Commit | Line | Data |
---|---|---|
252b5132 RH |
1 | \input texinfo |
2 | @setfilename ld.info | |
a2b64bed | 3 | @c Copyright 1991, 1992, 1993, 1994, 1995, 1996, 1997, 1998, 1999, 2000, |
8a308ae8 NC |
4 | @c 2001, 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009 |
5 | @c Free Software Foundation, Inc. | |
252b5132 | 6 | @syncodeindex ky cp |
dff70155 | 7 | @c man begin INCLUDE |
252b5132 RH |
8 | @include configdoc.texi |
9 | @c (configdoc.texi is generated by the Makefile) | |
c428fa83 | 10 | @include bfdver.texi |
dff70155 | 11 | @c man end |
252b5132 RH |
12 | |
13 | @c @smallbook | |
14 | ||
ff5dcc92 SC |
15 | @macro gcctabopt{body} |
16 | @code{\body\} | |
17 | @end macro | |
18 | ||
0285c67d NC |
19 | @c man begin NAME |
20 | @ifset man | |
21 | @c Configure for the generation of man pages | |
22 | @set UsesEnvVars | |
23 | @set GENERIC | |
0285c67d | 24 | @set ARM |
49fa1e15 | 25 | @set H8300 |
0285c67d | 26 | @set HPPA |
0285c67d | 27 | @set I960 |
0285c67d | 28 | @set M68HC11 |
7fb9f789 | 29 | @set M68K |
3c3bdf30 | 30 | @set MMIX |
2469cfa2 | 31 | @set MSP430 |
2a60a7a8 AM |
32 | @set POWERPC |
33 | @set POWERPC64 | |
49fa1e15 AM |
34 | @set Renesas |
35 | @set SPU | |
36 | @set TICOFF | |
2ca22b03 | 37 | @set WIN32 |
e0001a05 | 38 | @set XTENSA |
0285c67d NC |
39 | @end ifset |
40 | @c man end | |
41 | ||
252b5132 RH |
42 | @ifinfo |
43 | @format | |
44 | START-INFO-DIR-ENTRY | |
45 | * Ld: (ld). The GNU linker. | |
46 | END-INFO-DIR-ENTRY | |
47 | @end format | |
48 | @end ifinfo | |
49 | ||
0e9517a9 | 50 | @copying |
e49e529d JM |
51 | This file documents the @sc{gnu} linker LD |
52 | @ifset VERSION_PACKAGE | |
53 | @value{VERSION_PACKAGE} | |
54 | @end ifset | |
55 | version @value{VERSION}. | |
252b5132 | 56 | |
0e9517a9 | 57 | Copyright @copyright{} 1991, 92, 93, 94, 95, 96, 97, 98, 99, 2000, |
8a308ae8 | 58 | 2001, 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009 Free Software Foundation, Inc. |
252b5132 | 59 | |
cf055d54 | 60 | Permission is granted to copy, distribute and/or modify this document |
793c5807 | 61 | under the terms of the GNU Free Documentation License, Version 1.3 |
cf055d54 NC |
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 | |
36f63dca | 65 | section entitled ``GNU Free Documentation License''. |
0e9517a9 | 66 | @end copying |
252b5132 RH |
67 | @iftex |
68 | @finalout | |
69 | @setchapternewpage odd | |
71ba23f6 | 70 | @settitle The GNU linker |
252b5132 | 71 | @titlepage |
71ba23f6 | 72 | @title The GNU linker |
252b5132 | 73 | @sp 1 |
e49e529d JM |
74 | @subtitle @code{ld} |
75 | @ifset VERSION_PACKAGE | |
76 | @subtitle @value{VERSION_PACKAGE} | |
77 | @end ifset | |
252b5132 RH |
78 | @subtitle Version @value{VERSION} |
79 | @author Steve Chamberlain | |
80 | @author Ian Lance Taylor | |
252b5132 RH |
81 | @page |
82 | ||
83 | @tex | |
84 | {\parskip=0pt | |
704c465c NC |
85 | \hfill Red Hat Inc\par |
86 | \hfill nickc\@credhat.com, doc\@redhat.com\par | |
71ba23f6 | 87 | \hfill {\it The GNU linker}\par |
252b5132 RH |
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 | |
0285c67d | 94 | @c man begin COPYRIGHT |
9c8ebd6a | 95 | Copyright @copyright{} 1991, 92, 93, 94, 95, 96, 97, 98, 99, 2000, 2001, |
903249d7 | 96 | 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009 Free Software Foundation, Inc. |
252b5132 | 97 | |
0285c67d | 98 | Permission is granted to copy, distribute and/or modify this document |
793c5807 | 99 | under the terms of the GNU Free Documentation License, Version 1.3 |
0285c67d NC |
100 | or any later version published by the Free Software Foundation; |
101 | with no Invariant Sections, with no Front-Cover Texts, and with no | |
102 | Back-Cover Texts. A copy of the license is included in the | |
36f63dca | 103 | section entitled ``GNU Free Documentation License''. |
0285c67d | 104 | @c man end |
252b5132 | 105 | |
252b5132 RH |
106 | @end titlepage |
107 | @end iftex | |
4ecceb71 | 108 | @contents |
252b5132 RH |
109 | @c FIXME: Talk about importance of *order* of args, cmds to linker! |
110 | ||
84ec0e6d | 111 | @ifnottex |
252b5132 | 112 | @node Top |
71ba23f6 | 113 | @top LD |
e49e529d JM |
114 | This file documents the @sc{gnu} linker ld |
115 | @ifset VERSION_PACKAGE | |
116 | @value{VERSION_PACKAGE} | |
117 | @end ifset | |
118 | version @value{VERSION}. | |
252b5132 | 119 | |
cf055d54 | 120 | This document is distributed under the terms of the GNU Free |
793c5807 NC |
121 | Documentation License version 1.3. A copy of the license is included |
122 | in the section entitled ``GNU Free Documentation License''. | |
cf055d54 | 123 | |
252b5132 RH |
124 | @menu |
125 | * Overview:: Overview | |
126 | * Invocation:: Invocation | |
127 | * Scripts:: Linker Scripts | |
128 | @ifset GENERIC | |
129 | * Machine Dependent:: Machine Dependent Features | |
130 | @end ifset | |
131 | @ifclear GENERIC | |
132 | @ifset H8300 | |
133 | * H8/300:: ld and the H8/300 | |
134 | @end ifset | |
c2dcd04e NC |
135 | @ifset Renesas |
136 | * Renesas:: ld and other Renesas micros | |
252b5132 RH |
137 | @end ifset |
138 | @ifset I960 | |
139 | * i960:: ld and the Intel 960 family | |
140 | @end ifset | |
36f63dca NC |
141 | @ifset ARM |
142 | * ARM:: ld and the ARM family | |
143 | @end ifset | |
144 | @ifset HPPA | |
145 | * HPPA ELF32:: ld and HPPA 32-bit ELF | |
146 | @end ifset | |
93fd0973 SC |
147 | @ifset M68HC11 |
148 | * M68HC11/68HC12:: ld and the Motorola 68HC11 and 68HC12 families | |
149 | @end ifset | |
7fb9f789 NC |
150 | @ifset M68K |
151 | * M68K:: ld and Motorola 68K family | |
152 | @end ifset | |
2a60a7a8 AM |
153 | @ifset POWERPC |
154 | * PowerPC ELF32:: ld and PowerPC 32-bit ELF Support | |
155 | @end ifset | |
156 | @ifset POWERPC64 | |
157 | * PowerPC64 ELF64:: ld and PowerPC64 64-bit ELF Support | |
158 | @end ifset | |
49fa1e15 AM |
159 | @ifset SPU |
160 | * SPU ELF:: ld and SPU ELF Support | |
161 | @end ifset | |
74459f0e TW |
162 | @ifset TICOFF |
163 | * TI COFF:: ld and the TI COFF | |
164 | @end ifset | |
2ca22b03 NC |
165 | @ifset WIN32 |
166 | * Win32:: ld and WIN32 (cygwin/mingw) | |
167 | @end ifset | |
e0001a05 NC |
168 | @ifset XTENSA |
169 | * Xtensa:: ld and Xtensa Processors | |
170 | @end ifset | |
252b5132 RH |
171 | @end ifclear |
172 | @ifclear SingleFormat | |
173 | * BFD:: BFD | |
174 | @end ifclear | |
175 | @c Following blank line required for remaining bug in makeinfo conds/menus | |
176 | ||
177 | * Reporting Bugs:: Reporting Bugs | |
178 | * MRI:: MRI Compatible Script Files | |
704c465c | 179 | * GNU Free Documentation License:: GNU Free Documentation License |
370b66a1 | 180 | * LD Index:: LD Index |
252b5132 | 181 | @end menu |
84ec0e6d | 182 | @end ifnottex |
252b5132 RH |
183 | |
184 | @node Overview | |
185 | @chapter Overview | |
186 | ||
187 | @cindex @sc{gnu} linker | |
188 | @cindex what is this? | |
0285c67d | 189 | |
0879a67a | 190 | @ifset man |
0285c67d | 191 | @c man begin SYNOPSIS |
ff5dcc92 | 192 | ld [@b{options}] @var{objfile} @dots{} |
0285c67d NC |
193 | @c man end |
194 | ||
195 | @c man begin SEEALSO | |
196 | ar(1), nm(1), objcopy(1), objdump(1), readelf(1) and | |
197 | the Info entries for @file{binutils} and | |
198 | @file{ld}. | |
199 | @c man end | |
200 | @end ifset | |
201 | ||
202 | @c man begin DESCRIPTION | |
203 | ||
ff5dcc92 | 204 | @command{ld} combines a number of object and archive files, relocates |
252b5132 | 205 | their data and ties up symbol references. Usually the last step in |
ff5dcc92 | 206 | compiling a program is to run @command{ld}. |
252b5132 | 207 | |
ff5dcc92 | 208 | @command{ld} accepts Linker Command Language files written in |
252b5132 RH |
209 | a superset of AT&T's Link Editor Command Language syntax, |
210 | to provide explicit and total control over the linking process. | |
211 | ||
0285c67d NC |
212 | @ifset man |
213 | @c For the man only | |
ece2d90e | 214 | This man page does not describe the command language; see the |
71ba23f6 NC |
215 | @command{ld} entry in @code{info} for full details on the command |
216 | language and on other aspects of the GNU linker. | |
0285c67d NC |
217 | @end ifset |
218 | ||
252b5132 | 219 | @ifclear SingleFormat |
ff5dcc92 SC |
220 | This version of @command{ld} uses the general purpose BFD libraries |
221 | to operate on object files. This allows @command{ld} to read, combine, and | |
252b5132 RH |
222 | write object files in many different formats---for example, COFF or |
223 | @code{a.out}. Different formats may be linked together to produce any | |
224 | available kind of object file. @xref{BFD}, for more information. | |
225 | @end ifclear | |
226 | ||
227 | Aside from its flexibility, the @sc{gnu} linker is more helpful than other | |
228 | linkers in providing diagnostic information. Many linkers abandon | |
229 | execution immediately upon encountering an error; whenever possible, | |
ff5dcc92 | 230 | @command{ld} continues executing, allowing you to identify other errors |
252b5132 RH |
231 | (or, in some cases, to get an output file in spite of the error). |
232 | ||
0285c67d NC |
233 | @c man end |
234 | ||
252b5132 RH |
235 | @node Invocation |
236 | @chapter Invocation | |
237 | ||
0285c67d NC |
238 | @c man begin DESCRIPTION |
239 | ||
ff5dcc92 | 240 | The @sc{gnu} linker @command{ld} is meant to cover a broad range of situations, |
252b5132 RH |
241 | and to be as compatible as possible with other linkers. As a result, |
