3 @c Copyright (C) 1991-2020 Free Software Foundation, Inc.
6 @include configdoc.texi
7 @c (configdoc.texi is generated by the Makefile)
13 @macro gcctabopt{body}
19 @c Configure for the generation of man pages
46 @dircategory Software development
48 * Ld: (ld). The GNU linker.
53 This file documents the @sc{gnu} linker LD
54 @ifset VERSION_PACKAGE
55 @value{VERSION_PACKAGE}
57 version @value{VERSION}.
59 Copyright @copyright{} 1991-2020 Free Software Foundation, Inc.
61 Permission is granted to copy, distribute and/or modify this document
62 under the terms of the GNU Free Documentation License, Version 1.3
63 or any later version published by the Free Software Foundation;
64 with no Invariant Sections, with no Front-Cover Texts, and with no
65 Back-Cover Texts. A copy of the license is included in the
66 section entitled ``GNU Free Documentation License''.
70 @setchapternewpage odd
71 @settitle The GNU linker
76 @ifset VERSION_PACKAGE
77 @subtitle @value{VERSION_PACKAGE}
79 @subtitle Version @value{VERSION}
80 @author Steve Chamberlain
81 @author Ian Lance Taylor
86 \hfill Red Hat Inc\par
87 \hfill nickc\@credhat.com, doc\@redhat.com\par
88 \hfill {\it The GNU linker}\par
89 \hfill Edited by Jeffrey Osier (jeffrey\@cygnus.com)\par
91 \global\parindent=0pt % Steve likes it this way.
94 @vskip 0pt plus 1filll
95 @c man begin COPYRIGHT
96 Copyright @copyright{} 1991-2020 Free Software Foundation, Inc.
98 Permission is granted to copy, distribute and/or modify this document
99 under the terms of the GNU Free Documentation License, Version 1.3
100 or any later version published by the Free Software Foundation;
101 with no Invariant Sections, with no Front-Cover Texts, and with no
102 Back-Cover Texts. A copy of the license is included in the
103 section entitled ``GNU Free Documentation License''.
109 @c FIXME: Talk about importance of *order* of args, cmds to linker!
114 This file documents the @sc{gnu} linker ld
115 @ifset VERSION_PACKAGE
116 @value{VERSION_PACKAGE}
118 version @value{VERSION}.
120 This document is distributed under the terms of the GNU Free
121 Documentation License version 1.3. A copy of the license is included
122 in the section entitled ``GNU Free Documentation License''.
125 * Overview:: Overview
126 * Invocation:: Invocation
127 * Scripts:: Linker Scripts
129 * Machine Dependent:: Machine Dependent Features
133 * H8/300:: ld and the H8/300
136 * Renesas:: ld and other Renesas micros
139 * ARM:: ld and the ARM family
142 * M68HC11/68HC12:: ld and the Motorola 68HC11 and 68HC12 families
145 * HPPA ELF32:: ld and HPPA 32-bit ELF
148 * M68K:: ld and Motorola 68K family
151 * MIPS:: ld and MIPS family
154 * PowerPC ELF32:: ld and PowerPC 32-bit ELF Support
157 * PowerPC64 ELF64:: ld and PowerPC64 64-bit ELF Support
160 * S/390 ELF:: ld and S/390 ELF Support
163 * SPU ELF:: ld and SPU ELF Support
166 * TI COFF:: ld and the TI COFF
169 * Win32:: ld and WIN32 (cygwin/mingw)
172 * Xtensa:: ld and Xtensa Processors
175 @ifclear SingleFormat
178 @c Following blank line required for remaining bug in makeinfo conds/menus
180 * Reporting Bugs:: Reporting Bugs
181 * MRI:: MRI Compatible Script Files
182 * GNU Free Documentation License:: GNU Free Documentation License
183 * LD Index:: LD Index
190 @cindex @sc{gnu} linker
191 @cindex what is this?
194 @c man begin SYNOPSIS
195 ld [@b{options}] @var{objfile} @dots{}
199 ar(1), nm(1), objcopy(1), objdump(1), readelf(1) and
200 the Info entries for @file{binutils} and
205 @c man begin DESCRIPTION
207 @command{ld} combines a number of object and archive files, relocates
208 their data and ties up symbol references. Usually the last step in
209 compiling a program is to run @command{ld}.
211 @command{ld} accepts Linker Command Language files written in
212 a superset of AT&T's Link Editor Command Language syntax,
213 to provide explicit and total control over the linking process.
217 This man page does not describe the command language; see the
218 @command{ld} entry in @code{info} for full details on the command
219 language and on other aspects of the GNU linker.
222 @ifclear SingleFormat
223 This version of @command{ld} uses the general purpose BFD libraries
224 to operate on object files. This allows @command{ld} to read, combine, and
225 write object files in many different formats---for example, COFF or
226 @code{a.out}. Different formats may be linked together to produce any
227 available kind of object file. @xref{BFD}, for more information.
230 Aside from its flexibility, the @sc{gnu} linker is more helpful than other
231 linkers in providing diagnostic information. Many linkers abandon
232 execution immediately upon encountering an error; whenever possible,
233 @command{ld} continues executing, allowing you to identify other errors
234 (or, in some cases, to get an output file in spite of the error).
241 @c man begin DESCRIPTION
243 The @sc{gnu} linker @command{ld} is meant to cover a broad range of situations,
244 and to be as compatible as possible with other linkers. As a result,
245 you have many choices to control its behavior.
251 * Options:: Command-line Options
252 * Environment:: Environment Variables
256 @section Command-line Options
264 The linker supports a plethora of command-line options, but in actual
265 practice few of them are used in any particular context.
266 @cindex standard Unix system
267 For instance, a frequent use of @command{ld} is to link standard Unix
268 object files on a standard, supported Unix system. On such a system, to
269 link a file @code{hello.o}:
272 ld -o @var{output} /lib/crt0.o hello.o -lc
275 This tells @command{ld} to produce a file called @var{output} as the
276 result of linking the file @code{/lib/crt0.o} with @code{hello.o} and
277 the library @code{libc.a}, which will come from the standard search
278 directories. (See the discussion of the @samp{-l} option below.)
280 Some of the command-line options to @command{ld} may be specified at any
281 point in the command line. However, options which refer to files, such
282 as @samp{-l} or @samp{-T}, cause the file to be read at the point at
283 which the option appears in the command line, relative to the object
284 files and other file options. Repeating non-file options with a
285 different argument will either have no further effect, or override prior
286 occurrences (those further to the left on the command line) of that
287 option. Options which may be meaningfully specified more than once are
288 noted in the descriptions below.
291 Non-option arguments are object files or archives which are to be linked
292 together. They may follow, precede, or be mixed in with command-line
293 options, except that an object file argument may not be placed between
294 an option and its argument.
296 Usually the linker is invoked with at least one object file, but you can
297 specify other forms of binary input files using @samp{-l}, @samp{-R},
298 and the script command language. If @emph{no} binary input files at all
299 are specified, the linker does not produce any output, and issues the
300 message @samp{No input files}.
302 If the linker cannot recognize the format of an object file, it will
303 assume that it is a linker script. A script specified in this way
304 augments the main linker script used for the link (either the default
305 linker script or the one specified by using @samp{-T}). This feature
306 permits the linker to link against a file which appears to be an object
307 or an archive, but actually merely defines some symbol values, or uses
308 @code{INPUT} or @code{GROUP} to load other objects. Specifying a
309 script in this way merely augments the main linker script, with the
310 extra commands placed after the main script; use the @samp{-T} option
311 to replace the default linker script entirely, but note the effect of
312 the @code{INSERT} command. @xref{Scripts}.
314 For options whose names are a single letter,
315 option arguments must either follow the option letter without intervening
316 whitespace, or be given as separate arguments immediately following the
317 option that requires them.
319 For options whose names are multiple letters, either one dash or two can
320 precede the option name; for example, @samp{-trace-symbol} and
321 @samp{--trace-symbol} are equivalent. Note---there is one exception to
322 this rule. Multiple letter options that start with a lower case 'o' can
323 only be preceded by two dashes. This is to reduce confusion with the
324 @samp{-o} option. So for example @samp{-omagic} sets the output file
325 name to @samp{magic} whereas @samp{--omagic} sets the NMAGIC flag on the
328 Arguments to multiple-letter options must either be separated from the
329 option name by an equals sign, or be given as separate arguments
330 immediately following the option that requires them. For example,
331 @samp{--trace-symbol foo} and @samp{--trace-symbol=foo} are equivalent.
332 Unique abbreviations of the names of multiple-letter options are
335 Note---if the linker is being invoked indirectly, via a compiler driver
336 (e.g. @samp{gcc}) then all the linker command-line options should be
337 prefixed by @samp{-Wl,} (or whatever is appropriate for the particular
338 compiler driver) like this:
341 gcc -Wl,--start-group foo.o bar.o -Wl,--end-group
344 This is important, because otherwise the compiler driver program may
345 silently drop the linker options, resulting in a bad link. Confusion
346 may also arise when passing options that require values through a
347 driver, as the use of a space between option and argument acts as
348 a separator, and causes the driver to pass only the option to the linker
349 and the argument to the compiler. In this case, it is simplest to use
350 the joined forms of both single- and multiple-letter options, such as:
353 gcc foo.o bar.o -Wl,-eENTRY -Wl,-Map=a.map
356 Here is a table of the generic command-line switches accepted by the GNU
360 @include at-file.texi
362 @kindex -a @var{keyword}
363 @item -a @var{keyword}
364 This option is supported for HP/UX compatibility. The @var{keyword}
365 argument must be one of the strings @samp{archive}, @samp{shared}, or
366 @samp{default}. @samp{-aarchive} is functionally equivalent to
367 @samp{-Bstatic}, and the other two keywords are functionally equivalent
368 to @samp{-Bdynamic}. This option may be used any number of times.
370 @kindex --audit @var{AUDITLIB}
371 @item --audit @var{AUDITLIB}
372 Adds @var{AUDITLIB} to the @code{DT_AUDIT} entry of the dynamic section.
373 @var{AUDITLIB} is not checked for existence, nor will it use the DT_SONAME
374 specified in the library. If specified multiple times @code{DT_AUDIT}
375 will contain a colon separated list of audit interfaces to use. If the linker
376 finds an object with an audit entry while searching for shared libraries,
377 it will add a corresponding @code{DT_DEPAUDIT} entry in the output file.
378 This option is only meaningful on ELF platforms supporting the rtld-audit
381 @ifclear SingleFormat
382 @cindex binary input format
383 @kindex -b @var{format}
384 @kindex --format=@var{format}
387 @item -b @var{input-format}
388 @itemx --format=@var{input-format}
389 @command{ld} may be configured to support more than one kind of object
390 file. If your @command{ld} is configured this way, you can use the
391 @samp{-b} option to specify the binary format for input object files
392 that follow this option on the command line. Even when @command{ld} is
393 configured to support alternative object formats, you don't usually need
394 to specify this, as @command{ld} should be configured to expect as a
395 default input format the most usual format on each machine.
396 @var{input-format} is a text string, the name of a particular format
397 supported by the BFD libraries. (You can list the available binary
398 formats with @samp{objdump -i}.)
401 You may want to use this option if you are linking files with an unusual
402 binary format. You can also use @samp{-b} to switch formats explicitly (when
403 linking object files of different formats), by including
404 @samp{-b @var{input-format}} before each group of object files in a
407 The default format is taken from the environment variable
412 You can also define the input format from a script, using the command
415 see @ref{Format Commands}.
419 @kindex -c @var{MRI-cmdfile}
420 @kindex --mri-script=@var{MRI-cmdfile}
421 @cindex compatibility, MRI
422 @item -c @var{MRI-commandfile}
423 @itemx --mri-script=@var{MRI-commandfile}
424 For compatibility with linkers produced by MRI, @command{ld} accepts script
425 files written in an alternate, restricted command language, described in
427 @ref{MRI,,MRI Compatible Script Files}.
430 the MRI Compatible Script Files section of GNU ld documentation.
432 Introduce MRI script files with
433 the option @samp{-c}; use the @samp{-T} option to run linker
434 scripts written in the general-purpose @command{ld} scripting language.
435 If @var{MRI-cmdfile} does not exist, @command{ld} looks for it in the directories
436 specified by any @samp{-L} options.
438 @cindex common allocation
445 These three options are equivalent; multiple forms are supported for
446 compatibility with other linkers. They assign space to common symbols
447 even if a relocatable output file is specified (with @samp{-r}). The
448 script command @code{FORCE_COMMON_ALLOCATION} has the same effect.
449 @xref{Miscellaneous Commands}.
451 @kindex --depaudit @var{AUDITLIB}
452 @kindex -P @var{AUDITLIB}
453 @item --depaudit @var{AUDITLIB}
454 @itemx -P @var{AUDITLIB}
455 Adds @var{AUDITLIB} to the @code{DT_DEPAUDIT} entry of the dynamic section.
456 @var{AUDITLIB} is not checked for existence, nor will it use the DT_SONAME
457 specified in the library. If specified multiple times @code{DT_DEPAUDIT}
458 will contain a colon separated list of audit interfaces to use. This
459 option is only meaningful on ELF platforms supporting the rtld-audit interface.
460 The -P option is provided for Solaris compatibility.
462 @cindex entry point, from command line
463 @kindex -e @var{entry}
464 @kindex --entry=@var{entry}
466 @itemx --entry=@var{entry}
467 Use @var{entry} as the explicit symbol for beginning execution of your
468 program, rather than the default entry point. If there is no symbol
469 named @var{entry}, the linker will try to parse @var{entry} as a number,
470 and use that as the entry address (the number will be interpreted in
471 base 10; you may use a leading @samp{0x} for base 16, or a leading
472 @samp{0} for base 8). @xref{Entry Point}, for a discussion of defaults
473 and other ways of specifying the entry point.
475 @kindex --exclude-libs
476 @item --exclude-libs @var{lib},@var{lib},...
477 Specifies a list of archive libraries from which symbols should not be automatically
478 exported. The library names may be delimited by commas or colons. Specifying
479 @code{--exclude-libs ALL} excludes symbols in all archive libraries from
480 automatic export. This option is available only for the i386 PE targeted
481 port of the linker and for ELF targeted ports. For i386 PE, symbols
482 explicitly listed in a .def file are still exported, regardless of this
483 option. For ELF targeted ports, symbols affected by this option will
484 be treated as hidden.
486 @kindex --exclude-modules-for-implib
487 @item --exclude-modules-for-implib @var{module},@var{module},...
488 Specifies a list of object files or archive members, from which symbols
489 should not be automatically exported, but which should be copied wholesale
490 into the import library being generated during the link. The module names
491 may be delimited by commas or colons, and must match exactly the filenames
492 used by @command{ld} to open the files; for archive members, this is simply
493 the member name, but for object files the name listed must include and
494 match precisely any path used to specify the input file on the linker's
495 command-line. This option is available only for the i386 PE targeted port
496 of the linker. Symbols explicitly listed in a .def file are still exported,
497 regardless of this option.
499 @cindex dynamic symbol table
501 @kindex --export-dynamic
502 @kindex --no-export-dynamic
504 @itemx --export-dynamic
505 @itemx --no-export-dynamic
506 When creating a dynamically linked executable, using the @option{-E}
507 option or the @option{--export-dynamic} option causes the linker to add
508 all symbols to the dynamic symbol table. The dynamic symbol table is the
509 set of symbols which are visible from dynamic objects at run time.
511 If you do not use either of these options (or use the
512 @option{--no-export-dynamic} option to restore the default behavior), the
513 dynamic symbol table will normally contain only those symbols which are
514 referenced by some dynamic object mentioned in the link.
516 If you use @code{dlopen} to load a dynamic object which needs to refer
517 back to the symbols defined by the program, rather than some other
518 dynamic object, then you will probably need to use this option when
519 linking the program itself.
521 You can also use the dynamic list to control what symbols should
522 be added to the dynamic symbol table if the output format supports it.
523 See the description of @samp{--dynamic-list}.
525 Note that this option is specific to ELF targeted ports. PE targets
526 support a similar function to export all symbols from a DLL or EXE; see
527 the description of @samp{--export-all-symbols} below.
529 @ifclear SingleFormat
530 @cindex big-endian objects
534 Link big-endian objects. This affects the default output format.
536 @cindex little-endian objects
539 Link little-endian objects. This affects the default output format.
542 @kindex -f @var{name}
543 @kindex --auxiliary=@var{name}
545 @itemx --auxiliary=@var{name}
546 When creating an ELF shared object, set the internal DT_AUXILIARY field
547 to the specified name. This tells the dynamic linker that the symbol
548 table of the shared object should be used as an auxiliary filter on the
549 symbol table of the shared object @var{name}.
551 If you later link a program against this filter object, then, when you
552 run the program, the dynamic linker will see the DT_AUXILIARY field. If
553 the dynamic linker resolves any symbols from the filter object, it will
554 first check whether there is a definition in the shared object
555 @var{name}. If there is one, it will be used instead of the definition
556 in the filter object. The shared object @var{name} need not exist.
557 Thus the shared object @var{name} may be used to provide an alternative
558 implementation of certain functions, perhaps for debugging or for
559 machine-specific performance.
561 This option may be specified more than once. The DT_AUXILIARY entries
562 will be created in the order in which they appear on the command line.
564 @kindex -F @var{name}
565 @kindex --filter=@var{name}
567 @itemx --filter=@var{name}
568 When creating an ELF shared object, set the internal DT_FILTER field to
569 the specified name. This tells the dynamic linker that the symbol table
570 of the shared object which is being created should be used as a filter
571 on the symbol table of the shared object @var{name}.
573 If you later link a program against this filter object, then, when you
574 run the program, the dynamic linker will see the DT_FILTER field. The
575 dynamic linker will resolve symbols according to the symbol table of the
576 filter object as usual, but it will actually link to the definitions
577 found in the shared object @var{name}. Thus the filter object can be
578 used to select a subset of the symbols provided by the object
581 Some older linkers used the @option{-F} option throughout a compilation
582 toolchain for specifying object-file format for both input and output
584 @ifclear SingleFormat
585 The @sc{gnu} linker uses other mechanisms for this purpose: the
586 @option{-b}, @option{--format}, @option{--oformat} options, the
587 @code{TARGET} command in linker scripts, and the @code{GNUTARGET}
588 environment variable.
590 The @sc{gnu} linker will ignore the @option{-F} option when not
591 creating an ELF shared object.
593 @cindex finalization function
594 @kindex -fini=@var{name}
595 @item -fini=@var{name}
596 When creating an ELF executable or shared object, call NAME when the
597 executable or shared object is unloaded, by setting DT_FINI to the
598 address of the function. By default, the linker uses @code{_fini} as
599 the function to call.
603 Ignored. Provided for compatibility with other tools.
605 @kindex -G @var{value}
606 @kindex --gpsize=@var{value}
609 @itemx --gpsize=@var{value}
610 Set the maximum size of objects to be optimized using the GP register to
611 @var{size}. This is only meaningful for object file formats such as
612 MIPS ELF that support putting large and small objects into different
613 sections. This is ignored for other object file formats.
615 @cindex runtime library name
616 @kindex -h @var{name}
617 @kindex -soname=@var{name}
619 @itemx -soname=@var{name}
620 When creating an ELF shared object, set the internal DT_SONAME field to
621 the specified name. When an executable is linked with a shared object
622 which has a DT_SONAME field, then when the executable is run the dynamic
623 linker will attempt to load the shared object specified by the DT_SONAME
624 field rather than the using the file name given to the linker.
627 @cindex incremental link
629 Perform an incremental link (same as option @samp{-r}).
631 @cindex initialization function
632 @kindex -init=@var{name}
633 @item -init=@var{name}
634 When creating an ELF executable or shared object, call NAME when the
635 executable or shared object is loaded, by setting DT_INIT to the address
636 of the function. By default, the linker uses @code{_init} as the
639 @cindex archive files, from cmd line
640 @kindex -l @var{namespec}
641 @kindex --library=@var{namespec}
642 @item -l @var{namespec}
643 @itemx --library=@var{namespec}
644 Add the archive or object file specified by @var{namespec} to the
645 list of files to link. This option may be used any number of times.
646 If @var{namespec} is of the form @file{:@var{filename}}, @command{ld}
647 will search the library path for a file called @var{filename}, otherwise it
648 will search the library path for a file called @file{lib@var{namespec}.a}.
650 On systems which support shared libraries, @command{ld} may also search for
651 files other than @file{lib@var{namespec}.a}. Specifically, on ELF
652 and SunOS systems, @command{ld} will search a directory for a library
653 called @file{lib@var{namespec}.so} before searching for one called
654 @file{lib@var{namespec}.a}. (By convention, a @code{.so} extension
655 indicates a shared library.) Note that this behavior does not apply
656 to @file{:@var{filename}}, which always specifies a file called
659 The linker will search an archive only once, at the location where it is
660 specified on the command line. If the archive defines a symbol which
661 was undefined in some object which appeared before the archive on the
662 command line, the linker will include the appropriate file(s) from the
663 archive. However, an undefined symbol in an object appearing later on
664 the command line will not cause the linker to search the archive again.
666 See the @option{-(} option for a way to force the linker to search
667 archives multiple times.
669 You may list the same archive multiple times on the command line.
672 This type of archive searching is standard for Unix linkers. However,
673 if you are using @command{ld} on AIX, note that it is different from the
674 behaviour of the AIX linker.
677 @cindex search directory, from cmd line
679 @kindex --library-path=@var{dir}
680 @item -L @var{searchdir}
681 @itemx --library-path=@var{searchdir}
682 Add path @var{searchdir} to the list of paths that @command{ld} will search
683 for archive libraries and @command{ld} control scripts. You may use this
684 option any number of times. The directories are searched in the order
685 in which they are specified on the command line. Directories specified
686 on the command line are searched before the default directories. All
687 @option{-L} options apply to all @option{-l} options, regardless of the
688 order in which the options appear. @option{-L} options do not affect
689 how @command{ld} searches for a linker script unless @option{-T}
692 If @var{searchdir} begins with @code{=} or @code{$SYSROOT}, then this
693 prefix will be replaced by the @dfn{sysroot prefix}, controlled by the
694 @samp{--sysroot} option, or specified when the linker is configured.
697 The default set of paths searched (without being specified with
698 @samp{-L}) depends on which emulation mode @command{ld} is using, and in
699 some cases also on how it was configured. @xref{Environment}.
702 The paths can also be specified in a link script with the
703 @code{SEARCH_DIR} command. Directories specified this way are searched
704 at the point in which the linker script appears in the command line.
707 @kindex -m @var{emulation}
708 @item -m @var{emulation}
709 Emulate the @var{emulation} linker. You can list the available
710 emulations with the @samp{--verbose} or @samp{-V} options.
712 If the @samp{-m} option is not used, the emulation is taken from the
713 @code{LDEMULATION} environment variable, if that is defined.
715 Otherwise, the default emulation depends upon how the linker was
723 Print a link map to the standard output. A link map provides
724 information about the link, including the following:
728 Where object files are mapped into memory.
730 How common symbols are allocated.
732 All archive members included in the link, with a mention of the symbol
733 which caused the archive member to be brought in.
735 The values assigned to symbols.
737 Note - symbols whose values are computed by an expression which
738 involves a reference to a previous value of the same symbol may not
739 have correct result displayed in the link map. This is because the
740 linker discards intermediate results and only retains the final value
741 of an expression. Under such circumstances the linker will display
742 the final value enclosed by square brackets. Thus for example a
743 linker script containing:
751 will produce the following output in the link map if the @option{-M}
756 [0x0000000c] foo = (foo * 0x4)
757 [0x0000000c] foo = (foo + 0x8)
760 See @ref{Expressions} for more information about expressions in linker
764 How GNU properties are merged.
766 When the linker merges input .note.gnu.property sections into one output
767 .note.gnu.property section, some properties are removed or updated.
768 These actions are reported in the link map. For example:
771 Removed property 0xc0000002 to merge foo.o (0x1) and bar.o (not found)
774 This indicates that property 0xc0000002 is removed from output when
775 merging properties in @file{foo.o}, whose property 0xc0000002 value
776 is 0x1, and @file{bar.o}, which doesn't have property 0xc0000002.
779 Updated property 0xc0010001 (0x1) to merge foo.o (0x1) and bar.o (0x1)
782 This indicates that property 0xc0010001 value is updated to 0x1 in output
783 when merging properties in @file{foo.o}, whose 0xc0010001 property value
784 is 0x1, and @file{bar.o}, whose 0xc0010001 property value is 0x1.
787 @cindex link map discarded
788 @kindex --print-map-discarded
789 @kindex --no-print-map-discarded
790 @item --print-map-discarded
791 @itemx --no-print-map-discarded
792 Print (or do not print) the list of discarded and garbage collected sections
793 in the link map. Enabled by default.
796 @cindex read-only text
801 Turn off page alignment of sections, and disable linking against shared
802 libraries. If the output format supports Unix style magic numbers,
803 mark the output as @code{NMAGIC}.
807 @cindex read/write from cmd line
811 Set the text and data sections to be readable and writable. Also, do
812 not page-align the data segment, and disable linking against shared
813 libraries. If the output format supports Unix style magic numbers,
814 mark the output as @code{OMAGIC}. Note: Although a writable text section
815 is allowed for PE-COFF targets, it does not conform to the format
816 specification published by Microsoft.
821 This option negates most of the effects of the @option{-N} option. It
822 sets the text section to be read-only, and forces the data segment to
823 be page-aligned. Note - this option does not enable linking against
824 shared libraries. Use @option{-Bdynamic} for this.
826 @kindex -o @var{output}
827 @kindex --output=@var{output}
828 @cindex naming the output file
829 @item -o @var{output}
830 @itemx --output=@var{output}
831 Use @var{output} as the name for the program produced by @command{ld}; if this
832 option is not specified, the name @file{a.out} is used by default. The
833 script command @code{OUTPUT} can also specify the output file name.
835 @kindex -O @var{level}
836 @cindex generating optimized output
838 If @var{level} is a numeric values greater than zero @command{ld} optimizes
839 the output. This might take significantly longer and therefore probably
840 should only be enabled for the final binary. At the moment this
841 option only affects ELF shared library generation. Future releases of
842 the linker may make more use of this option. Also currently there is
843 no difference in the linker's behaviour for different non-zero values
844 of this option. Again this may change with future releases.
846 @kindex -plugin @var{name}
847 @item -plugin @var{name}
848 Involve a plugin in the linking process. The @var{name} parameter is
849 the absolute filename of the plugin. Usually this parameter is
850 automatically added by the complier, when using link time
851 optimization, but users can also add their own plugins if they so
854 Note that the location of the compiler originated plugins is different
855 from the place where the @command{ar}, @command{nm} and
856 @command{ranlib} programs search for their plugins. In order for
857 those commands to make use of a compiler based plugin it must first be
858 copied into the @file{$@{libdir@}/bfd-plugins} directory. All gcc
859 based linker plugins are backward compatible, so it is sufficient to
860 just copy in the newest one.
863 @cindex push state governing input file handling
865 The @option{--push-state} allows to preserve the current state of the
866 flags which govern the input file handling so that they can all be
867 restored with one corresponding @option{--pop-state} option.
869 The option which are covered are: @option{-Bdynamic}, @option{-Bstatic},
870 @option{-dn}, @option{-dy}, @option{-call_shared}, @option{-non_shared},
871 @option{-static}, @option{-N}, @option{-n}, @option{--whole-archive},
872 @option{--no-whole-archive}, @option{-r}, @option{-Ur},
873 @option{--copy-dt-needed-entries}, @option{--no-copy-dt-needed-entries},
874 @option{--as-needed}, @option{--no-as-needed}, and @option{-a}.
876 One target for this option are specifications for @file{pkg-config}. When
877 used with the @option{--libs} option all possibly needed libraries are
878 listed and then possibly linked with all the time. It is better to return
879 something as follows:
882 -Wl,--push-state,--as-needed -libone -libtwo -Wl,--pop-state
886 @cindex pop state governing input file handling
888 Undoes the effect of --push-state, restores the previous values of the
889 flags governing input file handling.
892 @kindex --emit-relocs
893 @cindex retain relocations in final executable
896 Leave relocation sections and contents in fully linked executables.
897 Post link analysis and optimization tools may need this information in
898 order to perform correct modifications of executables. This results
899 in larger executables.
901 This option is currently only supported on ELF platforms.
903 @kindex --force-dynamic
904 @cindex forcing the creation of dynamic sections
905 @item --force-dynamic
906 Force the output file to have dynamic sections. This option is specific
910 @cindex relocatable output
912 @kindex --relocatable
915 Generate relocatable output---i.e., generate an output file that can in
916 turn serve as input to @command{ld}. This is often called @dfn{partial
917 linking}. As a side effect, in environments that support standard Unix
918 magic numbers, this option also sets the output file's magic number to
920 @c ; see @option{-N}.
921 If this option is not specified, an absolute file is produced. When
922 linking C++ programs, this option @emph{will not} resolve references to
923 constructors; to do that, use @samp{-Ur}.
925 When an input file does not have the same format as the output file,
926 partial linking is only supported if that input file does not contain any
927 relocations. Different output formats can have further restrictions; for
928 example some @code{a.out}-based formats do not support partial linking
929 with input files in other formats at all.
931 This option does the same thing as @samp{-i}.
933 @kindex -R @var{file}
934 @kindex --just-symbols=@var{file}
935 @cindex symbol-only input
936 @item -R @var{filename}
937 @itemx --just-symbols=@var{filename}
938 Read symbol names and their addresses from @var{filename}, but do not
939 relocate it or include it in the output. This allows your output file
940 to refer symbolically to absolute locations of memory defined in other
941 programs. You may use this option more than once.
943 For compatibility with other ELF linkers, if the @option{-R} option is
944 followed by a directory name, rather than a file name, it is treated as
945 the @option{-rpath} option.
949 @cindex strip all symbols
952 Omit all symbol information from the output file.
955 @kindex --strip-debug
956 @cindex strip debugger symbols
959 Omit debugger symbol information (but not all symbols) from the output file.
961 @kindex --strip-discarded
962 @kindex --no-strip-discarded
963 @item --strip-discarded
964 @itemx --no-strip-discarded
965 Omit (or do not omit) global symbols defined in discarded sections.
970 @cindex input files, displaying
973 Print the names of the input files as @command{ld} processes them. If
974 @samp{-t} is given twice then members within archives are also printed.
975 @samp{-t} output is useful to generate a list of all the object files
976 and scripts involved in linking, for example, when packaging files for
979 @kindex -T @var{script}
980 @kindex --script=@var{script}
982 @item -T @var{scriptfile}
983 @itemx --script=@var{scriptfile}
984 Use @var{scriptfile} as the linker script. This script replaces
985 @command{ld}'s default linker script (rather than adding to it), so
986 @var{commandfile} must specify everything necessary to describe the
987 output file. @xref{Scripts}. If @var{scriptfile} does not exist in
988 the current directory, @code{ld} looks for it in the directories
989 specified by any preceding @samp{-L} options. Multiple @samp{-T}
992 @kindex -dT @var{script}
993 @kindex --default-script=@var{script}
995 @item -dT @var{scriptfile}
996 @itemx --default-script=@var{scriptfile}
997 Use @var{scriptfile} as the default linker script. @xref{Scripts}.
999 This option is similar to the @option{--script} option except that
1000 processing of the script is delayed until after the rest of the
1001 command line has been processed. This allows options placed after the
1002 @option{--default-script} option on the command line to affect the
1003 behaviour of the linker script, which can be important when the linker
1004 command line cannot be directly controlled by the user. (eg because
1005 the command line is being constructed by another tool, such as
1008 @kindex -u @var{symbol}
1009 @kindex --undefined=@var{symbol}
1010 @cindex undefined symbol
1011 @item -u @var{symbol}
1012 @itemx --undefined=@var{symbol}
1013 Force @var{symbol} to be entered in the output file as an undefined
1014 symbol. Doing this may, for example, trigger linking of additional
1015 modules from standard libraries. @samp{-u} may be repeated with
1016 different option arguments to enter additional undefined symbols. This
1017 option is equivalent to the @code{EXTERN} linker script command.
1019 If this option is being used to force additional modules to be pulled
1020 into the link, and if it is an error for the symbol to remain
1021 undefined, then the option @option{--require-defined} should be used
1024 @kindex --require-defined=@var{symbol}
1025 @cindex symbols, require defined
1026 @cindex defined symbol
1027 @item --require-defined=@var{symbol}
1028 Require that @var{symbol} is defined in the output file. This option
1029 is the same as option @option{--undefined} except that if @var{symbol}
1030 is not defined in the output file then the linker will issue an error
1031 and exit. The same effect can be achieved in a linker script by using
1032 @code{EXTERN}, @code{ASSERT} and @code{DEFINED} together. This option
1033 can be used multiple times to require additional symbols.
1036 @cindex constructors
1038 For anything other than C++ programs, this option is equivalent to
1039 @samp{-r}: it generates relocatable output---i.e., an output file that can in
1040 turn serve as input to @command{ld}. When linking C++ programs, @samp{-Ur}
1041 @emph{does} resolve references to constructors, unlike @samp{-r}.
1042 It does not work to use @samp{-Ur} on files that were themselves linked
1043 with @samp{-Ur}; once the constructor table has been built, it cannot
1044 be added to. Use @samp{-Ur} only for the last partial link, and
1045 @samp{-r} for the others.
1047 @kindex --orphan-handling=@var{MODE}
1048 @cindex orphan sections
1049 @cindex sections, orphan
1050 @item --orphan-handling=@var{MODE}
1051 Control how orphan sections are handled. An orphan section is one not
1052 specifically mentioned in a linker script. @xref{Orphan Sections}.
1054 @var{MODE} can have any of the following values:
1058 Orphan sections are placed into a suitable output section following
1059 the strategy described in @ref{Orphan Sections}. The option
1060 @samp{--unique} also affects how sections are placed.
1063 All orphan sections are discarded, by placing them in the
1064 @samp{/DISCARD/} section (@pxref{Output Section Discarding}).
1067 The linker will place the orphan section as for @code{place} and also
1071 The linker will exit with an error if any orphan section is found.
1074 The default if @samp{--orphan-handling} is not given is @code{place}.
1076 @kindex --unique[=@var{SECTION}]
1077 @item --unique[=@var{SECTION}]
1078 Creates a separate output section for every input section matching
1079 @var{SECTION}, or if the optional wildcard @var{SECTION} argument is
1080 missing, for every orphan input section. An orphan section is one not
1081 specifically mentioned in a linker script. You may use this option
1082 multiple times on the command line; It prevents the normal merging of
1083 input sections with the same name, overriding output section assignments
1093 Display the version number for @command{ld}. The @option{-V} option also
1094 lists the supported emulations.
1097 @kindex --discard-all
1098 @cindex deleting local symbols
1100 @itemx --discard-all
1101 Delete all local symbols.
1104 @kindex --discard-locals
1105 @cindex local symbols, deleting
1107 @itemx --discard-locals
1108 Delete all temporary local symbols. (These symbols start with
1109 system-specific local label prefixes, typically @samp{.L} for ELF systems
1110 or @samp{L} for traditional a.out systems.)
1112 @kindex -y @var{symbol}
1113 @kindex --trace-symbol=@var{symbol}
1114 @cindex symbol tracing
1115 @item -y @var{symbol}
1116 @itemx --trace-symbol=@var{symbol}
1117 Print the name of each linked file in which @var{symbol} appears. This
1118 option may be given any number of times. On many systems it is necessary
1119 to prepend an underscore.
1121 This option is useful when you have an undefined symbol in your link but
1122 don't know where the reference is coming from.
1124 @kindex -Y @var{path}
1126 Add @var{path} to the default library search path. This option exists
1127 for Solaris compatibility.
1129 @kindex -z @var{keyword}
1130 @item -z @var{keyword}
1131 The recognized keywords are:
1135 Always generate BND prefix in PLT entries. Supported for Linux/x86_64.
1137 @item call-nop=prefix-addr
1138 @itemx call-nop=suffix-nop
1139 @itemx call-nop=prefix-@var{byte}
1140 @itemx call-nop=suffix-@var{byte}
1141 Specify the 1-byte @code{NOP} padding when transforming indirect call
1142 to a locally defined function, foo, via its GOT slot.
1143 @option{call-nop=prefix-addr} generates @code{0x67 call foo}.
1144 @option{call-nop=suffix-nop} generates @code{call foo 0x90}.
1145 @option{call-nop=prefix-@var{byte}} generates @code{@var{byte} call foo}.
1146 @option{call-nop=suffix-@var{byte}} generates @code{call foo @var{byte}}.
1147 Supported for i386 and x86_64.
1149 @item cet-report=none
1150 @itemx cet-report=warning
1151 @itemx cet-report=error
1152 Specify how to report the missing GNU_PROPERTY_X86_FEATURE_1_IBT and
1153 GNU_PROPERTY_X86_FEATURE_1_SHSTK properties in input .note.gnu.property
1154 section. @option{cet-report=none}, which is the default, will make the
1155 linker not report missing properties in input files.
1156 @option{cet-report=warning} will make the linker issue a warning for
1157 missing properties in input files. @option{cet-report=error} will make
1158 the linker issue an error for missing properties in input files.
1159 Note that @option{ibt} will turn off the missing
1160 GNU_PROPERTY_X86_FEATURE_1_IBT property report and @option{shstk} will
1161 turn off the missing GNU_PROPERTY_X86_FEATURE_1_SHSTK property report.
1162 Supported for Linux/i386 and Linux/x86_64.
1166 Combine multiple dynamic relocation sections and sort to improve
1167 dynamic symbol lookup caching. Do not do this if @samp{nocombreloc}.
1171 Generate common symbols with STT_COMMON type during a relocatable
1172 link. Use STT_OBJECT type if @samp{nocommon}.
1174 @item common-page-size=@var{value}
1175 Set the page size most commonly used to @var{value}. Memory image
1176 layout will be optimized to minimize memory pages if the system is
1177 using pages of this size.
1180 Report unresolved symbol references from regular object files. This
1181 is done even if the linker is creating a non-symbolic shared library.
1182 This option is the inverse of @samp{-z undefs}.
1184 @item dynamic-undefined-weak
1185 @itemx nodynamic-undefined-weak
1186 Make undefined weak symbols dynamic when building a dynamic object,
1187 if they are referenced from a regular object file and not forced local
1188 by symbol visibility or versioning. Do not make them dynamic if
1189 @samp{nodynamic-undefined-weak}. If neither option is given, a target
1190 may default to either option being in force, or make some other
1191 selection of undefined weak symbols dynamic. Not all targets support
1195 Marks the object as requiring executable stack.
1198 This option is only meaningful when building a shared object. It makes
1199 the symbols defined by this shared object available for symbol resolution
1200 of subsequently loaded libraries.
