5812 not put it into the static import library used to resolve
5813 imports at link time. The symbol can still be imported using
5814 the `LoadLibrary/GetProcAddress' API at runtime or by by
5815 using the GNU ld extension of linking directly to the DLL
5816 without an import library.
5818 See ld/deffilep.y in the binutils sources for the full
5819 specification of other DEF file statements
5821 While linking a shared dll, `ld' is able to create a DEF file
5822 with the `--output-def <file>' command line option.
5825 Another way of marking symbols for export is to modify the
5826 source code itself, so that when building the DLL each symbol
5827 to be exported is declared as:
5829 __declspec(dllexport) int a_variable
5830 __declspec(dllexport) void a_function(int with_args)
5832 All such symbols will be exported from the DLL. If, however,
5833 any of the object files in the DLL contain symbols decorated
5834 in this way, then the normal auto-export behavior is
5835 disabled, unless the `--export-all-symbols' option is also
5838 Note that object files that wish to access these symbols must
5839 _not_ decorate them with dllexport. Instead, they should use
5842 __declspec(dllimport) int a_variable
5843 __declspec(dllimport) void a_function(int with_args)
5845 This complicates the structure of library header files,
5846 because when included by the library itself the header must
5847 declare the variables and functions as dllexport, but when
5848 included by client code the header must declare them as
5849 dllimport. There are a number of idioms that are typically
5850 used to do this; often client code can omit the __declspec()
5851 declaration completely. See `--enable-auto-import' and
5852 `automatic data imports' for more information.
5854 _automatic data imports_
5855 The standard Windows dll format supports data imports from dlls
5856 only by adding special decorations (dllimport/dllexport), which
5857 let the compiler produce specific assembler instructions to deal
5858 with this issue. This increases the effort necessary to port
5859 existing Un*x code to these platforms, especially for large c++
5860 libraries and applications. The auto-import feature, which was
5861 initially provided by Paul Sokolovsky, allows one to omit the
5862 decorations to achieve a behavior that conforms to that on
5863 POSIX/Un*x platforms. This feature is enabled with the
5864 `--enable-auto-import' command-line option, although it is enabled
5865 by default on cygwin/mingw. The `--enable-auto-import' option
5866 itself now serves mainly to suppress any warnings that are
5867 ordinarily emitted when linked objects trigger the feature's use.
5869 auto-import of variables does not always work flawlessly without
5870 additional assistance. Sometimes, you will see this message
5872 "variable '<var>' can't be auto-imported. Please read the
5873 documentation for ld's `--enable-auto-import' for details."
5875 The `--enable-auto-import' documentation explains why this error
5876 occurs, and several methods that can be used to overcome this
5877 difficulty. One of these methods is the _runtime pseudo-relocs_
5878 feature, described below.
5880 For complex variables imported from DLLs (such as structs or
5881 classes), object files typically contain a base address for the
5882 variable and an offset (_addend_) within the variable-to specify a
5883 particular field or public member, for instance. Unfortunately,
5884 the runtime loader used in win32 environments is incapable of
5885 fixing these references at runtime without the additional
5886 information supplied by dllimport/dllexport decorations. The
5887 standard auto-import feature described above is unable to resolve
5890 The `--enable-runtime-pseudo-relocs' switch allows these
5891 references to be resolved without error, while leaving the task of
5892 adjusting the references themselves (with their non-zero addends)
5893 to specialized code provided by the runtime environment. Recent
5894 versions of the cygwin and mingw environments and compilers
5895 provide this runtime support; older versions do not. However, the
5896 support is only necessary on the developer's platform; the
5897 compiled result will run without error on an older system.
5899 `--enable-runtime-pseudo-relocs' is not the default; it must be
5900 explicitly enabled as needed.
5902 _direct linking to a dll_
5903 The cygwin/mingw ports of `ld' support the direct linking,
5904 including data symbols, to a dll without the usage of any import
5905 libraries. This is much faster and uses much less memory than
5906 does the traditional import library method, especially when
5907 linking large libraries or applications. When `ld' creates an
5908 import lib, each function or variable exported from the dll is
5909 stored in its own bfd, even though a single bfd could contain many
5910 exports. The overhead involved in storing, loading, and
5911 processing so many bfd's is quite large, and explains the
5912 tremendous time, memory, and storage needed to link against
5913 particularly large or complex libraries when using import libs.
5915 Linking directly to a dll uses no extra command-line switches
5916 other than `-L' and `-l', because `ld' already searches for a
5917 number of names to match each library. All that is needed from
5918 the developer's perspective is an understanding of this search, in
5919 order to force ld to select the dll instead of an import library.
5921 For instance, when ld is called with the argument `-lxxx' it will
5922 attempt to find, in the first directory of its search path,
5932 before moving on to the next directory in the search path.
5934 (*) Actually, this is not `cygxxx.dll' but in fact is
5935 `<prefix>xxx.dll', where `<prefix>' is set by the `ld' option
5936 `--dll-search-prefix=<prefix>'. In the case of cygwin, the
5937 standard gcc spec file includes `--dll-search-prefix=cyg', so in
5938 effect we actually search for `cygxxx.dll'.
5940 Other win32-based unix environments, such as mingw or pw32, may
5941 use other `<prefix>'es, although at present only cygwin makes use
5942 of this feature. It was originally intended to help avoid name
5943 conflicts among dll's built for the various win32/un*x
5944 environments, so that (for example) two versions of a zlib dll
5945 could coexist on the same machine.
5947 The generic cygwin/mingw path layout uses a `bin' directory for
5948 applications and dll's and a `lib' directory for the import
5949 libraries (using cygwin nomenclature):
5954 libxxx.dll.a (in case of dll's)
5955 libxxx.a (in case of static archive)
5957 Linking directly to a dll without using the import library can be
5960 1. Use the dll directly by adding the `bin' path to the link line
5961 gcc -Wl,-verbose -o a.exe -L../bin/ -lxxx
5963 However, as the dll's often have version numbers appended to their
5964 names (`cygncurses-5.dll') this will often fail, unless one
5965 specifies `-L../bin -lncurses-5' to include the version. Import
5966 libs are generally not versioned, and do not have this difficulty.
5968 2. Create a symbolic link from the dll to a file in the `lib'
5969 directory according to the above mentioned search pattern. This
5970 should be used to avoid unwanted changes in the tools needed for
5973 ln -s bin/cygxxx.dll lib/[cyg|lib|]xxx.dll[.a]
5975 Then you can link without any make environment changes.
5977 gcc -Wl,-verbose -o a.exe -L../lib/ -lxxx
5979 This technique also avoids the version number problems, because
5980 the following is perfectly legal
5985 libxxx.dll.a -> ../bin/cygxxx-5.dll
5987 Linking directly to a dll without using an import lib will work
5988 even when auto-import features are exercised, and even when
5989 `--enable-runtime-pseudo-relocs' is used.
5991 Given the improvements in speed and memory usage, one might
5992 justifiably wonder why import libraries are used at all. There
5995 1. Until recently, the link-directly-to-dll functionality did _not_
5996 work with auto-imported data.
5998 2. Sometimes it is necessary to include pure static objects within
5999 the import library (which otherwise contains only bfd's for
6000 indirection symbols that point to the exports of a dll). Again,
6001 the import lib for the cygwin kernel makes use of this ability,
6002 and it is not possible to do this without an import lib.
6004 3. Symbol aliases can only be resolved using an import lib. This
6005 is critical when linking against OS-supplied dll's (eg, the win32
6006 API) in which symbols are usually exported as undecorated aliases
6007 of their stdcall-decorated assembly names.
6009 So, import libs are not going away. But the ability to replace
6010 true import libs with a simple symbolic link to (or a copy of) a
6011 dll, in many cases, is a useful addition to the suite of tools
6012 binutils makes available to the win32 developer. Given the
6013 massive improvements in memory requirements during linking, storage
6014 requirements, and linking speed, we expect that many developers
6015 will soon begin to use this feature whenever possible.
6019 _adding additional names_
6020 Sometimes, it is useful to export symbols with additional
6021 names. A symbol `foo' will be exported as `foo', but it can
6022 also be exported as `_foo' by using special directives in the
6023 DEF file when creating the dll. This will affect also the
6024 optional created import library. Consider the following DEF
6027 LIBRARY "xyz.dll" BASE=0x61000000
6033 The line `_foo = foo' maps the symbol `foo' to `_foo'.
6035 Another method for creating a symbol alias is to create it in
6036 the source code using the "weak" attribute:
6038 void foo () { /* Do something. */; }
6039 void _foo () __attribute__ ((weak, alias ("foo")));
6041 See the gcc manual for more information about attributes and
6045 Sometimes it is useful to rename exports. For instance, the
6046 cygwin kernel does this regularly. A symbol `_foo' can be
6047 exported as `foo' but not as `_foo' by using special
6048 directives in the DEF file. (This will also affect the import
6049 library, if it is created). In the following example:
6051 LIBRARY "xyz.dll" BASE=0x61000000
6056 The line `_foo = foo' maps the exported symbol `foo' to
6059 Note: using a DEF file disables the default auto-export behavior,
6060 unless the `--export-all-symbols' command line option is used.
6061 If, however, you are trying to rename symbols, then you should list
6062 _all_ desired exports in the DEF file, including the symbols that
6063 are not being renamed, and do _not_ use the `--export-all-symbols'
6064 option. If you list only the renamed symbols in the DEF file, and
6065 use `--export-all-symbols' to handle the other symbols, then the
6066 both the new names _and_ the original names for the renamed
6067 symbols will be exported. In effect, you'd be aliasing those
6068 symbols, not renaming them, which is probably not what you wanted.
6071 The Windows object format, PE, specifies a form of weak symbols
6072 called weak externals. When a weak symbol is linked and the
6073 symbol is not defined, the weak symbol becomes an alias for some
6074 other symbol. There are three variants of weak externals:
6075 * Definition is searched for in objects and libraries,
6076 historically called lazy externals.
6078 * Definition is searched for only in other objects, not in
6079 libraries. This form is not presently implemented.
6081 * No search; the symbol is an alias. This form is not presently
6083 As a GNU extension, weak symbols that do not specify an alternate
6084 symbol are supported. If the symbol is undefined when linking,
6085 the symbol uses a default value.
6087 _aligned common symbols_
6088 As a GNU extension to the PE file format, it is possible to
6089 specify the desired alignment for a common symbol. This
6090 information is conveyed from the assembler or compiler to the
6091 linker by means of GNU-specific commands carried in the object
6092 file's `.drectve' section, which are recognized by `ld' and
6093 respected when laying out the common symbols. Native tools will
6094 be able to process object files employing this GNU extension, but
6095 will fail to respect the alignment instructions, and may issue
6096 noisy warnings about unknown linker directives.
6099 File: ld.info, Node: Xtensa, Prev: WIN32, Up: Machine Dependent
6101 4.14 `ld' and Xtensa Processors
6102 ===============================
6104 The default `ld' behavior for Xtensa processors is to interpret
6105 `SECTIONS' commands so that lists of explicitly named sections in a
6106 specification with a wildcard file will be interleaved when necessary to
6107 keep literal pools within the range of PC-relative load offsets. For
6108 example, with the command:
6117 `ld' may interleave some of the `.literal' and `.text' sections from
6118 different object files to ensure that the literal pools are within the
6119 range of PC-relative load offsets. A valid interleaving might place
6120 the `.literal' sections from an initial group of files followed by the
6121 `.text' sections of that group of files. Then, the `.literal' sections
6122 from the rest of the files and the `.text' sections from the rest of
6123 the files would follow.
6125 Relaxation is enabled by default for the Xtensa version of `ld' and
6126 provides two important link-time optimizations. The first optimization
6127 is to combine identical literal values to reduce code size. A redundant
6128 literal will be removed and all the `L32R' instructions that use it
6129 will be changed to reference an identical literal, as long as the
6130 location of the replacement literal is within the offset range of all
6131 the `L32R' instructions. The second optimization is to remove
6132 unnecessary overhead from assembler-generated "longcall" sequences of
6133 `L32R'/`CALLXN' when the target functions are within range of direct
6134 `CALLN' instructions.