242 | you have many choices to control its behavior. | |
243 | ||
0285c67d NC |
244 | @c man end |
245 | ||
252b5132 RH |
246 | @ifset UsesEnvVars |
247 | @menu | |
248 | * Options:: Command Line Options | |
249 | * Environment:: Environment Variables | |
250 | @end menu | |
251 | ||
252 | @node Options | |
253 | @section Command Line Options | |
254 | @end ifset | |
255 | ||
256 | @cindex command line | |
257 | @cindex options | |
0285c67d NC |
258 | |
259 | @c man begin OPTIONS | |
260 | ||
252b5132 RH |
261 | The linker supports a plethora of command-line options, but in actual |
262 | practice few of them are used in any particular context. | |
263 | @cindex standard Unix system | |
ff5dcc92 | 264 | For instance, a frequent use of @command{ld} is to link standard Unix |
252b5132 RH |
265 | object files on a standard, supported Unix system. On such a system, to |
266 | link a file @code{hello.o}: | |
267 | ||
268 | @smallexample | |
269 | ld -o @var{output} /lib/crt0.o hello.o -lc | |
270 | @end smallexample | |
271 | ||
ff5dcc92 | 272 | This tells @command{ld} to produce a file called @var{output} as the |
252b5132 RH |
273 | result of linking the file @code{/lib/crt0.o} with @code{hello.o} and |
274 | the library @code{libc.a}, which will come from the standard search | |
275 | directories. (See the discussion of the @samp{-l} option below.) | |
276 | ||
ff5dcc92 | 277 | Some of the command-line options to @command{ld} may be specified at any |
511ab9e9 ILT |
278 | point in the command line. However, options which refer to files, such |
279 | as @samp{-l} or @samp{-T}, cause the file to be read at the point at | |
280 | which the option appears in the command line, relative to the object | |
281 | files and other file options. Repeating non-file options with a | |
282 | different argument will either have no further effect, or override prior | |
252b5132 RH |
283 | occurrences (those further to the left on the command line) of that |
284 | option. Options which may be meaningfully specified more than once are | |
285 | noted in the descriptions below. | |
286 | ||
287 | @cindex object files | |
511ab9e9 ILT |
288 | Non-option arguments are object files or archives which are to be linked |
289 | together. They may follow, precede, or be mixed in with command-line | |
290 | options, except that an object file argument may not be placed between | |
291 | an option and its argument. | |
252b5132 RH |
292 | |
293 | Usually the linker is invoked with at least one object file, but you can | |
294 | specify other forms of binary input files using @samp{-l}, @samp{-R}, | |
295 | and the script command language. If @emph{no} binary input files at all | |
296 | are specified, the linker does not produce any output, and issues the | |
297 | message @samp{No input files}. | |
298 | ||
36f63dca | 299 | If the linker cannot recognize the format of an object file, it will |
252b5132 RH |
300 | assume that it is a linker script. A script specified in this way |
301 | augments the main linker script used for the link (either the default | |
302 | linker script or the one specified by using @samp{-T}). This feature | |
303 | permits the linker to link against a file which appears to be an object | |
304 | or an archive, but actually merely defines some symbol values, or uses | |
53d25da6 AM |
305 | @code{INPUT} or @code{GROUP} to load other objects. Specifying a |
306 | script in this way merely augments the main linker script, with the | |
307 | extra commands placed after the main script; use the @samp{-T} option | |
308 | to replace the default linker script entirely, but note the effect of | |
309 | the @code{INSERT} command. @xref{Scripts}. | |
252b5132 RH |
310 | |
311 | For options whose names are a single letter, | |
312 | option arguments must either follow the option letter without intervening | |
313 | whitespace, or be given as separate arguments immediately following the | |
314 | option that requires them. | |
315 | ||
316 | For options whose names are multiple letters, either one dash or two can | |
e4897a32 | 317 | precede the option name; for example, @samp{-trace-symbol} and |
36f63dca | 318 | @samp{--trace-symbol} are equivalent. Note---there is one exception to |
e4897a32 | 319 | this rule. Multiple letter options that start with a lower case 'o' can |
ba1be17e | 320 | only be preceded by two dashes. This is to reduce confusion with the |
e4897a32 NC |
321 | @samp{-o} option. So for example @samp{-omagic} sets the output file |
322 | name to @samp{magic} whereas @samp{--omagic} sets the NMAGIC flag on the | |
323 | output. | |
324 | ||
325 | Arguments to multiple-letter options must either be separated from the | |
326 | option name by an equals sign, or be given as separate arguments | |
327 | immediately following the option that requires them. For example, | |
328 | @samp{--trace-symbol foo} and @samp{--trace-symbol=foo} are equivalent. | |
329 | Unique abbreviations of the names of multiple-letter options are | |
330 | accepted. | |
252b5132 | 331 | |
36f63dca NC |
332 | Note---if the linker is being invoked indirectly, via a compiler driver |
333 | (e.g. @samp{gcc}) then all the linker command line options should be | |
fa19fce0 NC |
334 | prefixed by @samp{-Wl,} (or whatever is appropriate for the particular |
335 | compiler driver) like this: | |
4e53152f NC |
336 | |
337 | @smallexample | |
2509a395 | 338 | gcc -Wl,--start-group foo.o bar.o -Wl,--end-group |
4e53152f NC |
339 | @end smallexample |
340 | ||
341 | This is important, because otherwise the compiler driver program may | |
2509a395 SL |
342 | silently drop the linker options, resulting in a bad link. Confusion |
343 | may also arise when passing options that require values through a | |
344 | driver, as the use of a space between option and argument acts as | |
345 | a separator, and causes the driver to pass only the option to the linker | |
346 | and the argument to the compiler. In this case, it is simplest to use | |
347 | the joined forms of both single- and multiple-letter options, such as: | |
348 | ||
349 | @smallexample | |
350 | gcc foo.o bar.o -Wl,-eENTRY -Wl,-Map=a.map | |
351 | @end smallexample | |
4e53152f NC |
352 | |
353 | Here is a table of the generic command line switches accepted by the GNU | |
354 | linker: | |
355 | ||
ff5dcc92 | 356 | @table @gcctabopt |
38fc1cb1 | 357 | @include at-file.texi |
dff70155 | 358 | |
2509a395 SL |
359 | @kindex -a @var{keyword} |
360 | @item -a @var{keyword} | |
252b5132 RH |
361 | This option is supported for HP/UX compatibility. The @var{keyword} |
362 | argument must be one of the strings @samp{archive}, @samp{shared}, or | |
363 | @samp{default}. @samp{-aarchive} is functionally equivalent to | |
364 | @samp{-Bstatic}, and the other two keywords are functionally equivalent | |
365 | to @samp{-Bdynamic}. This option may be used any number of times. | |
366 | ||
7ee314fa AM |
367 | @kindex --audit @var{AUDITLIB} |
368 | @item --audit @var{AUDITLIB} | |
369 | Adds @var{AUDITLIB} to the @code{DT_AUDIT} entry of the dynamic section. | |
370 | @var{AUDITLIB} is not checked for existence, nor will it use the DT_SONAME | |
371 | specified in the library. If specified multiple times @code{DT_AUDIT} | |
372 | will contain a colon separated list of audit interfaces to use. If the linker | |
373 | finds an object with an audit entry while searching for shared libraries, | |
374 | it will add a corresponding @code{DT_DEPAUDIT} entry in the output file. | |
375 | This option is only meaningful on ELF platforms supporting the rtld-audit | |
376 | interface. | |
377 | ||
252b5132 RH |
378 | @ifset I960 |
379 | @cindex architectures | |
2509a395 SL |
380 | @kindex -A @var{arch} |
381 | @item -A @var{architecture} | |
252b5132 RH |
382 | @kindex --architecture=@var{arch} |
383 | @itemx --architecture=@var{architecture} | |
ff5dcc92 SC |
384 | In the current release of @command{ld}, this option is useful only for the |
385 | Intel 960 family of architectures. In that @command{ld} configuration, the | |
252b5132 RH |
386 | @var{architecture} argument identifies the particular architecture in |
387 | the 960 family, enabling some safeguards and modifying the | |
ff5dcc92 | 388 | archive-library search path. @xref{i960,,@command{ld} and the Intel 960 |
252b5132 RH |
389 | family}, for details. |
390 | ||
ff5dcc92 | 391 | Future releases of @command{ld} may support similar functionality for |
252b5132 RH |
392 | other architecture families. |
393 | @end ifset | |
394 | ||
395 | @ifclear SingleFormat | |
396 | @cindex binary input format | |
397 | @kindex -b @var{format} | |
398 | @kindex --format=@var{format} | |
399 | @cindex input format | |
400 | @cindex input format | |
401 | @item -b @var{input-format} | |
402 | @itemx --format=@var{input-format} | |
ff5dcc92 SC |
403 | @command{ld} may be configured to support more than one kind of object |
404 | file. If your @command{ld} is configured this way, you can use the | |
252b5132 | 405 | @samp{-b} option to specify the binary format for input object files |
ff5dcc92 | 406 | that follow this option on the command line. Even when @command{ld} is |
252b5132 | 407 | configured to support alternative object formats, you don't usually need |
ff5dcc92 | 408 | to specify this, as @command{ld} should be configured to expect as a |
252b5132 RH |
409 | default input format the most usual format on each machine. |
410 | @var{input-format} is a text string, the name of a particular format | |
