* configure.ac (FLAGS_FOR_TARGET,target=cygwin): Fix for building
[deliverable/binutils-gdb.git] / ld / ld.texinfo
1 \input texinfo
2 @setfilename ld.info
3 @c Copyright 1991, 1992, 1993, 1994, 1995, 1996, 1997, 1998, 1999, 2000,
4 @c 2001, 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009, 2010, 2011, 2012
5 @c Free Software Foundation, Inc.
6 @syncodeindex ky cp
7 @c man begin INCLUDE
8 @include configdoc.texi
9 @c (configdoc.texi is generated by the Makefile)
10 @include bfdver.texi
11 @c man end
12
13 @c @smallbook
14
15 @macro gcctabopt{body}
16 @code{\body\}
17 @end macro
18
19 @c man begin NAME
20 @ifset man
21 @c Configure for the generation of man pages
22 @set UsesEnvVars
23 @set GENERIC
24 @set ARM
25 @set C6X
26 @set H8300
27 @set HPPA
28 @set I960
29 @set M68HC11
30 @set M68K
31 @set MMIX
32 @set MSP430
33 @set POWERPC
34 @set POWERPC64
35 @set Renesas
36 @set SPU
37 @set TICOFF
38 @set WIN32
39 @set XTENSA
40 @end ifset
41 @c man end
42
43 @ifnottex
44 @dircategory Software development
45 @direntry
46 * Ld: (ld). The GNU linker.
47 @end direntry
48 @end ifnottex
49
50 @copying
51 This file documents the @sc{gnu} linker LD
52 @ifset VERSION_PACKAGE
53 @value{VERSION_PACKAGE}
54 @end ifset
55 version @value{VERSION}.
56
57 Copyright @copyright{} 1991, 1992, 1993, 1994, 1995, 1996, 1997, 1998, 1999, 2000,
58 2001, 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009 Free Software Foundation, Inc.
59
60 Permission is granted to copy, distribute and/or modify this document
61 under the terms of the GNU Free Documentation License, Version 1.3
62 or any later version published by the Free Software Foundation;
63 with no Invariant Sections, with no Front-Cover Texts, and with no
64 Back-Cover Texts. A copy of the license is included in the
65 section entitled ``GNU Free Documentation License''.
66 @end copying
67 @iftex
68 @finalout
69 @setchapternewpage odd
70 @settitle The GNU linker
71 @titlepage
72 @title The GNU linker
73 @sp 1
74 @subtitle @code{ld}
75 @ifset VERSION_PACKAGE
76 @subtitle @value{VERSION_PACKAGE}
77 @end ifset
78 @subtitle Version @value{VERSION}
79 @author Steve Chamberlain
80 @author Ian Lance Taylor
81 @page
82
83 @tex
84 {\parskip=0pt
85 \hfill Red Hat Inc\par
86 \hfill nickc\@credhat.com, doc\@redhat.com\par
87 \hfill {\it The GNU linker}\par
88 \hfill Edited by Jeffrey Osier (jeffrey\@cygnus.com)\par
89 }
90 \global\parindent=0pt % Steve likes it this way.
91 @end tex
92
93 @vskip 0pt plus 1filll
94 @c man begin COPYRIGHT
95 Copyright @copyright{} 1991, 1992, 1993, 1994, 1995, 1996, 1997, 1998,
96 1999, 2000, 2001, 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009 Free
97 Software Foundation, Inc.
98
99 Permission is granted to copy, distribute and/or modify this document
100 under the terms of the GNU Free Documentation License, Version 1.3
101 or any later version published by the Free Software Foundation;
102 with no Invariant Sections, with no Front-Cover Texts, and with no
103 Back-Cover Texts. A copy of the license is included in the
104 section entitled ``GNU Free Documentation License''.
105 @c man end
106
107 @end titlepage
108 @end iftex
109 @contents
110 @c FIXME: Talk about importance of *order* of args, cmds to linker!
111
112 @ifnottex
113 @node Top
114 @top LD
115 This file documents the @sc{gnu} linker ld
116 @ifset VERSION_PACKAGE
117 @value{VERSION_PACKAGE}
118 @end ifset
119 version @value{VERSION}.
120
121 This document is distributed under the terms of the GNU Free
122 Documentation License version 1.3. A copy of the license is included
123 in the section entitled ``GNU Free Documentation License''.
124
125 @menu
126 * Overview:: Overview
127 * Invocation:: Invocation
128 * Scripts:: Linker Scripts
129 @ifset GENERIC
130 * Machine Dependent:: Machine Dependent Features
131 @end ifset
132 @ifclear GENERIC
133 @ifset H8300
134 * H8/300:: ld and the H8/300
135 @end ifset
136 @ifset Renesas
137 * Renesas:: ld and other Renesas micros
138 @end ifset
139 @ifset I960
140 * i960:: ld and the Intel 960 family
141 @end ifset
142 @ifset ARM
143 * ARM:: ld and the ARM family
144 @end ifset
145 @ifset HPPA
146 * HPPA ELF32:: ld and HPPA 32-bit ELF
147 @end ifset
148 @ifset M68HC11
149 * M68HC11/68HC12:: ld and the Motorola 68HC11 and 68HC12 families
150 @end ifset
151 @ifset M68K
152 * M68K:: ld and Motorola 68K family
153 @end ifset
154 @ifset POWERPC
155 * PowerPC ELF32:: ld and PowerPC 32-bit ELF Support
156 @end ifset
157 @ifset POWERPC64
158 * PowerPC64 ELF64:: ld and PowerPC64 64-bit ELF Support
159 @end ifset
160 @ifset SPU
161 * SPU ELF:: ld and SPU ELF Support
162 @end ifset
163 @ifset TICOFF
164 * TI COFF:: ld and the TI COFF
165 @end ifset
166 @ifset WIN32
167 * Win32:: ld and WIN32 (cygwin/mingw)
168 @end ifset
169 @ifset XTENSA
170 * Xtensa:: ld and Xtensa Processors
171 @end ifset
172 @end ifclear
173 @ifclear SingleFormat
174 * BFD:: BFD
175 @end ifclear
176 @c Following blank line required for remaining bug in makeinfo conds/menus
177
178 * Reporting Bugs:: Reporting Bugs
179 * MRI:: MRI Compatible Script Files
180 * GNU Free Documentation License:: GNU Free Documentation License
181 * LD Index:: LD Index
182 @end menu
183 @end ifnottex
184
185 @node Overview
186 @chapter Overview
187
188 @cindex @sc{gnu} linker
189 @cindex what is this?
190
191 @ifset man
192 @c man begin SYNOPSIS
193 ld [@b{options}] @var{objfile} @dots{}
194 @c man end
195
196 @c man begin SEEALSO
197 ar(1), nm(1), objcopy(1), objdump(1), readelf(1) and
198 the Info entries for @file{binutils} and
199 @file{ld}.
200 @c man end
201 @end ifset
202
203 @c man begin DESCRIPTION
204
205 @command{ld} combines a number of object and archive files, relocates
206 their data and ties up symbol references. Usually the last step in
207 compiling a program is to run @command{ld}.
208
209 @command{ld} accepts Linker Command Language files written in
210 a superset of AT&T's Link Editor Command Language syntax,
211 to provide explicit and total control over the linking process.
212
213 @ifset man
214 @c For the man only
215 This man page does not describe the command language; see the
216 @command{ld} entry in @code{info} for full details on the command
217 language and on other aspects of the GNU linker.
218 @end ifset
219
220 @ifclear SingleFormat
221 This version of @command{ld} uses the general purpose BFD libraries
222 to operate on object files. This allows @command{ld} to read, combine, and
223 write object files in many different formats---for example, COFF or
224 @code{a.out}. Different formats may be linked together to produce any
225 available kind of object file. @xref{BFD}, for more information.
226 @end ifclear
227
228 Aside from its flexibility, the @sc{gnu} linker is more helpful than other
229 linkers in providing diagnostic information. Many linkers abandon
230 execution immediately upon encountering an error; whenever possible,
231 @command{ld} continues executing, allowing you to identify other errors
232 (or, in some cases, to get an output file in spite of the error).
233
234 @c man end
235
236 @node Invocation
237 @chapter Invocation
238
239 @c man begin DESCRIPTION
240
241 The @sc{gnu} linker @command{ld} is meant to cover a broad range of situations,
242 and to be as compatible as possible with other linkers. As a result,
243 you have many choices to control its behavior.
244
245 @c man end
246
247 @ifset UsesEnvVars
248 @menu
249 * Options:: Command Line Options
250 * Environment:: Environment Variables
251 @end menu
252
253 @node Options
254 @section Command Line Options
255 @end ifset
256
257 @cindex command line
258 @cindex options
259
260 @c man begin OPTIONS
261
262 The linker supports a plethora of command-line options, but in actual
263 practice few of them are used in any particular context.
264 @cindex standard Unix system
265 For instance, a frequent use of @command{ld} is to link standard Unix
266 object files on a standard, supported Unix system. On such a system, to
267 link a file @code{hello.o}:
268
269 @smallexample
270 ld -o @var{output} /lib/crt0.o hello.o -lc
271 @end smallexample
272
273 This tells @command{ld} to produce a file called @var{output} as the
274 result of linking the file @code{/lib/crt0.o} with @code{hello.o} and
275 the library @code{libc.a}, which will come from the standard search
276 directories. (See the discussion of the @samp{-l} option below.)
277
278 Some of the command-line options to @command{ld} may be specified at any
279 point in the command line. However, options which refer to files, such
280 as @samp{-l} or @samp{-T}, cause the file to be read at the point at
281 which the option appears in the command line, relative to the object
282 files and other file options. Repeating non-file options with a
283 different argument will either have no further effect, or override prior
284 occurrences (those further to the left on the command line) of that
285 option. Options which may be meaningfully specified more than once are
286 noted in the descriptions below.
287
288 @cindex object files
289 Non-option arguments are object files or archives which are to be linked
290 together. They may follow, precede, or be mixed in with command-line
291 options, except that an object file argument may not be placed between
292 an option and its argument.
293
294 Usually the linker is invoked with at least one object file, but you can
295 specify other forms of binary input files using @samp{-l}, @samp{-R},
296 and the script command language. If @emph{no} binary input files at all
297 are specified, the linker does not produce any output, and issues the
298 message @samp{No input files}.
299
300 If the linker cannot recognize the format of an object file, it will
301 assume that it is a linker script. A script specified in this way
302 augments the main linker script used for the link (either the default
303 linker script or the one specified by using @samp{-T}). This feature
304 permits the linker to link against a file which appears to be an object
305 or an archive, but actually merely defines some symbol values, or uses
306 @code{INPUT} or @code{GROUP} to load other objects. Specifying a
307 script in this way merely augments the main linker script, with the
308 extra commands placed after the main script; use the @samp{-T} option
309 to replace the default linker script entirely, but note the effect of
310 the @code{INSERT} command. @xref{Scripts}.
311
312 For options whose names are a single letter,
313 option arguments must either follow the option letter without intervening
314 whitespace, or be given as separate arguments immediately following the
315 option that requires them.
316
317 For options whose names are multiple letters, either one dash or two can
318 precede the option name; for example, @samp{-trace-symbol} and
319 @samp{--trace-symbol} are equivalent. Note---there is one exception to
320 this rule. Multiple letter options that start with a lower case 'o' can
321 only be preceded by two dashes. This is to reduce confusion with the
322 @samp{-o} option. So for example @samp{-omagic} sets the output file
323 name to @samp{magic} whereas @samp{--omagic} sets the NMAGIC flag on the
324 output.
325
326 Arguments to multiple-letter options must either be separated from the
327 option name by an equals sign, or be given as separate arguments
328 immediately following the option that requires them. For example,
329 @samp{--trace-symbol foo} and @samp{--trace-symbol=foo} are equivalent.
330 Unique abbreviations of the names of multiple-letter options are
331 accepted.
332
333 Note---if the linker is being invoked indirectly, via a compiler driver
334 (e.g. @samp{gcc}) then all the linker command line options should be
335 prefixed by @samp{-Wl,} (or whatever is appropriate for the particular
336 compiler driver) like this:
337
338 @smallexample
339 gcc -Wl,--start-group foo.o bar.o -Wl,--end-group
340 @end smallexample
341
342 This is important, because otherwise the compiler driver program may
343 silently drop the linker options, resulting in a bad link. Confusion
344 may also arise when passing options that require values through a
345 driver, as the use of a space between option and argument acts as
346 a separator, and causes the driver to pass only the option to the linker
347 and the argument to the compiler. In this case, it is simplest to use
348 the joined forms of both single- and multiple-letter options, such as:
349
350 @smallexample
351 gcc foo.o bar.o -Wl,-eENTRY -Wl,-Map=a.map
352 @end smallexample
353
354 Here is a table of the generic command line switches accepted by the GNU
355 linker:
356
357 @table @gcctabopt
358 @include at-file.texi
359
360 @kindex -a @var{keyword}
361 @item -a @var{keyword}
362 This option is supported for HP/UX compatibility. The @var{keyword}
363 argument must be one of the strings @samp{archive}, @samp{shared}, or
364 @samp{default}. @samp{-aarchive} is functionally equivalent to
365 @samp{-Bstatic}, and the other two keywords are functionally equivalent
366 to @samp{-Bdynamic}. This option may be used any number of times.
367
368 @kindex --audit @var{AUDITLIB}
369 @item --audit @var{AUDITLIB}
370 Adds @var{AUDITLIB} to the @code{DT_AUDIT} entry of the dynamic section.
371 @var{AUDITLIB} is not checked for existence, nor will it use the DT_SONAME
372 specified in the library. If specified multiple times @code{DT_AUDIT}
373 will contain a colon separated list of audit interfaces to use. If the linker
374 finds an object with an audit entry while searching for shared libraries,
375 it will add a corresponding @code{DT_DEPAUDIT} entry in the output file.
376 This option is only meaningful on ELF platforms supporting the rtld-audit
377 interface.
378
379 @ifset I960
380 @cindex architectures
381 @kindex -A @var{arch}
382 @item -A @var{architecture}
383 @kindex --architecture=@var{arch}
384 @itemx --architecture=@var{architecture}
385 In the current release of @command{ld}, this option is useful only for the
386 Intel 960 family of architectures. In that @command{ld} configuration, the
387 @var{architecture} argument identifies the particular architecture in
388 the 960 family, enabling some safeguards and modifying the
389 archive-library search path. @xref{i960,,@command{ld} and the Intel 960
390 family}, for details.
391
392 Future releases of @command{ld} may support similar functionality for
393 other architecture families.
394 @end ifset
395
396 @ifclear SingleFormat
397 @cindex binary input format
398 @kindex -b @var{format}
399 @kindex --format=@var{format}
400 @cindex input format
401 @cindex input format
402 @item -b @var{input-format}
403 @itemx --format=@var{input-format}
404 @command{ld} may be configured to support more than one kind of object
405 file. If your @command{ld} is configured this way, you can use the
406 @samp{-b} option to specify the binary format for input object files
407 that follow this option on the command line. Even when @command{ld} is
408 configured to support alternative object formats, you don't usually need
409 to specify this, as @command{ld} should be configured to expect as a
410 default input format the most usual format on each machine.
411 @var{input-format} is a text string, the name of a particular format
412 supported by the BFD libraries. (You can list the available binary
413 formats with @samp{objdump -i}.)
414 @xref{BFD}.
415
416 You may want to use this option if you are linking files with an unusual
417 binary format. You can also use @samp{-b} to switch formats explicitly (when
418 linking object files of different formats), by including
419 @samp{-b @var{input-format}} before each group of object files in a
420 particular format.
421
422 The default format is taken from the environment variable
423 @code{GNUTARGET}.
424 @ifset UsesEnvVars
425 @xref{Environment}.
426 @end ifset
427 You can also define the input format from a script, using the command
428 @code{TARGET};
429 @ifclear man
430 see @ref{Format Commands}.
431 @end ifclear
432 @end ifclear
433
434 @kindex -c @var{MRI-cmdfile}
435 @kindex --mri-script=@var{MRI-cmdfile}
436 @cindex compatibility, MRI
437 @item -c @var{MRI-commandfile}
438 @itemx --mri-script=@var{MRI-commandfile}
439 For compatibility with linkers produced by MRI, @command{ld} accepts script
440 files written in an alternate, restricted command language, described in
441 @ifclear man
442 @ref{MRI,,MRI Compatible Script Files}.
443 @end ifclear
444 @ifset man
445 the MRI Compatible Script Files section of GNU ld documentation.
446 @end ifset
447 Introduce MRI script files with
448 the option @samp{-c}; use the @samp{-T} option to run linker
449 scripts written in the general-purpose @command{ld} scripting language.
450 If @var{MRI-cmdfile} does not exist, @command{ld} looks for it in the directories
451 specified by any @samp{-L} options.
452
453 @cindex common allocation
454 @kindex -d
455 @kindex -dc
456 @kindex -dp
457 @item -d
458 @itemx -dc
459 @itemx -dp
460 These three options are equivalent; multiple forms are supported for
461 compatibility with other linkers. They assign space to common symbols
462 even if a relocatable output file is specified (with @samp{-r}). The
463 script command @code{FORCE_COMMON_ALLOCATION} has the same effect.
464 @xref{Miscellaneous Commands}.
465
466 @kindex --depaudit @var{AUDITLIB}
467 @kindex -P @var{AUDITLIB}
468 @item --depaudit @var{AUDITLIB}
469 @itemx -P @var{AUDITLIB}
470 Adds @var{AUDITLIB} to the @code{DT_DEPAUDIT} entry of the dynamic section.
471 @var{AUDITLIB} is not checked for existence, nor will it use the DT_SONAME
472 specified in the library. If specified multiple times @code{DT_DEPAUDIT}
473 will contain a colon separated list of audit interfaces to use. This
474 option is only meaningful on ELF platforms supporting the rtld-audit interface.
475 The -P option is provided for Solaris compatibility.
476
477 @cindex entry point, from command line
478 @kindex -e @var{entry}
479 @kindex --entry=@var{entry}
480 @item -e @var{entry}
481 @itemx --entry=@var{entry}
482 Use @var{entry} as the explicit symbol for beginning execution of your
483 program, rather than the default entry point. If there is no symbol
484 named @var{entry}, the linker will try to parse @var{entry} as a number,
485 and use that as the entry address (the number will be interpreted in
486 base 10; you may use a leading @samp{0x} for base 16, or a leading
487 @samp{0} for base 8). @xref{Entry Point}, for a discussion of defaults
488 and other ways of specifying the entry point.
489
490 @kindex --exclude-libs
491 @item --exclude-libs @var{lib},@var{lib},...
492 Specifies a list of archive libraries from which symbols should not be automatically
493 exported. The library names may be delimited by commas or colons. Specifying
494 @code{--exclude-libs ALL} excludes symbols in all archive libraries from
495 automatic export. This option is available only for the i386 PE targeted
496 port of the linker and for ELF targeted ports. For i386 PE, symbols
497 explicitly listed in a .def file are still exported, regardless of this
498 option. For ELF targeted ports, symbols affected by this option will
499 be treated as hidden.
500
501 @kindex --exclude-modules-for-implib
502 @item --exclude-modules-for-implib @var{module},@var{module},...
503 Specifies a list of object files or archive members, from which symbols
504 should not be automatically exported, but which should be copied wholesale
505 into the import library being generated during the link. The module names
506 may be delimited by commas or colons, and must match exactly the filenames
507 used by @command{ld} to open the files; for archive members, this is simply
508 the member name, but for object files the name listed must include and
509 match precisely any path used to specify the input file on the linker's
510 command-line. This option is available only for the i386 PE targeted port
511 of the linker. Symbols explicitly listed in a .def file are still exported,
512 regardless of this option.
513
514 @cindex dynamic symbol table
515 @kindex -E
516 @kindex --export-dynamic
517 @kindex --no-export-dynamic
518 @item -E
519 @itemx --export-dynamic
520 @itemx --no-export-dynamic
521 When creating a dynamically linked executable, using the @option{-E}
522 option or the @option{--export-dynamic} option causes the linker to add
523 all symbols to the dynamic symbol table. The dynamic symbol table is the
524 set of symbols which are visible from dynamic objects at run time.
525
526 If you do not use either of these options (or use the
527 @option{--no-export-dynamic} option to restore the default behavior), the
528 dynamic symbol table will normally contain only those symbols which are
529 referenced by some dynamic object mentioned in the link.
530
531 If you use @code{dlopen} to load a dynamic object which needs to refer
532 back to the symbols defined by the program, rather than some other
533 dynamic object, then you will probably need to use this option when
534 linking the program itself.
535
536 You can also use the dynamic list to control what symbols should
537 be added to the dynamic symbol table if the output format supports it.
538 See the description of @samp{--dynamic-list}.
539
540 Note that this option is specific to ELF targeted ports. PE targets
541 support a similar function to export all symbols from a DLL or EXE; see
542 the description of @samp{--export-all-symbols} below.
543
544 @ifclear SingleFormat
545 @cindex big-endian objects
546 @cindex endianness
547 @kindex -EB
548 @item -EB
549 Link big-endian objects. This affects the default output format.
550
551 @cindex little-endian objects
552 @kindex -EL
553 @item -EL
554 Link little-endian objects. This affects the default output format.
555 @end ifclear
556
557 @kindex -f @var{name}
558 @kindex --auxiliary=@var{name}
559 @item -f @var{name}
560 @itemx --auxiliary=@var{name}
561 When creating an ELF shared object, set the internal DT_AUXILIARY field
562 to the specified name. This tells the dynamic linker that the symbol
563 table of the shared object should be used as an auxiliary filter on the
564 symbol table of the shared object @var{name}.
565
566 If you later link a program against this filter object, then, when you
567 run the program, the dynamic linker will see the DT_AUXILIARY field. If
568 the dynamic linker resolves any symbols from the filter object, it will
569 first check whether there is a definition in the shared object
570 @var{name}. If there is one, it will be used instead of the definition
571 in the filter object. The shared object @var{name} need not exist.
572 Thus the shared object @var{name} may be used to provide an alternative
573 implementation of certain functions, perhaps for debugging or for
574 machine specific performance.
575
576 This option may be specified more than once. The DT_AUXILIARY entries
577 will be created in the order in which they appear on the command line.
578
579 @kindex -F @var{name}
580 @kindex --filter=@var{name}
581 @item -F @var{name}
582 @itemx --filter=@var{name}
583 When creating an ELF shared object, set the internal DT_FILTER field to
584 the specified name. This tells the dynamic linker that the symbol table
585 of the shared object which is being created should be used as a filter
586 on the symbol table of the shared object @var{name}.
587
588 If you later link a program against this filter object, then, when you
589 run the program, the dynamic linker will see the DT_FILTER field. The
590 dynamic linker will resolve symbols according to the symbol table of the
591 filter object as usual, but it will actually link to the definitions
592 found in the shared object @var{name}. Thus the filter object can be
593 used to select a subset of the symbols provided by the object
594 @var{name}.
595
596 Some older linkers used the @option{-F} option throughout a compilation
597 toolchain for specifying object-file format for both input and output
598 object files.
599 @ifclear SingleFormat
600 The @sc{gnu} linker uses other mechanisms for this purpose: the
601 @option{-b}, @option{--format}, @option{--oformat} options, the
602 @code{TARGET} command in linker scripts, and the @code{GNUTARGET}
603 environment variable.
604 @end ifclear
605 The @sc{gnu} linker will ignore the @option{-F} option when not
606 creating an ELF shared object.
607
608 @cindex finalization function
609 @kindex -fini=@var{name}
610 @item -fini=@var{name}
611 When creating an ELF executable or shared object, call NAME when the
612 executable or shared object is unloaded, by setting DT_FINI to the
613 address of the function. By default, the linker uses @code{_fini} as
614 the function to call.
615
616 @kindex -g
617 @item -g
618 Ignored. Provided for compatibility with other tools.
619
620 @kindex -G @var{value}
621 @kindex --gpsize=@var{value}
622 @cindex object size
623 @item -G @var{value}
624 @itemx --gpsize=@var{value}
625 Set the maximum size of objects to be optimized using the GP register to
626 @var{size}. This is only meaningful for object file formats such as
627 MIPS ECOFF which supports putting large and small objects into different
628 sections. This is ignored for other object file formats.
629
630 @cindex runtime library name
631 @kindex -h @var{name}
632 @kindex -soname=@var{name}
633 @item -h @var{name}
634 @itemx -soname=@var{name}
635 When creating an ELF shared object, set the internal DT_SONAME field to
636 the specified name. When an executable is linked with a shared object
637 which has a DT_SONAME field, then when the executable is run the dynamic
638 linker will attempt to load the shared object specified by the DT_SONAME
639 field rather than the using the file name given to the linker.
640
641 @kindex -i
642 @cindex incremental link
643 @item -i
644 Perform an incremental link (same as option @samp{-r}).
645
646 @cindex initialization function
647 @kindex -init=@var{name}
648 @item -init=@var{name}
649 When creating an ELF executable or shared object, call NAME when the
650 executable or shared object is loaded, by setting DT_INIT to the address
651 of the function. By default, the linker uses @code{_init} as the
652 function to call.
653
654 @cindex archive files, from cmd line
655 @kindex -l @var{namespec}
656 @kindex --library=@var{namespec}
657 @item -l @var{namespec}
658 @itemx --library=@var{namespec}
659 Add the archive or object file specified by @var{namespec} to the
660 list of files to link. This option may be used any number of times.
661 If @var{namespec} is of the form @file{:@var{filename}}, @command{ld}
662 will search the library path for a file called @var{filename}, otherwise it
663 will search the library path for a file called @file{lib@var{namespec}.a}.
664
665 On systems which support shared libraries, @command{ld} may also search for
666 files other than @file{lib@var{namespec}.a}. Specifically, on ELF
667 and SunOS systems, @command{ld} will search a directory for a library
668 called @file{lib@var{namespec}.so} before searching for one called
669 @file{lib@var{namespec}.a}. (By convention, a @code{.so} extension
670 indicates a shared library.) Note that this behavior does not apply
671 to @file{:@var{filename}}, which always specifies a file called
672 @var{filename}.
673
674 The linker will search an archive only once, at the location where it is
675 specified on the command line. If the archive defines a symbol which
676 was undefined in some object which appeared before the archive on the
677 command line, the linker will include the appropriate file(s) from the
678 archive. However, an undefined symbol in an object appearing later on
679 the command line will not cause the linker to search the archive again.
680
681 See the @option{-(} option for a way to force the linker to search
682 archives multiple times.
683
684 You may list the same archive multiple times on the command line.
685
686 @ifset GENERIC
687 This type of archive searching is standard for Unix linkers. However,
688 if you are using @command{ld} on AIX, note that it is different from the
689 behaviour of the AIX linker.
690 @end ifset
691
692 @cindex search directory, from cmd line
693 @kindex -L @var{dir}
694 @kindex --library-path=@var{dir}
695 @item -L @var{searchdir}
696 @itemx --library-path=@var{searchdir}
697 Add path @var{searchdir} to the list of paths that @command{ld} will search
698 for archive libraries and @command{ld} control scripts. You may use this
699 option any number of times. The directories are searched in the order
700 in which they are specified on the command line. Directories specified
701 on the command line are searched before the default directories. All
702 @option{-L} options apply to all @option{-l} options, regardless of the
703 order in which the options appear. @option{-L} options do not affect
704 how @command{ld} searches for a linker script unless @option{-T}
705 option is specified.
706
707 If @var{searchdir} begins with @code{=}, then the @code{=} will be replaced
708 by the @dfn{sysroot prefix}, a path specified when the linker is configured.
709
710 @ifset UsesEnvVars
711 The default set of paths searched (without being specified with
712 @samp{-L}) depends on which emulation mode @command{ld} is using, and in
713 some cases also on how it was configured. @xref{Environment}.
714 @end ifset
715
716 The paths can also be specified in a link script with the
717 @code{SEARCH_DIR} command. Directories specified this way are searched
718 at the point in which the linker script appears in the command line.
719
720 @cindex emulation
721 @kindex -m @var{emulation}
722 @item -m @var{emulation}
723 Emulate the @var{emulation} linker. You can list the available
724 emulations with the @samp{--verbose} or @samp{-V} options.
725
726 If the @samp{-m} option is not used, the emulation is taken from the
727 @code{LDEMULATION} environment variable, if that is defined.
728
729 Otherwise, the default emulation depends upon how the linker was
730 configured.
731
732 @cindex link map
733 @kindex -M
734 @kindex --print-map
735 @item -M
736 @itemx --print-map
737 Print a link map to the standard output. A link map provides
738 information about the link, including the following:
739
740 @itemize @bullet
741 @item
742 Where object files are mapped into memory.
743 @item
744 How common symbols are allocated.
745 @item
746 All archive members included in the link, with a mention of the symbol
747 which caused the archive member to be brought in.
748 @item
749 The values assigned to symbols.
750
751 Note - symbols whose values are computed by an expression which
752 involves a reference to a previous value of the same symbol may not
753 have correct result displayed in the link map. This is because the
754 linker discards intermediate results and only retains the final value
755 of an expression. Under such circumstances the linker will display
756 the final value enclosed by square brackets. Thus for example a
757 linker script containing:
758
759 @smallexample
760 foo = 1
761 foo = foo * 4
762 foo = foo + 8
763 @end smallexample
764
765 will produce the following output in the link map if the @option{-M}
766 option is used:
767
768 @smallexample
769 0x00000001 foo = 0x1
770 [0x0000000c] foo = (foo * 0x4)
771 [0x0000000c] foo = (foo + 0x8)
772 @end smallexample
773
774 See @ref{Expressions} for more information about expressions in linker
775 scripts.
776 @end itemize
777
778 @kindex -n
779 @cindex read-only text
780 @cindex NMAGIC
781 @kindex --nmagic
782 @item -n
783 @itemx --nmagic
784 Turn off page alignment of sections, and disable linking against shared
785 libraries. If the output format supports Unix style magic numbers,
786 mark the output as @code{NMAGIC}.
787
788 @kindex -N
789 @kindex --omagic
790 @cindex read/write from cmd line
791 @cindex OMAGIC
792 @item -N
793 @itemx --omagic
794 Set the text and data sections to be readable and writable. Also, do
795 not page-align the data segment, and disable linking against shared
796 libraries. If the output format supports Unix style magic numbers,
797 mark the output as @code{OMAGIC}. Note: Although a writable text section
798 is allowed for PE-COFF targets, it does not conform to the format
799 specification published by Microsoft.
800
801 @kindex --no-omagic
802 @cindex OMAGIC
803 @item --no-omagic
804 This option negates most of the effects of the @option{-N} option. It
805 sets the text section to be read-only, and forces the data segment to
806 be page-aligned. Note - this option does not enable linking against
807 shared libraries. Use @option{-Bdynamic} for this.
808
809 @kindex -o @var{output}
810 @kindex --output=@var{output}
811 @cindex naming the output file
812 @item -o @var{output}
813 @itemx --output=@var{output}
814 Use @var{output} as the name for the program produced by @command{ld}; if this
815 option is not specified, the name @file{a.out} is used by default. The
816 script command @code{OUTPUT} can also specify the output file name.
817
818 @kindex -O @var{level}
819 @cindex generating optimized output
820 @item -O @var{level}
821 If @var{level} is a numeric values greater than zero @command{ld} optimizes
822 the output. This might take significantly longer and therefore probably
823 should only be enabled for the final binary. At the moment this
824 option only affects ELF shared library generation. Future releases of
825 the linker may make more use of this option. Also currently there is
826 no difference in the linker's behaviour for different non-zero values
827 of this option. Again this may change with future releases.
828
829 @kindex -q
830 @kindex --emit-relocs
831 @cindex retain relocations in final executable
832 @item -q
833 @itemx --emit-relocs
834 Leave relocation sections and contents in fully linked executables.
835 Post link analysis and optimization tools may need this information in
836 order to perform correct modifications of executables. This results
837 in larger executables.
838
839 This option is currently only supported on ELF platforms.
840
841 @kindex --force-dynamic
842 @cindex forcing the creation of dynamic sections
843 @item --force-dynamic
844 Force the output file to have dynamic sections. This option is specific
845 to VxWorks targets.
846
847 @cindex partial link
848 @cindex relocatable output
849 @kindex -r
850 @kindex --relocatable
851 @item -r
852 @itemx --relocatable
853 Generate relocatable output---i.e., generate an output file that can in
854 turn serve as input to @command{ld}. This is often called @dfn{partial
855 linking}. As a side effect, in environments that support standard Unix
856 magic numbers, this option also sets the output file's magic number to
857 @code{OMAGIC}.
858 @c ; see @option{-N}.
859 If this option is not specified, an absolute file is produced. When
860 linking C++ programs, this option @emph{will not} resolve references to
861 constructors; to do that, use @samp{-Ur}.
862
863 When an input file does not have the same format as the output file,
864 partial linking is only supported if that input file does not contain any
865 relocations. Different output formats can have further restrictions; for
866 example some @code{a.out}-based formats do not support partial linking
867 with input files in other formats at all.
868
869 This option does the same thing as @samp{-i}.
870
871 @kindex -R @var{file}
872 @kindex --just-symbols=@var{file}
873 @cindex symbol-only input
874 @item -R @var{filename}
875 @itemx --just-symbols=@var{filename}
876 Read symbol names and their addresses from @var{filename}, but do not
877 relocate it or include it in the output. This allows your output file
878 to refer symbolically to absolute locations of memory defined in other
879 programs. You may use this option more than once.
880
881 For compatibility with other ELF linkers, if the @option{-R} option is
882 followed by a directory name, rather than a file name, it is treated as
883 the @option{-rpath} option.
884
885 @kindex -s
886 @kindex --strip-all
887 @cindex strip all symbols
888 @item -s
889 @itemx --strip-all
890 Omit all symbol information from the output file.
891
892 @kindex -S
893 @kindex --strip-debug
894 @cindex strip debugger symbols
895 @item -S
896 @itemx --strip-debug
897 Omit debugger symbol information (but not all symbols) from the output file.
898
899 @kindex -t
900 @kindex --trace
901 @cindex input files, displaying
902 @item -t
903 @itemx --trace
904 Print the names of the input files as @command{ld} processes them.
905
906 @kindex -T @var{script}
907 @kindex --script=@var{script}
908 @cindex script files
909 @item -T @var{scriptfile}
910 @itemx --script=@var{scriptfile}
911 Use @var{scriptfile} as the linker script. This script replaces
912 @command{ld}'s default linker script (rather than adding to it), so
913 @var{commandfile} must specify everything necessary to describe the
914 output file. @xref{Scripts}. If @var{scriptfile} does not exist in
915 the current directory, @code{ld} looks for it in the directories
916 specified by any preceding @samp{-L} options. Multiple @samp{-T}
917 options accumulate.
918
919 @kindex -dT @var{script}
920 @kindex --default-script=@var{script}
921 @cindex script files
922 @item -dT @var{scriptfile}
923 @itemx --default-script=@var{scriptfile}
924 Use @var{scriptfile} as the default linker script. @xref{Scripts}.
925
926 This option is similar to the @option{--script} option except that
927 processing of the script is delayed until after the rest of the
928 command line has been processed. This allows options placed after the
929 @option{--default-script} option on the command line to affect the
930 behaviour of the linker script, which can be important when the linker
931 command line cannot be directly controlled by the user. (eg because
932 the command line is being constructed by another tool, such as
933 @samp{gcc}).
934
935 @kindex -u @var{symbol}
936 @kindex --undefined=@var{symbol}
937 @cindex undefined symbol
938 @item -u @var{symbol}
939 @itemx --undefined=@var{symbol}
940 Force @var{symbol} to be entered in the output file as an undefined
941 symbol. Doing this may, for example, trigger linking of additional
942 modules from standard libraries. @samp{-u} may be repeated with
943 different option arguments to enter additional undefined symbols. This
944 option is equivalent to the @code{EXTERN} linker script command.
945
946 @kindex -Ur
947 @cindex constructors
948 @item -Ur
949 For anything other than C++ programs, this option is equivalent to
950 @samp{-r}: it generates relocatable output---i.e., an output file that can in
951 turn serve as input to @command{ld}. When linking C++ programs, @samp{-Ur}
952 @emph{does} resolve references to constructors, unlike @samp{-r}.
953 It does not work to use @samp{-Ur} on files that were themselves linked
954 with @samp{-Ur}; once the constructor table has been built, it cannot
955 be added to. Use @samp{-Ur} only for the last partial link, and
956 @samp{-r} for the others.
957
958 @kindex --unique[=@var{SECTION}]
959 @item --unique[=@var{SECTION}]
960 Creates a separate output section for every input section matching
961 @var{SECTION}, or if the optional wildcard @var{SECTION} argument is
962 missing, for every orphan input section. An orphan section is one not
963 specifically mentioned in a linker script. You may use this option
964 multiple times on the command line; It prevents the normal merging of
965 input sections with the same name, overriding output section assignments
966 in a linker script.
967
968 @kindex -v
969 @kindex -V
970 @kindex --version
971 @cindex version
972 @item -v
973 @itemx --version
974 @itemx -V
975 Display the version number for @command{ld}. The @option{-V} option also
976 lists the supported emulations.
977
978 @kindex -x
979 @kindex --discard-all
980 @cindex deleting local symbols
981 @item -x
982 @itemx --discard-all
983 Delete all local symbols.
984
985 @kindex -X
986 @kindex --discard-locals
987 @cindex local symbols, deleting
988 @item -X
989 @itemx --discard-locals
990 Delete all temporary local symbols. (These symbols start with
991 system-specific local label prefixes, typically @samp{.L} for ELF systems
992 or @samp{L} for traditional a.out systems.)
993
994 @kindex -y @var{symbol}
995 @kindex --trace-symbol=@var{symbol}
996 @cindex symbol tracing
997 @item -y @var{symbol}
998 @itemx --trace-symbol=@var{symbol}
999 Print the name of each linked file in which @var{symbol} appears. This
1000 option may be given any number of times. On many systems it is necessary
1001 to prepend an underscore.
1002
1003 This option is useful when you have an undefined symbol in your link but
1004 don't know where the reference is coming from.
1005
1006 @kindex -Y @var{path}
1007 @item -Y @var{path}
1008 Add @var{path} to the default library search path. This option exists
1009 for Solaris compatibility.
1010
1011 @kindex -z @var{keyword}
1012 @item -z @var{keyword}
1013 The recognized keywords are:
1014 @table @samp
1015
1016 @item combreloc
1017 Combines multiple reloc sections and sorts them to make dynamic symbol
1018 lookup caching possible.
1019
1020 @item defs
1021 Disallows undefined symbols in object files. Undefined symbols in
1022 shared libraries are still allowed.
1023
1024 @item execstack
1025 Marks the object as requiring executable stack.
1026
1027 @item initfirst
1028 This option is only meaningful when building a shared object.
1029 It marks the object so that its runtime initialization will occur
1030 before the runtime initialization of any other objects brought into
1031 the process at the same time. Similarly the runtime finalization of
1032 the object will occur after the runtime finalization of any other
1033 objects.
1034
1035 @item interpose
1036 Marks the object that its symbol table interposes before all symbols
1037 but the primary executable.
1038
1039 @item lazy
1040 When generating an executable or shared library, mark it to tell the
1041 dynamic linker to defer function call resolution to the point when
1042 the function is called (lazy binding), rather than at load time.
1043 Lazy binding is the default.
1044
1045 @item loadfltr
1046 Marks the object that its filters be processed immediately at
1047 runtime.
1048
1049 @item muldefs
1050 Allows multiple definitions.
1051
1052 @item nocombreloc
1053 Disables multiple reloc sections combining.
1054
1055 @item nocopyreloc
1056 Disables production of copy relocs.
1057
1058 @item nodefaultlib
1059 Marks the object that the search for dependencies of this object will
1060 ignore any default library search paths.