6136 For each of these cases where an indirect call sequence can be
6137 optimized to a direct call, the linker will change the `CALLXN'
6138 instruction to a `CALLN' instruction, remove the `L32R' instruction,
6139 and remove the literal referenced by the `L32R' instruction if it is
6140 not used for anything else. Removing the `L32R' instruction always
6141 reduces code size but can potentially hurt performance by changing the
6142 alignment of subsequent branch targets. By default, the linker will
6143 always preserve alignments, either by switching some instructions
6144 between 24-bit encodings and the equivalent density instructions or by
6145 inserting a no-op in place of the `L32R' instruction that was removed.
6146 If code size is more important than performance, the `--size-opt'
6147 option can be used to prevent the linker from widening density
6148 instructions or inserting no-ops, except in a few cases where no-ops
6149 are required for correctness.
6151 The following Xtensa-specific command-line options can be used to
6155 When optimizing indirect calls to direct calls, optimize for code
6156 size more than performance. With this option, the linker will not
6157 insert no-ops or widen density instructions to preserve branch
6158 target alignment. There may still be some cases where no-ops are
6159 required to preserve the correctness of the code.
6162 File: ld.info, Node: BFD, Next: Reporting Bugs, Prev: Machine Dependent, Up: Top
6167 The linker accesses object and archive files using the BFD libraries.
6168 These libraries allow the linker to use the same routines to operate on
6169 object files whatever the object file format. A different object file
6170 format can be supported simply by creating a new BFD back end and adding
6171 it to the library. To conserve runtime memory, however, the linker and
6172 associated tools are usually configured to support only a subset of the
6173 object file formats available. You can use `objdump -i' (*note
6174 objdump: (binutils.info)objdump.) to list all the formats available for
6177 As with most implementations, BFD is a compromise between several
6178 conflicting requirements. The major factor influencing BFD design was
6179 efficiency: any time used converting between formats is time which
6180 would not have been spent had BFD not been involved. This is partly
6181 offset by abstraction payback; since BFD simplifies applications and
6182 back ends, more time and care may be spent optimizing algorithms for a
6185 One minor artifact of the BFD solution which you should bear in mind
6186 is the potential for information loss. There are two places where
6187 useful information can be lost using the BFD mechanism: during
6188 conversion and during output. *Note BFD information loss::.
6192 * BFD outline:: How it works: an outline of BFD
6195 File: ld.info, Node: BFD outline, Up: BFD
6197 5.1 How It Works: An Outline of BFD
6198 ===================================
6200 When an object file is opened, BFD subroutines automatically determine
6201 the format of the input object file. They then build a descriptor in
6202 memory with pointers to routines that will be used to access elements of
6203 the object file's data structures.
6205 As different information from the object files is required, BFD
6206 reads from different sections of the file and processes them. For
6207 example, a very common operation for the linker is processing symbol
6208 tables. Each BFD back end provides a routine for converting between
6209 the object file's representation of symbols and an internal canonical
6210 format. When the linker asks for the symbol table of an object file, it
6211 calls through a memory pointer to the routine from the relevant BFD
6212 back end which reads and converts the table into a canonical form. The
6213 linker then operates upon the canonical form. When the link is finished
6214 and the linker writes the output file's symbol table, another BFD back
6215 end routine is called to take the newly created symbol table and
6216 convert it into the chosen output format.
6220 * BFD information loss:: Information Loss
6221 * Canonical format:: The BFD canonical object-file format
6224 File: ld.info, Node: BFD information loss, Next: Canonical format, Up: BFD outline
6226 5.1.1 Information Loss
6227 ----------------------
6229 _Information can be lost during output._ The output formats supported
6230 by BFD do not provide identical facilities, and information which can
6231 be described in one form has nowhere to go in another format. One
6232 example of this is alignment information in `b.out'. There is nowhere
6233 in an `a.out' format file to store alignment information on the
6234 contained data, so when a file is linked from `b.out' and an `a.out'
6235 image is produced, alignment information will not propagate to the
6236 output file. (The linker will still use the alignment information
6237 internally, so the link is performed correctly).
6239 Another example is COFF section names. COFF files may contain an
6240 unlimited number of sections, each one with a textual section name. If
6241 the target of the link is a format which does not have many sections
6242 (e.g., `a.out') or has sections without names (e.g., the Oasys format),
6243 the link cannot be done simply. You can circumvent this problem by
6244 describing the desired input-to-output section mapping with the linker
6247 _Information can be lost during canonicalization._ The BFD internal
6248 canonical form of the external formats is not exhaustive; there are
6249 structures in input formats for which there is no direct representation
6250 internally. This means that the BFD back ends cannot maintain all
6251 possible data richness through the transformation between external to
6252 internal and back to external formats.
6254 This limitation is only a problem when an application reads one
6255 format and writes another. Each BFD back end is responsible for
6256 maintaining as much data as possible, and the internal BFD canonical
6257 form has structures which are opaque to the BFD core, and exported only
6258 to the back ends. When a file is read in one format, the canonical form
6259 is generated for BFD and the application. At the same time, the back
6260 end saves away any information which may otherwise be lost. If the data
6261 is then written back in the same format, the back end routine will be
6262 able to use the canonical form provided by the BFD core as well as the
6263 information it prepared earlier. Since there is a great deal of
6264 commonality between back ends, there is no information lost when
6265 linking or copying big endian COFF to little endian COFF, or `a.out' to
6266 `b.out'. When a mixture of formats is linked, the information is only
6267 lost from the files whose format differs from the destination.
6270 File: ld.info, Node: Canonical format, Prev: BFD information loss, Up: BFD outline
6272 5.1.2 The BFD canonical object-file format
6273 ------------------------------------------
6275 The greatest potential for loss of information occurs when there is the
6276 least overlap between the information provided by the source format,
6277 that stored by the canonical format, and that needed by the destination
6278 format. A brief description of the canonical form may help you
6279 understand which kinds of data you can count on preserving across
6283 Information stored on a per-file basis includes target machine
6284 architecture, particular implementation format type, a demand
6285 pageable bit, and a write protected bit. Information like Unix
6286 magic numbers is not stored here--only the magic numbers' meaning,
6287 so a `ZMAGIC' file would have both the demand pageable bit and the
6288 write protected text bit set. The byte order of the target is
6289 stored on a per-file basis, so that big- and little-endian object
6290 files may be used with one another.
6293 Each section in the input file contains the name of the section,
6294 the section's original address in the object file, size and
6295 alignment information, various flags, and pointers into other BFD
6299 Each symbol contains a pointer to the information for the object
6300 file which originally defined it, its name, its value, and various
6301 flag bits. When a BFD back end reads in a symbol table, it
6302 relocates all symbols to make them relative to the base of the
6303 section where they were defined. Doing this ensures that each
6304 symbol points to its containing section. Each symbol also has a
6305 varying amount of hidden private data for the BFD back end. Since
6306 the symbol points to the original file, the private data format
6307 for that symbol is accessible. `ld' can operate on a collection
6308 of symbols of wildly different formats without problems.
6310 Normal global and simple local symbols are maintained on output,
6311 so an output file (no matter its format) will retain symbols
6312 pointing to functions and to global, static, and common variables.
6313 Some symbol information is not worth retaining; in `a.out', type
6314 information is stored in the symbol table as long symbol names.
6315 This information would be useless to most COFF debuggers; the
6316 linker has command line switches to allow users to throw it away.
6318 There is one word of type information within the symbol, so if the
6319 format supports symbol type information within symbols (for
6320 example, COFF, IEEE, Oasys) and the type is simple enough to fit
6321 within one word (nearly everything but aggregates), the
6322 information will be preserved.
6325 Each canonical BFD relocation record contains a pointer to the
6326 symbol to relocate to, the offset of the data to relocate, the
6327 section the data is in, and a pointer to a relocation type
6328 descriptor. Relocation is performed by passing messages through
6329 the relocation type descriptor and the symbol pointer. Therefore,
6330 relocations can be performed on output data using a relocation
6331 method that is only available in one of the input formats. For
6332 instance, Oasys provides a byte relocation format. A relocation
6333 record requesting this relocation type would point indirectly to a
6334 routine to perform this, so the relocation may be performed on a
6335 byte being written to a 68k COFF file, even though 68k COFF has no
6336 such relocation type.
6339 Object formats can contain, for debugging purposes, some form of
6340 mapping between symbols, source line numbers, and addresses in the
6341 output file. These addresses have to be relocated along with the
6342 symbol information. Each symbol with an associated list of line
6343 number records points to the first record of the list. The head
6344 of a line number list consists of a pointer to the symbol, which
6345 allows finding out the address of the function whose line number
6346 is being described. The rest of the list is made up of pairs:
6347 offsets into the section and line numbers. Any format which can
6348 simply derive this information can pass it successfully between
6349 formats (COFF, IEEE and Oasys).
6352 File: ld.info, Node: Reporting Bugs, Next: MRI, Prev: BFD, Up: Top
6357 Your bug reports play an essential role in making `ld' reliable.
6359 Reporting a bug may help you by bringing a solution to your problem,
6360 or it may not. But in any case the principal function of a bug report
6361 is to help the entire community by making the next version of `ld' work
6362 better. Bug reports are your contribution to the maintenance of `ld'.
6364 In order for a bug report to serve its purpose, you must include the
6365 information that enables us to fix the bug.
6369 * Bug Criteria:: Have you found a bug?
6370 * Bug Reporting:: How to report bugs
6373 File: ld.info, Node: Bug Criteria, Next: Bug Reporting, Up: Reporting Bugs
6375 6.1 Have You Found a Bug?
6376 =========================
6378 If you are not sure whether you have found a bug, here are some
6381 * If the linker gets a fatal signal, for any input whatever, that is
6382 a `ld' bug. Reliable linkers never crash.
6384 * If `ld' produces an error message for valid input, that is a bug.
6386 * If `ld' does not produce an error message for invalid input, that
6387 may be a bug. In the general case, the linker can not verify that
6388 object files are correct.
6390 * If you are an experienced user of linkers, your suggestions for
6391 improvement of `ld' are welcome in any case.
6394 File: ld.info, Node: Bug Reporting, Prev: Bug Criteria, Up: Reporting Bugs
6396 6.2 How to Report Bugs
6397 ======================
6399 A number of companies and individuals offer support for GNU products.
6400 If you obtained `ld' from a support organization, we recommend you
6401 contact that organization first.
6403 You can find contact information for many support companies and
6404 individuals in the file `etc/SERVICE' in the GNU Emacs distribution.
6406 Otherwise, send bug reports for `ld' to
6407 `http://www.sourceware.org/bugzilla/'.
6409 The fundamental principle of reporting bugs usefully is this:
6410 *report all the facts*. If you are not sure whether to state a fact or
6411 leave it out, state it!
6413 Often people omit facts because they think they know what causes the
6414 problem and assume that some details do not matter. Thus, you might
6415 assume that the name of a symbol you use in an example does not matter.
6416 Well, probably it does not, but one cannot be sure. Perhaps the bug
6417 is a stray memory reference which happens to fetch from the location
6418 where that name is stored in memory; perhaps, if the name were
6419 different, the contents of that location would fool the linker into
6420 doing the right thing despite the bug. Play it safe and give a
6421 specific, complete example. That is the easiest thing for you to do,
6422 and the most helpful.
6424 Keep in mind that the purpose of a bug report is to enable us to fix
6425 the bug if it is new to us. Therefore, always write your bug reports
6426 on the assumption that the bug has not been reported previously.
6428 Sometimes people give a few sketchy facts and ask, "Does this ring a
6429 bell?" This cannot help us fix a bug, so it is basically useless. We
6430 respond by asking for enough details to enable us to investigate. You
6431 might as well expedite matters by sending them to begin with.
6433 To enable us to fix the bug, you should include all these things:
6435 * The version of `ld'. `ld' announces it if you start it with the
6436 `--version' argument.
6438 Without this, we will not know whether there is any point in
6439 looking for the bug in the current version of `ld'.
6441 * Any patches you may have applied to the `ld' source, including any
6442 patches made to the `BFD' library.
6444 * The type of machine you are using, and the operating system name
6447 * What compiler (and its version) was used to compile `ld'--e.g.