411 | supported by the BFD libraries. (You can list the available binary | |
412 | formats with @samp{objdump -i}.) | |
413 | @xref{BFD}. | |
414 | ||
415 | You may want to use this option if you are linking files with an unusual | |
416 | binary format. You can also use @samp{-b} to switch formats explicitly (when | |
417 | linking object files of different formats), by including | |
418 | @samp{-b @var{input-format}} before each group of object files in a | |
a1ab1d2a | 419 | particular format. |
252b5132 RH |
420 | |
421 | The default format is taken from the environment variable | |
422 | @code{GNUTARGET}. | |
423 | @ifset UsesEnvVars | |
424 | @xref{Environment}. | |
425 | @end ifset | |
426 | You can also define the input format from a script, using the command | |
0285c67d NC |
427 | @code{TARGET}; |
428 | @ifclear man | |
429 | see @ref{Format Commands}. | |
430 | @end ifclear | |
252b5132 RH |
431 | @end ifclear |
432 | ||
433 | @kindex -c @var{MRI-cmdfile} | |
434 | @kindex --mri-script=@var{MRI-cmdfile} | |
435 | @cindex compatibility, MRI | |
436 | @item -c @var{MRI-commandfile} | |
437 | @itemx --mri-script=@var{MRI-commandfile} | |
ff5dcc92 | 438 | For compatibility with linkers produced by MRI, @command{ld} accepts script |
252b5132 | 439 | files written in an alternate, restricted command language, described in |
0285c67d NC |
440 | @ifclear man |
441 | @ref{MRI,,MRI Compatible Script Files}. | |
442 | @end ifclear | |
443 | @ifset man | |
444 | the MRI Compatible Script Files section of GNU ld documentation. | |
445 | @end ifset | |
446 | Introduce MRI script files with | |
252b5132 | 447 | the option @samp{-c}; use the @samp{-T} option to run linker |
ff5dcc92 SC |
448 | scripts written in the general-purpose @command{ld} scripting language. |
449 | If @var{MRI-cmdfile} does not exist, @command{ld} looks for it in the directories | |
252b5132 RH |
450 | specified by any @samp{-L} options. |
451 | ||
452 | @cindex common allocation | |
453 | @kindex -d | |
454 | @kindex -dc | |
455 | @kindex -dp | |
a1ab1d2a | 456 | @item -d |
252b5132 RH |
457 | @itemx -dc |
458 | @itemx -dp | |
459 | These three options are equivalent; multiple forms are supported for | |
460 | compatibility with other linkers. They assign space to common symbols | |
461 | even if a relocatable output file is specified (with @samp{-r}). The | |
462 | script command @code{FORCE_COMMON_ALLOCATION} has the same effect. | |
463 | @xref{Miscellaneous Commands}. | |
464 | ||
7ee314fa AM |
465 | @kindex --depaudit @var{AUDITLIB} |
466 | @kindex -P @var{AUDITLIB} | |
467 | @item --depaudit @var{AUDITLIB} | |
468 | @itemx -P @var{AUDITLIB} | |
469 | Adds @var{AUDITLIB} to the @code{DT_DEPAUDIT} entry of the dynamic section. | |
470 | @var{AUDITLIB} is not checked for existence, nor will it use the DT_SONAME | |
471 | specified in the library. If specified multiple times @code{DT_DEPAUDIT} | |
472 | will contain a colon separated list of audit interfaces to use. This | |
473 | option is only meaningful on ELF platforms supporting the rtld-audit interface. | |
474 | The -P option is provided for Solaris compatibility. | |
475 | ||
252b5132 RH |
476 | @cindex entry point, from command line |
477 | @kindex -e @var{entry} | |
478 | @kindex --entry=@var{entry} | |
a1ab1d2a | 479 | @item -e @var{entry} |
252b5132 RH |
480 | @itemx --entry=@var{entry} |
481 | Use @var{entry} as the explicit symbol for beginning execution of your | |
482 | program, rather than the default entry point. If there is no symbol | |
483 | named @var{entry}, the linker will try to parse @var{entry} as a number, | |
484 | and use that as the entry address (the number will be interpreted in | |
485 | base 10; you may use a leading @samp{0x} for base 16, or a leading | |
486 | @samp{0} for base 8). @xref{Entry Point}, for a discussion of defaults | |
487 | and other ways of specifying the entry point. | |
488 | ||
b58f81ae DJ |
489 | @kindex --exclude-libs |
490 | @item --exclude-libs @var{lib},@var{lib},... | |
491 | Specifies a list of archive libraries from which symbols should not be automatically | |
e1c37eb5 | 492 | exported. The library names may be delimited by commas or colons. Specifying |
b58f81ae DJ |
493 | @code{--exclude-libs ALL} excludes symbols in all archive libraries from |
494 | automatic export. This option is available only for the i386 PE targeted | |
495 | port of the linker and for ELF targeted ports. For i386 PE, symbols | |
496 | explicitly listed in a .def file are still exported, regardless of this | |
497 | option. For ELF targeted ports, symbols affected by this option will | |
498 | be treated as hidden. | |
499 | ||
e1c37eb5 DK |
500 | @kindex --exclude-modules-for-implib |
501 | @item --exclude-modules-for-implib @var{module},@var{module},... | |
502 | Specifies a list of object files or archive members, from which symbols | |
503 | should not be automatically exported, but which should be copied wholesale | |
504 | into the import library being generated during the link. The module names | |
505 | may be delimited by commas or colons, and must match exactly the filenames | |
506 | used by @command{ld} to open the files; for archive members, this is simply | |
507 | the member name, but for object files the name listed must include and | |
508 | match precisely any path used to specify the input file on the linker's | |
509 | command-line. This option is available only for the i386 PE targeted port | |
510 | of the linker. Symbols explicitly listed in a .def file are still exported, | |
511 | regardless of this option. | |
512 | ||
252b5132 RH |
513 | @cindex dynamic symbol table |
514 | @kindex -E | |
515 | @kindex --export-dynamic | |
267e2722 | 516 | @kindex --no-export-dynamic |
252b5132 RH |
517 | @item -E |
518 | @itemx --export-dynamic | |
267e2722 CD |
519 | @itemx --no-export-dynamic |
520 | When creating a dynamically linked executable, using the @option{-E} | |
521 | option or the @option{--export-dynamic} option causes the linker to add | |
522 | all symbols to the dynamic symbol table. The dynamic symbol table is the | |
523 | set of symbols which are visible from dynamic objects at run time. | |
524 | ||
525 | If you do not use either of these options (or use the | |
526 | @option{--no-export-dynamic} option to restore the default behavior), the | |
527 | dynamic symbol table will normally contain only those symbols which are | |
528 | referenced by some dynamic object mentioned in the link. | |
252b5132 RH |
529 | |
530 | If you use @code{dlopen} to load a dynamic object which needs to refer | |
531 | back to the symbols defined by the program, rather than some other | |
532 | dynamic object, then you will probably need to use this option when | |
533 | linking the program itself. | |
534 | ||
55255dae | 535 | You can also use the dynamic list to control what symbols should |
cb840a31 | 536 | be added to the dynamic symbol table if the output format supports it. |
55255dae | 537 | See the description of @samp{--dynamic-list}. |
cb840a31 | 538 | |
8b747e1a DK |
539 | Note that this option is specific to ELF targeted ports. PE targets |
540 | support a similar function to export all symbols from a DLL or EXE; see | |
541 | the description of @samp{--export-all-symbols} below. | |
542 | ||
36f63dca | 543 | @ifclear SingleFormat |
252b5132 RH |
544 | @cindex big-endian objects |
545 | @cindex endianness | |
546 | @kindex -EB | |
547 | @item -EB | |
548 | Link big-endian objects. This affects the default output format. | |
549 | ||
550 | @cindex little-endian objects | |
551 | @kindex -EL | |
552 | @item -EL | |
553 | Link little-endian objects. This affects the default output format. | |
36f63dca | 554 | @end ifclear |
252b5132 | 555 | |
2509a395 SL |
556 | @kindex -f @var{name} |
557 | @kindex --auxiliary=@var{name} | |
558 | @item -f @var{name} | |
559 | @itemx --auxiliary=@var{name} | |
252b5132 RH |
560 | When creating an ELF shared object, set the internal DT_AUXILIARY field |
561 | to the specified name. This tells the dynamic linker that the symbol | |
562 | table of the shared object should be used as an auxiliary filter on the | |
563 | symbol table of the shared object @var{name}. | |
564 | ||
565 | If you later link a program against this filter object, then, when you | |
566 | run the program, the dynamic linker will see the DT_AUXILIARY field. If | |
567 | the dynamic linker resolves any symbols from the filter object, it will | |
568 | first check whether there is a definition in the shared object | |
569 | @var{name}. If there is one, it will be used instead of the definition | |
570 | in the filter object. The shared object @var{name} need not exist. | |
571 | Thus the shared object @var{name} may be used to provide an alternative | |
572 | implementation of certain functions, perhaps for debugging or for | |
573 | machine specific performance. | |
574 | ||
575 | This option may be specified more than once. The DT_AUXILIARY entries | |
576 | will be created in the order in which they appear on the command line. | |
577 | ||
2509a395 SL |
578 | @kindex -F @var{name} |
579 | @kindex --filter=@var{name} | |
252b5132 | 580 | @item -F @var{name} |
2509a395 | 581 | @itemx --filter=@var{name} |
252b5132 RH |
582 | When creating an ELF shared object, set the internal DT_FILTER field to |
583 | the specified name. This tells the dynamic linker that the symbol table | |