1203 This option is only meaningful when building a dynamic executable.
1204 This option marks the executable as requiring global auditing by
1205 setting the @code{DF_1_GLOBAUDIT} bit in the @code{DT_FLAGS_1} dynamic
1206 tag. Global auditing requires that any auditing library defined via
1207 the @option{--depaudit} or @option{-P} command-line options be run for
1208 all dynamic objects loaded by the application.
1211 Generate Intel Indirect Branch Tracking (IBT) enabled PLT entries.
1212 Supported for Linux/i386 and Linux/x86_64.
1215 Generate GNU_PROPERTY_X86_FEATURE_1_IBT in .note.gnu.property section
1216 to indicate compatibility with IBT. This also implies @option{ibtplt}.
1217 Supported for Linux/i386 and Linux/x86_64.
1220 This option is only meaningful when building a shared object.
1221 It marks the object so that its runtime initialization will occur
1222 before the runtime initialization of any other objects brought into
1223 the process at the same time. Similarly the runtime finalization of
1224 the object will occur after the runtime finalization of any other
1228 Specify that the dynamic loader should modify its symbol search order
1229 so that symbols in this shared library interpose all other shared
1230 libraries not so marked.
1233 When generating an executable or shared library, mark it to tell the
1234 dynamic linker to defer function call resolution to the point when
1235 the function is called (lazy binding), rather than at load time.
1236 Lazy binding is the default.
1239 Specify that the object's filters be processed immediately at runtime.
1241 @item max-page-size=@var{value}
1242 Set the maximum memory page size supported to @var{value}.
1245 Allow multiple definitions.
1248 Disable linker generated .dynbss variables used in place of variables
1249 defined in shared libraries. May result in dynamic text relocations.
1252 Specify that the dynamic loader search for dependencies of this object
1253 should ignore any default library search paths.
1256 Specify that the object shouldn't be unloaded at runtime.
1259 Specify that the object is not available to @code{dlopen}.
1262 Specify that the object can not be dumped by @code{dldump}.
1265 Marks the object as not requiring executable stack.
1267 @item noextern-protected-data
1268 Don't treat protected data symbols as external when building a shared
1269 library. This option overrides the linker backend default. It can be
1270 used to work around incorrect relocations against protected data symbols
1271 generated by compiler. Updates on protected data symbols by another
1272 module aren't visible to the resulting shared library. Supported for
1275 @item noreloc-overflow
1276 Disable relocation overflow check. This can be used to disable
1277 relocation overflow check if there will be no dynamic relocation
1278 overflow at run-time. Supported for x86_64.
1281 When generating an executable or shared library, mark it to tell the
1282 dynamic linker to resolve all symbols when the program is started, or
1283 when the shared library is loaded by dlopen, instead of deferring
1284 function call resolution to the point when the function is first
1288 Specify that the object requires @samp{$ORIGIN} handling in paths.
1292 Create an ELF @code{PT_GNU_RELRO} segment header in the object. This
1293 specifies a memory segment that should be made read-only after
1294 relocation, if supported. Specifying @samp{common-page-size} smaller
1295 than the system page size will render this protection ineffective.
1296 Don't create an ELF @code{PT_GNU_RELRO} segment if @samp{norelro}.
1299 @itemx noseparate-code
1300 Create separate code @code{PT_LOAD} segment header in the object. This
1301 specifies a memory segment that should contain only instructions and must
1302 be in wholly disjoint pages from any other data. Don't create separate
1303 code @code{PT_LOAD} segment if @samp{noseparate-code} is used.
1306 Generate GNU_PROPERTY_X86_FEATURE_1_SHSTK in .note.gnu.property section
1307 to indicate compatibility with Intel Shadow Stack. Supported for
1308 Linux/i386 and Linux/x86_64.
1310 @item stack-size=@var{value}
1311 Specify a stack size for an ELF @code{PT_GNU_STACK} segment.
1312 Specifying zero will override any default non-zero sized
1313 @code{PT_GNU_STACK} segment creation.
1318 Report an error if DT_TEXTREL is set, i.e., if the binary has dynamic
1319 relocations in read-only sections. Don't report an error if
1320 @samp{notext} or @samp{textoff}.
1323 Do not report unresolved symbol references from regular object files,
1324 either when creating an executable, or when creating a shared library.
1325 This option is the inverse of @samp{-z defs}.
1329 Other keywords are ignored for Solaris compatibility.
1332 @cindex groups of archives
1333 @item -( @var{archives} -)
1334 @itemx --start-group @var{archives} --end-group
1335 The @var{archives} should be a list of archive files. They may be
1336 either explicit file names, or @samp{-l} options.
1338 The specified archives are searched repeatedly until no new undefined
1339 references are created. Normally, an archive is searched only once in
1340 the order that it is specified on the command line. If a symbol in that
1341 archive is needed to resolve an undefined symbol referred to by an
1342 object in an archive that appears later on the command line, the linker
1343 would not be able to resolve that reference. By grouping the archives,
1344 they will all be searched repeatedly until all possible references are
1347 Using this option has a significant performance cost. It is best to use
1348 it only when there are unavoidable circular references between two or
1351 @kindex --accept-unknown-input-arch
1352 @kindex --no-accept-unknown-input-arch
1353 @item --accept-unknown-input-arch
1354 @itemx --no-accept-unknown-input-arch
1355 Tells the linker to accept input files whose architecture cannot be
1356 recognised. The assumption is that the user knows what they are doing
1357 and deliberately wants to link in these unknown input files. This was
1358 the default behaviour of the linker, before release 2.14. The default
1359 behaviour from release 2.14 onwards is to reject such input files, and
1360 so the @samp{--accept-unknown-input-arch} option has been added to
1361 restore the old behaviour.
1364 @kindex --no-as-needed
1366 @itemx --no-as-needed
1367 This option affects ELF DT_NEEDED tags for dynamic libraries mentioned
1368 on the command line after the @option{--as-needed} option. Normally
1369 the linker will add a DT_NEEDED tag for each dynamic library mentioned
1370 on the command line, regardless of whether the library is actually
1371 needed or not. @option{--as-needed} causes a DT_NEEDED tag to only be
1372 emitted for a library that @emph{at that point in the link} satisfies a
1373 non-weak undefined symbol reference from a regular object file or, if
1374 the library is not found in the DT_NEEDED lists of other needed libraries, a
1375 non-weak undefined symbol reference from another needed dynamic library.
1376 Object files or libraries appearing on the command line @emph{after}
1377 the library in question do not affect whether the library is seen as
1378 needed. This is similar to the rules for extraction of object files
1379 from archives. @option{--no-as-needed} restores the default behaviour.
1381 @kindex --add-needed
1382 @kindex --no-add-needed
1384 @itemx --no-add-needed
1385 These two options have been deprecated because of the similarity of
1386 their names to the @option{--as-needed} and @option{--no-as-needed}
1387 options. They have been replaced by @option{--copy-dt-needed-entries}
1388 and @option{--no-copy-dt-needed-entries}.
1390 @kindex -assert @var{keyword}
1391 @item -assert @var{keyword}
1392 This option is ignored for SunOS compatibility.
1396 @kindex -call_shared
1400 Link against dynamic libraries. This is only meaningful on platforms
1401 for which shared libraries are supported. This option is normally the
1402 default on such platforms. The different variants of this option are
1403 for compatibility with various systems. You may use this option
1404 multiple times on the command line: it affects library searching for
1405 @option{-l} options which follow it.
1409 Set the @code{DF_1_GROUP} flag in the @code{DT_FLAGS_1} entry in the dynamic
1410 section. This causes the runtime linker to handle lookups in this
1411 object and its dependencies to be performed only inside the group.
1412 @option{--unresolved-symbols=report-all} is implied. This option is
1413 only meaningful on ELF platforms which support shared libraries.
1423 Do not link against shared libraries. This is only meaningful on
1424 platforms for which shared libraries are supported. The different
1425 variants of this option are for compatibility with various systems. You
1426 may use this option multiple times on the command line: it affects
1427 library searching for @option{-l} options which follow it. This
1428 option also implies @option{--unresolved-symbols=report-all}. This
1429 option can be used with @option{-shared}. Doing so means that a
1430 shared library is being created but that all of the library's external
1431 references must be resolved by pulling in entries from static
1436 When creating a shared library, bind references to global symbols to the
1437 definition within the shared library, if any. Normally, it is possible
1438 for a program linked against a shared library to override the definition
1439 within the shared library. This option is only meaningful on ELF
1440 platforms which support shared libraries.
1442 @kindex -Bsymbolic-functions
1443 @item -Bsymbolic-functions
1444 When creating a shared library, bind references to global function
1445 symbols to the definition within the shared library, if any.
1446 This option is only meaningful on ELF platforms which support shared
1449 @kindex --dynamic-list=@var{dynamic-list-file}
1450 @item --dynamic-list=@var{dynamic-list-file}
1451 Specify the name of a dynamic list file to the linker. This is
1452 typically used when creating shared libraries to specify a list of
1453 global symbols whose references shouldn't be bound to the definition
1454 within the shared library, or creating dynamically linked executables
1455 to specify a list of symbols which should be added to the symbol table
1456 in the executable. This option is only meaningful on ELF platforms
1457 which support shared libraries.
1459 The format of the dynamic list is the same as the version node without
1460 scope and node name. See @ref{VERSION} for more information.
1462 @kindex --dynamic-list-data
1463 @item --dynamic-list-data
1464 Include all global data symbols to the dynamic list.
1466 @kindex --dynamic-list-cpp-new
1467 @item --dynamic-list-cpp-new
1468 Provide the builtin dynamic list for C++ operator new and delete. It
1469 is mainly useful for building shared libstdc++.
1471 @kindex --dynamic-list-cpp-typeinfo
1472 @item --dynamic-list-cpp-typeinfo
1473 Provide the builtin dynamic list for C++ runtime type identification.
1475 @kindex --check-sections
1476 @kindex --no-check-sections
1477 @item --check-sections
1478 @itemx --no-check-sections
1479 Asks the linker @emph{not} to check section addresses after they have
1480 been assigned to see if there are any overlaps. Normally the linker will
1481 perform this check, and if it finds any overlaps it will produce
1482 suitable error messages. The linker does know about, and does make
1483 allowances for sections in overlays. The default behaviour can be
1484 restored by using the command-line switch @option{--check-sections}.
1485 Section overlap is not usually checked for relocatable links. You can
1486 force checking in that case by using the @option{--check-sections}
1489 @kindex --copy-dt-needed-entries
1490 @kindex --no-copy-dt-needed-entries
1491 @item --copy-dt-needed-entries
1492 @itemx --no-copy-dt-needed-entries
1493 This option affects the treatment of dynamic libraries referred to
1494 by DT_NEEDED tags @emph{inside} ELF dynamic libraries mentioned on the
1495 command line. Normally the linker won't add a DT_NEEDED tag to the
1496 output binary for each library mentioned in a DT_NEEDED tag in an
1497 input dynamic library. With @option{--copy-dt-needed-entries}
1498 specified on the command line however any dynamic libraries that
1499 follow it will have their DT_NEEDED entries added. The default
1500 behaviour can be restored with @option{--no-copy-dt-needed-entries}.
1502 This option also has an effect on the resolution of symbols in dynamic
1503 libraries. With @option{--copy-dt-needed-entries} dynamic libraries
1504 mentioned on the command line will be recursively searched, following
1505 their DT_NEEDED tags to other libraries, in order to resolve symbols
1506 required by the output binary. With the default setting however
1507 the searching of dynamic libraries that follow it will stop with the
1508 dynamic library itself. No DT_NEEDED links will be traversed to resolve
1511 @cindex cross reference table
1514 Output a cross reference table. If a linker map file is being
1515 generated, the cross reference table is printed to the map file.
1516 Otherwise, it is printed on the standard output.
1518 The format of the table is intentionally simple, so that it may be
1519 easily processed by a script if necessary. The symbols are printed out,
1520 sorted by name. For each symbol, a list of file names is given. If the
1521 symbol is defined, the first file listed is the location of the
1522 definition. If the symbol is defined as a common value then any files
1523 where this happens appear next. Finally any files that reference the
1526 @cindex common allocation
1527 @kindex --no-define-common
1528 @item --no-define-common
1529 This option inhibits the assignment of addresses to common symbols.
1530 The script command @code{INHIBIT_COMMON_ALLOCATION} has the same effect.
1531 @xref{Miscellaneous Commands}.
1533 The @samp{--no-define-common} option allows decoupling
1534 the decision to assign addresses to Common symbols from the choice
1535 of the output file type; otherwise a non-Relocatable output type
1536 forces assigning addresses to Common symbols.
1537 Using @samp{--no-define-common} allows Common symbols that are referenced
1538 from a shared library to be assigned addresses only in the main program.
1539 This eliminates the unused duplicate space in the shared library,
1540 and also prevents any possible confusion over resolving to the wrong
1541 duplicate when there are many dynamic modules with specialized search
1542 paths for runtime symbol resolution.
1544 @cindex group allocation in linker script
1545 @cindex section groups
1547 @kindex --force-group-allocation
1548 @item --force-group-allocation
1549 This option causes the linker to place section group members like
1550 normal input sections, and to delete the section groups. This is the
1551 default behaviour for a final link but this option can be used to
1552 change the behaviour of a relocatable link (@samp{-r}). The script
1553 command @code{FORCE_GROUP_ALLOCATION} has the same
1554 effect. @xref{Miscellaneous Commands}.
1556 @cindex symbols, from command line
1557 @kindex --defsym=@var{symbol}=@var{exp}
1558 @item --defsym=@var{symbol}=@var{expression}
1559 Create a global symbol in the output file, containing the absolute
1560 address given by @var{expression}. You may use this option as many
1561 times as necessary to define multiple symbols in the command line. A
1562 limited form of arithmetic is supported for the @var{expression} in this
1563 context: you may give a hexadecimal constant or the name of an existing
1564 symbol, or use @code{+} and @code{-} to add or subtract hexadecimal
1565 constants or symbols. If you need more elaborate expressions, consider
1566 using the linker command language from a script (@pxref{Assignments}).
1567 @emph{Note:} there should be no white space between @var{symbol}, the
1568 equals sign (``@key{=}''), and @var{expression}.
1570 @cindex demangling, from command line
1571 @kindex --demangle[=@var{style}]
1572 @kindex --no-demangle
1573 @item --demangle[=@var{style}]
1574 @itemx --no-demangle
1575 These options control whether to demangle symbol names in error messages
1576 and other output. When the linker is told to demangle, it tries to
1577 present symbol names in a readable fashion: it strips leading
1578 underscores if they are used by the object file format, and converts C++
1579 mangled symbol names into user readable names. Different compilers have
1580 different mangling styles. The optional demangling style argument can be used
1581 to choose an appropriate demangling style for your compiler. The linker will
1582 demangle by default unless the environment variable @samp{COLLECT_NO_DEMANGLE}
1583 is set. These options may be used to override the default.
1585 @cindex dynamic linker, from command line
1586 @kindex -I@var{file}
1587 @kindex --dynamic-linker=@var{file}
1589 @itemx --dynamic-linker=@var{file}
1590 Set the name of the dynamic linker. This is only meaningful when
1591 generating dynamically linked ELF executables. The default dynamic
1592 linker is normally correct; don't use this unless you know what you are
1595 @kindex --no-dynamic-linker
1596 @item --no-dynamic-linker
1597 When producing an executable file, omit the request for a dynamic
1598 linker to be used at load-time. This is only meaningful for ELF
1599 executables that contain dynamic relocations, and usually requires
1600 entry point code that is capable of processing these relocations.
1602 @kindex --embedded-relocs
1603 @item --embedded-relocs
1604 This option is similar to the @option{--emit-relocs} option except
1605 that the relocs are stored in a target-specific section. This option
1606 is only supported by the @samp{BFIN}, @samp{CR16} and @emph{M68K}
1609 @kindex --disable-multiple-abs-defs
1610 @item --disable-multiple-abs-defs
1611 Do not allow multiple definitions with symbols included
1612 in filename invoked by -R or --just-symbols
1614 @kindex --fatal-warnings
1615 @kindex --no-fatal-warnings
1616 @item --fatal-warnings
1617 @itemx --no-fatal-warnings
1618 Treat all warnings as errors. The default behaviour can be restored
1619 with the option @option{--no-fatal-warnings}.
1621 @kindex --force-exe-suffix
1622 @item --force-exe-suffix
1623 Make sure that an output file has a .exe suffix.
1625 If a successfully built fully linked output file does not have a
1626 @code{.exe} or @code{.dll} suffix, this option forces the linker to copy
1627 the output file to one of the same name with a @code{.exe} suffix. This
1628 option is useful when using unmodified Unix makefiles on a Microsoft
1629 Windows host, since some versions of Windows won't run an image unless
1630 it ends in a @code{.exe} suffix.
1632 @kindex --gc-sections
1633 @kindex --no-gc-sections
1634 @cindex garbage collection
1636 @itemx --no-gc-sections
1637 Enable garbage collection of unused input sections. It is ignored on
1638 targets that do not support this option. The default behaviour (of not
1639 performing this garbage collection) can be restored by specifying
1640 @samp{--no-gc-sections} on the command line. Note that garbage
1641 collection for COFF and PE format targets is supported, but the
1642 implementation is currently considered to be experimental.
1644 @samp{--gc-sections} decides which input sections are used by
1645 examining symbols and relocations. The section containing the entry
1646 symbol and all sections containing symbols undefined on the
1647 command-line will be kept, as will sections containing symbols
1648 referenced by dynamic objects. Note that when building shared
1649 libraries, the linker must assume that any visible symbol is
1650 referenced. Once this initial set of sections has been determined,
1651 the linker recursively marks as used any section referenced by their
1652 relocations. See @samp{--entry}, @samp{--undefined}, and
1653 @samp{--gc-keep-exported}.
1655 This option can be set when doing a partial link (enabled with option
1656 @samp{-r}). In this case the root of symbols kept must be explicitly
1657 specified either by one of the options @samp{--entry},
1658 @samp{--undefined}, or @samp{--gc-keep-exported} or by a @code{ENTRY}
1659 command in the linker script.
1661 @kindex --print-gc-sections
1662 @kindex --no-print-gc-sections
1663 @cindex garbage collection
1664 @item --print-gc-sections
1665 @itemx --no-print-gc-sections
1666 List all sections removed by garbage collection. The listing is
1667 printed on stderr. This option is only effective if garbage
1668 collection has been enabled via the @samp{--gc-sections}) option. The
1669 default behaviour (of not listing the sections that are removed) can
1670 be restored by specifying @samp{--no-print-gc-sections} on the command
1673 @kindex --gc-keep-exported
1674 @cindex garbage collection
1675 @item --gc-keep-exported
1676 When @samp{--gc-sections} is enabled, this option prevents garbage
1677 collection of unused input sections that contain global symbols having
1678 default or protected visibility. This option is intended to be used for
1679 executables where unreferenced sections would otherwise be garbage
1680 collected regardless of the external visibility of contained symbols.
1681 Note that this option has no effect when linking shared objects since
1682 it is already the default behaviour. This option is only supported for
1685 @kindex --print-output-format
1686 @cindex output format
1687 @item --print-output-format
1688 Print the name of the default output format (perhaps influenced by
1689 other command-line options). This is the string that would appear
1690 in an @code{OUTPUT_FORMAT} linker script command (@pxref{File Commands}).
1692 @kindex --print-memory-usage
1693 @cindex memory usage
1694 @item --print-memory-usage
1695 Print used size, total size and used size of memory regions created with
1696 the @ref{MEMORY} command. This is useful on embedded targets to have a
1697 quick view of amount of free memory. The format of the output has one
1698 headline and one line per region. It is both human readable and easily
1699 parsable by tools. Here is an example of an output:
1702 Memory region Used Size Region Size %age Used
1703 ROM: 256 KB 1 MB 25.00%
1704 RAM: 32 B 2 GB 0.00%
1711 Print a summary of the command-line options on the standard output and exit.
1713 @kindex --target-help
1715 Print a summary of all target-specific options on the standard output and exit.
1717 @kindex -Map=@var{mapfile}
1718 @item -Map=@var{mapfile}
1719 Print a link map to the file @var{mapfile}. See the description of the
1720 @option{-M} option, above.
1722 @cindex memory usage
1723 @kindex --no-keep-memory
1724 @item --no-keep-memory
1725 @command{ld} normally optimizes for speed over memory usage by caching the
1726 symbol tables of input files in memory. This option tells @command{ld} to
1727 instead optimize for memory usage, by rereading the symbol tables as
1728 necessary. This may be required if @command{ld} runs out of memory space
1729 while linking a large executable.
1731 @kindex --no-undefined
1734 @item --no-undefined
1736 Report unresolved symbol references from regular object files. This
1737 is done even if the linker is creating a non-symbolic shared library.
1738 The switch @option{--[no-]allow-shlib-undefined} controls the
1739 behaviour for reporting unresolved references found in shared
1740 libraries being linked in.
1742 The effects of this option can be reverted by using @code{-z undefs}.
1744 @kindex --allow-multiple-definition
1746 @item --allow-multiple-definition
1748 Normally when a symbol is defined multiple times, the linker will
1749 report a fatal error. These options allow multiple definitions and the
1750 first definition will be used.
1752 @kindex --allow-shlib-undefined
1753 @kindex --no-allow-shlib-undefined
1754 @item --allow-shlib-undefined
1755 @itemx --no-allow-shlib-undefined
1756 Allows or disallows undefined symbols in shared libraries.
1757 This switch is similar to @option{--no-undefined} except that it
1758 determines the behaviour when the undefined symbols are in a
1759 shared library rather than a regular object file. It does not affect
1760 how undefined symbols in regular object files are handled.
1762 The default behaviour is to report errors for any undefined symbols
1763 referenced in shared libraries if the linker is being used to create
1764 an executable, but to allow them if the linker is being used to create
1767 The reasons for allowing undefined symbol references in shared
1768 libraries specified at link time are that:
1772 A shared library specified at link time may not be the same as the one
1773 that is available at load time, so the symbol might actually be
1774 resolvable at load time.
1776 There are some operating systems, eg BeOS and HPPA, where undefined
1777 symbols in shared libraries are normal.
1779 The BeOS kernel for example patches shared libraries at load time to
1780 select whichever function is most appropriate for the current
1781 architecture. This is used, for example, to dynamically select an
1782 appropriate memset function.
1785 @kindex --no-undefined-version
1786 @item --no-undefined-version
1787 Normally when a symbol has an undefined version, the linker will ignore
1788 it. This option disallows symbols with undefined version and a fatal error
1789 will be issued instead.
1791 @kindex --default-symver
1792 @item --default-symver
1793 Create and use a default symbol version (the soname) for unversioned
1796 @kindex --default-imported-symver
1797 @item --default-imported-symver
1798 Create and use a default symbol version (the soname) for unversioned
1801 @kindex --no-warn-mismatch
1802 @item --no-warn-mismatch
1803 Normally @command{ld} will give an error if you try to link together input
1804 files that are mismatched for some reason, perhaps because they have
1805 been compiled for different processors or for different endiannesses.
1806 This option tells @command{ld} that it should silently permit such possible
1807 errors. This option should only be used with care, in cases when you
1808 have taken some special action that ensures that the linker errors are
1811 @kindex --no-warn-search-mismatch
1812 @item --no-warn-search-mismatch
1813 Normally @command{ld} will give a warning if it finds an incompatible
1814 library during a library search. This option silences the warning.
1816 @kindex --no-whole-archive
1817 @item --no-whole-archive
1818 Turn off the effect of the @option{--whole-archive} option for subsequent
1821 @cindex output file after errors
1822 @kindex --noinhibit-exec
1823 @item --noinhibit-exec
1824 Retain the executable output file whenever it is still usable.
1825 Normally, the linker will not produce an output file if it encounters
1826 errors during the link process; it exits without writing an output file
1827 when it issues any error whatsoever.
1831 Only search library directories explicitly specified on the
1832 command line. Library directories specified in linker scripts
1833 (including linker scripts specified on the command line) are ignored.
1835 @ifclear SingleFormat
1836 @kindex --oformat=@var{output-format}
1837 @item --oformat=@var{output-format}
1838 @command{ld} may be configured to support more than one kind of object
1839 file. If your @command{ld} is configured this way, you can use the
1840 @samp{--oformat} option to specify the binary format for the output
1841 object file. Even when @command{ld} is configured to support alternative
1842 object formats, you don't usually need to specify this, as @command{ld}
1843 should be configured to produce as a default output format the most
1844 usual format on each machine. @var{output-format} is a text string, the
1845 name of a particular format supported by the BFD libraries. (You can
1846 list the available binary formats with @samp{objdump -i}.) The script
1847 command @code{OUTPUT_FORMAT} can also specify the output format, but
1848 this option overrides it. @xref{BFD}.
1851 @kindex --out-implib
1852 @item --out-implib @var{file}
1853 Create an import library in @var{file} corresponding to the executable
1854 the linker is generating (eg. a DLL or ELF program). This import
1855 library (which should be called @code{*.dll.a} or @code{*.a} for DLLs)
1856 may be used to link clients against the generated executable; this
1857 behaviour makes it possible to skip a separate import library creation
1858 step (eg. @code{dlltool} for DLLs). This option is only available for
1859 the i386 PE and ELF targetted ports of the linker.
1862 @kindex --pic-executable
1864 @itemx --pic-executable
1865 @cindex position independent executables
1866 Create a position independent executable. This is currently only supported on
1867 ELF platforms. Position independent executables are similar to shared
1868 libraries in that they are relocated by the dynamic linker to the virtual
1869 address the OS chooses for them (which can vary between invocations). Like
1870 normal dynamically linked executables they can be executed and symbols
1871 defined in the executable cannot be overridden by shared libraries.
1875 This option is ignored for Linux compatibility.
1879 This option is ignored for SVR4 compatibility.
1882 @cindex synthesizing linker
1883 @cindex relaxing addressing modes
1887 An option with machine dependent effects.
1889 This option is only supported on a few targets.
1892 @xref{H8/300,,@command{ld} and the H8/300}.
1895 @xref{Xtensa,, @command{ld} and Xtensa Processors}.
1898 @xref{M68HC11/68HC12,,@command{ld} and the 68HC11 and 68HC12}.
1901 @xref{Nios II,,@command{ld} and the Altera Nios II}.
1904 @xref{PowerPC ELF32,,@command{ld} and PowerPC 32-bit ELF Support}.
1907 On some platforms the @samp{--relax} option performs target-specific,
1908 global optimizations that become possible when the linker resolves
1909 addressing in the program, such as relaxing address modes,
1910 synthesizing new instructions, selecting shorter version of current
1911 instructions, and combining constant values.
1913 On some platforms these link time global optimizations may make symbolic
1914 debugging of the resulting executable impossible.
1916 This is known to be the case for the Matsushita MN10200 and MN10300
1917 family of processors.
1921 On platforms where this is not supported, @samp{--relax} is accepted,
1925 On platforms where @samp{--relax} is accepted the option
1926 @samp{--no-relax} can be used to disable the feature.
1928 @cindex retaining specified symbols
1929 @cindex stripping all but some symbols
1930 @cindex symbols, retaining selectively
1931 @kindex --retain-symbols-file=@var{filename}
1932 @item --retain-symbols-file=@var{filename}
1933 Retain @emph{only} the symbols listed in the file @var{filename},
1934 discarding all others. @var{filename} is simply a flat file, with one
1935 symbol name per line. This option is especially useful in environments
1939 where a large global symbol table is accumulated gradually, to conserve
1942 @samp{--retain-symbols-file} does @emph{not} discard undefined symbols,
1943 or symbols needed for relocations.
1945 You may only specify @samp{--retain-symbols-file} once in the command
1946 line. It overrides @samp{-s} and @samp{-S}.
1949 @item -rpath=@var{dir}
1950 @cindex runtime library search path
1951 @kindex -rpath=@var{dir}
1952 Add a directory to the runtime library search path. This is used when
1953 linking an ELF executable with shared objects. All @option{-rpath}
1954 arguments are concatenated and passed to the runtime linker, which uses
1955 them to locate shared objects at runtime.
1957 The @option{-rpath} option is also used when locating shared objects which
1958 are needed by shared objects explicitly included in the link; see the
1959 description of the @option{-rpath-link} option. Searching @option{-rpath}
1960 in this way is only supported by native linkers and cross linkers which
1961 have been configured with the @option{--with-sysroot} option.
1963 If @option{-rpath} is not used when linking an ELF executable, the
1964 contents of the environment variable @code{LD_RUN_PATH} will be used if it
1967 The @option{-rpath} option may also be used on SunOS. By default, on
1968 SunOS, the linker will form a runtime search path out of all the
1969 @option{-L} options it is given. If a @option{-rpath} option is used, the
1970 runtime search path will be formed exclusively using the @option{-rpath}
1971 options, ignoring the @option{-L} options. This can be useful when using
1972 gcc, which adds many @option{-L} options which may be on NFS mounted
1975 For compatibility with other ELF linkers, if the @option{-R} option is
1976 followed by a directory name, rather than a file name, it is treated as
1977 the @option{-rpath} option.
1981 @cindex link-time runtime library search path
1982 @kindex -rpath-link=@var{dir}
1983 @item -rpath-link=@var{dir}
1984 When using ELF or SunOS, one shared library may require another. This
1985 happens when an @code{ld -shared} link includes a shared library as one
1988 When the linker encounters such a dependency when doing a non-shared,
1989 non-relocatable link, it will automatically try to locate the required
1990 shared library and include it in the link, if it is not included
1991 explicitly. In such a case, the @option{-rpath-link} option
1992 specifies the first set of directories to search. The
1993 @option{-rpath-link} option may specify a sequence of directory names
1994 either by specifying a list of names separated by colons, or by
1995 appearing multiple times.
1997 The tokens @var{$ORIGIN} and @var{$LIB} can appear in these search
1998 directories. They will be replaced by the full path to the directory
1999 containing the program or shared object in the case of @var{$ORIGIN}
2000 and either @samp{lib} - for 32-bit binaries - or @samp{lib64} - for
2001 64-bit binaries - in the case of @var{$LIB}.
2003 The alternative form of these tokens - @var{$@{ORIGIN@}} and
2004 @var{$@{LIB@}} can also be used. The token @var{$PLATFORM} is not
2007 This option should be used with caution as it overrides the search path
2008 that may have been hard compiled into a shared library. In such a case it
2009 is possible to use unintentionally a different search path than the
2010 runtime linker would do.
2012 The linker uses the following search paths to locate required shared
2017 Any directories specified by @option{-rpath-link} options.
2019 Any directories specified by @option{-rpath} options. The difference
2020 between @option{-rpath} and @option{-rpath-link} is that directories
2021 specified by @option{-rpath} options are included in the executable and
2022 used at runtime, whereas the @option{-rpath-link} option is only effective
2023 at link time. Searching @option{-rpath} in this way is only supported
2024 by native linkers and cross linkers which have been configured with
2025 the @option{--with-sysroot} option.
2027 On an ELF system, for native linkers, if the @option{-rpath} and
2028 @option{-rpath-link} options were not used, search the contents of the
2029 environment variable @code{LD_RUN_PATH}.
2031 On SunOS, if the @option{-rpath} option was not used, search any
2032 directories specified using @option{-L} options.
2034 For a native linker, search the contents of the environment
2035 variable @code{LD_LIBRARY_PATH}.
2037 For a native ELF linker, the directories in @code{DT_RUNPATH} or
2038 @code{DT_RPATH} of a shared library are searched for shared
2039 libraries needed by it. The @code{DT_RPATH} entries are ignored if
2040 @code{DT_RUNPATH} entries exist.
2042 The default directories, normally @file{/lib} and @file{/usr/lib}.
2044 For a linker for a Linux system, if the file @file{/etc/ld.so.conf}
2045 exists, the list of directories found in that file. Note: the path
2046 to this file is prefixed with the @code{sysroot} value, if that is
2047 defined, and then any @code{prefix} string if the linker was
2048 configured with the @command{--prefix=<path>} option.
2050 For a native linker on a FreeBSD system, any directories specified by
2051 the @code{_PATH_ELF_HINTS} macro defined in the @file{elf-hints.h}
2054 Any directories specifed by a @code{SEARCH_DIR} command in the
2055 linker script being used.
2058 If the required shared library is not found, the linker will issue a
2059 warning and continue with the link.
2066 @cindex shared libraries
2067 Create a shared library. This is currently only supported on ELF, XCOFF
2068 and SunOS platforms. On SunOS, the linker will automatically create a
2069 shared library if the @option{-e} option is not used and there are
2070 undefined symbols in the link.
2072 @kindex --sort-common
2074 @itemx --sort-common=ascending
2075 @itemx --sort-common=descending
2076 This option tells @command{ld} to sort the common symbols by alignment in
2077 ascending or descending order when it places them in the appropriate output
2078 sections. The symbol alignments considered are sixteen-byte or larger,
2079 eight-byte, four-byte, two-byte, and one-byte. This is to prevent gaps
2080 between symbols due to alignment constraints. If no sorting order is
2081 specified, then descending order is assumed.
2083 @kindex --sort-section=name
2084 @item --sort-section=name
2085 This option will apply @code{SORT_BY_NAME} to all wildcard section
2086 patterns in the linker script.
2088 @kindex --sort-section=alignment
2089 @item --sort-section=alignment
2090 This option will apply @code{SORT_BY_ALIGNMENT} to all wildcard section
2091 patterns in the linker script.
2093 @kindex --spare-dynamic-tags
2094 @item --spare-dynamic-tags=@var{count}
2095 This option specifies the number of empty slots to leave in the
2096 .dynamic section of ELF shared objects. Empty slots may be needed by
2097 post processing tools, such as the prelinker. The default is 5.
2099 @kindex --split-by-file
2100 @item --split-by-file[=@var{size}]
2101 Similar to @option{--split-by-reloc} but creates a new output section for
2102 each input file when @var{size} is reached. @var{size} defaults to a
2103 size of 1 if not given.
2105 @kindex --split-by-reloc
2106 @item --split-by-reloc[=@var{count}]
2107 Tries to creates extra sections in the output file so that no single
2108 output section in the file contains more than @var{count} relocations.
2109 This is useful when generating huge relocatable files for downloading into
2110 certain real time kernels with the COFF object file format; since COFF
2111 cannot represent more than 65535 relocations in a single section. Note
2112 that this will fail to work with object file formats which do not
2113 support arbitrary sections. The linker will not split up individual
2114 input sections for redistribution, so if a single input section contains
2115 more than @var{count} relocations one output section will contain that
2116 many relocations. @var{count} defaults to a value of 32768.
2120 Compute and display statistics about the operation of the linker, such
2121 as execution time and memory usage.
2123 @kindex --sysroot=@var{directory}
2124 @item --sysroot=@var{directory}
2125 Use @var{directory} as the location of the sysroot, overriding the
2126 configure-time default. This option is only supported by linkers
2127 that were configured using @option{--with-sysroot}.
2131 This is used by COFF/PE based targets to create a task-linked object
2132 file where all of the global symbols have been converted to statics.
2134 @kindex --traditional-format
2135 @cindex traditional format
2136 @item --traditional-format
2137 For some targets, the output of @command{ld} is different in some ways from
2138 the output of some existing linker. This switch requests @command{ld} to
2139 use the traditional format instead.
2142 For example, on SunOS, @command{ld} combines duplicate entries in the
2143 symbol string table. This can reduce the size of an output file with
2144 full debugging information by over 30 percent. Unfortunately, the SunOS
2145 @code{dbx} program can not read the resulting program (@code{gdb} has no
2146 trouble). The @samp{--traditional-format} switch tells @command{ld} to not
2147 combine duplicate entries.
2149 @kindex --section-start=@var{sectionname}=@var{org}
2150 @item --section-start=@var{sectionname}=@var{org}
2151 Locate a section in the output file at the absolute
2152 address given by @var{org}. You may use this option as many
2153 times as necessary to locate multiple sections in the command
2155 @var{org} must be a single hexadecimal integer;
2156 for compatibility with other linkers, you may omit the leading
2157 @samp{0x} usually associated with hexadecimal values. @emph{Note:} there
2158 should be no white space between @var{sectionname}, the equals
2159 sign (``@key{=}''), and @var{org}.
2161 @kindex -Tbss=@var{org}
2162 @kindex -Tdata=@var{org}
2163 @kindex -Ttext=@var{org}
2164 @cindex segment origins, cmd line
2165 @item -Tbss=@var{org}
2166 @itemx -Tdata=@var{org}
2167 @itemx -Ttext=@var{org}
2168 Same as @option{--section-start}, with @code{.bss}, @code{.data} or
2169 @code{.text} as the @var{sectionname}.
2171 @kindex -Ttext-segment=@var{org}
2172 @item -Ttext-segment=@var{org}
2173 @cindex text segment origin, cmd line
2174 When creating an ELF executable, it will set the address of the first
2175 byte of the text segment.
2177 @kindex -Trodata-segment=@var{org}
2178 @item -Trodata-segment=@var{org}
2179 @cindex rodata segment origin, cmd line
2180 When creating an ELF executable or shared object for a target where
2181 the read-only data is in its own segment separate from the executable
2182 text, it will set the address of the first byte of the read-only data segment.
2184 @kindex -Tldata-segment=@var{org}
2185 @item -Tldata-segment=@var{org}
2186 @cindex ldata segment origin, cmd line
2187 When creating an ELF executable or shared object for x86-64 medium memory
2188 model, it will set the address of the first byte of the ldata segment.
2190 @kindex --unresolved-symbols
2191 @item --unresolved-symbols=@var{method}
2192 Determine how to handle unresolved symbols. There are four possible
2193 values for @samp{method}:
2197 Do not report any unresolved symbols.
2200 Report all unresolved symbols. This is the default.
2202 @item ignore-in-object-files
2203 Report unresolved symbols that are contained in shared libraries, but
2204 ignore them if they come from regular object files.
2206 @item ignore-in-shared-libs
2207 Report unresolved symbols that come from regular object files, but
2208 ignore them if they come from shared libraries. This can be useful
2209 when creating a dynamic binary and it is known that all the shared
2210 libraries that it should be referencing are included on the linker's
2214 The behaviour for shared libraries on their own can also be controlled
2215 by the @option{--[no-]allow-shlib-undefined} option.
2217 Normally the linker will generate an error message for each reported
2218 unresolved symbol but the option @option{--warn-unresolved-symbols}
2219 can change this to a warning.
2221 @kindex --verbose[=@var{NUMBER}]
2222 @cindex verbose[=@var{NUMBER}]
2224 @itemx --verbose[=@var{NUMBER}]
2225 Display the version number for @command{ld} and list the linker emulations
2226 supported. Display which input files can and cannot be opened. Display
2227 the linker script being used by the linker. If the optional @var{NUMBER}
2228 argument > 1, plugin symbol status will also be displayed.