6450 * The command arguments you gave the linker to link your example and
6451 observe the bug. To guarantee you will not omit something
6452 important, list them all. A copy of the Makefile (or the output
6453 from make) is sufficient.
6455 If we were to try to guess the arguments, we would probably guess
6456 wrong and then we might not encounter the bug.
6458 * A complete input file, or set of input files, that will reproduce
6459 the bug. It is generally most helpful to send the actual object
6460 files provided that they are reasonably small. Say no more than
6461 10K. For bigger files you can either make them available by FTP
6462 or HTTP or else state that you are willing to send the object
6463 file(s) to whomever requests them. (Note - your email will be
6464 going to a mailing list, so we do not want to clog it up with
6465 large attachments). But small attachments are best.
6467 If the source files were assembled using `gas' or compiled using
6468 `gcc', then it may be OK to send the source files rather than the
6469 object files. In this case, be sure to say exactly what version of
6470 `gas' or `gcc' was used to produce the object files. Also say how
6471 `gas' or `gcc' were configured.
6473 * A description of what behavior you observe that you believe is
6474 incorrect. For example, "It gets a fatal signal."
6476 Of course, if the bug is that `ld' gets a fatal signal, then we
6477 will certainly notice it. But if the bug is incorrect output, we
6478 might not notice unless it is glaringly wrong. You might as well
6479 not give us a chance to make a mistake.
6481 Even if the problem you experience is a fatal signal, you should
6482 still say so explicitly. Suppose something strange is going on,
6483 such as, your copy of `ld' is out of sync, or you have encountered
6484 a bug in the C library on your system. (This has happened!) Your
6485 copy might crash and ours would not. If you told us to expect a
6486 crash, then when ours fails to crash, we would know that the bug
6487 was not happening for us. If you had not told us to expect a
6488 crash, then we would not be able to draw any conclusion from our
6491 * If you wish to suggest changes to the `ld' source, send us context
6492 diffs, as generated by `diff' with the `-u', `-c', or `-p' option.
6493 Always send diffs from the old file to the new file. If you even
6494 discuss something in the `ld' source, refer to it by context, not
6497 The line numbers in our development sources will not match those
6498 in your sources. Your line numbers would convey no useful
6501 Here are some things that are not necessary:
6503 * A description of the envelope of the bug.
6505 Often people who encounter a bug spend a lot of time investigating
6506 which changes to the input file will make the bug go away and which
6507 changes will not affect it.
6509 This is often time consuming and not very useful, because the way
6510 we will find the bug is by running a single example under the
6511 debugger with breakpoints, not by pure deduction from a series of
6512 examples. We recommend that you save your time for something else.
6514 Of course, if you can find a simpler example to report _instead_
6515 of the original one, that is a convenience for us. Errors in the
6516 output will be easier to spot, running under the debugger will take
6517 less time, and so on.
6519 However, simplification is not vital; if you do not want to do
6520 this, report the bug anyway and send us the entire test case you
6523 * A patch for the bug.
6525 A patch for the bug does help us if it is a good one. But do not
6526 omit the necessary information, such as the test case, on the
6527 assumption that a patch is all we need. We might see problems
6528 with your patch and decide to fix the problem another way, or we
6529 might not understand it at all.
6531 Sometimes with a program as complicated as `ld' it is very hard to
6532 construct an example that will make the program follow a certain
6533 path through the code. If you do not send us the example, we will
6534 not be able to construct one, so we will not be able to verify
6535 that the bug is fixed.
6537 And if we cannot understand what bug you are trying to fix, or why
6538 your patch should be an improvement, we will not install it. A
6539 test case will help us to understand.
6541 * A guess about what the bug is or what it depends on.
6543 Such guesses are usually wrong. Even we cannot guess right about
6544 such things without first using the debugger to find the facts.
6547 File: ld.info, Node: MRI, Next: GNU Free Documentation License, Prev: Reporting Bugs, Up: Top
6549 Appendix A MRI Compatible Script Files
6550 **************************************
6552 To aid users making the transition to GNU `ld' from the MRI linker,
6553 `ld' can use MRI compatible linker scripts as an alternative to the
6554 more general-purpose linker scripting language described in *Note
6555 Scripts::. MRI compatible linker scripts have a much simpler command
6556 set than the scripting language otherwise used with `ld'. GNU `ld'
6557 supports the most commonly used MRI linker commands; these commands are
6560 In general, MRI scripts aren't of much use with the `a.out' object
6561 file format, since it only has three sections and MRI scripts lack some
6562 features to make use of them.
6564 You can specify a file containing an MRI-compatible script using the
6565 `-c' command-line option.
6567 Each command in an MRI-compatible script occupies its own line; each
6568 command line starts with the keyword that identifies the command (though
6569 blank lines are also allowed for punctuation). If a line of an
6570 MRI-compatible script begins with an unrecognized keyword, `ld' issues
6571 a warning message, but continues processing the script.
6573 Lines beginning with `*' are comments.
6575 You can write these commands using all upper-case letters, or all
6576 lower case; for example, `chip' is the same as `CHIP'. The following
6577 list shows only the upper-case form of each command.
6580 `ABSOLUTE SECNAME, SECNAME, ... SECNAME'
6581 Normally, `ld' includes in the output file all sections from all
6582 the input files. However, in an MRI-compatible script, you can
6583 use the `ABSOLUTE' command to restrict the sections that will be
6584 present in your output program. If the `ABSOLUTE' command is used
6585 at all in a script, then only the sections named explicitly in
6586 `ABSOLUTE' commands will appear in the linker output. You can
6587 still use other input sections (whatever you select on the command
6588 line, or using `LOAD') to resolve addresses in the output file.
6590 `ALIAS OUT-SECNAME, IN-SECNAME'
6591 Use this command to place the data from input section IN-SECNAME
6592 in a section called OUT-SECNAME in the linker output file.
6594 IN-SECNAME may be an integer.
6596 `ALIGN SECNAME = EXPRESSION'
6597 Align the section called SECNAME to EXPRESSION. The EXPRESSION
6598 should be a power of two.
6601 Use the value of EXPRESSION as the lowest address (other than
6602 absolute addresses) in the output file.
6605 `CHIP EXPRESSION, EXPRESSION'
6606 This command does nothing; it is accepted only for compatibility.
6609 This command does nothing whatever; it's only accepted for
6612 `FORMAT OUTPUT-FORMAT'
6613 Similar to the `OUTPUT_FORMAT' command in the more general linker
6614 language, but restricted to one of these output formats:
6616 1. S-records, if OUTPUT-FORMAT is `S'
6618 2. IEEE, if OUTPUT-FORMAT is `IEEE'
6620 3. COFF (the `coff-m68k' variant in BFD), if OUTPUT-FORMAT is
6624 Print (to the standard output file) a link map, as produced by the
6625 `ld' command-line option `-M'.
6627 The keyword `LIST' may be followed by anything on the same line,
6628 with no change in its effect.
6631 `LOAD FILENAME, FILENAME, ... FILENAME'
6632 Include one or more object file FILENAME in the link; this has the
6633 same effect as specifying FILENAME directly on the `ld' command
6637 OUTPUT-NAME is the name for the program produced by `ld'; the
6638 MRI-compatible command `NAME' is equivalent to the command-line
6639 option `-o' or the general script language command `OUTPUT'.
6641 `ORDER SECNAME, SECNAME, ... SECNAME'
6642 `ORDER SECNAME SECNAME SECNAME'
6643 Normally, `ld' orders the sections in its output file in the order
6644 in which they first appear in the input files. In an
6645 MRI-compatible script, you can override this ordering with the
6646 `ORDER' command. The sections you list with `ORDER' will appear
6647 first in your output file, in the order specified.
6649 `PUBLIC NAME=EXPRESSION'
6650 `PUBLIC NAME,EXPRESSION'
6651 `PUBLIC NAME EXPRESSION'
6652 Supply a value (EXPRESSION) for external symbol NAME used in the
6655 `SECT SECNAME, EXPRESSION'
6656 `SECT SECNAME=EXPRESSION'
6657 `SECT SECNAME EXPRESSION'
6658 You can use any of these three forms of the `SECT' command to
6659 specify the start address (EXPRESSION) for section SECNAME. If
6660 you have more than one `SECT' statement for the same SECNAME, only
6661 the _first_ sets the start address.
6664 File: ld.info, Node: GNU Free Documentation License, Next: LD Index, Prev: MRI, Up: Top
6666 Appendix B GNU Free Documentation License
6667 *****************************************
6669 Version 1.3, 3 November 2008
6671 Copyright (C) 2000, 2001, 2002, 2007, 2008 Free Software Foundation, Inc.
6674 Everyone is permitted to copy and distribute verbatim copies
6675 of this license document, but changing it is not allowed.
6679 The purpose of this License is to make a manual, textbook, or other
6680 functional and useful document "free" in the sense of freedom: to
6681 assure everyone the effective freedom to copy and redistribute it,
6682 with or without modifying it, either commercially or
6683 noncommercially. Secondarily, this License preserves for the
6684 author and publisher a way to get credit for their work, while not
6685 being considered responsible for modifications made by others.
6687 This License is a kind of "copyleft", which means that derivative
6688 works of the document must themselves be free in the same sense.
6689 It complements the GNU General Public License, which is a copyleft
6690 license designed for free software.
6692 We have designed this License in order to use it for manuals for
6693 free software, because free software needs free documentation: a
6694 free program should come with manuals providing the same freedoms
6695 that the software does. But this License is not limited to
6696 software manuals; it can be used for any textual work, regardless
6697 of subject matter or whether it is published as a printed book.
6698 We recommend this License principally for works whose purpose is
6699 instruction or reference.
6701 1. APPLICABILITY AND DEFINITIONS
6703 This License applies to any manual or other work, in any medium,
6704 that contains a notice placed by the copyright holder saying it
6705 can be distributed under the terms of this License. Such a notice
6706 grants a world-wide, royalty-free license, unlimited in duration,
6707 to use that work under the conditions stated herein. The
6708 "Document", below, refers to any such manual or work. Any member
6709 of the public is a licensee, and is addressed as "you". You
6710 accept the license if you copy, modify or distribute the work in a
6711 way requiring permission under copyright law.
6713 A "Modified Version" of the Document means any work containing the
6714 Document or a portion of it, either copied verbatim, or with
6715 modifications and/or translated into another language.
6717 A "Secondary Section" is a named appendix or a front-matter section
6718 of the Document that deals exclusively with the relationship of the
6719 publishers or authors of the Document to the Document's overall
6720 subject (or to related matters) and contains nothing that could
6721 fall directly within that overall subject. (Thus, if the Document
6722 is in part a textbook of mathematics, a Secondary Section may not
6723 explain any mathematics.) The relationship could be a matter of
6724 historical connection with the subject or with related matters, or
6725 of legal, commercial, philosophical, ethical or political position
6728 The "Invariant Sections" are certain Secondary Sections whose
6729 titles are designated, as being those of Invariant Sections, in
6730 the notice that says that the Document is released under this
6731 License. If a section does not fit the above definition of
6732 Secondary then it is not allowed to be designated as Invariant.
6733 The Document may contain zero Invariant Sections. If the Document
6734 does not identify any Invariant Sections then there are none.
6736 The "Cover Texts" are certain short passages of text that are
6737 listed, as Front-Cover Texts or Back-Cover Texts, in the notice
6738 that says that the Document is released under this License. A
6739 Front-Cover Text may be at most 5 words, and a Back-Cover Text may
6740 be at most 25 words.
6742 A "Transparent" copy of the Document means a machine-readable copy,
6743 represented in a format whose specification is available to the
6744 general public, that is suitable for revising the document
6745 straightforwardly with generic text editors or (for images
6746 composed of pixels) generic paint programs or (for drawings) some
6747 widely available drawing editor, and that is suitable for input to
6748 text formatters or for automatic translation to a variety of
6749 formats suitable for input to text formatters. A copy made in an
6750 otherwise Transparent file format whose markup, or absence of
6751 markup, has been arranged to thwart or discourage subsequent
6752 modification by readers is not Transparent. An image format is
6753 not Transparent if used for any substantial amount of text. A
6754 copy that is not "Transparent" is called "Opaque".