584 | of the shared object which is being created should be used as a filter | |
585 | on the symbol table of the shared object @var{name}. | |
586 | ||
587 | If you later link a program against this filter object, then, when you | |
588 | run the program, the dynamic linker will see the DT_FILTER field. The | |
589 | dynamic linker will resolve symbols according to the symbol table of the | |
590 | filter object as usual, but it will actually link to the definitions | |
591 | found in the shared object @var{name}. Thus the filter object can be | |
592 | used to select a subset of the symbols provided by the object | |
593 | @var{name}. | |
594 | ||
ff5dcc92 | 595 | Some older linkers used the @option{-F} option throughout a compilation |
252b5132 | 596 | toolchain for specifying object-file format for both input and output |
36f63dca NC |
597 | object files. |
598 | @ifclear SingleFormat | |
599 | The @sc{gnu} linker uses other mechanisms for this purpose: the | |
ece2d90e | 600 | @option{-b}, @option{--format}, @option{--oformat} options, the |
252b5132 | 601 | @code{TARGET} command in linker scripts, and the @code{GNUTARGET} |
36f63dca NC |
602 | environment variable. |
603 | @end ifclear | |
604 | The @sc{gnu} linker will ignore the @option{-F} option when not | |
605 | creating an ELF shared object. | |
252b5132 | 606 | |
3dbf70a2 | 607 | @cindex finalization function |
2509a395 SL |
608 | @kindex -fini=@var{name} |
609 | @item -fini=@var{name} | |
3dbf70a2 MM |
610 | When creating an ELF executable or shared object, call NAME when the |
611 | executable or shared object is unloaded, by setting DT_FINI to the | |
612 | address of the function. By default, the linker uses @code{_fini} as | |
613 | the function to call. | |
614 | ||
252b5132 RH |
615 | @kindex -g |
616 | @item -g | |
617 | Ignored. Provided for compatibility with other tools. | |
618 | ||
2509a395 SL |
619 | @kindex -G @var{value} |
620 | @kindex --gpsize=@var{value} | |
252b5132 | 621 | @cindex object size |
2509a395 | 622 | @item -G @var{value} |
252b5132 RH |
623 | @itemx --gpsize=@var{value} |
624 | Set the maximum size of objects to be optimized using the GP register to | |
625 | @var{size}. This is only meaningful for object file formats such as | |
626 | MIPS ECOFF which supports putting large and small objects into different | |
627 | sections. This is ignored for other object file formats. | |
628 | ||
629 | @cindex runtime library name | |
2509a395 | 630 | @kindex -h @var{name} |
252b5132 | 631 | @kindex -soname=@var{name} |
2509a395 | 632 | @item -h @var{name} |
252b5132 RH |
633 | @itemx -soname=@var{name} |
634 | When creating an ELF shared object, set the internal DT_SONAME field to | |
635 | the specified name. When an executable is linked with a shared object | |
636 | which has a DT_SONAME field, then when the executable is run the dynamic | |
637 | linker will attempt to load the shared object specified by the DT_SONAME | |
638 | field rather than the using the file name given to the linker. | |
639 | ||
640 | @kindex -i | |
641 | @cindex incremental link | |
642 | @item -i | |
643 | Perform an incremental link (same as option @samp{-r}). | |
644 | ||
3dbf70a2 | 645 | @cindex initialization function |
2509a395 SL |
646 | @kindex -init=@var{name} |
647 | @item -init=@var{name} | |
3dbf70a2 MM |
648 | When creating an ELF executable or shared object, call NAME when the |
649 | executable or shared object is loaded, by setting DT_INIT to the address | |
650 | of the function. By default, the linker uses @code{_init} as the | |
651 | function to call. | |
652 | ||
252b5132 | 653 | @cindex archive files, from cmd line |
2509a395 | 654 | @kindex -l @var{namespec} |
bcb674cf | 655 | @kindex --library=@var{namespec} |
2509a395 | 656 | @item -l @var{namespec} |
bcb674cf RS |
657 | @itemx --library=@var{namespec} |
658 | Add the archive or object file specified by @var{namespec} to the | |
659 | list of files to link. This option may be used any number of times. | |
660 | If @var{namespec} is of the form @file{:@var{filename}}, @command{ld} | |
07d8eb55 | 661 | will search the library path for a file called @var{filename}, otherwise it |
bcb674cf | 662 | will search the library path for a file called @file{lib@var{namespec}.a}. |
252b5132 | 663 | |
ff5dcc92 | 664 | On systems which support shared libraries, @command{ld} may also search for |
bcb674cf RS |
665 | files other than @file{lib@var{namespec}.a}. Specifically, on ELF |
666 | and SunOS systems, @command{ld} will search a directory for a library | |
667 | called @file{lib@var{namespec}.so} before searching for one called | |
668 | @file{lib@var{namespec}.a}. (By convention, a @code{.so} extension | |
669 | indicates a shared library.) Note that this behavior does not apply | |
670 | to @file{:@var{filename}}, which always specifies a file called | |
671 | @var{filename}. | |
252b5132 RH |
672 | |
673 | The linker will search an archive only once, at the location where it is | |
674 | specified on the command line. If the archive defines a symbol which | |
675 | was undefined in some object which appeared before the archive on the | |
676 | command line, the linker will include the appropriate file(s) from the | |
677 | archive. However, an undefined symbol in an object appearing later on | |
678 | the command line will not cause the linker to search the archive again. | |
679 | ||
ff5dcc92 | 680 | See the @option{-(} option for a way to force the linker to search |
252b5132 RH |
681 | archives multiple times. |
682 | ||
683 | You may list the same archive multiple times on the command line. | |
684 | ||
685 | @ifset GENERIC | |
686 | This type of archive searching is standard for Unix linkers. However, | |
ff5dcc92 | 687 | if you are using @command{ld} on AIX, note that it is different from the |
252b5132 RH |
688 | behaviour of the AIX linker. |
689 | @end ifset | |
690 | ||
691 | @cindex search directory, from cmd line | |
2509a395 | 692 | @kindex -L @var{dir} |
252b5132 | 693 | @kindex --library-path=@var{dir} |
2509a395 | 694 | @item -L @var{searchdir} |
252b5132 | 695 | @itemx --library-path=@var{searchdir} |
ff5dcc92 SC |
696 | Add path @var{searchdir} to the list of paths that @command{ld} will search |
697 | for archive libraries and @command{ld} control scripts. You may use this | |
252b5132 RH |
698 | option any number of times. The directories are searched in the order |
699 | in which they are specified on the command line. Directories specified | |
700 | on the command line are searched before the default directories. All | |
ff5dcc92 | 701 | @option{-L} options apply to all @option{-l} options, regardless of the |
7d24f02c KH |
702 | order in which the options appear. @option{-L} options do not affect |
703 | how @command{ld} searches for a linker script unless @option{-T} | |
704 | option is specified. | |
252b5132 | 705 | |
9c8ebd6a DJ |
706 | If @var{searchdir} begins with @code{=}, then the @code{=} will be replaced |
707 | by the @dfn{sysroot prefix}, a path specified when the linker is configured. | |
708 | ||
252b5132 RH |
709 | @ifset UsesEnvVars |
710 | The default set of paths searched (without being specified with | |
ff5dcc92 | 711 | @samp{-L}) depends on which emulation mode @command{ld} is using, and in |
252b5132 RH |
712 | some cases also on how it was configured. @xref{Environment}. |
713 | @end ifset | |
714 | ||
715 | The paths can also be specified in a link script with the | |
716 | @code{SEARCH_DIR} command. Directories specified this way are searched | |
717 | at the point in which the linker script appears in the command line. | |
718 | ||
719 | @cindex emulation | |
720 | @kindex -m @var{emulation} | |
2509a395 | 721 | @item -m @var{emulation} |
252b5132 RH |
722 | Emulate the @var{emulation} linker. You can list the available |
723 | emulations with the @samp{--verbose} or @samp{-V} options. | |
724 | ||
725 | If the @samp{-m} option is not used, the emulation is taken from the | |
726 | @code{LDEMULATION} environment variable, if that is defined. | |
727 | ||
728 | Otherwise, the default emulation depends upon how the linker was | |
729 | configured. | |
730 | ||
731 | @cindex link map | |
732 | @kindex -M | |
733 | @kindex --print-map | |
734 | @item -M | |
735 | @itemx --print-map | |
736 | Print a link map to the standard output. A link map provides | |
737 | information about the link, including the following: | |
738 | ||
739 | @itemize @bullet | |
740 | @item | |
3b83e13a | 741 | Where object files are mapped into memory. |
252b5132 RH |
742 | @item |
743 | How common symbols are allocated. | |
744 | @item | |
745 | All archive members included in the link, with a mention of the symbol | |
746 | which caused the archive member to be brought in. | |
3b83e13a NC |
747 | @item |
748 | The values assigned to symbols. | |
749 | ||
750 | Note - symbols whose values are computed by an expression which | |
751 | involves a reference to a previous value of the same symbol may not | |
752 | have correct result displayed in the link map. This is because the | |
753 | linker discards intermediate results and only retains the final value | |
754 | of an expression. Under such circumstances the linker will display | |
755 | the final value enclosed by square brackets. Thus for example a | |
756 | linker script containing: | |
757 | ||
758 | @smallexample | |
759 | foo = 1 | |
760 | foo = foo * 4 | |
761 | foo = foo + 8 | |