2230 @kindex --version-script=@var{version-scriptfile}
2231 @cindex version script, symbol versions
2232 @item --version-script=@var{version-scriptfile}
2233 Specify the name of a version script to the linker. This is typically
2234 used when creating shared libraries to specify additional information
2235 about the version hierarchy for the library being created. This option
2236 is only fully supported on ELF platforms which support shared libraries;
2237 see @ref{VERSION}. It is partially supported on PE platforms, which can
2238 use version scripts to filter symbol visibility in auto-export mode: any
2239 symbols marked @samp{local} in the version script will not be exported.
2242 @kindex --warn-common
2243 @cindex warnings, on combining symbols
2244 @cindex combining symbols, warnings on
2246 Warn when a common symbol is combined with another common symbol or with
2247 a symbol definition. Unix linkers allow this somewhat sloppy practice,
2248 but linkers on some other operating systems do not. This option allows
2249 you to find potential problems from combining global symbols.
2250 Unfortunately, some C libraries use this practice, so you may get some
2251 warnings about symbols in the libraries as well as in your programs.
2253 There are three kinds of global symbols, illustrated here by C examples:
2257 A definition, which goes in the initialized data section of the output
2261 An undefined reference, which does not allocate space.
2262 There must be either a definition or a common symbol for the
2266 A common symbol. If there are only (one or more) common symbols for a
2267 variable, it goes in the uninitialized data area of the output file.
2268 The linker merges multiple common symbols for the same variable into a
2269 single symbol. If they are of different sizes, it picks the largest
2270 size. The linker turns a common symbol into a declaration, if there is
2271 a definition of the same variable.
2274 The @samp{--warn-common} option can produce five kinds of warnings.
2275 Each warning consists of a pair of lines: the first describes the symbol
2276 just encountered, and the second describes the previous symbol
2277 encountered with the same name. One or both of the two symbols will be
2282 Turning a common symbol into a reference, because there is already a
2283 definition for the symbol.
2285 @var{file}(@var{section}): warning: common of `@var{symbol}'
2286 overridden by definition
2287 @var{file}(@var{section}): warning: defined here
2291 Turning a common symbol into a reference, because a later definition for
2292 the symbol is encountered. This is the same as the previous case,
2293 except that the symbols are encountered in a different order.
2295 @var{file}(@var{section}): warning: definition of `@var{symbol}'
2297 @var{file}(@var{section}): warning: common is here
2301 Merging a common symbol with a previous same-sized common symbol.
2303 @var{file}(@var{section}): warning: multiple common
2305 @var{file}(@var{section}): warning: previous common is here
2309 Merging a common symbol with a previous larger common symbol.
2311 @var{file}(@var{section}): warning: common of `@var{symbol}'
2312 overridden by larger common
2313 @var{file}(@var{section}): warning: larger common is here
2317 Merging a common symbol with a previous smaller common symbol. This is
2318 the same as the previous case, except that the symbols are
2319 encountered in a different order.
2321 @var{file}(@var{section}): warning: common of `@var{symbol}'
2322 overriding smaller common
2323 @var{file}(@var{section}): warning: smaller common is here
2327 @kindex --warn-constructors
2328 @item --warn-constructors
2329 Warn if any global constructors are used. This is only useful for a few
2330 object file formats. For formats like COFF or ELF, the linker can not
2331 detect the use of global constructors.
2333 @kindex --warn-multiple-gp
2334 @item --warn-multiple-gp
2335 Warn if multiple global pointer values are required in the output file.
2336 This is only meaningful for certain processors, such as the Alpha.
2337 Specifically, some processors put large-valued constants in a special
2338 section. A special register (the global pointer) points into the middle
2339 of this section, so that constants can be loaded efficiently via a
2340 base-register relative addressing mode. Since the offset in
2341 base-register relative mode is fixed and relatively small (e.g., 16
2342 bits), this limits the maximum size of the constant pool. Thus, in
2343 large programs, it is often necessary to use multiple global pointer
2344 values in order to be able to address all possible constants. This
2345 option causes a warning to be issued whenever this case occurs.
2348 @cindex warnings, on undefined symbols
2349 @cindex undefined symbols, warnings on
2351 Only warn once for each undefined symbol, rather than once per module
2354 @kindex --warn-section-align
2355 @cindex warnings, on section alignment
2356 @cindex section alignment, warnings on
2357 @item --warn-section-align
2358 Warn if the address of an output section is changed because of
2359 alignment. Typically, the alignment will be set by an input section.
2360 The address will only be changed if it not explicitly specified; that
2361 is, if the @code{SECTIONS} command does not specify a start address for
2362 the section (@pxref{SECTIONS}).
2364 @kindex --warn-shared-textrel
2365 @item --warn-shared-textrel
2366 Warn if the linker adds a DT_TEXTREL to a shared object.
2368 @kindex --warn-alternate-em
2369 @item --warn-alternate-em
2370 Warn if an object has alternate ELF machine code.
2372 @kindex --warn-unresolved-symbols
2373 @item --warn-unresolved-symbols
2374 If the linker is going to report an unresolved symbol (see the option
2375 @option{--unresolved-symbols}) it will normally generate an error.
2376 This option makes it generate a warning instead.
2378 @kindex --error-unresolved-symbols
2379 @item --error-unresolved-symbols
2380 This restores the linker's default behaviour of generating errors when
2381 it is reporting unresolved symbols.
2383 @kindex --whole-archive
2384 @cindex including an entire archive
2385 @item --whole-archive
2386 For each archive mentioned on the command line after the
2387 @option{--whole-archive} option, include every object file in the archive
2388 in the link, rather than searching the archive for the required object
2389 files. This is normally used to turn an archive file into a shared
2390 library, forcing every object to be included in the resulting shared
2391 library. This option may be used more than once.
2393 Two notes when using this option from gcc: First, gcc doesn't know
2394 about this option, so you have to use @option{-Wl,-whole-archive}.
2395 Second, don't forget to use @option{-Wl,-no-whole-archive} after your
2396 list of archives, because gcc will add its own list of archives to
2397 your link and you may not want this flag to affect those as well.
2399 @kindex --wrap=@var{symbol}
2400 @item --wrap=@var{symbol}
2401 Use a wrapper function for @var{symbol}. Any undefined reference to
2402 @var{symbol} will be resolved to @code{__wrap_@var{symbol}}. Any
2403 undefined reference to @code{__real_@var{symbol}} will be resolved to
2406 This can be used to provide a wrapper for a system function. The
2407 wrapper function should be called @code{__wrap_@var{symbol}}. If it
2408 wishes to call the system function, it should call
2409 @code{__real_@var{symbol}}.
2411 Here is a trivial example:
2415 __wrap_malloc (size_t c)
2417 printf ("malloc called with %zu\n", c);
2418 return __real_malloc (c);
2422 If you link other code with this file using @option{--wrap malloc}, then
2423 all calls to @code{malloc} will call the function @code{__wrap_malloc}
2424 instead. The call to @code{__real_malloc} in @code{__wrap_malloc} will
2425 call the real @code{malloc} function.
2427 You may wish to provide a @code{__real_malloc} function as well, so that
2428 links without the @option{--wrap} option will succeed. If you do this,
2429 you should not put the definition of @code{__real_malloc} in the same
2430 file as @code{__wrap_malloc}; if you do, the assembler may resolve the
2431 call before the linker has a chance to wrap it to @code{malloc}.
2433 Only undefined references are replaced by the linker. So, translation unit
2434 internal references to @var{symbol} are not resolved to
2435 @code{__wrap_@var{symbol}}. In the next example, the call to @code{f} in
2436 @code{g} is not resolved to @code{__wrap_f}.
2452 @kindex --eh-frame-hdr
2453 @kindex --no-eh-frame-hdr
2454 @item --eh-frame-hdr
2455 @itemx --no-eh-frame-hdr
2456 Request (@option{--eh-frame-hdr}) or suppress
2457 (@option{--no-eh-frame-hdr}) the creation of @code{.eh_frame_hdr}
2458 section and ELF @code{PT_GNU_EH_FRAME} segment header.
2460 @kindex --ld-generated-unwind-info
2461 @item --no-ld-generated-unwind-info
2462 Request creation of @code{.eh_frame} unwind info for linker
2463 generated code sections like PLT. This option is on by default
2464 if linker generated unwind info is supported.
2466 @kindex --enable-new-dtags
2467 @kindex --disable-new-dtags
2468 @item --enable-new-dtags
2469 @itemx --disable-new-dtags
2470 This linker can create the new dynamic tags in ELF. But the older ELF
2471 systems may not understand them. If you specify
2472 @option{--enable-new-dtags}, the new dynamic tags will be created as needed
2473 and older dynamic tags will be omitted.
2474 If you specify @option{--disable-new-dtags}, no new dynamic tags will be
2475 created. By default, the new dynamic tags are not created. Note that
2476 those options are only available for ELF systems.
2478 @kindex --hash-size=@var{number}
2479 @item --hash-size=@var{number}
2480 Set the default size of the linker's hash tables to a prime number
2481 close to @var{number}. Increasing this value can reduce the length of
2482 time it takes the linker to perform its tasks, at the expense of
2483 increasing the linker's memory requirements. Similarly reducing this
2484 value can reduce the memory requirements at the expense of speed.
2486 @kindex --hash-style=@var{style}
2487 @item --hash-style=@var{style}
2488 Set the type of linker's hash table(s). @var{style} can be either
2489 @code{sysv} for classic ELF @code{.hash} section, @code{gnu} for
2490 new style GNU @code{.gnu.hash} section or @code{both} for both
2491 the classic ELF @code{.hash} and new style GNU @code{.gnu.hash}
2492 hash tables. The default depends upon how the linker was configured,
2493 but for most Linux based systems it will be @code{both}.
2495 @kindex --compress-debug-sections=none
2496 @kindex --compress-debug-sections=zlib
2497 @kindex --compress-debug-sections=zlib-gnu
2498 @kindex --compress-debug-sections=zlib-gabi
2499 @item --compress-debug-sections=none
2500 @itemx --compress-debug-sections=zlib
2501 @itemx --compress-debug-sections=zlib-gnu
2502 @itemx --compress-debug-sections=zlib-gabi
2503 On ELF platforms, these options control how DWARF debug sections are
2504 compressed using zlib.
2506 @option{--compress-debug-sections=none} doesn't compress DWARF debug
2507 sections. @option{--compress-debug-sections=zlib-gnu} compresses
2508 DWARF debug sections and renames them to begin with @samp{.zdebug}
2509 instead of @samp{.debug}. @option{--compress-debug-sections=zlib-gabi}
2510 also compresses DWARF debug sections, but rather than renaming them it
2511 sets the SHF_COMPRESSED flag in the sections' headers.
2513 The @option{--compress-debug-sections=zlib} option is an alias for
2514 @option{--compress-debug-sections=zlib-gabi}.
2516 Note that this option overrides any compression in input debug
2517 sections, so if a binary is linked with @option{--compress-debug-sections=none}
2518 for example, then any compressed debug sections in input files will be
2519 uncompressed before they are copied into the output binary.
2521 The default compression behaviour varies depending upon the target
2522 involved and the configure options used to build the toolchain. The
2523 default can be determined by examining the output from the linker's
2524 @option{--help} option.
2526 @kindex --reduce-memory-overheads
2527 @item --reduce-memory-overheads
2528 This option reduces memory requirements at ld runtime, at the expense of
2529 linking speed. This was introduced to select the old O(n^2) algorithm
2530 for link map file generation, rather than the new O(n) algorithm which uses
2531 about 40% more memory for symbol storage.
2533 Another effect of the switch is to set the default hash table size to
2534 1021, which again saves memory at the cost of lengthening the linker's
2535 run time. This is not done however if the @option{--hash-size} switch
2538 The @option{--reduce-memory-overheads} switch may be also be used to
2539 enable other tradeoffs in future versions of the linker.
2542 @kindex --build-id=@var{style}
2544 @itemx --build-id=@var{style}
2545 Request the creation of a @code{.note.gnu.build-id} ELF note section
2546 or a @code{.buildid} COFF section. The contents of the note are
2547 unique bits identifying this linked file. @var{style} can be
2548 @code{uuid} to use 128 random bits, @code{sha1} to use a 160-bit
2549 @sc{SHA1} hash on the normative parts of the output contents,
2550 @code{md5} to use a 128-bit @sc{MD5} hash on the normative parts of
2551 the output contents, or @code{0x@var{hexstring}} to use a chosen bit
2552 string specified as an even number of hexadecimal digits (@code{-} and
2553 @code{:} characters between digit pairs are ignored). If @var{style}
2554 is omitted, @code{sha1} is used.
2556 The @code{md5} and @code{sha1} styles produces an identifier
2557 that is always the same in an identical output file, but will be
2558 unique among all nonidentical output files. It is not intended
2559 to be compared as a checksum for the file's contents. A linked
2560 file may be changed later by other tools, but the build ID bit
2561 string identifying the original linked file does not change.
2563 Passing @code{none} for @var{style} disables the setting from any
2564 @code{--build-id} options earlier on the command line.
2569 @subsection Options Specific to i386 PE Targets
2571 @c man begin OPTIONS
2573 The i386 PE linker supports the @option{-shared} option, which causes
2574 the output to be a dynamically linked library (DLL) instead of a
2575 normal executable. You should name the output @code{*.dll} when you
2576 use this option. In addition, the linker fully supports the standard
2577 @code{*.def} files, which may be specified on the linker command line
2578 like an object file (in fact, it should precede archives it exports
2579 symbols from, to ensure that they get linked in, just like a normal
2582 In addition to the options common to all targets, the i386 PE linker
2583 support additional command-line options that are specific to the i386
2584 PE target. Options that take values may be separated from their
2585 values by either a space or an equals sign.
2589 @kindex --add-stdcall-alias
2590 @item --add-stdcall-alias
2591 If given, symbols with a stdcall suffix (@@@var{nn}) will be exported
2592 as-is and also with the suffix stripped.
2593 [This option is specific to the i386 PE targeted port of the linker]
2596 @item --base-file @var{file}
2597 Use @var{file} as the name of a file in which to save the base
2598 addresses of all the relocations needed for generating DLLs with
2600 [This is an i386 PE specific option]
2604 Create a DLL instead of a regular executable. You may also use
2605 @option{-shared} or specify a @code{LIBRARY} in a given @code{.def}
2607 [This option is specific to the i386 PE targeted port of the linker]
2609 @kindex --enable-long-section-names
2610 @kindex --disable-long-section-names
2611 @item --enable-long-section-names
2612 @itemx --disable-long-section-names
2613 The PE variants of the COFF object format add an extension that permits
2614 the use of section names longer than eight characters, the normal limit
2615 for COFF. By default, these names are only allowed in object files, as
2616 fully-linked executable images do not carry the COFF string table required
2617 to support the longer names. As a GNU extension, it is possible to
2618 allow their use in executable images as well, or to (probably pointlessly!)
2619 disallow it in object files, by using these two options. Executable images
2620 generated with these long section names are slightly non-standard, carrying
2621 as they do a string table, and may generate confusing output when examined
2622 with non-GNU PE-aware tools, such as file viewers and dumpers. However,
2623 GDB relies on the use of PE long section names to find Dwarf-2 debug
2624 information sections in an executable image at runtime, and so if neither
2625 option is specified on the command-line, @command{ld} will enable long
2626 section names, overriding the default and technically correct behaviour,
2627 when it finds the presence of debug information while linking an executable
2628 image and not stripping symbols.
2629 [This option is valid for all PE targeted ports of the linker]
2631 @kindex --enable-stdcall-fixup
2632 @kindex --disable-stdcall-fixup
2633 @item --enable-stdcall-fixup
2634 @itemx --disable-stdcall-fixup
2635 If the link finds a symbol that it cannot resolve, it will attempt to
2636 do ``fuzzy linking'' by looking for another defined symbol that differs
2637 only in the format of the symbol name (cdecl vs stdcall) and will
2638 resolve that symbol by linking to the match. For example, the
2639 undefined symbol @code{_foo} might be linked to the function
2640 @code{_foo@@12}, or the undefined symbol @code{_bar@@16} might be linked
2641 to the function @code{_bar}. When the linker does this, it prints a
2642 warning, since it normally should have failed to link, but sometimes
2643 import libraries generated from third-party dlls may need this feature
2644 to be usable. If you specify @option{--enable-stdcall-fixup}, this
2645 feature is fully enabled and warnings are not printed. If you specify
2646 @option{--disable-stdcall-fixup}, this feature is disabled and such
2647 mismatches are considered to be errors.
2648 [This option is specific to the i386 PE targeted port of the linker]
2650 @kindex --leading-underscore
2651 @kindex --no-leading-underscore
2652 @item --leading-underscore
2653 @itemx --no-leading-underscore
2654 For most targets default symbol-prefix is an underscore and is defined
2655 in target's description. By this option it is possible to
2656 disable/enable the default underscore symbol-prefix.
2658 @cindex DLLs, creating
2659 @kindex --export-all-symbols
2660 @item --export-all-symbols
2661 If given, all global symbols in the objects used to build a DLL will
2662 be exported by the DLL. Note that this is the default if there
2663 otherwise wouldn't be any exported symbols. When symbols are
2664 explicitly exported via DEF files or implicitly exported via function
2665 attributes, the default is to not export anything else unless this
2666 option is given. Note that the symbols @code{DllMain@@12},
2667 @code{DllEntryPoint@@0}, @code{DllMainCRTStartup@@12}, and
2668 @code{impure_ptr} will not be automatically
2669 exported. Also, symbols imported from other DLLs will not be
2670 re-exported, nor will symbols specifying the DLL's internal layout
2671 such as those beginning with @code{_head_} or ending with
2672 @code{_iname}. In addition, no symbols from @code{libgcc},
2673 @code{libstd++}, @code{libmingw32}, or @code{crtX.o} will be exported.
2674 Symbols whose names begin with @code{__rtti_} or @code{__builtin_} will
2675 not be exported, to help with C++ DLLs. Finally, there is an
2676 extensive list of cygwin-private symbols that are not exported
2677 (obviously, this applies on when building DLLs for cygwin targets).
2678 These cygwin-excludes are: @code{_cygwin_dll_entry@@12},
2679 @code{_cygwin_crt0_common@@8}, @code{_cygwin_noncygwin_dll_entry@@12},
2680 @code{_fmode}, @code{_impure_ptr}, @code{cygwin_attach_dll},
2681 @code{cygwin_premain0}, @code{cygwin_premain1}, @code{cygwin_premain2},
2682 @code{cygwin_premain3}, and @code{environ}.
2683 [This option is specific to the i386 PE targeted port of the linker]
2685 @kindex --exclude-symbols
2686 @item --exclude-symbols @var{symbol},@var{symbol},...
2687 Specifies a list of symbols which should not be automatically
2688 exported. The symbol names may be delimited by commas or colons.
2689 [This option is specific to the i386 PE targeted port of the linker]
2691 @kindex --exclude-all-symbols
2692 @item --exclude-all-symbols
2693 Specifies no symbols should be automatically exported.
2694 [This option is specific to the i386 PE targeted port of the linker]
2696 @kindex --file-alignment
2697 @item --file-alignment
2698 Specify the file alignment. Sections in the file will always begin at
2699 file offsets which are multiples of this number. This defaults to
2701 [This option is specific to the i386 PE targeted port of the linker]
2705 @item --heap @var{reserve}
2706 @itemx --heap @var{reserve},@var{commit}
2707 Specify the number of bytes of memory to reserve (and optionally commit)
2708 to be used as heap for this program. The default is 1MB reserved, 4K
2710 [This option is specific to the i386 PE targeted port of the linker]
2713 @kindex --image-base
2714 @item --image-base @var{value}
2715 Use @var{value} as the base address of your program or dll. This is
2716 the lowest memory location that will be used when your program or dll
2717 is loaded. To reduce the need to relocate and improve performance of
2718 your dlls, each should have a unique base address and not overlap any
2719 other dlls. The default is 0x400000 for executables, and 0x10000000
2721 [This option is specific to the i386 PE targeted port of the linker]
2725 If given, the stdcall suffixes (@@@var{nn}) will be stripped from
2726 symbols before they are exported.
2727 [This option is specific to the i386 PE targeted port of the linker]
2729 @kindex --large-address-aware
2730 @item --large-address-aware
2731 If given, the appropriate bit in the ``Characteristics'' field of the COFF
2732 header is set to indicate that this executable supports virtual addresses
2733 greater than 2 gigabytes. This should be used in conjunction with the /3GB
2734 or /USERVA=@var{value} megabytes switch in the ``[operating systems]''
2735 section of the BOOT.INI. Otherwise, this bit has no effect.
2736 [This option is specific to PE targeted ports of the linker]
2738 @kindex --disable-large-address-aware
2739 @item --disable-large-address-aware
2740 Reverts the effect of a previous @samp{--large-address-aware} option.
2741 This is useful if @samp{--large-address-aware} is always set by the compiler
2742 driver (e.g. Cygwin gcc) and the executable does not support virtual
2743 addresses greater than 2 gigabytes.
2744 [This option is specific to PE targeted ports of the linker]
2746 @kindex --major-image-version
2747 @item --major-image-version @var{value}
2748 Sets the major number of the ``image version''. Defaults to 1.
2749 [This option is specific to the i386 PE targeted port of the linker]
2751 @kindex --major-os-version
2752 @item --major-os-version @var{value}
2753 Sets the major number of the ``os version''. Defaults to 4.
2754 [This option is specific to the i386 PE targeted port of the linker]
2756 @kindex --major-subsystem-version
2757 @item --major-subsystem-version @var{value}
2758 Sets the major number of the ``subsystem version''. Defaults to 4.
2759 [This option is specific to the i386 PE targeted port of the linker]
2761 @kindex --minor-image-version
2762 @item --minor-image-version @var{value}
2763 Sets the minor number of the ``image version''. Defaults to 0.
2764 [This option is specific to the i386 PE targeted port of the linker]
2766 @kindex --minor-os-version
2767 @item --minor-os-version @var{value}
2768 Sets the minor number of the ``os version''. Defaults to 0.
2769 [This option is specific to the i386 PE targeted port of the linker]
2771 @kindex --minor-subsystem-version
2772 @item --minor-subsystem-version @var{value}
2773 Sets the minor number of the ``subsystem version''. Defaults to 0.
2774 [This option is specific to the i386 PE targeted port of the linker]
2776 @cindex DEF files, creating
2777 @cindex DLLs, creating
2778 @kindex --output-def
2779 @item --output-def @var{file}
2780 The linker will create the file @var{file} which will contain a DEF
2781 file corresponding to the DLL the linker is generating. This DEF file
2782 (which should be called @code{*.def}) may be used to create an import
2783 library with @code{dlltool} or may be used as a reference to
2784 automatically or implicitly exported symbols.
2785 [This option is specific to the i386 PE targeted port of the linker]
2787 @cindex DLLs, creating
2788 @kindex --enable-auto-image-base
2789 @item --enable-auto-image-base
2790 @itemx --enable-auto-image-base=@var{value}
2791 Automatically choose the image base for DLLs, optionally starting with base
2792 @var{value}, unless one is specified using the @code{--image-base} argument.
2793 By using a hash generated from the dllname to create unique image bases
2794 for each DLL, in-memory collisions and relocations which can delay program
2795 execution are avoided.
2796 [This option is specific to the i386 PE targeted port of the linker]
2798 @kindex --disable-auto-image-base
2799 @item --disable-auto-image-base
2800 Do not automatically generate a unique image base. If there is no
2801 user-specified image base (@code{--image-base}) then use the platform
2803 [This option is specific to the i386 PE targeted port of the linker]
2805 @cindex DLLs, linking to
2806 @kindex --dll-search-prefix
2807 @item --dll-search-prefix @var{string}
2808 When linking dynamically to a dll without an import library,
2809 search for @code{<string><basename>.dll} in preference to
2810 @code{lib<basename>.dll}. This behaviour allows easy distinction
2811 between DLLs built for the various "subplatforms": native, cygwin,
2812 uwin, pw, etc. For instance, cygwin DLLs typically use
2813 @code{--dll-search-prefix=cyg}.
2814 [This option is specific to the i386 PE targeted port of the linker]
2816 @kindex --enable-auto-import
2817 @item --enable-auto-import
2818 Do sophisticated linking of @code{_symbol} to @code{__imp__symbol} for
2819 DATA imports from DLLs, thus making it possible to bypass the dllimport
2820 mechanism on the user side and to reference unmangled symbol names.
2821 [This option is specific to the i386 PE targeted port of the linker]
2823 The following remarks pertain to the original implementation of the
2824 feature and are obsolete nowadays for Cygwin and MinGW targets.
2826 Note: Use of the 'auto-import' extension will cause the text section
2827 of the image file to be made writable. This does not conform to the
2828 PE-COFF format specification published by Microsoft.
2830 Note - use of the 'auto-import' extension will also cause read only
2831 data which would normally be placed into the .rdata section to be
2832 placed into the .data section instead. This is in order to work
2833 around a problem with consts that is described here:
2834 http://www.cygwin.com/ml/cygwin/2004-09/msg01101.html
2836 Using 'auto-import' generally will 'just work' -- but sometimes you may
2839 "variable '<var>' can't be auto-imported. Please read the
2840 documentation for ld's @code{--enable-auto-import} for details."
2842 This message occurs when some (sub)expression accesses an address
2843 ultimately given by the sum of two constants (Win32 import tables only
2844 allow one). Instances where this may occur include accesses to member
2845 fields of struct variables imported from a DLL, as well as using a
2846 constant index into an array variable imported from a DLL. Any
2847 multiword variable (arrays, structs, long long, etc) may trigger
2848 this error condition. However, regardless of the exact data type
2849 of the offending exported variable, ld will always detect it, issue
2850 the warning, and exit.
2852 There are several ways to address this difficulty, regardless of the
2853 data type of the exported variable:
2855 One way is to use --enable-runtime-pseudo-reloc switch. This leaves the task
2856 of adjusting references in your client code for runtime environment, so
2857 this method works only when runtime environment supports this feature.
2859 A second solution is to force one of the 'constants' to be a variable --
2860 that is, unknown and un-optimizable at compile time. For arrays,
2861 there are two possibilities: a) make the indexee (the array's address)
2862 a variable, or b) make the 'constant' index a variable. Thus:
2865 extern type extern_array[];
2867 @{ volatile type *t=extern_array; t[1] @}
2873 extern type extern_array[];
2875 @{ volatile int t=1; extern_array[t] @}
2878 For structs (and most other multiword data types) the only option
2879 is to make the struct itself (or the long long, or the ...) variable:
2882 extern struct s extern_struct;
2883 extern_struct.field -->
2884 @{ volatile struct s *t=&extern_struct; t->field @}
2890 extern long long extern_ll;
2892 @{ volatile long long * local_ll=&extern_ll; *local_ll @}
2895 A third method of dealing with this difficulty is to abandon
2896 'auto-import' for the offending symbol and mark it with
2897 @code{__declspec(dllimport)}. However, in practice that
2898 requires using compile-time #defines to indicate whether you are
2899 building a DLL, building client code that will link to the DLL, or
2900 merely building/linking to a static library. In making the choice
2901 between the various methods of resolving the 'direct address with
2902 constant offset' problem, you should consider typical real-world usage:
2910 void main(int argc, char **argv)@{
2911 printf("%d\n",arr[1]);
2921 void main(int argc, char **argv)@{
2922 /* This workaround is for win32 and cygwin; do not "optimize" */
2923 volatile int *parr = arr;
2924 printf("%d\n",parr[1]);
2931 /* Note: auto-export is assumed (no __declspec(dllexport)) */
2932 #if (defined(_WIN32) || defined(__CYGWIN__)) && \
2933 !(defined(FOO_BUILD_DLL) || defined(FOO_STATIC))
2934 #define FOO_IMPORT __declspec(dllimport)
2938 extern FOO_IMPORT int arr[];
2941 void main(int argc, char **argv)@{
2942 printf("%d\n",arr[1]);
2946 A fourth way to avoid this problem is to re-code your
2947 library to use a functional interface rather than a data interface
2948 for the offending variables (e.g. set_foo() and get_foo() accessor
2951 @kindex --disable-auto-import
2952 @item --disable-auto-import
2953 Do not attempt to do sophisticated linking of @code{_symbol} to
2954 @code{__imp__symbol} for DATA imports from DLLs.
2955 [This option is specific to the i386 PE targeted port of the linker]
2957 @kindex --enable-runtime-pseudo-reloc
2958 @item --enable-runtime-pseudo-reloc
2959 If your code contains expressions described in --enable-auto-import section,
2960 that is, DATA imports from DLL with non-zero offset, this switch will create
2961 a vector of 'runtime pseudo relocations' which can be used by runtime
2962 environment to adjust references to such data in your client code.
2963 [This option is specific to the i386 PE targeted port of the linker]
2965 @kindex --disable-runtime-pseudo-reloc
2966 @item --disable-runtime-pseudo-reloc
2967 Do not create pseudo relocations for non-zero offset DATA imports from DLLs.
2968 [This option is specific to the i386 PE targeted port of the linker]
2970 @kindex --enable-extra-pe-debug
2971 @item --enable-extra-pe-debug
2972 Show additional debug info related to auto-import symbol thunking.
2973 [This option is specific to the i386 PE targeted port of the linker]
2975 @kindex --section-alignment
2976 @item --section-alignment
2977 Sets the section alignment. Sections in memory will always begin at
2978 addresses which are a multiple of this number. Defaults to 0x1000.
2979 [This option is specific to the i386 PE targeted port of the linker]
2983 @item --stack @var{reserve}
2984 @itemx --stack @var{reserve},@var{commit}
2985 Specify the number of bytes of memory to reserve (and optionally commit)
2986 to be used as stack for this program. The default is 2MB reserved, 4K
2988 [This option is specific to the i386 PE targeted port of the linker]
2991 @item --subsystem @var{which}
2992 @itemx --subsystem @var{which}:@var{major}
2993 @itemx --subsystem @var{which}:@var{major}.@var{minor}
2994 Specifies the subsystem under which your program will execute. The
2995 legal values for @var{which} are @code{native}, @code{windows},
2996 @code{console}, @code{posix}, and @code{xbox}. You may optionally set
2997 the subsystem version also. Numeric values are also accepted for
2999 [This option is specific to the i386 PE targeted port of the linker]
3001 The following options set flags in the @code{DllCharacteristics} field
3002 of the PE file header:
3003 [These options are specific to PE targeted ports of the linker]
3005 @kindex --high-entropy-va
3006 @item --high-entropy-va
3007 Image is compatible with 64-bit address space layout randomization
3009 This option also implies @option{--dynamicbase} and
3010 @option{--enable-reloc-section}.
3012 @kindex --dynamicbase
3014 The image base address may be relocated using address space layout
3015 randomization (ASLR). This feature was introduced with MS Windows
3016 Vista for i386 PE targets.
3017 This option also implies @option{--enable-reloc-section}.
3019 @kindex --forceinteg
3021 Code integrity checks are enforced.
3025 The image is compatible with the Data Execution Prevention.
3026 This feature was introduced with MS Windows XP SP2 for i386 PE targets.
3028 @kindex --no-isolation
3029 @item --no-isolation
3030 Although the image understands isolation, do not isolate the image.
3034 The image does not use SEH. No SE handler may be called from
3039 Do not bind this image.
3043 The driver uses the MS Windows Driver Model.
3047 The image is Terminal Server aware.
3049 @kindex --insert-timestamp
3050 @item --insert-timestamp
3051 @itemx --no-insert-timestamp
3052 Insert a real timestamp into the image. This is the default behaviour
3053 as it matches legacy code and it means that the image will work with
3054 other, proprietary tools. The problem with this default is that it
3055 will result in slightly different images being produced each time the
3056 same sources are linked. The option @option{--no-insert-timestamp}
3057 can be used to insert a zero value for the timestamp, this ensuring
3058 that binaries produced from identical sources will compare
3061 @kindex --enable-reloc-section
3062 @item --enable-reloc-section
3063 Create the base relocation table, which is necessary if the image
3064 is loaded at a different image base than specified in the PE header.
3070 @subsection Options specific to C6X uClinux targets
3072 @c man begin OPTIONS
3074 The C6X uClinux target uses a binary format called DSBT to support shared
3075 libraries. Each shared library in the system needs to have a unique index;
3076 all executables use an index of 0.
3081 @item --dsbt-size @var{size}
3082 This option sets the number of entries in the DSBT of the current executable
3083 or shared library to @var{size}. The default is to create a table with 64
3086 @kindex --dsbt-index
3087 @item --dsbt-index @var{index}
3088 This option sets the DSBT index of the current executable or shared library
3089 to @var{index}. The default is 0, which is appropriate for generating
3090 executables. If a shared library is generated with a DSBT index of 0, the
3091 @code{R_C6000_DSBT_INDEX} relocs are copied into the output file.
3093 @kindex --no-merge-exidx-entries
3094 The @samp{--no-merge-exidx-entries} switch disables the merging of adjacent
3095 exidx entries in frame unwind info.
3103 @subsection Options specific to C-SKY targets
3105 @c man begin OPTIONS
3109 @kindex --branch-stub on C-SKY
3111 This option enables linker branch relaxation by inserting branch stub
3112 sections when needed to extend the range of branches. This option is
3113 usually not required since C-SKY supports branch and call instructions that
3114 can access the full memory range and branch relaxation is normally handled by
3115 the compiler or assembler.
3117 @kindex --stub-group-size on C-SKY
3118 @item --stub-group-size=@var{N}
3119 This option allows finer control of linker branch stub creation.
3120 It sets the maximum size of a group of input sections that can
3121 be handled by one stub section. A negative value of @var{N} locates
3122 stub sections after their branches, while a positive value allows stub
3123 sections to appear either before or after the branches. Values of
3124 @samp{1} or @samp{-1} indicate that the
3125 linker should choose suitable defaults.
3133 @subsection Options specific to Motorola 68HC11 and 68HC12 targets
3135 @c man begin OPTIONS
3137 The 68HC11 and 68HC12 linkers support specific options to control the
3138 memory bank switching mapping and trampoline code generation.
3142 @kindex --no-trampoline
3143 @item --no-trampoline
3144 This option disables the generation of trampoline. By default a trampoline
3145 is generated for each far function which is called using a @code{jsr}
3146 instruction (this happens when a pointer to a far function is taken).
3148 @kindex --bank-window
3149 @item --bank-window @var{name}
3150 This option indicates to the linker the name of the memory region in
3151 the @samp{MEMORY} specification that describes the memory bank window.
3152 The definition of such region is then used by the linker to compute
3153 paging and addresses within the memory window.
3161 @subsection Options specific to Motorola 68K target
3163 @c man begin OPTIONS
3165 The following options are supported to control handling of GOT generation
3166 when linking for 68K targets.
3171 @item --got=@var{type}
3172 This option tells the linker which GOT generation scheme to use.
3173 @var{type} should be one of @samp{single}, @samp{negative},
3174 @samp{multigot} or @samp{target}. For more information refer to the
3175 Info entry for @file{ld}.
3183 @subsection Options specific to MIPS targets
3185 @c man begin OPTIONS
3187 The following options are supported to control microMIPS instruction
3188 generation and branch relocation checks for ISA mode transitions when
3189 linking for MIPS targets.
3197 These options control the choice of microMIPS instructions used in code
3198 generated by the linker, such as that in the PLT or lazy binding stubs,
3199 or in relaxation. If @samp{--insn32} is used, then the linker only uses
3200 32-bit instruction encodings. By default or if @samp{--no-insn32} is
3201 used, all instruction encodings are used, including 16-bit ones where
3204 @kindex --ignore-branch-isa
3205 @item --ignore-branch-isa
3206 @kindex --no-ignore-branch-isa
3207 @itemx --no-ignore-branch-isa
3208 These options control branch relocation checks for invalid ISA mode
3209 transitions. If @samp{--ignore-branch-isa} is used, then the linker
3210 accepts any branch relocations and any ISA mode transition required
3211 is lost in relocation calculation, except for some cases of @code{BAL}
3212 instructions which meet relaxation conditions and are converted to
3213 equivalent @code{JALX} instructions as the associated relocation is
3214 calculated. By default or if @samp{--no-ignore-branch-isa} is used
3215 a check is made causing the loss of an ISA mode transition to produce
3218 @kindex --compact-branches
3219 @item --compact-branches
3220 @kindex --no-compact-branches
3221 @item --compact-branches
3222 These options control the generation of compact instructions by the linker
3223 in the PLT entries for MIPS R6.
3232 @section Environment Variables
3234 @c man begin ENVIRONMENT
3236 You can change the behaviour of @command{ld} with the environment variables
3237 @ifclear SingleFormat
3240 @code{LDEMULATION} and @code{COLLECT_NO_DEMANGLE}.
3242 @ifclear SingleFormat
3244 @cindex default input format
3245 @code{GNUTARGET} determines the input-file object format if you don't
3246 use @samp{-b} (or its synonym @samp{--format}). Its value should be one
3247 of the BFD names for an input format (@pxref{BFD}). If there is no
3248 @code{GNUTARGET} in the environment, @command{ld} uses the natural format
3249 of the target. If @code{GNUTARGET} is set to @code{default} then BFD
3250 attempts to discover the input format by examining binary input files;
3251 this method often succeeds, but there are potential ambiguities, since
3252 there is no method of ensuring that the magic number used to specify
3253 object-file formats is unique. However, the configuration procedure for
3254 BFD on each system places the conventional format for that system first
3255 in the search-list, so ambiguities are resolved in favor of convention.
3259 @cindex default emulation
3260 @cindex emulation, default
3261 @code{LDEMULATION} determines the default emulation if you don't use the
3262 @samp{-m} option. The emulation can affect various aspects of linker
3263 behaviour, particularly the default linker script. You can list the
3264 available emulations with the @samp{--verbose} or @samp{-V} options. If
3265 the @samp{-m} option is not used, and the @code{LDEMULATION} environment
3266 variable is not defined, the default emulation depends upon how the
3267 linker was configured.
3269 @kindex COLLECT_NO_DEMANGLE
3270 @cindex demangling, default
3271 Normally, the linker will default to demangling symbols. However, if
3272 @code{COLLECT_NO_DEMANGLE} is set in the environment, then it will
3273 default to not demangling symbols. This environment variable is used in
3274 a similar fashion by the @code{gcc} linker wrapper program. The default
3275 may be overridden by the @samp{--demangle} and @samp{--no-demangle}
3282 @chapter Linker Scripts
3285 @cindex linker scripts
3286 @cindex command files
3287 Every link is controlled by a @dfn{linker script}. This script is
3288 written in the linker command language.