6756 Examples of suitable formats for Transparent copies include plain
6757 ASCII without markup, Texinfo input format, LaTeX input format,
6758 SGML or XML using a publicly available DTD, and
6759 standard-conforming simple HTML, PostScript or PDF designed for
6760 human modification. Examples of transparent image formats include
6761 PNG, XCF and JPG. Opaque formats include proprietary formats that
6762 can be read and edited only by proprietary word processors, SGML or
6763 XML for which the DTD and/or processing tools are not generally
6764 available, and the machine-generated HTML, PostScript or PDF
6765 produced by some word processors for output purposes only.
6767 The "Title Page" means, for a printed book, the title page itself,
6768 plus such following pages as are needed to hold, legibly, the
6769 material this License requires to appear in the title page. For
6770 works in formats which do not have any title page as such, "Title
6771 Page" means the text near the most prominent appearance of the
6772 work's title, preceding the beginning of the body of the text.
6774 The "publisher" means any person or entity that distributes copies
6775 of the Document to the public.
6777 A section "Entitled XYZ" means a named subunit of the Document
6778 whose title either is precisely XYZ or contains XYZ in parentheses
6779 following text that translates XYZ in another language. (Here XYZ
6780 stands for a specific section name mentioned below, such as
6781 "Acknowledgements", "Dedications", "Endorsements", or "History".)
6782 To "Preserve the Title" of such a section when you modify the
6783 Document means that it remains a section "Entitled XYZ" according
6786 The Document may include Warranty Disclaimers next to the notice
6787 which states that this License applies to the Document. These
6788 Warranty Disclaimers are considered to be included by reference in
6789 this License, but only as regards disclaiming warranties: any other
6790 implication that these Warranty Disclaimers may have is void and
6791 has no effect on the meaning of this License.
6795 You may copy and distribute the Document in any medium, either
6796 commercially or noncommercially, provided that this License, the
6797 copyright notices, and the license notice saying this License
6798 applies to the Document are reproduced in all copies, and that you
6799 add no other conditions whatsoever to those of this License. You
6800 may not use technical measures to obstruct or control the reading
6801 or further copying of the copies you make or distribute. However,
6802 you may accept compensation in exchange for copies. If you
6803 distribute a large enough number of copies you must also follow
6804 the conditions in section 3.
6806 You may also lend copies, under the same conditions stated above,
6807 and you may publicly display copies.
6809 3. COPYING IN QUANTITY
6811 If you publish printed copies (or copies in media that commonly
6812 have printed covers) of the Document, numbering more than 100, and
6813 the Document's license notice requires Cover Texts, you must
6814 enclose the copies in covers that carry, clearly and legibly, all
6815 these Cover Texts: Front-Cover Texts on the front cover, and
6816 Back-Cover Texts on the back cover. Both covers must also clearly
6817 and legibly identify you as the publisher of these copies. The
6818 front cover must present the full title with all words of the
6819 title equally prominent and visible. You may add other material
6820 on the covers in addition. Copying with changes limited to the
6821 covers, as long as they preserve the title of the Document and
6822 satisfy these conditions, can be treated as verbatim copying in
6825 If the required texts for either cover are too voluminous to fit
6826 legibly, you should put the first ones listed (as many as fit
6827 reasonably) on the actual cover, and continue the rest onto
6830 If you publish or distribute Opaque copies of the Document
6831 numbering more than 100, you must either include a
6832 machine-readable Transparent copy along with each Opaque copy, or
6833 state in or with each Opaque copy a computer-network location from
6834 which the general network-using public has access to download
6835 using public-standard network protocols a complete Transparent
6836 copy of the Document, free of added material. If you use the
6837 latter option, you must take reasonably prudent steps, when you
6838 begin distribution of Opaque copies in quantity, to ensure that
6839 this Transparent copy will remain thus accessible at the stated
6840 location until at least one year after the last time you
6841 distribute an Opaque copy (directly or through your agents or
6842 retailers) of that edition to the public.
6844 It is requested, but not required, that you contact the authors of
6845 the Document well before redistributing any large number of
6846 copies, to give them a chance to provide you with an updated
6847 version of the Document.
6851 You may copy and distribute a Modified Version of the Document
6852 under the conditions of sections 2 and 3 above, provided that you
6853 release the Modified Version under precisely this License, with
6854 the Modified Version filling the role of the Document, thus
6855 licensing distribution and modification of the Modified Version to
6856 whoever possesses a copy of it. In addition, you must do these
6857 things in the Modified Version:
6859 A. Use in the Title Page (and on the covers, if any) a title
6860 distinct from that of the Document, and from those of
6861 previous versions (which should, if there were any, be listed
6862 in the History section of the Document). You may use the
6863 same title as a previous version if the original publisher of
6864 that version gives permission.
6866 B. List on the Title Page, as authors, one or more persons or
6867 entities responsible for authorship of the modifications in
6868 the Modified Version, together with at least five of the
6869 principal authors of the Document (all of its principal
6870 authors, if it has fewer than five), unless they release you
6871 from this requirement.
6873 C. State on the Title page the name of the publisher of the
6874 Modified Version, as the publisher.
6876 D. Preserve all the copyright notices of the Document.
6878 E. Add an appropriate copyright notice for your modifications
6879 adjacent to the other copyright notices.
6881 F. Include, immediately after the copyright notices, a license
6882 notice giving the public permission to use the Modified
6883 Version under the terms of this License, in the form shown in
6886 G. Preserve in that license notice the full lists of Invariant
6887 Sections and required Cover Texts given in the Document's
6890 H. Include an unaltered copy of this License.
6892 I. Preserve the section Entitled "History", Preserve its Title,
6893 and add to it an item stating at least the title, year, new
6894 authors, and publisher of the Modified Version as given on
6895 the Title Page. If there is no section Entitled "History" in
6896 the Document, create one stating the title, year, authors,
6897 and publisher of the Document as given on its Title Page,
6898 then add an item describing the Modified Version as stated in
6899 the previous sentence.
6901 J. Preserve the network location, if any, given in the Document
6902 for public access to a Transparent copy of the Document, and
6903 likewise the network locations given in the Document for
6904 previous versions it was based on. These may be placed in
6905 the "History" section. You may omit a network location for a
6906 work that was published at least four years before the
6907 Document itself, or if the original publisher of the version
6908 it refers to gives permission.
6910 K. For any section Entitled "Acknowledgements" or "Dedications",
6911 Preserve the Title of the section, and preserve in the
6912 section all the substance and tone of each of the contributor
6913 acknowledgements and/or dedications given therein.
6915 L. Preserve all the Invariant Sections of the Document,
6916 unaltered in their text and in their titles. Section numbers
6917 or the equivalent are not considered part of the section
6920 M. Delete any section Entitled "Endorsements". Such a section
6921 may not be included in the Modified Version.
6923 N. Do not retitle any existing section to be Entitled
6924 "Endorsements" or to conflict in title with any Invariant
6927 O. Preserve any Warranty Disclaimers.
6929 If the Modified Version includes new front-matter sections or
6930 appendices that qualify as Secondary Sections and contain no
6931 material copied from the Document, you may at your option
6932 designate some or all of these sections as invariant. To do this,
6933 add their titles to the list of Invariant Sections in the Modified
6934 Version's license notice. These titles must be distinct from any
6935 other section titles.
6937 You may add a section Entitled "Endorsements", provided it contains
6938 nothing but endorsements of your Modified Version by various
6939 parties--for example, statements of peer review or that the text
6940 has been approved by an organization as the authoritative
6941 definition of a standard.
6943 You may add a passage of up to five words as a Front-Cover Text,
6944 and a passage of up to 25 words as a Back-Cover Text, to the end
6945 of the list of Cover Texts in the Modified Version. Only one
6946 passage of Front-Cover Text and one of Back-Cover Text may be
6947 added by (or through arrangements made by) any one entity. If the
6948 Document already includes a cover text for the same cover,
6949 previously added by you or by arrangement made by the same entity
6950 you are acting on behalf of, you may not add another; but you may
6951 replace the old one, on explicit permission from the previous
6952 publisher that added the old one.
6954 The author(s) and publisher(s) of the Document do not by this
6955 License give permission to use their names for publicity for or to
6956 assert or imply endorsement of any Modified Version.
6958 5. COMBINING DOCUMENTS
6960 You may combine the Document with other documents released under
6961 this License, under the terms defined in section 4 above for
6962 modified versions, provided that you include in the combination
6963 all of the Invariant Sections of all of the original documents,
6964 unmodified, and list them all as Invariant Sections of your
6965 combined work in its license notice, and that you preserve all
6966 their Warranty Disclaimers.
6968 The combined work need only contain one copy of this License, and
6969 multiple identical Invariant Sections may be replaced with a single
6970 copy. If there are multiple Invariant Sections with the same name
6971 but different contents, make the title of each such section unique
6972 by adding at the end of it, in parentheses, the name of the
6973 original author or publisher of that section if known, or else a
6974 unique number. Make the same adjustment to the section titles in
6975 the list of Invariant Sections in the license notice of the
6978 In the combination, you must combine any sections Entitled
6979 "History" in the various original documents, forming one section
6980 Entitled "History"; likewise combine any sections Entitled
6981 "Acknowledgements", and any sections Entitled "Dedications". You
6982 must delete all sections Entitled "Endorsements."
6984 6. COLLECTIONS OF DOCUMENTS
6986 You may make a collection consisting of the Document and other
6987 documents released under this License, and replace the individual
6988 copies of this License in the various documents with a single copy
6989 that is included in the collection, provided that you follow the
6990 rules of this License for verbatim copying of each of the
6991 documents in all other respects.
6993 You may extract a single document from such a collection, and
6994 distribute it individually under this License, provided you insert
6995 a copy of this License into the extracted document, and follow
6996 this License in all other respects regarding verbatim copying of
6999 7. AGGREGATION WITH INDEPENDENT WORKS
7001 A compilation of the Document or its derivatives with other
7002 separate and independent documents or works, in or on a volume of
7003 a storage or distribution medium, is called an "aggregate" if the
7004 copyright resulting from the compilation is not used to limit the
7005 legal rights of the compilation's users beyond what the individual
7006 works permit. When the Document is included in an aggregate, this
7007 License does not apply to the other works in the aggregate which
7008 are not themselves derivative works of the Document.
7010 If the Cover Text requirement of section 3 is applicable to these
7011 copies of the Document, then if the Document is less than one half
7012 of the entire aggregate, the Document's Cover Texts may be placed
7013 on covers that bracket the Document within the aggregate, or the
7014 electronic equivalent of covers if the Document is in electronic
7015 form. Otherwise they must appear on printed covers that bracket
7016 the whole aggregate.
7020 Translation is considered a kind of modification, so you may
7021 distribute translations of the Document under the terms of section
7022 4. Replacing Invariant Sections with translations requires special
7023 permission from their copyright holders, but you may include
7024 translations of some or all Invariant Sections in addition to the
7025 original versions of these Invariant Sections. You may include a
7026 translation of this License, and all the license notices in the
7027 Document, and any Warranty Disclaimers, provided that you also
7028 include the original English version of this License and the
7029 original versions of those notices and disclaimers. In case of a
7030 disagreement between the translation and the original version of
7031 this License or a notice or disclaimer, the original version will
7034 If a section in the Document is Entitled "Acknowledgements",
7035 "Dedications", or "History", the requirement (section 4) to
7036 Preserve its Title (section 1) will typically require changing the
7041 You may not copy, modify, sublicense, or distribute the Document
7042 except as expressly provided under this License. Any attempt
7043 otherwise to copy, modify, sublicense, or distribute it is void,
7044 and will automatically terminate your rights under this License.
7046 However, if you cease all violation of this License, then your
7047 license from a particular copyright holder is reinstated (a)
7048 provisionally, unless and until the copyright holder explicitly
7049 and finally terminates your license, and (b) permanently, if the
7050 copyright holder fails to notify you of the violation by some
7051 reasonable means prior to 60 days after the cessation.
7053 Moreover, your license from a particular copyright holder is
7054 reinstated permanently if the copyright holder notifies you of the
7055 violation by some reasonable means, this is the first time you have
7056 received notice of violation of this License (for any work) from
7057 that copyright holder, and you cure the violation prior to 30 days
7058 after your receipt of the notice.