1061
1062 @item nodelete
1063 Marks the object shouldn't be unloaded at runtime.
1064
1065 @item nodlopen
1066 Marks the object not available to @code{dlopen}.
1067
1068 @item nodump
1069 Marks the object can not be dumped by @code{dldump}.
1070
1071 @item noexecstack
1072 Marks the object as not requiring executable stack.
1073
1074 @item norelro
1075 Don't create an ELF @code{PT_GNU_RELRO} segment header in the object.
1076
1077 @item now
1078 When generating an executable or shared library, mark it to tell the
1079 dynamic linker to resolve all symbols when the program is started, or
1080 when the shared library is linked to using dlopen, instead of
1081 deferring function call resolution to the point when the function is
1082 first called.
1083
1084 @item origin
1085 Marks the object may contain $ORIGIN.
1086
1087 @item relro
1088 Create an ELF @code{PT_GNU_RELRO} segment header in the object.
1089
1090 @item max-page-size=@var{value}
1091 Set the emulation maximum page size to @var{value}.
1092
1093 @item common-page-size=@var{value}
1094 Set the emulation common page size to @var{value}.
1095
1096 @item stack-size=@var{value}
1097 Specify a stack size for in an ELF @code{PT_GNU_STACK} segment.
1098 Specifying zero will override any default non-zero sized
1099 @code{PT_GNU_STACK} segment creation.
1100
1101 @end table
1102
1103 Other keywords are ignored for Solaris compatibility.
1104
1105 @kindex -(
1106 @cindex groups of archives
1107 @item -( @var{archives} -)
1108 @itemx --start-group @var{archives} --end-group
1109 The @var{archives} should be a list of archive files. They may be
1110 either explicit file names, or @samp{-l} options.
1111
1112 The specified archives are searched repeatedly until no new undefined
1113 references are created. Normally, an archive is searched only once in
1114 the order that it is specified on the command line. If a symbol in that
1115 archive is needed to resolve an undefined symbol referred to by an
1116 object in an archive that appears later on the command line, the linker
1117 would not be able to resolve that reference. By grouping the archives,
1118 they all be searched repeatedly until all possible references are
1119 resolved.
1120
1121 Using this option has a significant performance cost. It is best to use
1122 it only when there are unavoidable circular references between two or
1123 more archives.
1124
1125 @kindex --accept-unknown-input-arch
1126 @kindex --no-accept-unknown-input-arch
1127 @item --accept-unknown-input-arch
1128 @itemx --no-accept-unknown-input-arch
1129 Tells the linker to accept input files whose architecture cannot be
1130 recognised. The assumption is that the user knows what they are doing
1131 and deliberately wants to link in these unknown input files. This was
1132 the default behaviour of the linker, before release 2.14. The default
1133 behaviour from release 2.14 onwards is to reject such input files, and
1134 so the @samp{--accept-unknown-input-arch} option has been added to
1135 restore the old behaviour.
1136
1137 @kindex --as-needed
1138 @kindex --no-as-needed
1139 @item --as-needed
1140 @itemx --no-as-needed
1141 This option affects ELF DT_NEEDED tags for dynamic libraries mentioned
1142 on the command line after the @option{--as-needed} option. Normally
1143 the linker will add a DT_NEEDED tag for each dynamic library mentioned
1144 on the command line, regardless of whether the library is actually
1145 needed or not. @option{--as-needed} causes a DT_NEEDED tag to only be
1146 emitted for a library that satisfies an undefined symbol reference
1147 from a regular object file or, if the library is not found in the
1148 DT_NEEDED lists of other libraries linked up to that point, an
1149 undefined symbol reference from another dynamic library.
1150 @option{--no-as-needed} restores the default behaviour.
1151
1152 @kindex --add-needed
1153 @kindex --no-add-needed
1154 @item --add-needed
1155 @itemx --no-add-needed
1156 These two options have been deprecated because of the similarity of
1157 their names to the @option{--as-needed} and @option{--no-as-needed}
1158 options. They have been replaced by @option{--copy-dt-needed-entries}
1159 and @option{--no-copy-dt-needed-entries}.
1160
1161 @kindex -assert @var{keyword}
1162 @item -assert @var{keyword}
1163 This option is ignored for SunOS compatibility.
1164
1165 @kindex -Bdynamic
1166 @kindex -dy
1167 @kindex -call_shared
1168 @item -Bdynamic
1169 @itemx -dy
1170 @itemx -call_shared
1171 Link against dynamic libraries. This is only meaningful on platforms
1172 for which shared libraries are supported. This option is normally the
1173 default on such platforms. The different variants of this option are
1174 for compatibility with various systems. You may use this option
1175 multiple times on the command line: it affects library searching for
1176 @option{-l} options which follow it.
1177
1178 @kindex -Bgroup
1179 @item -Bgroup
1180 Set the @code{DF_1_GROUP} flag in the @code{DT_FLAGS_1} entry in the dynamic
1181 section. This causes the runtime linker to handle lookups in this
1182 object and its dependencies to be performed only inside the group.
1183 @option{--unresolved-symbols=report-all} is implied. This option is
1184 only meaningful on ELF platforms which support shared libraries.
1185
1186 @kindex -Bstatic
1187 @kindex -dn
1188 @kindex -non_shared
1189 @kindex -static
1190 @item -Bstatic
1191 @itemx -dn
1192 @itemx -non_shared
1193 @itemx -static
1194 Do not link against shared libraries. This is only meaningful on
1195 platforms for which shared libraries are supported. The different
1196 variants of this option are for compatibility with various systems. You
1197 may use this option multiple times on the command line: it affects
1198 library searching for @option{-l} options which follow it. This
1199 option also implies @option{--unresolved-symbols=report-all}. This
1200 option can be used with @option{-shared}. Doing so means that a
1201 shared library is being created but that all of the library's external
1202 references must be resolved by pulling in entries from static
1203 libraries.
1204
1205 @kindex -Bsymbolic
1206 @item -Bsymbolic
1207 When creating a shared library, bind references to global symbols to the
1208 definition within the shared library, if any. Normally, it is possible
1209 for a program linked against a shared library to override the definition
1210 within the shared library. This option is only meaningful on ELF
1211 platforms which support shared libraries.
1212
1213 @kindex -Bsymbolic-functions
1214 @item -Bsymbolic-functions
1215 When creating a shared library, bind references to global function
1216 symbols to the definition within the shared library, if any.
1217 This option is only meaningful on ELF platforms which support shared
1218 libraries.
1219
1220 @kindex --dynamic-list=@var{dynamic-list-file}
1221 @item --dynamic-list=@var{dynamic-list-file}
1222 Specify the name of a dynamic list file to the linker. This is
1223 typically used when creating shared libraries to specify a list of
1224 global symbols whose references shouldn't be bound to the definition
1225 within the shared library, or creating dynamically linked executables
1226 to specify a list of symbols which should be added to the symbol table
1227 in the executable. This option is only meaningful on ELF platforms
1228 which support shared libraries.
1229
1230 The format of the dynamic list is the same as the version node without
1231 scope and node name. See @ref{VERSION} for more information.
1232
1233 @kindex --dynamic-list-data
1234 @item --dynamic-list-data
1235 Include all global data symbols to the dynamic list.
1236
1237 @kindex --dynamic-list-cpp-new
1238 @item --dynamic-list-cpp-new
1239 Provide the builtin dynamic list for C++ operator new and delete. It
1240 is mainly useful for building shared libstdc++.
1241
1242 @kindex --dynamic-list-cpp-typeinfo
1243 @item --dynamic-list-cpp-typeinfo
1244 Provide the builtin dynamic list for C++ runtime type identification.
1245
1246 @kindex --check-sections
1247 @kindex --no-check-sections
1248 @item --check-sections
1249 @itemx --no-check-sections
1250 Asks the linker @emph{not} to check section addresses after they have
1251 been assigned to see if there are any overlaps. Normally the linker will
1252 perform this check, and if it finds any overlaps it will produce
1253 suitable error messages. The linker does know about, and does make
1254 allowances for sections in overlays. The default behaviour can be
1255 restored by using the command line switch @option{--check-sections}.
1256 Section overlap is not usually checked for relocatable links. You can
1257 force checking in that case by using the @option{--check-sections}
1258 option.
1259
1260 @kindex --copy-dt-needed-entries
1261 @kindex --no-copy-dt-needed-entries
1262 @item --copy-dt-needed-entries
1263 @itemx --no-copy-dt-needed-entries
1264 This option affects the treatment of dynamic libraries referred to
1265 by DT_NEEDED tags @emph{inside} ELF dynamic libraries mentioned on the
1266 command line. Normally the linker won't add a DT_NEEDED tag to the
1267 output binary for each library mentioned in a DT_NEEDED tag in an
1268 input dynamic library. With @option{--copy-dt-needed-entries}
1269 specified on the command line however any dynamic libraries that
1270 follow it will have their DT_NEEDED entries added. The default
1271 behaviour can be restored with @option{--no-copy-dt-needed-entries}.
1272
1273 This option also has an effect on the resolution of symbols in dynamic
1274 libraries. With @option{--copy-dt-needed-entries} dynamic libraries
1275 mentioned on the command line will be recursively searched, following
1276 their DT_NEEDED tags to other libraries, in order to resolve symbols
1277 required by the output binary. With the default setting however
1278 the searching of dynamic libraries that follow it will stop with the
1279 dynamic library itself. No DT_NEEDED links will be traversed to resolve
1280 symbols.
1281
1282 @cindex cross reference table
1283 @kindex --cref
1284 @item --cref
1285 Output a cross reference table. If a linker map file is being
1286 generated, the cross reference table is printed to the map file.
1287 Otherwise, it is printed on the standard output.
1288
1289 The format of the table is intentionally simple, so that it may be
1290 easily processed by a script if necessary. The symbols are printed out,
1291 sorted by name. For each symbol, a list of file names is given. If the
1292 symbol is defined, the first file listed is the location of the
1293 definition. The remaining files contain references to the symbol.
1294
1295 @cindex common allocation
1296 @kindex --no-define-common
1297 @item --no-define-common
1298 This option inhibits the assignment of addresses to common symbols.
1299 The script command @code{INHIBIT_COMMON_ALLOCATION} has the same effect.
1300 @xref{Miscellaneous Commands}.
1301
1302 The @samp{--no-define-common} option allows decoupling
1303 the decision to assign addresses to Common symbols from the choice
1304 of the output file type; otherwise a non-Relocatable output type
1305 forces assigning addresses to Common symbols.
1306 Using @samp{--no-define-common} allows Common symbols that are referenced
1307 from a shared library to be assigned addresses only in the main program.
1308 This eliminates the unused duplicate space in the shared library,
1309 and also prevents any possible confusion over resolving to the wrong
1310 duplicate when there are many dynamic modules with specialized search
1311 paths for runtime symbol resolution.
1312
1313 @cindex symbols, from command line
1314 @kindex --defsym=@var{symbol}=@var{exp}
1315 @item --defsym=@var{symbol}=@var{expression}
1316 Create a global symbol in the output file, containing the absolute
1317 address given by @var{expression}. You may use this option as many
1318 times as necessary to define multiple symbols in the command line. A
1319 limited form of arithmetic is supported for the @var{expression} in this
1320 context: you may give a hexadecimal constant or the name of an existing
1321 symbol, or use @code{+} and @code{-} to add or subtract hexadecimal
1322 constants or symbols. If you need more elaborate expressions, consider
1323 using the linker command language from a script (@pxref{Assignments,,
1324 Assignment: Symbol Definitions}). @emph{Note:} there should be no white
1325 space between @var{symbol}, the equals sign (``@key{=}''), and
1326 @var{expression}.
1327
1328 @cindex demangling, from command line
1329 @kindex --demangle[=@var{style}]
1330 @kindex --no-demangle
1331 @item --demangle[=@var{style}]
1332 @itemx --no-demangle
1333 These options control whether to demangle symbol names in error messages
1334 and other output. When the linker is told to demangle, it tries to
1335 present symbol names in a readable fashion: it strips leading
1336 underscores if they are used by the object file format, and converts C++
1337 mangled symbol names into user readable names. Different compilers have
1338 different mangling styles. The optional demangling style argument can be used
1339 to choose an appropriate demangling style for your compiler. The linker will
1340 demangle by default unless the environment variable @samp{COLLECT_NO_DEMANGLE}
1341 is set. These options may be used to override the default.
1342
1343 @cindex dynamic linker, from command line
1344 @kindex -I@var{file}
1345 @kindex --dynamic-linker=@var{file}
1346 @item -I@var{file}
1347 @itemx --dynamic-linker=@var{file}
1348 Set the name of the dynamic linker. This is only meaningful when
1349 generating dynamically linked ELF executables. The default dynamic
1350 linker is normally correct; don't use this unless you know what you are
1351 doing.
1352
1353 @kindex --fatal-warnings
1354 @kindex --no-fatal-warnings
1355 @item --fatal-warnings
1356 @itemx --no-fatal-warnings
1357 Treat all warnings as errors. The default behaviour can be restored
1358 with the option @option{--no-fatal-warnings}.
1359
1360 @kindex --force-exe-suffix
1361 @item --force-exe-suffix
1362 Make sure that an output file has a .exe suffix.
1363
1364 If a successfully built fully linked output file does not have a
1365 @code{.exe} or @code{.dll} suffix, this option forces the linker to copy
1366 the output file to one of the same name with a @code{.exe} suffix. This
1367 option is useful when using unmodified Unix makefiles on a Microsoft
1368 Windows host, since some versions of Windows won't run an image unless
1369 it ends in a @code{.exe} suffix.
1370
1371 @kindex --gc-sections
1372 @kindex --no-gc-sections
1373 @cindex garbage collection
1374 @item --gc-sections
1375 @itemx --no-gc-sections
1376 Enable garbage collection of unused input sections. It is ignored on
1377 targets that do not support this option. The default behaviour (of not
1378 performing this garbage collection) can be restored by specifying
1379 @samp{--no-gc-sections} on the command line.
1380
1381 @samp{--gc-sections} decides which input sections are used by
1382 examining symbols and relocations. The section containing the entry
1383 symbol and all sections containing symbols undefined on the
1384 command-line will be kept, as will sections containing symbols
1385 referenced by dynamic objects. Note that when building shared
1386 libraries, the linker must assume that any visible symbol is
1387 referenced. Once this initial set of sections has been determined,
1388 the linker recursively marks as used any section referenced by their
1389 relocations. See @samp{--entry} and @samp{--undefined}.
1390
1391 This option can be set when doing a partial link (enabled with option
1392 @samp{-r}). In this case the root of symbols kept must be explicitly
1393 specified either by an @samp{--entry} or @samp{--undefined} option or by
1394 a @code{ENTRY} command in the linker script.
1395
1396 @kindex --print-gc-sections
1397 @kindex --no-print-gc-sections
1398 @cindex garbage collection
1399 @item --print-gc-sections
1400 @itemx --no-print-gc-sections
1401 List all sections removed by garbage collection. The listing is
1402 printed on stderr. This option is only effective if garbage
1403 collection has been enabled via the @samp{--gc-sections}) option. The
1404 default behaviour (of not listing the sections that are removed) can
1405 be restored by specifying @samp{--no-print-gc-sections} on the command
1406 line.
1407
1408 @kindex --print-output-format
1409 @cindex output format
1410 @item --print-output-format
1411 Print the name of the default output format (perhaps influenced by
1412 other command-line options). This is the string that would appear
1413 in an @code{OUTPUT_FORMAT} linker script command (@pxref{File Commands}).
1414
1415 @cindex help
1416 @cindex usage
1417 @kindex --help
1418 @item --help
1419 Print a summary of the command-line options on the standard output and exit.
1420
1421 @kindex --target-help
1422 @item --target-help
1423 Print a summary of all target specific options on the standard output and exit.
1424
1425 @kindex -Map=@var{mapfile}
1426 @item -Map=@var{mapfile}
1427 Print a link map to the file @var{mapfile}. See the description of the
1428 @option{-M} option, above.
1429
1430 @cindex memory usage
1431 @kindex --no-keep-memory
1432 @item --no-keep-memory
1433 @command{ld} normally optimizes for speed over memory usage by caching the
1434 symbol tables of input files in memory. This option tells @command{ld} to
1435 instead optimize for memory usage, by rereading the symbol tables as
1436 necessary. This may be required if @command{ld} runs out of memory space
1437 while linking a large executable.
1438
1439 @kindex --no-undefined
1440 @kindex -z defs
1441 @item --no-undefined
1442 @itemx -z defs
1443 Report unresolved symbol references from regular object files. This
1444 is done even if the linker is creating a non-symbolic shared library.
1445 The switch @option{--[no-]allow-shlib-undefined} controls the
1446 behaviour for reporting unresolved references found in shared
1447 libraries being linked in.
1448
1449 @kindex --allow-multiple-definition
1450 @kindex -z muldefs
1451 @item --allow-multiple-definition
1452 @itemx -z muldefs
1453 Normally when a symbol is defined multiple times, the linker will
1454 report a fatal error. These options allow multiple definitions and the
1455 first definition will be used.
1456
1457 @kindex --allow-shlib-undefined
1458 @kindex --no-allow-shlib-undefined
1459 @item --allow-shlib-undefined
1460 @itemx --no-allow-shlib-undefined
1461 Allows or disallows undefined symbols in shared libraries.
1462 This switch is similar to @option{--no-undefined} except that it
1463 determines the behaviour when the undefined symbols are in a
1464 shared library rather than a regular object file. It does not affect
1465 how undefined symbols in regular object files are handled.
1466
1467 The default behaviour is to report errors for any undefined symbols
1468 referenced in shared libraries if the linker is being used to create
1469 an executable, but to allow them if the linker is being used to create
1470 a shared library.
1471
1472 The reasons for allowing undefined symbol references in shared
1473 libraries specified at link time are that:
1474
1475 @itemize @bullet
1476 @item
1477 A shared library specified at link time may not be the same as the one
1478 that is available at load time, so the symbol might actually be
1479 resolvable at load time.
1480 @item
1481 There are some operating systems, eg BeOS and HPPA, where undefined
1482 symbols in shared libraries are normal.
1483
1484 The BeOS kernel for example patches shared libraries at load time to
1485 select whichever function is most appropriate for the current
1486 architecture. This is used, for example, to dynamically select an
1487 appropriate memset function.
1488 @end itemize
1489
1490 @kindex --no-undefined-version
1491 @item --no-undefined-version
1492 Normally when a symbol has an undefined version, the linker will ignore
1493 it. This option disallows symbols with undefined version and a fatal error
1494 will be issued instead.
1495
1496 @kindex --default-symver
1497 @item --default-symver
1498 Create and use a default symbol version (the soname) for unversioned
1499 exported symbols.
1500
1501 @kindex --default-imported-symver
1502 @item --default-imported-symver
1503 Create and use a default symbol version (the soname) for unversioned
1504 imported symbols.
1505
1506 @kindex --no-warn-mismatch
1507 @item --no-warn-mismatch
1508 Normally @command{ld} will give an error if you try to link together input
1509 files that are mismatched for some reason, perhaps because they have
1510 been compiled for different processors or for different endiannesses.
1511 This option tells @command{ld} that it should silently permit such possible
1512 errors. This option should only be used with care, in cases when you
1513 have taken some special action that ensures that the linker errors are
1514 inappropriate.
1515
1516 @kindex --no-warn-search-mismatch
1517 @item --no-warn-search-mismatch
1518 Normally @command{ld} will give a warning if it finds an incompatible
1519 library during a library search. This option silences the warning.
1520
1521 @kindex --no-whole-archive
1522 @item --no-whole-archive
1523 Turn off the effect of the @option{--whole-archive} option for subsequent
1524 archive files.
1525
1526 @cindex output file after errors
1527 @kindex --noinhibit-exec
1528 @item --noinhibit-exec
1529 Retain the executable output file whenever it is still usable.
1530 Normally, the linker will not produce an output file if it encounters
1531 errors during the link process; it exits without writing an output file
1532 when it issues any error whatsoever.
1533
1534 @kindex -nostdlib
1535 @item -nostdlib
1536 Only search library directories explicitly specified on the
1537 command line. Library directories specified in linker scripts
1538 (including linker scripts specified on the command line) are ignored.
1539
1540 @ifclear SingleFormat
1541 @kindex --oformat=@var{output-format}
1542 @item --oformat=@var{output-format}
1543 @command{ld} may be configured to support more than one kind of object
1544 file. If your @command{ld} is configured this way, you can use the
1545 @samp{--oformat} option to specify the binary format for the output
1546 object file. Even when @command{ld} is configured to support alternative
1547 object formats, you don't usually need to specify this, as @command{ld}
1548 should be configured to produce as a default output format the most
1549 usual format on each machine. @var{output-format} is a text string, the
1550 name of a particular format supported by the BFD libraries. (You can
1551 list the available binary formats with @samp{objdump -i}.) The script
1552 command @code{OUTPUT_FORMAT} can also specify the output format, but
1553 this option overrides it. @xref{BFD}.
1554 @end ifclear
1555
1556 @kindex -pie
1557 @kindex --pic-executable
1558 @item -pie
1559 @itemx --pic-executable
1560 @cindex position independent executables
1561 Create a position independent executable. This is currently only supported on
1562 ELF platforms. Position independent executables are similar to shared
1563 libraries in that they are relocated by the dynamic linker to the virtual
1564 address the OS chooses for them (which can vary between invocations). Like
1565 normal dynamically linked executables they can be executed and symbols
1566 defined in the executable cannot be overridden by shared libraries.
1567
1568 @kindex -qmagic
1569 @item -qmagic
1570 This option is ignored for Linux compatibility.
1571
1572 @kindex -Qy
1573 @item -Qy
1574 This option is ignored for SVR4 compatibility.
1575
1576 @kindex --relax
1577 @cindex synthesizing linker
1578 @cindex relaxing addressing modes
1579 @cindex --no-relax
1580 @item --relax
1581 @itemx --no-relax
1582 An option with machine dependent effects.
1583 @ifset GENERIC
1584 This option is only supported on a few targets.
1585 @end ifset
1586 @ifset H8300
1587 @xref{H8/300,,@command{ld} and the H8/300}.
1588 @end ifset
1589 @ifset I960
1590 @xref{i960,, @command{ld} and the Intel 960 family}.
1591 @end ifset
1592 @ifset XTENSA
1593 @xref{Xtensa,, @command{ld} and Xtensa Processors}.
1594 @end ifset
1595 @ifset M68HC11
1596 @xref{M68HC11/68HC12,,@command{ld} and the 68HC11 and 68HC12}.
1597 @end ifset
1598 @ifset POWERPC
1599 @xref{PowerPC ELF32,,@command{ld} and PowerPC 32-bit ELF Support}.
1600 @end ifset
1601
1602 On some platforms the @samp{--relax} option performs target specific,
1603 global optimizations that become possible when the linker resolves
1604 addressing in the program, such as relaxing address modes,
1605 synthesizing new instructions, selecting shorter version of current
1606 instructions, and combinig constant values.
1607
1608 On some platforms these link time global optimizations may make symbolic
1609 debugging of the resulting executable impossible.
1610 @ifset GENERIC
1611 This is known to be the case for the Matsushita MN10200 and MN10300
1612 family of processors.
1613 @end ifset
1614
1615 @ifset GENERIC
1616 On platforms where this is not supported, @samp{--relax} is accepted,
1617 but ignored.
1618 @end ifset
1619
1620 On platforms where @samp{--relax} is accepted the option
1621 @samp{--no-relax} can be used to disable the feature.
1622
1623 @cindex retaining specified symbols
1624 @cindex stripping all but some symbols
1625 @cindex symbols, retaining selectively
1626 @kindex --retain-symbols-file=@var{filename}
1627 @item --retain-symbols-file=@var{filename}
1628 Retain @emph{only} the symbols listed in the file @var{filename},
1629 discarding all others. @var{filename} is simply a flat file, with one
1630 symbol name per line. This option is especially useful in environments
1631 @ifset GENERIC
1632 (such as VxWorks)
1633 @end ifset
1634 where a large global symbol table is accumulated gradually, to conserve
1635 run-time memory.
1636
1637 @samp{--retain-symbols-file} does @emph{not} discard undefined symbols,
1638 or symbols needed for relocations.
1639
1640 You may only specify @samp{--retain-symbols-file} once in the command
1641 line. It overrides @samp{-s} and @samp{-S}.
1642
1643 @ifset GENERIC
1644 @item -rpath=@var{dir}
1645 @cindex runtime library search path
1646 @kindex -rpath=@var{dir}
1647 Add a directory to the runtime library search path. This is used when
1648 linking an ELF executable with shared objects. All @option{-rpath}
1649 arguments are concatenated and passed to the runtime linker, which uses
1650 them to locate shared objects at runtime. The @option{-rpath} option is
1651 also used when locating shared objects which are needed by shared
1652 objects explicitly included in the link; see the description of the
1653 @option{-rpath-link} option. If @option{-rpath} is not used when linking an
1654 ELF executable, the contents of the environment variable
1655 @code{LD_RUN_PATH} will be used if it is defined.
1656
1657 The @option{-rpath} option may also be used on SunOS. By default, on
1658 SunOS, the linker will form a runtime search patch out of all the
1659 @option{-L} options it is given. If a @option{-rpath} option is used, the
1660 runtime search path will be formed exclusively using the @option{-rpath}
1661 options, ignoring the @option{-L} options. This can be useful when using
1662 gcc, which adds many @option{-L} options which may be on NFS mounted
1663 file systems.
1664
1665 For compatibility with other ELF linkers, if the @option{-R} option is
1666 followed by a directory name, rather than a file name, it is treated as
1667 the @option{-rpath} option.
1668 @end ifset
1669
1670 @ifset GENERIC
1671 @cindex link-time runtime library search path
1672 @kindex -rpath-link=@var{dir}
1673 @item -rpath-link=@var{dir}
1674 When using ELF or SunOS, one shared library may require another. This
1675 happens when an @code{ld -shared} link includes a shared library as one
1676 of the input files.
1677
1678 When the linker encounters such a dependency when doing a non-shared,
1679 non-relocatable link, it will automatically try to locate the required
1680 shared library and include it in the link, if it is not included
1681 explicitly. In such a case, the @option{-rpath-link} option
1682 specifies the first set of directories to search. The
1683 @option{-rpath-link} option may specify a sequence of directory names
1684 either by specifying a list of names separated by colons, or by
1685 appearing multiple times.
1686
1687 This option should be used with caution as it overrides the search path
1688 that may have been hard compiled into a shared library. In such a case it
1689 is possible to use unintentionally a different search path than the
1690 runtime linker would do.
1691
1692 The linker uses the following search paths to locate required shared
1693 libraries:
1694 @enumerate
1695 @item
1696 Any directories specified by @option{-rpath-link} options.
1697 @item
1698 Any directories specified by @option{-rpath} options. The difference
1699 between @option{-rpath} and @option{-rpath-link} is that directories
1700 specified by @option{-rpath} options are included in the executable and
1701 used at runtime, whereas the @option{-rpath-link} option is only effective
1702 at link time. Searching @option{-rpath} in this way is only supported
1703 by native linkers and cross linkers which have been configured with
1704 the @option{--with-sysroot} option.
1705 @item
1706 On an ELF system, for native linkers, if the @option{-rpath} and
1707 @option{-rpath-link} options were not used, search the contents of the
1708 environment variable @code{LD_RUN_PATH}.
1709 @item
1710 On SunOS, if the @option{-rpath} option was not used, search any
1711 directories specified using @option{-L} options.
1712 @item
1713 For a native linker, the search the contents of the environment
1714 variable @code{LD_LIBRARY_PATH}.
1715 @item
1716 For a native ELF linker, the directories in @code{DT_RUNPATH} or
1717 @code{DT_RPATH} of a shared library are searched for shared
1718 libraries needed by it. The @code{DT_RPATH} entries are ignored if
1719 @code{DT_RUNPATH} entries exist.
1720 @item
1721 The default directories, normally @file{/lib} and @file{/usr/lib}.
1722 @item
1723 For a native linker on an ELF system, if the file @file{/etc/ld.so.conf}
1724 exists, the list of directories found in that file.
1725 @end enumerate
1726
1727 If the required shared library is not found, the linker will issue a
1728 warning and continue with the link.
1729 @end ifset
1730
1731 @kindex -shared
1732 @kindex -Bshareable
1733 @item -shared
1734 @itemx -Bshareable
1735 @cindex shared libraries
1736 Create a shared library. This is currently only supported on ELF, XCOFF
1737 and SunOS platforms. On SunOS, the linker will automatically create a
1738 shared library if the @option{-e} option is not used and there are
1739 undefined symbols in the link.
1740
1741 @kindex --sort-common
1742 @item --sort-common
1743 @itemx --sort-common=ascending
1744 @itemx --sort-common=descending
1745 This option tells @command{ld} to sort the common symbols by alignment in
1746 ascending or descending order when it places them in the appropriate output
1747 sections. The symbol alignments considered are sixteen-byte or larger,
1748 eight-byte, four-byte, two-byte, and one-byte. This is to prevent gaps
1749 between symbols due to alignment constraints. If no sorting order is
1750 specified, then descending order is assumed.
1751
1752 @kindex --sort-section=name
1753 @item --sort-section=name
1754 This option will apply @code{SORT_BY_NAME} to all wildcard section
1755 patterns in the linker script.
1756
1757 @kindex --sort-section=alignment
1758 @item --sort-section=alignment
1759 This option will apply @code{SORT_BY_ALIGNMENT} to all wildcard section
1760 patterns in the linker script.
1761
1762 @kindex --split-by-file
1763 @item --split-by-file[=@var{size}]
1764 Similar to @option{--split-by-reloc} but creates a new output section for
1765 each input file when @var{size} is reached. @var{size} defaults to a
1766 size of 1 if not given.
1767
1768 @kindex --split-by-reloc
1769 @item --split-by-reloc[=@var{count}]
1770 Tries to creates extra sections in the output file so that no single
1771 output section in the file contains more than @var{count} relocations.
1772 This is useful when generating huge relocatable files for downloading into
1773 certain real time kernels with the COFF object file format; since COFF
1774 cannot represent more than 65535 relocations in a single section. Note
1775 that this will fail to work with object file formats which do not
1776 support arbitrary sections. The linker will not split up individual
1777 input sections for redistribution, so if a single input section contains
1778 more than @var{count} relocations one output section will contain that
1779 many relocations. @var{count} defaults to a value of 32768.
1780
1781 @kindex --stats
1782 @item --stats
1783 Compute and display statistics about the operation of the linker, such
1784 as execution time and memory usage.
1785
1786 @kindex --sysroot=@var{directory}
1787 @item --sysroot=@var{directory}
1788 Use @var{directory} as the location of the sysroot, overriding the
1789 configure-time default. This option is only supported by linkers
1790 that were configured using @option{--with-sysroot}.
1791
1792 @kindex --traditional-format
1793 @cindex traditional format
1794 @item --traditional-format
1795 For some targets, the output of @command{ld} is different in some ways from
1796 the output of some existing linker. This switch requests @command{ld} to
1797 use the traditional format instead.
1798
1799 @cindex dbx
1800 For example, on SunOS, @command{ld} combines duplicate entries in the
1801 symbol string table. This can reduce the size of an output file with
1802 full debugging information by over 30 percent. Unfortunately, the SunOS
1803 @code{dbx} program can not read the resulting program (@code{gdb} has no
1804 trouble). The @samp{--traditional-format} switch tells @command{ld} to not
1805 combine duplicate entries.
1806
1807 @kindex --section-start=@var{sectionname}=@var{org}
1808 @item --section-start=@var{sectionname}=@var{org}
1809 Locate a section in the output file at the absolute
1810 address given by @var{org}. You may use this option as many
1811 times as necessary to locate multiple sections in the command
1812 line.
1813 @var{org} must be a single hexadecimal integer;
1814 for compatibility with other linkers, you may omit the leading
1815 @samp{0x} usually associated with hexadecimal values. @emph{Note:} there
1816 should be no white space between @var{sectionname}, the equals
1817 sign (``@key{=}''), and @var{org}.
1818
1819 @kindex -Tbss=@var{org}
1820 @kindex -Tdata=@var{org}
1821 @kindex -Ttext=@var{org}
1822 @cindex segment origins, cmd line
1823 @item -Tbss=@var{org}
1824 @itemx -Tdata=@var{org}
1825 @itemx -Ttext=@var{org}
1826 Same as @option{--section-start}, with @code{.bss}, @code{.data} or
1827 @code{.text} as the @var{sectionname}.
1828
1829 @kindex -Ttext-segment=@var{org}
1830 @item -Ttext-segment=@var{org}
1831 @cindex text segment origin, cmd line
1832 When creating an ELF executable or shared object, it will set the address
1833 of the first byte of the text segment.
1834
1835 @kindex --unresolved-symbols
1836 @item --unresolved-symbols=@var{method}
1837 Determine how to handle unresolved symbols. There are four possible
1838 values for @samp{method}:
1839
1840 @table @samp
1841 @item ignore-all
1842 Do not report any unresolved symbols.
1843
1844 @item report-all
1845 Report all unresolved symbols. This is the default.
1846
1847 @item ignore-in-object-files
1848 Report unresolved symbols that are contained in shared libraries, but
1849 ignore them if they come from regular object files.
1850
1851 @item ignore-in-shared-libs
1852 Report unresolved symbols that come from regular object files, but
1853 ignore them if they come from shared libraries. This can be useful
1854 when creating a dynamic binary and it is known that all the shared
1855 libraries that it should be referencing are included on the linker's
1856 command line.
1857 @end table
1858
1859 The behaviour for shared libraries on their own can also be controlled
1860 by the @option{--[no-]allow-shlib-undefined} option.
1861
1862 Normally the linker will generate an error message for each reported
1863 unresolved symbol but the option @option{--warn-unresolved-symbols}
1864 can change this to a warning.
1865
1866 @kindex --verbose[=@var{NUMBER}]
1867 @cindex verbose[=@var{NUMBER}]
1868 @item --dll-verbose
1869 @itemx --verbose[=@var{NUMBER}]
1870 Display the version number for @command{ld} and list the linker emulations
1871 supported. Display which input files can and cannot be opened. Display
1872 the linker script being used by the linker. If the optional @var{NUMBER}
1873 argument > 1, plugin symbol status will also be displayed.
1874
1875 @kindex --version-script=@var{version-scriptfile}
1876 @cindex version script, symbol versions
1877 @item --version-script=@var{version-scriptfile}
1878 Specify the name of a version script to the linker. This is typically
1879 used when creating shared libraries to specify additional information
1880 about the version hierarchy for the library being created. This option
1881 is only fully supported on ELF platforms which support shared libraries;
1882 see @ref{VERSION}. It is partially supported on PE platforms, which can
1883 use version scripts to filter symbol visibility in auto-export mode: any
1884 symbols marked @samp{local} in the version script will not be exported.
1885 @xref{WIN32}.
1886
1887 @kindex --warn-common
1888 @cindex warnings, on combining symbols
1889 @cindex combining symbols, warnings on
1890 @item --warn-common
1891 Warn when a common symbol is combined with another common symbol or with
1892 a symbol definition. Unix linkers allow this somewhat sloppy practise,
1893 but linkers on some other operating systems do not. This option allows
1894 you to find potential problems from combining global symbols.
1895 Unfortunately, some C libraries use this practise, so you may get some
1896 warnings about symbols in the libraries as well as in your programs.
1897
1898 There are three kinds of global symbols, illustrated here by C examples:
1899
1900 @table @samp
1901 @item int i = 1;
1902 A definition, which goes in the initialized data section of the output
1903 file.
1904
1905 @item extern int i;
1906 An undefined reference, which does not allocate space.
1907 There must be either a definition or a common symbol for the
1908 variable somewhere.
1909
1910 @item int i;
1911 A common symbol. If there are only (one or more) common symbols for a
1912 variable, it goes in the uninitialized data area of the output file.
1913 The linker merges multiple common symbols for the same variable into a
1914 single symbol. If they are of different sizes, it picks the largest
1915 size. The linker turns a common symbol into a declaration, if there is
1916 a definition of the same variable.
1917 @end table
1918
1919 The @samp{--warn-common} option can produce five kinds of warnings.
1920 Each warning consists of a pair of lines: the first describes the symbol
1921 just encountered, and the second describes the previous symbol
1922 encountered with the same name. One or both of the two symbols will be
1923 a common symbol.
1924
1925 @enumerate
1926 @item
1927 Turning a common symbol into a reference, because there is already a
1928 definition for the symbol.
1929 @smallexample
1930 @var{file}(@var{section}): warning: common of `@var{symbol}'
1931 overridden by definition
1932 @var{file}(@var{section}): warning: defined here
1933 @end smallexample
1934
1935 @item
1936 Turning a common symbol into a reference, because a later definition for
1937 the symbol is encountered. This is the same as the previous case,
1938 except that the symbols are encountered in a different order.
1939 @smallexample
1940 @var{file}(@var{section}): warning: definition of `@var{symbol}'
1941 overriding common
1942 @var{file}(@var{section}): warning: common is here
1943 @end smallexample
1944
1945 @item
1946 Merging a common symbol with a previous same-sized common symbol.
1947 @smallexample
1948 @var{file}(@var{section}): warning: multiple common
1949 of `@var{symbol}'
1950 @var{file}(@var{section}): warning: previous common is here
1951 @end smallexample
1952
1953 @item
1954 Merging a common symbol with a previous larger common symbol.
1955 @smallexample
1956 @var{file}(@var{section}): warning: common of `@var{symbol}'
1957 overridden by larger common
1958 @var{file}(@var{section}): warning: larger common is here
1959 @end smallexample
1960
1961 @item
1962 Merging a common symbol with a previous smaller common symbol. This is
1963 the same as the previous case, except that the symbols are
1964 encountered in a different order.
1965 @smallexample
1966 @var{file}(@var{section}): warning: common of `@var{symbol}'
1967 overriding smaller common
1968 @var{file}(@var{section}): warning: smaller common is here
1969 @end smallexample
1970 @end enumerate
1971
1972 @kindex --warn-constructors
1973 @item --warn-constructors
1974 Warn if any global constructors are used. This is only useful for a few
1975 object file formats. For formats like COFF or ELF, the linker can not
1976 detect the use of global constructors.
1977
1978 @kindex --warn-multiple-gp
1979 @item --warn-multiple-gp
1980 Warn if multiple global pointer values are required in the output file.
1981 This is only meaningful for certain processors, such as the Alpha.
1982 Specifically, some processors put large-valued constants in a special
1983 section. A special register (the global pointer) points into the middle
1984 of this section, so that constants can be loaded efficiently via a
1985 base-register relative addressing mode. Since the offset in
1986 base-register relative mode is fixed and relatively small (e.g., 16
1987 bits), this limits the maximum size of the constant pool. Thus, in
1988 large programs, it is often necessary to use multiple global pointer
1989 values in order to be able to address all possible constants. This
1990 option causes a warning to be issued whenever this case occurs.
1991
1992 @kindex --warn-once
1993 @cindex warnings, on undefined symbols
1994 @cindex undefined symbols, warnings on
1995 @item --warn-once
1996 Only warn once for each undefined symbol, rather than once per module
1997 which refers to it.
1998
1999 @kindex --warn-section-align
2000 @cindex warnings, on section alignment
2001 @cindex section alignment, warnings on
2002 @item --warn-section-align
2003 Warn if the address of an output section is changed because of
2004 alignment. Typically, the alignment will be set by an input section.
2005 The address will only be changed if it not explicitly specified; that
2006 is, if the @code{SECTIONS} command does not specify a start address for
2007 the section (@pxref{SECTIONS}).
2008
2009 @kindex --warn-shared-textrel
2010 @item --warn-shared-textrel
2011 Warn if the linker adds a DT_TEXTREL to a shared object.
2012
2013 @kindex --warn-alternate-em
2014 @item --warn-alternate-em
2015 Warn if an object has alternate ELF machine code.
2016
2017 @kindex --warn-unresolved-symbols
2018 @item --warn-unresolved-symbols
2019 If the linker is going to report an unresolved symbol (see the option
2020 @option{--unresolved-symbols}) it will normally generate an error.