762 | @end smallexample | |
763 | ||
764 | will produce the following output in the link map if the @option{-M} | |
765 | option is used: | |
766 | ||
767 | @smallexample | |
768 | 0x00000001 foo = 0x1 | |
769 | [0x0000000c] foo = (foo * 0x4) | |
770 | [0x0000000c] foo = (foo + 0x8) | |
771 | @end smallexample | |
772 | ||
773 | See @ref{Expressions} for more information about expressions in linker | |
774 | scripts. | |
252b5132 RH |
775 | @end itemize |
776 | ||
777 | @kindex -n | |
778 | @cindex read-only text | |
779 | @cindex NMAGIC | |
780 | @kindex --nmagic | |
781 | @item -n | |
782 | @itemx --nmagic | |
fa19fce0 | 783 | Turn off page alignment of sections, and mark the output as |
a1ab1d2a | 784 | @code{NMAGIC} if possible. |
252b5132 RH |
785 | |
786 | @kindex -N | |
787 | @kindex --omagic | |
788 | @cindex read/write from cmd line | |
789 | @cindex OMAGIC | |
a1ab1d2a | 790 | @item -N |
252b5132 RH |
791 | @itemx --omagic |
792 | Set the text and data sections to be readable and writable. Also, do | |
63fd3b82 NC |
793 | not page-align the data segment, and disable linking against shared |
794 | libraries. If the output format supports Unix style magic numbers, | |
4d8907ac DS |
795 | mark the output as @code{OMAGIC}. Note: Although a writable text section |
796 | is allowed for PE-COFF targets, it does not conform to the format | |
797 | specification published by Microsoft. | |
63fd3b82 NC |
798 | |
799 | @kindex --no-omagic | |
800 | @cindex OMAGIC | |
801 | @item --no-omagic | |
802 | This option negates most of the effects of the @option{-N} option. It | |
803 | sets the text section to be read-only, and forces the data segment to | |
804 | be page-aligned. Note - this option does not enable linking against | |
805 | shared libraries. Use @option{-Bdynamic} for this. | |
252b5132 RH |
806 | |
807 | @kindex -o @var{output} | |
808 | @kindex --output=@var{output} | |
809 | @cindex naming the output file | |
810 | @item -o @var{output} | |
811 | @itemx --output=@var{output} | |
ff5dcc92 | 812 | Use @var{output} as the name for the program produced by @command{ld}; if this |
252b5132 RH |
813 | option is not specified, the name @file{a.out} is used by default. The |
814 | script command @code{OUTPUT} can also specify the output file name. | |
815 | ||
816 | @kindex -O @var{level} | |
817 | @cindex generating optimized output | |
818 | @item -O @var{level} | |
ff5dcc92 | 819 | If @var{level} is a numeric values greater than zero @command{ld} optimizes |
252b5132 | 820 | the output. This might take significantly longer and therefore probably |
98c503ac NC |
821 | should only be enabled for the final binary. At the moment this |
822 | option only affects ELF shared library generation. Future releases of | |
823 | the linker may make more use of this option. Also currently there is | |
824 | no difference in the linker's behaviour for different non-zero values | |
825 | of this option. Again this may change with future releases. | |
252b5132 | 826 | |
a712da20 NC |
827 | @kindex -q |
828 | @kindex --emit-relocs | |
829 | @cindex retain relocations in final executable | |
830 | @item -q | |
831 | @itemx --emit-relocs | |
ba1be17e | 832 | Leave relocation sections and contents in fully linked executables. |
a712da20 NC |
833 | Post link analysis and optimization tools may need this information in |
834 | order to perform correct modifications of executables. This results | |
835 | in larger executables. | |
836 | ||
dbab7a7b NC |
837 | This option is currently only supported on ELF platforms. |
838 | ||
4f471f39 RS |
839 | @kindex --force-dynamic |
840 | @cindex forcing the creation of dynamic sections | |
841 | @item --force-dynamic | |
842 | Force the output file to have dynamic sections. This option is specific | |
843 | to VxWorks targets. | |
844 | ||
252b5132 RH |
845 | @cindex partial link |
846 | @cindex relocatable output | |
847 | @kindex -r | |
1049f94e | 848 | @kindex --relocatable |
252b5132 | 849 | @item -r |
1049f94e | 850 | @itemx --relocatable |
252b5132 | 851 | Generate relocatable output---i.e., generate an output file that can in |
ff5dcc92 | 852 | turn serve as input to @command{ld}. This is often called @dfn{partial |
252b5132 RH |
853 | linking}. As a side effect, in environments that support standard Unix |
854 | magic numbers, this option also sets the output file's magic number to | |
855 | @code{OMAGIC}. | |
ff5dcc92 | 856 | @c ; see @option{-N}. |
252b5132 RH |
857 | If this option is not specified, an absolute file is produced. When |
858 | linking C++ programs, this option @emph{will not} resolve references to | |
859 | constructors; to do that, use @samp{-Ur}. | |
860 | ||
62bf86b4 HPN |
861 | When an input file does not have the same format as the output file, |
862 | partial linking is only supported if that input file does not contain any | |
863 | relocations. Different output formats can have further restrictions; for | |
864 | example some @code{a.out}-based formats do not support partial linking | |
865 | with input files in other formats at all. | |
866 | ||
252b5132 RH |
867 | This option does the same thing as @samp{-i}. |
868 | ||
869 | @kindex -R @var{file} | |
870 | @kindex --just-symbols=@var{file} | |
871 | @cindex symbol-only input | |
872 | @item -R @var{filename} | |
873 | @itemx --just-symbols=@var{filename} | |
874 | Read symbol names and their addresses from @var{filename}, but do not | |
875 | relocate it or include it in the output. This allows your output file | |
876 | to refer symbolically to absolute locations of memory defined in other | |
877 | programs. You may use this option more than once. | |
878 | ||
ff5dcc92 | 879 | For compatibility with other ELF linkers, if the @option{-R} option is |
252b5132 | 880 | followed by a directory name, rather than a file name, it is treated as |
ff5dcc92 | 881 | the @option{-rpath} option. |
252b5132 RH |
882 | |
883 | @kindex -s | |
884 | @kindex --strip-all | |
885 | @cindex strip all symbols | |
a1ab1d2a | 886 | @item -s |
252b5132 RH |
887 | @itemx --strip-all |
888 | Omit all symbol information from the output file. | |
889 | ||
890 | @kindex -S | |
891 | @kindex --strip-debug | |
892 | @cindex strip debugger symbols | |
a1ab1d2a | 893 | @item -S |
252b5132 RH |
894 | @itemx --strip-debug |
895 | Omit debugger symbol information (but not all symbols) from the output file. | |
896 | ||
897 | @kindex -t | |
898 | @kindex --trace | |
899 | @cindex input files, displaying | |
a1ab1d2a | 900 | @item -t |
252b5132 | 901 | @itemx --trace |
ff5dcc92 | 902 | Print the names of the input files as @command{ld} processes them. |
252b5132 RH |
903 | |
904 | @kindex -T @var{script} | |
905 | @kindex --script=@var{script} | |
906 | @cindex script files | |
907 | @item -T @var{scriptfile} | |
908 | @itemx --script=@var{scriptfile} | |
909 | Use @var{scriptfile} as the linker script. This script replaces | |
ff5dcc92 | 910 | @command{ld}'s default linker script (rather than adding to it), so |
252b5132 | 911 | @var{commandfile} must specify everything necessary to describe the |
114283d8 NC |
912 | output file. @xref{Scripts}. If @var{scriptfile} does not exist in |
913 | the current directory, @code{ld} looks for it in the directories | |
914 | specified by any preceding @samp{-L} options. Multiple @samp{-T} | |
915 | options accumulate. | |
252b5132 | 916 | |
14be8564 L |
917 | @kindex -dT @var{script} |
918 | @kindex --default-script=@var{script} | |
919 | @cindex script files | |
920 | @item -dT @var{scriptfile} | |
921 | @itemx --default-script=@var{scriptfile} | |
922 | Use @var{scriptfile} as the default linker script. @xref{Scripts}. | |
923 | ||
924 | This option is similar to the @option{--script} option except that | |
925 | processing of the script is delayed until after the rest of the | |
926 | command line has been processed. This allows options placed after the | |
927 | @option{--default-script} option on the command line to affect the | |
928 | behaviour of the linker script, which can be important when the linker | |
929 | command line cannot be directly controlled by the user. (eg because | |
930 | the command line is being constructed by another tool, such as | |
931 | @samp{gcc}). | |
932 | ||
252b5132 RH |
933 | @kindex -u @var{symbol} |
934 | @kindex --undefined=@var{symbol} | |
935 | @cindex undefined symbol | |
936 | @item -u @var{symbol} | |
937 | @itemx --undefined=@var{symbol} | |
938 | Force @var{symbol} to be entered in the output file as an undefined | |
939 | symbol. Doing this may, for example, trigger linking of additional | |
940 | modules from standard libraries. @samp{-u} may be repeated with | |
941 | different option arguments to enter additional undefined symbols. This | |
942 | option is equivalent to the @code{EXTERN} linker script command. | |
943 | ||
944 | @kindex -Ur | |
945 | @cindex constructors | |
a1ab1d2a | 946 | @item -Ur |
252b5132 RH |
947 | For anything other than C++ programs, this option is equivalent to |
948 | @samp{-r}: it generates relocatable output---i.e., an output file that can in | |
ff5dcc92 | 949 | turn serve as input to @command{ld}. When linking C++ programs, @samp{-Ur} |
252b5132 RH |
950 | @emph{does} resolve references to constructors, unlike @samp{-r}. |
951 | It does not work to use @samp{-Ur} on files that were themselves linked | |
952 | with @samp{-Ur}; once the constructor table has been built, it cannot | |