3290 The main purpose of the linker script is to describe how the sections in
3291 the input files should be mapped into the output file, and to control
3292 the memory layout of the output file. Most linker scripts do nothing
3293 more than this. However, when necessary, the linker script can also
3294 direct the linker to perform many other operations, using the commands
3297 The linker always uses a linker script. If you do not supply one
3298 yourself, the linker will use a default script that is compiled into the
3299 linker executable. You can use the @samp{--verbose} command-line option
3300 to display the default linker script. Certain command-line options,
3301 such as @samp{-r} or @samp{-N}, will affect the default linker script.
3303 You may supply your own linker script by using the @samp{-T} command
3304 line option. When you do this, your linker script will replace the
3305 default linker script.
3307 You may also use linker scripts implicitly by naming them as input files
3308 to the linker, as though they were files to be linked. @xref{Implicit
3312 * Basic Script Concepts:: Basic Linker Script Concepts
3313 * Script Format:: Linker Script Format
3314 * Simple Example:: Simple Linker Script Example
3315 * Simple Commands:: Simple Linker Script Commands
3316 * Assignments:: Assigning Values to Symbols
3317 * SECTIONS:: SECTIONS Command
3318 * MEMORY:: MEMORY Command
3319 * PHDRS:: PHDRS Command
3320 * VERSION:: VERSION Command
3321 * Expressions:: Expressions in Linker Scripts
3322 * Implicit Linker Scripts:: Implicit Linker Scripts
3325 @node Basic Script Concepts
3326 @section Basic Linker Script Concepts
3327 @cindex linker script concepts
3328 We need to define some basic concepts and vocabulary in order to
3329 describe the linker script language.
3331 The linker combines input files into a single output file. The output
3332 file and each input file are in a special data format known as an
3333 @dfn{object file format}. Each file is called an @dfn{object file}.
3334 The output file is often called an @dfn{executable}, but for our
3335 purposes we will also call it an object file. Each object file has,
3336 among other things, a list of @dfn{sections}. We sometimes refer to a
3337 section in an input file as an @dfn{input section}; similarly, a section
3338 in the output file is an @dfn{output section}.
3340 Each section in an object file has a name and a size. Most sections
3341 also have an associated block of data, known as the @dfn{section
3342 contents}. A section may be marked as @dfn{loadable}, which means that
3343 the contents should be loaded into memory when the output file is run.
3344 A section with no contents may be @dfn{allocatable}, which means that an
3345 area in memory should be set aside, but nothing in particular should be
3346 loaded there (in some cases this memory must be zeroed out). A section
3347 which is neither loadable nor allocatable typically contains some sort
3348 of debugging information.
3350 Every loadable or allocatable output section has two addresses. The
3351 first is the @dfn{VMA}, or virtual memory address. This is the address
3352 the section will have when the output file is run. The second is the
3353 @dfn{LMA}, or load memory address. This is the address at which the
3354 section will be loaded. In most cases the two addresses will be the
3355 same. An example of when they might be different is when a data section
3356 is loaded into ROM, and then copied into RAM when the program starts up
3357 (this technique is often used to initialize global variables in a ROM
3358 based system). In this case the ROM address would be the LMA, and the
3359 RAM address would be the VMA.
3361 You can see the sections in an object file by using the @code{objdump}
3362 program with the @samp{-h} option.
3364 Every object file also has a list of @dfn{symbols}, known as the
3365 @dfn{symbol table}. A symbol may be defined or undefined. Each symbol
3366 has a name, and each defined symbol has an address, among other
3367 information. If you compile a C or C++ program into an object file, you
3368 will get a defined symbol for every defined function and global or
3369 static variable. Every undefined function or global variable which is
3370 referenced in the input file will become an undefined symbol.
3372 You can see the symbols in an object file by using the @code{nm}
3373 program, or by using the @code{objdump} program with the @samp{-t}
3377 @section Linker Script Format
3378 @cindex linker script format
3379 Linker scripts are text files.
3381 You write a linker script as a series of commands. Each command is
3382 either a keyword, possibly followed by arguments, or an assignment to a
3383 symbol. You may separate commands using semicolons. Whitespace is
3386 Strings such as file or format names can normally be entered directly.
3387 If the file name contains a character such as a comma which would
3388 otherwise serve to separate file names, you may put the file name in
3389 double quotes. There is no way to use a double quote character in a
3392 You may include comments in linker scripts just as in C, delimited by
3393 @samp{/*} and @samp{*/}. As in C, comments are syntactically equivalent
3396 @node Simple Example
3397 @section Simple Linker Script Example
3398 @cindex linker script example
3399 @cindex example of linker script
3400 Many linker scripts are fairly simple.
3402 The simplest possible linker script has just one command:
3403 @samp{SECTIONS}. You use the @samp{SECTIONS} command to describe the
3404 memory layout of the output file.
3406 The @samp{SECTIONS} command is a powerful command. Here we will
3407 describe a simple use of it. Let's assume your program consists only of
3408 code, initialized data, and uninitialized data. These will be in the
3409 @samp{.text}, @samp{.data}, and @samp{.bss} sections, respectively.
3410 Let's assume further that these are the only sections which appear in
3413 For this example, let's say that the code should be loaded at address
3414 0x10000, and that the data should start at address 0x8000000. Here is a
3415 linker script which will do that:
3420 .text : @{ *(.text) @}
3422 .data : @{ *(.data) @}
3423 .bss : @{ *(.bss) @}
3427 You write the @samp{SECTIONS} command as the keyword @samp{SECTIONS},
3428 followed by a series of symbol assignments and output section
3429 descriptions enclosed in curly braces.
3431 The first line inside the @samp{SECTIONS} command of the above example
3432 sets the value of the special symbol @samp{.}, which is the location
3433 counter. If you do not specify the address of an output section in some
3434 other way (other ways are described later), the address is set from the
3435 current value of the location counter. The location counter is then
3436 incremented by the size of the output section. At the start of the
3437 @samp{SECTIONS} command, the location counter has the value @samp{0}.
3439 The second line defines an output section, @samp{.text}. The colon is
3440 required syntax which may be ignored for now. Within the curly braces
3441 after the output section name, you list the names of the input sections
3442 which should be placed into this output section. The @samp{*} is a
3443 wildcard which matches any file name. The expression @samp{*(.text)}
3444 means all @samp{.text} input sections in all input files.
3446 Since the location counter is @samp{0x10000} when the output section
3447 @samp{.text} is defined, the linker will set the address of the
3448 @samp{.text} section in the output file to be @samp{0x10000}.
3450 The remaining lines define the @samp{.data} and @samp{.bss} sections in
3451 the output file. The linker will place the @samp{.data} output section
3452 at address @samp{0x8000000}. After the linker places the @samp{.data}
3453 output section, the value of the location counter will be
3454 @samp{0x8000000} plus the size of the @samp{.data} output section. The
3455 effect is that the linker will place the @samp{.bss} output section
3456 immediately after the @samp{.data} output section in memory.
3458 The linker will ensure that each output section has the required
3459 alignment, by increasing the location counter if necessary. In this
3460 example, the specified addresses for the @samp{.text} and @samp{.data}
3461 sections will probably satisfy any alignment constraints, but the linker
3462 may have to create a small gap between the @samp{.data} and @samp{.bss}
3465 That's it! That's a simple and complete linker script.
3467 @node Simple Commands
3468 @section Simple Linker Script Commands
3469 @cindex linker script simple commands
3470 In this section we describe the simple linker script commands.
3473 * Entry Point:: Setting the entry point
3474 * File Commands:: Commands dealing with files
3475 @ifclear SingleFormat
3476 * Format Commands:: Commands dealing with object file formats
3479 * REGION_ALIAS:: Assign alias names to memory regions
3480 * Miscellaneous Commands:: Other linker script commands
3484 @subsection Setting the Entry Point
3485 @kindex ENTRY(@var{symbol})
3486 @cindex start of execution
3487 @cindex first instruction
3489 The first instruction to execute in a program is called the @dfn{entry
3490 point}. You can use the @code{ENTRY} linker script command to set the
3491 entry point. The argument is a symbol name:
3496 There are several ways to set the entry point. The linker will set the
3497 entry point by trying each of the following methods in order, and
3498 stopping when one of them succeeds:
3501 the @samp{-e} @var{entry} command-line option;
3503 the @code{ENTRY(@var{symbol})} command in a linker script;
3505 the value of a target-specific symbol, if it is defined; For many
3506 targets this is @code{start}, but PE- and BeOS-based systems for example
3507 check a list of possible entry symbols, matching the first one found.
3509 the address of the first byte of the @samp{.text} section, if present;
3511 The address @code{0}.
3515 @subsection Commands Dealing with Files
3516 @cindex linker script file commands
3517 Several linker script commands deal with files.
3520 @item INCLUDE @var{filename}
3521 @kindex INCLUDE @var{filename}
3522 @cindex including a linker script
3523 Include the linker script @var{filename} at this point. The file will
3524 be searched for in the current directory, and in any directory specified
3525 with the @option{-L} option. You can nest calls to @code{INCLUDE} up to
3528 You can place @code{INCLUDE} directives at the top level, in @code{MEMORY} or
3529 @code{SECTIONS} commands, or in output section descriptions.
3531 @item INPUT(@var{file}, @var{file}, @dots{})
3532 @itemx INPUT(@var{file} @var{file} @dots{})
3533 @kindex INPUT(@var{files})
3534 @cindex input files in linker scripts
3535 @cindex input object files in linker scripts
3536 @cindex linker script input object files
3537 The @code{INPUT} command directs the linker to include the named files
3538 in the link, as though they were named on the command line.
3540 For example, if you always want to include @file{subr.o} any time you do
3541 a link, but you can't be bothered to put it on every link command line,
3542 then you can put @samp{INPUT (subr.o)} in your linker script.
3544 In fact, if you like, you can list all of your input files in the linker
3545 script, and then invoke the linker with nothing but a @samp{-T} option.
3547 In case a @dfn{sysroot prefix} is configured, and the filename starts
3548 with the @samp{/} character, and the script being processed was
3549 located inside the @dfn{sysroot prefix}, the filename will be looked
3550 for in the @dfn{sysroot prefix}. Otherwise, the linker will try to
3551 open the file in the current directory. If it is not found, the
3552 linker will search through the archive library search path.
3553 The @dfn{sysroot prefix} can also be forced by specifying @code{=}
3554 as the first character in the filename path, or prefixing the filename
3555 path with @code{$SYSROOT}. See also the description of @samp{-L} in
3556 @ref{Options,,Command-line Options}.
3558 If you use @samp{INPUT (-l@var{file})}, @command{ld} will transform the
3559 name to @code{lib@var{file}.a}, as with the command-line argument
3562 When you use the @code{INPUT} command in an implicit linker script, the
3563 files will be included in the link at the point at which the linker
3564 script file is included. This can affect archive searching.
3566 @item GROUP(@var{file}, @var{file}, @dots{})
3567 @itemx GROUP(@var{file} @var{file} @dots{})
3568 @kindex GROUP(@var{files})
3569 @cindex grouping input files
3570 The @code{GROUP} command is like @code{INPUT}, except that the named
3571 files should all be archives, and they are searched repeatedly until no
3572 new undefined references are created. See the description of @samp{-(}
3573 in @ref{Options,,Command-line Options}.
3575 @item AS_NEEDED(@var{file}, @var{file}, @dots{})
3576 @itemx AS_NEEDED(@var{file} @var{file} @dots{})
3577 @kindex AS_NEEDED(@var{files})
3578 This construct can appear only inside of the @code{INPUT} or @code{GROUP}
3579 commands, among other filenames. The files listed will be handled
3580 as if they appear directly in the @code{INPUT} or @code{GROUP} commands,
3581 with the exception of ELF shared libraries, that will be added only
3582 when they are actually needed. This construct essentially enables
3583 @option{--as-needed} option for all the files listed inside of it
3584 and restores previous @option{--as-needed} resp. @option{--no-as-needed}
3587 @item OUTPUT(@var{filename})
3588 @kindex OUTPUT(@var{filename})
3589 @cindex output file name in linker script
3590 The @code{OUTPUT} command names the output file. Using
3591 @code{OUTPUT(@var{filename})} in the linker script is exactly like using
3592 @samp{-o @var{filename}} on the command line (@pxref{Options,,Command
3593 Line Options}). If both are used, the command-line option takes
3596 You can use the @code{OUTPUT} command to define a default name for the
3597 output file other than the usual default of @file{a.out}.
3599 @item SEARCH_DIR(@var{path})
3600 @kindex SEARCH_DIR(@var{path})
3601 @cindex library search path in linker script
3602 @cindex archive search path in linker script
3603 @cindex search path in linker script
3604 The @code{SEARCH_DIR} command adds @var{path} to the list of paths where
3605 @command{ld} looks for archive libraries. Using
3606 @code{SEARCH_DIR(@var{path})} is exactly like using @samp{-L @var{path}}
3607 on the command line (@pxref{Options,,Command-line Options}). If both
3608 are used, then the linker will search both paths. Paths specified using
3609 the command-line option are searched first.
3611 @item STARTUP(@var{filename})
3612 @kindex STARTUP(@var{filename})
3613 @cindex first input file
3614 The @code{STARTUP} command is just like the @code{INPUT} command, except
3615 that @var{filename} will become the first input file to be linked, as
3616 though it were specified first on the command line. This may be useful
3617 when using a system in which the entry point is always the start of the
3621 @ifclear SingleFormat
3622 @node Format Commands
3623 @subsection Commands Dealing with Object File Formats
3624 A couple of linker script commands deal with object file formats.
3627 @item OUTPUT_FORMAT(@var{bfdname})
3628 @itemx OUTPUT_FORMAT(@var{default}, @var{big}, @var{little})
3629 @kindex OUTPUT_FORMAT(@var{bfdname})
3630 @cindex output file format in linker script
3631 The @code{OUTPUT_FORMAT} command names the BFD format to use for the
3632 output file (@pxref{BFD}). Using @code{OUTPUT_FORMAT(@var{bfdname})} is
3633 exactly like using @samp{--oformat @var{bfdname}} on the command line
3634 (@pxref{Options,,Command-line Options}). If both are used, the command
3635 line option takes precedence.
3637 You can use @code{OUTPUT_FORMAT} with three arguments to use different
3638 formats based on the @samp{-EB} and @samp{-EL} command-line options.
3639 This permits the linker script to set the output format based on the
3642 If neither @samp{-EB} nor @samp{-EL} are used, then the output format
3643 will be the first argument, @var{default}. If @samp{-EB} is used, the
3644 output format will be the second argument, @var{big}. If @samp{-EL} is
3645 used, the output format will be the third argument, @var{little}.
3647 For example, the default linker script for the MIPS ELF target uses this
3650 OUTPUT_FORMAT(elf32-bigmips, elf32-bigmips, elf32-littlemips)
3652 This says that the default format for the output file is
3653 @samp{elf32-bigmips}, but if the user uses the @samp{-EL} command-line
3654 option, the output file will be created in the @samp{elf32-littlemips}
3657 @item TARGET(@var{bfdname})
3658 @kindex TARGET(@var{bfdname})
3659 @cindex input file format in linker script
3660 The @code{TARGET} command names the BFD format to use when reading input
3661 files. It affects subsequent @code{INPUT} and @code{GROUP} commands.
3662 This command is like using @samp{-b @var{bfdname}} on the command line
3663 (@pxref{Options,,Command-line Options}). If the @code{TARGET} command
3664 is used but @code{OUTPUT_FORMAT} is not, then the last @code{TARGET}
3665 command is also used to set the format for the output file. @xref{BFD}.
3670 @subsection Assign alias names to memory regions
3671 @kindex REGION_ALIAS(@var{alias}, @var{region})
3672 @cindex region alias
3673 @cindex region names
3675 Alias names can be added to existing memory regions created with the
3676 @ref{MEMORY} command. Each name corresponds to at most one memory region.
3679 REGION_ALIAS(@var{alias}, @var{region})
3682 The @code{REGION_ALIAS} function creates an alias name @var{alias} for the
3683 memory region @var{region}. This allows a flexible mapping of output sections
3684 to memory regions. An example follows.
3686 Suppose we have an application for embedded systems which come with various
3687 memory storage devices. All have a general purpose, volatile memory @code{RAM}
3688 that allows code execution or data storage. Some may have a read-only,
3689 non-volatile memory @code{ROM} that allows code execution and read-only data
3690 access. The last variant is a read-only, non-volatile memory @code{ROM2} with
3691 read-only data access and no code execution capability. We have four output
3696 @code{.text} program code;
3698 @code{.rodata} read-only data;
3700 @code{.data} read-write initialized data;
3702 @code{.bss} read-write zero initialized data.
3705 The goal is to provide a linker command file that contains a system independent
3706 part defining the output sections and a system dependent part mapping the
3707 output sections to the memory regions available on the system. Our embedded
3708 systems come with three different memory setups @code{A}, @code{B} and
3710 @multitable @columnfractions .25 .25 .25 .25
3711 @item Section @tab Variant A @tab Variant B @tab Variant C
3712 @item .text @tab RAM @tab ROM @tab ROM
3713 @item .rodata @tab RAM @tab ROM @tab ROM2
3714 @item .data @tab RAM @tab RAM/ROM @tab RAM/ROM2
3715 @item .bss @tab RAM @tab RAM @tab RAM
3717 The notation @code{RAM/ROM} or @code{RAM/ROM2} means that this section is
3718 loaded into region @code{ROM} or @code{ROM2} respectively. Please note that
3719 the load address of the @code{.data} section starts in all three variants at
3720 the end of the @code{.rodata} section.
3722 The base linker script that deals with the output sections follows. It
3723 includes the system dependent @code{linkcmds.memory} file that describes the
3726 INCLUDE linkcmds.memory
3739 .data : AT (rodata_end)
3744 data_size = SIZEOF(.data);
3745 data_load_start = LOADADDR(.data);
3753 Now we need three different @code{linkcmds.memory} files to define memory
3754 regions and alias names. The content of @code{linkcmds.memory} for the three
3755 variants @code{A}, @code{B} and @code{C}:
3758 Here everything goes into the @code{RAM}.
3762 RAM : ORIGIN = 0, LENGTH = 4M
3765 REGION_ALIAS("REGION_TEXT", RAM);
3766 REGION_ALIAS("REGION_RODATA", RAM);
3767 REGION_ALIAS("REGION_DATA", RAM);
3768 REGION_ALIAS("REGION_BSS", RAM);
3771 Program code and read-only data go into the @code{ROM}. Read-write data goes
3772 into the @code{RAM}. An image of the initialized data is loaded into the
3773 @code{ROM} and will be copied during system start into the @code{RAM}.
3777 ROM : ORIGIN = 0, LENGTH = 3M
3778 RAM : ORIGIN = 0x10000000, LENGTH = 1M
3781 REGION_ALIAS("REGION_TEXT", ROM);
3782 REGION_ALIAS("REGION_RODATA", ROM);
3783 REGION_ALIAS("REGION_DATA", RAM);
3784 REGION_ALIAS("REGION_BSS", RAM);
3787 Program code goes into the @code{ROM}. Read-only data goes into the
3788 @code{ROM2}. Read-write data goes into the @code{RAM}. An image of the
3789 initialized data is loaded into the @code{ROM2} and will be copied during
3790 system start into the @code{RAM}.
3794 ROM : ORIGIN = 0, LENGTH = 2M
3795 ROM2 : ORIGIN = 0x10000000, LENGTH = 1M
3796 RAM : ORIGIN = 0x20000000, LENGTH = 1M
3799 REGION_ALIAS("REGION_TEXT", ROM);
3800 REGION_ALIAS("REGION_RODATA", ROM2);
3801 REGION_ALIAS("REGION_DATA", RAM);
3802 REGION_ALIAS("REGION_BSS", RAM);
3806 It is possible to write a common system initialization routine to copy the
3807 @code{.data} section from @code{ROM} or @code{ROM2} into the @code{RAM} if
3812 extern char data_start [];
3813 extern char data_size [];
3814 extern char data_load_start [];
3816 void copy_data(void)
3818 if (data_start != data_load_start)
3820 memcpy(data_start, data_load_start, (size_t) data_size);
3825 @node Miscellaneous Commands
3826 @subsection Other Linker Script Commands
3827 There are a few other linker scripts commands.
3830 @item ASSERT(@var{exp}, @var{message})
3832 @cindex assertion in linker script
3833 Ensure that @var{exp} is non-zero. If it is zero, then exit the linker
3834 with an error code, and print @var{message}.
3836 Note that assertions are checked before the final stages of linking
3837 take place. This means that expressions involving symbols PROVIDEd
3838 inside section definitions will fail if the user has not set values
3839 for those symbols. The only exception to this rule is PROVIDEd
3840 symbols that just reference dot. Thus an assertion like this:
3845 PROVIDE (__stack = .);
3846 PROVIDE (__stack_size = 0x100);
3847 ASSERT ((__stack > (_end + __stack_size)), "Error: No room left for the stack");
3851 will fail if @code{__stack_size} is not defined elsewhere. Symbols
3852 PROVIDEd outside of section definitions are evaluated earlier, so they
3853 can be used inside ASSERTions. Thus:
3856 PROVIDE (__stack_size = 0x100);
3859 PROVIDE (__stack = .);
3860 ASSERT ((__stack > (_end + __stack_size)), "Error: No room left for the stack");
3866 @item EXTERN(@var{symbol} @var{symbol} @dots{})
3868 @cindex undefined symbol in linker script
3869 Force @var{symbol} to be entered in the output file as an undefined
3870 symbol. Doing this may, for example, trigger linking of additional
3871 modules from standard libraries. You may list several @var{symbol}s for
3872 each @code{EXTERN}, and you may use @code{EXTERN} multiple times. This
3873 command has the same effect as the @samp{-u} command-line option.
3875 @item FORCE_COMMON_ALLOCATION
3876 @kindex FORCE_COMMON_ALLOCATION
3877 @cindex common allocation in linker script
3878 This command has the same effect as the @samp{-d} command-line option:
3879 to make @command{ld} assign space to common symbols even if a relocatable
3880 output file is specified (@samp{-r}).
3882 @item INHIBIT_COMMON_ALLOCATION
3883 @kindex INHIBIT_COMMON_ALLOCATION
3884 @cindex common allocation in linker script
3885 This command has the same effect as the @samp{--no-define-common}
3886 command-line option: to make @code{ld} omit the assignment of addresses
3887 to common symbols even for a non-relocatable output file.
3889 @item FORCE_GROUP_ALLOCATION
3890 @kindex FORCE_GROUP_ALLOCATION
3891 @cindex group allocation in linker script
3892 @cindex section groups
3894 This command has the same effect as the
3895 @samp{--force-group-allocation} command-line option: to make
3896 @command{ld} place section group members like normal input sections,
3897 and to delete the section groups even if a relocatable output file is
3898 specified (@samp{-r}).
3900 @item INSERT [ AFTER | BEFORE ] @var{output_section}
3902 @cindex insert user script into default script
3903 This command is typically used in a script specified by @samp{-T} to
3904 augment the default @code{SECTIONS} with, for example, overlays. It
3905 inserts all prior linker script statements after (or before)
3906 @var{output_section}, and also causes @samp{-T} to not override the
3907 default linker script. The exact insertion point is as for orphan
3908 sections. @xref{Location Counter}. The insertion happens after the
3909 linker has mapped input sections to output sections. Prior to the
3910 insertion, since @samp{-T} scripts are parsed before the default
3911 linker script, statements in the @samp{-T} script occur before the
3912 default linker script statements in the internal linker representation
3913 of the script. In particular, input section assignments will be made
3914 to @samp{-T} output sections before those in the default script. Here
3915 is an example of how a @samp{-T} script using @code{INSERT} might look:
3922 .ov1 @{ ov1*(.text) @}
3923 .ov2 @{ ov2*(.text) @}
3929 @item NOCROSSREFS(@var{section} @var{section} @dots{})
3930 @kindex NOCROSSREFS(@var{sections})
3931 @cindex cross references
3932 This command may be used to tell @command{ld} to issue an error about any
3933 references among certain output sections.
3935 In certain types of programs, particularly on embedded systems when
3936 using overlays, when one section is loaded into memory, another section
3937 will not be. Any direct references between the two sections would be
3938 errors. For example, it would be an error if code in one section called
3939 a function defined in the other section.
3941 The @code{NOCROSSREFS} command takes a list of output section names. If
3942 @command{ld} detects any cross references between the sections, it reports
3943 an error and returns a non-zero exit status. Note that the
3944 @code{NOCROSSREFS} command uses output section names, not input section
3947 @item NOCROSSREFS_TO(@var{tosection} @var{fromsection} @dots{})
3948 @kindex NOCROSSREFS_TO(@var{tosection} @var{fromsections})
3949 @cindex cross references
3950 This command may be used to tell @command{ld} to issue an error about any
3951 references to one section from a list of other sections.
3953 The @code{NOCROSSREFS} command is useful when ensuring that two or more
3954 output sections are entirely independent but there are situations where
3955 a one-way dependency is needed. For example, in a multi-core application
3956 there may be shared code that can be called from each core but for safety
3957 must never call back.
3959 The @code{NOCROSSREFS_TO} command takes a list of output section names.
3960 The first section can not be referenced from any of the other sections.
3961 If @command{ld} detects any references to the first section from any of
3962 the other sections, it reports an error and returns a non-zero exit
3963 status. Note that the @code{NOCROSSREFS_TO} command uses output section
3964 names, not input section names.
3966 @ifclear SingleFormat
3967 @item OUTPUT_ARCH(@var{bfdarch})
3968 @kindex OUTPUT_ARCH(@var{bfdarch})
3969 @cindex machine architecture
3970 @cindex architecture
3971 Specify a particular output machine architecture. The argument is one
3972 of the names used by the BFD library (@pxref{BFD}). You can see the
3973 architecture of an object file by using the @code{objdump} program with
3974 the @samp{-f} option.
3977 @item LD_FEATURE(@var{string})
3978 @kindex LD_FEATURE(@var{string})
3979 This command may be used to modify @command{ld} behavior. If
3980 @var{string} is @code{"SANE_EXPR"} then absolute symbols and numbers
3981 in a script are simply treated as numbers everywhere.
3982 @xref{Expression Section}.
3986 @section Assigning Values to Symbols
3987 @cindex assignment in scripts
3988 @cindex symbol definition, scripts
3989 @cindex variables, defining
3990 You may assign a value to a symbol in a linker script. This will define
3991 the symbol and place it into the symbol table with a global scope.
3994 * Simple Assignments:: Simple Assignments
3997 * PROVIDE_HIDDEN:: PROVIDE_HIDDEN
3998 * Source Code Reference:: How to use a linker script defined symbol in source code
4001 @node Simple Assignments
4002 @subsection Simple Assignments
4004 You may assign to a symbol using any of the C assignment operators:
4007 @item @var{symbol} = @var{expression} ;
4008 @itemx @var{symbol} += @var{expression} ;
4009 @itemx @var{symbol} -= @var{expression} ;
4010 @itemx @var{symbol} *= @var{expression} ;
4011 @itemx @var{symbol} /= @var{expression} ;
4012 @itemx @var{symbol} <<= @var{expression} ;
4013 @itemx @var{symbol} >>= @var{expression} ;
4014 @itemx @var{symbol} &= @var{expression} ;
4015 @itemx @var{symbol} |= @var{expression} ;
4018 The first case will define @var{symbol} to the value of
4019 @var{expression}. In the other cases, @var{symbol} must already be
4020 defined, and the value will be adjusted accordingly.
4022 The special symbol name @samp{.} indicates the location counter. You
4023 may only use this within a @code{SECTIONS} command. @xref{Location Counter}.
4025 The semicolon after @var{expression} is required.
4027 Expressions are defined below; see @ref{Expressions}.
4029 You may write symbol assignments as commands in their own right, or as
4030 statements within a @code{SECTIONS} command, or as part of an output
4031 section description in a @code{SECTIONS} command.
4033 The section of the symbol will be set from the section of the
4034 expression; for more information, see @ref{Expression Section}.
4036 Here is an example showing the three different places that symbol
4037 assignments may be used:
4048 _bdata = (. + 3) & ~ 3;
4049 .data : @{ *(.data) @}
4053 In this example, the symbol @samp{floating_point} will be defined as
4054 zero. The symbol @samp{_etext} will be defined as the address following
4055 the last @samp{.text} input section. The symbol @samp{_bdata} will be
4056 defined as the address following the @samp{.text} output section aligned
4057 upward to a 4 byte boundary.
4062 For ELF targeted ports, define a symbol that will be hidden and won't be
4063 exported. The syntax is @code{HIDDEN(@var{symbol} = @var{expression})}.
4065 Here is the example from @ref{Simple Assignments}, rewritten to use
4069 HIDDEN(floating_point = 0);
4077 HIDDEN(_bdata = (. + 3) & ~ 3);
4078 .data : @{ *(.data) @}
4082 In this case none of the three symbols will be visible outside this module.
4087 In some cases, it is desirable for a linker script to define a symbol
4088 only if it is referenced and is not defined by any object included in
4089 the link. For example, traditional linkers defined the symbol
4090 @samp{etext}. However, ANSI C requires that the user be able to use
4091 @samp{etext} as a function name without encountering an error. The
4092 @code{PROVIDE} keyword may be used to define a symbol, such as
4093 @samp{etext}, only if it is referenced but not defined. The syntax is
4094 @code{PROVIDE(@var{symbol} = @var{expression})}.
4096 Here is an example of using @code{PROVIDE} to define @samp{etext}:
4109 In this example, if the program defines @samp{_etext} (with a leading
4110 underscore), the linker will give a multiple definition error. If, on
4111 the other hand, the program defines @samp{etext} (with no leading
4112 underscore), the linker will silently use the definition in the program.
4113 If the program references @samp{etext} but does not define it, the
4114 linker will use the definition in the linker script.
4116 Note - the @code{PROVIDE} directive considers a common symbol to be
4117 defined, even though such a symbol could be combined with the symbol
4118 that the @code{PROVIDE} would create. This is particularly important
4119 when considering constructor and destructor list symbols such as
4120 @samp{__CTOR_LIST__} as these are often defined as common symbols.
4122 @node PROVIDE_HIDDEN
4123 @subsection PROVIDE_HIDDEN
4124 @cindex PROVIDE_HIDDEN
4125 Similar to @code{PROVIDE}. For ELF targeted ports, the symbol will be
4126 hidden and won't be exported.
4128 @node Source Code Reference
4129 @subsection Source Code Reference
4131 Accessing a linker script defined variable from source code is not
4132 intuitive. In particular a linker script symbol is not equivalent to
4133 a variable declaration in a high level language, it is instead a
4134 symbol that does not have a value.
4136 Before going further, it is important to note that compilers often
4137 transform names in the source code into different names when they are
4138 stored in the symbol table. For example, Fortran compilers commonly
4139 prepend or append an underscore, and C++ performs extensive @samp{name
4140 mangling}. Therefore there might be a discrepancy between the name
4141 of a variable as it is used in source code and the name of the same
4142 variable as it is defined in a linker script. For example in C a
4143 linker script variable might be referred to as:
4149 But in the linker script it might be defined as:
4155 In the remaining examples however it is assumed that no name
4156 transformation has taken place.
4158 When a symbol is declared in a high level language such as C, two
4159 things happen. The first is that the compiler reserves enough space
4160 in the program's memory to hold the @emph{value} of the symbol. The
4161 second is that the compiler creates an entry in the program's symbol
4162 table which holds the symbol's @emph{address}. ie the symbol table
4163 contains the address of the block of memory holding the symbol's
4164 value. So for example the following C declaration, at file scope:
4170 creates an entry called @samp{foo} in the symbol table. This entry
4171 holds the address of an @samp{int} sized block of memory where the
4172 number 1000 is initially stored.
4174 When a program references a symbol the compiler generates code that
4175 first accesses the symbol table to find the address of the symbol's
4176 memory block and then code to read the value from that memory block.
4183 looks up the symbol @samp{foo} in the symbol table, gets the address
4184 associated with this symbol and then writes the value 1 into that
4191 looks up the symbol @samp{foo} in the symbol table, gets its address
4192 and then copies this address into the block of memory associated with
4193 the variable @samp{a}.
4195 Linker scripts symbol declarations, by contrast, create an entry in
4196 the symbol table but do not assign any memory to them. Thus they are
4197 an address without a value. So for example the linker script definition:
4203 creates an entry in the symbol table called @samp{foo} which holds
4204 the address of memory location 1000, but nothing special is stored at
4205 address 1000. This means that you cannot access the @emph{value} of a
4206 linker script defined symbol - it has no value - all you can do is
4207 access the @emph{address} of a linker script defined symbol.
4209 Hence when you are using a linker script defined symbol in source code
4210 you should always take the address of the symbol, and never attempt to
4211 use its value. For example suppose you want to copy the contents of a
4212 section of memory called .ROM into a section called .FLASH and the
4213 linker script contains these declarations:
4217 start_of_ROM = .ROM;
4218 end_of_ROM = .ROM + sizeof (.ROM);
4219 start_of_FLASH = .FLASH;
4223 Then the C source code to perform the copy would be:
4227 extern char start_of_ROM, end_of_ROM, start_of_FLASH;
4229 memcpy (& start_of_FLASH, & start_of_ROM, & end_of_ROM - & start_of_ROM);
4233 Note the use of the @samp{&} operators. These are correct.
4234 Alternatively the symbols can be treated as the names of vectors or
4235 arrays and then the code will again work as expected:
4239 extern char start_of_ROM[], end_of_ROM[], start_of_FLASH[];
4241 memcpy (start_of_FLASH, start_of_ROM, end_of_ROM - start_of_ROM);
4245 Note how using this method does not require the use of @samp{&}
4249 @section SECTIONS Command
4251 The @code{SECTIONS} command tells the linker how to map input sections
4252 into output sections, and how to place the output sections in memory.
4254 The format of the @code{SECTIONS} command is:
4258 @var{sections-command}
4259 @var{sections-command}
4264 Each @var{sections-command} may of be one of the following:
4268 an @code{ENTRY} command (@pxref{Entry Point,,Entry command})
4270 a symbol assignment (@pxref{Assignments})
4272 an output section description
4274 an overlay description
4277 The @code{ENTRY} command and symbol assignments are permitted inside the
4278 @code{SECTIONS} command for convenience in using the location counter in
4279 those commands. This can also make the linker script easier to
4280 understand because you can use those commands at meaningful points in
4281 the layout of the output file.
4283 Output section descriptions and overlay descriptions are described
4286 If you do not use a @code{SECTIONS} command in your linker script, the
4287 linker will place each input section into an identically named output
4288 section in the order that the sections are first encountered in the
4289 input files. If all input sections are present in the first file, for
4290 example, the order of sections in the output file will match the order
4291 in the first input file. The first section will be at address zero.
4294 * Output Section Description:: Output section description
4295 * Output Section Name:: Output section name
4296 * Output Section Address:: Output section address
4297 * Input Section:: Input section description
4298 * Output Section Data:: Output section data
4299 * Output Section Keywords:: Output section keywords
4300 * Output Section Discarding:: Output section discarding
4301 * Output Section Attributes:: Output section attributes
4302 * Overlay Description:: Overlay description
4305 @node Output Section Description
4306 @subsection Output Section Description
4307 The full description of an output section looks like this:
4310 @var{section} [@var{address}] [(@var{type})] :
4312 [ALIGN(@var{section_align}) | ALIGN_WITH_INPUT]
4313 [SUBALIGN(@var{subsection_align})]
4316 @var{output-section-command}
4317 @var{output-section-command}
4319 @} [>@var{region}] [AT>@var{lma_region}] [:@var{phdr} :@var{phdr} @dots{}] [=@var{fillexp}] [,]
4323 Most output sections do not use most of the optional section attributes.
4325 The whitespace around @var{section} is required, so that the section
4326 name is unambiguous. The colon and the curly braces are also required.
4327 The comma at the end may be required if a @var{fillexp} is used and
4328 the next @var{sections-command} looks like a continuation of the expression.
4329 The line breaks and other white space are optional.
4331 Each @var{output-section-command} may be one of the following:
4335 a symbol assignment (@pxref{Assignments})
4337 an input section description (@pxref{Input Section})
4339 data values to include directly (@pxref{Output Section Data})
4341 a special output section keyword (@pxref{Output Section Keywords})
4344 @node Output Section Name
4345 @subsection Output Section Name
4346 @cindex name, section
4347 @cindex section name
4348 The name of the output section is @var{section}. @var{section} must
4349 meet the constraints of your output format. In formats which only
4350 support a limited number of sections, such as @code{a.out}, the name
4351 must be one of the names supported by the format (@code{a.out}, for
4352 example, allows only @samp{.text}, @samp{.data} or @samp{.bss}). If the
4353 output format supports any number of sections, but with numbers and not
4354 names (as is the case for Oasys), the name should be supplied as a
4355 quoted numeric string. A section name may consist of any sequence of
4356 characters, but a name which contains any unusual characters such as
4357 commas must be quoted.
4359 The output section name @samp{/DISCARD/} is special; @ref{Output Section
4362 @node Output Section Address
4363 @subsection Output Section Address
4364 @cindex address, section
4365 @cindex section address
4366 The @var{address} is an expression for the VMA (the virtual memory
4367 address) of the output section. This address is optional, but if it
4368 is provided then the output address will be set exactly as specified.
4370 If the output address is not specified then one will be chosen for the
4371 section, based on the heuristic below. This address will be adjusted
4372 to fit the alignment requirement of the output section. The
4373 alignment requirement is the strictest alignment of any input section
4374 contained within the output section.
4376 The output section address heuristic is as follows:
4380 If an output memory @var{region} is set for the section then it
4381 is added to this region and its address will be the next free address
4385 If the MEMORY command has been used to create a list of memory
4386 regions then the first region which has attributes compatible with the
4387 section is selected to contain it. The section's output address will
4388 be the next free address in that region; @ref{MEMORY}.
4391 If no memory regions were specified, or none match the section then
4392 the output address will be based on the current value of the location
4400 .text . : @{ *(.text) @}
4407 .text : @{ *(.text) @}
4411 are subtly different. The first will set the address of the
4412 @samp{.text} output section to the current value of the location
4413 counter. The second will set it to the current value of the location
4414 counter aligned to the strictest alignment of any of the @samp{.text}
4417 The @var{address} may be an arbitrary expression; @ref{Expressions}.
4418 For example, if you want to align the section on a 0x10 byte boundary,
4419 so that the lowest four bits of the section address are zero, you could
4420 do something like this:
4422 .text ALIGN(0x10) : @{ *(.text) @}
4425 This works because @code{ALIGN} returns the current location counter
4426 aligned upward to the specified value.