7060 Termination of your rights under this section does not terminate
7061 the licenses of parties who have received copies or rights from
7062 you under this License. If your rights have been terminated and
7063 not permanently reinstated, receipt of a copy of some or all of
7064 the same material does not give you any rights to use it.
7066 10. FUTURE REVISIONS OF THIS LICENSE
7068 The Free Software Foundation may publish new, revised versions of
7069 the GNU Free Documentation License from time to time. Such new
7070 versions will be similar in spirit to the present version, but may
7071 differ in detail to address new problems or concerns. See
7072 `http://www.gnu.org/copyleft/'.
7074 Each version of the License is given a distinguishing version
7075 number. If the Document specifies that a particular numbered
7076 version of this License "or any later version" applies to it, you
7077 have the option of following the terms and conditions either of
7078 that specified version or of any later version that has been
7079 published (not as a draft) by the Free Software Foundation. If
7080 the Document does not specify a version number of this License,
7081 you may choose any version ever published (not as a draft) by the
7082 Free Software Foundation. If the Document specifies that a proxy
7083 can decide which future versions of this License can be used, that
7084 proxy's public statement of acceptance of a version permanently
7085 authorizes you to choose that version for the Document.
7089 "Massive Multiauthor Collaboration Site" (or "MMC Site") means any
7090 World Wide Web server that publishes copyrightable works and also
7091 provides prominent facilities for anybody to edit those works. A
7092 public wiki that anybody can edit is an example of such a server.
7093 A "Massive Multiauthor Collaboration" (or "MMC") contained in the
7094 site means any set of copyrightable works thus published on the MMC
7097 "CC-BY-SA" means the Creative Commons Attribution-Share Alike 3.0
7098 license published by Creative Commons Corporation, a not-for-profit
7099 corporation with a principal place of business in San Francisco,
7100 California, as well as future copyleft versions of that license
7101 published by that same organization.
7103 "Incorporate" means to publish or republish a Document, in whole or
7104 in part, as part of another Document.
7106 An MMC is "eligible for relicensing" if it is licensed under this
7107 License, and if all works that were first published under this
7108 License somewhere other than this MMC, and subsequently
7109 incorporated in whole or in part into the MMC, (1) had no cover
7110 texts or invariant sections, and (2) were thus incorporated prior
7111 to November 1, 2008.
7113 The operator of an MMC Site may republish an MMC contained in the
7114 site under CC-BY-SA on the same site at any time before August 1,
7115 2009, provided the MMC is eligible for relicensing.
7118 ADDENDUM: How to use this License for your documents
7119 ====================================================
7121 To use this License in a document you have written, include a copy of
7122 the License in the document and put the following copyright and license
7123 notices just after the title page:
7125 Copyright (C) YEAR YOUR NAME.
7126 Permission is granted to copy, distribute and/or modify this document
7127 under the terms of the GNU Free Documentation License, Version 1.3
7128 or any later version published by the Free Software Foundation;
7129 with no Invariant Sections, no Front-Cover Texts, and no Back-Cover
7130 Texts. A copy of the license is included in the section entitled ``GNU
7131 Free Documentation License''.
7133 If you have Invariant Sections, Front-Cover Texts and Back-Cover
7134 Texts, replace the "with...Texts." line with this:
7136 with the Invariant Sections being LIST THEIR TITLES, with
7137 the Front-Cover Texts being LIST, and with the Back-Cover Texts
7140 If you have Invariant Sections without Cover Texts, or some other
7141 combination of the three, merge those two alternatives to suit the
7144 If your document contains nontrivial examples of program code, we
7145 recommend releasing these examples in parallel under your choice of
7146 free software license, such as the GNU General Public License, to
7147 permit their use in free software.
7150 File: ld.info, Node: LD Index, Prev: GNU Free Documentation License, Up: Top
7158 * ": Symbols. (line 6)
7159 * -(: Options. (line 696)
7160 * --accept-unknown-input-arch: Options. (line 714)
7161 * --add-needed: Options. (line 738)
7162 * --add-stdcall-alias: Options. (line 1583)
7163 * --allow-multiple-definition: Options. (line 989)
7164 * --allow-shlib-undefined: Options. (line 995)
7165 * --architecture=ARCH: Options. (line 123)
7166 * --as-needed: Options. (line 724)
7167 * --audit AUDITLIB: Options. (line 112)
7168 * --auxiliary=NAME: Options. (line 255)
7169 * --bank-window: Options. (line 2015)
7170 * --base-file: Options. (line 1588)
7171 * --be8: ARM. (line 28)
7172 * --bss-plt: PowerPC ELF32. (line 16)
7173 * --build-id: Options. (line 1545)
7174 * --build-id=STYLE: Options. (line 1545)
7175 * --check-sections: Options. (line 817)
7176 * --copy-dt-needed-entries: Options. (line 829)
7177 * --cref: Options. (line 849)
7178 * --default-imported-symver: Options. (line 1032)
7179 * --default-script=SCRIPT: Options. (line 541)
7180 * --default-symver: Options. (line 1028)
7181 * --defsym=SYMBOL=EXP: Options. (line 877)
7182 * --demangle[=STYLE]: Options. (line 890)
7183 * --depaudit AUDITLIB: Options. (line 177)
7184 * --disable-auto-image-base: Options. (line 1767)
7185 * --disable-auto-import: Options. (line 1902)
7186 * --disable-long-section-names: Options. (line 1598)
7187 * --disable-new-dtags: Options. (line 1508)
7188 * --disable-runtime-pseudo-reloc: Options. (line 1915)
7189 * --disable-stdcall-fixup: Options. (line 1620)
7190 * --discard-all: Options. (line 587)
7191 * --discard-locals: Options. (line 591)
7192 * --dll: Options. (line 1593)
7193 * --dll-search-prefix: Options. (line 1773)
7194 * --dotsyms: PowerPC64 ELF64. (line 33)
7195 * --dsbt-index: Options. (line 1992)
7196 * --dsbt-size: Options. (line 1987)
7197 * --dynamic-linker=FILE: Options. (line 903)
7198 * --dynamic-list-cpp-new: Options. (line 809)
7199 * --dynamic-list-cpp-typeinfo: Options. (line 813)
7200 * --dynamic-list-data: Options. (line 806)
7201 * --dynamic-list=DYNAMIC-LIST-FILE: Options. (line 793)
7202 * --dynamicbase: Options. (line 1951)
7203 * --eh-frame-hdr: Options. (line 1499)
7204 * --emit-relocs: Options. (line 476)
7205 * --emit-stack-syms: SPU ELF. (line 46)
7206 * --emit-stub-syms <1>: PowerPC64 ELF64. (line 29)
7207 * --emit-stub-syms <2>: PowerPC ELF32. (line 47)
7208 * --emit-stub-syms: SPU ELF. (line 15)
7209 * --enable-auto-image-base: Options. (line 1759)
7210 * --enable-auto-import: Options. (line 1782)
7211 * --enable-extra-pe-debug: Options. (line 1920)
7212 * --enable-long-section-names: Options. (line 1598)
7213 * --enable-new-dtags: Options. (line 1508)
7214 * --enable-runtime-pseudo-reloc: Options. (line 1907)
7215 * --enable-stdcall-fixup: Options. (line 1620)
7216 * --entry=ENTRY: Options. (line 187)
7217 * --error-unresolved-symbols: Options. (line 1452)
7218 * --exclude-all-symbols: Options. (line 1674)
7219 * --exclude-libs: Options. (line 197)
7220 * --exclude-modules-for-implib: Options. (line 208)
7221 * --exclude-symbols: Options. (line 1668)
7222 * --export-all-symbols: Options. (line 1644)
7223 * --export-dynamic: Options. (line 221)
7224 * --extra-overlay-stubs: SPU ELF. (line 19)
7225 * --fatal-warnings: Options. (line 910)
7226 * --file-alignment: Options. (line 1678)
7227 * --filter=NAME: Options. (line 276)
7228 * --fix-arm1176: ARM. (line 111)
7229 * --fix-cortex-a8: i960. (line 39)
7230 * --fix-v4bx: ARM. (line 49)
7231 * --fix-v4bx-interworking: ARM. (line 62)
7232 * --force-dynamic: Options. (line 485)
7233 * --force-exe-suffix: Options. (line 915)
7234 * --forceinteg: Options. (line 1956)
7235 * --format=FORMAT: Options. (line 134)
7236 * --format=VERSION: TI COFF. (line 6)
7237 * --gc-sections: Options. (line 925)
7238 * --got: Options. (line 2028)
7239 * --got=TYPE: M68K. (line 6)
7240 * --gpsize=VALUE: Options. (line 309)
7241 * --hash-size=NUMBER: Options. (line 1517)
7242 * --hash-style=STYLE: Options. (line 1525)
7243 * --heap: Options. (line 1684)
7244 * --help: Options. (line 962)
7245 * --image-base: Options. (line 1691)
7246 * --just-symbols=FILE: Options. (line 508)
7247 * --kill-at: Options. (line 1700)
7248 * --large-address-aware: Options. (line 1705)
7249 * --ld-generated-unwind-info: Options. (line 1503)
7250 * --leading-underscore: Options. (line 1638)
7251 * --library-path=DIR: Options. (line 368)
7252 * --library=NAMESPEC: Options. (line 335)
7253 * --local-store=lo:hi: SPU ELF. (line 24)
7254 * --major-image-version: Options. (line 1714)
7255 * --major-os-version: Options. (line 1719)
7256 * --major-subsystem-version: Options. (line 1723)
7257 * --merge-exidx-entries: i960. (line 48)
7258 * --minor-image-version: Options. (line 1728)
7259 * --minor-os-version: Options. (line 1733)
7260 * --minor-subsystem-version: Options. (line 1737)
7261 * --mri-script=MRI-CMDFILE: Options. (line 158)
7262 * --multi-subspace: HPPA ELF32. (line 6)
7263 * --nmagic: Options. (line 439)
7264 * --no-accept-unknown-input-arch: Options. (line 714)
7265 * --no-add-needed: Options. (line 738)
7266 * --no-allow-shlib-undefined: Options. (line 995)
7267 * --no-as-needed: Options. (line 724)
7268 * --no-bind: Options. (line 1970)
7269 * --no-check-sections: Options. (line 817)
7270 * --no-copy-dt-needed-entries: Options. (line 829)
7271 * --no-define-common: Options. (line 861)
7272 * --no-demangle: Options. (line 890)
7273 * --no-dotsyms: PowerPC64 ELF64. (line 33)
7274 * --no-enum-size-warning: ARM. (line 120)
7275 * --no-export-dynamic: Options. (line 221)
7276 * --no-fatal-warnings: Options. (line 910)
7277 * --no-fix-arm1176: ARM. (line 111)
7278 * --no-fix-cortex-a8: i960. (line 39)
7279 * --no-gc-sections: Options. (line 925)
7280 * --no-isolation: Options. (line 1963)
7281 * --no-keep-memory: Options. (line 974)
7282 * --no-leading-underscore: Options. (line 1638)
7283 * --no-merge-exidx-entries <1>: i960. (line 48)