953 | be added to. Use @samp{-Ur} only for the last partial link, and | |
954 | @samp{-r} for the others. | |
955 | ||
577a0623 AM |
956 | @kindex --unique[=@var{SECTION}] |
957 | @item --unique[=@var{SECTION}] | |
958 | Creates a separate output section for every input section matching | |
959 | @var{SECTION}, or if the optional wildcard @var{SECTION} argument is | |
960 | missing, for every orphan input section. An orphan section is one not | |
961 | specifically mentioned in a linker script. You may use this option | |
962 | multiple times on the command line; It prevents the normal merging of | |
963 | input sections with the same name, overriding output section assignments | |
964 | in a linker script. | |
a854a4a7 | 965 | |
252b5132 RH |
966 | @kindex -v |
967 | @kindex -V | |
968 | @kindex --version | |
969 | @cindex version | |
970 | @item -v | |
971 | @itemx --version | |
972 | @itemx -V | |
ff5dcc92 | 973 | Display the version number for @command{ld}. The @option{-V} option also |
252b5132 RH |
974 | lists the supported emulations. |
975 | ||
976 | @kindex -x | |
977 | @kindex --discard-all | |
978 | @cindex deleting local symbols | |
979 | @item -x | |
980 | @itemx --discard-all | |
981 | Delete all local symbols. | |
982 | ||
983 | @kindex -X | |
984 | @kindex --discard-locals | |
985 | @cindex local symbols, deleting | |
a1ab1d2a | 986 | @item -X |
252b5132 | 987 | @itemx --discard-locals |
3c68c38f BW |
988 | Delete all temporary local symbols. (These symbols start with |
989 | system-specific local label prefixes, typically @samp{.L} for ELF systems | |
990 | or @samp{L} for traditional a.out systems.) | |
252b5132 RH |
991 | |
992 | @kindex -y @var{symbol} | |
993 | @kindex --trace-symbol=@var{symbol} | |
994 | @cindex symbol tracing | |
995 | @item -y @var{symbol} | |
996 | @itemx --trace-symbol=@var{symbol} | |
997 | Print the name of each linked file in which @var{symbol} appears. This | |
998 | option may be given any number of times. On many systems it is necessary | |
999 | to prepend an underscore. | |
1000 | ||
1001 | This option is useful when you have an undefined symbol in your link but | |
1002 | don't know where the reference is coming from. | |
1003 | ||
1004 | @kindex -Y @var{path} | |
1005 | @item -Y @var{path} | |
1006 | Add @var{path} to the default library search path. This option exists | |
1007 | for Solaris compatibility. | |
1008 | ||
1009 | @kindex -z @var{keyword} | |
1010 | @item -z @var{keyword} | |
cd6d6c15 NC |
1011 | The recognized keywords are: |
1012 | @table @samp | |
1013 | ||
1014 | @item combreloc | |
1015 | Combines multiple reloc sections and sorts them to make dynamic symbol | |
1016 | lookup caching possible. | |
1017 | ||
1018 | @item defs | |
560e09e9 | 1019 | Disallows undefined symbols in object files. Undefined symbols in |
07f3b6ad | 1020 | shared libraries are still allowed. |
cd6d6c15 | 1021 | |
6aa29e7b JJ |
1022 | @item execstack |
1023 | Marks the object as requiring executable stack. | |
1024 | ||
cd6d6c15 NC |
1025 | @item initfirst |
1026 | This option is only meaningful when building a shared object. | |
1027 | It marks the object so that its runtime initialization will occur | |
1028 | before the runtime initialization of any other objects brought into | |
1029 | the process at the same time. Similarly the runtime finalization of | |
1030 | the object will occur after the runtime finalization of any other | |
1031 | objects. | |
1032 | ||
1033 | @item interpose | |
1034 | Marks the object that its symbol table interposes before all symbols | |
1035 | but the primary executable. | |
1036 | ||
5fa222e4 AM |
1037 | @item lazy |
1038 | When generating an executable or shared library, mark it to tell the | |
1039 | dynamic linker to defer function call resolution to the point when | |
1040 | the function is called (lazy binding), rather than at load time. | |
1041 | Lazy binding is the default. | |
1042 | ||
cd6d6c15 NC |
1043 | @item loadfltr |
1044 | Marks the object that its filters be processed immediately at | |
1045 | runtime. | |
1046 | ||
1047 | @item muldefs | |
1048 | Allows multiple definitions. | |
1049 | ||
1050 | @item nocombreloc | |
1051 | Disables multiple reloc sections combining. | |
1052 | ||
1053 | @item nocopyreloc | |
1054 | Disables production of copy relocs. | |
1055 | ||
1056 | @item nodefaultlib | |
1057 | Marks the object that the search for dependencies of this object will | |
1058 | ignore any default library search paths. | |
1059 | ||
1060 | @item nodelete | |
1061 | Marks the object shouldn't be unloaded at runtime. | |
1062 | ||
1063 | @item nodlopen | |
1064 | Marks the object not available to @code{dlopen}. | |
1065 | ||
1066 | @item nodump | |
1067 | Marks the object can not be dumped by @code{dldump}. | |
1068 | ||
6aa29e7b JJ |
1069 | @item noexecstack |
1070 | Marks the object as not requiring executable stack. | |
1071 | ||
1072 | @item norelro | |
1073 | Don't create an ELF @code{PT_GNU_RELRO} segment header in the object. | |
1074 | ||
cd6d6c15 NC |
1075 | @item now |
1076 | When generating an executable or shared library, mark it to tell the | |
1077 | dynamic linker to resolve all symbols when the program is started, or | |
1078 | when the shared library is linked to using dlopen, instead of | |
1079 | deferring function call resolution to the point when the function is | |
1080 | first called. | |
1081 | ||
1082 | @item origin | |
1083 | Marks the object may contain $ORIGIN. | |
1084 | ||
6aa29e7b JJ |
1085 | @item relro |
1086 | Create an ELF @code{PT_GNU_RELRO} segment header in the object. | |
1087 | ||
24718e3b L |
1088 | @item max-page-size=@var{value} |
1089 | Set the emulation maximum page size to @var{value}. | |
1090 | ||
1091 | @item common-page-size=@var{value} | |
1092 | Set the emulation common page size to @var{value}. | |
1093 | ||
cd6d6c15 NC |
1094 | @end table |
1095 | ||
ece2d90e | 1096 | Other keywords are ignored for Solaris compatibility. |
252b5132 RH |
1097 | |
1098 | @kindex -( | |
1099 | @cindex groups of archives | |
1100 | @item -( @var{archives} -) | |
1101 | @itemx --start-group @var{archives} --end-group | |
1102 | The @var{archives} should be a list of archive files. They may be | |
1103 | either explicit file names, or @samp{-l} options. | |
1104 | ||
1105 | The specified archives are searched repeatedly until no new undefined | |
1106 | references are created. Normally, an archive is searched only once in | |
1107 | the order that it is specified on the command line. If a symbol in that | |
1108 | archive is needed to resolve an undefined symbol referred to by an | |
1109 | object in an archive that appears later on the command line, the linker | |
1110 | would not be able to resolve that reference. By grouping the archives, | |
1111 | they all be searched repeatedly until all possible references are | |
1112 | resolved. | |
1113 | ||
1114 | Using this option has a significant performance cost. It is best to use | |
1115 | it only when there are unavoidable circular references between two or | |
1116 | more archives. | |
1117 | ||
69da35b5 NC |
1118 | @kindex --accept-unknown-input-arch |
1119 | @kindex --no-accept-unknown-input-arch | |
1120 | @item --accept-unknown-input-arch | |
1121 | @itemx --no-accept-unknown-input-arch | |
1122 | Tells the linker to accept input files whose architecture cannot be | |
2ca22b03 | 1123 | recognised. The assumption is that the user knows what they are doing |
69da35b5 NC |
1124 | and deliberately wants to link in these unknown input files. This was |
1125 | the default behaviour of the linker, before release 2.14. The default | |
1126 | behaviour from release 2.14 onwards is to reject such input files, and | |
1127 | so the @samp{--accept-unknown-input-arch} option has been added to | |
1128 | restore the old behaviour. | |
2ca22b03 | 1129 | |
4a43e768 AM |
1130 | @kindex --as-needed |
1131 | @kindex --no-as-needed | |
1132 | @item --as-needed | |
1133 | @itemx --no-as-needed | |
1134 | This option affects ELF DT_NEEDED tags for dynamic libraries mentioned | |
1135 | on the command line after the @option{--as-needed} option. Normally, | |
1136 | the linker will add a DT_NEEDED tag for each dynamic library mentioned | |
1137 | on the command line, regardless of whether the library is actually | |
010e5ae2 AM |
1138 | needed. @option{--as-needed} causes a DT_NEEDED tag to only be emitted |
1139 | for a library that satisfies a symbol reference from regular objects | |
1140 | which is undefined at the point that the library was linked, or, if | |
1141 | the library is not found in the DT_NEEDED lists of other libraries | |
1142 | linked up to that point, a reference from another dynamic library. | |
4a43e768 AM |
1143 | @option{--no-as-needed} restores the default behaviour. |
1144 | ||
e56f61be L |
1145 | @kindex --add-needed |
1146 | @kindex --no-add-needed | |
1147 | @item --add-needed | |
1148 | @itemx --no-add-needed | |
1149 | This option affects the treatment of dynamic libraries from ELF | |
1150 | DT_NEEDED tags in dynamic libraries mentioned on the command line after | |
1151 | the @option{--no-add-needed} option. Normally, the linker will add | |
1152 | a DT_NEEDED tag for each dynamic library from DT_NEEDED tags. | |
1153 | @option{--no-add-needed} causes DT_NEEDED tags will never be emitted | |
1154 | for those libraries from DT_NEEDED tags. @option{--add-needed} restores | |
1155 | the default behaviour. | |