2021 This option makes it generate a warning instead.
2022
2023 @kindex --error-unresolved-symbols
2024 @item --error-unresolved-symbols
2025 This restores the linker's default behaviour of generating errors when
2026 it is reporting unresolved symbols.
2027
2028 @kindex --whole-archive
2029 @cindex including an entire archive
2030 @item --whole-archive
2031 For each archive mentioned on the command line after the
2032 @option{--whole-archive} option, include every object file in the archive
2033 in the link, rather than searching the archive for the required object
2034 files. This is normally used to turn an archive file into a shared
2035 library, forcing every object to be included in the resulting shared
2036 library. This option may be used more than once.
2037
2038 Two notes when using this option from gcc: First, gcc doesn't know
2039 about this option, so you have to use @option{-Wl,-whole-archive}.
2040 Second, don't forget to use @option{-Wl,-no-whole-archive} after your
2041 list of archives, because gcc will add its own list of archives to
2042 your link and you may not want this flag to affect those as well.
2043
2044 @kindex --wrap=@var{symbol}
2045 @item --wrap=@var{symbol}
2046 Use a wrapper function for @var{symbol}. Any undefined reference to
2047 @var{symbol} will be resolved to @code{__wrap_@var{symbol}}. Any
2048 undefined reference to @code{__real_@var{symbol}} will be resolved to
2049 @var{symbol}.
2050
2051 This can be used to provide a wrapper for a system function. The
2052 wrapper function should be called @code{__wrap_@var{symbol}}. If it
2053 wishes to call the system function, it should call
2054 @code{__real_@var{symbol}}.
2055
2056 Here is a trivial example:
2057
2058 @smallexample
2059 void *
2060 __wrap_malloc (size_t c)
2061 @{
2062 printf ("malloc called with %zu\n", c);
2063 return __real_malloc (c);
2064 @}
2065 @end smallexample
2066
2067 If you link other code with this file using @option{--wrap malloc}, then
2068 all calls to @code{malloc} will call the function @code{__wrap_malloc}
2069 instead. The call to @code{__real_malloc} in @code{__wrap_malloc} will
2070 call the real @code{malloc} function.
2071
2072 You may wish to provide a @code{__real_malloc} function as well, so that
2073 links without the @option{--wrap} option will succeed. If you do this,
2074 you should not put the definition of @code{__real_malloc} in the same
2075 file as @code{__wrap_malloc}; if you do, the assembler may resolve the
2076 call before the linker has a chance to wrap it to @code{malloc}.
2077
2078 @kindex --eh-frame-hdr
2079 @item --eh-frame-hdr
2080 Request creation of @code{.eh_frame_hdr} section and ELF
2081 @code{PT_GNU_EH_FRAME} segment header.
2082
2083 @kindex --ld-generated-unwind-info
2084 @item --no-ld-generated-unwind-info
2085 Request creation of @code{.eh_frame} unwind info for linker
2086 generated code sections like PLT. This option is on by default
2087 if linker generated unwind info is supported.
2088
2089 @kindex --enable-new-dtags
2090 @kindex --disable-new-dtags
2091 @item --enable-new-dtags
2092 @itemx --disable-new-dtags
2093 This linker can create the new dynamic tags in ELF. But the older ELF
2094 systems may not understand them. If you specify
2095 @option{--enable-new-dtags}, the dynamic tags will be created as needed.
2096 If you specify @option{--disable-new-dtags}, no new dynamic tags will be
2097 created. By default, the new dynamic tags are not created. Note that
2098 those options are only available for ELF systems.
2099
2100 @kindex --hash-size=@var{number}
2101 @item --hash-size=@var{number}
2102 Set the default size of the linker's hash tables to a prime number
2103 close to @var{number}. Increasing this value can reduce the length of
2104 time it takes the linker to perform its tasks, at the expense of
2105 increasing the linker's memory requirements. Similarly reducing this
2106 value can reduce the memory requirements at the expense of speed.
2107
2108 @kindex --hash-style=@var{style}
2109 @item --hash-style=@var{style}
2110 Set the type of linker's hash table(s). @var{style} can be either
2111 @code{sysv} for classic ELF @code{.hash} section, @code{gnu} for
2112 new style GNU @code{.gnu.hash} section or @code{both} for both
2113 the classic ELF @code{.hash} and new style GNU @code{.gnu.hash}
2114 hash tables. The default is @code{sysv}.
2115
2116 @kindex --reduce-memory-overheads
2117 @item --reduce-memory-overheads
2118 This option reduces memory requirements at ld runtime, at the expense of
2119 linking speed. This was introduced to select the old O(n^2) algorithm
2120 for link map file generation, rather than the new O(n) algorithm which uses
2121 about 40% more memory for symbol storage.
2122
2123 Another effect of the switch is to set the default hash table size to
2124 1021, which again saves memory at the cost of lengthening the linker's
2125 run time. This is not done however if the @option{--hash-size} switch
2126 has been used.
2127
2128 The @option{--reduce-memory-overheads} switch may be also be used to
2129 enable other tradeoffs in future versions of the linker.
2130
2131 @kindex --build-id
2132 @kindex --build-id=@var{style}
2133 @item --build-id
2134 @itemx --build-id=@var{style}
2135 Request creation of @code{.note.gnu.build-id} ELF note section.
2136 The contents of the note are unique bits identifying this linked
2137 file. @var{style} can be @code{uuid} to use 128 random bits,
2138 @code{sha1} to use a 160-bit @sc{SHA1} hash on the normative
2139 parts of the output contents, @code{md5} to use a 128-bit
2140 @sc{MD5} hash on the normative parts of the output contents, or
2141 @code{0x@var{hexstring}} to use a chosen bit string specified as
2142 an even number of hexadecimal digits (@code{-} and @code{:}
2143 characters between digit pairs are ignored). If @var{style} is
2144 omitted, @code{sha1} is used.
2145
2146 The @code{md5} and @code{sha1} styles produces an identifier
2147 that is always the same in an identical output file, but will be
2148 unique among all nonidentical output files. It is not intended
2149 to be compared as a checksum for the file's contents. A linked
2150 file may be changed later by other tools, but the build ID bit
2151 string identifying the original linked file does not change.
2152
2153 Passing @code{none} for @var{style} disables the setting from any
2154 @code{--build-id} options earlier on the command line.
2155 @end table
2156
2157 @c man end
2158
2159 @subsection Options Specific to i386 PE Targets
2160
2161 @c man begin OPTIONS
2162
2163 The i386 PE linker supports the @option{-shared} option, which causes
2164 the output to be a dynamically linked library (DLL) instead of a
2165 normal executable. You should name the output @code{*.dll} when you
2166 use this option. In addition, the linker fully supports the standard
2167 @code{*.def} files, which may be specified on the linker command line
2168 like an object file (in fact, it should precede archives it exports
2169 symbols from, to ensure that they get linked in, just like a normal
2170 object file).
2171
2172 In addition to the options common to all targets, the i386 PE linker
2173 support additional command line options that are specific to the i386
2174 PE target. Options that take values may be separated from their
2175 values by either a space or an equals sign.
2176
2177 @table @gcctabopt
2178
2179 @kindex --add-stdcall-alias
2180 @item --add-stdcall-alias
2181 If given, symbols with a stdcall suffix (@@@var{nn}) will be exported
2182 as-is and also with the suffix stripped.
2183 [This option is specific to the i386 PE targeted port of the linker]
2184
2185 @kindex --base-file
2186 @item --base-file @var{file}
2187 Use @var{file} as the name of a file in which to save the base
2188 addresses of all the relocations needed for generating DLLs with
2189 @file{dlltool}.
2190 [This is an i386 PE specific option]
2191
2192 @kindex --dll
2193 @item --dll
2194 Create a DLL instead of a regular executable. You may also use
2195 @option{-shared} or specify a @code{LIBRARY} in a given @code{.def}
2196 file.
2197 [This option is specific to the i386 PE targeted port of the linker]
2198
2199 @kindex --enable-long-section-names
2200 @kindex --disable-long-section-names
2201 @item --enable-long-section-names
2202 @itemx --disable-long-section-names
2203 The PE variants of the Coff object format add an extension that permits
2204 the use of section names longer than eight characters, the normal limit
2205 for Coff. By default, these names are only allowed in object files, as
2206 fully-linked executable images do not carry the Coff string table required
2207 to support the longer names. As a GNU extension, it is possible to
2208 allow their use in executable images as well, or to (probably pointlessly!)
2209 disallow it in object files, by using these two options. Executable images
2210 generated with these long section names are slightly non-standard, carrying
2211 as they do a string table, and may generate confusing output when examined
2212 with non-GNU PE-aware tools, such as file viewers and dumpers. However,
2213 GDB relies on the use of PE long section names to find Dwarf-2 debug
2214 information sections in an executable image at runtime, and so if neither
2215 option is specified on the command-line, @command{ld} will enable long
2216 section names, overriding the default and technically correct behaviour,
2217 when it finds the presence of debug information while linking an executable
2218 image and not stripping symbols.
2219 [This option is valid for all PE targeted ports of the linker]
2220
2221 @kindex --enable-stdcall-fixup
2222 @kindex --disable-stdcall-fixup
2223 @item --enable-stdcall-fixup
2224 @itemx --disable-stdcall-fixup
2225 If the link finds a symbol that it cannot resolve, it will attempt to
2226 do ``fuzzy linking'' by looking for another defined symbol that differs
2227 only in the format of the symbol name (cdecl vs stdcall) and will
2228 resolve that symbol by linking to the match. For example, the
2229 undefined symbol @code{_foo} might be linked to the function
2230 @code{_foo@@12}, or the undefined symbol @code{_bar@@16} might be linked
2231 to the function @code{_bar}. When the linker does this, it prints a
2232 warning, since it normally should have failed to link, but sometimes
2233 import libraries generated from third-party dlls may need this feature
2234 to be usable. If you specify @option{--enable-stdcall-fixup}, this
2235 feature is fully enabled and warnings are not printed. If you specify
2236 @option{--disable-stdcall-fixup}, this feature is disabled and such
2237 mismatches are considered to be errors.
2238 [This option is specific to the i386 PE targeted port of the linker]
2239
2240 @kindex --leading-underscore
2241 @kindex --no-leading-underscore
2242 @item --leading-underscore
2243 @itemx --no-leading-underscore
2244 For most targets default symbol-prefix is an underscore and is defined
2245 in target's description. By this option it is possible to
2246 disable/enable the default underscore symbol-prefix.
2247
2248 @cindex DLLs, creating
2249 @kindex --export-all-symbols
2250 @item --export-all-symbols
2251 If given, all global symbols in the objects used to build a DLL will
2252 be exported by the DLL. Note that this is the default if there
2253 otherwise wouldn't be any exported symbols. When symbols are
2254 explicitly exported via DEF files or implicitly exported via function
2255 attributes, the default is to not export anything else unless this
2256 option is given. Note that the symbols @code{DllMain@@12},
2257 @code{DllEntryPoint@@0}, @code{DllMainCRTStartup@@12}, and
2258 @code{impure_ptr} will not be automatically
2259 exported. Also, symbols imported from other DLLs will not be
2260 re-exported, nor will symbols specifying the DLL's internal layout
2261 such as those beginning with @code{_head_} or ending with
2262 @code{_iname}. In addition, no symbols from @code{libgcc},
2263 @code{libstd++}, @code{libmingw32}, or @code{crtX.o} will be exported.
2264 Symbols whose names begin with @code{__rtti_} or @code{__builtin_} will
2265 not be exported, to help with C++ DLLs. Finally, there is an
2266 extensive list of cygwin-private symbols that are not exported
2267 (obviously, this applies on when building DLLs for cygwin targets).
2268 These cygwin-excludes are: @code{_cygwin_dll_entry@@12},
2269 @code{_cygwin_crt0_common@@8}, @code{_cygwin_noncygwin_dll_entry@@12},
2270 @code{_fmode}, @code{_impure_ptr}, @code{cygwin_attach_dll},
2271 @code{cygwin_premain0}, @code{cygwin_premain1}, @code{cygwin_premain2},
2272 @code{cygwin_premain3}, and @code{environ}.
2273 [This option is specific to the i386 PE targeted port of the linker]
2274
2275 @kindex --exclude-symbols
2276 @item --exclude-symbols @var{symbol},@var{symbol},...
2277 Specifies a list of symbols which should not be automatically
2278 exported. The symbol names may be delimited by commas or colons.
2279 [This option is specific to the i386 PE targeted port of the linker]
2280
2281 @kindex --exclude-all-symbols
2282 @item --exclude-all-symbols
2283 Specifies no symbols should be automatically exported.
2284 [This option is specific to the i386 PE targeted port of the linker]
2285
2286 @kindex --file-alignment
2287 @item --file-alignment
2288 Specify the file alignment. Sections in the file will always begin at
2289 file offsets which are multiples of this number. This defaults to
2290 512.
2291 [This option is specific to the i386 PE targeted port of the linker]
2292
2293 @cindex heap size
2294 @kindex --heap
2295 @item --heap @var{reserve}
2296 @itemx --heap @var{reserve},@var{commit}
2297 Specify the number of bytes of memory to reserve (and optionally commit)
2298 to be used as heap for this program. The default is 1Mb reserved, 4K
2299 committed.
2300 [This option is specific to the i386 PE targeted port of the linker]
2301
2302 @cindex image base
2303 @kindex --image-base
2304 @item --image-base @var{value}
2305 Use @var{value} as the base address of your program or dll. This is
2306 the lowest memory location that will be used when your program or dll
2307 is loaded. To reduce the need to relocate and improve performance of
2308 your dlls, each should have a unique base address and not overlap any
2309 other dlls. The default is 0x400000 for executables, and 0x10000000
2310 for dlls.
2311 [This option is specific to the i386 PE targeted port of the linker]
2312
2313 @kindex --kill-at
2314 @item --kill-at
2315 If given, the stdcall suffixes (@@@var{nn}) will be stripped from
2316 symbols before they are exported.
2317 [This option is specific to the i386 PE targeted port of the linker]
2318
2319 @kindex --large-address-aware
2320 @item --large-address-aware
2321 If given, the appropriate bit in the ``Characteristics'' field of the COFF
2322 header is set to indicate that this executable supports virtual addresses
2323 greater than 2 gigabytes. This should be used in conjunction with the /3GB
2324 or /USERVA=@var{value} megabytes switch in the ``[operating systems]''
2325 section of the BOOT.INI. Otherwise, this bit has no effect.
2326 [This option is specific to PE targeted ports of the linker]
2327
2328 @kindex --major-image-version
2329 @item --major-image-version @var{value}
2330 Sets the major number of the ``image version''. Defaults to 1.
2331 [This option is specific to the i386 PE targeted port of the linker]
2332
2333 @kindex --major-os-version
2334 @item --major-os-version @var{value}
2335 Sets the major number of the ``os version''. Defaults to 4.
2336 [This option is specific to the i386 PE targeted port of the linker]
2337
2338 @kindex --major-subsystem-version
2339 @item --major-subsystem-version @var{value}
2340 Sets the major number of the ``subsystem version''. Defaults to 4.
2341 [This option is specific to the i386 PE targeted port of the linker]
2342
2343 @kindex --minor-image-version
2344 @item --minor-image-version @var{value}
2345 Sets the minor number of the ``image version''. Defaults to 0.
2346 [This option is specific to the i386 PE targeted port of the linker]
2347
2348 @kindex --minor-os-version
2349 @item --minor-os-version @var{value}
2350 Sets the minor number of the ``os version''. Defaults to 0.
2351 [This option is specific to the i386 PE targeted port of the linker]
2352
2353 @kindex --minor-subsystem-version
2354 @item --minor-subsystem-version @var{value}
2355 Sets the minor number of the ``subsystem version''. Defaults to 0.
2356 [This option is specific to the i386 PE targeted port of the linker]
2357
2358 @cindex DEF files, creating
2359 @cindex DLLs, creating
2360 @kindex --output-def
2361 @item --output-def @var{file}
2362 The linker will create the file @var{file} which will contain a DEF
2363 file corresponding to the DLL the linker is generating. This DEF file
2364 (which should be called @code{*.def}) may be used to create an import
2365 library with @code{dlltool} or may be used as a reference to
2366 automatically or implicitly exported symbols.
2367 [This option is specific to the i386 PE targeted port of the linker]
2368
2369 @cindex DLLs, creating
2370 @kindex --out-implib
2371 @item --out-implib @var{file}
2372 The linker will create the file @var{file} which will contain an
2373 import lib corresponding to the DLL the linker is generating. This
2374 import lib (which should be called @code{*.dll.a} or @code{*.a}
2375 may be used to link clients against the generated DLL; this behaviour
2376 makes it possible to skip a separate @code{dlltool} import library
2377 creation step.
2378 [This option is specific to the i386 PE targeted port of the linker]
2379
2380 @kindex --enable-auto-image-base
2381 @item --enable-auto-image-base
2382 Automatically choose the image base for DLLs, unless one is specified
2383 using the @code{--image-base} argument. By using a hash generated
2384 from the dllname to create unique image bases for each DLL, in-memory
2385 collisions and relocations which can delay program execution are
2386 avoided.
2387 [This option is specific to the i386 PE targeted port of the linker]
2388
2389 @kindex --disable-auto-image-base
2390 @item --disable-auto-image-base
2391 Do not automatically generate a unique image base. If there is no
2392 user-specified image base (@code{--image-base}) then use the platform
2393 default.
2394 [This option is specific to the i386 PE targeted port of the linker]
2395
2396 @cindex DLLs, linking to
2397 @kindex --dll-search-prefix
2398 @item --dll-search-prefix @var{string}
2399 When linking dynamically to a dll without an import library,
2400 search for @code{<string><basename>.dll} in preference to
2401 @code{lib<basename>.dll}. This behaviour allows easy distinction
2402 between DLLs built for the various "subplatforms": native, cygwin,
2403 uwin, pw, etc. For instance, cygwin DLLs typically use
2404 @code{--dll-search-prefix=cyg}.
2405 [This option is specific to the i386 PE targeted port of the linker]
2406
2407 @kindex --enable-auto-import
2408 @item --enable-auto-import
2409 Do sophisticated linking of @code{_symbol} to @code{__imp__symbol} for
2410 DATA imports from DLLs, and create the necessary thunking symbols when
2411 building the import libraries with those DATA exports. Note: Use of the
2412 'auto-import' extension will cause the text section of the image file
2413 to be made writable. This does not conform to the PE-COFF format
2414 specification published by Microsoft.
2415
2416 Note - use of the 'auto-import' extension will also cause read only
2417 data which would normally be placed into the .rdata section to be
2418 placed into the .data section instead. This is in order to work
2419 around a problem with consts that is described here:
2420 http://www.cygwin.com/ml/cygwin/2004-09/msg01101.html
2421
2422 Using 'auto-import' generally will 'just work' -- but sometimes you may
2423 see this message:
2424
2425 "variable '<var>' can't be auto-imported. Please read the
2426 documentation for ld's @code{--enable-auto-import} for details."
2427
2428 This message occurs when some (sub)expression accesses an address
2429 ultimately given by the sum of two constants (Win32 import tables only
2430 allow one). Instances where this may occur include accesses to member
2431 fields of struct variables imported from a DLL, as well as using a
2432 constant index into an array variable imported from a DLL. Any
2433 multiword variable (arrays, structs, long long, etc) may trigger
2434 this error condition. However, regardless of the exact data type
2435 of the offending exported variable, ld will always detect it, issue
2436 the warning, and exit.
2437
2438 There are several ways to address this difficulty, regardless of the
2439 data type of the exported variable:
2440
2441 One way is to use --enable-runtime-pseudo-reloc switch. This leaves the task
2442 of adjusting references in your client code for runtime environment, so
2443 this method works only when runtime environment supports this feature.
2444
2445 A second solution is to force one of the 'constants' to be a variable --
2446 that is, unknown and un-optimizable at compile time. For arrays,
2447 there are two possibilities: a) make the indexee (the array's address)
2448 a variable, or b) make the 'constant' index a variable. Thus:
2449
2450 @example
2451 extern type extern_array[];
2452 extern_array[1] -->
2453 @{ volatile type *t=extern_array; t[1] @}
2454 @end example
2455
2456 or
2457
2458 @example
2459 extern type extern_array[];
2460 extern_array[1] -->
2461 @{ volatile int t=1; extern_array[t] @}
2462 @end example
2463
2464 For structs (and most other multiword data types) the only option
2465 is to make the struct itself (or the long long, or the ...) variable:
2466
2467 @example
2468 extern struct s extern_struct;
2469 extern_struct.field -->
2470 @{ volatile struct s *t=&extern_struct; t->field @}
2471 @end example
2472
2473 or
2474
2475 @example
2476 extern long long extern_ll;
2477 extern_ll -->
2478 @{ volatile long long * local_ll=&extern_ll; *local_ll @}
2479 @end example
2480
2481 A third method of dealing with this difficulty is to abandon
2482 'auto-import' for the offending symbol and mark it with
2483 @code{__declspec(dllimport)}. However, in practise that
2484 requires using compile-time #defines to indicate whether you are
2485 building a DLL, building client code that will link to the DLL, or
2486 merely building/linking to a static library. In making the choice
2487 between the various methods of resolving the 'direct address with
2488 constant offset' problem, you should consider typical real-world usage:
2489
2490 Original:
2491 @example
2492 --foo.h
2493 extern int arr[];
2494 --foo.c
2495 #include "foo.h"
2496 void main(int argc, char **argv)@{
2497 printf("%d\n",arr[1]);
2498 @}
2499 @end example
2500
2501 Solution 1:
2502 @example
2503 --foo.h
2504 extern int arr[];
2505 --foo.c
2506 #include "foo.h"
2507 void main(int argc, char **argv)@{
2508 /* This workaround is for win32 and cygwin; do not "optimize" */
2509 volatile int *parr = arr;
2510 printf("%d\n",parr[1]);
2511 @}
2512 @end example
2513
2514 Solution 2:
2515 @example
2516 --foo.h
2517 /* Note: auto-export is assumed (no __declspec(dllexport)) */
2518 #if (defined(_WIN32) || defined(__CYGWIN__)) && \
2519 !(defined(FOO_BUILD_DLL) || defined(FOO_STATIC))
2520 #define FOO_IMPORT __declspec(dllimport)
2521 #else
2522 #define FOO_IMPORT
2523 #endif
2524 extern FOO_IMPORT int arr[];
2525 --foo.c
2526 #include "foo.h"
2527 void main(int argc, char **argv)@{
2528 printf("%d\n",arr[1]);
2529 @}
2530 @end example
2531
2532 A fourth way to avoid this problem is to re-code your
2533 library to use a functional interface rather than a data interface
2534 for the offending variables (e.g. set_foo() and get_foo() accessor
2535 functions).
2536 [This option is specific to the i386 PE targeted port of the linker]
2537
2538 @kindex --disable-auto-import
2539 @item --disable-auto-import
2540 Do not attempt to do sophisticated linking of @code{_symbol} to
2541 @code{__imp__symbol} for DATA imports from DLLs.
2542 [This option is specific to the i386 PE targeted port of the linker]
2543
2544 @kindex --enable-runtime-pseudo-reloc
2545 @item --enable-runtime-pseudo-reloc
2546 If your code contains expressions described in --enable-auto-import section,
2547 that is, DATA imports from DLL with non-zero offset, this switch will create
2548 a vector of 'runtime pseudo relocations' which can be used by runtime
2549 environment to adjust references to such data in your client code.
2550 [This option is specific to the i386 PE targeted port of the linker]
2551
2552 @kindex --disable-runtime-pseudo-reloc
2553 @item --disable-runtime-pseudo-reloc
2554 Do not create pseudo relocations for non-zero offset DATA imports from
2555 DLLs. This is the default.
2556 [This option is specific to the i386 PE targeted port of the linker]
2557
2558 @kindex --enable-extra-pe-debug
2559 @item --enable-extra-pe-debug
2560 Show additional debug info related to auto-import symbol thunking.
2561 [This option is specific to the i386 PE targeted port of the linker]
2562
2563 @kindex --section-alignment
2564 @item --section-alignment
2565 Sets the section alignment. Sections in memory will always begin at
2566 addresses which are a multiple of this number. Defaults to 0x1000.
2567 [This option is specific to the i386 PE targeted port of the linker]
2568
2569 @cindex stack size
2570 @kindex --stack
2571 @item --stack @var{reserve}
2572 @itemx --stack @var{reserve},@var{commit}
2573 Specify the number of bytes of memory to reserve (and optionally commit)
2574 to be used as stack for this program. The default is 2Mb reserved, 4K
2575 committed.
2576 [This option is specific to the i386 PE targeted port of the linker]
2577
2578 @kindex --subsystem
2579 @item --subsystem @var{which}
2580 @itemx --subsystem @var{which}:@var{major}
2581 @itemx --subsystem @var{which}:@var{major}.@var{minor}
2582 Specifies the subsystem under which your program will execute. The
2583 legal values for @var{which} are @code{native}, @code{windows},
2584 @code{console}, @code{posix}, and @code{xbox}. You may optionally set
2585 the subsystem version also. Numeric values are also accepted for
2586 @var{which}.
2587 [This option is specific to the i386 PE targeted port of the linker]
2588
2589 The following options set flags in the @code{DllCharacteristics} field
2590 of the PE file header:
2591 [These options are specific to PE targeted ports of the linker]
2592
2593 @kindex --dynamicbase
2594 @item --dynamicbase
2595 The image base address may be relocated using address space layout
2596 randomization (ASLR). This feature was introduced with MS Windows
2597 Vista for i386 PE targets.
2598
2599 @kindex --forceinteg
2600 @item --forceinteg
2601 Code integrity checks are enforced.
2602
2603 @kindex --nxcompat
2604 @item --nxcompat
2605 The image is compatible with the Data Execution Prevention.
2606 This feature was introduced with MS Windows XP SP2 for i386 PE targets.
2607
2608 @kindex --no-isolation
2609 @item --no-isolation
2610 Although the image understands isolation, do not isolate the image.
2611
2612 @kindex --no-seh
2613 @item --no-seh
2614 The image does not use SEH. No SE handler may be called from
2615 this image.
2616
2617 @kindex --no-bind
2618 @item --no-bind
2619 Do not bind this image.
2620
2621 @kindex --wdmdriver
2622 @item --wdmdriver
2623 The driver uses the MS Windows Driver Model.
2624
2625 @kindex --tsaware
2626 @item --tsaware
2627 The image is Terminal Server aware.
2628
2629 @end table
2630
2631 @c man end
2632
2633 @ifset C6X
2634 @subsection Options specific to C6X uClinux targets
2635
2636 @c man begin OPTIONS
2637
2638 The C6X uClinux target uses a binary format called DSBT to support shared
2639 libraries. Each shared library in the system needs to have a unique index;
2640 all executables use an index of 0.
2641
2642 @table @gcctabopt
2643
2644 @kindex --dsbt-size
2645 @item --dsbt-size @var{size}
2646 This option sets the number of entires in the DSBT of the current executable
2647 or shared library to @var{size}. The default is to create a table with 64
2648 entries.
2649
2650 @kindex --dsbt-index
2651 @item --dsbt-index @var{index}
2652 This option sets the DSBT index of the current executable or shared library
2653 to @var{index}. The default is 0, which is appropriate for generating
2654 executables. If a shared library is generated with a DSBT index of 0, the
2655 @code{R_C6000_DSBT_INDEX} relocs are copied into the output file.
2656
2657 @kindex --no-merge-exidx-entries
2658 The @samp{--no-merge-exidx-entries} switch disables the merging of adjacent
2659 exidx entries in frame unwind info.
2660
2661 @end table
2662
2663 @c man end
2664 @end ifset
2665
2666 @ifset M68HC11
2667 @subsection Options specific to Motorola 68HC11 and 68HC12 targets
2668
2669 @c man begin OPTIONS
2670
2671 The 68HC11 and 68HC12 linkers support specific options to control the
2672 memory bank switching mapping and trampoline code generation.
2673
2674 @table @gcctabopt
2675
2676 @kindex --no-trampoline
2677 @item --no-trampoline
2678 This option disables the generation of trampoline. By default a trampoline
2679 is generated for each far function which is called using a @code{jsr}
2680 instruction (this happens when a pointer to a far function is taken).
2681
2682 @kindex --bank-window
2683 @item --bank-window @var{name}
2684 This option indicates to the linker the name of the memory region in
2685 the @samp{MEMORY} specification that describes the memory bank window.
2686 The definition of such region is then used by the linker to compute
2687 paging and addresses within the memory window.
2688
2689 @end table
2690
2691 @c man end
2692 @end ifset
2693
2694 @ifset M68K
2695 @subsection Options specific to Motorola 68K target
2696
2697 @c man begin OPTIONS
2698
2699 The following options are supported to control handling of GOT generation
2700 when linking for 68K targets.
2701
2702 @table @gcctabopt
2703
2704 @kindex --got
2705 @item --got=@var{type}
2706 This option tells the linker which GOT generation scheme to use.
2707 @var{type} should be one of @samp{single}, @samp{negative},
2708 @samp{multigot} or @samp{target}. For more information refer to the
2709 Info entry for @file{ld}.
2710
2711 @end table
2712
2713 @c man end
2714 @end ifset
2715
2716 @ifset UsesEnvVars
2717 @node Environment
2718 @section Environment Variables
2719
2720 @c man begin ENVIRONMENT
2721
2722 You can change the behaviour of @command{ld} with the environment variables
2723 @ifclear SingleFormat
2724 @code{GNUTARGET},
2725 @end ifclear
2726 @code{LDEMULATION} and @code{COLLECT_NO_DEMANGLE}.
2727
2728 @ifclear SingleFormat
2729 @kindex GNUTARGET
2730 @cindex default input format
2731 @code{GNUTARGET} determines the input-file object format if you don't
2732 use @samp{-b} (or its synonym @samp{--format}). Its value should be one
2733 of the BFD names for an input format (@pxref{BFD}). If there is no
2734 @code{GNUTARGET} in the environment, @command{ld} uses the natural format
2735 of the target. If @code{GNUTARGET} is set to @code{default} then BFD
2736 attempts to discover the input format by examining binary input files;
2737 this method often succeeds, but there are potential ambiguities, since
2738 there is no method of ensuring that the magic number used to specify
2739 object-file formats is unique. However, the configuration procedure for
2740 BFD on each system places the conventional format for that system first
2741 in the search-list, so ambiguities are resolved in favor of convention.
2742 @end ifclear
2743
2744 @kindex LDEMULATION
2745 @cindex default emulation
2746 @cindex emulation, default
2747 @code{LDEMULATION} determines the default emulation if you don't use the
2748 @samp{-m} option. The emulation can affect various aspects of linker
2749 behaviour, particularly the default linker script. You can list the
2750 available emulations with the @samp{--verbose} or @samp{-V} options. If
2751 the @samp{-m} option is not used, and the @code{LDEMULATION} environment
2752 variable is not defined, the default emulation depends upon how the
2753 linker was configured.
2754
2755 @kindex COLLECT_NO_DEMANGLE
2756 @cindex demangling, default
2757 Normally, the linker will default to demangling symbols. However, if
2758 @code{COLLECT_NO_DEMANGLE} is set in the environment, then it will
2759 default to not demangling symbols. This environment variable is used in
2760 a similar fashion by the @code{gcc} linker wrapper program. The default
2761 may be overridden by the @samp{--demangle} and @samp{--no-demangle}
2762 options.
2763
2764 @c man end
2765 @end ifset
2766
2767 @node Scripts
2768 @chapter Linker Scripts
2769
2770 @cindex scripts
2771 @cindex linker scripts
2772 @cindex command files
2773 Every link is controlled by a @dfn{linker script}. This script is
2774 written in the linker command language.
2775
2776 The main purpose of the linker script is to describe how the sections in
2777 the input files should be mapped into the output file, and to control
2778 the memory layout of the output file. Most linker scripts do nothing
2779 more than this. However, when necessary, the linker script can also
2780 direct the linker to perform many other operations, using the commands
2781 described below.
2782
2783 The linker always uses a linker script. If you do not supply one
2784 yourself, the linker will use a default script that is compiled into the
2785 linker executable. You can use the @samp{--verbose} command line option
2786 to display the default linker script. Certain command line options,
2787 such as @samp{-r} or @samp{-N}, will affect the default linker script.
2788
2789 You may supply your own linker script by using the @samp{-T} command
2790 line option. When you do this, your linker script will replace the
2791 default linker script.
2792
2793 You may also use linker scripts implicitly by naming them as input files
2794 to the linker, as though they were files to be linked. @xref{Implicit
2795 Linker Scripts}.
2796
2797 @menu
2798 * Basic Script Concepts:: Basic Linker Script Concepts
2799 * Script Format:: Linker Script Format
2800 * Simple Example:: Simple Linker Script Example
2801 * Simple Commands:: Simple Linker Script Commands
2802 * Assignments:: Assigning Values to Symbols
2803 * SECTIONS:: SECTIONS Command
2804 * MEMORY:: MEMORY Command
2805 * PHDRS:: PHDRS Command
2806 * VERSION:: VERSION Command
2807 * Expressions:: Expressions in Linker Scripts
2808 * Implicit Linker Scripts:: Implicit Linker Scripts
2809 @end menu
2810
2811 @node Basic Script Concepts
2812 @section Basic Linker Script Concepts
2813 @cindex linker script concepts
2814 We need to define some basic concepts and vocabulary in order to
2815 describe the linker script language.
2816
2817 The linker combines input files into a single output file. The output
2818 file and each input file are in a special data format known as an
2819 @dfn{object file format}. Each file is called an @dfn{object file}.
2820 The output file is often called an @dfn{executable}, but for our
2821 purposes we will also call it an object file. Each object file has,
2822 among other things, a list of @dfn{sections}. We sometimes refer to a
2823 section in an input file as an @dfn{input section}; similarly, a section
2824 in the output file is an @dfn{output section}.
2825
2826 Each section in an object file has a name and a size. Most sections
2827 also have an associated block of data, known as the @dfn{section
2828 contents}. A section may be marked as @dfn{loadable}, which mean that
2829 the contents should be loaded into memory when the output file is run.
2830 A section with no contents may be @dfn{allocatable}, which means that an
2831 area in memory should be set aside, but nothing in particular should be
2832 loaded there (in some cases this memory must be zeroed out). A section
2833 which is neither loadable nor allocatable typically contains some sort
2834 of debugging information.
2835
2836 Every loadable or allocatable output section has two addresses. The
2837 first is the @dfn{VMA}, or virtual memory address. This is the address
2838 the section will have when the output file is run. The second is the
2839 @dfn{LMA}, or load memory address. This is the address at which the
2840 section will be loaded. In most cases the two addresses will be the
2841 same. An example of when they might be different is when a data section
2842 is loaded into ROM, and then copied into RAM when the program starts up
2843 (this technique is often used to initialize global variables in a ROM
2844 based system). In this case the ROM address would be the LMA, and the
2845 RAM address would be the VMA.
2846
2847 You can see the sections in an object file by using the @code{objdump}
2848 program with the @samp{-h} option.
2849
2850 Every object file also has a list of @dfn{symbols}, known as the
2851 @dfn{symbol table}. A symbol may be defined or undefined. Each symbol
2852 has a name, and each defined symbol has an address, among other
2853 information. If you compile a C or C++ program into an object file, you
2854 will get a defined symbol for every defined function and global or
2855 static variable. Every undefined function or global variable which is
2856 referenced in the input file will become an undefined symbol.
2857
2858 You can see the symbols in an object file by using the @code{nm}
2859 program, or by using the @code{objdump} program with the @samp{-t}
2860 option.
2861
2862 @node Script Format
2863 @section Linker Script Format
2864 @cindex linker script format
2865 Linker scripts are text files.
2866
2867 You write a linker script as a series of commands. Each command is
2868 either a keyword, possibly followed by arguments, or an assignment to a
2869 symbol. You may separate commands using semicolons. Whitespace is
2870 generally ignored.
2871
2872 Strings such as file or format names can normally be entered directly.
2873 If the file name contains a character such as a comma which would
2874 otherwise serve to separate file names, you may put the file name in
2875 double quotes. There is no way to use a double quote character in a
2876 file name.
2877
2878 You may include comments in linker scripts just as in C, delimited by
2879 @samp{/*} and @samp{*/}. As in C, comments are syntactically equivalent
2880 to whitespace.
2881
2882 @node Simple Example
2883 @section Simple Linker Script Example
2884 @cindex linker script example
2885 @cindex example of linker script
2886 Many linker scripts are fairly simple.
2887
2888 The simplest possible linker script has just one command:
2889 @samp{SECTIONS}. You use the @samp{SECTIONS} command to describe the
2890 memory layout of the output file.
2891
2892 The @samp{SECTIONS} command is a powerful command. Here we will
2893 describe a simple use of it. Let's assume your program consists only of
2894 code, initialized data, and uninitialized data. These will be in the
2895 @samp{.text}, @samp{.data}, and @samp{.bss} sections, respectively.
2896 Let's assume further that these are the only sections which appear in
2897 your input files.
2898
2899 For this example, let's say that the code should be loaded at address
2900 0x10000, and that the data should start at address 0x8000000. Here is a
2901 linker script which will do that:
2902 @smallexample
2903 SECTIONS
2904 @{
2905 . = 0x10000;
2906 .text : @{ *(.text) @}
2907 . = 0x8000000;
2908 .data : @{ *(.data) @}
2909 .bss : @{ *(.bss) @}
2910 @}
2911 @end smallexample
2912
2913 You write the @samp{SECTIONS} command as the keyword @samp{SECTIONS},
2914 followed by a series of symbol assignments and output section
2915 descriptions enclosed in curly braces.
2916
2917 The first line inside the @samp{SECTIONS} command of the above example
2918 sets the value of the special symbol @samp{.}, which is the location
2919 counter. If you do not specify the address of an output section in some
2920 other way (other ways are described later), the address is set from the
2921 current value of the location counter. The location counter is then
2922 incremented by the size of the output section. At the start of the
2923 @samp{SECTIONS} command, the location counter has the value @samp{0}.
2924
2925 The second line defines an output section, @samp{.text}. The colon is
2926 required syntax which may be ignored for now. Within the curly braces
2927 after the output section name, you list the names of the input sections
2928 which should be placed into this output section. The @samp{*} is a
2929 wildcard which matches any file name. The expression @samp{*(.text)}
2930 means all @samp{.text} input sections in all input files.
2931
2932 Since the location counter is @samp{0x10000} when the output section
2933 @samp{.text} is defined, the linker will set the address of the
2934 @samp{.text} section in the output file to be @samp{0x10000}.
2935
2936 The remaining lines define the @samp{.data} and @samp{.bss} sections in
2937 the output file. The linker will place the @samp{.data} output section
2938 at address @samp{0x8000000}. After the linker places the @samp{.data}
2939 output section, the value of the location counter will be
2940 @samp{0x8000000} plus the size of the @samp{.data} output section. The
2941 effect is that the linker will place the @samp{.bss} output section
2942 immediately after the @samp{.data} output section in memory.
2943
2944 The linker will ensure that each output section has the required
2945 alignment, by increasing the location counter if necessary. In this
2946 example, the specified addresses for the @samp{.text} and @samp{.data}
2947 sections will probably satisfy any alignment constraints, but the linker
2948 may have to create a small gap between the @samp{.data} and @samp{.bss}
2949 sections.