4428 Specifying @var{address} for a section will change the value of the
4429 location counter, provided that the section is non-empty. (Empty
4430 sections are ignored).
4433 @subsection Input Section Description
4434 @cindex input sections
4435 @cindex mapping input sections to output sections
4436 The most common output section command is an input section description.
4438 The input section description is the most basic linker script operation.
4439 You use output sections to tell the linker how to lay out your program
4440 in memory. You use input section descriptions to tell the linker how to
4441 map the input files into your memory layout.
4444 * Input Section Basics:: Input section basics
4445 * Input Section Wildcards:: Input section wildcard patterns
4446 * Input Section Common:: Input section for common symbols
4447 * Input Section Keep:: Input section and garbage collection
4448 * Input Section Example:: Input section example
4451 @node Input Section Basics
4452 @subsubsection Input Section Basics
4453 @cindex input section basics
4454 An input section description consists of a file name optionally followed
4455 by a list of section names in parentheses.
4457 The file name and the section name may be wildcard patterns, which we
4458 describe further below (@pxref{Input Section Wildcards}).
4460 The most common input section description is to include all input
4461 sections with a particular name in the output section. For example, to
4462 include all input @samp{.text} sections, you would write:
4467 Here the @samp{*} is a wildcard which matches any file name. To exclude a list
4468 @cindex EXCLUDE_FILE
4469 of files from matching the file name wildcard, EXCLUDE_FILE may be used to
4470 match all files except the ones specified in the EXCLUDE_FILE list. For
4473 EXCLUDE_FILE (*crtend.o *otherfile.o) *(.ctors)
4476 will cause all .ctors sections from all files except @file{crtend.o}
4477 and @file{otherfile.o} to be included. The EXCLUDE_FILE can also be
4478 placed inside the section list, for example:
4480 *(EXCLUDE_FILE (*crtend.o *otherfile.o) .ctors)
4483 The result of this is identically to the previous example. Supporting
4484 two syntaxes for EXCLUDE_FILE is useful if the section list contains
4485 more than one section, as described below.
4487 There are two ways to include more than one section:
4493 The difference between these is the order in which the @samp{.text} and
4494 @samp{.rdata} input sections will appear in the output section. In the
4495 first example, they will be intermingled, appearing in the same order as
4496 they are found in the linker input. In the second example, all
4497 @samp{.text} input sections will appear first, followed by all
4498 @samp{.rdata} input sections.
4500 When using EXCLUDE_FILE with more than one section, if the exclusion
4501 is within the section list then the exclusion only applies to the
4502 immediately following section, for example:
4504 *(EXCLUDE_FILE (*somefile.o) .text .rdata)
4507 will cause all @samp{.text} sections from all files except
4508 @file{somefile.o} to be included, while all @samp{.rdata} sections
4509 from all files, including @file{somefile.o}, will be included. To
4510 exclude the @samp{.rdata} sections from @file{somefile.o} the example
4511 could be modified to:
4513 *(EXCLUDE_FILE (*somefile.o) .text EXCLUDE_FILE (*somefile.o) .rdata)
4516 Alternatively, placing the EXCLUDE_FILE outside of the section list,
4517 before the input file selection, will cause the exclusion to apply for
4518 all sections. Thus the previous example can be rewritten as:
4520 EXCLUDE_FILE (*somefile.o) *(.text .rdata)
4523 You can specify a file name to include sections from a particular file.
4524 You would do this if one or more of your files contain special data that
4525 needs to be at a particular location in memory. For example:
4530 To refine the sections that are included based on the section flags
4531 of an input section, INPUT_SECTION_FLAGS may be used.
4533 Here is a simple example for using Section header flags for ELF sections:
4538 .text : @{ INPUT_SECTION_FLAGS (SHF_MERGE & SHF_STRINGS) *(.text) @}
4539 .text2 : @{ INPUT_SECTION_FLAGS (!SHF_WRITE) *(.text) @}
4544 In this example, the output section @samp{.text} will be comprised of any
4545 input section matching the name *(.text) whose section header flags
4546 @code{SHF_MERGE} and @code{SHF_STRINGS} are set. The output section
4547 @samp{.text2} will be comprised of any input section matching the name *(.text)
4548 whose section header flag @code{SHF_WRITE} is clear.
4550 You can also specify files within archives by writing a pattern
4551 matching the archive, a colon, then the pattern matching the file,
4552 with no whitespace around the colon.
4556 matches file within archive
4558 matches the whole archive
4560 matches file but not one in an archive
4563 Either one or both of @samp{archive} and @samp{file} can contain shell
4564 wildcards. On DOS based file systems, the linker will assume that a
4565 single letter followed by a colon is a drive specifier, so
4566 @samp{c:myfile.o} is a simple file specification, not @samp{myfile.o}
4567 within an archive called @samp{c}. @samp{archive:file} filespecs may
4568 also be used within an @code{EXCLUDE_FILE} list, but may not appear in
4569 other linker script contexts. For instance, you cannot extract a file
4570 from an archive by using @samp{archive:file} in an @code{INPUT}
4573 If you use a file name without a list of sections, then all sections in
4574 the input file will be included in the output section. This is not
4575 commonly done, but it may by useful on occasion. For example:
4580 When you use a file name which is not an @samp{archive:file} specifier
4581 and does not contain any wild card
4582 characters, the linker will first see if you also specified the file
4583 name on the linker command line or in an @code{INPUT} command. If you
4584 did not, the linker will attempt to open the file as an input file, as
4585 though it appeared on the command line. Note that this differs from an
4586 @code{INPUT} command, because the linker will not search for the file in
4587 the archive search path.
4589 @node Input Section Wildcards
4590 @subsubsection Input Section Wildcard Patterns
4591 @cindex input section wildcards
4592 @cindex wildcard file name patterns
4593 @cindex file name wildcard patterns
4594 @cindex section name wildcard patterns
4595 In an input section description, either the file name or the section
4596 name or both may be wildcard patterns.
4598 The file name of @samp{*} seen in many examples is a simple wildcard
4599 pattern for the file name.
4601 The wildcard patterns are like those used by the Unix shell.
4605 matches any number of characters
4607 matches any single character
4609 matches a single instance of any of the @var{chars}; the @samp{-}
4610 character may be used to specify a range of characters, as in
4611 @samp{[a-z]} to match any lower case letter
4613 quotes the following character
4616 When a file name is matched with a wildcard, the wildcard characters
4617 will not match a @samp{/} character (used to separate directory names on
4618 Unix). A pattern consisting of a single @samp{*} character is an
4619 exception; it will always match any file name, whether it contains a
4620 @samp{/} or not. In a section name, the wildcard characters will match
4621 a @samp{/} character.
4623 File name wildcard patterns only match files which are explicitly
4624 specified on the command line or in an @code{INPUT} command. The linker
4625 does not search directories to expand wildcards.
4627 If a file name matches more than one wildcard pattern, or if a file name
4628 appears explicitly and is also matched by a wildcard pattern, the linker
4629 will use the first match in the linker script. For example, this
4630 sequence of input section descriptions is probably in error, because the
4631 @file{data.o} rule will not be used:
4633 .data : @{ *(.data) @}
4634 .data1 : @{ data.o(.data) @}
4637 @cindex SORT_BY_NAME
4638 Normally, the linker will place files and sections matched by wildcards
4639 in the order in which they are seen during the link. You can change
4640 this by using the @code{SORT_BY_NAME} keyword, which appears before a wildcard
4641 pattern in parentheses (e.g., @code{SORT_BY_NAME(.text*)}). When the
4642 @code{SORT_BY_NAME} keyword is used, the linker will sort the files or sections
4643 into ascending order by name before placing them in the output file.
4645 @cindex SORT_BY_ALIGNMENT
4646 @code{SORT_BY_ALIGNMENT} is similar to @code{SORT_BY_NAME}.
4647 @code{SORT_BY_ALIGNMENT} will sort sections into descending order of
4648 alignment before placing them in the output file. Placing larger
4649 alignments before smaller alignments can reduce the amount of padding
4652 @cindex SORT_BY_INIT_PRIORITY
4653 @code{SORT_BY_INIT_PRIORITY} is also similar to @code{SORT_BY_NAME}.
4654 @code{SORT_BY_INIT_PRIORITY} will sort sections into ascending
4655 numerical order of the GCC init_priority attribute encoded in the
4656 section name before placing them in the output file. In
4657 @code{.init_array.NNNNN} and @code{.fini_array.NNNNN}, @code{NNNNN} is
4658 the init_priority. In @code{.ctors.NNNNN} and @code{.dtors.NNNNN},
4659 @code{NNNNN} is 65535 minus the init_priority.
4662 @code{SORT} is an alias for @code{SORT_BY_NAME}.
4664 When there are nested section sorting commands in linker script, there
4665 can be at most 1 level of nesting for section sorting commands.
4669 @code{SORT_BY_NAME} (@code{SORT_BY_ALIGNMENT} (wildcard section pattern)).
4670 It will sort the input sections by name first, then by alignment if two
4671 sections have the same name.
4673 @code{SORT_BY_ALIGNMENT} (@code{SORT_BY_NAME} (wildcard section pattern)).
4674 It will sort the input sections by alignment first, then by name if two
4675 sections have the same alignment.
4677 @code{SORT_BY_NAME} (@code{SORT_BY_NAME} (wildcard section pattern)) is
4678 treated the same as @code{SORT_BY_NAME} (wildcard section pattern).
4680 @code{SORT_BY_ALIGNMENT} (@code{SORT_BY_ALIGNMENT} (wildcard section pattern))
4681 is treated the same as @code{SORT_BY_ALIGNMENT} (wildcard section pattern).
4683 All other nested section sorting commands are invalid.
4686 When both command-line section sorting option and linker script
4687 section sorting command are used, section sorting command always
4688 takes precedence over the command-line option.
4690 If the section sorting command in linker script isn't nested, the
4691 command-line option will make the section sorting command to be
4692 treated as nested sorting command.
4696 @code{SORT_BY_NAME} (wildcard section pattern ) with
4697 @option{--sort-sections alignment} is equivalent to
4698 @code{SORT_BY_NAME} (@code{SORT_BY_ALIGNMENT} (wildcard section pattern)).
4700 @code{SORT_BY_ALIGNMENT} (wildcard section pattern) with
4701 @option{--sort-section name} is equivalent to
4702 @code{SORT_BY_ALIGNMENT} (@code{SORT_BY_NAME} (wildcard section pattern)).
4705 If the section sorting command in linker script is nested, the
4706 command-line option will be ignored.
4709 @code{SORT_NONE} disables section sorting by ignoring the command-line
4710 section sorting option.
4712 If you ever get confused about where input sections are going, use the
4713 @samp{-M} linker option to generate a map file. The map file shows
4714 precisely how input sections are mapped to output sections.
4716 This example shows how wildcard patterns might be used to partition
4717 files. This linker script directs the linker to place all @samp{.text}
4718 sections in @samp{.text} and all @samp{.bss} sections in @samp{.bss}.
4719 The linker will place the @samp{.data} section from all files beginning
4720 with an upper case character in @samp{.DATA}; for all other files, the
4721 linker will place the @samp{.data} section in @samp{.data}.
4725 .text : @{ *(.text) @}
4726 .DATA : @{ [A-Z]*(.data) @}
4727 .data : @{ *(.data) @}
4728 .bss : @{ *(.bss) @}
4733 @node Input Section Common
4734 @subsubsection Input Section for Common Symbols
4735 @cindex common symbol placement
4736 @cindex uninitialized data placement
4737 A special notation is needed for common symbols, because in many object
4738 file formats common symbols do not have a particular input section. The
4739 linker treats common symbols as though they are in an input section
4740 named @samp{COMMON}.
4742 You may use file names with the @samp{COMMON} section just as with any
4743 other input sections. You can use this to place common symbols from a
4744 particular input file in one section while common symbols from other
4745 input files are placed in another section.
4747 In most cases, common symbols in input files will be placed in the
4748 @samp{.bss} section in the output file. For example:
4750 .bss @{ *(.bss) *(COMMON) @}
4753 @cindex scommon section
4754 @cindex small common symbols
4755 Some object file formats have more than one type of common symbol. For
4756 example, the MIPS ELF object file format distinguishes standard common
4757 symbols and small common symbols. In this case, the linker will use a
4758 different special section name for other types of common symbols. In
4759 the case of MIPS ELF, the linker uses @samp{COMMON} for standard common
4760 symbols and @samp{.scommon} for small common symbols. This permits you
4761 to map the different types of common symbols into memory at different
4765 You will sometimes see @samp{[COMMON]} in old linker scripts. This
4766 notation is now considered obsolete. It is equivalent to
4769 @node Input Section Keep
4770 @subsubsection Input Section and Garbage Collection
4772 @cindex garbage collection
4773 When link-time garbage collection is in use (@samp{--gc-sections}),
4774 it is often useful to mark sections that should not be eliminated.
4775 This is accomplished by surrounding an input section's wildcard entry
4776 with @code{KEEP()}, as in @code{KEEP(*(.init))} or
4777 @code{KEEP(SORT_BY_NAME(*)(.ctors))}.
4779 @node Input Section Example
4780 @subsubsection Input Section Example
4781 The following example is a complete linker script. It tells the linker
4782 to read all of the sections from file @file{all.o} and place them at the
4783 start of output section @samp{outputa} which starts at location
4784 @samp{0x10000}. All of section @samp{.input1} from file @file{foo.o}
4785 follows immediately, in the same output section. All of section
4786 @samp{.input2} from @file{foo.o} goes into output section
4787 @samp{outputb}, followed by section @samp{.input1} from @file{foo1.o}.
4788 All of the remaining @samp{.input1} and @samp{.input2} sections from any
4789 files are written to output section @samp{outputc}.
4817 If an output section's name is the same as the input section's name
4818 and is representable as a C identifier, then the linker will
4819 automatically @pxref{PROVIDE} two symbols: __start_SECNAME and
4820 __stop_SECNAME, where SECNAME is the name of the section. These
4821 indicate the start address and end address of the output section
4822 respectively. Note: most section names are not representable as
4823 C identifiers because they contain a @samp{.} character.
4825 @node Output Section Data
4826 @subsection Output Section Data
4828 @cindex section data
4829 @cindex output section data
4830 @kindex BYTE(@var{expression})
4831 @kindex SHORT(@var{expression})
4832 @kindex LONG(@var{expression})
4833 @kindex QUAD(@var{expression})
4834 @kindex SQUAD(@var{expression})
4835 You can include explicit bytes of data in an output section by using
4836 @code{BYTE}, @code{SHORT}, @code{LONG}, @code{QUAD}, or @code{SQUAD} as
4837 an output section command. Each keyword is followed by an expression in
4838 parentheses providing the value to store (@pxref{Expressions}). The
4839 value of the expression is stored at the current value of the location
4842 The @code{BYTE}, @code{SHORT}, @code{LONG}, and @code{QUAD} commands
4843 store one, two, four, and eight bytes (respectively). After storing the
4844 bytes, the location counter is incremented by the number of bytes
4847 For example, this will store the byte 1 followed by the four byte value
4848 of the symbol @samp{addr}:
4854 When using a 64 bit host or target, @code{QUAD} and @code{SQUAD} are the
4855 same; they both store an 8 byte, or 64 bit, value. When both host and
4856 target are 32 bits, an expression is computed as 32 bits. In this case
4857 @code{QUAD} stores a 32 bit value zero extended to 64 bits, and
4858 @code{SQUAD} stores a 32 bit value sign extended to 64 bits.
4860 If the object file format of the output file has an explicit endianness,
4861 which is the normal case, the value will be stored in that endianness.
4862 When the object file format does not have an explicit endianness, as is
4863 true of, for example, S-records, the value will be stored in the
4864 endianness of the first input object file.
4866 Note---these commands only work inside a section description and not
4867 between them, so the following will produce an error from the linker:
4869 SECTIONS @{@ .text : @{@ *(.text) @}@ LONG(1) .data : @{@ *(.data) @}@ @}@
4871 whereas this will work:
4873 SECTIONS @{@ .text : @{@ *(.text) ; LONG(1) @}@ .data : @{@ *(.data) @}@ @}@
4876 @kindex FILL(@var{expression})
4877 @cindex holes, filling
4878 @cindex unspecified memory
4879 You may use the @code{FILL} command to set the fill pattern for the
4880 current section. It is followed by an expression in parentheses. Any
4881 otherwise unspecified regions of memory within the section (for example,
4882 gaps left due to the required alignment of input sections) are filled
4883 with the value of the expression, repeated as
4884 necessary. A @code{FILL} statement covers memory locations after the
4885 point at which it occurs in the section definition; by including more
4886 than one @code{FILL} statement, you can have different fill patterns in
4887 different parts of an output section.
4889 This example shows how to fill unspecified regions of memory with the
4895 The @code{FILL} command is similar to the @samp{=@var{fillexp}} output
4896 section attribute, but it only affects the
4897 part of the section following the @code{FILL} command, rather than the
4898 entire section. If both are used, the @code{FILL} command takes
4899 precedence. @xref{Output Section Fill}, for details on the fill
4902 @node Output Section Keywords
4903 @subsection Output Section Keywords
4904 There are a couple of keywords which can appear as output section
4908 @kindex CREATE_OBJECT_SYMBOLS
4909 @cindex input filename symbols
4910 @cindex filename symbols
4911 @item CREATE_OBJECT_SYMBOLS
4912 The command tells the linker to create a symbol for each input file.
4913 The name of each symbol will be the name of the corresponding input
4914 file. The section of each symbol will be the output section in which
4915 the @code{CREATE_OBJECT_SYMBOLS} command appears.
4917 This is conventional for the a.out object file format. It is not
4918 normally used for any other object file format.
4920 @kindex CONSTRUCTORS
4921 @cindex C++ constructors, arranging in link
4922 @cindex constructors, arranging in link
4924 When linking using the a.out object file format, the linker uses an
4925 unusual set construct to support C++ global constructors and
4926 destructors. When linking object file formats which do not support
4927 arbitrary sections, such as ECOFF and XCOFF, the linker will
4928 automatically recognize C++ global constructors and destructors by name.
4929 For these object file formats, the @code{CONSTRUCTORS} command tells the
4930 linker to place constructor information in the output section where the
4931 @code{CONSTRUCTORS} command appears. The @code{CONSTRUCTORS} command is
4932 ignored for other object file formats.
4934 The symbol @w{@code{__CTOR_LIST__}} marks the start of the global
4935 constructors, and the symbol @w{@code{__CTOR_END__}} marks the end.
4936 Similarly, @w{@code{__DTOR_LIST__}} and @w{@code{__DTOR_END__}} mark
4937 the start and end of the global destructors. The
4938 first word in the list is the number of entries, followed by the address
4939 of each constructor or destructor, followed by a zero word. The
4940 compiler must arrange to actually run the code. For these object file
4941 formats @sc{gnu} C++ normally calls constructors from a subroutine
4942 @code{__main}; a call to @code{__main} is automatically inserted into
4943 the startup code for @code{main}. @sc{gnu} C++ normally runs
4944 destructors either by using @code{atexit}, or directly from the function
4947 For object file formats such as @code{COFF} or @code{ELF} which support
4948 arbitrary section names, @sc{gnu} C++ will normally arrange to put the
4949 addresses of global constructors and destructors into the @code{.ctors}
4950 and @code{.dtors} sections. Placing the following sequence into your
4951 linker script will build the sort of table which the @sc{gnu} C++
4952 runtime code expects to see.
4956 LONG((__CTOR_END__ - __CTOR_LIST__) / 4 - 2)
4961 LONG((__DTOR_END__ - __DTOR_LIST__) / 4 - 2)
4967 If you are using the @sc{gnu} C++ support for initialization priority,
4968 which provides some control over the order in which global constructors
4969 are run, you must sort the constructors at link time to ensure that they
4970 are executed in the correct order. When using the @code{CONSTRUCTORS}
4971 command, use @samp{SORT_BY_NAME(CONSTRUCTORS)} instead. When using the
4972 @code{.ctors} and @code{.dtors} sections, use @samp{*(SORT_BY_NAME(.ctors))} and
4973 @samp{*(SORT_BY_NAME(.dtors))} instead of just @samp{*(.ctors)} and
4976 Normally the compiler and linker will handle these issues automatically,
4977 and you will not need to concern yourself with them. However, you may
4978 need to consider this if you are using C++ and writing your own linker
4983 @node Output Section Discarding
4984 @subsection Output Section Discarding
4985 @cindex discarding sections
4986 @cindex sections, discarding
4987 @cindex removing sections
4988 The linker will not normally create output sections with no contents.
4989 This is for convenience when referring to input sections that may or
4990 may not be present in any of the input files. For example:
4992 .foo : @{ *(.foo) @}
4995 will only create a @samp{.foo} section in the output file if there is a
4996 @samp{.foo} section in at least one input file, and if the input
4997 sections are not all empty. Other link script directives that allocate
4998 space in an output section will also create the output section. So
4999 too will assignments to dot even if the assignment does not create
5000 space, except for @samp{. = 0}, @samp{. = . + 0}, @samp{. = sym},
5001 @samp{. = . + sym} and @samp{. = ALIGN (. != 0, expr, 1)} when
5002 @samp{sym} is an absolute symbol of value 0 defined in the script.
5003 This allows you to force output of an empty section with @samp{. = .}.
5005 The linker will ignore address assignments (@pxref{Output Section Address})
5006 on discarded output sections, except when the linker script defines
5007 symbols in the output section. In that case the linker will obey
5008 the address assignments, possibly advancing dot even though the
5009 section is discarded.
5012 The special output section name @samp{/DISCARD/} may be used to discard
5013 input sections. Any input sections which are assigned to an output
5014 section named @samp{/DISCARD/} are not included in the output file.
5016 Note, sections that match the @samp{/DISCARD/} output section will be
5017 discarded even if they are in an ELF section group which has other
5018 members which are not being discarded. This is deliberate.
5019 Discarding takes precedence over grouping.
5021 @node Output Section Attributes
5022 @subsection Output Section Attributes
5023 @cindex output section attributes
5024 We showed above that the full description of an output section looked
5029 @var{section} [@var{address}] [(@var{type})] :
5031 [ALIGN(@var{section_align}) | ALIGN_WITH_INPUT]
5032 [SUBALIGN(@var{subsection_align})]
5035 @var{output-section-command}
5036 @var{output-section-command}
5038 @} [>@var{region}] [AT>@var{lma_region}] [:@var{phdr} :@var{phdr} @dots{}] [=@var{fillexp}]
5042 We've already described @var{section}, @var{address}, and
5043 @var{output-section-command}. In this section we will describe the
5044 remaining section attributes.
5047 * Output Section Type:: Output section type
5048 * Output Section LMA:: Output section LMA
5049 * Forced Output Alignment:: Forced Output Alignment
5050 * Forced Input Alignment:: Forced Input Alignment
5051 * Output Section Constraint:: Output section constraint
5052 * Output Section Region:: Output section region
5053 * Output Section Phdr:: Output section phdr
5054 * Output Section Fill:: Output section fill
5057 @node Output Section Type
5058 @subsubsection Output Section Type
5059 Each output section may have a type. The type is a keyword in
5060 parentheses. The following types are defined:
5064 The section should be marked as not loadable, so that it will not be
5065 loaded into memory when the program is run.
5070 These type names are supported for backward compatibility, and are
5071 rarely used. They all have the same effect: the section should be
5072 marked as not allocatable, so that no memory is allocated for the
5073 section when the program is run.
5077 @cindex prevent unnecessary loading
5078 @cindex loading, preventing
5079 The linker normally sets the attributes of an output section based on
5080 the input sections which map into it. You can override this by using
5081 the section type. For example, in the script sample below, the
5082 @samp{ROM} section is addressed at memory location @samp{0} and does not
5083 need to be loaded when the program is run.
5087 ROM 0 (NOLOAD) : @{ @dots{} @}
5093 @node Output Section LMA
5094 @subsubsection Output Section LMA
5095 @kindex AT>@var{lma_region}
5096 @kindex AT(@var{lma})
5097 @cindex load address
5098 @cindex section load address
5099 Every section has a virtual address (VMA) and a load address (LMA); see
5100 @ref{Basic Script Concepts}. The virtual address is specified by the
5101 @pxref{Output Section Address} described earlier. The load address is
5102 specified by the @code{AT} or @code{AT>} keywords. Specifying a load
5103 address is optional.
5105 The @code{AT} keyword takes an expression as an argument. This
5106 specifies the exact load address of the section. The @code{AT>} keyword
5107 takes the name of a memory region as an argument. @xref{MEMORY}. The
5108 load address of the section is set to the next free address in the
5109 region, aligned to the section's alignment requirements.
5111 If neither @code{AT} nor @code{AT>} is specified for an allocatable
5112 section, the linker will use the following heuristic to determine the
5117 If the section has a specific VMA address, then this is used as
5118 the LMA address as well.
5121 If the section is not allocatable then its LMA is set to its VMA.
5124 Otherwise if a memory region can be found that is compatible
5125 with the current section, and this region contains at least one
5126 section, then the LMA is set so the difference between the
5127 VMA and LMA is the same as the difference between the VMA and LMA of
5128 the last section in the located region.
5131 If no memory regions have been declared then a default region
5132 that covers the entire address space is used in the previous step.
5135 If no suitable region could be found, or there was no previous
5136 section then the LMA is set equal to the VMA.
5139 @cindex ROM initialized data
5140 @cindex initialized data in ROM
5141 This feature is designed to make it easy to build a ROM image. For
5142 example, the following linker script creates three output sections: one
5143 called @samp{.text}, which starts at @code{0x1000}, one called
5144 @samp{.mdata}, which is loaded at the end of the @samp{.text} section
5145 even though its VMA is @code{0x2000}, and one called @samp{.bss} to hold
5146 uninitialized data at address @code{0x3000}. The symbol @code{_data} is
5147 defined with the value @code{0x2000}, which shows that the location
5148 counter holds the VMA value, not the LMA value.
5154 .text 0x1000 : @{ *(.text) _etext = . ; @}
5156 AT ( ADDR (.text) + SIZEOF (.text) )
5157 @{ _data = . ; *(.data); _edata = . ; @}
5159 @{ _bstart = . ; *(.bss) *(COMMON) ; _bend = . ;@}
5164 The run-time initialization code for use with a program generated with
5165 this linker script would include something like the following, to copy
5166 the initialized data from the ROM image to its runtime address. Notice
5167 how this code takes advantage of the symbols defined by the linker
5172 extern char _etext, _data, _edata, _bstart, _bend;
5173 char *src = &_etext;
5176 /* ROM has data at end of text; copy it. */
5177 while (dst < &_edata)
5181 for (dst = &_bstart; dst< &_bend; dst++)
5186 @node Forced Output Alignment
5187 @subsubsection Forced Output Alignment
5188 @kindex ALIGN(@var{section_align})
5189 @cindex forcing output section alignment
5190 @cindex output section alignment
5191 You can increase an output section's alignment by using ALIGN. As an
5192 alternative you can enforce that the difference between the VMA and LMA remains
5193 intact throughout this output section with the ALIGN_WITH_INPUT attribute.
5195 @node Forced Input Alignment
5196 @subsubsection Forced Input Alignment
5197 @kindex SUBALIGN(@var{subsection_align})
5198 @cindex forcing input section alignment
5199 @cindex input section alignment
5200 You can force input section alignment within an output section by using
5201 SUBALIGN. The value specified overrides any alignment given by input
5202 sections, whether larger or smaller.
5204 @node Output Section Constraint
5205 @subsubsection Output Section Constraint
5208 @cindex constraints on output sections
5209 You can specify that an output section should only be created if all
5210 of its input sections are read-only or all of its input sections are
5211 read-write by using the keyword @code{ONLY_IF_RO} and
5212 @code{ONLY_IF_RW} respectively.
5214 @node Output Section Region
5215 @subsubsection Output Section Region
5216 @kindex >@var{region}
5217 @cindex section, assigning to memory region
5218 @cindex memory regions and sections
5219 You can assign a section to a previously defined region of memory by
5220 using @samp{>@var{region}}. @xref{MEMORY}.
5222 Here is a simple example:
5225 MEMORY @{ rom : ORIGIN = 0x1000, LENGTH = 0x1000 @}
5226 SECTIONS @{ ROM : @{ *(.text) @} >rom @}
5230 @node Output Section Phdr
5231 @subsubsection Output Section Phdr
5233 @cindex section, assigning to program header
5234 @cindex program headers and sections
5235 You can assign a section to a previously defined program segment by
5236 using @samp{:@var{phdr}}. @xref{PHDRS}. If a section is assigned to
5237 one or more segments, then all subsequent allocated sections will be
5238 assigned to those segments as well, unless they use an explicitly
5239 @code{:@var{phdr}} modifier. You can use @code{:NONE} to tell the
5240 linker to not put the section in any segment at all.
5242 Here is a simple example:
5245 PHDRS @{ text PT_LOAD ; @}
5246 SECTIONS @{ .text : @{ *(.text) @} :text @}
5250 @node Output Section Fill
5251 @subsubsection Output Section Fill
5252 @kindex =@var{fillexp}
5253 @cindex section fill pattern
5254 @cindex fill pattern, entire section
5255 You can set the fill pattern for an entire section by using
5256 @samp{=@var{fillexp}}. @var{fillexp} is an expression
5257 (@pxref{Expressions}). Any otherwise unspecified regions of memory
5258 within the output section (for example, gaps left due to the required
5259 alignment of input sections) will be filled with the value, repeated as
5260 necessary. If the fill expression is a simple hex number, ie. a string
5261 of hex digit starting with @samp{0x} and without a trailing @samp{k} or @samp{M}, then
5262 an arbitrarily long sequence of hex digits can be used to specify the
5263 fill pattern; Leading zeros become part of the pattern too. For all
5264 other cases, including extra parentheses or a unary @code{+}, the fill
5265 pattern is the four least significant bytes of the value of the
5266 expression. In all cases, the number is big-endian.
5268 You can also change the fill value with a @code{FILL} command in the
5269 output section commands; (@pxref{Output Section Data}).
5271 Here is a simple example:
5274 SECTIONS @{ .text : @{ *(.text) @} =0x90909090 @}
5278 @node Overlay Description
5279 @subsection Overlay Description
5282 An overlay description provides an easy way to describe sections which
5283 are to be loaded as part of a single memory image but are to be run at
5284 the same memory address. At run time, some sort of overlay manager will
5285 copy the overlaid sections in and out of the runtime memory address as
5286 required, perhaps by simply manipulating addressing bits. This approach
5287 can be useful, for example, when a certain region of memory is faster
5290 Overlays are described using the @code{OVERLAY} command. The
5291 @code{OVERLAY} command is used within a @code{SECTIONS} command, like an
5292 output section description. The full syntax of the @code{OVERLAY}
5293 command is as follows:
5296 OVERLAY [@var{start}] : [NOCROSSREFS] [AT ( @var{ldaddr} )]
5300 @var{output-section-command}
5301 @var{output-section-command}
5303 @} [:@var{phdr}@dots{}] [=@var{fill}]
5306 @var{output-section-command}
5307 @var{output-section-command}
5309 @} [:@var{phdr}@dots{}] [=@var{fill}]
5311 @} [>@var{region}] [:@var{phdr}@dots{}] [=@var{fill}] [,]
5315 Everything is optional except @code{OVERLAY} (a keyword), and each
5316 section must have a name (@var{secname1} and @var{secname2} above). The
5317 section definitions within the @code{OVERLAY} construct are identical to
5318 those within the general @code{SECTIONS} construct (@pxref{SECTIONS}),
5319 except that no addresses and no memory regions may be defined for
5320 sections within an @code{OVERLAY}.
5322 The comma at the end may be required if a @var{fill} is used and
5323 the next @var{sections-command} looks like a continuation of the expression.
5325 The sections are all defined with the same starting address. The load
5326 addresses of the sections are arranged such that they are consecutive in
5327 memory starting at the load address used for the @code{OVERLAY} as a
5328 whole (as with normal section definitions, the load address is optional,
5329 and defaults to the start address; the start address is also optional,
5330 and defaults to the current value of the location counter).
5332 If the @code{NOCROSSREFS} keyword is used, and there are any
5333 references among the sections, the linker will report an error. Since
5334 the sections all run at the same address, it normally does not make
5335 sense for one section to refer directly to another.
5336 @xref{Miscellaneous Commands, NOCROSSREFS}.
5338 For each section within the @code{OVERLAY}, the linker automatically
5339 provides two symbols. The symbol @code{__load_start_@var{secname}} is
5340 defined as the starting load address of the section. The symbol
5341 @code{__load_stop_@var{secname}} is defined as the final load address of
5342 the section. Any characters within @var{secname} which are not legal
5343 within C identifiers are removed. C (or assembler) code may use these
5344 symbols to move the overlaid sections around as necessary.
5346 At the end of the overlay, the value of the location counter is set to
5347 the start address of the overlay plus the size of the largest section.
5349 Here is an example. Remember that this would appear inside a
5350 @code{SECTIONS} construct.
5353 OVERLAY 0x1000 : AT (0x4000)
5355 .text0 @{ o1/*.o(.text) @}
5356 .text1 @{ o2/*.o(.text) @}
5361 This will define both @samp{.text0} and @samp{.text1} to start at
5362 address 0x1000. @samp{.text0} will be loaded at address 0x4000, and
5363 @samp{.text1} will be loaded immediately after @samp{.text0}. The
5364 following symbols will be defined if referenced: @code{__load_start_text0},
5365 @code{__load_stop_text0}, @code{__load_start_text1},
5366 @code{__load_stop_text1}.
5368 C code to copy overlay @code{.text1} into the overlay area might look
5373 extern char __load_start_text1, __load_stop_text1;
5374 memcpy ((char *) 0x1000, &__load_start_text1,
5375 &__load_stop_text1 - &__load_start_text1);
5379 Note that the @code{OVERLAY} command is just syntactic sugar, since
5380 everything it does can be done using the more basic commands. The above
5381 example could have been written identically as follows.
5385 .text0 0x1000 : AT (0x4000) @{ o1/*.o(.text) @}
5386 PROVIDE (__load_start_text0 = LOADADDR (.text0));
5387 PROVIDE (__load_stop_text0 = LOADADDR (.text0) + SIZEOF (.text0));
5388 .text1 0x1000 : AT (0x4000 + SIZEOF (.text0)) @{ o2/*.o(.text) @}
5389 PROVIDE (__load_start_text1 = LOADADDR (.text1));
5390 PROVIDE (__load_stop_text1 = LOADADDR (.text1) + SIZEOF (.text1));
5391 . = 0x1000 + MAX (SIZEOF (.text0), SIZEOF (.text1));
5396 @section MEMORY Command
5398 @cindex memory regions
5399 @cindex regions of memory
5400 @cindex allocating memory
5401 @cindex discontinuous memory
5402 The linker's default configuration permits allocation of all available
5403 memory. You can override this by using the @code{MEMORY} command.
5405 The @code{MEMORY} command describes the location and size of blocks of
5406 memory in the target. You can use it to describe which memory regions
5407 may be used by the linker, and which memory regions it must avoid. You
5408 can then assign sections to particular memory regions. The linker will
5409 set section addresses based on the memory regions, and will warn about
5410 regions that become too full. The linker will not shuffle sections
5411 around to fit into the available regions.
5413 A linker script may contain many uses of the @code{MEMORY} command,
5414 however, all memory blocks defined are treated as if they were
5415 specified inside a single @code{MEMORY} command. The syntax for
5421 @var{name} [(@var{attr})] : ORIGIN = @var{origin}, LENGTH = @var{len}
5427 The @var{name} is a name used in the linker script to refer to the
5428 region. The region name has no meaning outside of the linker script.
5429 Region names are stored in a separate name space, and will not conflict
5430 with symbol names, file names, or section names. Each memory region
5431 must have a distinct name within the @code{MEMORY} command. However you can
5432 add later alias names to existing memory regions with the @ref{REGION_ALIAS}
5435 @cindex memory region attributes
5436 The @var{attr} string is an optional list of attributes that specify
5437 whether to use a particular memory region for an input section which is
5438 not explicitly mapped in the linker script. As described in
5439 @ref{SECTIONS}, if you do not specify an output section for some input
5440 section, the linker will create an output section with the same name as
5441 the input section. If you define region attributes, the linker will use
5442 them to select the memory region for the output section that it creates.
5444 The @var{attr} string must consist only of the following characters:
5459 Invert the sense of any of the attributes that follow
5462 If an unmapped section matches any of the listed attributes other than
5463 @samp{!}, it will be placed in the memory region. The @samp{!}
5464 attribute reverses the test for the characters that follow, so that an
5465 unmapped section will be placed in the memory region only if it does
5466 not match any of the attributes listed afterwards. Thus an attribute
5467 string of @samp{RW!X} will match any unmapped section that has either
5468 or both of the @samp{R} and @samp{W} attributes, but only as long as
5469 the section does not also have the @samp{X} attribute.
5474 The @var{origin} is an numerical expression for the start address of
5475 the memory region. The expression must evaluate to a constant and it
5476 cannot involve any symbols. The keyword @code{ORIGIN} may be
5477 abbreviated to @code{org} or @code{o} (but not, for example,
5483 The @var{len} is an expression for the size in bytes of the memory
5484 region. As with the @var{origin} expression, the expression must
5485 be numerical only and must evaluate to a constant. The keyword
5486 @code{LENGTH} may be abbreviated to @code{len} or @code{l}.
5488 In the following example, we specify that there are two memory regions
5489 available for allocation: one starting at @samp{0} for 256 kilobytes,
5490 and the other starting at @samp{0x40000000} for four megabytes. The
5491 linker will place into the @samp{rom} memory region every section which
5492 is not explicitly mapped into a memory region, and is either read-only
5493 or executable. The linker will place other sections which are not
5494 explicitly mapped into a memory region into the @samp{ram} memory
5501 rom (rx) : ORIGIN = 0, LENGTH = 256K
5502 ram (!rx) : org = 0x40000000, l = 4M
5507 Once you define a memory region, you can direct the linker to place
5508 specific output sections into that memory region by using the
5509 @samp{>@var{region}} output section attribute. For example, if you have
5510 a memory region named @samp{mem}, you would use @samp{>mem} in the
5511 output section definition. @xref{Output Section Region}. If no address
5512 was specified for the output section, the linker will set the address to
5513 the next available address within the memory region. If the combined
5514 output sections directed to a memory region are too large for the
5515 region, the linker will issue an error message.