7284 * --no-merge-exidx-entries: Options. (line 1999)
7285 * --no-multi-toc: PowerPC64 ELF64. (line 74)
7286 * --no-omagic: Options. (line 454)
7287 * --no-opd-optimize: PowerPC64 ELF64. (line 48)
7288 * --no-overlays: SPU ELF. (line 9)
7289 * --no-plt-align: PowerPC64 ELF64. (line 96)
7290 * --no-plt-static-chain: PowerPC64 ELF64. (line 103)
7291 * --no-plt-thread-safe: PowerPC64 ELF64. (line 109)
7292 * --no-print-gc-sections: Options. (line 947)
7293 * --no-seh: Options. (line 1966)
7294 * --no-tls-optimize <1>: PowerPC64 ELF64. (line 43)
7295 * --no-tls-optimize: PowerPC ELF32. (line 51)
7296 * --no-toc-optimize: PowerPC64 ELF64. (line 60)
7297 * --no-toc-sort: PowerPC64 ELF64. (line 86)
7298 * --no-trampoline: Options. (line 2009)
7299 * --no-undefined: Options. (line 981)
7300 * --no-undefined-version: Options. (line 1023)
7301 * --no-warn-mismatch: Options. (line 1036)
7302 * --no-warn-search-mismatch: Options. (line 1045)
7303 * --no-wchar-size-warning: ARM. (line 127)
7304 * --no-whole-archive: Options. (line 1049)
7305 * --noinhibit-exec: Options. (line 1053)
7306 * --non-overlapping-opd: PowerPC64 ELF64. (line 54)
7307 * --nxcompat: Options. (line 1959)
7308 * --oformat=OUTPUT-FORMAT: Options. (line 1065)
7309 * --omagic: Options. (line 445)
7310 * --out-implib: Options. (line 1750)
7311 * --output-def: Options. (line 1742)
7312 * --output=OUTPUT: Options. (line 460)
7313 * --pic-executable: Options. (line 1078)
7314 * --pic-veneer: ARM. (line 133)
7315 * --plt-align: PowerPC64 ELF64. (line 96)
7316 * --plt-static-chain: PowerPC64 ELF64. (line 103)
7317 * --plt-thread-safe: PowerPC64 ELF64. (line 109)
7318 * --plugin: SPU ELF. (line 6)
7319 * --print-gc-sections: Options. (line 947)
7320 * --print-map: Options. (line 402)
7321 * --print-output-format: Options. (line 956)
7322 * --reduce-memory-overheads: Options. (line 1531)
7323 * --relax: Options. (line 1094)
7324 * --relax on i960: i960. (line 31)
7325 * --relax on PowerPC: PowerPC ELF32. (line 6)
7326 * --relax on Xtensa: Xtensa. (line 27)
7327 * --relocatable: Options. (line 489)
7328 * --retain-symbols-file=FILENAME: Options. (line 1120)
7329 * --script=SCRIPT: Options. (line 532)
7330 * --sdata-got: PowerPC ELF32. (line 33)
7331 * --section-alignment: Options. (line 1925)
7332 * --section-start=SECTIONNAME=ORG: Options. (line 1276)
7333 * --secure-plt: PowerPC ELF32. (line 26)
7334 * --sort-common: Options. (line 1218)
7335 * --sort-section=alignment: Options. (line 1233)
7336 * --sort-section=name: Options. (line 1229)
7337 * --split-by-file: Options. (line 1237)
7338 * --split-by-reloc: Options. (line 1242)
7339 * --stack: Options. (line 1931)
7340 * --stack-analysis: SPU ELF. (line 29)
7341 * --stats: Options. (line 1255)
7342 * --strip-all: Options. (line 519)
7343 * --strip-debug: Options. (line 523)
7344 * --stub-group-size: PowerPC64 ELF64. (line 6)
7345 * --stub-group-size=N <1>: HPPA ELF32. (line 12)
7346 * --stub-group-size=N: ARM. (line 138)
7347 * --subsystem: Options. (line 1938)
7348 * --support-old-code: ARM. (line 6)
7349 * --sysroot=DIRECTORY: Options. (line 1259)
7350 * --target-help: Options. (line 966)
7351 * --target1-abs: ARM. (line 32)
7352 * --target1-rel: ARM. (line 32)
7353 * --target2=TYPE: ARM. (line 37)
7354 * --thumb-entry=ENTRY: ARM. (line 17)
7355 * --trace: Options. (line 528)
7356 * --trace-symbol=SYMBOL: Options. (line 597)
7357 * --traditional-format: Options. (line 1264)
7358 * --tsaware: Options. (line 1976)
7359 * --undefined=SYMBOL: Options. (line 554)
7360 * --unique[=SECTION]: Options. (line 572)
7361 * --unresolved-symbols: Options. (line 1295)
7362 * --use-blx: ARM. (line 74)
7363 * --use-nul-prefixed-import-tables: ARM. (line 23)
7364 * --verbose[=NUMBER]: Options. (line 1324)
7365 * --version: Options. (line 581)
7366 * --version-script=VERSION-SCRIPTFILE: Options. (line 1332)
7367 * --vfp11-denorm-fix: ARM. (line 83)
7368 * --warn-alternate-em: Options. (line 1444)
7369 * --warn-common: Options. (line 1343)
7370 * --warn-constructors: Options. (line 1411)
7371 * --warn-multiple-gp: Options. (line 1416)
7372 * --warn-once: Options. (line 1430)
7373 * --warn-section-align: Options. (line 1434)
7374 * --warn-shared-textrel: Options. (line 1441)
7375 * --warn-unresolved-symbols: Options. (line 1447)
7376 * --wdmdriver: Options. (line 1973)
7377 * --whole-archive: Options. (line 1456)
7378 * --wrap=SYMBOL: Options. (line 1470)
7379 * -A ARCH: Options. (line 122)
7380 * -a KEYWORD: Options. (line 105)
7381 * -assert KEYWORD: Options. (line 745)
7382 * -b FORMAT: Options. (line 134)
7383 * -Bdynamic: Options. (line 748)
7384 * -Bgroup: Options. (line 758)
7385 * -Bshareable: Options. (line 1211)
7386 * -Bstatic: Options. (line 765)
7387 * -Bsymbolic: Options. (line 780)
7388 * -Bsymbolic-functions: Options. (line 787)
7389 * -c MRI-CMDFILE: Options. (line 158)
7390 * -call_shared: Options. (line 748)
7391 * -d: Options. (line 168)
7392 * -dc: Options. (line 168)
7393 * -dn: Options. (line 765)
7394 * -dp: Options. (line 168)
7395 * -dT SCRIPT: Options. (line 541)
7396 * -dy: Options. (line 748)
7397 * -E: Options. (line 221)
7398 * -e ENTRY: Options. (line 187)
7399 * -EB: Options. (line 248)
7400 * -EL: Options. (line 251)
7401 * -f NAME: Options. (line 255)
7402 * -F NAME: Options. (line 276)
7403 * -fini=NAME: Options. (line 300)
7404 * -g: Options. (line 306)
7405 * -G VALUE: Options. (line 309)
7406 * -h NAME: Options. (line 317)
7407 * -i: Options. (line 326)
7408 * -IFILE: Options. (line 903)
7409 * -init=NAME: Options. (line 329)
7410 * -L DIR: Options. (line 368)
7411 * -l NAMESPEC: Options. (line 335)
7412 * -M: Options. (line 402)
7413 * -m EMULATION: Options. (line 392)
7414 * -Map=MAPFILE: Options. (line 970)
7415 * -N: Options. (line 445)
7416 * -n: Options. (line 439)
7417 * -no-relax: Options. (line 1094)
7418 * -non_shared: Options. (line 765)
7419 * -nostdlib: Options. (line 1059)
7420 * -O LEVEL: Options. (line 466)
7421 * -o OUTPUT: Options. (line 460)
7422 * -P AUDITLIB: Options. (line 177)
7423 * -pie: Options. (line 1078)
7424 * -q: Options. (line 476)
7425 * -qmagic: Options. (line 1088)
7426 * -Qy: Options. (line 1091)
7427 * -r: Options. (line 489)
7428 * -R FILE: Options. (line 508)
7429 * -rpath-link=DIR: Options. (line 1156)
7430 * -rpath=DIR: Options. (line 1134)
7431 * -S: Options. (line 523)
7432 * -s: Options. (line 519)
7433 * -shared: Options. (line 1211)
7434 * -soname=NAME: Options. (line 317)
7435 * -static: Options. (line 765)
7436 * -t: Options. (line 528)
7437 * -T SCRIPT: Options. (line 532)
7438 * -Tbss=ORG: Options. (line 1285)
7439 * -Tdata=ORG: Options. (line 1285)
7440 * -Ttext-segment=ORG: Options. (line 1291)
7441 * -Ttext=ORG: Options. (line 1285)
7442 * -u SYMBOL: Options. (line 554)
7443 * -Ur: Options. (line 562)
7444 * -v: Options. (line 581)
7445 * -V: Options. (line 581)
7446 * -x: Options. (line 587)
7447 * -X: Options. (line 591)
7448 * -Y PATH: Options. (line 606)
7449 * -y SYMBOL: Options. (line 597)
7450 * -z defs: Options. (line 981)
7451 * -z KEYWORD: Options. (line 610)
7452 * -z muldefs: Options. (line 989)
7453 * .: Location Counter. (line 6)
7454 * /DISCARD/: Output Section Discarding.
7456 * :PHDR: Output Section Phdr.
7458 * =FILLEXP: Output Section Fill.
7460 * >REGION: Output Section Region.
7462 * [COMMON]: Input Section Common.
7464 * ABSOLUTE (MRI): MRI. (line 33)
7465 * absolute and relocatable symbols: Expression Section. (line 6)
7466 * absolute expressions: Expression Section. (line 6)
7467 * ABSOLUTE(EXP): Builtin Functions. (line 10)
7468 * ADDR(SECTION): Builtin Functions. (line 17)
7469 * address, section: Output Section Address.
7471 * ALIAS (MRI): MRI. (line 44)
7472 * ALIGN (MRI): MRI. (line 50)
7473 * align expression: Builtin Functions. (line 38)
7474 * align location counter: Builtin Functions. (line 38)
7475 * ALIGN(ALIGN): Builtin Functions. (line 38)
7476 * ALIGN(EXP,ALIGN): Builtin Functions. (line 38)
7477 * ALIGN(SECTION_ALIGN): Forced Output Alignment.
7479 * aligned common symbols: WIN32. (line 424)
7480 * ALIGNOF(SECTION): Builtin Functions. (line 64)
7481 * allocating memory: MEMORY. (line 6)
7482 * architecture: Miscellaneous Commands.
7484 * architectures: Options. (line 122)
7485 * archive files, from cmd line: Options. (line 335)
7486 * archive search path in linker script: File Commands. (line 74)
7487 * arithmetic: Expressions. (line 6)
7488 * arithmetic operators: Operators. (line 6)
7489 * ARM interworking support: ARM. (line 6)
7490 * ARM1176 erratum workaround: ARM. (line 111)
7491 * AS_NEEDED(FILES): File Commands. (line 54)
7492 * ASSERT: Miscellaneous Commands.
7494 * assertion in linker script: Miscellaneous Commands.
7496 * assignment in scripts: Assignments. (line 6)
7497 * AT(LMA): Output Section LMA. (line 6)
7498 * AT>LMA_REGION: Output Section LMA. (line 6)
7499 * automatic data imports: WIN32. (line 191)
7500 * back end: BFD. (line 6)
7501 * BASE (MRI): MRI. (line 54)
7502 * BE8: ARM. (line 28)
7503 * BFD canonical format: Canonical format. (line 11)
7504 * BFD requirements: BFD. (line 16)
7505 * big-endian objects: Options. (line 248)
7506 * binary input format: Options. (line 134)
7507 * BLOCK(EXP): Builtin Functions. (line 77)
7508 * bug criteria: Bug Criteria. (line 6)
7509 * bug reports: Bug Reporting. (line 6)
7510 * bugs in ld: Reporting Bugs. (line 6)
7511 * BYTE(EXPRESSION): Output Section Data.
7513 * C++ constructors, arranging in link: Output Section Keywords.
7515 * CHIP (MRI): MRI. (line 58)
7516 * COLLECT_NO_DEMANGLE: Environment. (line 29)
7517 * combining symbols, warnings on: Options. (line 1343)
7518 * command files: Scripts. (line 6)
7519 * command line: Options. (line 6)
7520 * common allocation: Options. (line 861)
7521 * common allocation in linker script: Miscellaneous Commands.
7523 * common symbol placement: Input Section Common.
7525 * COMMONPAGESIZE: Symbolic Constants. (line 13)
7526 * compatibility, MRI: Options. (line 158)
7527 * CONSTANT: Symbolic Constants. (line 6)
7528 * constants in linker scripts: Constants. (line 6)
7529 * constraints on output sections: Output Section Constraint.
7531 * CONSTRUCTORS: Output Section Keywords.
7533 * constructors: Options. (line 562)
7534 * constructors, arranging in link: Output Section Keywords.