2950
2951 That's it! That's a simple and complete linker script.
2952
2953 @node Simple Commands
2954 @section Simple Linker Script Commands
2955 @cindex linker script simple commands
2956 In this section we describe the simple linker script commands.
2957
2958 @menu
2959 * Entry Point:: Setting the entry point
2960 * File Commands:: Commands dealing with files
2961 @ifclear SingleFormat
2962 * Format Commands:: Commands dealing with object file formats
2963 @end ifclear
2964
2965 * REGION_ALIAS:: Assign alias names to memory regions
2966 * Miscellaneous Commands:: Other linker script commands
2967 @end menu
2968
2969 @node Entry Point
2970 @subsection Setting the Entry Point
2971 @kindex ENTRY(@var{symbol})
2972 @cindex start of execution
2973 @cindex first instruction
2974 @cindex entry point
2975 The first instruction to execute in a program is called the @dfn{entry
2976 point}. You can use the @code{ENTRY} linker script command to set the
2977 entry point. The argument is a symbol name:
2978 @smallexample
2979 ENTRY(@var{symbol})
2980 @end smallexample
2981
2982 There are several ways to set the entry point. The linker will set the
2983 entry point by trying each of the following methods in order, and
2984 stopping when one of them succeeds:
2985 @itemize @bullet
2986 @item
2987 the @samp{-e} @var{entry} command-line option;
2988 @item
2989 the @code{ENTRY(@var{symbol})} command in a linker script;
2990 @item
2991 the value of a target specific symbol, if it is defined; For many
2992 targets this is @code{start}, but PE and BeOS based systems for example
2993 check a list of possible entry symbols, matching the first one found.
2994 @item
2995 the address of the first byte of the @samp{.text} section, if present;
2996 @item
2997 The address @code{0}.
2998 @end itemize
2999
3000 @node File Commands
3001 @subsection Commands Dealing with Files
3002 @cindex linker script file commands
3003 Several linker script commands deal with files.
3004
3005 @table @code
3006 @item INCLUDE @var{filename}
3007 @kindex INCLUDE @var{filename}
3008 @cindex including a linker script
3009 Include the linker script @var{filename} at this point. The file will
3010 be searched for in the current directory, and in any directory specified
3011 with the @option{-L} option. You can nest calls to @code{INCLUDE} up to
3012 10 levels deep.
3013
3014 You can place @code{INCLUDE} directives at the top level, in @code{MEMORY} or
3015 @code{SECTIONS} commands, or in output section descriptions.
3016
3017 @item INPUT(@var{file}, @var{file}, @dots{})
3018 @itemx INPUT(@var{file} @var{file} @dots{})
3019 @kindex INPUT(@var{files})
3020 @cindex input files in linker scripts
3021 @cindex input object files in linker scripts
3022 @cindex linker script input object files
3023 The @code{INPUT} command directs the linker to include the named files
3024 in the link, as though they were named on the command line.
3025
3026 For example, if you always want to include @file{subr.o} any time you do
3027 a link, but you can't be bothered to put it on every link command line,
3028 then you can put @samp{INPUT (subr.o)} in your linker script.
3029
3030 In fact, if you like, you can list all of your input files in the linker
3031 script, and then invoke the linker with nothing but a @samp{-T} option.
3032
3033 In case a @dfn{sysroot prefix} is configured, and the filename starts
3034 with the @samp{/} character, and the script being processed was
3035 located inside the @dfn{sysroot prefix}, the filename will be looked
3036 for in the @dfn{sysroot prefix}. Otherwise, the linker will try to
3037 open the file in the current directory. If it is not found, the
3038 linker will search through the archive library search path. See the
3039 description of @samp{-L} in @ref{Options,,Command Line Options}.
3040
3041 If you use @samp{INPUT (-l@var{file})}, @command{ld} will transform the
3042 name to @code{lib@var{file}.a}, as with the command line argument
3043 @samp{-l}.
3044
3045 When you use the @code{INPUT} command in an implicit linker script, the
3046 files will be included in the link at the point at which the linker
3047 script file is included. This can affect archive searching.
3048
3049 @item GROUP(@var{file}, @var{file}, @dots{})
3050 @itemx GROUP(@var{file} @var{file} @dots{})
3051 @kindex GROUP(@var{files})
3052 @cindex grouping input files
3053 The @code{GROUP} command is like @code{INPUT}, except that the named
3054 files should all be archives, and they are searched repeatedly until no
3055 new undefined references are created. See the description of @samp{-(}
3056 in @ref{Options,,Command Line Options}.
3057
3058 @item AS_NEEDED(@var{file}, @var{file}, @dots{})
3059 @itemx AS_NEEDED(@var{file} @var{file} @dots{})
3060 @kindex AS_NEEDED(@var{files})
3061 This construct can appear only inside of the @code{INPUT} or @code{GROUP}
3062 commands, among other filenames. The files listed will be handled
3063 as if they appear directly in the @code{INPUT} or @code{GROUP} commands,
3064 with the exception of ELF shared libraries, that will be added only
3065 when they are actually needed. This construct essentially enables
3066 @option{--as-needed} option for all the files listed inside of it
3067 and restores previous @option{--as-needed} resp. @option{--no-as-needed}
3068 setting afterwards.
3069
3070 @item OUTPUT(@var{filename})
3071 @kindex OUTPUT(@var{filename})
3072 @cindex output file name in linker script
3073 The @code{OUTPUT} command names the output file. Using
3074 @code{OUTPUT(@var{filename})} in the linker script is exactly like using
3075 @samp{-o @var{filename}} on the command line (@pxref{Options,,Command
3076 Line Options}). If both are used, the command line option takes
3077 precedence.
3078
3079 You can use the @code{OUTPUT} command to define a default name for the
3080 output file other than the usual default of @file{a.out}.
3081
3082 @item SEARCH_DIR(@var{path})
3083 @kindex SEARCH_DIR(@var{path})
3084 @cindex library search path in linker script
3085 @cindex archive search path in linker script
3086 @cindex search path in linker script
3087 The @code{SEARCH_DIR} command adds @var{path} to the list of paths where
3088 @command{ld} looks for archive libraries. Using
3089 @code{SEARCH_DIR(@var{path})} is exactly like using @samp{-L @var{path}}
3090 on the command line (@pxref{Options,,Command Line Options}). If both
3091 are used, then the linker will search both paths. Paths specified using
3092 the command line option are searched first.
3093
3094 @item STARTUP(@var{filename})
3095 @kindex STARTUP(@var{filename})
3096 @cindex first input file
3097 The @code{STARTUP} command is just like the @code{INPUT} command, except
3098 that @var{filename} will become the first input file to be linked, as
3099 though it were specified first on the command line. This may be useful
3100 when using a system in which the entry point is always the start of the
3101 first file.
3102 @end table
3103
3104 @ifclear SingleFormat
3105 @node Format Commands
3106 @subsection Commands Dealing with Object File Formats
3107 A couple of linker script commands deal with object file formats.
3108
3109 @table @code
3110 @item OUTPUT_FORMAT(@var{bfdname})
3111 @itemx OUTPUT_FORMAT(@var{default}, @var{big}, @var{little})
3112 @kindex OUTPUT_FORMAT(@var{bfdname})
3113 @cindex output file format in linker script
3114 The @code{OUTPUT_FORMAT} command names the BFD format to use for the
3115 output file (@pxref{BFD}). Using @code{OUTPUT_FORMAT(@var{bfdname})} is
3116 exactly like using @samp{--oformat @var{bfdname}} on the command line
3117 (@pxref{Options,,Command Line Options}). If both are used, the command
3118 line option takes precedence.
3119
3120 You can use @code{OUTPUT_FORMAT} with three arguments to use different
3121 formats based on the @samp{-EB} and @samp{-EL} command line options.
3122 This permits the linker script to set the output format based on the
3123 desired endianness.
3124
3125 If neither @samp{-EB} nor @samp{-EL} are used, then the output format
3126 will be the first argument, @var{default}. If @samp{-EB} is used, the
3127 output format will be the second argument, @var{big}. If @samp{-EL} is
3128 used, the output format will be the third argument, @var{little}.
3129
3130 For example, the default linker script for the MIPS ELF target uses this
3131 command:
3132 @smallexample
3133 OUTPUT_FORMAT(elf32-bigmips, elf32-bigmips, elf32-littlemips)
3134 @end smallexample
3135 This says that the default format for the output file is
3136 @samp{elf32-bigmips}, but if the user uses the @samp{-EL} command line
3137 option, the output file will be created in the @samp{elf32-littlemips}
3138 format.
3139
3140 @item TARGET(@var{bfdname})
3141 @kindex TARGET(@var{bfdname})
3142 @cindex input file format in linker script
3143 The @code{TARGET} command names the BFD format to use when reading input
3144 files. It affects subsequent @code{INPUT} and @code{GROUP} commands.
3145 This command is like using @samp{-b @var{bfdname}} on the command line
3146 (@pxref{Options,,Command Line Options}). If the @code{TARGET} command
3147 is used but @code{OUTPUT_FORMAT} is not, then the last @code{TARGET}
3148 command is also used to set the format for the output file. @xref{BFD}.
3149 @end table
3150 @end ifclear
3151
3152 @node REGION_ALIAS
3153 @subsection Assign alias names to memory regions
3154 @kindex REGION_ALIAS(@var{alias}, @var{region})
3155 @cindex region alias
3156 @cindex region names
3157
3158 Alias names can be added to existing memory regions created with the
3159 @ref{MEMORY} command. Each name corresponds to at most one memory region.
3160
3161 @smallexample
3162 REGION_ALIAS(@var{alias}, @var{region})
3163 @end smallexample
3164
3165 The @code{REGION_ALIAS} function creates an alias name @var{alias} for the
3166 memory region @var{region}. This allows a flexible mapping of output sections
3167 to memory regions. An example follows.
3168
3169 Suppose we have an application for embedded systems which come with various
3170 memory storage devices. All have a general purpose, volatile memory @code{RAM}
3171 that allows code execution or data storage. Some may have a read-only,
3172 non-volatile memory @code{ROM} that allows code execution and read-only data
3173 access. The last variant is a read-only, non-volatile memory @code{ROM2} with
3174 read-only data access and no code execution capability. We have four output
3175 sections:
3176
3177 @itemize @bullet
3178 @item
3179 @code{.text} program code;
3180 @item
3181 @code{.rodata} read-only data;
3182 @item
3183 @code{.data} read-write initialized data;
3184 @item
3185 @code{.bss} read-write zero initialized data.
3186 @end itemize
3187
3188 The goal is to provide a linker command file that contains a system independent
3189 part defining the output sections and a system dependent part mapping the
3190 output sections to the memory regions available on the system. Our embedded
3191 systems come with three different memory setups @code{A}, @code{B} and
3192 @code{C}:
3193 @multitable @columnfractions .25 .25 .25 .25
3194 @item Section @tab Variant A @tab Variant B @tab Variant C
3195 @item .text @tab RAM @tab ROM @tab ROM
3196 @item .rodata @tab RAM @tab ROM @tab ROM2
3197 @item .data @tab RAM @tab RAM/ROM @tab RAM/ROM2
3198 @item .bss @tab RAM @tab RAM @tab RAM
3199 @end multitable
3200 The notation @code{RAM/ROM} or @code{RAM/ROM2} means that this section is
3201 loaded into region @code{ROM} or @code{ROM2} respectively. Please note that
3202 the load address of the @code{.data} section starts in all three variants at
3203 the end of the @code{.rodata} section.
3204
3205 The base linker script that deals with the output sections follows. It
3206 includes the system dependent @code{linkcmds.memory} file that describes the
3207 memory layout:
3208 @smallexample
3209 INCLUDE linkcmds.memory
3210
3211 SECTIONS
3212 @{
3213 .text :
3214 @{
3215 *(.text)
3216 @} > REGION_TEXT
3217 .rodata :
3218 @{
3219 *(.rodata)
3220 rodata_end = .;
3221 @} > REGION_RODATA
3222 .data : AT (rodata_end)
3223 @{
3224 data_start = .;
3225 *(.data)
3226 @} > REGION_DATA
3227 data_size = SIZEOF(.data);
3228 data_load_start = LOADADDR(.data);
3229 .bss :
3230 @{
3231 *(.bss)
3232 @} > REGION_BSS
3233 @}
3234 @end smallexample
3235
3236 Now we need three different @code{linkcmds.memory} files to define memory
3237 regions and alias names. The content of @code{linkcmds.memory} for the three
3238 variants @code{A}, @code{B} and @code{C}:
3239 @table @code
3240 @item A
3241 Here everything goes into the @code{RAM}.
3242 @smallexample
3243 MEMORY
3244 @{
3245 RAM : ORIGIN = 0, LENGTH = 4M
3246 @}
3247
3248 REGION_ALIAS("REGION_TEXT", RAM);
3249 REGION_ALIAS("REGION_RODATA", RAM);
3250 REGION_ALIAS("REGION_DATA", RAM);
3251 REGION_ALIAS("REGION_BSS", RAM);
3252 @end smallexample
3253 @item B
3254 Program code and read-only data go into the @code{ROM}. Read-write data goes
3255 into the @code{RAM}. An image of the initialized data is loaded into the
3256 @code{ROM} and will be copied during system start into the @code{RAM}.
3257 @smallexample
3258 MEMORY
3259 @{
3260 ROM : ORIGIN = 0, LENGTH = 3M
3261 RAM : ORIGIN = 0x10000000, LENGTH = 1M
3262 @}
3263
3264 REGION_ALIAS("REGION_TEXT", ROM);
3265 REGION_ALIAS("REGION_RODATA", ROM);
3266 REGION_ALIAS("REGION_DATA", RAM);
3267 REGION_ALIAS("REGION_BSS", RAM);
3268 @end smallexample
3269 @item C
3270 Program code goes into the @code{ROM}. Read-only data goes into the
3271 @code{ROM2}. Read-write data goes into the @code{RAM}. An image of the
3272 initialized data is loaded into the @code{ROM2} and will be copied during
3273 system start into the @code{RAM}.
3274 @smallexample
3275 MEMORY
3276 @{
3277 ROM : ORIGIN = 0, LENGTH = 2M
3278 ROM2 : ORIGIN = 0x10000000, LENGTH = 1M
3279 RAM : ORIGIN = 0x20000000, LENGTH = 1M
3280 @}
3281
3282 REGION_ALIAS("REGION_TEXT", ROM);
3283 REGION_ALIAS("REGION_RODATA", ROM2);
3284 REGION_ALIAS("REGION_DATA", RAM);
3285 REGION_ALIAS("REGION_BSS", RAM);
3286 @end smallexample
3287 @end table
3288
3289 It is possible to write a common system initialization routine to copy the
3290 @code{.data} section from @code{ROM} or @code{ROM2} into the @code{RAM} if
3291 necessary:
3292 @smallexample
3293 #include <string.h>
3294
3295 extern char data_start [];
3296 extern char data_size [];
3297 extern char data_load_start [];
3298
3299 void copy_data(void)
3300 @{
3301 if (data_start != data_load_start)
3302 @{
3303 memcpy(data_start, data_load_start, (size_t) data_size);
3304 @}
3305 @}
3306 @end smallexample
3307
3308 @node Miscellaneous Commands
3309 @subsection Other Linker Script Commands
3310 There are a few other linker scripts commands.
3311
3312 @table @code
3313 @item ASSERT(@var{exp}, @var{message})
3314 @kindex ASSERT
3315 @cindex assertion in linker script
3316 Ensure that @var{exp} is non-zero. If it is zero, then exit the linker
3317 with an error code, and print @var{message}.
3318
3319 @item EXTERN(@var{symbol} @var{symbol} @dots{})
3320 @kindex EXTERN
3321 @cindex undefined symbol in linker script
3322 Force @var{symbol} to be entered in the output file as an undefined
3323 symbol. Doing this may, for example, trigger linking of additional
3324 modules from standard libraries. You may list several @var{symbol}s for
3325 each @code{EXTERN}, and you may use @code{EXTERN} multiple times. This
3326 command has the same effect as the @samp{-u} command-line option.
3327
3328 @item FORCE_COMMON_ALLOCATION
3329 @kindex FORCE_COMMON_ALLOCATION
3330 @cindex common allocation in linker script
3331 This command has the same effect as the @samp{-d} command-line option:
3332 to make @command{ld} assign space to common symbols even if a relocatable
3333 output file is specified (@samp{-r}).
3334
3335 @item INHIBIT_COMMON_ALLOCATION
3336 @kindex INHIBIT_COMMON_ALLOCATION
3337 @cindex common allocation in linker script
3338 This command has the same effect as the @samp{--no-define-common}
3339 command-line option: to make @code{ld} omit the assignment of addresses
3340 to common symbols even for a non-relocatable output file.
3341
3342 @item INSERT [ AFTER | BEFORE ] @var{output_section}
3343 @kindex INSERT
3344 @cindex insert user script into default script
3345 This command is typically used in a script specified by @samp{-T} to
3346 augment the default @code{SECTIONS} with, for example, overlays. It
3347 inserts all prior linker script statements after (or before)
3348 @var{output_section}, and also causes @samp{-T} to not override the
3349 default linker script. The exact insertion point is as for orphan
3350 sections. @xref{Location Counter}. The insertion happens after the
3351 linker has mapped input sections to output sections. Prior to the
3352 insertion, since @samp{-T} scripts are parsed before the default
3353 linker script, statements in the @samp{-T} script occur before the
3354 default linker script statements in the internal linker representation
3355 of the script. In particular, input section assignments will be made
3356 to @samp{-T} output sections before those in the default script. Here
3357 is an example of how a @samp{-T} script using @code{INSERT} might look:
3358
3359 @smallexample
3360 SECTIONS
3361 @{
3362 OVERLAY :
3363 @{
3364 .ov1 @{ ov1*(.text) @}
3365 .ov2 @{ ov2*(.text) @}
3366 @}
3367 @}
3368 INSERT AFTER .text;
3369 @end smallexample
3370
3371 @item NOCROSSREFS(@var{section} @var{section} @dots{})
3372 @kindex NOCROSSREFS(@var{sections})
3373 @cindex cross references
3374 This command may be used to tell @command{ld} to issue an error about any
3375 references among certain output sections.
3376
3377 In certain types of programs, particularly on embedded systems when
3378 using overlays, when one section is loaded into memory, another section
3379 will not be. Any direct references between the two sections would be
3380 errors. For example, it would be an error if code in one section called
3381 a function defined in the other section.
3382
3383 The @code{NOCROSSREFS} command takes a list of output section names. If
3384 @command{ld} detects any cross references between the sections, it reports
3385 an error and returns a non-zero exit status. Note that the
3386 @code{NOCROSSREFS} command uses output section names, not input section
3387 names.
3388
3389 @ifclear SingleFormat
3390 @item OUTPUT_ARCH(@var{bfdarch})
3391 @kindex OUTPUT_ARCH(@var{bfdarch})
3392 @cindex machine architecture
3393 @cindex architecture
3394 Specify a particular output machine architecture. The argument is one
3395 of the names used by the BFD library (@pxref{BFD}). You can see the
3396 architecture of an object file by using the @code{objdump} program with
3397 the @samp{-f} option.
3398 @end ifclear
3399
3400 @item LD_FEATURE(@var{string})
3401 @kindex LD_FEATURE(@var{string})
3402 This command may be used to modify @command{ld} behavior. If
3403 @var{string} is @code{"SANE_EXPR"} then absolute symbols and numbers
3404 in a script are simply treated as numbers everywhere.
3405 @xref{Expression Section}.
3406 @end table
3407
3408 @node Assignments
3409 @section Assigning Values to Symbols
3410 @cindex assignment in scripts
3411 @cindex symbol definition, scripts
3412 @cindex variables, defining
3413 You may assign a value to a symbol in a linker script. This will define
3414 the symbol and place it into the symbol table with a global scope.
3415
3416 @menu
3417 * Simple Assignments:: Simple Assignments
3418 * HIDDEN:: HIDDEN
3419 * PROVIDE:: PROVIDE
3420 * PROVIDE_HIDDEN:: PROVIDE_HIDDEN
3421 * Source Code Reference:: How to use a linker script defined symbol in source code
3422 @end menu
3423
3424 @node Simple Assignments
3425 @subsection Simple Assignments
3426
3427 You may assign to a symbol using any of the C assignment operators:
3428
3429 @table @code
3430 @item @var{symbol} = @var{expression} ;
3431 @itemx @var{symbol} += @var{expression} ;
3432 @itemx @var{symbol} -= @var{expression} ;
3433 @itemx @var{symbol} *= @var{expression} ;
3434 @itemx @var{symbol} /= @var{expression} ;
3435 @itemx @var{symbol} <<= @var{expression} ;
3436 @itemx @var{symbol} >>= @var{expression} ;
3437 @itemx @var{symbol} &= @var{expression} ;
3438 @itemx @var{symbol} |= @var{expression} ;
3439 @end table
3440
3441 The first case will define @var{symbol} to the value of
3442 @var{expression}. In the other cases, @var{symbol} must already be
3443 defined, and the value will be adjusted accordingly.
3444
3445 The special symbol name @samp{.} indicates the location counter. You
3446 may only use this within a @code{SECTIONS} command. @xref{Location Counter}.
3447
3448 The semicolon after @var{expression} is required.
3449
3450 Expressions are defined below; see @ref{Expressions}.
3451
3452 You may write symbol assignments as commands in their own right, or as
3453 statements within a @code{SECTIONS} command, or as part of an output
3454 section description in a @code{SECTIONS} command.
3455
3456 The section of the symbol will be set from the section of the
3457 expression; for more information, see @ref{Expression Section}.
3458
3459 Here is an example showing the three different places that symbol
3460 assignments may be used:
3461
3462 @smallexample
3463 floating_point = 0;
3464 SECTIONS
3465 @{
3466 .text :
3467 @{
3468 *(.text)
3469 _etext = .;
3470 @}
3471 _bdata = (. + 3) & ~ 3;
3472 .data : @{ *(.data) @}
3473 @}
3474 @end smallexample
3475 @noindent
3476 In this example, the symbol @samp{floating_point} will be defined as
3477 zero. The symbol @samp{_etext} will be defined as the address following
3478 the last @samp{.text} input section. The symbol @samp{_bdata} will be
3479 defined as the address following the @samp{.text} output section aligned
3480 upward to a 4 byte boundary.
3481
3482 @node HIDDEN
3483 @subsection HIDDEN
3484 @cindex HIDDEN
3485 For ELF targeted ports, define a symbol that will be hidden and won't be
3486 exported. The syntax is @code{HIDDEN(@var{symbol} = @var{expression})}.
3487
3488 Here is the example from @ref{Simple Assignments}, rewritten to use
3489 @code{HIDDEN}:
3490
3491 @smallexample
3492 HIDDEN(floating_point = 0);
3493 SECTIONS
3494 @{
3495 .text :
3496 @{
3497 *(.text)
3498 HIDDEN(_etext = .);
3499 @}
3500 HIDDEN(_bdata = (. + 3) & ~ 3);
3501 .data : @{ *(.data) @}
3502 @}
3503 @end smallexample
3504 @noindent
3505 In this case none of the three symbols will be visible outside this module.
3506
3507 @node PROVIDE
3508 @subsection PROVIDE
3509 @cindex PROVIDE
3510 In some cases, it is desirable for a linker script to define a symbol
3511 only if it is referenced and is not defined by any object included in
3512 the link. For example, traditional linkers defined the symbol
3513 @samp{etext}. However, ANSI C requires that the user be able to use
3514 @samp{etext} as a function name without encountering an error. The
3515 @code{PROVIDE} keyword may be used to define a symbol, such as
3516 @samp{etext}, only if it is referenced but not defined. The syntax is
3517 @code{PROVIDE(@var{symbol} = @var{expression})}.
3518
3519 Here is an example of using @code{PROVIDE} to define @samp{etext}:
3520 @smallexample
3521 SECTIONS
3522 @{
3523 .text :
3524 @{
3525 *(.text)
3526 _etext = .;
3527 PROVIDE(etext = .);
3528 @}
3529 @}
3530 @end smallexample
3531
3532 In this example, if the program defines @samp{_etext} (with a leading
3533 underscore), the linker will give a multiple definition error. If, on
3534 the other hand, the program defines @samp{etext} (with no leading
3535 underscore), the linker will silently use the definition in the program.
3536 If the program references @samp{etext} but does not define it, the
3537 linker will use the definition in the linker script.
3538
3539 @node PROVIDE_HIDDEN
3540 @subsection PROVIDE_HIDDEN
3541 @cindex PROVIDE_HIDDEN
3542 Similar to @code{PROVIDE}. For ELF targeted ports, the symbol will be
3543 hidden and won't be exported.
3544
3545 @node Source Code Reference
3546 @subsection Source Code Reference
3547
3548 Accessing a linker script defined variable from source code is not
3549 intuitive. In particular a linker script symbol is not equivalent to
3550 a variable declaration in a high level language, it is instead a
3551 symbol that does not have a value.
3552
3553 Before going further, it is important to note that compilers often
3554 transform names in the source code into different names when they are
3555 stored in the symbol table. For example, Fortran compilers commonly
3556 prepend or append an underscore, and C++ performs extensive @samp{name
3557 mangling}. Therefore there might be a discrepancy between the name
3558 of a variable as it is used in source code and the name of the same
3559 variable as it is defined in a linker script. For example in C a
3560 linker script variable might be referred to as:
3561
3562 @smallexample
3563 extern int foo;
3564 @end smallexample
3565
3566 But in the linker script it might be defined as:
3567
3568 @smallexample
3569 _foo = 1000;
3570 @end smallexample
3571
3572 In the remaining examples however it is assumed that no name
3573 transformation has taken place.
3574
3575 When a symbol is declared in a high level language such as C, two
3576 things happen. The first is that the compiler reserves enough space
3577 in the program's memory to hold the @emph{value} of the symbol. The
3578 second is that the compiler creates an entry in the program's symbol
3579 table which holds the symbol's @emph{address}. ie the symbol table
3580 contains the address of the block of memory holding the symbol's
3581 value. So for example the following C declaration, at file scope:
3582
3583 @smallexample
3584 int foo = 1000;
3585 @end smallexample
3586
3587 creates a entry called @samp{foo} in the symbol table. This entry
3588 holds the address of an @samp{int} sized block of memory where the
3589 number 1000 is initially stored.
3590
3591 When a program references a symbol the compiler generates code that
3592 first accesses the symbol table to find the address of the symbol's
3593 memory block and then code to read the value from that memory block.
3594 So:
3595
3596 @smallexample
3597 foo = 1;
3598 @end smallexample
3599
3600 looks up the symbol @samp{foo} in the symbol table, gets the address
3601 associated with this symbol and then writes the value 1 into that
3602 address. Whereas:
3603
3604 @smallexample
3605 int * a = & foo;
3606 @end smallexample
3607
3608 looks up the symbol @samp{foo} in the symbol table, gets it address
3609 and then copies this address into the block of memory associated with
3610 the variable @samp{a}.
3611
3612 Linker scripts symbol declarations, by contrast, create an entry in
3613 the symbol table but do not assign any memory to them. Thus they are
3614 an address without a value. So for example the linker script definition:
3615
3616 @smallexample
3617 foo = 1000;
3618 @end smallexample
3619
3620 creates an entry in the symbol table called @samp{foo} which holds
3621 the address of memory location 1000, but nothing special is stored at
3622 address 1000. This means that you cannot access the @emph{value} of a
3623 linker script defined symbol - it has no value - all you can do is
3624 access the @emph{address} of a linker script defined symbol.
3625
3626 Hence when you are using a linker script defined symbol in source code
3627 you should always take the address of the symbol, and never attempt to
3628 use its value. For example suppose you want to copy the contents of a
3629 section of memory called .ROM into a section called .FLASH and the
3630 linker script contains these declarations:
3631
3632 @smallexample
3633 @group
3634 start_of_ROM = .ROM;
3635 end_of_ROM = .ROM + sizeof (.ROM) - 1;
3636 start_of_FLASH = .FLASH;
3637 @end group
3638 @end smallexample
3639
3640 Then the C source code to perform the copy would be:
3641
3642 @smallexample
3643 @group
3644 extern char start_of_ROM, end_of_ROM, start_of_FLASH;
3645
3646 memcpy (& start_of_FLASH, & start_of_ROM, & end_of_ROM - & start_of_ROM);
3647 @end group
3648 @end smallexample
3649
3650 Note the use of the @samp{&} operators. These are correct.
3651
3652 @node SECTIONS
3653 @section SECTIONS Command
3654 @kindex SECTIONS
3655 The @code{SECTIONS} command tells the linker how to map input sections
3656 into output sections, and how to place the output sections in memory.
3657
3658 The format of the @code{SECTIONS} command is:
3659 @smallexample
3660 SECTIONS
3661 @{
3662 @var{sections-command}
3663 @var{sections-command}
3664 @dots{}
3665 @}
3666 @end smallexample
3667
3668 Each @var{sections-command} may of be one of the following:
3669
3670 @itemize @bullet
3671 @item
3672 an @code{ENTRY} command (@pxref{Entry Point,,Entry command})
3673 @item
3674 a symbol assignment (@pxref{Assignments})
3675 @item
3676 an output section description
3677 @item
3678 an overlay description
3679 @end itemize
3680
3681 The @code{ENTRY} command and symbol assignments are permitted inside the
3682 @code{SECTIONS} command for convenience in using the location counter in
3683 those commands. This can also make the linker script easier to
3684 understand because you can use those commands at meaningful points in
3685 the layout of the output file.
3686
3687 Output section descriptions and overlay descriptions are described
3688 below.
3689
3690 If you do not use a @code{SECTIONS} command in your linker script, the
3691 linker will place each input section into an identically named output
3692 section in the order that the sections are first encountered in the
3693 input files. If all input sections are present in the first file, for
3694 example, the order of sections in the output file will match the order
3695 in the first input file. The first section will be at address zero.
3696
3697 @menu
3698 * Output Section Description:: Output section description
3699 * Output Section Name:: Output section name
3700 * Output Section Address:: Output section address
3701 * Input Section:: Input section description
3702 * Output Section Data:: Output section data
3703 * Output Section Keywords:: Output section keywords
3704 * Output Section Discarding:: Output section discarding
3705 * Output Section Attributes:: Output section attributes
3706 * Overlay Description:: Overlay description
3707 @end menu
3708
3709 @node Output Section Description
3710 @subsection Output Section Description
3711 The full description of an output section looks like this:
3712 @smallexample
3713 @group
3714 @var{section} [@var{address}] [(@var{type})] :
3715 [AT(@var{lma})]
3716 [ALIGN(@var{section_align})]
3717 [SUBALIGN(@var{subsection_align})]
3718 [@var{constraint}]
3719 @{
3720 @var{output-section-command}
3721 @var{output-section-command}
3722 @dots{}
3723 @} [>@var{region}] [AT>@var{lma_region}] [:@var{phdr} :@var{phdr} @dots{}] [=@var{fillexp}]
3724 @end group
3725 @end smallexample
3726
3727 Most output sections do not use most of the optional section attributes.
3728
3729 The whitespace around @var{section} is required, so that the section
3730 name is unambiguous. The colon and the curly braces are also required.
3731 The line breaks and other white space are optional.
3732
3733 Each @var{output-section-command} may be one of the following:
3734
3735 @itemize @bullet
3736 @item
3737 a symbol assignment (@pxref{Assignments})
3738 @item
3739 an input section description (@pxref{Input Section})
3740 @item
3741 data values to include directly (@pxref{Output Section Data})
3742 @item
3743 a special output section keyword (@pxref{Output Section Keywords})
3744 @end itemize
3745
3746 @node Output Section Name
3747 @subsection Output Section Name
3748 @cindex name, section
3749 @cindex section name
3750 The name of the output section is @var{section}. @var{section} must
3751 meet the constraints of your output format. In formats which only
3752 support a limited number of sections, such as @code{a.out}, the name
3753 must be one of the names supported by the format (@code{a.out}, for
3754 example, allows only @samp{.text}, @samp{.data} or @samp{.bss}). If the
3755 output format supports any number of sections, but with numbers and not
3756 names (as is the case for Oasys), the name should be supplied as a
3757 quoted numeric string. A section name may consist of any sequence of
3758 characters, but a name which contains any unusual characters such as
3759 commas must be quoted.
3760
3761 The output section name @samp{/DISCARD/} is special; @ref{Output Section
3762 Discarding}.
3763
3764 @node Output Section Address
3765 @subsection Output Section Address
3766 @cindex address, section
3767 @cindex section address
3768 The @var{address} is an expression for the VMA (the virtual memory
3769 address) of the output section. This address is optional, but if it
3770 is provided then the output address will be set exactly as specified.
3771
3772 If the output address is not specified then one will be chosen for the
3773 section, based on the heuristic below. This address will be adjusted
3774 to fit the alignment requirement of the output section. The
3775 alignment requirement is the strictest alignment of any input section
3776 contained within the output section.
3777
3778 The output section address heuristic is as follows:
3779
3780 @itemize @bullet
3781 @item
3782 If an output memory @var{region} is set for the section then it
3783 is added to this region and its address will be the next free address
3784 in that region.
3785
3786 @item
3787 If the MEMORY command has been used to create a list of memory
3788 regions then the first region which has attributes compatible with the
3789 section is selected to contain it. The section's output address will
3790 be the next free address in that region; @ref{MEMORY}.
3791
3792 @item
3793 If no memory regions were specified, or none match the section then
3794 the output address will be based on the current value of the location
3795 counter.
3796 @end itemize
3797
3798 @noindent
3799 For example:
3800
3801 @smallexample
3802 .text . : @{ *(.text) @}
3803 @end smallexample
3804
3805 @noindent
3806 and
3807
3808 @smallexample
3809 .text : @{ *(.text) @}
3810 @end smallexample
3811
3812 @noindent
3813 are subtly different. The first will set the address of the
3814 @samp{.text} output section to the current value of the location
3815 counter. The second will set it to the current value of the location
3816 counter aligned to the strictest alignment of any of the @samp{.text}
3817 input sections.
3818
3819 The @var{address} may be an arbitrary expression; @ref{Expressions}.
3820 For example, if you want to align the section on a 0x10 byte boundary,
3821 so that the lowest four bits of the section address are zero, you could
3822 do something like this:
3823 @smallexample
3824 .text ALIGN(0x10) : @{ *(.text) @}
3825 @end smallexample
3826 @noindent
3827 This works because @code{ALIGN} returns the current location counter
3828 aligned upward to the specified value.
3829
3830 Specifying @var{address} for a section will change the value of the
3831 location counter, provided that the section is non-empty. (Empty
3832 sections are ignored).
3833
3834 @node Input Section
3835 @subsection Input Section Description
3836 @cindex input sections
3837 @cindex mapping input sections to output sections
3838 The most common output section command is an input section description.
3839
3840 The input section description is the most basic linker script operation.
3841 You use output sections to tell the linker how to lay out your program
3842 in memory. You use input section descriptions to tell the linker how to
3843 map the input files into your memory layout.
3844
3845 @menu
3846 * Input Section Basics:: Input section basics
3847 * Input Section Wildcards:: Input section wildcard patterns
3848 * Input Section Common:: Input section for common symbols
3849 * Input Section Keep:: Input section and garbage collection
3850 * Input Section Example:: Input section example
3851 @end menu
3852
3853 @node Input Section Basics
3854 @subsubsection Input Section Basics
3855 @cindex input section basics
3856 An input section description consists of a file name optionally followed
3857 by a list of section names in parentheses.
3858
3859 The file name and the section name may be wildcard patterns, which we
3860 describe further below (@pxref{Input Section Wildcards}).
3861
3862 The most common input section description is to include all input
3863 sections with a particular name in the output section. For example, to
3864 include all input @samp{.text} sections, you would write:
3865 @smallexample
3866 *(.text)
3867 @end smallexample
3868 @noindent
3869 Here the @samp{*} is a wildcard which matches any file name. To exclude a list
3870 of files from matching the file name wildcard, EXCLUDE_FILE may be used to
3871 match all files except the ones specified in the EXCLUDE_FILE list. For
3872 example:
3873 @smallexample
3874 *(EXCLUDE_FILE (*crtend.o *otherfile.o) .ctors)
3875 @end smallexample
3876 will cause all .ctors sections from all files except @file{crtend.o} and
3877 @file{otherfile.o} to be included.
3878
3879 There are two ways to include more than one section:
3880 @smallexample
3881 *(.text .rdata)
3882 *(.text) *(.rdata)
3883 @end smallexample
3884 @noindent
3885 The difference between these is the order in which the @samp{.text} and
3886 @samp{.rdata} input sections will appear in the output section. In the
3887 first example, they will be intermingled, appearing in the same order as
3888 they are found in the linker input. In the second example, all
3889 @samp{.text} input sections will appear first, followed by all
3890 @samp{.rdata} input sections.
3891
3892 You can specify a file name to include sections from a particular file.
3893 You would do this if one or more of your files contain special data that
3894 needs to be at a particular location in memory. For example:
3895 @smallexample
3896 data.o(.data)
3897 @end smallexample
3898
3899 To refine the sections that are included based on the section flags
3900 of an input section, INPUT_SECTION_FLAGS may be used.
3901
3902 Here is a simple example for using Section header flags for ELF sections:
3903
3904 @smallexample
3905 @group
3906 SECTIONS @{
3907 .text : @{ INPUT_SECTION_FLAGS (SHF_MERGE & SHF_STRINGS) *(.text) @}
3908 .text2 : @{ INPUT_SECTION_FLAGS (!SHF_WRITE) *(.text) @}
3909 @}
3910 @end group
3911 @end smallexample
3912
3913 In this example, the output section @samp{.text} will be comprised of any
3914 input section matching the name *(.text) whose section header flags
3915 @code{SHF_MERGE} and @code{SHF_STRINGS} are set. The output section
3916 @samp{.text2} will be comprised of any input section matching the name *(.text)
3917 whose section header flag @code{SHF_WRITE} is clear.
3918
3919 You can also specify files within archives by writing a pattern
3920 matching the archive, a colon, then the pattern matching the file,
3921 with no whitespace around the colon.
3922
3923 @table @samp
3924 @item archive:file
3925 matches file within archive
3926 @item archive:
3927 matches the whole archive
3928 @item :file
3929 matches file but not one in an archive
3930 @end table
3931
3932 Either one or both of @samp{archive} and @samp{file} can contain shell
3933 wildcards. On DOS based file systems, the linker will assume that a
3934 single letter followed by a colon is a drive specifier, so
3935 @samp{c:myfile.o} is a simple file specification, not @samp{myfile.o}
3936 within an archive called @samp{c}. @samp{archive:file} filespecs may
3937 also be used within an @code{EXCLUDE_FILE} list, but may not appear in
3938 other linker script contexts. For instance, you cannot extract a file
3939 from an archive by using @samp{archive:file} in an @code{INPUT}
3940 command.
3941
3942 If you use a file name without a list of sections, then all sections in
3943 the input file will be included in the output section. This is not
3944 commonly done, but it may by useful on occasion. For example:
3945 @smallexample
3946 data.o
3947 @end smallexample
3948
3949 When you use a file name which is not an @samp{archive:file} specifier
3950 and does not contain any wild card
3951 characters, the linker will first see if you also specified the file
3952 name on the linker command line or in an @code{INPUT} command. If you
3953 did not, the linker will attempt to open the file as an input file, as
3954 though it appeared on the command line. Note that this differs from an
3955 @code{INPUT} command, because the linker will not search for the file in
3956 the archive search path.
3957
3958 @node Input Section Wildcards
3959 @subsubsection Input Section Wildcard Patterns
3960 @cindex input section wildcards
3961 @cindex wildcard file name patterns
3962 @cindex file name wildcard patterns
3963 @cindex section name wildcard patterns
3964 In an input section description, either the file name or the section
3965 name or both may be wildcard patterns.
3966
3967 The file name of @samp{*} seen in many examples is a simple wildcard
3968 pattern for the file name.
3969
3970 The wildcard patterns are like those used by the Unix shell.
3971
3972 @table @samp
3973 @item *
3974 matches any number of characters
3975 @item ?
3976 matches any single character
3977 @item [@var{chars}]
3978 matches a single instance of any of the @var{chars}; the @samp{-}
3979 character may be used to specify a range of characters, as in
3980 @samp{[a-z]} to match any lower case letter
3981 @item \
3982 quotes the following character
3983 @end table
3984
3985 When a file name is matched with a wildcard, the wildcard characters
3986 will not match a @samp{/} character (used to separate directory names on
3987 Unix). A pattern consisting of a single @samp{*} character is an
3988 exception; it will always match any file name, whether it contains a
3989 @samp{/} or not. In a section name, the wildcard characters will match
3990 a @samp{/} character.