5517 It is possible to access the origin and length of a memory in an
5518 expression via the @code{ORIGIN(@var{memory})} and
5519 @code{LENGTH(@var{memory})} functions:
5523 _fstack = ORIGIN(ram) + LENGTH(ram) - 4;
5528 @section PHDRS Command
5530 @cindex program headers
5531 @cindex ELF program headers
5532 @cindex program segments
5533 @cindex segments, ELF
5534 The ELF object file format uses @dfn{program headers}, also knows as
5535 @dfn{segments}. The program headers describe how the program should be
5536 loaded into memory. You can print them out by using the @code{objdump}
5537 program with the @samp{-p} option.
5539 When you run an ELF program on a native ELF system, the system loader
5540 reads the program headers in order to figure out how to load the
5541 program. This will only work if the program headers are set correctly.
5542 This manual does not describe the details of how the system loader
5543 interprets program headers; for more information, see the ELF ABI.
5545 The linker will create reasonable program headers by default. However,
5546 in some cases, you may need to specify the program headers more
5547 precisely. You may use the @code{PHDRS} command for this purpose. When
5548 the linker sees the @code{PHDRS} command in the linker script, it will
5549 not create any program headers other than the ones specified.
5551 The linker only pays attention to the @code{PHDRS} command when
5552 generating an ELF output file. In other cases, the linker will simply
5553 ignore @code{PHDRS}.
5555 This is the syntax of the @code{PHDRS} command. The words @code{PHDRS},
5556 @code{FILEHDR}, @code{AT}, and @code{FLAGS} are keywords.
5562 @var{name} @var{type} [ FILEHDR ] [ PHDRS ] [ AT ( @var{address} ) ]
5563 [ FLAGS ( @var{flags} ) ] ;
5568 The @var{name} is used only for reference in the @code{SECTIONS} command
5569 of the linker script. It is not put into the output file. Program
5570 header names are stored in a separate name space, and will not conflict
5571 with symbol names, file names, or section names. Each program header
5572 must have a distinct name. The headers are processed in order and it
5573 is usual for them to map to sections in ascending load address order.
5575 Certain program header types describe segments of memory which the
5576 system loader will load from the file. In the linker script, you
5577 specify the contents of these segments by placing allocatable output
5578 sections in the segments. You use the @samp{:@var{phdr}} output section
5579 attribute to place a section in a particular segment. @xref{Output
5582 It is normal to put certain sections in more than one segment. This
5583 merely implies that one segment of memory contains another. You may
5584 repeat @samp{:@var{phdr}}, using it once for each segment which should
5585 contain the section.
5587 If you place a section in one or more segments using @samp{:@var{phdr}},
5588 then the linker will place all subsequent allocatable sections which do
5589 not specify @samp{:@var{phdr}} in the same segments. This is for
5590 convenience, since generally a whole set of contiguous sections will be
5591 placed in a single segment. You can use @code{:NONE} to override the
5592 default segment and tell the linker to not put the section in any
5597 You may use the @code{FILEHDR} and @code{PHDRS} keywords after
5598 the program header type to further describe the contents of the segment.
5599 The @code{FILEHDR} keyword means that the segment should include the ELF
5600 file header. The @code{PHDRS} keyword means that the segment should
5601 include the ELF program headers themselves. If applied to a loadable
5602 segment (@code{PT_LOAD}), all prior loadable segments must have one of
5605 The @var{type} may be one of the following. The numbers indicate the
5606 value of the keyword.
5609 @item @code{PT_NULL} (0)
5610 Indicates an unused program header.
5612 @item @code{PT_LOAD} (1)
5613 Indicates that this program header describes a segment to be loaded from
5616 @item @code{PT_DYNAMIC} (2)
5617 Indicates a segment where dynamic linking information can be found.
5619 @item @code{PT_INTERP} (3)
5620 Indicates a segment where the name of the program interpreter may be
5623 @item @code{PT_NOTE} (4)
5624 Indicates a segment holding note information.
5626 @item @code{PT_SHLIB} (5)
5627 A reserved program header type, defined but not specified by the ELF
5630 @item @code{PT_PHDR} (6)
5631 Indicates a segment where the program headers may be found.
5633 @item @code{PT_TLS} (7)
5634 Indicates a segment containing thread local storage.
5636 @item @var{expression}
5637 An expression giving the numeric type of the program header. This may
5638 be used for types not defined above.
5641 You can specify that a segment should be loaded at a particular address
5642 in memory by using an @code{AT} expression. This is identical to the
5643 @code{AT} command used as an output section attribute (@pxref{Output
5644 Section LMA}). The @code{AT} command for a program header overrides the
5645 output section attribute.
5647 The linker will normally set the segment flags based on the sections
5648 which comprise the segment. You may use the @code{FLAGS} keyword to
5649 explicitly specify the segment flags. The value of @var{flags} must be
5650 an integer. It is used to set the @code{p_flags} field of the program
5653 Here is an example of @code{PHDRS}. This shows a typical set of program
5654 headers used on a native ELF system.
5660 headers PT_PHDR PHDRS ;
5662 text PT_LOAD FILEHDR PHDRS ;
5664 dynamic PT_DYNAMIC ;
5670 .interp : @{ *(.interp) @} :text :interp
5671 .text : @{ *(.text) @} :text
5672 .rodata : @{ *(.rodata) @} /* defaults to :text */
5674 . = . + 0x1000; /* move to a new page in memory */
5675 .data : @{ *(.data) @} :data
5676 .dynamic : @{ *(.dynamic) @} :data :dynamic
5683 @section VERSION Command
5684 @kindex VERSION @{script text@}
5685 @cindex symbol versions
5686 @cindex version script
5687 @cindex versions of symbols
5688 The linker supports symbol versions when using ELF. Symbol versions are
5689 only useful when using shared libraries. The dynamic linker can use
5690 symbol versions to select a specific version of a function when it runs
5691 a program that may have been linked against an earlier version of the
5694 You can include a version script directly in the main linker script, or
5695 you can supply the version script as an implicit linker script. You can
5696 also use the @samp{--version-script} linker option.
5698 The syntax of the @code{VERSION} command is simply
5700 VERSION @{ version-script-commands @}
5703 The format of the version script commands is identical to that used by
5704 Sun's linker in Solaris 2.5. The version script defines a tree of
5705 version nodes. You specify the node names and interdependencies in the
5706 version script. You can specify which symbols are bound to which
5707 version nodes, and you can reduce a specified set of symbols to local
5708 scope so that they are not globally visible outside of the shared
5711 The easiest way to demonstrate the version script language is with a few
5737 This example version script defines three version nodes. The first
5738 version node defined is @samp{VERS_1.1}; it has no other dependencies.
5739 The script binds the symbol @samp{foo1} to @samp{VERS_1.1}. It reduces
5740 a number of symbols to local scope so that they are not visible outside
5741 of the shared library; this is done using wildcard patterns, so that any
5742 symbol whose name begins with @samp{old}, @samp{original}, or @samp{new}
5743 is matched. The wildcard patterns available are the same as those used
5744 in the shell when matching filenames (also known as ``globbing'').
5745 However, if you specify the symbol name inside double quotes, then the
5746 name is treated as literal, rather than as a glob pattern.
5748 Next, the version script defines node @samp{VERS_1.2}. This node
5749 depends upon @samp{VERS_1.1}. The script binds the symbol @samp{foo2}
5750 to the version node @samp{VERS_1.2}.
5752 Finally, the version script defines node @samp{VERS_2.0}. This node
5753 depends upon @samp{VERS_1.2}. The scripts binds the symbols @samp{bar1}
5754 and @samp{bar2} are bound to the version node @samp{VERS_2.0}.
5756 When the linker finds a symbol defined in a library which is not
5757 specifically bound to a version node, it will effectively bind it to an
5758 unspecified base version of the library. You can bind all otherwise
5759 unspecified symbols to a given version node by using @samp{global: *;}
5760 somewhere in the version script. Note that it's slightly crazy to use
5761 wildcards in a global spec except on the last version node. Global
5762 wildcards elsewhere run the risk of accidentally adding symbols to the
5763 set exported for an old version. That's wrong since older versions
5764 ought to have a fixed set of symbols.
5766 The names of the version nodes have no specific meaning other than what
5767 they might suggest to the person reading them. The @samp{2.0} version
5768 could just as well have appeared in between @samp{1.1} and @samp{1.2}.
5769 However, this would be a confusing way to write a version script.
5771 Node name can be omitted, provided it is the only version node
5772 in the version script. Such version script doesn't assign any versions to
5773 symbols, only selects which symbols will be globally visible out and which
5777 @{ global: foo; bar; local: *; @};
5780 When you link an application against a shared library that has versioned
5781 symbols, the application itself knows which version of each symbol it
5782 requires, and it also knows which version nodes it needs from each
5783 shared library it is linked against. Thus at runtime, the dynamic
5784 loader can make a quick check to make sure that the libraries you have
5785 linked against do in fact supply all of the version nodes that the
5786 application will need to resolve all of the dynamic symbols. In this
5787 way it is possible for the dynamic linker to know with certainty that
5788 all external symbols that it needs will be resolvable without having to
5789 search for each symbol reference.
5791 The symbol versioning is in effect a much more sophisticated way of
5792 doing minor version checking that SunOS does. The fundamental problem
5793 that is being addressed here is that typically references to external
5794 functions are bound on an as-needed basis, and are not all bound when
5795 the application starts up. If a shared library is out of date, a
5796 required interface may be missing; when the application tries to use
5797 that interface, it may suddenly and unexpectedly fail. With symbol
5798 versioning, the user will get a warning when they start their program if
5799 the libraries being used with the application are too old.
5801 There are several GNU extensions to Sun's versioning approach. The
5802 first of these is the ability to bind a symbol to a version node in the
5803 source file where the symbol is defined instead of in the versioning
5804 script. This was done mainly to reduce the burden on the library
5805 maintainer. You can do this by putting something like:
5807 __asm__(".symver original_foo,foo@@VERS_1.1");
5810 in the C source file. This renames the function @samp{original_foo} to
5811 be an alias for @samp{foo} bound to the version node @samp{VERS_1.1}.
5812 The @samp{local:} directive can be used to prevent the symbol
5813 @samp{original_foo} from being exported. A @samp{.symver} directive
5814 takes precedence over a version script.
5816 The second GNU extension is to allow multiple versions of the same
5817 function to appear in a given shared library. In this way you can make
5818 an incompatible change to an interface without increasing the major
5819 version number of the shared library, while still allowing applications
5820 linked against the old interface to continue to function.
5822 To do this, you must use multiple @samp{.symver} directives in the
5823 source file. Here is an example:
5826 __asm__(".symver original_foo,foo@@");
5827 __asm__(".symver old_foo,foo@@VERS_1.1");
5828 __asm__(".symver old_foo1,foo@@VERS_1.2");
5829 __asm__(".symver new_foo,foo@@@@VERS_2.0");
5832 In this example, @samp{foo@@} represents the symbol @samp{foo} bound to the
5833 unspecified base version of the symbol. The source file that contains this
5834 example would define 4 C functions: @samp{original_foo}, @samp{old_foo},
5835 @samp{old_foo1}, and @samp{new_foo}.
5837 When you have multiple definitions of a given symbol, there needs to be
5838 some way to specify a default version to which external references to
5839 this symbol will be bound. You can do this with the
5840 @samp{foo@@@@VERS_2.0} type of @samp{.symver} directive. You can only
5841 declare one version of a symbol as the default in this manner; otherwise
5842 you would effectively have multiple definitions of the same symbol.
5844 If you wish to bind a reference to a specific version of the symbol
5845 within the shared library, you can use the aliases of convenience
5846 (i.e., @samp{old_foo}), or you can use the @samp{.symver} directive to
5847 specifically bind to an external version of the function in question.
5849 You can also specify the language in the version script:
5852 VERSION extern "lang" @{ version-script-commands @}
5855 The supported @samp{lang}s are @samp{C}, @samp{C++}, and @samp{Java}.
5856 The linker will iterate over the list of symbols at the link time and
5857 demangle them according to @samp{lang} before matching them to the
5858 patterns specified in @samp{version-script-commands}. The default
5859 @samp{lang} is @samp{C}.
5861 Demangled names may contains spaces and other special characters. As
5862 described above, you can use a glob pattern to match demangled names,
5863 or you can use a double-quoted string to match the string exactly. In
5864 the latter case, be aware that minor differences (such as differing
5865 whitespace) between the version script and the demangler output will
5866 cause a mismatch. As the exact string generated by the demangler
5867 might change in the future, even if the mangled name does not, you
5868 should check that all of your version directives are behaving as you
5869 expect when you upgrade.
5872 @section Expressions in Linker Scripts
5875 The syntax for expressions in the linker script language is identical to
5876 that of C expressions. All expressions are evaluated as integers. All
5877 expressions are evaluated in the same size, which is 32 bits if both the
5878 host and target are 32 bits, and is otherwise 64 bits.
5880 You can use and set symbol values in expressions.
5882 The linker defines several special purpose builtin functions for use in
5886 * Constants:: Constants
5887 * Symbolic Constants:: Symbolic constants
5888 * Symbols:: Symbol Names
5889 * Orphan Sections:: Orphan Sections
5890 * Location Counter:: The Location Counter
5891 * Operators:: Operators
5892 * Evaluation:: Evaluation
5893 * Expression Section:: The Section of an Expression
5894 * Builtin Functions:: Builtin Functions
5898 @subsection Constants
5899 @cindex integer notation
5900 @cindex constants in linker scripts
5901 All constants are integers.
5903 As in C, the linker considers an integer beginning with @samp{0} to be
5904 octal, and an integer beginning with @samp{0x} or @samp{0X} to be
5905 hexadecimal. Alternatively the linker accepts suffixes of @samp{h} or
5906 @samp{H} for hexadecimal, @samp{o} or @samp{O} for octal, @samp{b} or
5907 @samp{B} for binary and @samp{d} or @samp{D} for decimal. Any integer
5908 value without a prefix or a suffix is considered to be decimal.
5910 @cindex scaled integers
5911 @cindex K and M integer suffixes
5912 @cindex M and K integer suffixes
5913 @cindex suffixes for integers
5914 @cindex integer suffixes
5915 In addition, you can use the suffixes @code{K} and @code{M} to scale a
5919 @c END TEXI2ROFF-KILL
5920 @code{1024} or @code{1024*1024}
5924 ${\rm 1024}$ or ${\rm 1024}^2$
5926 @c END TEXI2ROFF-KILL
5927 respectively. For example, the following
5928 all refer to the same quantity:
5937 Note - the @code{K} and @code{M} suffixes cannot be used in
5938 conjunction with the base suffixes mentioned above.
5940 @node Symbolic Constants
5941 @subsection Symbolic Constants
5942 @cindex symbolic constants
5944 It is possible to refer to target-specific constants via the use of
5945 the @code{CONSTANT(@var{name})} operator, where @var{name} is one of:
5950 The target's maximum page size.
5952 @item COMMONPAGESIZE
5953 @kindex COMMONPAGESIZE
5954 The target's default page size.
5960 .text ALIGN (CONSTANT (MAXPAGESIZE)) : @{ *(.text) @}
5963 will create a text section aligned to the largest page boundary
5964 supported by the target.
5967 @subsection Symbol Names
5968 @cindex symbol names
5970 @cindex quoted symbol names
5972 Unless quoted, symbol names start with a letter, underscore, or period
5973 and may include letters, digits, underscores, periods, and hyphens.
5974 Unquoted symbol names must not conflict with any keywords. You can
5975 specify a symbol which contains odd characters or has the same name as a
5976 keyword by surrounding the symbol name in double quotes:
5979 "with a space" = "also with a space" + 10;
5982 Since symbols can contain many non-alphabetic characters, it is safest
5983 to delimit symbols with spaces. For example, @samp{A-B} is one symbol,
5984 whereas @samp{A - B} is an expression involving subtraction.
5986 @node Orphan Sections
5987 @subsection Orphan Sections
5989 Orphan sections are sections present in the input files which
5990 are not explicitly placed into the output file by the linker
5991 script. The linker will still copy these sections into the
5992 output file by either finding, or creating a suitable output section
5993 in which to place the orphaned input section.
5995 If the name of an orphaned input section exactly matches the name of
5996 an existing output section, then the orphaned input section will be
5997 placed at the end of that output section.
5999 If there is no output section with a matching name then new output
6000 sections will be created. Each new output section will have the same
6001 name as the orphan section placed within it. If there are multiple
6002 orphan sections with the same name, these will all be combined into
6003 one new output section.
6005 If new output sections are created to hold orphaned input sections,
6006 then the linker must decide where to place these new output sections
6007 in relation to existing output sections. On most modern targets, the
6008 linker attempts to place orphan sections after sections of the same
6009 attribute, such as code vs data, loadable vs non-loadable, etc. If no
6010 sections with matching attributes are found, or your target lacks this
6011 support, the orphan section is placed at the end of the file.
6013 The command-line options @samp{--orphan-handling} and @samp{--unique}
6014 (@pxref{Options,,Command-line Options}) can be used to control which
6015 output sections an orphan is placed in.
6017 @node Location Counter
6018 @subsection The Location Counter
6021 @cindex location counter
6022 @cindex current output location
6023 The special linker variable @dfn{dot} @samp{.} always contains the
6024 current output location counter. Since the @code{.} always refers to a
6025 location in an output section, it may only appear in an expression
6026 within a @code{SECTIONS} command. The @code{.} symbol may appear
6027 anywhere that an ordinary symbol is allowed in an expression.
6030 Assigning a value to @code{.} will cause the location counter to be
6031 moved. This may be used to create holes in the output section. The
6032 location counter may not be moved backwards inside an output section,
6033 and may not be moved backwards outside of an output section if so
6034 doing creates areas with overlapping LMAs.
6050 In the previous example, the @samp{.text} section from @file{file1} is
6051 located at the beginning of the output section @samp{output}. It is
6052 followed by a 1000 byte gap. Then the @samp{.text} section from
6053 @file{file2} appears, also with a 1000 byte gap following before the
6054 @samp{.text} section from @file{file3}. The notation @samp{= 0x12345678}
6055 specifies what data to write in the gaps (@pxref{Output Section Fill}).
6057 @cindex dot inside sections
6058 Note: @code{.} actually refers to the byte offset from the start of the
6059 current containing object. Normally this is the @code{SECTIONS}
6060 statement, whose start address is 0, hence @code{.} can be used as an
6061 absolute address. If @code{.} is used inside a section description
6062 however, it refers to the byte offset from the start of that section,
6063 not an absolute address. Thus in a script like this:
6081 The @samp{.text} section will be assigned a starting address of 0x100
6082 and a size of exactly 0x200 bytes, even if there is not enough data in
6083 the @samp{.text} input sections to fill this area. (If there is too
6084 much data, an error will be produced because this would be an attempt to
6085 move @code{.} backwards). The @samp{.data} section will start at 0x500
6086 and it will have an extra 0x600 bytes worth of space after the end of
6087 the values from the @samp{.data} input sections and before the end of
6088 the @samp{.data} output section itself.
6090 @cindex dot outside sections
6091 Setting symbols to the value of the location counter outside of an
6092 output section statement can result in unexpected values if the linker
6093 needs to place orphan sections. For example, given the following:
6099 .text: @{ *(.text) @}
6103 .data: @{ *(.data) @}
6108 If the linker needs to place some input section, e.g. @code{.rodata},
6109 not mentioned in the script, it might choose to place that section
6110 between @code{.text} and @code{.data}. You might think the linker
6111 should place @code{.rodata} on the blank line in the above script, but
6112 blank lines are of no particular significance to the linker. As well,
6113 the linker doesn't associate the above symbol names with their
6114 sections. Instead, it assumes that all assignments or other
6115 statements belong to the previous output section, except for the
6116 special case of an assignment to @code{.}. I.e., the linker will
6117 place the orphan @code{.rodata} section as if the script was written
6124 .text: @{ *(.text) @}
6128 .rodata: @{ *(.rodata) @}
6129 .data: @{ *(.data) @}
6134 This may or may not be the script author's intention for the value of
6135 @code{start_of_data}. One way to influence the orphan section
6136 placement is to assign the location counter to itself, as the linker
6137 assumes that an assignment to @code{.} is setting the start address of
6138 a following output section and thus should be grouped with that
6139 section. So you could write:
6145 .text: @{ *(.text) @}
6150 .data: @{ *(.data) @}
6155 Now, the orphan @code{.rodata} section will be placed between
6156 @code{end_of_text} and @code{start_of_data}.
6160 @subsection Operators
6161 @cindex operators for arithmetic
6162 @cindex arithmetic operators
6163 @cindex precedence in expressions
6164 The linker recognizes the standard C set of arithmetic operators, with
6165 the standard bindings and precedence levels:
6168 @c END TEXI2ROFF-KILL
6170 precedence associativity Operators Notes
6176 5 left == != > < <= >=
6182 11 right &= += -= *= /= (2)
6186 (1) Prefix operators
6187 (2) @xref{Assignments}.
6191 \vskip \baselineskip
6192 %"lispnarrowing" is the extra indent used generally for smallexample
6193 \hskip\lispnarrowing\vbox{\offinterlineskip
6196 {\vrule#&\strut\hfil\ #\ \hfil&\vrule#&\strut\hfil\ #\ \hfil&\vrule#&\strut\hfil\ {\tt #}\ \hfil&\vrule#\cr
6197 height2pt&\omit&&\omit&&\omit&\cr
6198 &Precedence&& Associativity &&{\rm Operators}&\cr
6199 height2pt&\omit&&\omit&&\omit&\cr
6201 height2pt&\omit&&\omit&&\omit&\cr
6203 % '176 is tilde, '~' in tt font
6204 &1&&left&&\qquad- \char'176\ !\qquad\dag&\cr
6205 &2&&left&&* / \%&\cr
6208 &5&&left&&== != > < <= >=&\cr
6211 &8&&left&&{\&\&}&\cr
6214 &11&&right&&\qquad\&= += -= *= /=\qquad\ddag&\cr
6216 height2pt&\omit&&\omit&&\omit&\cr}
6221 @obeylines@parskip=0pt@parindent=0pt
6222 @dag@quad Prefix operators.
6223 @ddag@quad @xref{Assignments}.
6226 @c END TEXI2ROFF-KILL
6229 @subsection Evaluation
6230 @cindex lazy evaluation
6231 @cindex expression evaluation order
6232 The linker evaluates expressions lazily. It only computes the value of
6233 an expression when absolutely necessary.
6235 The linker needs some information, such as the value of the start
6236 address of the first section, and the origins and lengths of memory
6237 regions, in order to do any linking at all. These values are computed
6238 as soon as possible when the linker reads in the linker script.
6240 However, other values (such as symbol values) are not known or needed
6241 until after storage allocation. Such values are evaluated later, when
6242 other information (such as the sizes of output sections) is available
6243 for use in the symbol assignment expression.
6245 The sizes of sections cannot be known until after allocation, so
6246 assignments dependent upon these are not performed until after
6249 Some expressions, such as those depending upon the location counter
6250 @samp{.}, must be evaluated during section allocation.
6252 If the result of an expression is required, but the value is not
6253 available, then an error results. For example, a script like the
6259 .text 9+this_isnt_constant :
6265 will cause the error message @samp{non constant expression for initial
6268 @node Expression Section
6269 @subsection The Section of an Expression
6270 @cindex expression sections
6271 @cindex absolute expressions
6272 @cindex relative expressions
6273 @cindex absolute and relocatable symbols
6274 @cindex relocatable and absolute symbols
6275 @cindex symbols, relocatable and absolute
6276 Addresses and symbols may be section relative, or absolute. A section
6277 relative symbol is relocatable. If you request relocatable output
6278 using the @samp{-r} option, a further link operation may change the
6279 value of a section relative symbol. On the other hand, an absolute
6280 symbol will retain the same value throughout any further link
6283 Some terms in linker expressions are addresses. This is true of
6284 section relative symbols and for builtin functions that return an
6285 address, such as @code{ADDR}, @code{LOADADDR}, @code{ORIGIN} and
6286 @code{SEGMENT_START}. Other terms are simply numbers, or are builtin
6287 functions that return a non-address value, such as @code{LENGTH}.
6288 One complication is that unless you set @code{LD_FEATURE ("SANE_EXPR")}
6289 (@pxref{Miscellaneous Commands}), numbers and absolute symbols are treated
6290 differently depending on their location, for compatibility with older
6291 versions of @code{ld}. Expressions appearing outside an output
6292 section definition treat all numbers as absolute addresses.
6293 Expressions appearing inside an output section definition treat
6294 absolute symbols as numbers. If @code{LD_FEATURE ("SANE_EXPR")} is
6295 given, then absolute symbols and numbers are simply treated as numbers
6298 In the following simple example,
6305 __executable_start = 0x100;
6309 __data_start = 0x10;
6317 both @code{.} and @code{__executable_start} are set to the absolute
6318 address 0x100 in the first two assignments, then both @code{.} and
6319 @code{__data_start} are set to 0x10 relative to the @code{.data}
6320 section in the second two assignments.
6322 For expressions involving numbers, relative addresses and absolute
6323 addresses, ld follows these rules to evaluate terms:
6327 Unary operations on an absolute address or number, and binary
6328 operations on two absolute addresses or two numbers, or between one
6329 absolute address and a number, apply the operator to the value(s).
6331 Unary operations on a relative address, and binary operations on two
6332 relative addresses in the same section or between one relative address
6333 and a number, apply the operator to the offset part of the address(es).
6335 Other binary operations, that is, between two relative addresses not
6336 in the same section, or between a relative address and an absolute
6337 address, first convert any non-absolute term to an absolute address
6338 before applying the operator.
6341 The result section of each sub-expression is as follows:
6345 An operation involving only numbers results in a number.
6347 The result of comparisons, @samp{&&} and @samp{||} is also a number.
6349 The result of other binary arithmetic and logical operations on two
6350 relative addresses in the same section or two absolute addresses
6351 (after above conversions) is also a number when
6352 @code{LD_FEATURE ("SANE_EXPR")} or inside an output section definition
6353 but an absolute address otherwise.
6355 The result of other operations on relative addresses or one
6356 relative address and a number, is a relative address in the same
6357 section as the relative operand(s).
6359 The result of other operations on absolute addresses (after above
6360 conversions) is an absolute address.
6363 You can use the builtin function @code{ABSOLUTE} to force an expression
6364 to be absolute when it would otherwise be relative. For example, to
6365 create an absolute symbol set to the address of the end of the output
6366 section @samp{.data}:
6370 .data : @{ *(.data) _edata = ABSOLUTE(.); @}
6374 If @samp{ABSOLUTE} were not used, @samp{_edata} would be relative to the
6375 @samp{.data} section.
6377 Using @code{LOADADDR} also forces an expression absolute, since this
6378 particular builtin function returns an absolute address.
6380 @node Builtin Functions
6381 @subsection Builtin Functions
6382 @cindex functions in expressions
6383 The linker script language includes a number of builtin functions for
6384 use in linker script expressions.
6387 @item ABSOLUTE(@var{exp})
6388 @kindex ABSOLUTE(@var{exp})
6389 @cindex expression, absolute
6390 Return the absolute (non-relocatable, as opposed to non-negative) value
6391 of the expression @var{exp}. Primarily useful to assign an absolute
6392 value to a symbol within a section definition, where symbol values are
6393 normally section relative. @xref{Expression Section}.
6395 @item ADDR(@var{section})
6396 @kindex ADDR(@var{section})
6397 @cindex section address in expression
6398 Return the address (VMA) of the named @var{section}. Your
6399 script must previously have defined the location of that section. In
6400 the following example, @code{start_of_output_1}, @code{symbol_1} and
6401 @code{symbol_2} are assigned equivalent values, except that
6402 @code{symbol_1} will be relative to the @code{.output1} section while
6403 the other two will be absolute:
6409 start_of_output_1 = ABSOLUTE(.);
6414 symbol_1 = ADDR(.output1);
6415 symbol_2 = start_of_output_1;
6421 @item ALIGN(@var{align})
6422 @itemx ALIGN(@var{exp},@var{align})
6423 @kindex ALIGN(@var{align})
6424 @kindex ALIGN(@var{exp},@var{align})
6425 @cindex round up location counter
6426 @cindex align location counter
6427 @cindex round up expression
6428 @cindex align expression
6429 Return the location counter (@code{.}) or arbitrary expression aligned
6430 to the next @var{align} boundary. The single operand @code{ALIGN}
6431 doesn't change the value of the location counter---it just does
6432 arithmetic on it. The two operand @code{ALIGN} allows an arbitrary
6433 expression to be aligned upwards (@code{ALIGN(@var{align})} is
6434 equivalent to @code{ALIGN(ABSOLUTE(.), @var{align})}).
6436 Here is an example which aligns the output @code{.data} section to the
6437 next @code{0x2000} byte boundary after the preceding section and sets a
6438 variable within the section to the next @code{0x8000} boundary after the
6443 .data ALIGN(0x2000): @{
6445 variable = ALIGN(0x8000);
6451 The first use of @code{ALIGN} in this example specifies the location of
6452 a section because it is used as the optional @var{address} attribute of
6453 a section definition (@pxref{Output Section Address}). The second use
6454 of @code{ALIGN} is used to defines the value of a symbol.
6456 The builtin function @code{NEXT} is closely related to @code{ALIGN}.
6458 @item ALIGNOF(@var{section})
6459 @kindex ALIGNOF(@var{section})
6460 @cindex section alignment
6461 Return the alignment in bytes of the named @var{section}, if that section has
6462 been allocated. If the section has not been allocated when this is
6463 evaluated, the linker will report an error. In the following example,
6464 the alignment of the @code{.output} section is stored as the first
6465 value in that section.
6470 LONG (ALIGNOF (.output))
6477 @item BLOCK(@var{exp})
6478 @kindex BLOCK(@var{exp})
6479 This is a synonym for @code{ALIGN}, for compatibility with older linker
6480 scripts. It is most often seen when setting the address of an output
6483 @item DATA_SEGMENT_ALIGN(@var{maxpagesize}, @var{commonpagesize})
6484 @kindex DATA_SEGMENT_ALIGN(@var{maxpagesize}, @var{commonpagesize})
6485 This is equivalent to either
6487 (ALIGN(@var{maxpagesize}) + (. & (@var{maxpagesize} - 1)))
6491 (ALIGN(@var{maxpagesize})
6492 + ((. + @var{commonpagesize} - 1) & (@var{maxpagesize} - @var{commonpagesize})))
6495 depending on whether the latter uses fewer @var{commonpagesize} sized pages
6496 for the data segment (area between the result of this expression and
6497 @code{DATA_SEGMENT_END}) than the former or not.
6498 If the latter form is used, it means @var{commonpagesize} bytes of runtime
6499 memory will be saved at the expense of up to @var{commonpagesize} wasted
6500 bytes in the on-disk file.
6502 This expression can only be used directly in @code{SECTIONS} commands, not in
6503 any output section descriptions and only once in the linker script.
6504 @var{commonpagesize} should be less or equal to @var{maxpagesize} and should
6505 be the system page size the object wants to be optimized for while still
6506 running on system page sizes up to @var{maxpagesize}. Note however
6507 that @samp{-z relro} protection will not be effective if the system
6508 page size is larger than @var{commonpagesize}.
6513 . = DATA_SEGMENT_ALIGN(0x10000, 0x2000);
6516 @item DATA_SEGMENT_END(@var{exp})
6517 @kindex DATA_SEGMENT_END(@var{exp})
6518 This defines the end of data segment for @code{DATA_SEGMENT_ALIGN}
6519 evaluation purposes.
6522 . = DATA_SEGMENT_END(.);
6525 @item DATA_SEGMENT_RELRO_END(@var{offset}, @var{exp})
6526 @kindex DATA_SEGMENT_RELRO_END(@var{offset}, @var{exp})
6527 This defines the end of the @code{PT_GNU_RELRO} segment when
6528 @samp{-z relro} option is used.
6529 When @samp{-z relro} option is not present, @code{DATA_SEGMENT_RELRO_END}
6530 does nothing, otherwise @code{DATA_SEGMENT_ALIGN} is padded so that
6531 @var{exp} + @var{offset} is aligned to the @var{commonpagesize}
6532 argument given to @code{DATA_SEGMENT_ALIGN}. If present in the linker
6533 script, it must be placed between @code{DATA_SEGMENT_ALIGN} and
6534 @code{DATA_SEGMENT_END}. Evaluates to the second argument plus any
6535 padding needed at the end of the @code{PT_GNU_RELRO} segment due to
6539 . = DATA_SEGMENT_RELRO_END(24, .);
6542 @item DEFINED(@var{symbol})
6543 @kindex DEFINED(@var{symbol})
6544 @cindex symbol defaults
6545 Return 1 if @var{symbol} is in the linker global symbol table and is
6546 defined before the statement using DEFINED in the script, otherwise
6547 return 0. You can use this function to provide
6548 default values for symbols. For example, the following script fragment
6549 shows how to set a global symbol @samp{begin} to the first location in
6550 the @samp{.text} section---but if a symbol called @samp{begin} already
6551 existed, its value is preserved:
6557 begin = DEFINED(begin) ? begin : . ;
6565 @item LENGTH(@var{memory})
6566 @kindex LENGTH(@var{memory})
6567 Return the length of the memory region named @var{memory}.
6569 @item LOADADDR(@var{section})
6570 @kindex LOADADDR(@var{section})
6571 @cindex section load address in expression
6572 Return the absolute LMA of the named @var{section}. (@pxref{Output
6575 @item LOG2CEIL(@var{exp})
6576 @kindex LOG2CEIL(@var{exp})
6577 Return the binary logarithm of @var{exp} rounded towards infinity.
6578 @code{LOG2CEIL(0)} returns 0.
6581 @item MAX(@var{exp1}, @var{exp2})
6582 Returns the maximum of @var{exp1} and @var{exp2}.
6585 @item MIN(@var{exp1}, @var{exp2})
6586 Returns the minimum of @var{exp1} and @var{exp2}.
6588 @item NEXT(@var{exp})
6589 @kindex NEXT(@var{exp})
6590 @cindex unallocated address, next
6591 Return the next unallocated address that is a multiple of @var{exp}.
6592 This function is closely related to @code{ALIGN(@var{exp})}; unless you
6593 use the @code{MEMORY} command to define discontinuous memory for the
6594 output file, the two functions are equivalent.
6596 @item ORIGIN(@var{memory})
6597 @kindex ORIGIN(@var{memory})
6598 Return the origin of the memory region named @var{memory}.
6600 @item SEGMENT_START(@var{segment}, @var{default})
6601 @kindex SEGMENT_START(@var{segment}, @var{default})
6602 Return the base address of the named @var{segment}. If an explicit
6603 value has already been given for this segment (with a command-line
6604 @samp{-T} option) then that value will be returned otherwise the value
6605 will be @var{default}. At present, the @samp{-T} command-line option
6606 can only be used to set the base address for the ``text'', ``data'', and
6607 ``bss'' sections, but you can use @code{SEGMENT_START} with any segment
6610 @item SIZEOF(@var{section})
6611 @kindex SIZEOF(@var{section})
6612 @cindex section size
6613 Return the size in bytes of the named @var{section}, if that section has
6614 been allocated. If the section has not been allocated when this is
6615 evaluated, the linker will report an error. In the following example,
6616 @code{symbol_1} and @code{symbol_2} are assigned identical values:
6625 symbol_1 = .end - .start ;
6626 symbol_2 = SIZEOF(.output);
6631 @item SIZEOF_HEADERS
6632 @itemx sizeof_headers
6633 @kindex SIZEOF_HEADERS
6635 Return the size in bytes of the output file's headers. This is
6636 information which appears at the start of the output file. You can use
6637 this number when setting the start address of the first section, if you
6638 choose, to facilitate paging.
6640 @cindex not enough room for program headers
6641 @cindex program headers, not enough room
6642 When producing an ELF output file, if the linker script uses the
6643 @code{SIZEOF_HEADERS} builtin function, the linker must compute the
6644 number of program headers before it has determined all the section
6645 addresses and sizes. If the linker later discovers that it needs
6646 additional program headers, it will report an error @samp{not enough
6647 room for program headers}. To avoid this error, you must avoid using
6648 the @code{SIZEOF_HEADERS} function, or you must rework your linker
6649 script to avoid forcing the linker to use additional program headers, or
6650 you must define the program headers yourself using the @code{PHDRS}
6651 command (@pxref{PHDRS}).
6654 @node Implicit Linker Scripts
6655 @section Implicit Linker Scripts
6656 @cindex implicit linker scripts
6657 If you specify a linker input file which the linker can not recognize as
6658 an object file or an archive file, it will try to read the file as a
6659 linker script. If the file can not be parsed as a linker script, the
6660 linker will report an error.
6662 An implicit linker script will not replace the default linker script.
6664 Typically an implicit linker script would contain only symbol
6665 assignments, or the @code{INPUT}, @code{GROUP}, or @code{VERSION}
6668 Any input files read because of an implicit linker script will be read
6669 at the position in the command line where the implicit linker script was
6670 read. This can affect archive searching.
6673 @node Machine Dependent
6674 @chapter Machine Dependent Features
6676 @cindex machine dependencies
6677 @command{ld} has additional features on some platforms; the following
6678 sections describe them. Machines where @command{ld} has no additional
6679 functionality are not listed.
6683 * H8/300:: @command{ld} and the H8/300
6686 * M68HC11/68HC12:: @code{ld} and the Motorola 68HC11 and 68HC12 families
6689 * ARM:: @command{ld} and the ARM family
6692 * HPPA ELF32:: @command{ld} and HPPA 32-bit ELF
6695 * M68K:: @command{ld} and the Motorola 68K family
6698 * MIPS:: @command{ld} and the MIPS family
6701 * MMIX:: @command{ld} and MMIX
6704 * MSP430:: @command{ld} and MSP430
6707 * NDS32:: @command{ld} and NDS32
6710 * Nios II:: @command{ld} and the Altera Nios II
6713 * PowerPC ELF32:: @command{ld} and PowerPC 32-bit ELF Support
6716 * PowerPC64 ELF64:: @command{ld} and PowerPC64 64-bit ELF Support
6719 * S/390 ELF:: @command{ld} and S/390 ELF Support
6722 * SPU ELF:: @command{ld} and SPU ELF Support
6725 * TI COFF:: @command{ld} and TI COFF
6728 * WIN32:: @command{ld} and WIN32 (cygwin/mingw)
6731 * Xtensa:: @command{ld} and Xtensa Processors
6742 @section @command{ld} and the H8/300
6744 @cindex H8/300 support
6745 For the H8/300, @command{ld} can perform these global optimizations when
6746 you specify the @samp{--relax} command-line option.