7536 * Cortex-A8 erratum workaround: i960. (line 39)
7537 * crash of linker: Bug Criteria. (line 9)
7538 * CREATE_OBJECT_SYMBOLS: Output Section Keywords.
7540 * creating a DEF file: WIN32. (line 158)
7541 * cross reference table: Options. (line 849)
7542 * cross references: Miscellaneous Commands.
7544 * current output location: Location Counter. (line 6)
7545 * data: Output Section Data.
7547 * DATA_SEGMENT_ALIGN(MAXPAGESIZE, COMMONPAGESIZE): Builtin Functions.
7549 * DATA_SEGMENT_END(EXP): Builtin Functions. (line 103)
7550 * DATA_SEGMENT_RELRO_END(OFFSET, EXP): Builtin Functions. (line 109)
7551 * dbx: Options. (line 1269)
7552 * DEF files, creating: Options. (line 1742)
7553 * default emulation: Environment. (line 21)
7554 * default input format: Environment. (line 9)
7555 * DEFINED(SYMBOL): Builtin Functions. (line 120)
7556 * deleting local symbols: Options. (line 587)
7557 * demangling, default: Environment. (line 29)
7558 * demangling, from command line: Options. (line 890)
7559 * direct linking to a dll: WIN32. (line 239)
7560 * discarding sections: Output Section Discarding.
7562 * discontinuous memory: MEMORY. (line 6)
7563 * DLLs, creating: Options. (line 1742)
7564 * DLLs, linking to: Options. (line 1773)
7565 * dot: Location Counter. (line 6)
7566 * dot inside sections: Location Counter. (line 36)
7567 * dot outside sections: Location Counter. (line 66)
7568 * dynamic linker, from command line: Options. (line 903)
7569 * dynamic symbol table: Options. (line 221)
7570 * ELF program headers: PHDRS. (line 6)
7571 * emulation: Options. (line 392)
7572 * emulation, default: Environment. (line 21)
7573 * END (MRI): MRI. (line 62)
7574 * endianness: Options. (line 248)
7575 * entry point: Entry Point. (line 6)
7576 * entry point, from command line: Options. (line 187)
7577 * entry point, thumb: ARM. (line 17)
7578 * ENTRY(SYMBOL): Entry Point. (line 6)
7579 * error on valid input: Bug Criteria. (line 12)
7580 * example of linker script: Simple Example. (line 6)
7581 * exporting DLL symbols: WIN32. (line 19)
7582 * expression evaluation order: Evaluation. (line 6)
7583 * expression sections: Expression Section. (line 6)
7584 * expression, absolute: Builtin Functions. (line 10)
7585 * expressions: Expressions. (line 6)
7586 * EXTERN: Miscellaneous Commands.
7588 * fatal signal: Bug Criteria. (line 9)
7589 * file name wildcard patterns: Input Section Wildcards.
7591 * FILEHDR: PHDRS. (line 62)
7592 * filename symbols: Output Section Keywords.
7594 * fill pattern, entire section: Output Section Fill.
7596 * FILL(EXPRESSION): Output Section Data.
7598 * finalization function: Options. (line 300)
7599 * first input file: File Commands. (line 82)
7600 * first instruction: Entry Point. (line 6)
7601 * FIX_V4BX: ARM. (line 49)
7602 * FIX_V4BX_INTERWORKING: ARM. (line 62)
7603 * FORCE_COMMON_ALLOCATION: Miscellaneous Commands.
7605 * forcing input section alignment: Forced Input Alignment.
7607 * forcing output section alignment: Forced Output Alignment.
7609 * forcing the creation of dynamic sections: Options. (line 485)
7610 * FORMAT (MRI): MRI. (line 66)
7611 * functions in expressions: Builtin Functions. (line 6)
7612 * garbage collection <1>: Input Section Keep. (line 6)
7613 * garbage collection: Options. (line 925)
7614 * generating optimized output: Options. (line 466)
7615 * GNU linker: Overview. (line 6)
7616 * GNUTARGET: Environment. (line 9)
7617 * GROUP(FILES): File Commands. (line 47)
7618 * grouping input files: File Commands. (line 47)
7619 * groups of archives: Options. (line 696)
7620 * H8/300 support: H8/300. (line 6)
7621 * header size: Builtin Functions. (line 183)
7622 * heap size: Options. (line 1684)
7623 * help: Options. (line 962)
7624 * HIDDEN: HIDDEN. (line 6)
7625 * holes: Location Counter. (line 12)
7626 * holes, filling: Output Section Data.
7628 * HPPA multiple sub-space stubs: HPPA ELF32. (line 6)
7629 * HPPA stub grouping: HPPA ELF32. (line 12)
7630 * i960 support: i960. (line 6)
7631 * image base: Options. (line 1691)
7632 * implicit linker scripts: Implicit Linker Scripts.
7634 * import libraries: WIN32. (line 10)
7635 * INCLUDE FILENAME: File Commands. (line 9)
7636 * including a linker script: File Commands. (line 9)
7637 * including an entire archive: Options. (line 1456)
7638 * incremental link: Options. (line 326)
7639 * INHIBIT_COMMON_ALLOCATION: Miscellaneous Commands.
7641 * initialization function: Options. (line 329)
7642 * initialized data in ROM: Output Section LMA. (line 39)
7643 * input file format in linker script: Format Commands. (line 35)
7644 * input filename symbols: Output Section Keywords.
7646 * input files in linker scripts: File Commands. (line 19)
7647 * input files, displaying: Options. (line 528)
7648 * input format: Options. (line 134)
7649 * input object files in linker scripts: File Commands. (line 19)
7650 * input section alignment: Forced Input Alignment.
7652 * input section basics: Input Section Basics.
7654 * input section wildcards: Input Section Wildcards.
7656 * input sections: Input Section. (line 6)
7657 * INPUT(FILES): File Commands. (line 19)
7658 * INSERT: Miscellaneous Commands.
7660 * insert user script into default script: Miscellaneous Commands.
7662 * integer notation: Constants. (line 6)
7663 * integer suffixes: Constants. (line 15)
7664 * internal object-file format: Canonical format. (line 11)
7665 * invalid input: Bug Criteria. (line 14)
7666 * K and M integer suffixes: Constants. (line 15)
7667 * KEEP: Input Section Keep. (line 6)
7668 * l =: MEMORY. (line 74)
7669 * lazy evaluation: Evaluation. (line 6)
7670 * ld bugs, reporting: Bug Reporting. (line 6)
7671 * LD_FEATURE(STRING): Miscellaneous Commands.
7673 * LDEMULATION: Environment. (line 21)
7674 * len =: MEMORY. (line 74)
7675 * LENGTH =: MEMORY. (line 74)
7676 * LENGTH(MEMORY): Builtin Functions. (line 137)
7677 * library search path in linker script: File Commands. (line 74)
7678 * link map: Options. (line 402)
7679 * link-time runtime library search path: Options. (line 1156)
7680 * linker crash: Bug Criteria. (line 9)
7681 * linker script concepts: Basic Script Concepts.
7683 * linker script example: Simple Example. (line 6)
7684 * linker script file commands: File Commands. (line 6)
7685 * linker script format: Script Format. (line 6)
7686 * linker script input object files: File Commands. (line 19)
7687 * linker script simple commands: Simple Commands. (line 6)
7688 * linker scripts: Scripts. (line 6)
7689 * LIST (MRI): MRI. (line 77)
7690 * little-endian objects: Options. (line 251)
7691 * LOAD (MRI): MRI. (line 84)
7692 * load address: Output Section LMA. (line 6)
7693 * LOADADDR(SECTION): Builtin Functions. (line 140)
7694 * loading, preventing: Output Section Type.
7696 * local symbols, deleting: Options. (line 591)
7697 * location counter: Location Counter. (line 6)
7698 * LONG(EXPRESSION): Output Section Data.
7700 * M and K integer suffixes: Constants. (line 15)
7701 * M68HC11 and 68HC12 support: M68HC11/68HC12. (line 6)
7702 * machine architecture: Miscellaneous Commands.
7704 * machine dependencies: Machine Dependent. (line 6)
7705 * mapping input sections to output sections: Input Section. (line 6)
7706 * MAX: Builtin Functions. (line 143)
7707 * MAXPAGESIZE: Symbolic Constants. (line 10)
7708 * MEMORY: MEMORY. (line 6)
7709 * memory region attributes: MEMORY. (line 34)
7710 * memory regions: MEMORY. (line 6)
7711 * memory regions and sections: Output Section Region.
7713 * memory usage: Options. (line 974)
7714 * MIN: Builtin Functions. (line 146)
7715 * Motorola 68K GOT generation: M68K. (line 6)
7716 * MRI compatibility: MRI. (line 6)
7717 * MSP430 extra sections: MSP430. (line 11)
7718 * NAME (MRI): MRI. (line 90)
7719 * name, section: Output Section Name.
7721 * names: Symbols. (line 6)
7722 * naming the output file: Options. (line 460)
7723 * NEXT(EXP): Builtin Functions. (line 150)
7724 * NMAGIC: Options. (line 439)
7725 * NO_ENUM_SIZE_WARNING: ARM. (line 120)
7726 * NO_WCHAR_SIZE_WARNING: ARM. (line 127)
7727 * NOCROSSREFS(SECTIONS): Miscellaneous Commands.
7729 * NOLOAD: Output Section Type.
7731 * not enough room for program headers: Builtin Functions. (line 188)
7732 * o =: MEMORY. (line 69)
7733 * objdump -i: BFD. (line 6)
7734 * object file management: BFD. (line 6)
7735 * object files: Options. (line 29)
7736 * object formats available: BFD. (line 6)
7737 * object size: Options. (line 309)
7738 * OMAGIC: Options. (line 454)
7739 * ONLY_IF_RO: Output Section Constraint.
7741 * ONLY_IF_RW: Output Section Constraint.
7743 * opening object files: BFD outline. (line 6)
7744 * operators for arithmetic: Operators. (line 6)
7745 * options: Options. (line 6)
7746 * ORDER (MRI): MRI. (line 95)
7747 * org =: MEMORY. (line 69)
7748 * ORIGIN =: MEMORY. (line 69)
7749 * ORIGIN(MEMORY): Builtin Functions. (line 156)
7750 * orphan: Orphan Sections. (line 6)
7751 * output file after errors: Options. (line 1053)
7752 * output file format in linker script: Format Commands. (line 10)
7753 * output file name in linker script: File Commands. (line 64)
7754 * output format: Options. (line 956)
7755 * output section alignment: Forced Output Alignment.
7757 * output section attributes: Output Section Attributes.
7759 * output section data: Output Section Data.
7761 * OUTPUT(FILENAME): File Commands. (line 64)
7762 * OUTPUT_ARCH(BFDARCH): Miscellaneous Commands.
7764 * OUTPUT_FORMAT(BFDNAME): Format Commands. (line 10)
7765 * OVERLAY: Overlay Description.
7767 * overlays: Overlay Description.
7769 * partial link: Options. (line 489)
7770 * PE import table prefixing: ARM. (line 23)
7771 * PHDRS: PHDRS. (line 62)
7772 * PIC_VENEER: ARM. (line 133)
7773 * position independent executables: Options. (line 1080)
7774 * PowerPC ELF32 options: PowerPC ELF32. (line 16)
7775 * PowerPC GOT: PowerPC ELF32. (line 33)
7776 * PowerPC long branches: PowerPC ELF32. (line 6)
7777 * PowerPC PLT: PowerPC ELF32. (line 16)
7778 * PowerPC stub symbols: PowerPC ELF32. (line 47)
7779 * PowerPC TLS optimization: PowerPC ELF32. (line 51)
7780 * PowerPC64 dot symbols: PowerPC64 ELF64. (line 33)
7781 * PowerPC64 ELF64 options: PowerPC64 ELF64. (line 6)
7782 * PowerPC64 multi-TOC: PowerPC64 ELF64. (line 74)
7783 * PowerPC64 OPD optimization: PowerPC64 ELF64. (line 48)
7784 * PowerPC64 OPD spacing: PowerPC64 ELF64. (line 54)
7785 * PowerPC64 PLT call stub static chain: PowerPC64 ELF64. (line 103)
7786 * PowerPC64 PLT call stub thread safety: PowerPC64 ELF64. (line 109)
7787 * PowerPC64 PLT stub alignment: PowerPC64 ELF64. (line 96)
7788 * PowerPC64 stub grouping: PowerPC64 ELF64. (line 6)
7789 * PowerPC64 stub symbols: PowerPC64 ELF64. (line 29)
7790 * PowerPC64 TLS optimization: PowerPC64 ELF64. (line 43)
7791 * PowerPC64 TOC optimization: PowerPC64 ELF64. (line 60)
7792 * PowerPC64 TOC sorting: PowerPC64 ELF64. (line 86)
7793 * precedence in expressions: Operators. (line 6)
7794 * prevent unnecessary loading: Output Section Type.