3991
3992 File name wildcard patterns only match files which are explicitly
3993 specified on the command line or in an @code{INPUT} command. The linker
3994 does not search directories to expand wildcards.
3995
3996 If a file name matches more than one wildcard pattern, or if a file name
3997 appears explicitly and is also matched by a wildcard pattern, the linker
3998 will use the first match in the linker script. For example, this
3999 sequence of input section descriptions is probably in error, because the
4000 @file{data.o} rule will not be used:
4001 @smallexample
4002 .data : @{ *(.data) @}
4003 .data1 : @{ data.o(.data) @}
4004 @end smallexample
4005
4006 @cindex SORT_BY_NAME
4007 Normally, the linker will place files and sections matched by wildcards
4008 in the order in which they are seen during the link. You can change
4009 this by using the @code{SORT_BY_NAME} keyword, which appears before a wildcard
4010 pattern in parentheses (e.g., @code{SORT_BY_NAME(.text*)}). When the
4011 @code{SORT_BY_NAME} keyword is used, the linker will sort the files or sections
4012 into ascending order by name before placing them in the output file.
4013
4014 @cindex SORT_BY_ALIGNMENT
4015 @code{SORT_BY_ALIGNMENT} is very similar to @code{SORT_BY_NAME}. The
4016 difference is @code{SORT_BY_ALIGNMENT} will sort sections into
4017 ascending order by alignment before placing them in the output file.
4018
4019 @cindex SORT_BY_INIT_PRIORITY
4020 @code{SORT_BY_INIT_PRIORITY} is very similar to @code{SORT_BY_NAME}. The
4021 difference is @code{SORT_BY_INIT_PRIORITY} will sort sections into
4022 ascending order by numerical value of the GCC init_priority attribute
4023 encoded in the section name before placing them in the output file.
4024
4025 @cindex SORT
4026 @code{SORT} is an alias for @code{SORT_BY_NAME}.
4027
4028 When there are nested section sorting commands in linker script, there
4029 can be at most 1 level of nesting for section sorting commands.
4030
4031 @enumerate
4032 @item
4033 @code{SORT_BY_NAME} (@code{SORT_BY_ALIGNMENT} (wildcard section pattern)).
4034 It will sort the input sections by name first, then by alignment if 2
4035 sections have the same name.
4036 @item
4037 @code{SORT_BY_ALIGNMENT} (@code{SORT_BY_NAME} (wildcard section pattern)).
4038 It will sort the input sections by alignment first, then by name if 2
4039 sections have the same alignment.
4040 @item
4041 @code{SORT_BY_NAME} (@code{SORT_BY_NAME} (wildcard section pattern)) is
4042 treated the same as @code{SORT_BY_NAME} (wildcard section pattern).
4043 @item
4044 @code{SORT_BY_ALIGNMENT} (@code{SORT_BY_ALIGNMENT} (wildcard section pattern))
4045 is treated the same as @code{SORT_BY_ALIGNMENT} (wildcard section pattern).
4046 @item
4047 All other nested section sorting commands are invalid.
4048 @end enumerate
4049
4050 When both command line section sorting option and linker script
4051 section sorting command are used, section sorting command always
4052 takes precedence over the command line option.
4053
4054 If the section sorting command in linker script isn't nested, the
4055 command line option will make the section sorting command to be
4056 treated as nested sorting command.
4057
4058 @enumerate
4059 @item
4060 @code{SORT_BY_NAME} (wildcard section pattern ) with
4061 @option{--sort-sections alignment} is equivalent to
4062 @code{SORT_BY_NAME} (@code{SORT_BY_ALIGNMENT} (wildcard section pattern)).
4063 @item
4064 @code{SORT_BY_ALIGNMENT} (wildcard section pattern) with
4065 @option{--sort-section name} is equivalent to
4066 @code{SORT_BY_ALIGNMENT} (@code{SORT_BY_NAME} (wildcard section pattern)).
4067 @end enumerate
4068
4069 If the section sorting command in linker script is nested, the
4070 command line option will be ignored.
4071
4072 @cindex SORT_NONE
4073 @code{SORT_NONE} disables section sorting by ignoring the command line
4074 section sorting option.
4075
4076 If you ever get confused about where input sections are going, use the
4077 @samp{-M} linker option to generate a map file. The map file shows
4078 precisely how input sections are mapped to output sections.
4079
4080 This example shows how wildcard patterns might be used to partition
4081 files. This linker script directs the linker to place all @samp{.text}
4082 sections in @samp{.text} and all @samp{.bss} sections in @samp{.bss}.
4083 The linker will place the @samp{.data} section from all files beginning
4084 with an upper case character in @samp{.DATA}; for all other files, the
4085 linker will place the @samp{.data} section in @samp{.data}.
4086 @smallexample
4087 @group
4088 SECTIONS @{
4089 .text : @{ *(.text) @}
4090 .DATA : @{ [A-Z]*(.data) @}
4091 .data : @{ *(.data) @}
4092 .bss : @{ *(.bss) @}
4093 @}
4094 @end group
4095 @end smallexample
4096
4097 @node Input Section Common
4098 @subsubsection Input Section for Common Symbols
4099 @cindex common symbol placement
4100 @cindex uninitialized data placement
4101 A special notation is needed for common symbols, because in many object
4102 file formats common symbols do not have a particular input section. The
4103 linker treats common symbols as though they are in an input section
4104 named @samp{COMMON}.
4105
4106 You may use file names with the @samp{COMMON} section just as with any
4107 other input sections. You can use this to place common symbols from a
4108 particular input file in one section while common symbols from other
4109 input files are placed in another section.
4110
4111 In most cases, common symbols in input files will be placed in the
4112 @samp{.bss} section in the output file. For example:
4113 @smallexample
4114 .bss @{ *(.bss) *(COMMON) @}
4115 @end smallexample
4116
4117 @cindex scommon section
4118 @cindex small common symbols
4119 Some object file formats have more than one type of common symbol. For
4120 example, the MIPS ELF object file format distinguishes standard common
4121 symbols and small common symbols. In this case, the linker will use a
4122 different special section name for other types of common symbols. In
4123 the case of MIPS ELF, the linker uses @samp{COMMON} for standard common
4124 symbols and @samp{.scommon} for small common symbols. This permits you
4125 to map the different types of common symbols into memory at different
4126 locations.
4127
4128 @cindex [COMMON]
4129 You will sometimes see @samp{[COMMON]} in old linker scripts. This
4130 notation is now considered obsolete. It is equivalent to
4131 @samp{*(COMMON)}.
4132
4133 @node Input Section Keep
4134 @subsubsection Input Section and Garbage Collection
4135 @cindex KEEP
4136 @cindex garbage collection
4137 When link-time garbage collection is in use (@samp{--gc-sections}),
4138 it is often useful to mark sections that should not be eliminated.
4139 This is accomplished by surrounding an input section's wildcard entry
4140 with @code{KEEP()}, as in @code{KEEP(*(.init))} or
4141 @code{KEEP(SORT_BY_NAME(*)(.ctors))}.
4142
4143 @node Input Section Example
4144 @subsubsection Input Section Example
4145 The following example is a complete linker script. It tells the linker
4146 to read all of the sections from file @file{all.o} and place them at the
4147 start of output section @samp{outputa} which starts at location
4148 @samp{0x10000}. All of section @samp{.input1} from file @file{foo.o}
4149 follows immediately, in the same output section. All of section
4150 @samp{.input2} from @file{foo.o} goes into output section
4151 @samp{outputb}, followed by section @samp{.input1} from @file{foo1.o}.
4152 All of the remaining @samp{.input1} and @samp{.input2} sections from any
4153 files are written to output section @samp{outputc}.
4154
4155 @smallexample
4156 @group
4157 SECTIONS @{
4158 outputa 0x10000 :
4159 @{
4160 all.o
4161 foo.o (.input1)
4162 @}
4163 @end group
4164 @group
4165 outputb :
4166 @{
4167 foo.o (.input2)
4168 foo1.o (.input1)
4169 @}
4170 @end group
4171 @group
4172 outputc :
4173 @{
4174 *(.input1)
4175 *(.input2)
4176 @}
4177 @}
4178 @end group
4179 @end smallexample
4180
4181 @node Output Section Data
4182 @subsection Output Section Data
4183 @cindex data
4184 @cindex section data
4185 @cindex output section data
4186 @kindex BYTE(@var{expression})
4187 @kindex SHORT(@var{expression})
4188 @kindex LONG(@var{expression})
4189 @kindex QUAD(@var{expression})
4190 @kindex SQUAD(@var{expression})
4191 You can include explicit bytes of data in an output section by using
4192 @code{BYTE}, @code{SHORT}, @code{LONG}, @code{QUAD}, or @code{SQUAD} as
4193 an output section command. Each keyword is followed by an expression in
4194 parentheses providing the value to store (@pxref{Expressions}). The
4195 value of the expression is stored at the current value of the location
4196 counter.
4197
4198 The @code{BYTE}, @code{SHORT}, @code{LONG}, and @code{QUAD} commands
4199 store one, two, four, and eight bytes (respectively). After storing the
4200 bytes, the location counter is incremented by the number of bytes
4201 stored.
4202
4203 For example, this will store the byte 1 followed by the four byte value
4204 of the symbol @samp{addr}:
4205 @smallexample
4206 BYTE(1)
4207 LONG(addr)
4208 @end smallexample
4209
4210 When using a 64 bit host or target, @code{QUAD} and @code{SQUAD} are the
4211 same; they both store an 8 byte, or 64 bit, value. When both host and
4212 target are 32 bits, an expression is computed as 32 bits. In this case
4213 @code{QUAD} stores a 32 bit value zero extended to 64 bits, and
4214 @code{SQUAD} stores a 32 bit value sign extended to 64 bits.
4215
4216 If the object file format of the output file has an explicit endianness,
4217 which is the normal case, the value will be stored in that endianness.
4218 When the object file format does not have an explicit endianness, as is
4219 true of, for example, S-records, the value will be stored in the
4220 endianness of the first input object file.
4221
4222 Note---these commands only work inside a section description and not
4223 between them, so the following will produce an error from the linker:
4224 @smallexample
4225 SECTIONS @{@ .text : @{@ *(.text) @}@ LONG(1) .data : @{@ *(.data) @}@ @}@
4226 @end smallexample
4227 whereas this will work:
4228 @smallexample
4229 SECTIONS @{@ .text : @{@ *(.text) ; LONG(1) @}@ .data : @{@ *(.data) @}@ @}@
4230 @end smallexample
4231
4232 @kindex FILL(@var{expression})
4233 @cindex holes, filling
4234 @cindex unspecified memory
4235 You may use the @code{FILL} command to set the fill pattern for the
4236 current section. It is followed by an expression in parentheses. Any
4237 otherwise unspecified regions of memory within the section (for example,
4238 gaps left due to the required alignment of input sections) are filled
4239 with the value of the expression, repeated as
4240 necessary. A @code{FILL} statement covers memory locations after the
4241 point at which it occurs in the section definition; by including more
4242 than one @code{FILL} statement, you can have different fill patterns in
4243 different parts of an output section.
4244
4245 This example shows how to fill unspecified regions of memory with the
4246 value @samp{0x90}:
4247 @smallexample
4248 FILL(0x90909090)
4249 @end smallexample
4250
4251 The @code{FILL} command is similar to the @samp{=@var{fillexp}} output
4252 section attribute, but it only affects the
4253 part of the section following the @code{FILL} command, rather than the
4254 entire section. If both are used, the @code{FILL} command takes
4255 precedence. @xref{Output Section Fill}, for details on the fill
4256 expression.
4257
4258 @node Output Section Keywords
4259 @subsection Output Section Keywords
4260 There are a couple of keywords which can appear as output section
4261 commands.
4262
4263 @table @code
4264 @kindex CREATE_OBJECT_SYMBOLS
4265 @cindex input filename symbols
4266 @cindex filename symbols
4267 @item CREATE_OBJECT_SYMBOLS
4268 The command tells the linker to create a symbol for each input file.
4269 The name of each symbol will be the name of the corresponding input
4270 file. The section of each symbol will be the output section in which
4271 the @code{CREATE_OBJECT_SYMBOLS} command appears.
4272
4273 This is conventional for the a.out object file format. It is not
4274 normally used for any other object file format.
4275
4276 @kindex CONSTRUCTORS
4277 @cindex C++ constructors, arranging in link
4278 @cindex constructors, arranging in link
4279 @item CONSTRUCTORS
4280 When linking using the a.out object file format, the linker uses an
4281 unusual set construct to support C++ global constructors and
4282 destructors. When linking object file formats which do not support
4283 arbitrary sections, such as ECOFF and XCOFF, the linker will
4284 automatically recognize C++ global constructors and destructors by name.
4285 For these object file formats, the @code{CONSTRUCTORS} command tells the
4286 linker to place constructor information in the output section where the
4287 @code{CONSTRUCTORS} command appears. The @code{CONSTRUCTORS} command is
4288 ignored for other object file formats.
4289
4290 The symbol @w{@code{__CTOR_LIST__}} marks the start of the global
4291 constructors, and the symbol @w{@code{__CTOR_END__}} marks the end.
4292 Similarly, @w{@code{__DTOR_LIST__}} and @w{@code{__DTOR_END__}} mark
4293 the start and end of the global destructors. The
4294 first word in the list is the number of entries, followed by the address
4295 of each constructor or destructor, followed by a zero word. The
4296 compiler must arrange to actually run the code. For these object file
4297 formats @sc{gnu} C++ normally calls constructors from a subroutine
4298 @code{__main}; a call to @code{__main} is automatically inserted into
4299 the startup code for @code{main}. @sc{gnu} C++ normally runs
4300 destructors either by using @code{atexit}, or directly from the function
4301 @code{exit}.
4302
4303 For object file formats such as @code{COFF} or @code{ELF} which support
4304 arbitrary section names, @sc{gnu} C++ will normally arrange to put the
4305 addresses of global constructors and destructors into the @code{.ctors}
4306 and @code{.dtors} sections. Placing the following sequence into your
4307 linker script will build the sort of table which the @sc{gnu} C++
4308 runtime code expects to see.
4309
4310 @smallexample
4311 __CTOR_LIST__ = .;
4312 LONG((__CTOR_END__ - __CTOR_LIST__) / 4 - 2)
4313 *(.ctors)
4314 LONG(0)
4315 __CTOR_END__ = .;
4316 __DTOR_LIST__ = .;
4317 LONG((__DTOR_END__ - __DTOR_LIST__) / 4 - 2)
4318 *(.dtors)
4319 LONG(0)
4320 __DTOR_END__ = .;
4321 @end smallexample
4322
4323 If you are using the @sc{gnu} C++ support for initialization priority,
4324 which provides some control over the order in which global constructors
4325 are run, you must sort the constructors at link time to ensure that they
4326 are executed in the correct order. When using the @code{CONSTRUCTORS}
4327 command, use @samp{SORT_BY_NAME(CONSTRUCTORS)} instead. When using the
4328 @code{.ctors} and @code{.dtors} sections, use @samp{*(SORT_BY_NAME(.ctors))} and
4329 @samp{*(SORT_BY_NAME(.dtors))} instead of just @samp{*(.ctors)} and
4330 @samp{*(.dtors)}.
4331
4332 Normally the compiler and linker will handle these issues automatically,
4333 and you will not need to concern yourself with them. However, you may
4334 need to consider this if you are using C++ and writing your own linker
4335 scripts.
4336
4337 @end table
4338
4339 @node Output Section Discarding
4340 @subsection Output Section Discarding
4341 @cindex discarding sections
4342 @cindex sections, discarding
4343 @cindex removing sections
4344 The linker will not create output sections with no contents. This is
4345 for convenience when referring to input sections that may or may not
4346 be present in any of the input files. For example:
4347 @smallexample
4348 .foo : @{ *(.foo) @}
4349 @end smallexample
4350 @noindent
4351 will only create a @samp{.foo} section in the output file if there is a
4352 @samp{.foo} section in at least one input file, and if the input
4353 sections are not all empty. Other link script directives that allocate
4354 space in an output section will also create the output section.
4355
4356 The linker will ignore address assignments (@pxref{Output Section Address})
4357 on discarded output sections, except when the linker script defines
4358 symbols in the output section. In that case the linker will obey
4359 the address assignments, possibly advancing dot even though the
4360 section is discarded.
4361
4362 @cindex /DISCARD/
4363 The special output section name @samp{/DISCARD/} may be used to discard
4364 input sections. Any input sections which are assigned to an output
4365 section named @samp{/DISCARD/} are not included in the output file.
4366
4367 @node Output Section Attributes
4368 @subsection Output Section Attributes
4369 @cindex output section attributes
4370 We showed above that the full description of an output section looked
4371 like this:
4372
4373 @smallexample
4374 @group
4375 @var{section} [@var{address}] [(@var{type})] :
4376 [AT(@var{lma})]
4377 [ALIGN(@var{section_align})]
4378 [SUBALIGN(@var{subsection_align})]
4379 [@var{constraint}]
4380 @{
4381 @var{output-section-command}
4382 @var{output-section-command}
4383 @dots{}
4384 @} [>@var{region}] [AT>@var{lma_region}] [:@var{phdr} :@var{phdr} @dots{}] [=@var{fillexp}]
4385 @end group
4386 @end smallexample
4387
4388 We've already described @var{section}, @var{address}, and
4389 @var{output-section-command}. In this section we will describe the
4390 remaining section attributes.
4391
4392 @menu
4393 * Output Section Type:: Output section type
4394 * Output Section LMA:: Output section LMA
4395 * Forced Output Alignment:: Forced Output Alignment
4396 * Forced Input Alignment:: Forced Input Alignment
4397 * Output Section Constraint:: Output section constraint
4398 * Output Section Region:: Output section region
4399 * Output Section Phdr:: Output section phdr
4400 * Output Section Fill:: Output section fill
4401 @end menu
4402
4403 @node Output Section Type
4404 @subsubsection Output Section Type
4405 Each output section may have a type. The type is a keyword in
4406 parentheses. The following types are defined:
4407
4408 @table @code
4409 @item NOLOAD
4410 The section should be marked as not loadable, so that it will not be
4411 loaded into memory when the program is run.
4412 @item DSECT
4413 @itemx COPY
4414 @itemx INFO
4415 @itemx OVERLAY
4416 These type names are supported for backward compatibility, and are
4417 rarely used. They all have the same effect: the section should be
4418 marked as not allocatable, so that no memory is allocated for the
4419 section when the program is run.
4420 @end table
4421
4422 @kindex NOLOAD
4423 @cindex prevent unnecessary loading
4424 @cindex loading, preventing
4425 The linker normally sets the attributes of an output section based on
4426 the input sections which map into it. You can override this by using
4427 the section type. For example, in the script sample below, the
4428 @samp{ROM} section is addressed at memory location @samp{0} and does not
4429 need to be loaded when the program is run.
4430 @smallexample
4431 @group
4432 SECTIONS @{
4433 ROM 0 (NOLOAD) : @{ @dots{} @}
4434 @dots{}
4435 @}
4436 @end group
4437 @end smallexample
4438
4439 @node Output Section LMA
4440 @subsubsection Output Section LMA
4441 @kindex AT>@var{lma_region}
4442 @kindex AT(@var{lma})
4443 @cindex load address
4444 @cindex section load address
4445 Every section has a virtual address (VMA) and a load address (LMA); see
4446 @ref{Basic Script Concepts}. The virtual address is specified by the
4447 @pxref{Output Section Address} described earlier. The load address is
4448 specified by the @code{AT} or @code{AT>} keywords. Specifying a load
4449 address is optional.
4450
4451 The @code{AT} keyword takes an expression as an argument. This
4452 specifies the exact load address of the section. The @code{AT>} keyword
4453 takes the name of a memory region as an argument. @xref{MEMORY}. The
4454 load address of the section is set to the next free address in the
4455 region, aligned to the section's alignment requirements.
4456
4457 If neither @code{AT} nor @code{AT>} is specified for an allocatable
4458 section, the linker will use the following heuristic to determine the
4459 load address:
4460
4461 @itemize @bullet
4462 @item
4463 If the section has a specific VMA address, then this is used as
4464 the LMA address as well.
4465
4466 @item
4467 If the section is not allocatable then its LMA is set to its VMA.
4468
4469 @item
4470 Otherwise if a memory region can be found that is compatible
4471 with the current section, and this region contains at least one
4472 section, then the LMA is set so the difference between the
4473 VMA and LMA is the same as the difference between the VMA and LMA of
4474 the last section in the located region.
4475
4476 @item
4477 If no memory regions have been declared then a default region
4478 that covers the entire address space is used in the previous step.
4479
4480 @item
4481 If no suitable region could be found, or there was no previous
4482 section then the LMA is set equal to the VMA.
4483 @end itemize
4484
4485 @cindex ROM initialized data
4486 @cindex initialized data in ROM
4487 This feature is designed to make it easy to build a ROM image. For
4488 example, the following linker script creates three output sections: one
4489 called @samp{.text}, which starts at @code{0x1000}, one called
4490 @samp{.mdata}, which is loaded at the end of the @samp{.text} section
4491 even though its VMA is @code{0x2000}, and one called @samp{.bss} to hold
4492 uninitialized data at address @code{0x3000}. The symbol @code{_data} is
4493 defined with the value @code{0x2000}, which shows that the location
4494 counter holds the VMA value, not the LMA value.
4495
4496 @smallexample
4497 @group
4498 SECTIONS
4499 @{
4500 .text 0x1000 : @{ *(.text) _etext = . ; @}
4501 .mdata 0x2000 :
4502 AT ( ADDR (.text) + SIZEOF (.text) )
4503 @{ _data = . ; *(.data); _edata = . ; @}
4504 .bss 0x3000 :
4505 @{ _bstart = . ; *(.bss) *(COMMON) ; _bend = . ;@}
4506 @}
4507 @end group
4508 @end smallexample
4509
4510 The run-time initialization code for use with a program generated with
4511 this linker script would include something like the following, to copy
4512 the initialized data from the ROM image to its runtime address. Notice
4513 how this code takes advantage of the symbols defined by the linker
4514 script.
4515
4516 @smallexample
4517 @group
4518 extern char _etext, _data, _edata, _bstart, _bend;
4519 char *src = &_etext;
4520 char *dst = &_data;
4521
4522 /* ROM has data at end of text; copy it. */
4523 while (dst < &_edata)
4524 *dst++ = *src++;
4525
4526 /* Zero bss. */
4527 for (dst = &_bstart; dst< &_bend; dst++)
4528 *dst = 0;
4529 @end group
4530 @end smallexample
4531
4532 @node Forced Output Alignment
4533 @subsubsection Forced Output Alignment
4534 @kindex ALIGN(@var{section_align})
4535 @cindex forcing output section alignment
4536 @cindex output section alignment
4537 You can increase an output section's alignment by using ALIGN.
4538
4539 @node Forced Input Alignment
4540 @subsubsection Forced Input Alignment
4541 @kindex SUBALIGN(@var{subsection_align})
4542 @cindex forcing input section alignment
4543 @cindex input section alignment
4544 You can force input section alignment within an output section by using
4545 SUBALIGN. The value specified overrides any alignment given by input
4546 sections, whether larger or smaller.
4547
4548 @node Output Section Constraint
4549 @subsubsection Output Section Constraint
4550 @kindex ONLY_IF_RO
4551 @kindex ONLY_IF_RW
4552 @cindex constraints on output sections
4553 You can specify that an output section should only be created if all
4554 of its input sections are read-only or all of its input sections are
4555 read-write by using the keyword @code{ONLY_IF_RO} and
4556 @code{ONLY_IF_RW} respectively.
4557
4558 @node Output Section Region
4559 @subsubsection Output Section Region
4560 @kindex >@var{region}
4561 @cindex section, assigning to memory region
4562 @cindex memory regions and sections
4563 You can assign a section to a previously defined region of memory by
4564 using @samp{>@var{region}}. @xref{MEMORY}.
4565
4566 Here is a simple example:
4567 @smallexample
4568 @group
4569 MEMORY @{ rom : ORIGIN = 0x1000, LENGTH = 0x1000 @}
4570 SECTIONS @{ ROM : @{ *(.text) @} >rom @}
4571 @end group
4572 @end smallexample
4573
4574 @node Output Section Phdr
4575 @subsubsection Output Section Phdr
4576 @kindex :@var{phdr}
4577 @cindex section, assigning to program header
4578 @cindex program headers and sections
4579 You can assign a section to a previously defined program segment by
4580 using @samp{:@var{phdr}}. @xref{PHDRS}. If a section is assigned to
4581 one or more segments, then all subsequent allocated sections will be
4582 assigned to those segments as well, unless they use an explicitly
4583 @code{:@var{phdr}} modifier. You can use @code{:NONE} to tell the
4584 linker to not put the section in any segment at all.
4585
4586 Here is a simple example:
4587 @smallexample
4588 @group
4589 PHDRS @{ text PT_LOAD ; @}
4590 SECTIONS @{ .text : @{ *(.text) @} :text @}
4591 @end group
4592 @end smallexample
4593
4594 @node Output Section Fill
4595 @subsubsection Output Section Fill
4596 @kindex =@var{fillexp}
4597 @cindex section fill pattern
4598 @cindex fill pattern, entire section
4599 You can set the fill pattern for an entire section by using
4600 @samp{=@var{fillexp}}. @var{fillexp} is an expression
4601 (@pxref{Expressions}). Any otherwise unspecified regions of memory
4602 within the output section (for example, gaps left due to the required
4603 alignment of input sections) will be filled with the value, repeated as
4604 necessary. If the fill expression is a simple hex number, ie. a string
4605 of hex digit starting with @samp{0x} and without a trailing @samp{k} or @samp{M}, then
4606 an arbitrarily long sequence of hex digits can be used to specify the
4607 fill pattern; Leading zeros become part of the pattern too. For all
4608 other cases, including extra parentheses or a unary @code{+}, the fill
4609 pattern is the four least significant bytes of the value of the
4610 expression. In all cases, the number is big-endian.
4611
4612 You can also change the fill value with a @code{FILL} command in the
4613 output section commands; (@pxref{Output Section Data}).
4614
4615 Here is a simple example:
4616 @smallexample
4617 @group
4618 SECTIONS @{ .text : @{ *(.text) @} =0x90909090 @}
4619 @end group
4620 @end smallexample
4621
4622 @node Overlay Description
4623 @subsection Overlay Description
4624 @kindex OVERLAY
4625 @cindex overlays
4626 An overlay description provides an easy way to describe sections which
4627 are to be loaded as part of a single memory image but are to be run at
4628 the same memory address. At run time, some sort of overlay manager will
4629 copy the overlaid sections in and out of the runtime memory address as
4630 required, perhaps by simply manipulating addressing bits. This approach
4631 can be useful, for example, when a certain region of memory is faster
4632 than another.
4633
4634 Overlays are described using the @code{OVERLAY} command. The
4635 @code{OVERLAY} command is used within a @code{SECTIONS} command, like an
4636 output section description. The full syntax of the @code{OVERLAY}
4637 command is as follows:
4638 @smallexample
4639 @group
4640 OVERLAY [@var{start}] : [NOCROSSREFS] [AT ( @var{ldaddr} )]
4641 @{
4642 @var{secname1}
4643 @{
4644 @var{output-section-command}
4645 @var{output-section-command}
4646 @dots{}
4647 @} [:@var{phdr}@dots{}] [=@var{fill}]
4648 @var{secname2}
4649 @{
4650 @var{output-section-command}
4651 @var{output-section-command}
4652 @dots{}
4653 @} [:@var{phdr}@dots{}] [=@var{fill}]
4654 @dots{}
4655 @} [>@var{region}] [:@var{phdr}@dots{}] [=@var{fill}]
4656 @end group
4657 @end smallexample
4658
4659 Everything is optional except @code{OVERLAY} (a keyword), and each
4660 section must have a name (@var{secname1} and @var{secname2} above). The
4661 section definitions within the @code{OVERLAY} construct are identical to
4662 those within the general @code{SECTIONS} contruct (@pxref{SECTIONS}),
4663 except that no addresses and no memory regions may be defined for
4664 sections within an @code{OVERLAY}.
4665
4666 The sections are all defined with the same starting address. The load
4667 addresses of the sections are arranged such that they are consecutive in
4668 memory starting at the load address used for the @code{OVERLAY} as a
4669 whole (as with normal section definitions, the load address is optional,
4670 and defaults to the start address; the start address is also optional,
4671 and defaults to the current value of the location counter).
4672
4673 If the @code{NOCROSSREFS} keyword is used, and there any references
4674 among the sections, the linker will report an error. Since the sections
4675 all run at the same address, it normally does not make sense for one
4676 section to refer directly to another. @xref{Miscellaneous Commands,
4677 NOCROSSREFS}.
4678
4679 For each section within the @code{OVERLAY}, the linker automatically
4680 provides two symbols. The symbol @code{__load_start_@var{secname}} is
4681 defined as the starting load address of the section. The symbol
4682 @code{__load_stop_@var{secname}} is defined as the final load address of
4683 the section. Any characters within @var{secname} which are not legal
4684 within C identifiers are removed. C (or assembler) code may use these
4685 symbols to move the overlaid sections around as necessary.
4686
4687 At the end of the overlay, the value of the location counter is set to
4688 the start address of the overlay plus the size of the largest section.
4689
4690 Here is an example. Remember that this would appear inside a
4691 @code{SECTIONS} construct.
4692 @smallexample
4693 @group
4694 OVERLAY 0x1000 : AT (0x4000)
4695 @{
4696 .text0 @{ o1/*.o(.text) @}
4697 .text1 @{ o2/*.o(.text) @}
4698 @}
4699 @end group
4700 @end smallexample
4701 @noindent
4702 This will define both @samp{.text0} and @samp{.text1} to start at
4703 address 0x1000. @samp{.text0} will be loaded at address 0x4000, and
4704 @samp{.text1} will be loaded immediately after @samp{.text0}. The
4705 following symbols will be defined if referenced: @code{__load_start_text0},
4706 @code{__load_stop_text0}, @code{__load_start_text1},
4707 @code{__load_stop_text1}.
4708
4709 C code to copy overlay @code{.text1} into the overlay area might look
4710 like the following.
4711
4712 @smallexample
4713 @group
4714 extern char __load_start_text1, __load_stop_text1;
4715 memcpy ((char *) 0x1000, &__load_start_text1,
4716 &__load_stop_text1 - &__load_start_text1);
4717 @end group
4718 @end smallexample
4719
4720 Note that the @code{OVERLAY} command is just syntactic sugar, since
4721 everything it does can be done using the more basic commands. The above
4722 example could have been written identically as follows.
4723
4724 @smallexample
4725 @group
4726 .text0 0x1000 : AT (0x4000) @{ o1/*.o(.text) @}
4727 PROVIDE (__load_start_text0 = LOADADDR (.text0));
4728 PROVIDE (__load_stop_text0 = LOADADDR (.text0) + SIZEOF (.text0));
4729 .text1 0x1000 : AT (0x4000 + SIZEOF (.text0)) @{ o2/*.o(.text) @}
4730 PROVIDE (__load_start_text1 = LOADADDR (.text1));
4731 PROVIDE (__load_stop_text1 = LOADADDR (.text1) + SIZEOF (.text1));
4732 . = 0x1000 + MAX (SIZEOF (.text0), SIZEOF (.text1));
4733 @end group
4734 @end smallexample
4735
4736 @node MEMORY
4737 @section MEMORY Command
4738 @kindex MEMORY
4739 @cindex memory regions
4740 @cindex regions of memory
4741 @cindex allocating memory
4742 @cindex discontinuous memory
4743 The linker's default configuration permits allocation of all available
4744 memory. You can override this by using the @code{MEMORY} command.
4745
4746 The @code{MEMORY} command describes the location and size of blocks of
4747 memory in the target. You can use it to describe which memory regions
4748 may be used by the linker, and which memory regions it must avoid. You
4749 can then assign sections to particular memory regions. The linker will
4750 set section addresses based on the memory regions, and will warn about
4751 regions that become too full. The linker will not shuffle sections
4752 around to fit into the available regions.
4753
4754 A linker script may contain at most one use of the @code{MEMORY}
4755 command. However, you can define as many blocks of memory within it as
4756 you wish. The syntax is:
4757 @smallexample
4758 @group
4759 MEMORY
4760 @{
4761 @var{name} [(@var{attr})] : ORIGIN = @var{origin}, LENGTH = @var{len}
4762 @dots{}
4763 @}
4764 @end group
4765 @end smallexample
4766
4767 The @var{name} is a name used in the linker script to refer to the
4768 region. The region name has no meaning outside of the linker script.
4769 Region names are stored in a separate name space, and will not conflict
4770 with symbol names, file names, or section names. Each memory region
4771 must have a distinct name within the @code{MEMORY} command. However you can
4772 add later alias names to existing memory regions with the @ref{REGION_ALIAS}
4773 command.
4774
4775 @cindex memory region attributes
4776 The @var{attr} string is an optional list of attributes that specify
4777 whether to use a particular memory region for an input section which is
4778 not explicitly mapped in the linker script. As described in
4779 @ref{SECTIONS}, if you do not specify an output section for some input
4780 section, the linker will create an output section with the same name as
4781 the input section. If you define region attributes, the linker will use
4782 them to select the memory region for the output section that it creates.
4783
4784 The @var{attr} string must consist only of the following characters:
4785 @table @samp
4786 @item R
4787 Read-only section
4788 @item W
4789 Read/write section
4790 @item X
4791 Executable section
4792 @item A
4793 Allocatable section
4794 @item I
4795 Initialized section
4796 @item L
4797 Same as @samp{I}
4798 @item !
4799 Invert the sense of any of the attributes that follow
4800 @end table
4801
4802 If a unmapped section matches any of the listed attributes other than
4803 @samp{!}, it will be placed in the memory region. The @samp{!}
4804 attribute reverses this test, so that an unmapped section will be placed
4805 in the memory region only if it does not match any of the listed
4806 attributes.
4807
4808 @kindex ORIGIN =
4809 @kindex o =
4810 @kindex org =
4811 The @var{origin} is an numerical expression for the start address of
4812 the memory region. The expression must evaluate to a constant and it
4813 cannot involve any symbols. The keyword @code{ORIGIN} may be
4814 abbreviated to @code{org} or @code{o} (but not, for example,
4815 @code{ORG}).
4816
4817 @kindex LENGTH =
4818 @kindex len =
4819 @kindex l =
4820 The @var{len} is an expression for the size in bytes of the memory
4821 region. As with the @var{origin} expression, the expression must
4822 be numerical only and must evaluate to a constant. The keyword
4823 @code{LENGTH} may be abbreviated to @code{len} or @code{l}.
4824
4825 In the following example, we specify that there are two memory regions
4826 available for allocation: one starting at @samp{0} for 256 kilobytes,
4827 and the other starting at @samp{0x40000000} for four megabytes. The
4828 linker will place into the @samp{rom} memory region every section which
4829 is not explicitly mapped into a memory region, and is either read-only
4830 or executable. The linker will place other sections which are not
4831 explicitly mapped into a memory region into the @samp{ram} memory
4832 region.
4833
4834 @smallexample
4835 @group
4836 MEMORY
4837 @{
4838 rom (rx) : ORIGIN = 0, LENGTH = 256K
4839 ram (!rx) : org = 0x40000000, l = 4M
4840 @}
4841 @end group
4842 @end smallexample
4843
4844 Once you define a memory region, you can direct the linker to place
4845 specific output sections into that memory region by using the
4846 @samp{>@var{region}} output section attribute. For example, if you have
4847 a memory region named @samp{mem}, you would use @samp{>mem} in the
4848 output section definition. @xref{Output Section Region}. If no address
4849 was specified for the output section, the linker will set the address to
4850 the next available address within the memory region. If the combined
4851 output sections directed to a memory region are too large for the
4852 region, the linker will issue an error message.
4853
4854 It is possible to access the origin and length of a memory in an
4855 expression via the @code{ORIGIN(@var{memory})} and
4856 @code{LENGTH(@var{memory})} functions:
4857
4858 @smallexample
4859 @group
4860 _fstack = ORIGIN(ram) + LENGTH(ram) - 4;
4861 @end group
4862 @end smallexample
4863
4864 @node PHDRS
4865 @section PHDRS Command
4866 @kindex PHDRS
4867 @cindex program headers
4868 @cindex ELF program headers
4869 @cindex program segments
4870 @cindex segments, ELF
4871 The ELF object file format uses @dfn{program headers}, also knows as
4872 @dfn{segments}. The program headers describe how the program should be
4873 loaded into memory. You can print them out by using the @code{objdump}
4874 program with the @samp{-p} option.
4875
4876 When you run an ELF program on a native ELF system, the system loader
4877 reads the program headers in order to figure out how to load the
4878 program. This will only work if the program headers are set correctly.
4879 This manual does not describe the details of how the system loader
4880 interprets program headers; for more information, see the ELF ABI.
4881
4882 The linker will create reasonable program headers by default. However,
4883 in some cases, you may need to specify the program headers more
4884 precisely. You may use the @code{PHDRS} command for this purpose. When
4885 the linker sees the @code{PHDRS} command in the linker script, it will
4886 not create any program headers other than the ones specified.
4887
4888 The linker only pays attention to the @code{PHDRS} command when
4889 generating an ELF output file. In other cases, the linker will simply
4890 ignore @code{PHDRS}.
4891
4892 This is the syntax of the @code{PHDRS} command. The words @code{PHDRS},
4893 @code{FILEHDR}, @code{AT}, and @code{FLAGS} are keywords.
4894
4895 @smallexample
4896 @group
4897 PHDRS
4898 @{
4899 @var{name} @var{type} [ FILEHDR ] [ PHDRS ] [ AT ( @var{address} ) ]
4900 [ FLAGS ( @var{flags} ) ] ;
4901 @}
4902 @end group
4903 @end smallexample
4904
4905 The @var{name} is used only for reference in the @code{SECTIONS} command
4906 of the linker script. It is not put into the output file. Program
4907 header names are stored in a separate name space, and will not conflict
4908 with symbol names, file names, or section names. Each program header
4909 must have a distinct name. The headers are processed in order and it
4910 is usual for them to map to sections in ascending load address order.
4911
4912 Certain program header types describe segments of memory which the
4913 system loader will load from the file. In the linker script, you
4914 specify the contents of these segments by placing allocatable output
4915 sections in the segments. You use the @samp{:@var{phdr}} output section
4916 attribute to place a section in a particular segment. @xref{Output
4917 Section Phdr}.
4918
4919 It is normal to put certain sections in more than one segment. This
4920 merely implies that one segment of memory contains another. You may
4921 repeat @samp{:@var{phdr}}, using it once for each segment which should
4922 contain the section.
4923
4924 If you place a section in one or more segments using @samp{:@var{phdr}},
4925 then the linker will place all subsequent allocatable sections which do
4926 not specify @samp{:@var{phdr}} in the same segments. This is for
4927 convenience, since generally a whole set of contiguous sections will be
4928 placed in a single segment. You can use @code{:NONE} to override the
4929 default segment and tell the linker to not put the section in any
4930 segment at all.
4931
4932 @kindex FILEHDR
4933 @kindex PHDRS
4934 You may use the @code{FILEHDR} and @code{PHDRS} keywords after
4935 the program header type to further describe the contents of the segment.
4936 The @code{FILEHDR} keyword means that the segment should include the ELF
4937 file header. The @code{PHDRS} keyword means that the segment should
4938 include the ELF program headers themselves. If applied to a loadable
4939 segment (@code{PT_LOAD}), all prior loadable segments must have one of
4940 these keywords.
4941
4942 The @var{type} may be one of the following. The numbers indicate the
4943 value of the keyword.
4944
4945 @table @asis
4946 @item @code{PT_NULL} (0)
4947 Indicates an unused program header.
4948
4949 @item @code{PT_LOAD} (1)
4950 Indicates that this program header describes a segment to be loaded from
4951 the file.
4952
4953 @item @code{PT_DYNAMIC} (2)
4954 Indicates a segment where dynamic linking information can be found.
4955
4956 @item @code{PT_INTERP} (3)
4957 Indicates a segment where the name of the program interpreter may be
4958 found.