6749 @cindex relaxing on H8/300
6750 @item relaxing address modes
6751 @command{ld} finds all @code{jsr} and @code{jmp} instructions whose
6752 targets are within eight bits, and turns them into eight-bit
6753 program-counter relative @code{bsr} and @code{bra} instructions,
6756 @cindex synthesizing on H8/300
6757 @item synthesizing instructions
6758 @c FIXME: specifically mov.b, or any mov instructions really? -> mov.b only, at least on H8, H8H, H8S
6759 @command{ld} finds all @code{mov.b} instructions which use the
6760 sixteen-bit absolute address form, but refer to the top
6761 page of memory, and changes them to use the eight-bit address form.
6762 (That is: the linker turns @samp{mov.b @code{@@}@var{aa}:16} into
6763 @samp{mov.b @code{@@}@var{aa}:8} whenever the address @var{aa} is in the
6764 top page of memory).
6766 @command{ld} finds all @code{mov} instructions which use the register
6767 indirect with 32-bit displacement addressing mode, but use a small
6768 displacement inside 16-bit displacement range, and changes them to use
6769 the 16-bit displacement form. (That is: the linker turns @samp{mov.b
6770 @code{@@}@var{d}:32,ERx} into @samp{mov.b @code{@@}@var{d}:16,ERx}
6771 whenever the displacement @var{d} is in the 16 bit signed integer
6772 range. Only implemented in ELF-format ld).
6774 @item bit manipulation instructions
6775 @command{ld} finds all bit manipulation instructions like @code{band, bclr,
6776 biand, bild, bior, bist, bixor, bld, bnot, bor, bset, bst, btst, bxor}
6777 which use 32 bit and 16 bit absolute address form, but refer to the top
6778 page of memory, and changes them to use the 8 bit address form.
6779 (That is: the linker turns @samp{bset #xx:3,@code{@@}@var{aa}:32} into
6780 @samp{bset #xx:3,@code{@@}@var{aa}:8} whenever the address @var{aa} is in
6781 the top page of memory).
6783 @item system control instructions
6784 @command{ld} finds all @code{ldc.w, stc.w} instructions which use the
6785 32 bit absolute address form, but refer to the top page of memory, and
6786 changes them to use 16 bit address form.
6787 (That is: the linker turns @samp{ldc.w @code{@@}@var{aa}:32,ccr} into
6788 @samp{ldc.w @code{@@}@var{aa}:16,ccr} whenever the address @var{aa} is in
6789 the top page of memory).
6799 @c This stuff is pointless to say unless you're especially concerned
6800 @c with Renesas chips; don't enable it for generic case, please.
6802 @chapter @command{ld} and Other Renesas Chips
6804 @command{ld} also supports the Renesas (formerly Hitachi) H8/300H,
6805 H8/500, and SH chips. No special features, commands, or command-line
6806 options are required for these chips.
6820 @node M68HC11/68HC12
6821 @section @command{ld} and the Motorola 68HC11 and 68HC12 families
6823 @cindex M68HC11 and 68HC12 support
6825 @subsection Linker Relaxation
6827 For the Motorola 68HC11, @command{ld} can perform these global
6828 optimizations when you specify the @samp{--relax} command-line option.
6831 @cindex relaxing on M68HC11
6832 @item relaxing address modes
6833 @command{ld} finds all @code{jsr} and @code{jmp} instructions whose
6834 targets are within eight bits, and turns them into eight-bit
6835 program-counter relative @code{bsr} and @code{bra} instructions,
6838 @command{ld} also looks at all 16-bit extended addressing modes and
6839 transforms them in a direct addressing mode when the address is in
6840 page 0 (between 0 and 0x0ff).
6842 @item relaxing gcc instruction group
6843 When @command{gcc} is called with @option{-mrelax}, it can emit group
6844 of instructions that the linker can optimize to use a 68HC11 direct
6845 addressing mode. These instructions consists of @code{bclr} or
6846 @code{bset} instructions.
6850 @subsection Trampoline Generation
6852 @cindex trampoline generation on M68HC11
6853 @cindex trampoline generation on M68HC12
6854 For 68HC11 and 68HC12, @command{ld} can generate trampoline code to
6855 call a far function using a normal @code{jsr} instruction. The linker
6856 will also change the relocation to some far function to use the
6857 trampoline address instead of the function address. This is typically the
6858 case when a pointer to a function is taken. The pointer will in fact
6859 point to the function trampoline.
6867 @section @command{ld} and the ARM family
6869 @cindex ARM interworking support
6870 @kindex --support-old-code
6871 For the ARM, @command{ld} will generate code stubs to allow functions calls
6872 between ARM and Thumb code. These stubs only work with code that has
6873 been compiled and assembled with the @samp{-mthumb-interwork} command
6874 line option. If it is necessary to link with old ARM object files or
6875 libraries, which have not been compiled with the -mthumb-interwork
6876 option then the @samp{--support-old-code} command-line switch should be
6877 given to the linker. This will make it generate larger stub functions
6878 which will work with non-interworking aware ARM code. Note, however,
6879 the linker does not support generating stubs for function calls to
6880 non-interworking aware Thumb code.
6882 @cindex thumb entry point
6883 @cindex entry point, thumb
6884 @kindex --thumb-entry=@var{entry}
6885 The @samp{--thumb-entry} switch is a duplicate of the generic
6886 @samp{--entry} switch, in that it sets the program's starting address.
6887 But it also sets the bottom bit of the address, so that it can be
6888 branched to using a BX instruction, and the program will start
6889 executing in Thumb mode straight away.
6891 @cindex PE import table prefixing
6892 @kindex --use-nul-prefixed-import-tables
6893 The @samp{--use-nul-prefixed-import-tables} switch is specifying, that
6894 the import tables idata4 and idata5 have to be generated with a zero
6895 element prefix for import libraries. This is the old style to generate
6896 import tables. By default this option is turned off.
6900 The @samp{--be8} switch instructs @command{ld} to generate BE8 format
6901 executables. This option is only valid when linking big-endian
6902 objects - ie ones which have been assembled with the @option{-EB}
6903 option. The resulting image will contain big-endian data and
6907 @kindex --target1-rel
6908 @kindex --target1-abs
6909 The @samp{R_ARM_TARGET1} relocation is typically used for entries in the
6910 @samp{.init_array} section. It is interpreted as either @samp{R_ARM_REL32}
6911 or @samp{R_ARM_ABS32}, depending on the target. The @samp{--target1-rel}
6912 and @samp{--target1-abs} switches override the default.
6915 @kindex --target2=@var{type}
6916 The @samp{--target2=type} switch overrides the default definition of the
6917 @samp{R_ARM_TARGET2} relocation. Valid values for @samp{type}, their
6918 meanings, and target defaults are as follows:
6921 @samp{R_ARM_REL32} (arm*-*-elf, arm*-*-eabi)
6923 @samp{R_ARM_ABS32} (arm*-*-symbianelf)
6925 @samp{R_ARM_GOT_PREL} (arm*-*-linux, arm*-*-*bsd)
6930 The @samp{R_ARM_V4BX} relocation (defined by the ARM AAELF
6931 specification) enables objects compiled for the ARMv4 architecture to be
6932 interworking-safe when linked with other objects compiled for ARMv4t, but
6933 also allows pure ARMv4 binaries to be built from the same ARMv4 objects.
6935 In the latter case, the switch @option{--fix-v4bx} must be passed to the
6936 linker, which causes v4t @code{BX rM} instructions to be rewritten as
6937 @code{MOV PC,rM}, since v4 processors do not have a @code{BX} instruction.
6939 In the former case, the switch should not be used, and @samp{R_ARM_V4BX}
6940 relocations are ignored.
6942 @cindex FIX_V4BX_INTERWORKING
6943 @kindex --fix-v4bx-interworking
6944 Replace @code{BX rM} instructions identified by @samp{R_ARM_V4BX}
6945 relocations with a branch to the following veneer:
6953 This allows generation of libraries/applications that work on ARMv4 cores
6954 and are still interworking safe. Note that the above veneer clobbers the
6955 condition flags, so may cause incorrect program behavior in rare cases.
6959 The @samp{--use-blx} switch enables the linker to use ARM/Thumb
6960 BLX instructions (available on ARMv5t and above) in various
6961 situations. Currently it is used to perform calls via the PLT from Thumb
6962 code using BLX rather than using BX and a mode-switching stub before
6963 each PLT entry. This should lead to such calls executing slightly faster.
6965 This option is enabled implicitly for SymbianOS, so there is no need to
6966 specify it if you are using that target.
6968 @cindex VFP11_DENORM_FIX
6969 @kindex --vfp11-denorm-fix
6970 The @samp{--vfp11-denorm-fix} switch enables a link-time workaround for a
6971 bug in certain VFP11 coprocessor hardware, which sometimes allows
6972 instructions with denorm operands (which must be handled by support code)
6973 to have those operands overwritten by subsequent instructions before
6974 the support code can read the intended values.
6976 The bug may be avoided in scalar mode if you allow at least one
6977 intervening instruction between a VFP11 instruction which uses a register
6978 and another instruction which writes to the same register, or at least two
6979 intervening instructions if vector mode is in use. The bug only affects
6980 full-compliance floating-point mode: you do not need this workaround if
6981 you are using "runfast" mode. Please contact ARM for further details.
6983 If you know you are using buggy VFP11 hardware, you can
6984 enable this workaround by specifying the linker option
6985 @samp{--vfp-denorm-fix=scalar} if you are using the VFP11 scalar
6986 mode only, or @samp{--vfp-denorm-fix=vector} if you are using
6987 vector mode (the latter also works for scalar code). The default is
6988 @samp{--vfp-denorm-fix=none}.
6990 If the workaround is enabled, instructions are scanned for
6991 potentially-troublesome sequences, and a veneer is created for each
6992 such sequence which may trigger the erratum. The veneer consists of the
6993 first instruction of the sequence and a branch back to the subsequent
6994 instruction. The original instruction is then replaced with a branch to
6995 the veneer. The extra cycles required to call and return from the veneer
6996 are sufficient to avoid the erratum in both the scalar and vector cases.
6998 @cindex ARM1176 erratum workaround
6999 @kindex --fix-arm1176
7000 @kindex --no-fix-arm1176
7001 The @samp{--fix-arm1176} switch enables a link-time workaround for an erratum
7002 in certain ARM1176 processors. The workaround is enabled by default if you
7003 are targeting ARM v6 (excluding ARM v6T2) or earlier. It can be disabled
7004 unconditionally by specifying @samp{--no-fix-arm1176}.
7006 Further information is available in the ``ARM1176JZ-S and ARM1176JZF-S
7007 Programmer Advice Notice'' available on the ARM documentation website at:
7008 http://infocenter.arm.com/.
7010 @cindex STM32L4xx erratum workaround
7011 @kindex --fix-stm32l4xx-629360
7013 The @samp{--fix-stm32l4xx-629360} switch enables a link-time
7014 workaround for a bug in the bus matrix / memory controller for some of
7015 the STM32 Cortex-M4 based products (STM32L4xx). When accessing
7016 off-chip memory via the affected bus for bus reads of 9 words or more,
7017 the bus can generate corrupt data and/or abort. These are only
7018 core-initiated accesses (not DMA), and might affect any access:
7019 integer loads such as LDM, POP and floating-point loads such as VLDM,
7020 VPOP. Stores are not affected.
7022 The bug can be avoided by splitting memory accesses into the
7023 necessary chunks to keep bus reads below 8 words.
7025 The workaround is not enabled by default, this is equivalent to use
7026 @samp{--fix-stm32l4xx-629360=none}. If you know you are using buggy
7027 STM32L4xx hardware, you can enable the workaround by specifying the
7028 linker option @samp{--fix-stm32l4xx-629360}, or the equivalent
7029 @samp{--fix-stm32l4xx-629360=default}.
7031 If the workaround is enabled, instructions are scanned for
7032 potentially-troublesome sequences, and a veneer is created for each
7033 such sequence which may trigger the erratum. The veneer consists in a
7034 replacement sequence emulating the behaviour of the original one and a
7035 branch back to the subsequent instruction. The original instruction is
7036 then replaced with a branch to the veneer.
7038 The workaround does not always preserve the memory access order for
7039 the LDMDB instruction, when the instruction loads the PC.
7041 The workaround is not able to handle problematic instructions when
7042 they are in the middle of an IT block, since a branch is not allowed
7043 there. In that case, the linker reports a warning and no replacement
7046 The workaround is not able to replace problematic instructions with a
7047 PC-relative branch instruction if the @samp{.text} section is too
7048 large. In that case, when the branch that replaces the original code
7049 cannot be encoded, the linker reports a warning and no replacement
7052 @cindex NO_ENUM_SIZE_WARNING
7053 @kindex --no-enum-size-warning
7054 The @option{--no-enum-size-warning} switch prevents the linker from
7055 warning when linking object files that specify incompatible EABI
7056 enumeration size attributes. For example, with this switch enabled,
7057 linking of an object file using 32-bit enumeration values with another
7058 using enumeration values fitted into the smallest possible space will
7061 @cindex NO_WCHAR_SIZE_WARNING
7062 @kindex --no-wchar-size-warning
7063 The @option{--no-wchar-size-warning} switch prevents the linker from
7064 warning when linking object files that specify incompatible EABI
7065 @code{wchar_t} size attributes. For example, with this switch enabled,
7066 linking of an object file using 32-bit @code{wchar_t} values with another
7067 using 16-bit @code{wchar_t} values will not be diagnosed.
7070 @kindex --pic-veneer
7071 The @samp{--pic-veneer} switch makes the linker use PIC sequences for
7072 ARM/Thumb interworking veneers, even if the rest of the binary
7073 is not PIC. This avoids problems on uClinux targets where
7074 @samp{--emit-relocs} is used to generate relocatable binaries.
7076 @cindex STUB_GROUP_SIZE
7077 @kindex --stub-group-size=@var{N}
7078 The linker will automatically generate and insert small sequences of
7079 code into a linked ARM ELF executable whenever an attempt is made to
7080 perform a function call to a symbol that is too far away. The
7081 placement of these sequences of instructions - called stubs - is
7082 controlled by the command-line option @option{--stub-group-size=N}.
7083 The placement is important because a poor choice can create a need for
7084 duplicate stubs, increasing the code size. The linker will try to
7085 group stubs together in order to reduce interruptions to the flow of
7086 code, but it needs guidance as to how big these groups should be and
7087 where they should be placed.
7089 The value of @samp{N}, the parameter to the
7090 @option{--stub-group-size=} option controls where the stub groups are
7091 placed. If it is negative then all stubs are placed after the first
7092 branch that needs them. If it is positive then the stubs can be
7093 placed either before or after the branches that need them. If the
7094 value of @samp{N} is 1 (either +1 or -1) then the linker will choose
7095 exactly where to place groups of stubs, using its built in heuristics.
7096 A value of @samp{N} greater than 1 (or smaller than -1) tells the
7097 linker that a single group of stubs can service at most @samp{N} bytes
7098 from the input sections.
7100 The default, if @option{--stub-group-size=} is not specified, is
7103 Farcalls stubs insertion is fully supported for the ARM-EABI target
7104 only, because it relies on object files properties not present
7107 @cindex Cortex-A8 erratum workaround
7108 @kindex --fix-cortex-a8
7109 @kindex --no-fix-cortex-a8
7110 The @samp{--fix-cortex-a8} switch enables a link-time workaround for an erratum in certain Cortex-A8 processors. The workaround is enabled by default if you are targeting the ARM v7-A architecture profile. It can be enabled otherwise by specifying @samp{--fix-cortex-a8}, or disabled unconditionally by specifying @samp{--no-fix-cortex-a8}.
7112 The erratum only affects Thumb-2 code. Please contact ARM for further details.
7114 @cindex Cortex-A53 erratum 835769 workaround
7115 @kindex --fix-cortex-a53-835769
7116 @kindex --no-fix-cortex-a53-835769
7117 The @samp{--fix-cortex-a53-835769} switch enables a link-time workaround for erratum 835769 present on certain early revisions of Cortex-A53 processors. The workaround is disabled by default. It can be enabled by specifying @samp{--fix-cortex-a53-835769}, or disabled unconditionally by specifying @samp{--no-fix-cortex-a53-835769}.
7119 Please contact ARM for further details.
7121 @kindex --merge-exidx-entries
7122 @kindex --no-merge-exidx-entries
7123 @cindex Merging exidx entries
7124 The @samp{--no-merge-exidx-entries} switch disables the merging of adjacent exidx entries in debuginfo.
7127 @cindex 32-bit PLT entries
7128 The @samp{--long-plt} option enables the use of 16 byte PLT entries
7129 which support up to 4Gb of code. The default is to use 12 byte PLT
7130 entries which only support 512Mb of code.
7132 @kindex --no-apply-dynamic-relocs
7133 @cindex AArch64 rela addend
7134 The @samp{--no-apply-dynamic-relocs} option makes AArch64 linker do not apply
7135 link-time values for dynamic relocations.
7137 @cindex Placement of SG veneers
7138 All SG veneers are placed in the special output section @code{.gnu.sgstubs}.
7139 Its start address must be set, either with the command-line option
7140 @samp{--section-start} or in a linker script, to indicate where to place these
7143 @kindex --cmse-implib
7144 @cindex Secure gateway import library
7145 The @samp{--cmse-implib} option requests that the import libraries
7146 specified by the @samp{--out-implib} and @samp{--in-implib} options are
7147 secure gateway import libraries, suitable for linking a non-secure
7148 executable against secure code as per ARMv8-M Security Extensions.
7150 @kindex --in-implib=@var{file}
7151 @cindex Input import library
7152 The @samp{--in-implib=file} specifies an input import library whose symbols
7153 must keep the same address in the executable being produced. A warning is
7154 given if no @samp{--out-implib} is given but new symbols have been introduced
7155 in the executable that should be listed in its import library. Otherwise, if
7156 @samp{--out-implib} is specified, the symbols are added to the output import
7157 library. A warning is also given if some symbols present in the input import
7158 library have disappeared from the executable. This option is only effective
7159 for Secure Gateway import libraries, ie. when @samp{--cmse-implib} is
7173 @section @command{ld} and HPPA 32-bit ELF Support
7174 @cindex HPPA multiple sub-space stubs
7175 @kindex --multi-subspace
7176 When generating a shared library, @command{ld} will by default generate
7177 import stubs suitable for use with a single sub-space application.
7178 The @samp{--multi-subspace} switch causes @command{ld} to generate export
7179 stubs, and different (larger) import stubs suitable for use with
7180 multiple sub-spaces.
7182 @cindex HPPA stub grouping
7183 @kindex --stub-group-size=@var{N}
7184 Long branch stubs and import/export stubs are placed by @command{ld} in
7185 stub sections located between groups of input sections.
7186 @samp{--stub-group-size} specifies the maximum size of a group of input
7187 sections handled by one stub section. Since branch offsets are signed,
7188 a stub section may serve two groups of input sections, one group before
7189 the stub section, and one group after it. However, when using
7190 conditional branches that require stubs, it may be better (for branch
7191 prediction) that stub sections only serve one group of input sections.
7192 A negative value for @samp{N} chooses this scheme, ensuring that
7193 branches to stubs always use a negative offset. Two special values of
7194 @samp{N} are recognized, @samp{1} and @samp{-1}. These both instruct
7195 @command{ld} to automatically size input section groups for the branch types
7196 detected, with the same behaviour regarding stub placement as other
7197 positive or negative values of @samp{N} respectively.
7199 Note that @samp{--stub-group-size} does not split input sections. A
7200 single input section larger than the group size specified will of course
7201 create a larger group (of one section). If input sections are too
7202 large, it may not be possible for a branch to reach its stub.
7215 @section @command{ld} and the Motorola 68K family
7217 @cindex Motorola 68K GOT generation
7218 @kindex --got=@var{type}
7219 The @samp{--got=@var{type}} option lets you choose the GOT generation scheme.
7220 The choices are @samp{single}, @samp{negative}, @samp{multigot} and
7221 @samp{target}. When @samp{target} is selected the linker chooses
7222 the default GOT generation scheme for the current target.
7223 @samp{single} tells the linker to generate a single GOT with
7224 entries only at non-negative offsets.
7225 @samp{negative} instructs the linker to generate a single GOT with
7226 entries at both negative and positive offsets. Not all environments
7228 @samp{multigot} allows the linker to generate several GOTs in the
7229 output file. All GOT references from a single input object
7230 file access the same GOT, but references from different input object
7231 files might access different GOTs. Not all environments support such GOTs.
7244 @section @command{ld} and the MIPS family
7246 @cindex MIPS microMIPS instruction choice selection
7249 The @samp{--insn32} and @samp{--no-insn32} options control the choice of
7250 microMIPS instructions used in code generated by the linker, such as that
7251 in the PLT or lazy binding stubs, or in relaxation. If @samp{--insn32} is
7252 used, then the linker only uses 32-bit instruction encodings. By default
7253 or if @samp{--no-insn32} is used, all instruction encodings are used,
7254 including 16-bit ones where possible.
7256 @cindex MIPS branch relocation check control
7257 @kindex --ignore-branch-isa
7258 @kindex --no-ignore-branch-isa
7259 The @samp{--ignore-branch-isa} and @samp{--no-ignore-branch-isa} options
7260 control branch relocation checks for invalid ISA mode transitions. If
7261 @samp{--ignore-branch-isa} is used, then the linker accepts any branch
7262 relocations and any ISA mode transition required is lost in relocation
7263 calculation, except for some cases of @code{BAL} instructions which meet
7264 relaxation conditions and are converted to equivalent @code{JALX}
7265 instructions as the associated relocation is calculated. By default
7266 or if @samp{--no-ignore-branch-isa} is used a check is made causing
7267 the loss of an ISA mode transition to produce an error.
7280 @section @code{ld} and MMIX
7281 For MMIX, there is a choice of generating @code{ELF} object files or
7282 @code{mmo} object files when linking. The simulator @code{mmix}
7283 understands the @code{mmo} format. The binutils @code{objcopy} utility
7284 can translate between the two formats.
7286 There is one special section, the @samp{.MMIX.reg_contents} section.
7287 Contents in this section is assumed to correspond to that of global
7288 registers, and symbols referring to it are translated to special symbols,
7289 equal to registers. In a final link, the start address of the
7290 @samp{.MMIX.reg_contents} section corresponds to the first allocated
7291 global register multiplied by 8. Register @code{$255} is not included in
7292 this section; it is always set to the program entry, which is at the
7293 symbol @code{Main} for @code{mmo} files.
7295 Global symbols with the prefix @code{__.MMIX.start.}, for example
7296 @code{__.MMIX.start..text} and @code{__.MMIX.start..data} are special.
7297 The default linker script uses these to set the default start address
7300 Initial and trailing multiples of zero-valued 32-bit words in a section,
7301 are left out from an mmo file.
7314 @section @code{ld} and MSP430
7315 For the MSP430 it is possible to select the MPU architecture. The flag @samp{-m [mpu type]}
7316 will select an appropriate linker script for selected MPU type. (To get a list of known MPUs
7317 just pass @samp{-m help} option to the linker).
7319 @cindex MSP430 extra sections
7320 The linker will recognize some extra sections which are MSP430 specific:
7323 @item @samp{.vectors}
7324 Defines a portion of ROM where interrupt vectors located.
7326 @item @samp{.bootloader}
7327 Defines the bootloader portion of the ROM (if applicable). Any code
7328 in this section will be uploaded to the MPU.
7330 @item @samp{.infomem}
7331 Defines an information memory section (if applicable). Any code in
7332 this section will be uploaded to the MPU.
7334 @item @samp{.infomemnobits}
7335 This is the same as the @samp{.infomem} section except that any code
7336 in this section will not be uploaded to the MPU.
7338 @item @samp{.noinit}
7339 Denotes a portion of RAM located above @samp{.bss} section.
7341 The last two sections are used by gcc.
7345 @cindex MSP430 Options
7346 @kindex --code-region
7347 @item --code-region=[either,lower,upper,none]
7348 This will transform .text* sections to [either,lower,upper].text* sections. The
7349 argument passed to GCC for -mcode-region is propagated to the linker
7352 @kindex --data-region
7353 @item --data-region=[either,lower,upper,none]
7354 This will transform .data*, .bss* and .rodata* sections to
7355 [either,lower,upper].[data,bss,rodata]* sections. The argument passed to GCC
7356 for -mdata-region is propagated to the linker using this option.
7358 @kindex --disable-sec-transformation
7359 @item --disable-sec-transformation
7360 Prevent the transformation of sections as specified by the @code{--code-region}
7361 and @code{--data-region} options.
7362 This is useful if you are compiling and linking using a single call to the GCC
7363 wrapper, and want to compile the source files using -m[code,data]-region but
7364 not transform the sections for prebuilt libraries and objects.
7378 @section @code{ld} and NDS32
7379 @kindex relaxing on NDS32
7380 For NDS32, there are some options to select relaxation behavior. The linker
7381 relaxes objects according to these options.
7384 @item @samp{--m[no-]fp-as-gp}
7385 Disable/enable fp-as-gp relaxation.
7387 @item @samp{--mexport-symbols=FILE}
7388 Exporting symbols and their address into FILE as linker script.
7390 @item @samp{--m[no-]ex9}
7391 Disable/enable link-time EX9 relaxation.
7393 @item @samp{--mexport-ex9=FILE}
7394 Export the EX9 table after linking.
7396 @item @samp{--mimport-ex9=FILE}
7397 Import the Ex9 table for EX9 relaxation.
7399 @item @samp{--mupdate-ex9}
7400 Update the existing EX9 table.
7402 @item @samp{--mex9-limit=NUM}
7403 Maximum number of entries in the ex9 table.
7405 @item @samp{--mex9-loop-aware}
7406 Avoid generating the EX9 instruction inside the loop.
7408 @item @samp{--m[no-]ifc}
7409 Disable/enable the link-time IFC optimization.
7411 @item @samp{--mifc-loop-aware}
7412 Avoid generating the IFC instruction inside the loop.
7426 @section @command{ld} and the Altera Nios II
7427 @cindex Nios II call relaxation
7428 @kindex --relax on Nios II
7430 Call and immediate jump instructions on Nios II processors are limited to
7431 transferring control to addresses in the same 256MB memory segment,
7432 which may result in @command{ld} giving
7433 @samp{relocation truncated to fit} errors with very large programs.
7434 The command-line option @option{--relax} enables the generation of
7435 trampolines that can access the entire 32-bit address space for calls
7436 outside the normal @code{call} and @code{jmpi} address range. These
7437 trampolines are inserted at section boundaries, so may not themselves
7438 be reachable if an input section and its associated call trampolines are
7441 The @option{--relax} option is enabled by default unless @option{-r}
7442 is also specified. You can disable trampoline generation by using the
7443 @option{--no-relax} linker option. You can also disable this optimization
7444 locally by using the @samp{set .noat} directive in assembly-language
7445 source files, as the linker-inserted trampolines use the @code{at}
7446 register as a temporary.
7448 Note that the linker @option{--relax} option is independent of assembler
7449 relaxation options, and that using the GNU assembler's @option{-relax-all}
7450 option interferes with the linker's more selective call instruction relaxation.
7463 @section @command{ld} and PowerPC 32-bit ELF Support
7464 @cindex PowerPC long branches
7465 @kindex --relax on PowerPC
7466 Branches on PowerPC processors are limited to a signed 26-bit
7467 displacement, which may result in @command{ld} giving
7468 @samp{relocation truncated to fit} errors with very large programs.
7469 @samp{--relax} enables the generation of trampolines that can access
7470 the entire 32-bit address space. These trampolines are inserted at
7471 section boundaries, so may not themselves be reachable if an input
7472 section exceeds 33M in size. You may combine @samp{-r} and
7473 @samp{--relax} to add trampolines in a partial link. In that case
7474 both branches to undefined symbols and inter-section branches are also
7475 considered potentially out of range, and trampolines inserted.
7477 @cindex PowerPC ELF32 options
7482 Current PowerPC GCC accepts a @samp{-msecure-plt} option that
7483 generates code capable of using a newer PLT and GOT layout that has
7484 the security advantage of no executable section ever needing to be
7485 writable and no writable section ever being executable. PowerPC
7486 @command{ld} will generate this layout, including stubs to access the
7487 PLT, if all input files (including startup and static libraries) were
7488 compiled with @samp{-msecure-plt}. @samp{--bss-plt} forces the old
7489 BSS PLT (and GOT layout) which can give slightly better performance.
7491 @kindex --secure-plt
7493 @command{ld} will use the new PLT and GOT layout if it is linking new
7494 @samp{-fpic} or @samp{-fPIC} code, but does not do so automatically
7495 when linking non-PIC code. This option requests the new PLT and GOT
7496 layout. A warning will be given if some object file requires the old
7502 The new secure PLT and GOT are placed differently relative to other
7503 sections compared to older BSS PLT and GOT placement. The location of
7504 @code{.plt} must change because the new secure PLT is an initialized
7505 section while the old PLT is uninitialized. The reason for the
7506 @code{.got} change is more subtle: The new placement allows
7507 @code{.got} to be read-only in applications linked with
7508 @samp{-z relro -z now}. However, this placement means that
7509 @code{.sdata} cannot always be used in shared libraries, because the
7510 PowerPC ABI accesses @code{.sdata} in shared libraries from the GOT
7511 pointer. @samp{--sdata-got} forces the old GOT placement. PowerPC
7512 GCC doesn't use @code{.sdata} in shared libraries, so this option is
7513 really only useful for other compilers that may do so.
7515 @cindex PowerPC stub symbols
7516 @kindex --emit-stub-syms
7517 @item --emit-stub-syms
7518 This option causes @command{ld} to label linker stubs with a local
7519 symbol that encodes the stub type and destination.
7521 @cindex PowerPC TLS optimization
7522 @kindex --no-tls-optimize
7523 @item --no-tls-optimize
7524 PowerPC @command{ld} normally performs some optimization of code
7525 sequences used to access Thread-Local Storage. Use this option to
7526 disable the optimization.
7539 @node PowerPC64 ELF64
7540 @section @command{ld} and PowerPC64 64-bit ELF Support
7542 @cindex PowerPC64 ELF64 options
7544 @cindex PowerPC64 stub grouping
7545 @kindex --stub-group-size
7546 @item --stub-group-size
7547 Long branch stubs, PLT call stubs and TOC adjusting stubs are placed
7548 by @command{ld} in stub sections located between groups of input sections.
7549 @samp{--stub-group-size} specifies the maximum size of a group of input
7550 sections handled by one stub section. Since branch offsets are signed,
7551 a stub section may serve two groups of input sections, one group before
7552 the stub section, and one group after it. However, when using
7553 conditional branches that require stubs, it may be better (for branch
7554 prediction) that stub sections only serve one group of input sections.
7555 A negative value for @samp{N} chooses this scheme, ensuring that
7556 branches to stubs always use a negative offset. Two special values of
7557 @samp{N} are recognized, @samp{1} and @samp{-1}. These both instruct
7558 @command{ld} to automatically size input section groups for the branch types
7559 detected, with the same behaviour regarding stub placement as other
7560 positive or negative values of @samp{N} respectively.
7562 Note that @samp{--stub-group-size} does not split input sections. A
7563 single input section larger than the group size specified will of course
7564 create a larger group (of one section). If input sections are too
7565 large, it may not be possible for a branch to reach its stub.
7567 @cindex PowerPC64 stub symbols
7568 @kindex --emit-stub-syms
7569 @item --emit-stub-syms
7570 This option causes @command{ld} to label linker stubs with a local
7571 symbol that encodes the stub type and destination.
7573 @cindex PowerPC64 dot symbols
7575 @kindex --no-dotsyms
7578 These two options control how @command{ld} interprets version patterns
7579 in a version script. Older PowerPC64 compilers emitted both a
7580 function descriptor symbol with the same name as the function, and a
7581 code entry symbol with the name prefixed by a dot (@samp{.}). To
7582 properly version a function @samp{foo}, the version script thus needs
7583 to control both @samp{foo} and @samp{.foo}. The option
7584 @samp{--dotsyms}, on by default, automatically adds the required
7585 dot-prefixed patterns. Use @samp{--no-dotsyms} to disable this
7588 @cindex PowerPC64 register save/restore functions
7589 @kindex --save-restore-funcs
7590 @kindex --no-save-restore-funcs
7591 @item --save-restore-funcs
7592 @itemx --no-save-restore-funcs
7593 These two options control whether PowerPC64 @command{ld} automatically
7594 provides out-of-line register save and restore functions used by
7595 @samp{-Os} code. The default is to provide any such referenced
7596 function for a normal final link, and to not do so for a relocatable
7599 @cindex PowerPC64 TLS optimization
7600 @kindex --no-tls-optimize
7601 @item --no-tls-optimize
7602 PowerPC64 @command{ld} normally performs some optimization of code
7603 sequences used to access Thread-Local Storage. Use this option to
7604 disable the optimization.
7606 @cindex PowerPC64 __tls_get_addr optimization
7607 @kindex --tls-get-addr-optimize
7608 @kindex --no-tls-get-addr-optimize
7609 @kindex --tls-get-addr-regsave
7610 @kindex --no-tls-get-addr-regsave
7611 @item --tls-get-addr-optimize
7612 @itemx --no-tls-get-addr-optimize
7613 These options control how PowerPC64 @command{ld} uses a special
7614 stub to call __tls_get_addr. PowerPC64 glibc 2.22 and later support
7615 an optimization that allows the second and subsequent calls to
7616 @code{__tls_get_addr} for a given symbol to be resolved by the special
7617 stub without calling in to glibc. By default the linker enables
7618 generation of the stub when glibc advertises the availability of
7620 Using @option{--tls-get-addr-optimize} with an older glibc won't do
7621 much besides slow down your applications, but may be useful if linking
7622 an application against an older glibc with the expectation that it
7623 will normally be used on systems having a newer glibc.
7624 @option{--tls-get-addr-regsave} forces generation of a stub that saves
7625 and restores volatile registers around the call into glibc. Normally,
7626 this is done when the linker detects a call to __tls_get_addr_desc.
7627 Such calls then go via the register saving stub to __tls_get_addr_opt.
7628 @option{--no-tls-get-addr-regsave} disables generation of the
7631 @cindex PowerPC64 OPD optimization
7632 @kindex --no-opd-optimize
7633 @item --no-opd-optimize
7634 PowerPC64 @command{ld} normally removes @code{.opd} section entries
7635 corresponding to deleted link-once functions, or functions removed by
7636 the action of @samp{--gc-sections} or linker script @code{/DISCARD/}.
7637 Use this option to disable @code{.opd} optimization.
7639 @cindex PowerPC64 OPD spacing
7640 @kindex --non-overlapping-opd
7641 @item --non-overlapping-opd
7642 Some PowerPC64 compilers have an option to generate compressed
7643 @code{.opd} entries spaced 16 bytes apart, overlapping the third word,
7644 the static chain pointer (unused in C) with the first word of the next
7645 entry. This option expands such entries to the full 24 bytes.
7647 @cindex PowerPC64 TOC optimization
7648 @kindex --no-toc-optimize
7649 @item --no-toc-optimize
7650 PowerPC64 @command{ld} normally removes unused @code{.toc} section
7651 entries. Such entries are detected by examining relocations that
7652 reference the TOC in code sections. A reloc in a deleted code section
7653 marks a TOC word as unneeded, while a reloc in a kept code section
7654 marks a TOC word as needed. Since the TOC may reference itself, TOC
7655 relocs are also examined. TOC words marked as both needed and
7656 unneeded will of course be kept. TOC words without any referencing
7657 reloc are assumed to be part of a multi-word entry, and are kept or
7658 discarded as per the nearest marked preceding word. This works
7659 reliably for compiler generated code, but may be incorrect if assembly
7660 code is used to insert TOC entries. Use this option to disable the
7663 @cindex PowerPC64 multi-TOC
7664 @kindex --no-multi-toc
7665 @item --no-multi-toc
7666 If given any toc option besides @code{-mcmodel=medium} or
7667 @code{-mcmodel=large}, PowerPC64 GCC generates code for a TOC model
7669 entries are accessed with a 16-bit offset from r2. This limits the
7670 total TOC size to 64K. PowerPC64 @command{ld} extends this limit by
7671 grouping code sections such that each group uses less than 64K for its
7672 TOC entries, then inserts r2 adjusting stubs between inter-group
7673 calls. @command{ld} does not split apart input sections, so cannot
7674 help if a single input file has a @code{.toc} section that exceeds
7675 64K, most likely from linking multiple files with @command{ld -r}.
7676 Use this option to turn off this feature.
7678 @cindex PowerPC64 TOC sorting
7679 @kindex --no-toc-sort
7681 By default, @command{ld} sorts TOC sections so that those whose file
7682 happens to have a section called @code{.init} or @code{.fini} are
7683 placed first, followed by TOC sections referenced by code generated
7684 with PowerPC64 gcc's @code{-mcmodel=small}, and lastly TOC sections
7685 referenced only by code generated with PowerPC64 gcc's
7686 @code{-mcmodel=medium} or @code{-mcmodel=large} options. Doing this
7687 results in better TOC grouping for multi-TOC. Use this option to turn
7690 @cindex PowerPC64 PLT stub alignment
7692 @kindex --no-plt-align
7694 @itemx --no-plt-align
7695 Use these options to control whether individual PLT call stubs are
7696 aligned to a 32-byte boundary, or to the specified power of two
7697 boundary when using @code{--plt-align=}. A negative value may be
7698 specified to pad PLT call stubs so that they do not cross the
7699 specified power of two boundary (or the minimum number of boundaries
7700 if a PLT stub is so large that it must cross a boundary). By default
7701 PLT call stubs are aligned to 32-byte boundaries.
7703 @cindex PowerPC64 PLT call stub static chain
7704 @kindex --plt-static-chain
7705 @kindex --no-plt-static-chain
7706 @item --plt-static-chain
7707 @itemx --no-plt-static-chain
7708 Use these options to control whether PLT call stubs load the static
7709 chain pointer (r11). @code{ld} defaults to not loading the static
7710 chain since there is never any need to do so on a PLT call.
7712 @cindex PowerPC64 PLT call stub thread safety
7713 @kindex --plt-thread-safe
7714 @kindex --no-plt-thread-safe
7715 @item --plt-thread-safe
7716 @itemx --no-plt-thread-safe
7717 With power7's weakly ordered memory model, it is possible when using
7718 lazy binding for ld.so to update a plt entry in one thread and have
7719 another thread see the individual plt entry words update in the wrong
7720 order, despite ld.so carefully writing in the correct order and using
7721 memory write barriers. To avoid this we need some sort of read
7722 barrier in the call stub, or use LD_BIND_NOW=1. By default, @code{ld}
7723 looks for calls to commonly used functions that create threads, and if
7724 seen, adds the necessary barriers. Use these options to change the
7727 @cindex PowerPC64 ELFv2 PLT localentry optimization
7728 @kindex --plt-localentry
7729 @kindex --no-plt-localentry
7730 @item --plt-localentry
7731 @itemx --no-localentry
7732 ELFv2 functions with localentry:0 are those with a single entry point,
7733 ie. global entry == local entry, and that have no requirement on r2
7734 (the TOC/GOT pointer) or r12, and guarantee r2 is unchanged on return.