1156 | ||
252b5132 RH |
1157 | @kindex -assert @var{keyword} |
1158 | @item -assert @var{keyword} | |
1159 | This option is ignored for SunOS compatibility. | |
1160 | ||
1161 | @kindex -Bdynamic | |
1162 | @kindex -dy | |
1163 | @kindex -call_shared | |
1164 | @item -Bdynamic | |
1165 | @itemx -dy | |
1166 | @itemx -call_shared | |
1167 | Link against dynamic libraries. This is only meaningful on platforms | |
1168 | for which shared libraries are supported. This option is normally the | |
1169 | default on such platforms. The different variants of this option are | |
1170 | for compatibility with various systems. You may use this option | |
1171 | multiple times on the command line: it affects library searching for | |
da8bce14 | 1172 | @option{-l} options which follow it. |
252b5132 | 1173 | |
a1ab1d2a UD |
1174 | @kindex -Bgroup |
1175 | @item -Bgroup | |
1176 | Set the @code{DF_1_GROUP} flag in the @code{DT_FLAGS_1} entry in the dynamic | |
1177 | section. This causes the runtime linker to handle lookups in this | |
1178 | object and its dependencies to be performed only inside the group. | |
560e09e9 NC |
1179 | @option{--unresolved-symbols=report-all} is implied. This option is |
1180 | only meaningful on ELF platforms which support shared libraries. | |
a1ab1d2a | 1181 | |
252b5132 RH |
1182 | @kindex -Bstatic |
1183 | @kindex -dn | |
1184 | @kindex -non_shared | |
1185 | @kindex -static | |
a1ab1d2a | 1186 | @item -Bstatic |
252b5132 RH |
1187 | @itemx -dn |
1188 | @itemx -non_shared | |
1189 | @itemx -static | |
1190 | Do not link against shared libraries. This is only meaningful on | |
1191 | platforms for which shared libraries are supported. The different | |
1192 | variants of this option are for compatibility with various systems. You | |
1193 | may use this option multiple times on the command line: it affects | |
560e09e9 | 1194 | library searching for @option{-l} options which follow it. This |
e9156f74 NC |
1195 | option also implies @option{--unresolved-symbols=report-all}. This |
1196 | option can be used with @option{-shared}. Doing so means that a | |
1197 | shared library is being created but that all of the library's external | |
1198 | references must be resolved by pulling in entries from static | |
ece2d90e | 1199 | libraries. |
252b5132 RH |
1200 | |
1201 | @kindex -Bsymbolic | |
1202 | @item -Bsymbolic | |
1203 | When creating a shared library, bind references to global symbols to the | |
1204 | definition within the shared library, if any. Normally, it is possible | |
1205 | for a program linked against a shared library to override the definition | |
1206 | within the shared library. This option is only meaningful on ELF | |
1207 | platforms which support shared libraries. | |
1208 | ||
40b36307 L |
1209 | @kindex -Bsymbolic-functions |
1210 | @item -Bsymbolic-functions | |
1211 | When creating a shared library, bind references to global function | |
c0065db7 | 1212 | symbols to the definition within the shared library, if any. |
40b36307 L |
1213 | This option is only meaningful on ELF platforms which support shared |
1214 | libraries. | |
1215 | ||
55255dae L |
1216 | @kindex --dynamic-list=@var{dynamic-list-file} |
1217 | @item --dynamic-list=@var{dynamic-list-file} | |
1218 | Specify the name of a dynamic list file to the linker. This is | |
1219 | typically used when creating shared libraries to specify a list of | |
1220 | global symbols whose references shouldn't be bound to the definition | |
1221 | within the shared library, or creating dynamically linked executables | |
1222 | to specify a list of symbols which should be added to the symbol table | |
1223 | in the executable. This option is only meaningful on ELF platforms | |
1224 | which support shared libraries. | |
1225 | ||
1226 | The format of the dynamic list is the same as the version node without | |
1227 | scope and node name. See @ref{VERSION} for more information. | |
1228 | ||
40b36307 L |
1229 | @kindex --dynamic-list-data |
1230 | @item --dynamic-list-data | |
1231 | Include all global data symbols to the dynamic list. | |
1232 | ||
1233 | @kindex --dynamic-list-cpp-new | |
1234 | @item --dynamic-list-cpp-new | |
1235 | Provide the builtin dynamic list for C++ operator new and delete. It | |
1236 | is mainly useful for building shared libstdc++. | |
1237 | ||
0b8a70d9 L |
1238 | @kindex --dynamic-list-cpp-typeinfo |
1239 | @item --dynamic-list-cpp-typeinfo | |
1240 | Provide the builtin dynamic list for C++ runtime type identification. | |
1241 | ||
252b5132 RH |
1242 | @kindex --check-sections |
1243 | @kindex --no-check-sections | |
1244 | @item --check-sections | |
308b1ffd | 1245 | @itemx --no-check-sections |
252b5132 | 1246 | Asks the linker @emph{not} to check section addresses after they have |
7d816a17 | 1247 | been assigned to see if there are any overlaps. Normally the linker will |
252b5132 RH |
1248 | perform this check, and if it finds any overlaps it will produce |
1249 | suitable error messages. The linker does know about, and does make | |
1250 | allowances for sections in overlays. The default behaviour can be | |
560e09e9 | 1251 | restored by using the command line switch @option{--check-sections}. |
02b0b1aa NS |
1252 | Section overlap is not usually checked for relocatable links. You can |
1253 | force checking in that case by using the @option{--check-sections} | |
1254 | option. | |
252b5132 RH |
1255 | |
1256 | @cindex cross reference table | |
1257 | @kindex --cref | |
1258 | @item --cref | |
1259 | Output a cross reference table. If a linker map file is being | |
1260 | generated, the cross reference table is printed to the map file. | |
1261 | Otherwise, it is printed on the standard output. | |
1262 | ||
1263 | The format of the table is intentionally simple, so that it may be | |
1264 | easily processed by a script if necessary. The symbols are printed out, | |
1265 | sorted by name. For each symbol, a list of file names is given. If the | |
1266 | symbol is defined, the first file listed is the location of the | |
1267 | definition. The remaining files contain references to the symbol. | |
1268 | ||
4818e05f AM |
1269 | @cindex common allocation |
1270 | @kindex --no-define-common | |
1271 | @item --no-define-common | |
1272 | This option inhibits the assignment of addresses to common symbols. | |
1273 | The script command @code{INHIBIT_COMMON_ALLOCATION} has the same effect. | |
1274 | @xref{Miscellaneous Commands}. | |
1275 | ||
1276 | The @samp{--no-define-common} option allows decoupling | |
1277 | the decision to assign addresses to Common symbols from the choice | |
1278 | of the output file type; otherwise a non-Relocatable output type | |
1279 | forces assigning addresses to Common symbols. | |
1280 | Using @samp{--no-define-common} allows Common symbols that are referenced | |
1281 | from a shared library to be assigned addresses only in the main program. | |
1282 | This eliminates the unused duplicate space in the shared library, | |
1283 | and also prevents any possible confusion over resolving to the wrong | |
1284 | duplicate when there are many dynamic modules with specialized search | |
1285 | paths for runtime symbol resolution. | |
1286 | ||
252b5132 | 1287 | @cindex symbols, from command line |
2509a395 SL |
1288 | @kindex --defsym=@var{symbol}=@var{exp} |
1289 | @item --defsym=@var{symbol}=@var{expression} | |
252b5132 RH |
1290 | Create a global symbol in the output file, containing the absolute |
1291 | address given by @var{expression}. You may use this option as many | |
1292 | times as necessary to define multiple symbols in the command line. A | |
1293 | limited form of arithmetic is supported for the @var{expression} in this | |
1294 | context: you may give a hexadecimal constant or the name of an existing | |
1295 | symbol, or use @code{+} and @code{-} to add or subtract hexadecimal | |
1296 | constants or symbols. If you need more elaborate expressions, consider | |
1297 | using the linker command language from a script (@pxref{Assignments,, | |
1298 | Assignment: Symbol Definitions}). @emph{Note:} there should be no white | |
1299 | space between @var{symbol}, the equals sign (``@key{=}''), and | |
1300 | @var{expression}. | |
1301 | ||
1302 | @cindex demangling, from command line | |
28c309a2 | 1303 | @kindex --demangle[=@var{style}] |
252b5132 | 1304 | @kindex --no-demangle |
28c309a2 | 1305 | @item --demangle[=@var{style}] |
252b5132 RH |
1306 | @itemx --no-demangle |
1307 | These options control whether to demangle symbol names in error messages | |
1308 | and other output. When the linker is told to demangle, it tries to | |
1309 | present symbol names in a readable fashion: it strips leading | |
1310 | underscores if they are used by the object file format, and converts C++ | |
a1ab1d2a UD |
1311 | mangled symbol names into user readable names. Different compilers have |
1312 | different mangling styles. The optional demangling style argument can be used | |
1313 | to choose an appropriate demangling style for your compiler. The linker will | |
28c309a2 NC |
1314 | demangle by default unless the environment variable @samp{COLLECT_NO_DEMANGLE} |
1315 | is set. These options may be used to override the default. | |
252b5132 RH |
1316 | |
1317 | @cindex dynamic linker, from command line | |
506eee22 | 1318 | @kindex -I@var{file} |
2509a395 SL |
1319 | @kindex --dynamic-linker=@var{file} |
1320 | @item -I@var{file} | |