4959
4960 @item @code{PT_NOTE} (4)
4961 Indicates a segment holding note information.
4962
4963 @item @code{PT_SHLIB} (5)
4964 A reserved program header type, defined but not specified by the ELF
4965 ABI.
4966
4967 @item @code{PT_PHDR} (6)
4968 Indicates a segment where the program headers may be found.
4969
4970 @item @var{expression}
4971 An expression giving the numeric type of the program header. This may
4972 be used for types not defined above.
4973 @end table
4974
4975 You can specify that a segment should be loaded at a particular address
4976 in memory by using an @code{AT} expression. This is identical to the
4977 @code{AT} command used as an output section attribute (@pxref{Output
4978 Section LMA}). The @code{AT} command for a program header overrides the
4979 output section attribute.
4980
4981 The linker will normally set the segment flags based on the sections
4982 which comprise the segment. You may use the @code{FLAGS} keyword to
4983 explicitly specify the segment flags. The value of @var{flags} must be
4984 an integer. It is used to set the @code{p_flags} field of the program
4985 header.
4986
4987 Here is an example of @code{PHDRS}. This shows a typical set of program
4988 headers used on a native ELF system.
4989
4990 @example
4991 @group
4992 PHDRS
4993 @{
4994 headers PT_PHDR PHDRS ;
4995 interp PT_INTERP ;
4996 text PT_LOAD FILEHDR PHDRS ;
4997 data PT_LOAD ;
4998 dynamic PT_DYNAMIC ;
4999 @}
5000
5001 SECTIONS
5002 @{
5003 . = SIZEOF_HEADERS;
5004 .interp : @{ *(.interp) @} :text :interp
5005 .text : @{ *(.text) @} :text
5006 .rodata : @{ *(.rodata) @} /* defaults to :text */
5007 @dots{}
5008 . = . + 0x1000; /* move to a new page in memory */
5009 .data : @{ *(.data) @} :data
5010 .dynamic : @{ *(.dynamic) @} :data :dynamic
5011 @dots{}
5012 @}
5013 @end group
5014 @end example
5015
5016 @node VERSION
5017 @section VERSION Command
5018 @kindex VERSION @{script text@}
5019 @cindex symbol versions
5020 @cindex version script
5021 @cindex versions of symbols
5022 The linker supports symbol versions when using ELF. Symbol versions are
5023 only useful when using shared libraries. The dynamic linker can use
5024 symbol versions to select a specific version of a function when it runs
5025 a program that may have been linked against an earlier version of the
5026 shared library.
5027
5028 You can include a version script directly in the main linker script, or
5029 you can supply the version script as an implicit linker script. You can
5030 also use the @samp{--version-script} linker option.
5031
5032 The syntax of the @code{VERSION} command is simply
5033 @smallexample
5034 VERSION @{ version-script-commands @}
5035 @end smallexample
5036
5037 The format of the version script commands is identical to that used by
5038 Sun's linker in Solaris 2.5. The version script defines a tree of
5039 version nodes. You specify the node names and interdependencies in the
5040 version script. You can specify which symbols are bound to which
5041 version nodes, and you can reduce a specified set of symbols to local
5042 scope so that they are not globally visible outside of the shared
5043 library.
5044
5045 The easiest way to demonstrate the version script language is with a few
5046 examples.
5047
5048 @smallexample
5049 VERS_1.1 @{
5050 global:
5051 foo1;
5052 local:
5053 old*;
5054 original*;
5055 new*;
5056 @};
5057
5058 VERS_1.2 @{
5059 foo2;
5060 @} VERS_1.1;
5061
5062 VERS_2.0 @{
5063 bar1; bar2;
5064 extern "C++" @{
5065 ns::*;
5066 "f(int, double)";
5067 @};
5068 @} VERS_1.2;
5069 @end smallexample
5070
5071 This example version script defines three version nodes. The first
5072 version node defined is @samp{VERS_1.1}; it has no other dependencies.
5073 The script binds the symbol @samp{foo1} to @samp{VERS_1.1}. It reduces
5074 a number of symbols to local scope so that they are not visible outside
5075 of the shared library; this is done using wildcard patterns, so that any
5076 symbol whose name begins with @samp{old}, @samp{original}, or @samp{new}
5077 is matched. The wildcard patterns available are the same as those used
5078 in the shell when matching filenames (also known as ``globbing'').
5079 However, if you specify the symbol name inside double quotes, then the
5080 name is treated as literal, rather than as a glob pattern.
5081
5082 Next, the version script defines node @samp{VERS_1.2}. This node
5083 depends upon @samp{VERS_1.1}. The script binds the symbol @samp{foo2}
5084 to the version node @samp{VERS_1.2}.
5085
5086 Finally, the version script defines node @samp{VERS_2.0}. This node
5087 depends upon @samp{VERS_1.2}. The scripts binds the symbols @samp{bar1}
5088 and @samp{bar2} are bound to the version node @samp{VERS_2.0}.
5089
5090 When the linker finds a symbol defined in a library which is not
5091 specifically bound to a version node, it will effectively bind it to an
5092 unspecified base version of the library. You can bind all otherwise
5093 unspecified symbols to a given version node by using @samp{global: *;}
5094 somewhere in the version script. Note that it's slightly crazy to use
5095 wildcards in a global spec except on the last version node. Global
5096 wildcards elsewhere run the risk of accidentally adding symbols to the
5097 set exported for an old version. That's wrong since older versions
5098 ought to have a fixed set of symbols.
5099
5100 The names of the version nodes have no specific meaning other than what
5101 they might suggest to the person reading them. The @samp{2.0} version
5102 could just as well have appeared in between @samp{1.1} and @samp{1.2}.
5103 However, this would be a confusing way to write a version script.
5104
5105 Node name can be omitted, provided it is the only version node
5106 in the version script. Such version script doesn't assign any versions to
5107 symbols, only selects which symbols will be globally visible out and which
5108 won't.
5109
5110 @smallexample
5111 @{ global: foo; bar; local: *; @};
5112 @end smallexample
5113
5114 When you link an application against a shared library that has versioned
5115 symbols, the application itself knows which version of each symbol it
5116 requires, and it also knows which version nodes it needs from each
5117 shared library it is linked against. Thus at runtime, the dynamic
5118 loader can make a quick check to make sure that the libraries you have
5119 linked against do in fact supply all of the version nodes that the
5120 application will need to resolve all of the dynamic symbols. In this
5121 way it is possible for the dynamic linker to know with certainty that
5122 all external symbols that it needs will be resolvable without having to
5123 search for each symbol reference.
5124
5125 The symbol versioning is in effect a much more sophisticated way of
5126 doing minor version checking that SunOS does. The fundamental problem
5127 that is being addressed here is that typically references to external
5128 functions are bound on an as-needed basis, and are not all bound when
5129 the application starts up. If a shared library is out of date, a
5130 required interface may be missing; when the application tries to use
5131 that interface, it may suddenly and unexpectedly fail. With symbol
5132 versioning, the user will get a warning when they start their program if
5133 the libraries being used with the application are too old.
5134
5135 There are several GNU extensions to Sun's versioning approach. The
5136 first of these is the ability to bind a symbol to a version node in the
5137 source file where the symbol is defined instead of in the versioning
5138 script. This was done mainly to reduce the burden on the library
5139 maintainer. You can do this by putting something like:
5140 @smallexample
5141 __asm__(".symver original_foo,foo@@VERS_1.1");
5142 @end smallexample
5143 @noindent
5144 in the C source file. This renames the function @samp{original_foo} to
5145 be an alias for @samp{foo} bound to the version node @samp{VERS_1.1}.
5146 The @samp{local:} directive can be used to prevent the symbol
5147 @samp{original_foo} from being exported. A @samp{.symver} directive
5148 takes precedence over a version script.
5149
5150 The second GNU extension is to allow multiple versions of the same
5151 function to appear in a given shared library. In this way you can make
5152 an incompatible change to an interface without increasing the major
5153 version number of the shared library, while still allowing applications
5154 linked against the old interface to continue to function.
5155
5156 To do this, you must use multiple @samp{.symver} directives in the
5157 source file. Here is an example:
5158
5159 @smallexample
5160 __asm__(".symver original_foo,foo@@");
5161 __asm__(".symver old_foo,foo@@VERS_1.1");
5162 __asm__(".symver old_foo1,foo@@VERS_1.2");
5163 __asm__(".symver new_foo,foo@@@@VERS_2.0");
5164 @end smallexample
5165
5166 In this example, @samp{foo@@} represents the symbol @samp{foo} bound to the
5167 unspecified base version of the symbol. The source file that contains this
5168 example would define 4 C functions: @samp{original_foo}, @samp{old_foo},
5169 @samp{old_foo1}, and @samp{new_foo}.
5170
5171 When you have multiple definitions of a given symbol, there needs to be
5172 some way to specify a default version to which external references to
5173 this symbol will be bound. You can do this with the
5174 @samp{foo@@@@VERS_2.0} type of @samp{.symver} directive. You can only
5175 declare one version of a symbol as the default in this manner; otherwise
5176 you would effectively have multiple definitions of the same symbol.
5177
5178 If you wish to bind a reference to a specific version of the symbol
5179 within the shared library, you can use the aliases of convenience
5180 (i.e., @samp{old_foo}), or you can use the @samp{.symver} directive to
5181 specifically bind to an external version of the function in question.
5182
5183 You can also specify the language in the version script:
5184
5185 @smallexample
5186 VERSION extern "lang" @{ version-script-commands @}
5187 @end smallexample
5188
5189 The supported @samp{lang}s are @samp{C}, @samp{C++}, and @samp{Java}.
5190 The linker will iterate over the list of symbols at the link time and
5191 demangle them according to @samp{lang} before matching them to the
5192 patterns specified in @samp{version-script-commands}. The default
5193 @samp{lang} is @samp{C}.
5194
5195 Demangled names may contains spaces and other special characters. As
5196 described above, you can use a glob pattern to match demangled names,
5197 or you can use a double-quoted string to match the string exactly. In
5198 the latter case, be aware that minor differences (such as differing
5199 whitespace) between the version script and the demangler output will
5200 cause a mismatch. As the exact string generated by the demangler
5201 might change in the future, even if the mangled name does not, you
5202 should check that all of your version directives are behaving as you
5203 expect when you upgrade.
5204
5205 @node Expressions
5206 @section Expressions in Linker Scripts
5207 @cindex expressions
5208 @cindex arithmetic
5209 The syntax for expressions in the linker script language is identical to
5210 that of C expressions. All expressions are evaluated as integers. All
5211 expressions are evaluated in the same size, which is 32 bits if both the
5212 host and target are 32 bits, and is otherwise 64 bits.
5213
5214 You can use and set symbol values in expressions.
5215
5216 The linker defines several special purpose builtin functions for use in
5217 expressions.
5218
5219 @menu
5220 * Constants:: Constants
5221 * Symbolic Constants:: Symbolic constants
5222 * Symbols:: Symbol Names
5223 * Orphan Sections:: Orphan Sections
5224 * Location Counter:: The Location Counter
5225 * Operators:: Operators
5226 * Evaluation:: Evaluation
5227 * Expression Section:: The Section of an Expression
5228 * Builtin Functions:: Builtin Functions
5229 @end menu
5230
5231 @node Constants
5232 @subsection Constants
5233 @cindex integer notation
5234 @cindex constants in linker scripts
5235 All constants are integers.
5236
5237 As in C, the linker considers an integer beginning with @samp{0} to be
5238 octal, and an integer beginning with @samp{0x} or @samp{0X} to be
5239 hexadecimal. Alternatively the linker accepts suffixes of @samp{h} or
5240 @samp{H} for hexadeciaml, @samp{o} or @samp{O} for octal, @samp{b} or
5241 @samp{B} for binary and @samp{d} or @samp{D} for decimal. Any integer
5242 value without a prefix or a suffix is considered to be decimal.
5243
5244 @cindex scaled integers
5245 @cindex K and M integer suffixes
5246 @cindex M and K integer suffixes
5247 @cindex suffixes for integers
5248 @cindex integer suffixes
5249 In addition, you can use the suffixes @code{K} and @code{M} to scale a
5250 constant by
5251 @c TEXI2ROFF-KILL
5252 @ifnottex
5253 @c END TEXI2ROFF-KILL
5254 @code{1024} or @code{1024*1024}
5255 @c TEXI2ROFF-KILL
5256 @end ifnottex
5257 @tex
5258 ${\rm 1024}$ or ${\rm 1024}^2$
5259 @end tex
5260 @c END TEXI2ROFF-KILL
5261 respectively. For example, the following
5262 all refer to the same quantity:
5263
5264 @smallexample
5265 _fourk_1 = 4K;
5266 _fourk_2 = 4096;
5267 _fourk_3 = 0x1000;
5268 _fourk_4 = 10000o;
5269 @end smallexample
5270
5271 Note - the @code{K} and @code{M} suffixes cannot be used in
5272 conjunction with the base suffixes mentioned above.
5273
5274 @node Symbolic Constants
5275 @subsection Symbolic Constants
5276 @cindex symbolic constants
5277 @kindex CONSTANT
5278 It is possible to refer to target specific constants via the use of
5279 the @code{CONSTANT(@var{name})} operator, where @var{name} is one of:
5280
5281 @table @code
5282 @item MAXPAGESIZE
5283 @kindex MAXPAGESIZE
5284 The target's maximum page size.
5285
5286 @item COMMONPAGESIZE
5287 @kindex COMMONPAGESIZE
5288 The target's default page size.
5289 @end table
5290
5291 So for example:
5292
5293 @smallexample
5294 .text ALIGN (CONSTANT (MAXPAGESIZE)) : @{ *(.text) @}
5295 @end smallexample
5296
5297 will create a text section aligned to the largest page boundary
5298 supported by the target.
5299
5300 @node Symbols
5301 @subsection Symbol Names
5302 @cindex symbol names
5303 @cindex names
5304 @cindex quoted symbol names
5305 @kindex "
5306 Unless quoted, symbol names start with a letter, underscore, or period
5307 and may include letters, digits, underscores, periods, and hyphens.
5308 Unquoted symbol names must not conflict with any keywords. You can
5309 specify a symbol which contains odd characters or has the same name as a
5310 keyword by surrounding the symbol name in double quotes:
5311 @smallexample
5312 "SECTION" = 9;
5313 "with a space" = "also with a space" + 10;
5314 @end smallexample
5315
5316 Since symbols can contain many non-alphabetic characters, it is safest
5317 to delimit symbols with spaces. For example, @samp{A-B} is one symbol,
5318 whereas @samp{A - B} is an expression involving subtraction.
5319
5320 @node Orphan Sections
5321 @subsection Orphan Sections
5322 @cindex orphan
5323 Orphan sections are sections present in the input files which
5324 are not explicitly placed into the output file by the linker
5325 script. The linker will still copy these sections into the
5326 output file, but it has to guess as to where they should be
5327 placed. The linker uses a simple heuristic to do this. It
5328 attempts to place orphan sections after non-orphan sections of the
5329 same attribute, such as code vs data, loadable vs non-loadable, etc.
5330 If there is not enough room to do this then it places
5331 at the end of the file.
5332
5333 For ELF targets, the attribute of the section includes section type as
5334 well as section flag.
5335
5336 If an orphaned section's name is representable as a C identifier then
5337 the linker will automatically @pxref{PROVIDE} two symbols:
5338 __start_SECNAME and __stop_SECNAME, where SECNAME is the name of the
5339 section. These indicate the start address and end address of the
5340 orphaned section respectively. Note: most section names are not
5341 representable as C identifiers because they contain a @samp{.}
5342 character.
5343
5344 @node Location Counter
5345 @subsection The Location Counter
5346 @kindex .
5347 @cindex dot
5348 @cindex location counter
5349 @cindex current output location
5350 The special linker variable @dfn{dot} @samp{.} always contains the
5351 current output location counter. Since the @code{.} always refers to a
5352 location in an output section, it may only appear in an expression
5353 within a @code{SECTIONS} command. The @code{.} symbol may appear
5354 anywhere that an ordinary symbol is allowed in an expression.
5355
5356 @cindex holes
5357 Assigning a value to @code{.} will cause the location counter to be
5358 moved. This may be used to create holes in the output section. The
5359 location counter may not be moved backwards inside an output section,
5360 and may not be moved backwards outside of an output section if so
5361 doing creates areas with overlapping LMAs.
5362
5363 @smallexample
5364 SECTIONS
5365 @{
5366 output :
5367 @{
5368 file1(.text)
5369 . = . + 1000;
5370 file2(.text)
5371 . += 1000;
5372 file3(.text)
5373 @} = 0x12345678;
5374 @}
5375 @end smallexample
5376 @noindent
5377 In the previous example, the @samp{.text} section from @file{file1} is
5378 located at the beginning of the output section @samp{output}. It is
5379 followed by a 1000 byte gap. Then the @samp{.text} section from
5380 @file{file2} appears, also with a 1000 byte gap following before the
5381 @samp{.text} section from @file{file3}. The notation @samp{= 0x12345678}
5382 specifies what data to write in the gaps (@pxref{Output Section Fill}).
5383
5384 @cindex dot inside sections
5385 Note: @code{.} actually refers to the byte offset from the start of the
5386 current containing object. Normally this is the @code{SECTIONS}
5387 statement, whose start address is 0, hence @code{.} can be used as an
5388 absolute address. If @code{.} is used inside a section description
5389 however, it refers to the byte offset from the start of that section,
5390 not an absolute address. Thus in a script like this:
5391
5392 @smallexample
5393 SECTIONS
5394 @{
5395 . = 0x100
5396 .text: @{
5397 *(.text)
5398 . = 0x200
5399 @}
5400 . = 0x500
5401 .data: @{
5402 *(.data)
5403 . += 0x600
5404 @}
5405 @}
5406 @end smallexample
5407
5408 The @samp{.text} section will be assigned a starting address of 0x100
5409 and a size of exactly 0x200 bytes, even if there is not enough data in
5410 the @samp{.text} input sections to fill this area. (If there is too
5411 much data, an error will be produced because this would be an attempt to
5412 move @code{.} backwards). The @samp{.data} section will start at 0x500
5413 and it will have an extra 0x600 bytes worth of space after the end of
5414 the values from the @samp{.data} input sections and before the end of
5415 the @samp{.data} output section itself.
5416
5417 @cindex dot outside sections
5418 Setting symbols to the value of the location counter outside of an
5419 output section statement can result in unexpected values if the linker
5420 needs to place orphan sections. For example, given the following:
5421
5422 @smallexample
5423 SECTIONS
5424 @{
5425 start_of_text = . ;
5426 .text: @{ *(.text) @}
5427 end_of_text = . ;
5428
5429 start_of_data = . ;
5430 .data: @{ *(.data) @}
5431 end_of_data = . ;
5432 @}
5433 @end smallexample
5434
5435 If the linker needs to place some input section, e.g. @code{.rodata},
5436 not mentioned in the script, it might choose to place that section
5437 between @code{.text} and @code{.data}. You might think the linker
5438 should place @code{.rodata} on the blank line in the above script, but
5439 blank lines are of no particular significance to the linker. As well,
5440 the linker doesn't associate the above symbol names with their
5441 sections. Instead, it assumes that all assignments or other
5442 statements belong to the previous output section, except for the
5443 special case of an assignment to @code{.}. I.e., the linker will
5444 place the orphan @code{.rodata} section as if the script was written
5445 as follows:
5446
5447 @smallexample
5448 SECTIONS
5449 @{
5450 start_of_text = . ;
5451 .text: @{ *(.text) @}
5452 end_of_text = . ;
5453
5454 start_of_data = . ;
5455 .rodata: @{ *(.rodata) @}
5456 .data: @{ *(.data) @}
5457 end_of_data = . ;
5458 @}
5459 @end smallexample
5460
5461 This may or may not be the script author's intention for the value of
5462 @code{start_of_data}. One way to influence the orphan section
5463 placement is to assign the location counter to itself, as the linker
5464 assumes that an assignment to @code{.} is setting the start address of
5465 a following output section and thus should be grouped with that
5466 section. So you could write:
5467
5468 @smallexample
5469 SECTIONS
5470 @{
5471 start_of_text = . ;
5472 .text: @{ *(.text) @}
5473 end_of_text = . ;
5474
5475 . = . ;
5476 start_of_data = . ;
5477 .data: @{ *(.data) @}
5478 end_of_data = . ;
5479 @}
5480 @end smallexample
5481
5482 Now, the orphan @code{.rodata} section will be placed between
5483 @code{end_of_text} and @code{start_of_data}.
5484
5485 @need 2000
5486 @node Operators
5487 @subsection Operators
5488 @cindex operators for arithmetic
5489 @cindex arithmetic operators
5490 @cindex precedence in expressions
5491 The linker recognizes the standard C set of arithmetic operators, with
5492 the standard bindings and precedence levels:
5493 @c TEXI2ROFF-KILL
5494 @ifnottex
5495 @c END TEXI2ROFF-KILL
5496 @smallexample
5497 precedence associativity Operators Notes
5498 (highest)
5499 1 left ! - ~ (1)
5500 2 left * / %
5501 3 left + -
5502 4 left >> <<
5503 5 left == != > < <= >=
5504 6 left &
5505 7 left |
5506 8 left &&
5507 9 left ||
5508 10 right ? :
5509 11 right &= += -= *= /= (2)
5510 (lowest)
5511 @end smallexample
5512 Notes:
5513 (1) Prefix operators
5514 (2) @xref{Assignments}.
5515 @c TEXI2ROFF-KILL
5516 @end ifnottex
5517 @tex
5518 \vskip \baselineskip
5519 %"lispnarrowing" is the extra indent used generally for smallexample
5520 \hskip\lispnarrowing\vbox{\offinterlineskip
5521 \hrule
5522 \halign
5523 {\vrule#&\strut\hfil\ #\ \hfil&\vrule#&\strut\hfil\ #\ \hfil&\vrule#&\strut\hfil\ {\tt #}\ \hfil&\vrule#\cr
5524 height2pt&\omit&&\omit&&\omit&\cr
5525 &Precedence&& Associativity &&{\rm Operators}&\cr
5526 height2pt&\omit&&\omit&&\omit&\cr
5527 \noalign{\hrule}
5528 height2pt&\omit&&\omit&&\omit&\cr
5529 &highest&&&&&\cr
5530 % '176 is tilde, '~' in tt font
5531 &1&&left&&\qquad- \char'176\ !\qquad\dag&\cr
5532 &2&&left&&* / \%&\cr
5533 &3&&left&&+ -&\cr
5534 &4&&left&&>> <<&\cr
5535 &5&&left&&== != > < <= >=&\cr
5536 &6&&left&&\&&\cr
5537 &7&&left&&|&\cr
5538 &8&&left&&{\&\&}&\cr
5539 &9&&left&&||&\cr
5540 &10&&right&&? :&\cr
5541 &11&&right&&\qquad\&= += -= *= /=\qquad\ddag&\cr
5542 &lowest&&&&&\cr
5543 height2pt&\omit&&\omit&&\omit&\cr}
5544 \hrule}
5545 @end tex
5546 @iftex
5547 {
5548 @obeylines@parskip=0pt@parindent=0pt
5549 @dag@quad Prefix operators.
5550 @ddag@quad @xref{Assignments}.
5551 }
5552 @end iftex
5553 @c END TEXI2ROFF-KILL
5554
5555 @node Evaluation
5556 @subsection Evaluation
5557 @cindex lazy evaluation
5558 @cindex expression evaluation order
5559 The linker evaluates expressions lazily. It only computes the value of
5560 an expression when absolutely necessary.
5561
5562 The linker needs some information, such as the value of the start
5563 address of the first section, and the origins and lengths of memory
5564 regions, in order to do any linking at all. These values are computed
5565 as soon as possible when the linker reads in the linker script.
5566
5567 However, other values (such as symbol values) are not known or needed
5568 until after storage allocation. Such values are evaluated later, when
5569 other information (such as the sizes of output sections) is available
5570 for use in the symbol assignment expression.
5571
5572 The sizes of sections cannot be known until after allocation, so
5573 assignments dependent upon these are not performed until after
5574 allocation.
5575
5576 Some expressions, such as those depending upon the location counter
5577 @samp{.}, must be evaluated during section allocation.
5578
5579 If the result of an expression is required, but the value is not
5580 available, then an error results. For example, a script like the
5581 following
5582 @smallexample
5583 @group
5584 SECTIONS
5585 @{
5586 .text 9+this_isnt_constant :
5587 @{ *(.text) @}
5588 @}
5589 @end group
5590 @end smallexample
5591 @noindent
5592 will cause the error message @samp{non constant expression for initial
5593 address}.
5594
5595 @node Expression Section
5596 @subsection The Section of an Expression
5597 @cindex expression sections
5598 @cindex absolute expressions
5599 @cindex relative expressions
5600 @cindex absolute and relocatable symbols
5601 @cindex relocatable and absolute symbols
5602 @cindex symbols, relocatable and absolute
5603 Addresses and symbols may be section relative, or absolute. A section
5604 relative symbol is relocatable. If you request relocatable output
5605 using the @samp{-r} option, a further link operation may change the
5606 value of a section relative symbol. On the other hand, an absolute
5607 symbol will retain the same value throughout any further link
5608 operations.
5609
5610 Some terms in linker expressions are addresses. This is true of
5611 section relative symbols and for builtin functions that return an
5612 address, such as @code{ADDR}, @code{LOADADDR}, @code{ORIGIN} and
5613 @code{SEGMENT_START}. Other terms are simply numbers, or are builtin
5614 functions that return a non-address value, such as @code{LENGTH}.
5615 One complication is that unless you set @code{LD_FEATURE ("SANE_EXPR")}
5616 (@pxref{Miscellaneous Commands}), numbers and absolute symbols are treated
5617 differently depending on their location, for compatibility with older
5618 versions of @code{ld}. Expressions appearing outside an output
5619 section definition treat all numbers as absolute addresses.
5620 Expressions appearing inside an output section definition treat
5621 absolute symbols as numbers. If @code{LD_FEATURE ("SANE_EXPR")} is
5622 given, then absolute symbols and numbers are simply treated as numbers
5623 everywhere.
5624
5625 In the following simple example,
5626
5627 @smallexample
5628 @group
5629 SECTIONS
5630 @{
5631 . = 0x100;
5632 __executable_start = 0x100;
5633 .data :
5634 @{
5635 . = 0x10;
5636 __data_start = 0x10;
5637 *(.data)
5638 @}
5639 @dots{}
5640 @}
5641 @end group
5642 @end smallexample
5643
5644 both @code{.} and @code{__executable_start} are set to the absolute
5645 address 0x100 in the first two assignments, then both @code{.} and
5646 @code{__data_start} are set to 0x10 relative to the @code{.data}
5647 section in the second two assignments.
5648
5649 For expressions involving numbers, relative addresses and absolute
5650 addresses, ld follows these rules to evaluate terms:
5651
5652 @itemize @bullet
5653 @item
5654 Unary operations on an absolute address or number, and binary
5655 operations on two absolute addresses or two numbers, or between one
5656 absolute address and a number, apply the operator to the value(s).
5657 @item
5658 Unary operations on a relative address, and binary operations on two
5659 relative addresses in the same section or between one relative address
5660 and a number, apply the operator to the offset part of the address(es).
5661 @item
5662 Other binary operations, that is, between two relative addresses not
5663 in the same section, or between a relative address and an absolute
5664 address, first convert any non-absolute term to an absolute address
5665 before applying the operator.
5666 @end itemize
5667
5668 The result section of each sub-expression is as follows:
5669
5670 @itemize @bullet
5671 @item
5672 An operation involving only numbers results in a number.
5673 @item
5674 The result of comparisons, @samp{&&} and @samp{||} is also a number.
5675 @item
5676 The result of other binary arithmetic and logical operations on two
5677 relative addresses in the same section or two absolute addresess
5678 (after above conversions) is also a number.
5679 @item
5680 The result of other operations on relative addresses or one
5681 relative address and a number, is a relative address in the same
5682 section as the relative operand(s).
5683 @item
5684 The result of other operations on absolute addresses (after above
5685 conversions) is an absolute address.
5686 @end itemize
5687
5688 You can use the builtin function @code{ABSOLUTE} to force an expression
5689 to be absolute when it would otherwise be relative. For example, to
5690 create an absolute symbol set to the address of the end of the output
5691 section @samp{.data}:
5692 @smallexample
5693 SECTIONS
5694 @{
5695 .data : @{ *(.data) _edata = ABSOLUTE(.); @}
5696 @}
5697 @end smallexample
5698 @noindent
5699 If @samp{ABSOLUTE} were not used, @samp{_edata} would be relative to the
5700 @samp{.data} section.
5701
5702 Using @code{LOADADDR} also forces an expression absolute, since this
5703 particular builtin function returns an absolute address.
5704
5705 @node Builtin Functions
5706 @subsection Builtin Functions
5707 @cindex functions in expressions
5708 The linker script language includes a number of builtin functions for
5709 use in linker script expressions.
5710
5711 @table @code
5712 @item ABSOLUTE(@var{exp})
5713 @kindex ABSOLUTE(@var{exp})
5714 @cindex expression, absolute
5715 Return the absolute (non-relocatable, as opposed to non-negative) value
5716 of the expression @var{exp}. Primarily useful to assign an absolute
5717 value to a symbol within a section definition, where symbol values are
5718 normally section relative. @xref{Expression Section}.
5719
5720 @item ADDR(@var{section})
5721 @kindex ADDR(@var{section})
5722 @cindex section address in expression
5723 Return the address (VMA) of the named @var{section}. Your
5724 script must previously have defined the location of that section. In
5725 the following example, @code{start_of_output_1}, @code{symbol_1} and
5726 @code{symbol_2} are assigned equivalent values, except that
5727 @code{symbol_1} will be relative to the @code{.output1} section while
5728 the other two will be absolute:
5729 @smallexample
5730 @group
5731 SECTIONS @{ @dots{}
5732 .output1 :
5733 @{
5734 start_of_output_1 = ABSOLUTE(.);
5735 @dots{}
5736 @}
5737 .output :
5738 @{
5739 symbol_1 = ADDR(.output1);
5740 symbol_2 = start_of_output_1;
5741 @}
5742 @dots{} @}
5743 @end group
5744 @end smallexample
5745
5746 @item ALIGN(@var{align})
5747 @itemx ALIGN(@var{exp},@var{align})
5748 @kindex ALIGN(@var{align})
5749 @kindex ALIGN(@var{exp},@var{align})
5750 @cindex round up location counter
5751 @cindex align location counter
5752 @cindex round up expression
5753 @cindex align expression
5754 Return the location counter (@code{.}) or arbitrary expression aligned
5755 to the next @var{align} boundary. The single operand @code{ALIGN}
5756 doesn't change the value of the location counter---it just does
5757 arithmetic on it. The two operand @code{ALIGN} allows an arbitrary
5758 expression to be aligned upwards (@code{ALIGN(@var{align})} is
5759 equivalent to @code{ALIGN(., @var{align})}).
5760
5761 Here is an example which aligns the output @code{.data} section to the
5762 next @code{0x2000} byte boundary after the preceding section and sets a
5763 variable within the section to the next @code{0x8000} boundary after the
5764 input sections:
5765 @smallexample
5766 @group
5767 SECTIONS @{ @dots{}
5768 .data ALIGN(0x2000): @{
5769 *(.data)
5770 variable = ALIGN(0x8000);
5771 @}
5772 @dots{} @}
5773 @end group
5774 @end smallexample
5775 @noindent
5776 The first use of @code{ALIGN} in this example specifies the location of
5777 a section because it is used as the optional @var{address} attribute of
5778 a section definition (@pxref{Output Section Address}). The second use
5779 of @code{ALIGN} is used to defines the value of a symbol.
5780
5781 The builtin function @code{NEXT} is closely related to @code{ALIGN}.
5782
5783 @item ALIGNOF(@var{section})
5784 @kindex ALIGNOF(@var{section})
5785 @cindex section alignment
5786 Return the alignment in bytes of the named @var{section}, if that section has
5787 been allocated. If the section has not been allocated when this is
5788 evaluated, the linker will report an error. In the following example,
5789 the alignment of the @code{.output} section is stored as the first
5790 value in that section.
5791 @smallexample
5792 @group
5793 SECTIONS@{ @dots{}
5794 .output @{
5795 LONG (ALIGNOF (.output))
5796 @dots{}
5797 @}
5798 @dots{} @}
5799 @end group
5800 @end smallexample
5801
5802 @item BLOCK(@var{exp})
5803 @kindex BLOCK(@var{exp})
5804 This is a synonym for @code{ALIGN}, for compatibility with older linker
5805 scripts. It is most often seen when setting the address of an output
5806 section.
5807
5808 @item DATA_SEGMENT_ALIGN(@var{maxpagesize}, @var{commonpagesize})
5809 @kindex DATA_SEGMENT_ALIGN(@var{maxpagesize}, @var{commonpagesize})
5810 This is equivalent to either
5811 @smallexample
5812 (ALIGN(@var{maxpagesize}) + (. & (@var{maxpagesize} - 1)))
5813 @end smallexample
5814 or
5815 @smallexample
5816 (ALIGN(@var{maxpagesize}) + (. & (@var{maxpagesize} - @var{commonpagesize})))
5817 @end smallexample
5818 @noindent
5819 depending on whether the latter uses fewer @var{commonpagesize} sized pages
5820 for the data segment (area between the result of this expression and
5821 @code{DATA_SEGMENT_END}) than the former or not.
5822 If the latter form is used, it means @var{commonpagesize} bytes of runtime
5823 memory will be saved at the expense of up to @var{commonpagesize} wasted
5824 bytes in the on-disk file.
5825
5826 This expression can only be used directly in @code{SECTIONS} commands, not in
5827 any output section descriptions and only once in the linker script.
5828 @var{commonpagesize} should be less or equal to @var{maxpagesize} and should
5829 be the system page size the object wants to be optimized for (while still
5830 working on system page sizes up to @var{maxpagesize}).
5831
5832 @noindent
5833 Example:
5834 @smallexample
5835 . = DATA_SEGMENT_ALIGN(0x10000, 0x2000);
5836 @end smallexample
5837
5838 @item DATA_SEGMENT_END(@var{exp})
5839 @kindex DATA_SEGMENT_END(@var{exp})
5840 This defines the end of data segment for @code{DATA_SEGMENT_ALIGN}
5841 evaluation purposes.
5842
5843 @smallexample
5844 . = DATA_SEGMENT_END(.);
5845 @end smallexample
5846
5847 @item DATA_SEGMENT_RELRO_END(@var{offset}, @var{exp})
5848 @kindex DATA_SEGMENT_RELRO_END(@var{offset}, @var{exp})
5849 This defines the end of the @code{PT_GNU_RELRO} segment when
5850 @samp{-z relro} option is used. Second argument is returned.
5851 When @samp{-z relro} option is not present, @code{DATA_SEGMENT_RELRO_END}
5852 does nothing, otherwise @code{DATA_SEGMENT_ALIGN} is padded so that
5853 @var{exp} + @var{offset} is aligned to the most commonly used page
5854 boundary for particular target. If present in the linker script,
5855 it must always come in between @code{DATA_SEGMENT_ALIGN} and
5856 @code{DATA_SEGMENT_END}.
5857
5858 @smallexample
5859 . = DATA_SEGMENT_RELRO_END(24, .);
5860 @end smallexample
5861
5862 @item DEFINED(@var{symbol})
5863 @kindex DEFINED(@var{symbol})
5864 @cindex symbol defaults
5865 Return 1 if @var{symbol} is in the linker global symbol table and is
5866 defined before the statement using DEFINED in the script, otherwise
5867 return 0. You can use this function to provide
5868 default values for symbols. For example, the following script fragment
5869 shows how to set a global symbol @samp{begin} to the first location in
5870 the @samp{.text} section---but if a symbol called @samp{begin} already
5871 existed, its value is preserved:
5872
5873 @smallexample
5874 @group
5875 SECTIONS @{ @dots{}
5876 .text : @{
5877 begin = DEFINED(begin) ? begin : . ;
5878 @dots{}
5879 @}
5880 @dots{}
5881 @}
5882 @end group
5883 @end smallexample
5884
5885 @item LENGTH(@var{memory})
5886 @kindex LENGTH(@var{memory})
5887 Return the length of the memory region named @var{memory}.
5888
5889 @item LOADADDR(@var{section})
5890 @kindex LOADADDR(@var{section})
5891 @cindex section load address in expression
5892 Return the absolute LMA of the named @var{section}. (@pxref{Output
5893 Section LMA}).
5894
5895 @kindex MAX
5896 @item MAX(@var{exp1}, @var{exp2})
5897 Returns the maximum of @var{exp1} and @var{exp2}.
5898
5899 @kindex MIN
5900 @item MIN(@var{exp1}, @var{exp2})
5901 Returns the minimum of @var{exp1} and @var{exp2}.
5902
5903 @item NEXT(@var{exp})
5904 @kindex NEXT(@var{exp})
5905 @cindex unallocated address, next
5906 Return the next unallocated address that is a multiple of @var{exp}.
5907 This function is closely related to @code{ALIGN(@var{exp})}; unless you
5908 use the @code{MEMORY} command to define discontinuous memory for the
5909 output file, the two functions are equivalent.
5910
5911 @item ORIGIN(@var{memory})
5912 @kindex ORIGIN(@var{memory})
5913 Return the origin of the memory region named @var{memory}.
5914
5915 @item SEGMENT_START(@var{segment}, @var{default})
5916 @kindex SEGMENT_START(@var{segment}, @var{default})
5917 Return the base address of the named @var{segment}. If an explicit
5918 value has been given for this segment (with a command-line @samp{-T}
5919 option) that value will be returned; otherwise the value will be
5920 @var{default}. At present, the @samp{-T} command-line option can only
5921 be used to set the base address for the ``text'', ``data'', and
5922 ``bss'' sections, but you can use @code{SEGMENT_START} with any segment
5923 name.
5924
5925 @item SIZEOF(@var{section})
5926 @kindex SIZEOF(@var{section})
5927 @cindex section size
5928 Return the size in bytes of the named @var{section}, if that section has
5929 been allocated. If the section has not been allocated when this is
5930 evaluated, the linker will report an error. In the following example,
5931 @code{symbol_1} and @code{symbol_2} are assigned identical values:
5932 @smallexample
5933 @group
5934 SECTIONS@{ @dots{}
5935 .output @{
5936 .start = . ;
5937 @dots{}
5938 .end = . ;
5939 @}
5940 symbol_1 = .end - .start ;
5941 symbol_2 = SIZEOF(.output);
5942 @dots{} @}
5943 @end group
5944 @end smallexample
5945
5946 @item SIZEOF_HEADERS
5947 @itemx sizeof_headers
5948 @kindex SIZEOF_HEADERS
5949 @cindex header size
5950 Return the size in bytes of the output file's headers. This is
5951 information which appears at the start of the output file. You can use
5952 this number when setting the start address of the first section, if you
5953 choose, to facilitate paging.
5954
5955 @cindex not enough room for program headers
5956 @cindex program headers, not enough room
5957 When producing an ELF output file, if the linker script uses the
5958 @code{SIZEOF_HEADERS} builtin function, the linker must compute the
5959 number of program headers before it has determined all the section
5960 addresses and sizes. If the linker later discovers that it needs
5961 additional program headers, it will report an error @samp{not enough
5962 room for program headers}. To avoid this error, you must avoid using
5963 the @code{SIZEOF_HEADERS} function, or you must rework your linker
5964 script to avoid forcing the linker to use additional program headers, or
5965 you must define the program headers yourself using the @code{PHDRS}
5966 command (@pxref{PHDRS}).
5967 @end table
5968
5969 @node Implicit Linker Scripts
5970 @section Implicit Linker Scripts
5971 @cindex implicit linker scripts
5972 If you specify a linker input file which the linker can not recognize as
5973 an object file or an archive file, it will try to read the file as a
5974 linker script. If the file can not be parsed as a linker script, the
5975 linker will report an error.