7735 Such an external function can be called via the PLT without saving r2
7736 or restoring it on return, avoiding a common load-hit-store for small
7737 functions. The optimization is attractive, with up to 40% reduction
7738 in execution time for a small function, but can result in symbol
7739 interposition failures. Also, minor changes in a shared library,
7740 including system libraries, can cause a function that was localentry:0
7741 to become localentry:8. This will result in a dynamic loader
7742 complaint and failure to run. The option is experimental, use with
7743 care. @option{--no-plt-localentry} is the default.
7757 @section @command{ld} and S/390 ELF Support
7759 @cindex S/390 ELF options
7763 @kindex --s390-pgste
7765 This option marks the result file with a @code{PT_S390_PGSTE}
7766 segment. The Linux kernel is supposed to allocate 4k page tables for
7767 binaries marked that way.
7781 @section @command{ld} and SPU ELF Support
7783 @cindex SPU ELF options
7789 This option marks an executable as a PIC plugin module.
7791 @cindex SPU overlays
7792 @kindex --no-overlays
7794 Normally, @command{ld} recognizes calls to functions within overlay
7795 regions, and redirects such calls to an overlay manager via a stub.
7796 @command{ld} also provides a built-in overlay manager. This option
7797 turns off all this special overlay handling.
7799 @cindex SPU overlay stub symbols
7800 @kindex --emit-stub-syms
7801 @item --emit-stub-syms
7802 This option causes @command{ld} to label overlay stubs with a local
7803 symbol that encodes the stub type and destination.
7805 @cindex SPU extra overlay stubs
7806 @kindex --extra-overlay-stubs
7807 @item --extra-overlay-stubs
7808 This option causes @command{ld} to add overlay call stubs on all
7809 function calls out of overlay regions. Normally stubs are not added
7810 on calls to non-overlay regions.
7812 @cindex SPU local store size
7813 @kindex --local-store=lo:hi
7814 @item --local-store=lo:hi
7815 @command{ld} usually checks that a final executable for SPU fits in
7816 the address range 0 to 256k. This option may be used to change the
7817 range. Disable the check entirely with @option{--local-store=0:0}.
7820 @kindex --stack-analysis
7821 @item --stack-analysis
7822 SPU local store space is limited. Over-allocation of stack space
7823 unnecessarily limits space available for code and data, while
7824 under-allocation results in runtime failures. If given this option,
7825 @command{ld} will provide an estimate of maximum stack usage.
7826 @command{ld} does this by examining symbols in code sections to
7827 determine the extents of functions, and looking at function prologues
7828 for stack adjusting instructions. A call-graph is created by looking
7829 for relocations on branch instructions. The graph is then searched
7830 for the maximum stack usage path. Note that this analysis does not
7831 find calls made via function pointers, and does not handle recursion
7832 and other cycles in the call graph. Stack usage may be
7833 under-estimated if your code makes such calls. Also, stack usage for
7834 dynamic allocation, e.g. alloca, will not be detected. If a link map
7835 is requested, detailed information about each function's stack usage
7836 and calls will be given.
7839 @kindex --emit-stack-syms
7840 @item --emit-stack-syms
7841 This option, if given along with @option{--stack-analysis} will result
7842 in @command{ld} emitting stack sizing symbols for each function.
7843 These take the form @code{__stack_<function_name>} for global
7844 functions, and @code{__stack_<number>_<function_name>} for static
7845 functions. @code{<number>} is the section id in hex. The value of
7846 such symbols is the stack requirement for the corresponding function.
7847 The symbol size will be zero, type @code{STT_NOTYPE}, binding
7848 @code{STB_LOCAL}, and section @code{SHN_ABS}.
7862 @section @command{ld}'s Support for Various TI COFF Versions
7863 @cindex TI COFF versions
7864 @kindex --format=@var{version}
7865 The @samp{--format} switch allows selection of one of the various
7866 TI COFF versions. The latest of this writing is 2; versions 0 and 1 are
7867 also supported. The TI COFF versions also vary in header byte-order
7868 format; @command{ld} will read any version or byte order, but the output
7869 header format depends on the default specified by the specific target.
7882 @section @command{ld} and WIN32 (cygwin/mingw)
7884 This section describes some of the win32 specific @command{ld} issues.
7885 See @ref{Options,,Command-line Options} for detailed description of the
7886 command-line options mentioned here.
7889 @cindex import libraries
7890 @item import libraries
7891 The standard Windows linker creates and uses so-called import
7892 libraries, which contains information for linking to dll's. They are
7893 regular static archives and are handled as any other static
7894 archive. The cygwin and mingw ports of @command{ld} have specific
7895 support for creating such libraries provided with the
7896 @samp{--out-implib} command-line option.
7898 @item exporting DLL symbols
7899 @cindex exporting DLL symbols
7900 The cygwin/mingw @command{ld} has several ways to export symbols for dll's.
7903 @item using auto-export functionality
7904 @cindex using auto-export functionality
7905 By default @command{ld} exports symbols with the auto-export functionality,
7906 which is controlled by the following command-line options:
7909 @item --export-all-symbols [This is the default]
7910 @item --exclude-symbols
7911 @item --exclude-libs
7912 @item --exclude-modules-for-implib
7913 @item --version-script
7916 When auto-export is in operation, @command{ld} will export all the non-local
7917 (global and common) symbols it finds in a DLL, with the exception of a few
7918 symbols known to belong to the system's runtime and libraries. As it will
7919 often not be desirable to export all of a DLL's symbols, which may include
7920 private functions that are not part of any public interface, the command-line
7921 options listed above may be used to filter symbols out from the list for
7922 exporting. The @samp{--output-def} option can be used in order to see the
7923 final list of exported symbols with all exclusions taken into effect.
7925 If @samp{--export-all-symbols} is not given explicitly on the
7926 command line, then the default auto-export behavior will be @emph{disabled}
7927 if either of the following are true:
7930 @item A DEF file is used.
7931 @item Any symbol in any object file was marked with the __declspec(dllexport) attribute.
7934 @item using a DEF file
7935 @cindex using a DEF file
7936 Another way of exporting symbols is using a DEF file. A DEF file is
7937 an ASCII file containing definitions of symbols which should be
7938 exported when a dll is created. Usually it is named @samp{<dll
7939 name>.def} and is added as any other object file to the linker's
7940 command line. The file's name must end in @samp{.def} or @samp{.DEF}.
7943 gcc -o <output> <objectfiles> <dll name>.def
7946 Using a DEF file turns off the normal auto-export behavior, unless the
7947 @samp{--export-all-symbols} option is also used.
7949 Here is an example of a DEF file for a shared library called @samp{xyz.dll}:
7952 LIBRARY "xyz.dll" BASE=0x20000000
7958 another_foo = abc.dll.afoo
7964 This example defines a DLL with a non-default base address and seven
7965 symbols in the export table. The third exported symbol @code{_bar} is an
7966 alias for the second. The fourth symbol, @code{another_foo} is resolved
7967 by "forwarding" to another module and treating it as an alias for
7968 @code{afoo} exported from the DLL @samp{abc.dll}. The final symbol
7969 @code{var1} is declared to be a data object. The @samp{doo} symbol in
7970 export library is an alias of @samp{foo}, which gets the string name
7971 in export table @samp{foo2}. The @samp{eoo} symbol is an data export
7972 symbol, which gets in export table the name @samp{var1}.
7974 The optional @code{LIBRARY <name>} command indicates the @emph{internal}
7975 name of the output DLL. If @samp{<name>} does not include a suffix,
7976 the default library suffix, @samp{.DLL} is appended.
7978 When the .DEF file is used to build an application, rather than a
7979 library, the @code{NAME <name>} command should be used instead of
7980 @code{LIBRARY}. If @samp{<name>} does not include a suffix, the default
7981 executable suffix, @samp{.EXE} is appended.
7983 With either @code{LIBRARY <name>} or @code{NAME <name>} the optional
7984 specification @code{BASE = <number>} may be used to specify a
7985 non-default base address for the image.
7987 If neither @code{LIBRARY <name>} nor @code{NAME <name>} is specified,
7988 or they specify an empty string, the internal name is the same as the
7989 filename specified on the command line.
7991 The complete specification of an export symbol is:
7995 ( ( ( <name1> [ = <name2> ] )
7996 | ( <name1> = <module-name> . <external-name>))
7997 [ @@ <integer> ] [NONAME] [DATA] [CONSTANT] [PRIVATE] [== <name3>] ) *
8000 Declares @samp{<name1>} as an exported symbol from the DLL, or declares
8001 @samp{<name1>} as an exported alias for @samp{<name2>}; or declares
8002 @samp{<name1>} as a "forward" alias for the symbol
8003 @samp{<external-name>} in the DLL @samp{<module-name>}.
8004 Optionally, the symbol may be exported by the specified ordinal
8005 @samp{<integer>} alias. The optional @samp{<name3>} is the to be used
8006 string in import/export table for the symbol.
8008 The optional keywords that follow the declaration indicate:
8010 @code{NONAME}: Do not put the symbol name in the DLL's export table. It
8011 will still be exported by its ordinal alias (either the value specified
8012 by the .def specification or, otherwise, the value assigned by the
8013 linker). The symbol name, however, does remain visible in the import
8014 library (if any), unless @code{PRIVATE} is also specified.
8016 @code{DATA}: The symbol is a variable or object, rather than a function.
8017 The import lib will export only an indirect reference to @code{foo} as
8018 the symbol @code{_imp__foo} (ie, @code{foo} must be resolved as
8021 @code{CONSTANT}: Like @code{DATA}, but put the undecorated @code{foo} as
8022 well as @code{_imp__foo} into the import library. Both refer to the
8023 read-only import address table's pointer to the variable, not to the
8024 variable itself. This can be dangerous. If the user code fails to add
8025 the @code{dllimport} attribute and also fails to explicitly add the
8026 extra indirection that the use of the attribute enforces, the
8027 application will behave unexpectedly.
8029 @code{PRIVATE}: Put the symbol in the DLL's export table, but do not put
8030 it into the static import library used to resolve imports at link time. The
8031 symbol can still be imported using the @code{LoadLibrary/GetProcAddress}
8032 API at runtime or by using the GNU ld extension of linking directly to
8033 the DLL without an import library.
8035 See ld/deffilep.y in the binutils sources for the full specification of
8036 other DEF file statements
8038 @cindex creating a DEF file
8039 While linking a shared dll, @command{ld} is able to create a DEF file
8040 with the @samp{--output-def <file>} command-line option.
8042 @item Using decorations
8043 @cindex Using decorations
8044 Another way of marking symbols for export is to modify the source code
8045 itself, so that when building the DLL each symbol to be exported is
8049 __declspec(dllexport) int a_variable
8050 __declspec(dllexport) void a_function(int with_args)
8053 All such symbols will be exported from the DLL. If, however,
8054 any of the object files in the DLL contain symbols decorated in
8055 this way, then the normal auto-export behavior is disabled, unless
8056 the @samp{--export-all-symbols} option is also used.
8058 Note that object files that wish to access these symbols must @emph{not}
8059 decorate them with dllexport. Instead, they should use dllimport,
8063 __declspec(dllimport) int a_variable
8064 __declspec(dllimport) void a_function(int with_args)
8067 This complicates the structure of library header files, because
8068 when included by the library itself the header must declare the
8069 variables and functions as dllexport, but when included by client
8070 code the header must declare them as dllimport. There are a number
8071 of idioms that are typically used to do this; often client code can
8072 omit the __declspec() declaration completely. See
8073 @samp{--enable-auto-import} and @samp{automatic data imports} for more
8077 @cindex automatic data imports
8078 @item automatic data imports
8079 The standard Windows dll format supports data imports from dlls only
8080 by adding special decorations (dllimport/dllexport), which let the
8081 compiler produce specific assembler instructions to deal with this
8082 issue. This increases the effort necessary to port existing Un*x
8083 code to these platforms, especially for large
8084 c++ libraries and applications. The auto-import feature, which was
8085 initially provided by Paul Sokolovsky, allows one to omit the
8086 decorations to achieve a behavior that conforms to that on POSIX/Un*x
8087 platforms. This feature is enabled with the @samp{--enable-auto-import}
8088 command-line option, although it is enabled by default on cygwin/mingw.
8089 The @samp{--enable-auto-import} option itself now serves mainly to
8090 suppress any warnings that are ordinarily emitted when linked objects
8091 trigger the feature's use.
8093 auto-import of variables does not always work flawlessly without
8094 additional assistance. Sometimes, you will see this message
8096 "variable '<var>' can't be auto-imported. Please read the
8097 documentation for ld's @code{--enable-auto-import} for details."
8099 The @samp{--enable-auto-import} documentation explains why this error
8100 occurs, and several methods that can be used to overcome this difficulty.
8101 One of these methods is the @emph{runtime pseudo-relocs} feature, described
8104 @cindex runtime pseudo-relocation
8105 For complex variables imported from DLLs (such as structs or classes),
8106 object files typically contain a base address for the variable and an
8107 offset (@emph{addend}) within the variable--to specify a particular
8108 field or public member, for instance. Unfortunately, the runtime loader used
8109 in win32 environments is incapable of fixing these references at runtime
8110 without the additional information supplied by dllimport/dllexport decorations.
8111 The standard auto-import feature described above is unable to resolve these
8114 The @samp{--enable-runtime-pseudo-relocs} switch allows these references to
8115 be resolved without error, while leaving the task of adjusting the references
8116 themselves (with their non-zero addends) to specialized code provided by the
8117 runtime environment. Recent versions of the cygwin and mingw environments and
8118 compilers provide this runtime support; older versions do not. However, the
8119 support is only necessary on the developer's platform; the compiled result will
8120 run without error on an older system.
8122 @samp{--enable-runtime-pseudo-relocs} is not the default; it must be explicitly
8125 @cindex direct linking to a dll
8126 @item direct linking to a dll
8127 The cygwin/mingw ports of @command{ld} support the direct linking,
8128 including data symbols, to a dll without the usage of any import
8129 libraries. This is much faster and uses much less memory than does the
8130 traditional import library method, especially when linking large
8131 libraries or applications. When @command{ld} creates an import lib, each
8132 function or variable exported from the dll is stored in its own bfd, even
8133 though a single bfd could contain many exports. The overhead involved in
8134 storing, loading, and processing so many bfd's is quite large, and explains the
8135 tremendous time, memory, and storage needed to link against particularly
8136 large or complex libraries when using import libs.
8138 Linking directly to a dll uses no extra command-line switches other than
8139 @samp{-L} and @samp{-l}, because @command{ld} already searches for a number
8140 of names to match each library. All that is needed from the developer's
8141 perspective is an understanding of this search, in order to force ld to
8142 select the dll instead of an import library.
8145 For instance, when ld is called with the argument @samp{-lxxx} it will attempt
8146 to find, in the first directory of its search path,
8159 before moving on to the next directory in the search path.
8161 (*) Actually, this is not @samp{cygxxx.dll} but in fact is @samp{<prefix>xxx.dll},
8162 where @samp{<prefix>} is set by the @command{ld} option
8163 @samp{--dll-search-prefix=<prefix>}. In the case of cygwin, the standard gcc spec
8164 file includes @samp{--dll-search-prefix=cyg}, so in effect we actually search for
8167 Other win32-based unix environments, such as mingw or pw32, may use other
8168 @samp{<prefix>}es, although at present only cygwin makes use of this feature. It
8169 was originally intended to help avoid name conflicts among dll's built for the
8170 various win32/un*x environments, so that (for example) two versions of a zlib dll
8171 could coexist on the same machine.
8173 The generic cygwin/mingw path layout uses a @samp{bin} directory for
8174 applications and dll's and a @samp{lib} directory for the import
8175 libraries (using cygwin nomenclature):
8181 libxxx.dll.a (in case of dll's)
8182 libxxx.a (in case of static archive)
8185 Linking directly to a dll without using the import library can be
8188 1. Use the dll directly by adding the @samp{bin} path to the link line
8190 gcc -Wl,-verbose -o a.exe -L../bin/ -lxxx
8193 However, as the dll's often have version numbers appended to their names
8194 (@samp{cygncurses-5.dll}) this will often fail, unless one specifies
8195 @samp{-L../bin -lncurses-5} to include the version. Import libs are generally
8196 not versioned, and do not have this difficulty.
8198 2. Create a symbolic link from the dll to a file in the @samp{lib}
8199 directory according to the above mentioned search pattern. This
8200 should be used to avoid unwanted changes in the tools needed for
8204 ln -s bin/cygxxx.dll lib/[cyg|lib|]xxx.dll[.a]
8207 Then you can link without any make environment changes.
8210 gcc -Wl,-verbose -o a.exe -L../lib/ -lxxx
8213 This technique also avoids the version number problems, because the following is
8220 libxxx.dll.a -> ../bin/cygxxx-5.dll
8223 Linking directly to a dll without using an import lib will work
8224 even when auto-import features are exercised, and even when
8225 @samp{--enable-runtime-pseudo-relocs} is used.
8227 Given the improvements in speed and memory usage, one might justifiably
8228 wonder why import libraries are used at all. There are three reasons:
8230 1. Until recently, the link-directly-to-dll functionality did @emph{not}
8231 work with auto-imported data.
8233 2. Sometimes it is necessary to include pure static objects within the
8234 import library (which otherwise contains only bfd's for indirection
8235 symbols that point to the exports of a dll). Again, the import lib
8236 for the cygwin kernel makes use of this ability, and it is not
8237 possible to do this without an import lib.
8239 3. Symbol aliases can only be resolved using an import lib. This is
8240 critical when linking against OS-supplied dll's (eg, the win32 API)
8241 in which symbols are usually exported as undecorated aliases of their
8242 stdcall-decorated assembly names.
8244 So, import libs are not going away. But the ability to replace
8245 true import libs with a simple symbolic link to (or a copy of)
8246 a dll, in many cases, is a useful addition to the suite of tools
8247 binutils makes available to the win32 developer. Given the
8248 massive improvements in memory requirements during linking, storage
8249 requirements, and linking speed, we expect that many developers
8250 will soon begin to use this feature whenever possible.
8252 @item symbol aliasing
8254 @item adding additional names
8255 Sometimes, it is useful to export symbols with additional names.
8256 A symbol @samp{foo} will be exported as @samp{foo}, but it can also be
8257 exported as @samp{_foo} by using special directives in the DEF file
8258 when creating the dll. This will affect also the optional created
8259 import library. Consider the following DEF file:
8262 LIBRARY "xyz.dll" BASE=0x61000000
8269 The line @samp{_foo = foo} maps the symbol @samp{foo} to @samp{_foo}.
8271 Another method for creating a symbol alias is to create it in the
8272 source code using the "weak" attribute:
8275 void foo () @{ /* Do something. */; @}
8276 void _foo () __attribute__ ((weak, alias ("foo")));
8279 See the gcc manual for more information about attributes and weak
8282 @item renaming symbols
8283 Sometimes it is useful to rename exports. For instance, the cygwin
8284 kernel does this regularly. A symbol @samp{_foo} can be exported as
8285 @samp{foo} but not as @samp{_foo} by using special directives in the
8286 DEF file. (This will also affect the import library, if it is
8287 created). In the following example:
8290 LIBRARY "xyz.dll" BASE=0x61000000
8296 The line @samp{_foo = foo} maps the exported symbol @samp{foo} to
8300 Note: using a DEF file disables the default auto-export behavior,
8301 unless the @samp{--export-all-symbols} command-line option is used.
8302 If, however, you are trying to rename symbols, then you should list
8303 @emph{all} desired exports in the DEF file, including the symbols
8304 that are not being renamed, and do @emph{not} use the
8305 @samp{--export-all-symbols} option. If you list only the
8306 renamed symbols in the DEF file, and use @samp{--export-all-symbols}
8307 to handle the other symbols, then the both the new names @emph{and}
8308 the original names for the renamed symbols will be exported.
8309 In effect, you'd be aliasing those symbols, not renaming them,
8310 which is probably not what you wanted.
8312 @cindex weak externals
8313 @item weak externals
8314 The Windows object format, PE, specifies a form of weak symbols called
8315 weak externals. When a weak symbol is linked and the symbol is not
8316 defined, the weak symbol becomes an alias for some other symbol. There
8317 are three variants of weak externals:
8319 @item Definition is searched for in objects and libraries, historically
8320 called lazy externals.
8321 @item Definition is searched for only in other objects, not in libraries.
8322 This form is not presently implemented.
8323 @item No search; the symbol is an alias. This form is not presently
8326 As a GNU extension, weak symbols that do not specify an alternate symbol
8327 are supported. If the symbol is undefined when linking, the symbol
8328 uses a default value.
8330 @cindex aligned common symbols
8331 @item aligned common symbols
8332 As a GNU extension to the PE file format, it is possible to specify the
8333 desired alignment for a common symbol. This information is conveyed from
8334 the assembler or compiler to the linker by means of GNU-specific commands
8335 carried in the object file's @samp{.drectve} section, which are recognized
8336 by @command{ld} and respected when laying out the common symbols. Native
8337 tools will be able to process object files employing this GNU extension,
8338 but will fail to respect the alignment instructions, and may issue noisy
8339 warnings about unknown linker directives.
8354 @section @code{ld} and Xtensa Processors
8356 @cindex Xtensa processors
8357 The default @command{ld} behavior for Xtensa processors is to interpret
8358 @code{SECTIONS} commands so that lists of explicitly named sections in a
8359 specification with a wildcard file will be interleaved when necessary to
8360 keep literal pools within the range of PC-relative load offsets. For
8361 example, with the command:
8373 @command{ld} may interleave some of the @code{.literal}
8374 and @code{.text} sections from different object files to ensure that the
8375 literal pools are within the range of PC-relative load offsets. A valid
8376 interleaving might place the @code{.literal} sections from an initial
8377 group of files followed by the @code{.text} sections of that group of
8378 files. Then, the @code{.literal} sections from the rest of the files
8379 and the @code{.text} sections from the rest of the files would follow.
8381 @cindex @option{--relax} on Xtensa
8382 @cindex relaxing on Xtensa
8383 Relaxation is enabled by default for the Xtensa version of @command{ld} and
8384 provides two important link-time optimizations. The first optimization
8385 is to combine identical literal values to reduce code size. A redundant
8386 literal will be removed and all the @code{L32R} instructions that use it
8387 will be changed to reference an identical literal, as long as the
8388 location of the replacement literal is within the offset range of all
8389 the @code{L32R} instructions. The second optimization is to remove
8390 unnecessary overhead from assembler-generated ``longcall'' sequences of
8391 @code{L32R}/@code{CALLX@var{n}} when the target functions are within
8392 range of direct @code{CALL@var{n}} instructions.
8394 For each of these cases where an indirect call sequence can be optimized
8395 to a direct call, the linker will change the @code{CALLX@var{n}}
8396 instruction to a @code{CALL@var{n}} instruction, remove the @code{L32R}
8397 instruction, and remove the literal referenced by the @code{L32R}
8398 instruction if it is not used for anything else. Removing the
8399 @code{L32R} instruction always reduces code size but can potentially
8400 hurt performance by changing the alignment of subsequent branch targets.
8401 By default, the linker will always preserve alignments, either by
8402 switching some instructions between 24-bit encodings and the equivalent
8403 density instructions or by inserting a no-op in place of the @code{L32R}
8404 instruction that was removed. If code size is more important than
8405 performance, the @option{--size-opt} option can be used to prevent the
8406 linker from widening density instructions or inserting no-ops, except in
8407 a few cases where no-ops are required for correctness.
8409 The following Xtensa-specific command-line options can be used to
8412 @cindex Xtensa options
8415 When optimizing indirect calls to direct calls, optimize for code size
8416 more than performance. With this option, the linker will not insert
8417 no-ops or widen density instructions to preserve branch target
8418 alignment. There may still be some cases where no-ops are required to
8419 preserve the correctness of the code.
8427 @ifclear SingleFormat
8432 @cindex object file management
8433 @cindex object formats available
8435 The linker accesses object and archive files using the BFD libraries.
8436 These libraries allow the linker to use the same routines to operate on
8437 object files whatever the object file format. A different object file
8438 format can be supported simply by creating a new BFD back end and adding
8439 it to the library. To conserve runtime memory, however, the linker and
8440 associated tools are usually configured to support only a subset of the
8441 object file formats available. You can use @code{objdump -i}
8442 (@pxref{objdump,,objdump,binutils.info,The GNU Binary Utilities}) to
8443 list all the formats available for your configuration.
8445 @cindex BFD requirements
8446 @cindex requirements for BFD
8447 As with most implementations, BFD is a compromise between
8448 several conflicting requirements. The major factor influencing
8449 BFD design was efficiency: any time used converting between
8450 formats is time which would not have been spent had BFD not
8451 been involved. This is partly offset by abstraction payback; since
8452 BFD simplifies applications and back ends, more time and care
8453 may be spent optimizing algorithms for a greater speed.
8455 One minor artifact of the BFD solution which you should bear in
8456 mind is the potential for information loss. There are two places where
8457 useful information can be lost using the BFD mechanism: during
8458 conversion and during output. @xref{BFD information loss}.
8461 * BFD outline:: How it works: an outline of BFD
8465 @section How It Works: An Outline of BFD
8466 @cindex opening object files
8467 @include bfdsumm.texi
8470 @node Reporting Bugs
8471 @chapter Reporting Bugs
8472 @cindex bugs in @command{ld}
8473 @cindex reporting bugs in @command{ld}
8475 Your bug reports play an essential role in making @command{ld} reliable.
8477 Reporting a bug may help you by bringing a solution to your problem, or
8478 it may not. But in any case the principal function of a bug report is
8479 to help the entire community by making the next version of @command{ld}
8480 work better. Bug reports are your contribution to the maintenance of
8483 In order for a bug report to serve its purpose, you must include the
8484 information that enables us to fix the bug.
8487 * Bug Criteria:: Have you found a bug?
8488 * Bug Reporting:: How to report bugs
8492 @section Have You Found a Bug?
8493 @cindex bug criteria
8495 If you are not sure whether you have found a bug, here are some guidelines:
8498 @cindex fatal signal
8499 @cindex linker crash
8500 @cindex crash of linker
8502 If the linker gets a fatal signal, for any input whatever, that is a
8503 @command{ld} bug. Reliable linkers never crash.
8505 @cindex error on valid input
8507 If @command{ld} produces an error message for valid input, that is a bug.
8509 @cindex invalid input
8511 If @command{ld} does not produce an error message for invalid input, that
8512 may be a bug. In the general case, the linker can not verify that
8513 object files are correct.
8516 If you are an experienced user of linkers, your suggestions for
8517 improvement of @command{ld} are welcome in any case.
8521 @section How to Report Bugs
8523 @cindex @command{ld} bugs, reporting
8525 A number of companies and individuals offer support for @sc{gnu}
8526 products. If you obtained @command{ld} from a support organization, we
8527 recommend you contact that organization first.
8529 You can find contact information for many support companies and
8530 individuals in the file @file{etc/SERVICE} in the @sc{gnu} Emacs
8534 Otherwise, send bug reports for @command{ld} to
8538 The fundamental principle of reporting bugs usefully is this:
8539 @strong{report all the facts}. If you are not sure whether to state a
8540 fact or leave it out, state it!
8542 Often people omit facts because they think they know what causes the
8543 problem and assume that some details do not matter. Thus, you might
8544 assume that the name of a symbol you use in an example does not
8545 matter. Well, probably it does not, but one cannot be sure. Perhaps
8546 the bug is a stray memory reference which happens to fetch from the
8547 location where that name is stored in memory; perhaps, if the name
8548 were different, the contents of that location would fool the linker
8549 into doing the right thing despite the bug. Play it safe and give a
8550 specific, complete example. That is the easiest thing for you to do,
8551 and the most helpful.
8553 Keep in mind that the purpose of a bug report is to enable us to fix
8554 the bug if it is new to us. Therefore, always write your bug reports
8555 on the assumption that the bug has not been reported previously.
8557 Sometimes people give a few sketchy facts and ask, ``Does this ring a
8558 bell?'' This cannot help us fix a bug, so it is basically useless. We
8559 respond by asking for enough details to enable us to investigate.
8560 You might as well expedite matters by sending them to begin with.
8562 To enable us to fix the bug, you should include all these things:
8566 The version of @command{ld}. @command{ld} announces it if you start it with
8567 the @samp{--version} argument.
8569 Without this, we will not know whether there is any point in looking for
8570 the bug in the current version of @command{ld}.
8573 Any patches you may have applied to the @command{ld} source, including any
8574 patches made to the @code{BFD} library.
8577 The type of machine you are using, and the operating system name and
8581 What compiler (and its version) was used to compile @command{ld}---e.g.
8585 The command arguments you gave the linker to link your example and
8586 observe the bug. To guarantee you will not omit something important,
8587 list them all. A copy of the Makefile (or the output from make) is
8590 If we were to try to guess the arguments, we would probably guess wrong
8591 and then we might not encounter the bug.
8594 A complete input file, or set of input files, that will reproduce the
8595 bug. It is generally most helpful to send the actual object files
8596 provided that they are reasonably small. Say no more than 10K. For
8597 bigger files you can either make them available by FTP or HTTP or else
8598 state that you are willing to send the object file(s) to whomever
8599 requests them. (Note - your email will be going to a mailing list, so
8600 we do not want to clog it up with large attachments). But small
8601 attachments are best.
8603 If the source files were assembled using @code{gas} or compiled using
8604 @code{gcc}, then it may be OK to send the source files rather than the
8605 object files. In this case, be sure to say exactly what version of
8606 @code{gas} or @code{gcc} was used to produce the object files. Also say
8607 how @code{gas} or @code{gcc} were configured.
8610 A description of what behavior you observe that you believe is
8611 incorrect. For example, ``It gets a fatal signal.''
8613 Of course, if the bug is that @command{ld} gets a fatal signal, then we
8614 will certainly notice it. But if the bug is incorrect output, we might
8615 not notice unless it is glaringly wrong. You might as well not give us
8616 a chance to make a mistake.
8618 Even if the problem you experience is a fatal signal, you should still
8619 say so explicitly. Suppose something strange is going on, such as, your
8620 copy of @command{ld} is out of sync, or you have encountered a bug in the
8621 C library on your system. (This has happened!) Your copy might crash
8622 and ours would not. If you told us to expect a crash, then when ours
8623 fails to crash, we would know that the bug was not happening for us. If
8624 you had not told us to expect a crash, then we would not be able to draw
8625 any conclusion from our observations.
8628 If you wish to suggest changes to the @command{ld} source, send us context
8629 diffs, as generated by @code{diff} with the @samp{-u}, @samp{-c}, or
8630 @samp{-p} option. Always send diffs from the old file to the new file.
8631 If you even discuss something in the @command{ld} source, refer to it by
8632 context, not by line number.
8634 The line numbers in our development sources will not match those in your
8635 sources. Your line numbers would convey no useful information to us.
8638 Here are some things that are not necessary:
8642 A description of the envelope of the bug.
8644 Often people who encounter a bug spend a lot of time investigating
8645 which changes to the input file will make the bug go away and which
8646 changes will not affect it.
8648 This is often time consuming and not very useful, because the way we
8649 will find the bug is by running a single example under the debugger
8650 with breakpoints, not by pure deduction from a series of examples.
8651 We recommend that you save your time for something else.
8653 Of course, if you can find a simpler example to report @emph{instead}
8654 of the original one, that is a convenience for us. Errors in the
8655 output will be easier to spot, running under the debugger will take
8656 less time, and so on.
8658 However, simplification is not vital; if you do not want to do this,
8659 report the bug anyway and send us the entire test case you used.
8662 A patch for the bug.
8664 A patch for the bug does help us if it is a good one. But do not omit
8665 the necessary information, such as the test case, on the assumption that
8666 a patch is all we need. We might see problems with your patch and decide
8667 to fix the problem another way, or we might not understand it at all.
8669 Sometimes with a program as complicated as @command{ld} it is very hard to
8670 construct an example that will make the program follow a certain path
8671 through the code. If you do not send us the example, we will not be
8672 able to construct one, so we will not be able to verify that the bug is
8675 And if we cannot understand what bug you are trying to fix, or why your
8676 patch should be an improvement, we will not install it. A test case will
8677 help us to understand.
8680 A guess about what the bug is or what it depends on.
8682 Such guesses are usually wrong. Even we cannot guess right about such
8683 things without first using the debugger to find the facts.
8687 @appendix MRI Compatible Script Files
8688 @cindex MRI compatibility
8689 To aid users making the transition to @sc{gnu} @command{ld} from the MRI
8690 linker, @command{ld} can use MRI compatible linker scripts as an
8691 alternative to the more general-purpose linker scripting language
8692 described in @ref{Scripts}. MRI compatible linker scripts have a much
8693 simpler command set than the scripting language otherwise used with
8694 @command{ld}. @sc{gnu} @command{ld} supports the most commonly used MRI
8695 linker commands; these commands are described here.
8697 In general, MRI scripts aren't of much use with the @code{a.out} object
8698 file format, since it only has three sections and MRI scripts lack some
8699 features to make use of them.
8701 You can specify a file containing an MRI-compatible script using the
8702 @samp{-c} command-line option.
8704 Each command in an MRI-compatible script occupies its own line; each
8705 command line starts with the keyword that identifies the command (though
8706 blank lines are also allowed for punctuation). If a line of an
8707 MRI-compatible script begins with an unrecognized keyword, @command{ld}
8708 issues a warning message, but continues processing the script.
8710 Lines beginning with @samp{*} are comments.
8712 You can write these commands using all upper-case letters, or all
8713 lower case; for example, @samp{chip} is the same as @samp{CHIP}.
8714 The following list shows only the upper-case form of each command.
8717 @cindex @code{ABSOLUTE} (MRI)
8718 @item ABSOLUTE @var{secname}
8719 @itemx ABSOLUTE @var{secname}, @var{secname}, @dots{} @var{secname}
8720 Normally, @command{ld} includes in the output file all sections from all
8721 the input files. However, in an MRI-compatible script, you can use the
8722 @code{ABSOLUTE} command to restrict the sections that will be present in
8723 your output program. If the @code{ABSOLUTE} command is used at all in a
8724 script, then only the sections named explicitly in @code{ABSOLUTE}
8725 commands will appear in the linker output. You can still use other
8726 input sections (whatever you select on the command line, or using
8727 @code{LOAD}) to resolve addresses in the output file.
8729 @cindex @code{ALIAS} (MRI)
8730 @item ALIAS @var{out-secname}, @var{in-secname}
8731 Use this command to place the data from input section @var{in-secname}
8732 in a section called @var{out-secname} in the linker output file.
8734 @var{in-secname} may be an integer.
8736 @cindex @code{ALIGN} (MRI)
8737 @item ALIGN @var{secname} = @var{expression}
8738 Align the section called @var{secname} to @var{expression}. The
8739 @var{expression} should be a power of two.
8741 @cindex @code{BASE} (MRI)
8742 @item BASE @var{expression}
8743 Use the value of @var{expression} as the lowest address (other than
8744 absolute addresses) in the output file.
8746 @cindex @code{CHIP} (MRI)
8747 @item CHIP @var{expression}
8748 @itemx CHIP @var{expression}, @var{expression}
8749 This command does nothing; it is accepted only for compatibility.
8751 @cindex @code{END} (MRI)
8753 This command does nothing whatever; it's only accepted for compatibility.
8755 @cindex @code{FORMAT} (MRI)
8756 @item FORMAT @var{output-format}
8757 Similar to the @code{OUTPUT_FORMAT} command in the more general linker
8758 language, but restricted to S-records, if @var{output-format} is @samp{S}
8760 @cindex @code{LIST} (MRI)
8761 @item LIST @var{anything}@dots{}
8762 Print (to the standard output file) a link map, as produced by the
8763 @command{ld} command-line option @samp{-M}.
8765 The keyword @code{LIST} may be followed by anything on the
8766 same line, with no change in its effect.
8768 @cindex @code{LOAD} (MRI)
8769 @item LOAD @var{filename}
8770 @itemx LOAD @var{filename}, @var{filename}, @dots{} @var{filename}
8771 Include one or more object file @var{filename} in the link; this has the
8772 same effect as specifying @var{filename} directly on the @command{ld}
8775 @cindex @code{NAME} (MRI)
8776 @item NAME @var{output-name}
8777 @var{output-name} is the name for the program produced by @command{ld}; the
8778 MRI-compatible command @code{NAME} is equivalent to the command-line
8779 option @samp{-o} or the general script language command @code{OUTPUT}.
8781 @cindex @code{ORDER} (MRI)
8782 @item ORDER @var{secname}, @var{secname}, @dots{} @var{secname}
8783 @itemx ORDER @var{secname} @var{secname} @var{secname}
8784 Normally, @command{ld} orders the sections in its output file in the
8785 order in which they first appear in the input files. In an MRI-compatible
8786 script, you can override this ordering with the @code{ORDER} command. The
8787 sections you list with @code{ORDER} will appear first in your output
8788 file, in the order specified.
8790 @cindex @code{PUBLIC} (MRI)
8791 @item PUBLIC @var{name}=@var{expression}
8792 @itemx PUBLIC @var{name},@var{expression}
8793 @itemx PUBLIC @var{name} @var{expression}
8794 Supply a value (@var{expression}) for external symbol
8795 @var{name} used in the linker input files.
8797 @cindex @code{SECT} (MRI)
8798 @item SECT @var{secname}, @var{expression}
8799 @itemx SECT @var{secname}=@var{expression}
8800 @itemx SECT @var{secname} @var{expression}
8801 You can use any of these three forms of the @code{SECT} command to
8802 specify the start address (@var{expression}) for section @var{secname}.
8803 If you have more than one @code{SECT} statement for the same
8804 @var{secname}, only the @emph{first} sets the start address.
8807 @node GNU Free Documentation License
8808 @appendix GNU Free Documentation License
8812 @unnumbered LD Index
8817 % I think something like @@colophon should be in texinfo. In the
8819 \long\def\colophon{\hbox to0pt{}\vfill
8820 \centerline{The body of this manual is set in}
8821 \centerline{\fontname\tenrm,}
8822 \centerline{with headings in {\bf\fontname\tenbf}}
8823 \centerline{and examples in {\tt\fontname\tentt}.}
8824 \centerline{{\it\fontname\tenit\/} and}
8825 \centerline{{\sl\fontname\tensl\/}}
8826 \centerline{are used for emphasis.}\vfill}
8828 % Blame: doc@@cygnus.com, 28mar91.