7796 * program headers: PHDRS. (line 6)
7797 * program headers and sections: Output Section Phdr.
7799 * program headers, not enough room: Builtin Functions. (line 188)
7800 * program segments: PHDRS. (line 6)
7801 * PROVIDE: PROVIDE. (line 6)
7802 * PROVIDE_HIDDEN: PROVIDE_HIDDEN. (line 6)
7803 * PUBLIC (MRI): MRI. (line 103)
7804 * QUAD(EXPRESSION): Output Section Data.
7806 * quoted symbol names: Symbols. (line 6)
7807 * read-only text: Options. (line 439)
7808 * read/write from cmd line: Options. (line 445)
7809 * region alias: REGION_ALIAS. (line 6)
7810 * region names: REGION_ALIAS. (line 6)
7811 * REGION_ALIAS(ALIAS, REGION): REGION_ALIAS. (line 6)
7812 * regions of memory: MEMORY. (line 6)
7813 * relative expressions: Expression Section. (line 6)
7814 * relaxing addressing modes: Options. (line 1094)
7815 * relaxing on H8/300: H8/300. (line 9)
7816 * relaxing on i960: i960. (line 31)
7817 * relaxing on M68HC11: M68HC11/68HC12. (line 12)
7818 * relaxing on Xtensa: Xtensa. (line 27)
7819 * relocatable and absolute symbols: Expression Section. (line 6)
7820 * relocatable output: Options. (line 489)
7821 * removing sections: Output Section Discarding.
7823 * reporting bugs in ld: Reporting Bugs. (line 6)
7824 * requirements for BFD: BFD. (line 16)
7825 * retain relocations in final executable: Options. (line 476)
7826 * retaining specified symbols: Options. (line 1120)
7827 * ROM initialized data: Output Section LMA. (line 39)
7828 * round up expression: Builtin Functions. (line 38)
7829 * round up location counter: Builtin Functions. (line 38)
7830 * runtime library name: Options. (line 317)
7831 * runtime library search path: Options. (line 1134)
7832 * runtime pseudo-relocation: WIN32. (line 217)
7833 * scaled integers: Constants. (line 15)
7834 * scommon section: Input Section Common.
7836 * script files: Options. (line 541)
7837 * scripts: Scripts. (line 6)
7838 * search directory, from cmd line: Options. (line 368)
7839 * search path in linker script: File Commands. (line 74)
7840 * SEARCH_DIR(PATH): File Commands. (line 74)
7841 * SECT (MRI): MRI. (line 109)
7842 * section address: Output Section Address.
7844 * section address in expression: Builtin Functions. (line 17)
7845 * section alignment: Builtin Functions. (line 64)
7846 * section alignment, warnings on: Options. (line 1434)
7847 * section data: Output Section Data.
7849 * section fill pattern: Output Section Fill.
7851 * section load address: Output Section LMA. (line 6)
7852 * section load address in expression: Builtin Functions. (line 140)
7853 * section name: Output Section Name.
7855 * section name wildcard patterns: Input Section Wildcards.
7857 * section size: Builtin Functions. (line 167)
7858 * section, assigning to memory region: Output Section Region.
7860 * section, assigning to program header: Output Section Phdr.
7862 * SECTIONS: SECTIONS. (line 6)
7863 * sections, discarding: Output Section Discarding.
7865 * segment origins, cmd line: Options. (line 1285)
7866 * SEGMENT_START(SEGMENT, DEFAULT): Builtin Functions. (line 159)
7867 * segments, ELF: PHDRS. (line 6)
7868 * shared libraries: Options. (line 1213)
7869 * SHORT(EXPRESSION): Output Section Data.
7871 * SIZEOF(SECTION): Builtin Functions. (line 167)
7872 * SIZEOF_HEADERS: Builtin Functions. (line 183)
7873 * small common symbols: Input Section Common.
7875 * SORT: Input Section Wildcards.
7877 * SORT_BY_ALIGNMENT: Input Section Wildcards.
7879 * SORT_BY_INIT_PRIORITY: Input Section Wildcards.
7881 * SORT_BY_NAME: Input Section Wildcards.
7883 * SORT_NONE: Input Section Wildcards.
7885 * SPU: SPU ELF. (line 46)
7886 * SPU ELF options: SPU ELF. (line 6)
7887 * SPU extra overlay stubs: SPU ELF. (line 19)
7888 * SPU local store size: SPU ELF. (line 24)
7889 * SPU overlay stub symbols: SPU ELF. (line 15)
7890 * SPU overlays: SPU ELF. (line 9)
7891 * SPU plugins: SPU ELF. (line 6)
7892 * SQUAD(EXPRESSION): Output Section Data.
7894 * stack size: Options. (line 1931)
7895 * standard Unix system: Options. (line 7)
7896 * start of execution: Entry Point. (line 6)
7897 * STARTUP(FILENAME): File Commands. (line 82)
7898 * strip all symbols: Options. (line 519)
7899 * strip debugger symbols: Options. (line 523)
7900 * stripping all but some symbols: Options. (line 1120)
7901 * STUB_GROUP_SIZE: ARM. (line 138)
7902 * SUBALIGN(SUBSECTION_ALIGN): Forced Input Alignment.
7904 * suffixes for integers: Constants. (line 15)
7905 * symbol defaults: Builtin Functions. (line 120)
7906 * symbol definition, scripts: Assignments. (line 6)
7907 * symbol names: Symbols. (line 6)
7908 * symbol tracing: Options. (line 597)
7909 * symbol versions: VERSION. (line 6)
7910 * symbol-only input: Options. (line 508)
7911 * symbolic constants: Symbolic Constants. (line 6)
7912 * symbols, from command line: Options. (line 877)
7913 * symbols, relocatable and absolute: Expression Section. (line 6)
7914 * symbols, retaining selectively: Options. (line 1120)
7915 * synthesizing linker: Options. (line 1094)
7916 * synthesizing on H8/300: H8/300. (line 14)
7917 * TARGET(BFDNAME): Format Commands. (line 35)
7918 * TARGET1: ARM. (line 32)
7919 * TARGET2: ARM. (line 37)
7920 * text segment origin, cmd line: Options. (line 1292)
7921 * thumb entry point: ARM. (line 17)
7922 * TI COFF versions: TI COFF. (line 6)
7923 * traditional format: Options. (line 1264)
7924 * trampoline generation on M68HC11: M68HC11/68HC12. (line 31)
7925 * trampoline generation on M68HC12: M68HC11/68HC12. (line 31)
7926 * unallocated address, next: Builtin Functions. (line 150)
7927 * undefined symbol: Options. (line 554)
7928 * undefined symbol in linker script: Miscellaneous Commands.
7930 * undefined symbols, warnings on: Options. (line 1430)
7931 * uninitialized data placement: Input Section Common.
7933 * unspecified memory: Output Section Data.
7935 * usage: Options. (line 962)
7936 * USE_BLX: ARM. (line 74)
7937 * using a DEF file: WIN32. (line 57)
7938 * using auto-export functionality: WIN32. (line 22)
7939 * Using decorations: WIN32. (line 162)
7940 * variables, defining: Assignments. (line 6)
7941 * verbose[=NUMBER]: Options. (line 1324)
7942 * version: Options. (line 581)
7943 * version script: VERSION. (line 6)
7944 * version script, symbol versions: Options. (line 1332)
7945 * VERSION {script text}: VERSION. (line 6)
7946 * versions of symbols: VERSION. (line 6)
7947 * VFP11_DENORM_FIX: ARM. (line 83)
7948 * warnings, on combining symbols: Options. (line 1343)
7949 * warnings, on section alignment: Options. (line 1434)
7950 * warnings, on undefined symbols: Options. (line 1430)
7951 * weak externals: WIN32. (line 407)
7952 * what is this?: Overview. (line 6)
7953 * wildcard file name patterns: Input Section Wildcards.
7955 * Xtensa options: Xtensa. (line 56)
7956 * Xtensa processors: Xtensa. (line 6)
7962 Node: Overview
\7f1600
7963 Node: Invocation
\7f2714
7964 Node: Options
\7f3122
7965 Node: Environment
\7f94794
7966 Node: Scripts
\7f96554
7967 Node: Basic Script Concepts
\7f98288
7968 Node: Script Format
\7f100995
7969 Node: Simple Example
\7f101858
7970 Node: Simple Commands
\7f104954
7971 Node: Entry Point
\7f105460
7972 Node: File Commands
\7f106393
7973 Node: Format Commands
\7f110394
7974 Node: REGION_ALIAS
\7f112350
7975 Node: Miscellaneous Commands
\7f117182
7976 Node: Assignments
\7f120790
7977 Node: Simple Assignments
\7f121301
7978 Node: HIDDEN
\7f123036
7979 Node: PROVIDE
\7f123666
7980 Node: PROVIDE_HIDDEN
\7f124859
7981 Node: Source Code Reference
\7f125103
7982 Node: SECTIONS
\7f128683
7983 Node: Output Section Description
\7f130574
7984 Node: Output Section Name
\7f131661
7985 Node: Output Section Address
\7f132537
7986 Node: Input Section
\7f134772
7987 Node: Input Section Basics
\7f135573
7988 Node: Input Section Wildcards
\7f139479
7989 Node: Input Section Common
\7f144573
7990 Node: Input Section Keep
\7f146055
7991 Node: Input Section Example
\7f146545
7992 Node: Output Section Data
\7f147513
7993 Node: Output Section Keywords
\7f150290
7994 Node: Output Section Discarding
\7f153859
7995 Node: Output Section Attributes
\7f155040
7996 Node: Output Section Type
\7f156141
7997 Node: Output Section LMA
\7f157212
7998 Node: Forced Output Alignment
\7f160283
7999 Node: Forced Input Alignment
\7f160551
8000 Node: Output Section Constraint
\7f160940
8001 Node: Output Section Region
\7f161368
8002 Node: Output Section Phdr
\7f161801
8003 Node: Output Section Fill
\7f162465
8004 Node: Overlay Description
\7f163607
8005 Node: MEMORY
\7f167910
8006 Node: PHDRS
\7f172245
8007 Node: VERSION
\7f177499
8008 Node: Expressions
\7f185592
8009 Node: Constants
\7f186521
8010 Node: Symbolic Constants
\7f187396
8011 Node: Symbols
\7f187947
8012 Node: Orphan Sections
\7f188694
8013 Node: Location Counter
\7f189859
8014 Node: Operators
\7f194295
8015 Node: Evaluation
\7f195217
8016 Node: Expression Section
\7f196581
8017 Node: Builtin Functions
\7f200238
8018 Node: Implicit Linker Scripts
\7f208199
8019 Node: Machine Dependent
\7f208974
8020 Node: H8/300
\7f209990
8022 Node: M68HC11/68HC12
\7f213819
8024 Node: HPPA ELF32
\7f223236
8027 Node: MSP430
\7f226933
8028 Node: PowerPC ELF32
\7f227982
8029 Node: PowerPC64 ELF64
\7f230818
8030 Node: SPU ELF
\7f236974
8031 Node: TI COFF
\7f239606
8032 Node: WIN32
\7f240132
8033 Node: Xtensa
\7f260257
8035 Node: BFD outline
\7f264677
8036 Node: BFD information loss
\7f265963
8037 Node: Canonical format
\7f268480
8038 Node: Reporting Bugs
\7f272837
8039 Node: Bug Criteria
\7f273531
8040 Node: Bug Reporting
\7f274230
8042 Node: GNU Free Documentation License
\7f285912
8043 Node: LD Index
\7f311068