5976
5977 An implicit linker script will not replace the default linker script.
5978
5979 Typically an implicit linker script would contain only symbol
5980 assignments, or the @code{INPUT}, @code{GROUP}, or @code{VERSION}
5981 commands.
5982
5983 Any input files read because of an implicit linker script will be read
5984 at the position in the command line where the implicit linker script was
5985 read. This can affect archive searching.
5986
5987 @ifset GENERIC
5988 @node Machine Dependent
5989 @chapter Machine Dependent Features
5990
5991 @cindex machine dependencies
5992 @command{ld} has additional features on some platforms; the following
5993 sections describe them. Machines where @command{ld} has no additional
5994 functionality are not listed.
5995
5996 @menu
5997 @ifset H8300
5998 * H8/300:: @command{ld} and the H8/300
5999 @end ifset
6000 @ifset I960
6001 * i960:: @command{ld} and the Intel 960 family
6002 @end ifset
6003 @ifset ARM
6004 * ARM:: @command{ld} and the ARM family
6005 @end ifset
6006 @ifset HPPA
6007 * HPPA ELF32:: @command{ld} and HPPA 32-bit ELF
6008 @end ifset
6009 @ifset M68K
6010 * M68K:: @command{ld} and the Motorola 68K family
6011 @end ifset
6012 @ifset MMIX
6013 * MMIX:: @command{ld} and MMIX
6014 @end ifset
6015 @ifset MSP430
6016 * MSP430:: @command{ld} and MSP430
6017 @end ifset
6018 @ifset M68HC11
6019 * M68HC11/68HC12:: @code{ld} and the Motorola 68HC11 and 68HC12 families
6020 @end ifset
6021 @ifset POWERPC
6022 * PowerPC ELF32:: @command{ld} and PowerPC 32-bit ELF Support
6023 @end ifset
6024 @ifset POWERPC64
6025 * PowerPC64 ELF64:: @command{ld} and PowerPC64 64-bit ELF Support
6026 @end ifset
6027 @ifset SPU
6028 * SPU ELF:: @command{ld} and SPU ELF Support
6029 @end ifset
6030 @ifset TICOFF
6031 * TI COFF:: @command{ld} and TI COFF
6032 @end ifset
6033 @ifset WIN32
6034 * WIN32:: @command{ld} and WIN32 (cygwin/mingw)
6035 @end ifset
6036 @ifset XTENSA
6037 * Xtensa:: @command{ld} and Xtensa Processors
6038 @end ifset
6039 @end menu
6040 @end ifset
6041
6042 @ifset H8300
6043 @ifclear GENERIC
6044 @raisesections
6045 @end ifclear
6046
6047 @node H8/300
6048 @section @command{ld} and the H8/300
6049
6050 @cindex H8/300 support
6051 For the H8/300, @command{ld} can perform these global optimizations when
6052 you specify the @samp{--relax} command-line option.
6053
6054 @table @emph
6055 @cindex relaxing on H8/300
6056 @item relaxing address modes
6057 @command{ld} finds all @code{jsr} and @code{jmp} instructions whose
6058 targets are within eight bits, and turns them into eight-bit
6059 program-counter relative @code{bsr} and @code{bra} instructions,
6060 respectively.
6061
6062 @cindex synthesizing on H8/300
6063 @item synthesizing instructions
6064 @c FIXME: specifically mov.b, or any mov instructions really?
6065 @command{ld} finds all @code{mov.b} instructions which use the
6066 sixteen-bit absolute address form, but refer to the top
6067 page of memory, and changes them to use the eight-bit address form.
6068 (That is: the linker turns @samp{mov.b @code{@@}@var{aa}:16} into
6069 @samp{mov.b @code{@@}@var{aa}:8} whenever the address @var{aa} is in the
6070 top page of memory).
6071
6072 @item bit manipulation instructions
6073 @command{ld} finds all bit manipulation instructions like @code{band, bclr,
6074 biand, bild, bior, bist, bixor, bld, bnot, bor, bset, bst, btst, bxor}
6075 which use 32 bit and 16 bit absolute address form, but refer to the top
6076 page of memory, and changes them to use the 8 bit address form.
6077 (That is: the linker turns @samp{bset #xx:3,@code{@@}@var{aa}:32} into
6078 @samp{bset #xx:3,@code{@@}@var{aa}:8} whenever the address @var{aa} is in
6079 the top page of memory).
6080
6081 @item system control instructions
6082 @command{ld} finds all @code{ldc.w, stc.w} instructions which use the
6083 32 bit absolute address form, but refer to the top page of memory, and
6084 changes them to use 16 bit address form.
6085 (That is: the linker turns @samp{ldc.w @code{@@}@var{aa}:32,ccr} into
6086 @samp{ldc.w @code{@@}@var{aa}:16,ccr} whenever the address @var{aa} is in
6087 the top page of memory).
6088 @end table
6089
6090 @ifclear GENERIC
6091 @lowersections
6092 @end ifclear
6093 @end ifset
6094
6095 @ifclear GENERIC
6096 @ifset Renesas
6097 @c This stuff is pointless to say unless you're especially concerned
6098 @c with Renesas chips; don't enable it for generic case, please.
6099 @node Renesas
6100 @chapter @command{ld} and Other Renesas Chips
6101
6102 @command{ld} also supports the Renesas (formerly Hitachi) H8/300H,
6103 H8/500, and SH chips. No special features, commands, or command-line
6104 options are required for these chips.
6105 @end ifset
6106 @end ifclear
6107
6108 @ifset I960
6109 @ifclear GENERIC
6110 @raisesections
6111 @end ifclear
6112
6113 @node i960
6114 @section @command{ld} and the Intel 960 Family
6115
6116 @cindex i960 support
6117
6118 You can use the @samp{-A@var{architecture}} command line option to
6119 specify one of the two-letter names identifying members of the 960
6120 family; the option specifies the desired output target, and warns of any
6121 incompatible instructions in the input files. It also modifies the
6122 linker's search strategy for archive libraries, to support the use of
6123 libraries specific to each particular architecture, by including in the
6124 search loop names suffixed with the string identifying the architecture.
6125
6126 For example, if your @command{ld} command line included @w{@samp{-ACA}} as
6127 well as @w{@samp{-ltry}}, the linker would look (in its built-in search
6128 paths, and in any paths you specify with @samp{-L}) for a library with
6129 the names
6130
6131 @smallexample
6132 @group
6133 try
6134 libtry.a
6135 tryca
6136 libtryca.a
6137 @end group
6138 @end smallexample
6139
6140 @noindent
6141 The first two possibilities would be considered in any event; the last
6142 two are due to the use of @w{@samp{-ACA}}.
6143
6144 You can meaningfully use @samp{-A} more than once on a command line, since
6145 the 960 architecture family allows combination of target architectures; each
6146 use will add another pair of name variants to search for when @w{@samp{-l}}
6147 specifies a library.
6148
6149 @cindex @option{--relax} on i960
6150 @cindex relaxing on i960
6151 @command{ld} supports the @samp{--relax} option for the i960 family. If
6152 you specify @samp{--relax}, @command{ld} finds all @code{balx} and
6153 @code{calx} instructions whose targets are within 24 bits, and turns
6154 them into 24-bit program-counter relative @code{bal} and @code{cal}
6155 instructions, respectively. @command{ld} also turns @code{cal}
6156 instructions into @code{bal} instructions when it determines that the
6157 target subroutine is a leaf routine (that is, the target subroutine does
6158 not itself call any subroutines).
6159
6160 @cindex Cortex-A8 erratum workaround
6161 @kindex --fix-cortex-a8
6162 @kindex --no-fix-cortex-a8
6163 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}.
6164
6165 The erratum only affects Thumb-2 code. Please contact ARM for further details.
6166
6167 @kindex --merge-exidx-entries
6168 @kindex --no-merge-exidx-entries
6169 The @samp{--no-merge-exidx-entries} switch disables the merging of adjacent exidx entries in debuginfo.
6170
6171 @ifclear GENERIC
6172 @lowersections
6173 @end ifclear
6174 @end ifset
6175
6176 @ifset ARM
6177 @ifclear GENERIC
6178 @raisesections
6179 @end ifclear
6180
6181 @ifset M68HC11
6182 @ifclear GENERIC
6183 @raisesections
6184 @end ifclear
6185
6186 @node M68HC11/68HC12
6187 @section @command{ld} and the Motorola 68HC11 and 68HC12 families
6188
6189 @cindex M68HC11 and 68HC12 support
6190
6191 @subsection Linker Relaxation
6192
6193 For the Motorola 68HC11, @command{ld} can perform these global
6194 optimizations when you specify the @samp{--relax} command-line option.
6195
6196 @table @emph
6197 @cindex relaxing on M68HC11
6198 @item relaxing address modes
6199 @command{ld} finds all @code{jsr} and @code{jmp} instructions whose
6200 targets are within eight bits, and turns them into eight-bit
6201 program-counter relative @code{bsr} and @code{bra} instructions,
6202 respectively.
6203
6204 @command{ld} also looks at all 16-bit extended addressing modes and
6205 transforms them in a direct addressing mode when the address is in
6206 page 0 (between 0 and 0x0ff).
6207
6208 @item relaxing gcc instruction group
6209 When @command{gcc} is called with @option{-mrelax}, it can emit group
6210 of instructions that the linker can optimize to use a 68HC11 direct
6211 addressing mode. These instructions consists of @code{bclr} or
6212 @code{bset} instructions.
6213
6214 @end table
6215
6216 @subsection Trampoline Generation
6217
6218 @cindex trampoline generation on M68HC11
6219 @cindex trampoline generation on M68HC12
6220 For 68HC11 and 68HC12, @command{ld} can generate trampoline code to
6221 call a far function using a normal @code{jsr} instruction. The linker
6222 will also change the relocation to some far function to use the
6223 trampoline address instead of the function address. This is typically the
6224 case when a pointer to a function is taken. The pointer will in fact
6225 point to the function trampoline.
6226
6227 @ifclear GENERIC
6228 @lowersections
6229 @end ifclear
6230 @end ifset
6231
6232 @node ARM
6233 @section @command{ld} and the ARM family
6234
6235 @cindex ARM interworking support
6236 @kindex --support-old-code
6237 For the ARM, @command{ld} will generate code stubs to allow functions calls
6238 between ARM and Thumb code. These stubs only work with code that has
6239 been compiled and assembled with the @samp{-mthumb-interwork} command
6240 line option. If it is necessary to link with old ARM object files or
6241 libraries, which have not been compiled with the -mthumb-interwork
6242 option then the @samp{--support-old-code} command line switch should be
6243 given to the linker. This will make it generate larger stub functions
6244 which will work with non-interworking aware ARM code. Note, however,
6245 the linker does not support generating stubs for function calls to
6246 non-interworking aware Thumb code.
6247
6248 @cindex thumb entry point
6249 @cindex entry point, thumb
6250 @kindex --thumb-entry=@var{entry}
6251 The @samp{--thumb-entry} switch is a duplicate of the generic
6252 @samp{--entry} switch, in that it sets the program's starting address.
6253 But it also sets the bottom bit of the address, so that it can be
6254 branched to using a BX instruction, and the program will start
6255 executing in Thumb mode straight away.
6256
6257 @cindex PE import table prefixing
6258 @kindex --use-nul-prefixed-import-tables
6259 The @samp{--use-nul-prefixed-import-tables} switch is specifying, that
6260 the import tables idata4 and idata5 have to be generated with a zero
6261 elememt prefix for import libraries. This is the old style to generate
6262 import tables. By default this option is turned off.
6263
6264 @cindex BE8
6265 @kindex --be8
6266 The @samp{--be8} switch instructs @command{ld} to generate BE8 format
6267 executables. This option is only valid when linking big-endian objects.
6268 The resulting image will contain big-endian data and little-endian code.
6269
6270 @cindex TARGET1
6271 @kindex --target1-rel
6272 @kindex --target1-abs
6273 The @samp{R_ARM_TARGET1} relocation is typically used for entries in the
6274 @samp{.init_array} section. It is interpreted as either @samp{R_ARM_REL32}
6275 or @samp{R_ARM_ABS32}, depending on the target. The @samp{--target1-rel}
6276 and @samp{--target1-abs} switches override the default.
6277
6278 @cindex TARGET2
6279 @kindex --target2=@var{type}
6280 The @samp{--target2=type} switch overrides the default definition of the
6281 @samp{R_ARM_TARGET2} relocation. Valid values for @samp{type}, their
6282 meanings, and target defaults are as follows:
6283 @table @samp
6284 @item rel
6285 @samp{R_ARM_REL32} (arm*-*-elf, arm*-*-eabi)
6286 @item abs
6287 @samp{R_ARM_ABS32} (arm*-*-symbianelf)
6288 @item got-rel
6289 @samp{R_ARM_GOT_PREL} (arm*-*-linux, arm*-*-*bsd)
6290 @end table
6291
6292 @cindex FIX_V4BX
6293 @kindex --fix-v4bx
6294 The @samp{R_ARM_V4BX} relocation (defined by the ARM AAELF
6295 specification) enables objects compiled for the ARMv4 architecture to be
6296 interworking-safe when linked with other objects compiled for ARMv4t, but
6297 also allows pure ARMv4 binaries to be built from the same ARMv4 objects.
6298
6299 In the latter case, the switch @option{--fix-v4bx} must be passed to the
6300 linker, which causes v4t @code{BX rM} instructions to be rewritten as
6301 @code{MOV PC,rM}, since v4 processors do not have a @code{BX} instruction.
6302
6303 In the former case, the switch should not be used, and @samp{R_ARM_V4BX}
6304 relocations are ignored.
6305
6306 @cindex FIX_V4BX_INTERWORKING
6307 @kindex --fix-v4bx-interworking
6308 Replace @code{BX rM} instructions identified by @samp{R_ARM_V4BX}
6309 relocations with a branch to the following veneer:
6310
6311 @smallexample
6312 TST rM, #1
6313 MOVEQ PC, rM
6314 BX Rn
6315 @end smallexample
6316
6317 This allows generation of libraries/applications that work on ARMv4 cores
6318 and are still interworking safe. Note that the above veneer clobbers the
6319 condition flags, so may cause incorrect progrm behavior in rare cases.
6320
6321 @cindex USE_BLX
6322 @kindex --use-blx
6323 The @samp{--use-blx} switch enables the linker to use ARM/Thumb
6324 BLX instructions (available on ARMv5t and above) in various
6325 situations. Currently it is used to perform calls via the PLT from Thumb
6326 code using BLX rather than using BX and a mode-switching stub before
6327 each PLT entry. This should lead to such calls executing slightly faster.
6328
6329 This option is enabled implicitly for SymbianOS, so there is no need to
6330 specify it if you are using that target.
6331
6332 @cindex VFP11_DENORM_FIX
6333 @kindex --vfp11-denorm-fix
6334 The @samp{--vfp11-denorm-fix} switch enables a link-time workaround for a
6335 bug in certain VFP11 coprocessor hardware, which sometimes allows
6336 instructions with denorm operands (which must be handled by support code)
6337 to have those operands overwritten by subsequent instructions before
6338 the support code can read the intended values.
6339
6340 The bug may be avoided in scalar mode if you allow at least one
6341 intervening instruction between a VFP11 instruction which uses a register
6342 and another instruction which writes to the same register, or at least two
6343 intervening instructions if vector mode is in use. The bug only affects
6344 full-compliance floating-point mode: you do not need this workaround if
6345 you are using "runfast" mode. Please contact ARM for further details.
6346
6347 If you know you are using buggy VFP11 hardware, you can
6348 enable this workaround by specifying the linker option
6349 @samp{--vfp-denorm-fix=scalar} if you are using the VFP11 scalar
6350 mode only, or @samp{--vfp-denorm-fix=vector} if you are using
6351 vector mode (the latter also works for scalar code). The default is
6352 @samp{--vfp-denorm-fix=none}.
6353
6354 If the workaround is enabled, instructions are scanned for
6355 potentially-troublesome sequences, and a veneer is created for each
6356 such sequence which may trigger the erratum. The veneer consists of the
6357 first instruction of the sequence and a branch back to the subsequent
6358 instruction. The original instruction is then replaced with a branch to
6359 the veneer. The extra cycles required to call and return from the veneer
6360 are sufficient to avoid the erratum in both the scalar and vector cases.
6361
6362 @cindex ARM1176 erratum workaround
6363 @kindex --fix-arm1176
6364 @kindex --no-fix-arm1176
6365 The @samp{--fix-arm1176} switch enables a link-time workaround for an erratum
6366 in certain ARM1176 processors. The workaround is enabled by default if you
6367 are targetting ARM v6 (excluding ARM v6T2) or earlier. It can be disabled
6368 unconditionally by specifying @samp{--no-fix-arm1176}.
6369
6370 Further information is available in the ``ARM1176JZ-S and ARM1176JZF-S
6371 Programmer Advice Notice'' available on the ARM documentaion website at:
6372 http://infocenter.arm.com/.
6373
6374 @cindex NO_ENUM_SIZE_WARNING
6375 @kindex --no-enum-size-warning
6376 The @option{--no-enum-size-warning} switch prevents the linker from
6377 warning when linking object files that specify incompatible EABI
6378 enumeration size attributes. For example, with this switch enabled,
6379 linking of an object file using 32-bit enumeration values with another
6380 using enumeration values fitted into the smallest possible space will
6381 not be diagnosed.
6382
6383 @cindex NO_WCHAR_SIZE_WARNING
6384 @kindex --no-wchar-size-warning
6385 The @option{--no-wchar-size-warning} switch prevents the linker from
6386 warning when linking object files that specify incompatible EABI
6387 @code{wchar_t} size attributes. For example, with this switch enabled,
6388 linking of an object file using 32-bit @code{wchar_t} values with another
6389 using 16-bit @code{wchar_t} values will not be diagnosed.
6390
6391 @cindex PIC_VENEER
6392 @kindex --pic-veneer
6393 The @samp{--pic-veneer} switch makes the linker use PIC sequences for
6394 ARM/Thumb interworking veneers, even if the rest of the binary
6395 is not PIC. This avoids problems on uClinux targets where
6396 @samp{--emit-relocs} is used to generate relocatable binaries.
6397
6398 @cindex STUB_GROUP_SIZE
6399 @kindex --stub-group-size=@var{N}
6400 The linker will automatically generate and insert small sequences of
6401 code into a linked ARM ELF executable whenever an attempt is made to
6402 perform a function call to a symbol that is too far away. The
6403 placement of these sequences of instructions - called stubs - is
6404 controlled by the command line option @option{--stub-group-size=N}.
6405 The placement is important because a poor choice can create a need for
6406 duplicate stubs, increasing the code sizw. The linker will try to
6407 group stubs together in order to reduce interruptions to the flow of
6408 code, but it needs guidance as to how big these groups should be and
6409 where they should be placed.
6410
6411 The value of @samp{N}, the parameter to the
6412 @option{--stub-group-size=} option controls where the stub groups are
6413 placed. If it is negative then all stubs are placed after the first
6414 branch that needs them. If it is positive then the stubs can be
6415 placed either before or after the branches that need them. If the
6416 value of @samp{N} is 1 (either +1 or -1) then the linker will choose
6417 exactly where to place groups of stubs, using its built in heuristics.
6418 A value of @samp{N} greater than 1 (or smaller than -1) tells the
6419 linker that a single group of stubs can service at most @samp{N} bytes
6420 from the input sections.
6421
6422 The default, if @option{--stub-group-size=} is not specified, is
6423 @samp{N = +1}.
6424
6425 Farcalls stubs insertion is fully supported for the ARM-EABI target
6426 only, because it relies on object files properties not present
6427 otherwise.
6428
6429 @ifclear GENERIC
6430 @lowersections
6431 @end ifclear
6432 @end ifset
6433
6434 @ifset HPPA
6435 @ifclear GENERIC
6436 @raisesections
6437 @end ifclear
6438
6439 @node HPPA ELF32
6440 @section @command{ld} and HPPA 32-bit ELF Support
6441 @cindex HPPA multiple sub-space stubs
6442 @kindex --multi-subspace
6443 When generating a shared library, @command{ld} will by default generate
6444 import stubs suitable for use with a single sub-space application.
6445 The @samp{--multi-subspace} switch causes @command{ld} to generate export
6446 stubs, and different (larger) import stubs suitable for use with
6447 multiple sub-spaces.
6448
6449 @cindex HPPA stub grouping
6450 @kindex --stub-group-size=@var{N}
6451 Long branch stubs and import/export stubs are placed by @command{ld} in
6452 stub sections located between groups of input sections.
6453 @samp{--stub-group-size} specifies the maximum size of a group of input
6454 sections handled by one stub section. Since branch offsets are signed,
6455 a stub section may serve two groups of input sections, one group before
6456 the stub section, and one group after it. However, when using
6457 conditional branches that require stubs, it may be better (for branch
6458 prediction) that stub sections only serve one group of input sections.
6459 A negative value for @samp{N} chooses this scheme, ensuring that
6460 branches to stubs always use a negative offset. Two special values of
6461 @samp{N} are recognized, @samp{1} and @samp{-1}. These both instruct
6462 @command{ld} to automatically size input section groups for the branch types
6463 detected, with the same behaviour regarding stub placement as other
6464 positive or negative values of @samp{N} respectively.
6465
6466 Note that @samp{--stub-group-size} does not split input sections. A
6467 single input section larger than the group size specified will of course
6468 create a larger group (of one section). If input sections are too
6469 large, it may not be possible for a branch to reach its stub.
6470
6471 @ifclear GENERIC
6472 @lowersections
6473 @end ifclear
6474 @end ifset
6475
6476 @ifset M68K
6477 @ifclear GENERIC
6478 @raisesections
6479 @end ifclear
6480
6481 @node M68K
6482 @section @command{ld} and the Motorola 68K family
6483
6484 @cindex Motorola 68K GOT generation
6485 @kindex --got=@var{type}
6486 The @samp{--got=@var{type}} option lets you choose the GOT generation scheme.
6487 The choices are @samp{single}, @samp{negative}, @samp{multigot} and
6488 @samp{target}. When @samp{target} is selected the linker chooses
6489 the default GOT generation scheme for the current target.
6490 @samp{single} tells the linker to generate a single GOT with
6491 entries only at non-negative offsets.
6492 @samp{negative} instructs the linker to generate a single GOT with
6493 entries at both negative and positive offsets. Not all environments
6494 support such GOTs.
6495 @samp{multigot} allows the linker to generate several GOTs in the
6496 output file. All GOT references from a single input object
6497 file access the same GOT, but references from different input object
6498 files might access different GOTs. Not all environments support such GOTs.
6499
6500 @ifclear GENERIC
6501 @lowersections
6502 @end ifclear
6503 @end ifset
6504
6505 @ifset MMIX
6506 @ifclear GENERIC
6507 @raisesections
6508 @end ifclear
6509
6510 @node MMIX
6511 @section @code{ld} and MMIX
6512 For MMIX, there is a choice of generating @code{ELF} object files or
6513 @code{mmo} object files when linking. The simulator @code{mmix}
6514 understands the @code{mmo} format. The binutils @code{objcopy} utility
6515 can translate between the two formats.
6516
6517 There is one special section, the @samp{.MMIX.reg_contents} section.
6518 Contents in this section is assumed to correspond to that of global
6519 registers, and symbols referring to it are translated to special symbols,
6520 equal to registers. In a final link, the start address of the
6521 @samp{.MMIX.reg_contents} section corresponds to the first allocated
6522 global register multiplied by 8. Register @code{$255} is not included in
6523 this section; it is always set to the program entry, which is at the
6524 symbol @code{Main} for @code{mmo} files.
6525
6526 Global symbols with the prefix @code{__.MMIX.start.}, for example
6527 @code{__.MMIX.start..text} and @code{__.MMIX.start..data} are special.
6528 The default linker script uses these to set the default start address
6529 of a section.
6530
6531 Initial and trailing multiples of zero-valued 32-bit words in a section,
6532 are left out from an mmo file.
6533
6534 @ifclear GENERIC
6535 @lowersections
6536 @end ifclear
6537 @end ifset
6538
6539 @ifset MSP430
6540 @ifclear GENERIC
6541 @raisesections
6542 @end ifclear
6543
6544 @node MSP430
6545 @section @code{ld} and MSP430
6546 For the MSP430 it is possible to select the MPU architecture. The flag @samp{-m [mpu type]}
6547 will select an appropriate linker script for selected MPU type. (To get a list of known MPUs
6548 just pass @samp{-m help} option to the linker).
6549
6550 @cindex MSP430 extra sections
6551 The linker will recognize some extra sections which are MSP430 specific:
6552
6553 @table @code
6554 @item @samp{.vectors}
6555 Defines a portion of ROM where interrupt vectors located.
6556
6557 @item @samp{.bootloader}
6558 Defines the bootloader portion of the ROM (if applicable). Any code
6559 in this section will be uploaded to the MPU.
6560
6561 @item @samp{.infomem}
6562 Defines an information memory section (if applicable). Any code in
6563 this section will be uploaded to the MPU.
6564
6565 @item @samp{.infomemnobits}
6566 This is the same as the @samp{.infomem} section except that any code
6567 in this section will not be uploaded to the MPU.
6568
6569 @item @samp{.noinit}
6570 Denotes a portion of RAM located above @samp{.bss} section.
6571
6572 The last two sections are used by gcc.
6573 @end table
6574
6575 @ifclear GENERIC
6576 @lowersections
6577 @end ifclear
6578 @end ifset
6579
6580 @ifset POWERPC
6581 @ifclear GENERIC
6582 @raisesections
6583 @end ifclear
6584
6585 @node PowerPC ELF32
6586 @section @command{ld} and PowerPC 32-bit ELF Support
6587 @cindex PowerPC long branches
6588 @kindex --relax on PowerPC
6589 Branches on PowerPC processors are limited to a signed 26-bit
6590 displacement, which may result in @command{ld} giving
6591 @samp{relocation truncated to fit} errors with very large programs.
6592 @samp{--relax} enables the generation of trampolines that can access
6593 the entire 32-bit address space. These trampolines are inserted at
6594 section boundaries, so may not themselves be reachable if an input
6595 section exceeds 33M in size. You may combine @samp{-r} and
6596 @samp{--relax} to add trampolines in a partial link. In that case
6597 both branches to undefined symbols and inter-section branches are also
6598 considered potentially out of range, and trampolines inserted.
6599
6600 @cindex PowerPC ELF32 options
6601 @table @option
6602 @cindex PowerPC PLT
6603 @kindex --bss-plt
6604 @item --bss-plt
6605 Current PowerPC GCC accepts a @samp{-msecure-plt} option that
6606 generates code capable of using a newer PLT and GOT layout that has
6607 the security advantage of no executable section ever needing to be
6608 writable and no writable section ever being executable. PowerPC
6609 @command{ld} will generate this layout, including stubs to access the
6610 PLT, if all input files (including startup and static libraries) were
6611 compiled with @samp{-msecure-plt}. @samp{--bss-plt} forces the old
6612 BSS PLT (and GOT layout) which can give slightly better performance.
6613
6614 @kindex --secure-plt
6615 @item --secure-plt
6616 @command{ld} will use the new PLT and GOT layout if it is linking new
6617 @samp{-fpic} or @samp{-fPIC} code, but does not do so automatically
6618 when linking non-PIC code. This option requests the new PLT and GOT
6619 layout. A warning will be given if some object file requires the old
6620 style BSS PLT.
6621
6622 @cindex PowerPC GOT
6623 @kindex --sdata-got
6624 @item --sdata-got
6625 The new secure PLT and GOT are placed differently relative to other
6626 sections compared to older BSS PLT and GOT placement. The location of
6627 @code{.plt} must change because the new secure PLT is an initialized
6628 section while the old PLT is uninitialized. The reason for the
6629 @code{.got} change is more subtle: The new placement allows
6630 @code{.got} to be read-only in applications linked with
6631 @samp{-z relro -z now}. However, this placement means that
6632 @code{.sdata} cannot always be used in shared libraries, because the
6633 PowerPC ABI accesses @code{.sdata} in shared libraries from the GOT
6634 pointer. @samp{--sdata-got} forces the old GOT placement. PowerPC
6635 GCC doesn't use @code{.sdata} in shared libraries, so this option is
6636 really only useful for other compilers that may do so.
6637
6638 @cindex PowerPC stub symbols
6639 @kindex --emit-stub-syms
6640 @item --emit-stub-syms
6641 This option causes @command{ld} to label linker stubs with a local
6642 symbol that encodes the stub type and destination.
6643
6644 @cindex PowerPC TLS optimization
6645 @kindex --no-tls-optimize
6646 @item --no-tls-optimize
6647 PowerPC @command{ld} normally performs some optimization of code
6648 sequences used to access Thread-Local Storage. Use this option to
6649 disable the optimization.
6650 @end table
6651
6652 @ifclear GENERIC
6653 @lowersections
6654 @end ifclear
6655 @end ifset
6656
6657 @ifset POWERPC64
6658 @ifclear GENERIC
6659 @raisesections
6660 @end ifclear
6661
6662 @node PowerPC64 ELF64
6663 @section @command{ld} and PowerPC64 64-bit ELF Support
6664
6665 @cindex PowerPC64 ELF64 options
6666 @table @option
6667 @cindex PowerPC64 stub grouping
6668 @kindex --stub-group-size
6669 @item --stub-group-size
6670 Long branch stubs, PLT call stubs and TOC adjusting stubs are placed
6671 by @command{ld} in stub sections located between groups of input sections.
6672 @samp{--stub-group-size} specifies the maximum size of a group of input
6673 sections handled by one stub section. Since branch offsets are signed,
6674 a stub section may serve two groups of input sections, one group before
6675 the stub section, and one group after it. However, when using
6676 conditional branches that require stubs, it may be better (for branch
6677 prediction) that stub sections only serve one group of input sections.
6678 A negative value for @samp{N} chooses this scheme, ensuring that
6679 branches to stubs always use a negative offset. Two special values of
6680 @samp{N} are recognized, @samp{1} and @samp{-1}. These both instruct
6681 @command{ld} to automatically size input section groups for the branch types
6682 detected, with the same behaviour regarding stub placement as other
6683 positive or negative values of @samp{N} respectively.
6684
6685 Note that @samp{--stub-group-size} does not split input sections. A
6686 single input section larger than the group size specified will of course
6687 create a larger group (of one section). If input sections are too
6688 large, it may not be possible for a branch to reach its stub.
6689
6690 @cindex PowerPC64 stub symbols
6691 @kindex --emit-stub-syms
6692 @item --emit-stub-syms
6693 This option causes @command{ld} to label linker stubs with a local
6694 symbol that encodes the stub type and destination.
6695
6696 @cindex PowerPC64 dot symbols
6697 @kindex --dotsyms
6698 @kindex --no-dotsyms
6699 @item --dotsyms, --no-dotsyms
6700 These two options control how @command{ld} interprets version patterns
6701 in a version script. Older PowerPC64 compilers emitted both a
6702 function descriptor symbol with the same name as the function, and a
6703 code entry symbol with the name prefixed by a dot (@samp{.}). To
6704 properly version a function @samp{foo}, the version script thus needs
6705 to control both @samp{foo} and @samp{.foo}. The option
6706 @samp{--dotsyms}, on by default, automatically adds the required
6707 dot-prefixed patterns. Use @samp{--no-dotsyms} to disable this
6708 feature.
6709
6710 @cindex PowerPC64 TLS optimization
6711 @kindex --no-tls-optimize
6712 @item --no-tls-optimize
6713 PowerPC64 @command{ld} normally performs some optimization of code
6714 sequences used to access Thread-Local Storage. Use this option to
6715 disable the optimization.
6716
6717 @cindex PowerPC64 OPD optimization
6718 @kindex --no-opd-optimize
6719 @item --no-opd-optimize
6720 PowerPC64 @command{ld} normally removes @code{.opd} section entries
6721 corresponding to deleted link-once functions, or functions removed by
6722 the action of @samp{--gc-sections} or linker script @code{/DISCARD/}.
6723 Use this option to disable @code{.opd} optimization.
6724
6725 @cindex PowerPC64 OPD spacing
6726 @kindex --non-overlapping-opd
6727 @item --non-overlapping-opd
6728 Some PowerPC64 compilers have an option to generate compressed
6729 @code{.opd} entries spaced 16 bytes apart, overlapping the third word,
6730 the static chain pointer (unused in C) with the first word of the next
6731 entry. This option expands such entries to the full 24 bytes.
6732
6733 @cindex PowerPC64 TOC optimization
6734 @kindex --no-toc-optimize
6735 @item --no-toc-optimize
6736 PowerPC64 @command{ld} normally removes unused @code{.toc} section
6737 entries. Such entries are detected by examining relocations that
6738 reference the TOC in code sections. A reloc in a deleted code section
6739 marks a TOC word as unneeded, while a reloc in a kept code section
6740 marks a TOC word as needed. Since the TOC may reference itself, TOC
6741 relocs are also examined. TOC words marked as both needed and
6742 unneeded will of course be kept. TOC words without any referencing
6743 reloc are assumed to be part of a multi-word entry, and are kept or
6744 discarded as per the nearest marked preceding word. This works
6745 reliably for compiler generated code, but may be incorrect if assembly
6746 code is used to insert TOC entries. Use this option to disable the
6747 optimization.
6748
6749 @cindex PowerPC64 multi-TOC
6750 @kindex --no-multi-toc
6751 @item --no-multi-toc
6752 If given any toc option besides @code{-mcmodel=medium} or
6753 @code{-mcmodel=large}, PowerPC64 GCC generates code for a TOC model
6754 where TOC
6755 entries are accessed with a 16-bit offset from r2. This limits the
6756 total TOC size to 64K. PowerPC64 @command{ld} extends this limit by
6757 grouping code sections such that each group uses less than 64K for its
6758 TOC entries, then inserts r2 adjusting stubs between inter-group
6759 calls. @command{ld} does not split apart input sections, so cannot
6760 help if a single input file has a @code{.toc} section that exceeds
6761 64K, most likely from linking multiple files with @command{ld -r}.
6762 Use this option to turn off this feature.
6763
6764 @cindex PowerPC64 TOC sorting
6765 @kindex --no-toc-sort
6766 @item --no-toc-sort
6767 By default, @command{ld} sorts TOC sections so that those whose file
6768 happens to have a section called @code{.init} or @code{.fini} are
6769 placed first, followed by TOC sections referenced by code generated
6770 with PowerPC64 gcc's @code{-mcmodel=small}, and lastly TOC sections
6771 referenced only by code generated with PowerPC64 gcc's
6772 @code{-mcmodel=medium} or @code{-mcmodel=large} options. Doing this
6773 results in better TOC grouping for multi-TOC. Use this option to turn
6774 off this feature.
6775
6776 @cindex PowerPC64 PLT stub alignment
6777 @kindex --plt-align
6778 @kindex --no-plt-align
6779 @item --plt-align
6780 @itemx --no-plt-align
6781 Use these options to control whether individual PLT call stubs are
6782 aligned to a 32-byte boundary, or to the specified power of two
6783 boundary when using @code{--plt-align=}. By default PLT call stubs
6784 are packed tightly.
6785
6786 @cindex PowerPC64 PLT call stub static chain
6787 @kindex --plt-static-chain
6788 @kindex --no-plt-static-chain
6789 @item --plt-static-chain
6790 @itemx --no-plt-static-chain
6791 Use these options to control whether PLT call stubs load the static
6792 chain pointer (r11). @code{ld} defaults to not loading the static
6793 chain since there is never any need to do so on a PLT call.
6794
6795 @cindex PowerPC64 PLT call stub thread safety
6796 @kindex --plt-thread-safe
6797 @kindex --no-plt-thread-safe
6798 @item --plt-thread-safe
6799 @itemx --no-thread-safe
6800 With power7's weakly ordered memory model, it is possible when using
6801 lazy binding for ld.so to update a plt entry in one thread and have
6802 another thread see the individual plt entry words update in the wrong
6803 order, despite ld.so carefully writing in the correct order and using
6804 memory write barriers. To avoid this we need some sort of read
6805 barrier in the call stub, or use LD_BIND_NOW=1. By default, @code{ld}
6806 looks for calls to commonly used functions that create threads, and if
6807 seen, adds the necessary barriers. Use these options to change the
6808 default behaviour.
6809 @end table
6810
6811 @ifclear GENERIC
6812 @lowersections
6813 @end ifclear
6814 @end ifset
6815
6816 @ifset SPU
6817 @ifclear GENERIC
6818 @raisesections
6819 @end ifclear
6820
6821 @node SPU ELF
6822 @section @command{ld} and SPU ELF Support
6823
6824 @cindex SPU ELF options
6825 @table @option
6826
6827 @cindex SPU plugins
6828 @kindex --plugin
6829 @item --plugin
6830 This option marks an executable as a PIC plugin module.
6831
6832 @cindex SPU overlays
6833 @kindex --no-overlays
6834 @item --no-overlays
6835 Normally, @command{ld} recognizes calls to functions within overlay
6836 regions, and redirects such calls to an overlay manager via a stub.
6837 @command{ld} also provides a built-in overlay manager. This option
6838 turns off all this special overlay handling.
6839
6840 @cindex SPU overlay stub symbols
6841 @kindex --emit-stub-syms
6842 @item --emit-stub-syms
6843 This option causes @command{ld} to label overlay stubs with a local
6844 symbol that encodes the stub type and destination.
6845
6846 @cindex SPU extra overlay stubs
6847 @kindex --extra-overlay-stubs
6848 @item --extra-overlay-stubs
6849 This option causes @command{ld} to add overlay call stubs on all
6850 function calls out of overlay regions. Normally stubs are not added
6851 on calls to non-overlay regions.
6852
6853 @cindex SPU local store size
6854 @kindex --local-store=lo:hi
6855 @item --local-store=lo:hi
6856 @command{ld} usually checks that a final executable for SPU fits in
6857 the address range 0 to 256k. This option may be used to change the
6858 range. Disable the check entirely with @option{--local-store=0:0}.
6859
6860 @cindex SPU
6861 @kindex --stack-analysis
6862 @item --stack-analysis
6863 SPU local store space is limited. Over-allocation of stack space
6864 unnecessarily limits space available for code and data, while
6865 under-allocation results in runtime failures. If given this option,
6866 @command{ld} will provide an estimate of maximum stack usage.
6867 @command{ld} does this by examining symbols in code sections to
6868 determine the extents of functions, and looking at function prologues
6869 for stack adjusting instructions. A call-graph is created by looking
6870 for relocations on branch instructions. The graph is then searched
6871 for the maximum stack usage path. Note that this analysis does not
6872 find calls made via function pointers, and does not handle recursion
6873 and other cycles in the call graph. Stack usage may be
6874 under-estimated if your code makes such calls. Also, stack usage for
6875 dynamic allocation, e.g. alloca, will not be detected. If a link map
6876 is requested, detailed information about each function's stack usage
6877 and calls will be given.
6878
6879 @cindex SPU
6880 @kindex --emit-stack-syms
6881 @item --emit-stack-syms
6882 This option, if given along with @option{--stack-analysis} will result
6883 in @command{ld} emitting stack sizing symbols for each function.
6884 These take the form @code{__stack_<function_name>} for global
6885 functions, and @code{__stack_<number>_<function_name>} for static
6886 functions. @code{<number>} is the section id in hex. The value of
6887 such symbols is the stack requirement for the corresponding function.
6888 The symbol size will be zero, type @code{STT_NOTYPE}, binding
6889 @code{STB_LOCAL}, and section @code{SHN_ABS}.
6890 @end table
6891
6892 @ifclear GENERIC
6893 @lowersections
6894 @end ifclear
6895 @end ifset
6896
6897 @ifset TICOFF
6898 @ifclear GENERIC
6899 @raisesections
6900 @end ifclear
6901
6902 @node TI COFF
6903 @section @command{ld}'s Support for Various TI COFF Versions
6904 @cindex TI COFF versions
6905 @kindex --format=@var{version}
6906 The @samp{--format} switch allows selection of one of the various
6907 TI COFF versions. The latest of this writing is 2; versions 0 and 1 are
6908 also supported. The TI COFF versions also vary in header byte-order
6909 format; @command{ld} will read any version or byte order, but the output
6910 header format depends on the default specified by the specific target.
6911
6912 @ifclear GENERIC
6913 @lowersections
6914 @end ifclear
6915 @end ifset
6916
6917 @ifset WIN32
6918 @ifclear GENERIC
6919 @raisesections
6920 @end ifclear
6921
6922 @node WIN32
6923 @section @command{ld} and WIN32 (cygwin/mingw)
6924
6925 This section describes some of the win32 specific @command{ld} issues.
6926 See @ref{Options,,Command Line Options} for detailed description of the
6927 command line options mentioned here.
6928
6929 @table @emph
6930 @cindex import libraries
6931 @item import libraries
6932 The standard Windows linker creates and uses so-called import
6933 libraries, which contains information for linking to dll's. They are
6934 regular static archives and are handled as any other static
6935 archive. The cygwin and mingw ports of @command{ld} have specific
6936 support for creating such libraries provided with the
6937 @samp{--out-implib} command line option.
6938
6939 @item exporting DLL symbols
6940 @cindex exporting DLL symbols
6941 The cygwin/mingw @command{ld} has several ways to export symbols for dll's.
6942
6943 @table @emph
6944 @item using auto-export functionality
6945 @cindex using auto-export functionality
6946 By default @command{ld} exports symbols with the auto-export functionality,
6947 which is controlled by the following command line options:
6948
6949 @itemize
6950 @item --export-all-symbols [This is the default]
6951 @item --exclude-symbols
6952 @item --exclude-libs
6953 @item --exclude-modules-for-implib
6954 @item --version-script
6955 @end itemize
6956
6957 When auto-export is in operation, @command{ld} will export all the non-local
6958 (global and common) symbols it finds in a DLL, with the exception of a few
6959 symbols known to belong to the system's runtime and libraries. As it will
6960 often not be desirable to export all of a DLL's symbols, which may include
6961 private functions that are not part of any public interface, the command-line
6962 options listed above may be used to filter symbols out from the list for
6963 exporting. The @samp{--output-def} option can be used in order to see the
6964 final list of exported symbols with all exclusions taken into effect.
6965
6966 If @samp{--export-all-symbols} is not given explicitly on the
6967 command line, then the default auto-export behavior will be @emph{disabled}
6968 if either of the following are true:
6969
6970 @itemize
6971 @item A DEF file is used.
6972 @item Any symbol in any object file was marked with the __declspec(dllexport) attribute.
6973 @end itemize
6974
6975 @item using a DEF file
6976 @cindex using a DEF file
6977 Another way of exporting symbols is using a DEF file. A DEF file is
6978 an ASCII file containing definitions of symbols which should be
6979 exported when a dll is created. Usually it is named @samp{<dll
6980 name>.def} and is added as any other object file to the linker's
6981 command line. The file's name must end in @samp{.def} or @samp{.DEF}.
6982
6983 @example
6984 gcc -o <output> <objectfiles> <dll name>.def
6985 @end example
6986
6987 Using a DEF file turns off the normal auto-export behavior, unless the
6988 @samp{--export-all-symbols} option is also used.
6989
6990 Here is an example of a DEF file for a shared library called @samp{xyz.dll}:
6991
6992 @example
6993 LIBRARY "xyz.dll" BASE=0x20000000
6994
6995 EXPORTS
6996 foo
6997 bar
6998 _bar = bar
6999 another_foo = abc.dll.afoo
7000 var1 DATA
7001 doo = foo == foo2
7002 eoo DATA == var1
7003 @end example
7004
7005 This example defines a DLL with a non-default base address and seven
7006 symbols in the export table. The third exported symbol @code{_bar} is an
7007 alias for the second. The fourth symbol, @code{another_foo} is resolved
7008 by "forwarding" to another module and treating it as an alias for
7009 @code{afoo} exported from the DLL @samp{abc.dll}. The final symbol
7010 @code{var1} is declared to be a data object. The @samp{doo} symbol in
7011 export library is an alias of @samp{foo}, which gets the string name
7012 in export table @samp{foo2}. The @samp{eoo} symbol is an data export
7013 symbol, which gets in export table the name @samp{var1}.
7014
7015 The optional @code{LIBRARY <name>} command indicates the @emph{internal}
7016 name of the output DLL. If @samp{<name>} does not include a suffix,
7017 the default library suffix, @samp{.DLL} is appended.
7018
7019 When the .DEF file is used to build an application, rather than a
7020 library, the @code{NAME <name>} command should be used instead of
7021 @code{LIBRARY}. If @samp{<name>} does not include a suffix, the default
7022 executable suffix, @samp{.EXE} is appended.
7023
7024 With either @code{LIBRARY <name>} or @code{NAME <name>} the optional
7025 specification @code{BASE = <number>} may be used to specify a
7026 non-default base address for the image.
7027
7028 If neither @code{LIBRARY <name>} nor @code{NAME <name>} is specified,
7029 or they specify an empty string, the internal name is the same as the
7030 filename specified on the command line.
7031
7032 The complete specification of an export symbol is:
7033
7034 @example
7035 EXPORTS
7036 ( ( ( <name1> [ = <name2> ] )
7037 | ( <name1> = <module-name> . <external-name>))
7038 [ @@ <integer> ] [NONAME] [DATA] [CONSTANT] [PRIVATE] [== <name3>] ) *
7039 @end example
7040
7041 Declares @samp{<name1>} as an exported symbol from the DLL, or declares
7042 @samp{<name1>} as an exported alias for @samp{<name2>}; or declares
7043 @samp{<name1>} as a "forward" alias for the symbol
7044 @samp{<external-name>} in the DLL @samp{<module-name>}.
7045 Optionally, the symbol may be exported by the specified ordinal
7046 @samp{<integer>} alias. The optional @samp{<name3>} is the to be used
7047 string in import/export table for the symbol.
7048
7049 The optional keywords that follow the declaration indicate:
7050
7051 @code{NONAME}: Do not put the symbol name in the DLL's export table. It
7052 will still be exported by its ordinal alias (either the value specified
7053 by the .def specification or, otherwise, the value assigned by the
7054 linker). The symbol name, however, does remain visible in the import
7055 library (if any), unless @code{PRIVATE} is also specified.
7056
7057 @code{DATA}: The symbol is a variable or object, rather than a function.
7058 The import lib will export only an indirect reference to @code{foo} as
7059 the symbol @code{_imp__foo} (ie, @code{foo} must be resolved as
7060 @code{*_imp__foo}).
7061
7062 @code{CONSTANT}: Like @code{DATA}, but put the undecorated @code{foo} as
7063 well as @code{_imp__foo} into the import library. Both refer to the
7064 read-only import address table's pointer to the variable, not to the
7065 variable itself. This can be dangerous. If the user code fails to add
7066 the @code{dllimport} attribute and also fails to explicitly add the
7067 extra indirection that the use of the attribute enforces, the
7068 application will behave unexpectedly.
7069
7070 @code{PRIVATE}: Put the symbol in the DLL's export table, but do not put
7071 it into the static import library used to resolve imports at link time. The
7072 symbol can still be imported using the @code{LoadLibrary/GetProcAddress}
7073 API at runtime or by by using the GNU ld extension of linking directly to
7074 the DLL without an import library.
7075
7076 See ld/deffilep.y in the binutils sources for the full specification of
7077 other DEF file statements
7078
7079 @cindex creating a DEF file
7080 While linking a shared dll, @command{ld} is able to create a DEF file
7081 with the @samp{--output-def <file>} command line option.
7082
7083 @item Using decorations
7084 @cindex Using decorations
7085 Another way of marking symbols for export is to modify the source code
7086 itself, so that when building the DLL each symbol to be exported is
7087 declared as:
7088
7089 @example
7090 __declspec(dllexport) int a_variable
7091 __declspec(dllexport) void a_function(int with_args)
7092 @end example
7093
7094 All such symbols will be exported from the DLL. If, however,
7095 any of the object files in the DLL contain symbols decorated in
7096 this way, then the normal auto-export behavior is disabled, unless
7097 the @samp{--export-all-symbols} option is also used.
7098
7099 Note that object files that wish to access these symbols must @emph{not}
7100 decorate them with dllexport. Instead, they should use dllimport,
7101 instead:
7102
7103 @example
7104 __declspec(dllimport) int a_variable
7105 __declspec(dllimport) void a_function(int with_args)
7106 @end example
7107
7108 This complicates the structure of library header files, because
7109 when included by the library itself the header must declare the
7110 variables and functions as dllexport, but when included by client
7111 code the header must declare them as dllimport. There are a number
7112 of idioms that are typically used to do this; often client code can
7113 omit the __declspec() declaration completely. See
7114 @samp{--enable-auto-import} and @samp{automatic data imports} for more
7115 information.
7116 @end table
7117
7118 @cindex automatic data imports
7119 @item automatic data imports
7120 The standard Windows dll format supports data imports from dlls only
7121 by adding special decorations (dllimport/dllexport), which let the
7122 compiler produce specific assembler instructions to deal with this
7123 issue. This increases the effort necessary to port existing Un*x
7124 code to these platforms, especially for large
7125 c++ libraries and applications. The auto-import feature, which was
7126 initially provided by Paul Sokolovsky, allows one to omit the
7127 decorations to achieve a behavior that conforms to that on POSIX/Un*x
7128 platforms. This feature is enabled with the @samp{--enable-auto-import}
7129 command-line option, although it is enabled by default on cygwin/mingw.
7130 The @samp{--enable-auto-import} option itself now serves mainly to
7131 suppress any warnings that are ordinarily emitted when linked objects
7132 trigger the feature's use.
7133
7134 auto-import of variables does not always work flawlessly without
7135 additional assistance. Sometimes, you will see this message
7136
7137 "variable '<var>' can't be auto-imported. Please read the
7138 documentation for ld's @code{--enable-auto-import} for details."
7139
7140 The @samp{--enable-auto-import} documentation explains why this error
7141 occurs, and several methods that can be used to overcome this difficulty.
7142 One of these methods is the @emph{runtime pseudo-relocs} feature, described
7143 below.
7144
7145 @cindex runtime pseudo-relocation
7146 For complex variables imported from DLLs (such as structs or classes),
7147 object files typically contain a base address for the variable and an
7148 offset (@emph{addend}) within the variable--to specify a particular
7149 field or public member, for instance. Unfortunately, the runtime loader used
7150 in win32 environments is incapable of fixing these references at runtime
7151 without the additional information supplied by dllimport/dllexport decorations.
7152 The standard auto-import feature described above is unable to resolve these
7153 references.
7154
7155 The @samp{--enable-runtime-pseudo-relocs} switch allows these references to
7156 be resolved without error, while leaving the task of adjusting the references
7157 themselves (with their non-zero addends) to specialized code provided by the
7158 runtime environment. Recent versions of the cygwin and mingw environments and
7159 compilers provide this runtime support; older versions do not. However, the
7160 support is only necessary on the developer's platform; the compiled result will
7161 run without error on an older system.
7162
7163 @samp{--enable-runtime-pseudo-relocs} is not the default; it must be explicitly
7164 enabled as needed.
7165
7166 @cindex direct linking to a dll
7167 @item direct linking to a dll
7168 The cygwin/mingw ports of @command{ld} support the direct linking,
7169 including data symbols, to a dll without the usage of any import
7170 libraries. This is much faster and uses much less memory than does the
7171 traditional import library method, especially when linking large
7172 libraries or applications. When @command{ld} creates an import lib, each
7173 function or variable exported from the dll is stored in its own bfd, even
7174 though a single bfd could contain many exports. The overhead involved in
7175 storing, loading, and processing so many bfd's is quite large, and explains the
7176 tremendous time, memory, and storage needed to link against particularly
7177 large or complex libraries when using import libs.
7178
7179 Linking directly to a dll uses no extra command-line switches other than
7180 @samp{-L} and @samp{-l}, because @command{ld} already searches for a number
7181 of names to match each library. All that is needed from the developer's
7182 perspective is an understanding of this search, in order to force ld to
7183 select the dll instead of an import library.
7184
7185
7186 For instance, when ld is called with the argument @samp{-lxxx} it will attempt
7187 to find, in the first directory of its search path,
7188
7189 @example
7190 libxxx.dll.a
7191 xxx.dll.a
7192 libxxx.a
7193 xxx.lib
7194 cygxxx.dll (*)
7195 libxxx.dll
7196 xxx.dll
7197 @end example
7198
7199 before moving on to the next directory in the search path.
7200
7201 (*) Actually, this is not @samp{cygxxx.dll} but in fact is @samp{<prefix>xxx.dll},
7202 where @samp{<prefix>} is set by the @command{ld} option
7203 @samp{--dll-search-prefix=<prefix>}. In the case of cygwin, the standard gcc spec
7204 file includes @samp{--dll-search-prefix=cyg}, so in effect we actually search for
7205 @samp{cygxxx.dll}.
7206
7207 Other win32-based unix environments, such as mingw or pw32, may use other
7208 @samp{<prefix>}es, although at present only cygwin makes use of this feature. It
7209 was originally intended to help avoid name conflicts among dll's built for the
7210 various win32/un*x environments, so that (for example) two versions of a zlib dll
7211 could coexist on the same machine.
7212
7213 The generic cygwin/mingw path layout uses a @samp{bin} directory for
7214 applications and dll's and a @samp{lib} directory for the import
7215 libraries (using cygwin nomenclature):
7216
7217 @example
7218 bin/
7219 cygxxx.dll
7220 lib/
7221 libxxx.dll.a (in case of dll's)
7222 libxxx.a (in case of static archive)
7223 @end example
7224
7225 Linking directly to a dll without using the import library can be
7226 done two ways:
7227
7228 1. Use the dll directly by adding the @samp{bin} path to the link line
7229 @example
7230 gcc -Wl,-verbose -o a.exe -L../bin/ -lxxx
7231 @end example
7232
7233 However, as the dll's often have version numbers appended to their names
7234 (@samp{cygncurses-5.dll}) this will often fail, unless one specifies
7235 @samp{-L../bin -lncurses-5} to include the version. Import libs are generally
7236 not versioned, and do not have this difficulty.
7237
7238 2. Create a symbolic link from the dll to a file in the @samp{lib}
7239 directory according to the above mentioned search pattern. This
7240 should be used to avoid unwanted changes in the tools needed for
7241 making the app/dll.
7242
7243 @example
7244 ln -s bin/cygxxx.dll lib/[cyg|lib|]xxx.dll[.a]
7245 @end example
7246
7247 Then you can link without any make environment changes.
7248
7249 @example
7250 gcc -Wl,-verbose -o a.exe -L../lib/ -lxxx
7251 @end example
7252
7253 This technique also avoids the version number problems, because the following is
7254 perfectly legal
7255
7256 @example
7257 bin/
7258 cygxxx-5.dll
7259 lib/
7260 libxxx.dll.a -> ../bin/cygxxx-5.dll
7261 @end example
7262
7263 Linking directly to a dll without using an import lib will work
7264 even when auto-import features are exercised, and even when
7265 @samp{--enable-runtime-pseudo-relocs} is used.
7266
7267 Given the improvements in speed and memory usage, one might justifiably
7268 wonder why import libraries are used at all. There are three reasons:
7269
7270 1. Until recently, the link-directly-to-dll functionality did @emph{not}
7271 work with auto-imported data.
7272
7273 2. Sometimes it is necessary to include pure static objects within the
7274 import library (which otherwise contains only bfd's for indirection
7275 symbols that point to the exports of a dll). Again, the import lib
7276 for the cygwin kernel makes use of this ability, and it is not
7277 possible to do this without an import lib.
7278
7279 3. Symbol aliases can only be resolved using an import lib. This is
7280 critical when linking against OS-supplied dll's (eg, the win32 API)
7281 in which symbols are usually exported as undecorated aliases of their
7282 stdcall-decorated assembly names.
7283
7284 So, import libs are not going away. But the ability to replace
7285 true import libs with a simple symbolic link to (or a copy of)
7286 a dll, in many cases, is a useful addition to the suite of tools
7287 binutils makes available to the win32 developer. Given the
7288 massive improvements in memory requirements during linking, storage
7289 requirements, and linking speed, we expect that many developers
7290 will soon begin to use this feature whenever possible.
7291
7292 @item symbol aliasing
7293 @table @emph
7294 @item adding additional names
7295 Sometimes, it is useful to export symbols with additional names.
7296 A symbol @samp{foo} will be exported as @samp{foo}, but it can also be
7297 exported as @samp{_foo} by using special directives in the DEF file
7298 when creating the dll. This will affect also the optional created
7299 import library. Consider the following DEF file:
7300
7301 @example
7302 LIBRARY "xyz.dll" BASE=0x61000000
7303
7304 EXPORTS
7305 foo
7306 _foo = foo
7307 @end example
7308
7309 The line @samp{_foo = foo} maps the symbol @samp{foo} to @samp{_foo}.
7310
7311 Another method for creating a symbol alias is to create it in the
7312 source code using the "weak" attribute:
7313
7314 @example
7315 void foo () @{ /* Do something. */; @}
7316 void _foo () __attribute__ ((weak, alias ("foo")));
7317 @end example
7318
7319 See the gcc manual for more information about attributes and weak
7320 symbols.
7321
7322 @item renaming symbols
7323 Sometimes it is useful to rename exports. For instance, the cygwin
7324 kernel does this regularly. A symbol @samp{_foo} can be exported as
7325 @samp{foo} but not as @samp{_foo} by using special directives in the
7326 DEF file. (This will also affect the import library, if it is
7327 created). In the following example:
7328
7329 @example
7330 LIBRARY "xyz.dll" BASE=0x61000000
7331
7332 EXPORTS
7333 _foo = foo
7334 @end example
7335
7336 The line @samp{_foo = foo} maps the exported symbol @samp{foo} to
7337 @samp{_foo}.
7338 @end table
7339
7340 Note: using a DEF file disables the default auto-export behavior,
7341 unless the @samp{--export-all-symbols} command line option is used.
7342 If, however, you are trying to rename symbols, then you should list
7343 @emph{all} desired exports in the DEF file, including the symbols
7344 that are not being renamed, and do @emph{not} use the
7345 @samp{--export-all-symbols} option. If you list only the
7346 renamed symbols in the DEF file, and use @samp{--export-all-symbols}
7347 to handle the other symbols, then the both the new names @emph{and}
7348 the original names for the renamed symbols will be exported.
7349 In effect, you'd be aliasing those symbols, not renaming them,
7350 which is probably not what you wanted.
7351
7352 @cindex weak externals
7353 @item weak externals
7354 The Windows object format, PE, specifies a form of weak symbols called
7355 weak externals. When a weak symbol is linked and the symbol is not
7356 defined, the weak symbol becomes an alias for some other symbol. There
7357 are three variants of weak externals:
7358 @itemize
7359 @item Definition is searched for in objects and libraries, historically
7360 called lazy externals.
7361 @item Definition is searched for only in other objects, not in libraries.
7362 This form is not presently implemented.
7363 @item No search; the symbol is an alias. This form is not presently
7364 implemented.
7365 @end itemize
7366 As a GNU extension, weak symbols that do not specify an alternate symbol
7367 are supported. If the symbol is undefined when linking, the symbol
7368 uses a default value.
7369
7370 @cindex aligned common symbols
7371 @item aligned common symbols
7372 As a GNU extension to the PE file format, it is possible to specify the
7373 desired alignment for a common symbol. This information is conveyed from
7374 the assembler or compiler to the linker by means of GNU-specific commands
7375 carried in the object file's @samp{.drectve} section, which are recognized
7376 by @command{ld} and respected when laying out the common symbols. Native
7377 tools will be able to process object files employing this GNU extension,
7378 but will fail to respect the alignment instructions, and may issue noisy
7379 warnings about unknown linker directives.
7380 @end table
7381
7382 @ifclear GENERIC
7383 @lowersections
7384 @end ifclear
7385 @end ifset
7386
7387 @ifset XTENSA
7388 @ifclear GENERIC
7389 @raisesections
7390 @end ifclear
7391
7392 @node Xtensa
7393 @section @code{ld} and Xtensa Processors
7394
7395 @cindex Xtensa processors
7396 The default @command{ld} behavior for Xtensa processors is to interpret
7397 @code{SECTIONS} commands so that lists of explicitly named sections in a
7398 specification with a wildcard file will be interleaved when necessary to
7399 keep literal pools within the range of PC-relative load offsets. For
7400 example, with the command:
7401
7402 @smallexample
7403 SECTIONS
7404 @{
7405 .text : @{
7406 *(.literal .text)
7407 @}
7408 @}
7409 @end smallexample
7410
7411 @noindent
7412 @command{ld} may interleave some of the @code{.literal}
7413 and @code{.text} sections from different object files to ensure that the
7414 literal pools are within the range of PC-relative load offsets. A valid
7415 interleaving might place the @code{.literal} sections from an initial
7416 group of files followed by the @code{.text} sections of that group of
7417 files. Then, the @code{.literal} sections from the rest of the files
7418 and the @code{.text} sections from the rest of the files would follow.
7419
7420 @cindex @option{--relax} on Xtensa
7421 @cindex relaxing on Xtensa
7422 Relaxation is enabled by default for the Xtensa version of @command{ld} and
7423 provides two important link-time optimizations. The first optimization
7424 is to combine identical literal values to reduce code size. A redundant
7425 literal will be removed and all the @code{L32R} instructions that use it
7426 will be changed to reference an identical literal, as long as the
7427 location of the replacement literal is within the offset range of all
7428 the @code{L32R} instructions. The second optimization is to remove
7429 unnecessary overhead from assembler-generated ``longcall'' sequences of
7430 @code{L32R}/@code{CALLX@var{n}} when the target functions are within
7431 range of direct @code{CALL@var{n}} instructions.
7432
7433 For each of these cases where an indirect call sequence can be optimized
7434 to a direct call, the linker will change the @code{CALLX@var{n}}
7435 instruction to a @code{CALL@var{n}} instruction, remove the @code{L32R}
7436 instruction, and remove the literal referenced by the @code{L32R}
7437 instruction if it is not used for anything else. Removing the
7438 @code{L32R} instruction always reduces code size but can potentially
7439 hurt performance by changing the alignment of subsequent branch targets.
7440 By default, the linker will always preserve alignments, either by
7441 switching some instructions between 24-bit encodings and the equivalent
7442 density instructions or by inserting a no-op in place of the @code{L32R}
7443 instruction that was removed. If code size is more important than
7444 performance, the @option{--size-opt} option can be used to prevent the
7445 linker from widening density instructions or inserting no-ops, except in
7446 a few cases where no-ops are required for correctness.
7447
7448 The following Xtensa-specific command-line options can be used to
7449 control the linker:
7450
7451 @cindex Xtensa options
7452 @table @option
7453 @item --size-opt
7454 When optimizing indirect calls to direct calls, optimize for code size
7455 more than performance. With this option, the linker will not insert
7456 no-ops or widen density instructions to preserve branch target
7457 alignment. There may still be some cases where no-ops are required to
7458 preserve the correctness of the code.
7459 @end table
7460
7461 @ifclear GENERIC
7462 @lowersections
7463 @end ifclear
7464 @end ifset
7465
7466 @ifclear SingleFormat
7467 @node BFD
7468 @chapter BFD
7469
7470 @cindex back end
7471 @cindex object file management
7472 @cindex object formats available
7473 @kindex objdump -i
7474 The linker accesses object and archive files using the BFD libraries.
7475 These libraries allow the linker to use the same routines to operate on
7476 object files whatever the object file format. A different object file
7477 format can be supported simply by creating a new BFD back end and adding
7478 it to the library. To conserve runtime memory, however, the linker and
7479 associated tools are usually configured to support only a subset of the
7480 object file formats available. You can use @code{objdump -i}
7481 (@pxref{objdump,,objdump,binutils.info,The GNU Binary Utilities}) to
7482 list all the formats available for your configuration.
7483
7484 @cindex BFD requirements
7485 @cindex requirements for BFD
7486 As with most implementations, BFD is a compromise between
7487 several conflicting requirements. The major factor influencing
7488 BFD design was efficiency: any time used converting between
7489 formats is time which would not have been spent had BFD not
7490 been involved. This is partly offset by abstraction payback; since
7491 BFD simplifies applications and back ends, more time and care
7492 may be spent optimizing algorithms for a greater speed.
7493
7494 One minor artifact of the BFD solution which you should bear in
7495 mind is the potential for information loss. There are two places where
7496 useful information can be lost using the BFD mechanism: during
7497 conversion and during output. @xref{BFD information loss}.
7498
7499 @menu
7500 * BFD outline:: How it works: an outline of BFD
7501 @end menu
7502
7503 @node BFD outline
7504 @section How It Works: An Outline of BFD
7505 @cindex opening object files
7506 @include bfdsumm.texi
7507 @end ifclear
7508
7509 @node Reporting Bugs
7510 @chapter Reporting Bugs
7511 @cindex bugs in @command{ld}
7512 @cindex reporting bugs in @command{ld}
7513
7514 Your bug reports play an essential role in making @command{ld} reliable.
7515
7516 Reporting a bug may help you by bringing a solution to your problem, or
7517 it may not. But in any case the principal function of a bug report is
7518 to help the entire community by making the next version of @command{ld}
7519 work better. Bug reports are your contribution to the maintenance of
7520 @command{ld}.
7521
7522 In order for a bug report to serve its purpose, you must include the
7523 information that enables us to fix the bug.
7524
7525 @menu
7526 * Bug Criteria:: Have you found a bug?
7527 * Bug Reporting:: How to report bugs
7528 @end menu
7529
7530 @node Bug Criteria
7531 @section Have You Found a Bug?
7532 @cindex bug criteria
7533
7534 If you are not sure whether you have found a bug, here are some guidelines:
7535
7536 @itemize @bullet
7537 @cindex fatal signal
7538 @cindex linker crash
7539 @cindex crash of linker
7540 @item
7541 If the linker gets a fatal signal, for any input whatever, that is a
7542 @command{ld} bug. Reliable linkers never crash.
7543
7544 @cindex error on valid input
7545 @item
7546 If @command{ld} produces an error message for valid input, that is a bug.
7547
7548 @cindex invalid input
7549 @item
7550 If @command{ld} does not produce an error message for invalid input, that
7551 may be a bug. In the general case, the linker can not verify that
7552 object files are correct.
7553
7554 @item
7555 If you are an experienced user of linkers, your suggestions for
7556 improvement of @command{ld} are welcome in any case.
7557 @end itemize
7558
7559 @node Bug Reporting
7560 @section How to Report Bugs
7561 @cindex bug reports
7562 @cindex @command{ld} bugs, reporting
7563
7564 A number of companies and individuals offer support for @sc{gnu}
7565 products. If you obtained @command{ld} from a support organization, we
7566 recommend you contact that organization first.
7567
7568 You can find contact information for many support companies and
7569 individuals in the file @file{etc/SERVICE} in the @sc{gnu} Emacs
7570 distribution.
7571
7572 @ifset BUGURL
7573 Otherwise, send bug reports for @command{ld} to
7574 @value{BUGURL}.
7575 @end ifset
7576
7577 The fundamental principle of reporting bugs usefully is this:
7578 @strong{report all the facts}. If you are not sure whether to state a
7579 fact or leave it out, state it!
7580
7581 Often people omit facts because they think they know what causes the
7582 problem and assume that some details do not matter. Thus, you might
7583 assume that the name of a symbol you use in an example does not
7584 matter. Well, probably it does not, but one cannot be sure. Perhaps
7585 the bug is a stray memory reference which happens to fetch from the
7586 location where that name is stored in memory; perhaps, if the name
7587 were different, the contents of that location would fool the linker
7588 into doing the right thing despite the bug. Play it safe and give a
7589 specific, complete example. That is the easiest thing for you to do,
7590 and the most helpful.
7591
7592 Keep in mind that the purpose of a bug report is to enable us to fix
7593 the bug if it is new to us. Therefore, always write your bug reports
7594 on the assumption that the bug has not been reported previously.
7595
7596 Sometimes people give a few sketchy facts and ask, ``Does this ring a
7597 bell?'' This cannot help us fix a bug, so it is basically useless. We
7598 respond by asking for enough details to enable us to investigate.
7599 You might as well expedite matters by sending them to begin with.
7600
7601 To enable us to fix the bug, you should include all these things:
7602
7603 @itemize @bullet
7604 @item
7605 The version of @command{ld}. @command{ld} announces it if you start it with
7606 the @samp{--version} argument.
7607
7608 Without this, we will not know whether there is any point in looking for
7609 the bug in the current version of @command{ld}.
7610
7611 @item
7612 Any patches you may have applied to the @command{ld} source, including any
7613 patches made to the @code{BFD} library.
7614
7615 @item
7616 The type of machine you are using, and the operating system name and
7617 version number.
7618
7619 @item
7620 What compiler (and its version) was used to compile @command{ld}---e.g.
7621 ``@code{gcc-2.7}''.
7622
7623 @item
7624 The command arguments you gave the linker to link your example and
7625 observe the bug. To guarantee you will not omit something important,
7626 list them all. A copy of the Makefile (or the output from make) is
7627 sufficient.
7628
7629 If we were to try to guess the arguments, we would probably guess wrong
7630 and then we might not encounter the bug.
7631
7632 @item
7633 A complete input file, or set of input files, that will reproduce the
7634 bug. It is generally most helpful to send the actual object files
7635 provided that they are reasonably small. Say no more than 10K. For
7636 bigger files you can either make them available by FTP or HTTP or else
7637 state that you are willing to send the object file(s) to whomever
7638 requests them. (Note - your email will be going to a mailing list, so
7639 we do not want to clog it up with large attachments). But small
7640 attachments are best.
7641
7642 If the source files were assembled using @code{gas} or compiled using
7643 @code{gcc}, then it may be OK to send the source files rather than the
7644 object files. In this case, be sure to say exactly what version of
7645 @code{gas} or @code{gcc} was used to produce the object files. Also say
7646 how @code{gas} or @code{gcc} were configured.
7647
7648 @item
7649 A description of what behavior you observe that you believe is
7650 incorrect. For example, ``It gets a fatal signal.''
7651
7652 Of course, if the bug is that @command{ld} gets a fatal signal, then we
7653 will certainly notice it. But if the bug is incorrect output, we might
7654 not notice unless it is glaringly wrong. You might as well not give us
7655 a chance to make a mistake.
7656
7657 Even if the problem you experience is a fatal signal, you should still
7658 say so explicitly. Suppose something strange is going on, such as, your
7659 copy of @command{ld} is out of sync, or you have encountered a bug in the
7660 C library on your system. (This has happened!) Your copy might crash
7661 and ours would not. If you told us to expect a crash, then when ours
7662 fails to crash, we would know that the bug was not happening for us. If
7663 you had not told us to expect a crash, then we would not be able to draw
7664 any conclusion from our observations.
7665
7666 @item
7667 If you wish to suggest changes to the @command{ld} source, send us context
7668 diffs, as generated by @code{diff} with the @samp{-u}, @samp{-c}, or
7669 @samp{-p} option. Always send diffs from the old file to the new file.
7670 If you even discuss something in the @command{ld} source, refer to it by
7671 context, not by line number.
7672
7673 The line numbers in our development sources will not match those in your
7674 sources. Your line numbers would convey no useful information to us.
7675 @end itemize
7676
7677 Here are some things that are not necessary:
7678
7679 @itemize @bullet
7680 @item
7681 A description of the envelope of the bug.
7682
7683 Often people who encounter a bug spend a lot of time investigating
7684 which changes to the input file will make the bug go away and which
7685 changes will not affect it.
7686
7687 This is often time consuming and not very useful, because the way we
7688 will find the bug is by running a single example under the debugger
7689 with breakpoints, not by pure deduction from a series of examples.
7690 We recommend that you save your time for something else.
7691
7692 Of course, if you can find a simpler example to report @emph{instead}
7693 of the original one, that is a convenience for us. Errors in the
7694 output will be easier to spot, running under the debugger will take
7695 less time, and so on.
7696
7697 However, simplification is not vital; if you do not want to do this,
7698 report the bug anyway and send us the entire test case you used.
7699
7700 @item
7701 A patch for the bug.
7702
7703 A patch for the bug does help us if it is a good one. But do not omit
7704 the necessary information, such as the test case, on the assumption that
7705 a patch is all we need. We might see problems with your patch and decide
7706 to fix the problem another way, or we might not understand it at all.
7707
7708 Sometimes with a program as complicated as @command{ld} it is very hard to
7709 construct an example that will make the program follow a certain path
7710 through the code. If you do not send us the example, we will not be
7711 able to construct one, so we will not be able to verify that the bug is
7712 fixed.
7713
7714 And if we cannot understand what bug you are trying to fix, or why your
7715 patch should be an improvement, we will not install it. A test case will
7716 help us to understand.
7717
7718 @item
7719 A guess about what the bug is or what it depends on.
7720
7721 Such guesses are usually wrong. Even we cannot guess right about such
7722 things without first using the debugger to find the facts.
7723 @end itemize
7724
7725 @node MRI
7726 @appendix MRI Compatible Script Files
7727 @cindex MRI compatibility
7728 To aid users making the transition to @sc{gnu} @command{ld} from the MRI
7729 linker, @command{ld} can use MRI compatible linker scripts as an
7730 alternative to the more general-purpose linker scripting language
7731 described in @ref{Scripts}. MRI compatible linker scripts have a much
7732 simpler command set than the scripting language otherwise used with
7733 @command{ld}. @sc{gnu} @command{ld} supports the most commonly used MRI
7734 linker commands; these commands are described here.
7735
7736 In general, MRI scripts aren't of much use with the @code{a.out} object
7737 file format, since it only has three sections and MRI scripts lack some
7738 features to make use of them.
7739
7740 You can specify a file containing an MRI-compatible script using the
7741 @samp{-c} command-line option.
7742
7743 Each command in an MRI-compatible script occupies its own line; each
7744 command line starts with the keyword that identifies the command (though
7745 blank lines are also allowed for punctuation). If a line of an
7746 MRI-compatible script begins with an unrecognized keyword, @command{ld}
7747 issues a warning message, but continues processing the script.
7748
7749 Lines beginning with @samp{*} are comments.
7750
7751 You can write these commands using all upper-case letters, or all
7752 lower case; for example, @samp{chip} is the same as @samp{CHIP}.
7753 The following list shows only the upper-case form of each command.
7754
7755 @table @code
7756 @cindex @code{ABSOLUTE} (MRI)
7757 @item ABSOLUTE @var{secname}
7758 @itemx ABSOLUTE @var{secname}, @var{secname}, @dots{} @var{secname}
7759 Normally, @command{ld} includes in the output file all sections from all
7760 the input files. However, in an MRI-compatible script, you can use the
7761 @code{ABSOLUTE} command to restrict the sections that will be present in
7762 your output program. If the @code{ABSOLUTE} command is used at all in a
7763 script, then only the sections named explicitly in @code{ABSOLUTE}
7764 commands will appear in the linker output. You can still use other
7765 input sections (whatever you select on the command line, or using
7766 @code{LOAD}) to resolve addresses in the output file.
7767
7768 @cindex @code{ALIAS} (MRI)
7769 @item ALIAS @var{out-secname}, @var{in-secname}
7770 Use this command to place the data from input section @var{in-secname}
7771 in a section called @var{out-secname} in the linker output file.
7772
7773 @var{in-secname} may be an integer.
7774
7775 @cindex @code{ALIGN} (MRI)
7776 @item ALIGN @var{secname} = @var{expression}
7777 Align the section called @var{secname} to @var{expression}. The
7778 @var{expression} should be a power of two.
7779
7780 @cindex @code{BASE} (MRI)
7781 @item BASE @var{expression}
7782 Use the value of @var{expression} as the lowest address (other than
7783 absolute addresses) in the output file.
7784
7785 @cindex @code{CHIP} (MRI)
7786 @item CHIP @var{expression}
7787 @itemx CHIP @var{expression}, @var{expression}
7788 This command does nothing; it is accepted only for compatibility.
7789
7790 @cindex @code{END} (MRI)
7791 @item END
7792 This command does nothing whatever; it's only accepted for compatibility.
7793
7794 @cindex @code{FORMAT} (MRI)
7795 @item FORMAT @var{output-format}
7796 Similar to the @code{OUTPUT_FORMAT} command in the more general linker
7797 language, but restricted to one of these output formats:
7798
7799 @enumerate
7800 @item
7801 S-records, if @var{output-format} is @samp{S}
7802
7803 @item
7804 IEEE, if @var{output-format} is @samp{IEEE}
7805
7806 @item
7807 COFF (the @samp{coff-m68k} variant in BFD), if @var{output-format} is
7808 @samp{COFF}
7809 @end enumerate
7810
7811 @cindex @code{LIST} (MRI)
7812 @item LIST @var{anything}@dots{}
7813 Print (to the standard output file) a link map, as produced by the
7814 @command{ld} command-line option @samp{-M}.
7815
7816 The keyword @code{LIST} may be followed by anything on the
7817 same line, with no change in its effect.
7818
7819 @cindex @code{LOAD} (MRI)
7820 @item LOAD @var{filename}
7821 @itemx LOAD @var{filename}, @var{filename}, @dots{} @var{filename}
7822 Include one or more object file @var{filename} in the link; this has the
7823 same effect as specifying @var{filename} directly on the @command{ld}
7824 command line.
7825
7826 @cindex @code{NAME} (MRI)
7827 @item NAME @var{output-name}
7828 @var{output-name} is the name for the program produced by @command{ld}; the
7829 MRI-compatible command @code{NAME} is equivalent to the command-line
7830 option @samp{-o} or the general script language command @code{OUTPUT}.
7831
7832 @cindex @code{ORDER} (MRI)
7833 @item ORDER @var{secname}, @var{secname}, @dots{} @var{secname}
7834 @itemx ORDER @var{secname} @var{secname} @var{secname}
7835 Normally, @command{ld} orders the sections in its output file in the
7836 order in which they first appear in the input files. In an MRI-compatible
7837 script, you can override this ordering with the @code{ORDER} command. The
7838 sections you list with @code{ORDER} will appear first in your output
7839 file, in the order specified.
7840
7841 @cindex @code{PUBLIC} (MRI)
7842 @item PUBLIC @var{name}=@var{expression}
7843 @itemx PUBLIC @var{name},@var{expression}
7844 @itemx PUBLIC @var{name} @var{expression}
7845 Supply a value (@var{expression}) for external symbol
7846 @var{name} used in the linker input files.
7847
7848 @cindex @code{SECT} (MRI)
7849 @item SECT @var{secname}, @var{expression}
7850 @itemx SECT @var{secname}=@var{expression}
7851 @itemx SECT @var{secname} @var{expression}
7852 You can use any of these three forms of the @code{SECT} command to
7853 specify the start address (@var{expression}) for section @var{secname}.
7854 If you have more than one @code{SECT} statement for the same
7855 @var{secname}, only the @emph{first} sets the start address.
7856 @end table
7857
7858 @node GNU Free Documentation License
7859 @appendix GNU Free Documentation License
7860 @include fdl.texi
7861
7862 @node LD Index
7863 @unnumbered LD Index
7864
7865 @printindex cp
7866
7867 @tex
7868 % I think something like @colophon should be in texinfo. In the
7869 % meantime:
7870 \long\def\colophon{\hbox to0pt{}\vfill
7871 \centerline{The body of this manual is set in}
7872 \centerline{\fontname\tenrm,}
7873 \centerline{with headings in {\bf\fontname\tenbf}}
7874 \centerline{and examples in {\tt\fontname\tentt}.}
7875 \centerline{{\it\fontname\tenit\/} and}
7876 \centerline{{\sl\fontname\tensl\/}}
7877 \centerline{are used for emphasis.}\vfill}
7878 \page\colophon
7879 % Blame: doc@cygnus.com, 28mar91.
7880 @end tex
7881
7882 @bye
This page took 0.253427 seconds and 4 git commands to generate.