1321 | @itemx --dynamic-linker=@var{file} | |
252b5132 RH |
1322 | Set the name of the dynamic linker. This is only meaningful when |
1323 | generating dynamically linked ELF executables. The default dynamic | |
1324 | linker is normally correct; don't use this unless you know what you are | |
1325 | doing. | |
1326 | ||
7ce691ae | 1327 | @kindex --fatal-warnings |
0fe58ccd | 1328 | @kindex --no-fatal-warnings |
7ce691ae | 1329 | @item --fatal-warnings |
0fe58ccd NC |
1330 | @itemx --no-fatal-warnings |
1331 | Treat all warnings as errors. The default behaviour can be restored | |
1332 | with the option @option{--no-fatal-warnings}. | |
7ce691ae | 1333 | |
252b5132 RH |
1334 | @kindex --force-exe-suffix |
1335 | @item --force-exe-suffix | |
1336 | Make sure that an output file has a .exe suffix. | |
1337 | ||
1338 | If a successfully built fully linked output file does not have a | |
1339 | @code{.exe} or @code{.dll} suffix, this option forces the linker to copy | |
1340 | the output file to one of the same name with a @code{.exe} suffix. This | |
1341 | option is useful when using unmodified Unix makefiles on a Microsoft | |
1342 | Windows host, since some versions of Windows won't run an image unless | |
1343 | it ends in a @code{.exe} suffix. | |
1344 | ||
1345 | @kindex --gc-sections | |
1346 | @kindex --no-gc-sections | |
1347 | @cindex garbage collection | |
c17d87de NC |
1348 | @item --gc-sections |
1349 | @itemx --no-gc-sections | |
252b5132 | 1350 | Enable garbage collection of unused input sections. It is ignored on |
ac69cbc6 | 1351 | targets that do not support this option. The default behaviour (of not |
b3549761 NC |
1352 | performing this garbage collection) can be restored by specifying |
1353 | @samp{--no-gc-sections} on the command line. | |
252b5132 | 1354 | |
d5465ba2 AM |
1355 | @samp{--gc-sections} decides which input sections are used by |
1356 | examining symbols and relocations. The section containing the entry | |
1357 | symbol and all sections containing symbols undefined on the | |
1358 | command-line will be kept, as will sections containing symbols | |
1359 | referenced by dynamic objects. Note that when building shared | |
1360 | libraries, the linker must assume that any visible symbol is | |
1361 | referenced. Once this initial set of sections has been determined, | |
1362 | the linker recursively marks as used any section referenced by their | |
1363 | relocations. See @samp{--entry} and @samp{--undefined}. | |
1364 | ||
ac69cbc6 TG |
1365 | This option can be set when doing a partial link (enabled with option |
1366 | @samp{-r}). In this case the root of symbols kept must be explicitely | |
1367 | specified either by an @samp{--entry} or @samp{--undefined} option or by | |
1368 | a @code{ENTRY} command in the linker script. | |
1369 | ||
c17d87de NC |
1370 | @kindex --print-gc-sections |
1371 | @kindex --no-print-gc-sections | |
1372 | @cindex garbage collection | |
1373 | @item --print-gc-sections | |
1374 | @itemx --no-print-gc-sections | |
1375 | List all sections removed by garbage collection. The listing is | |
1376 | printed on stderr. This option is only effective if garbage | |
1377 | collection has been enabled via the @samp{--gc-sections}) option. The | |
1378 | default behaviour (of not listing the sections that are removed) can | |
1379 | be restored by specifying @samp{--no-print-gc-sections} on the command | |
1380 | line. | |
1381 | ||
252b5132 RH |
1382 | @cindex help |
1383 | @cindex usage | |
1384 | @kindex --help | |
1385 | @item --help | |
1386 | Print a summary of the command-line options on the standard output and exit. | |
1387 | ||
ea20a7da CC |
1388 | @kindex --target-help |
1389 | @item --target-help | |
1390 | Print a summary of all target specific options on the standard output and exit. | |
1391 | ||
2509a395 SL |
1392 | @kindex -Map=@var{mapfile} |
1393 | @item -Map=@var{mapfile} | |
252b5132 | 1394 | Print a link map to the file @var{mapfile}. See the description of the |
560e09e9 | 1395 | @option{-M} option, above. |
252b5132 RH |
1396 | |
1397 | @cindex memory usage | |
1398 | @kindex --no-keep-memory | |
1399 | @item --no-keep-memory | |
ff5dcc92 SC |
1400 | @command{ld} normally optimizes for speed over memory usage by caching the |
1401 | symbol tables of input files in memory. This option tells @command{ld} to | |
252b5132 | 1402 | instead optimize for memory usage, by rereading the symbol tables as |
ff5dcc92 | 1403 | necessary. This may be required if @command{ld} runs out of memory space |
252b5132 RH |
1404 | while linking a large executable. |
1405 | ||
1406 | @kindex --no-undefined | |
a1ab1d2a | 1407 | @kindex -z defs |
252b5132 | 1408 | @item --no-undefined |
a1ab1d2a | 1409 | @itemx -z defs |
560e09e9 NC |
1410 | Report unresolved symbol references from regular object files. This |
1411 | is done even if the linker is creating a non-symbolic shared library. | |
1412 | The switch @option{--[no-]allow-shlib-undefined} controls the | |
1413 | behaviour for reporting unresolved references found in shared | |
ece2d90e | 1414 | libraries being linked in. |
252b5132 | 1415 | |
aa713662 L |
1416 | @kindex --allow-multiple-definition |
1417 | @kindex -z muldefs | |
1418 | @item --allow-multiple-definition | |
1419 | @itemx -z muldefs | |
1420 | Normally when a symbol is defined multiple times, the linker will | |
1421 | report a fatal error. These options allow multiple definitions and the | |
1422 | first definition will be used. | |
1423 | ||
b79e8c78 | 1424 | @kindex --allow-shlib-undefined |
ae9a127f | 1425 | @kindex --no-allow-shlib-undefined |
b79e8c78 | 1426 | @item --allow-shlib-undefined |
ae9a127f | 1427 | @itemx --no-allow-shlib-undefined |
903249d7 | 1428 | Allows or disallows undefined symbols in shared libraries. |
560e09e9 NC |
1429 | This switch is similar to @option{--no-undefined} except that it |
1430 | determines the behaviour when the undefined symbols are in a | |
1431 | shared library rather than a regular object file. It does not affect | |
1432 | how undefined symbols in regular object files are handled. | |
1433 | ||
903249d7 NC |
1434 | The default behaviour is to report errors for any undefined symbols |
1435 | referenced in shared libraries if the linker is being used to create | |
1436 | an executable, but to allow them if the linker is being used to create | |
1437 | a shared library. | |
1438 | ||
1439 | The reasons for allowing undefined symbol references in shared | |
1440 | libraries specified at link time are that: | |
1441 | ||
1442 | @itemize @bullet | |
1443 | @item | |
1444 | A shared library specified at link time may not be the same as the one | |
1445 | that is available at load time, so the symbol might actually be | |
1446 | resolvable at load time. | |
1447 | @item | |
1448 | There are some operating systems, eg BeOS and HPPA, where undefined | |
1449 | symbols in shared libraries are normal. | |
1450 | ||
1451 | The BeOS kernel for example patches shared libraries at load time to | |
1452 | select whichever function is most appropriate for the current | |
1453 | architecture. This is used, for example, to dynamically select an | |
1454 | appropriate memset function. | |
1455 | @end itemize | |
b79e8c78 | 1456 | |
31941635 L |
1457 | @kindex --no-undefined-version |
1458 | @item --no-undefined-version | |
1459 | Normally when a symbol has an undefined version, the linker will ignore | |
1460 | it. This option disallows symbols with undefined version and a fatal error | |
1461 | will be issued instead. | |
1462 | ||
3e3b46e5 PB |
1463 | @kindex --default-symver |
1464 | @item --default-symver | |
1465 | Create and use a default symbol version (the soname) for unversioned | |
fc0e6df6 PB |
1466 | exported symbols. |
1467 | ||
1468 | @kindex --default-imported-symver | |
1469 | @item --default-imported-symver | |
1470 | Create and use a default symbol version (the soname) for unversioned | |
1471 | imported symbols. | |
3e3b46e5 | 1472 | |
252b5132 RH |
1473 | @kindex --no-warn-mismatch |
1474 | @item --no-warn-mismatch | |
ff5dcc92 | 1475 | Normally @command{ld} will give an error if you try to link together input |
252b5132 RH |
1476 | files that are mismatched for some reason, perhaps because they have |
1477 | been compiled for different processors or for different endiannesses. | |
ff5dcc92 | 1478 | This option tells @command{ld} that it should silently permit such possible |
252b5132 RH |
1479 | errors. This option should only be used with care, in cases when you |
1480 | have taken some special action that ensures that the linker errors are | |
1481 | inappropriate. | |
1482 | ||
fe7929ce AM |
1483 | @kindex --no-warn-search-mismatch |
1484 | @item --no-warn-search-mismatch | |
1485 | Normally @command{ld} will give a warning if it finds an incompatible | |
1486 | library during a library search. This option silences the warning. | |
1487 | ||
252b5132 RH |
1488 | @kindex --no-whole-archive |
1489 | @item --no-whole-archive | |
ff5dcc92 | 1490 | Turn off the effect of the @option{--whole-archive} option for subsequent |
252b5132 RH |
1491 | archive files. |
1492 | ||
1493 | @cindex output file after errors | |
1494 | @kindex --noinhibit-exec | |
1495 | @item --noinhibit-exec | |
1496 | Retain the executable output file whenever it is still usable. | |
1497 | Normally, the linker will not produce an output file if it encounters | |
1498 | errors during the link process; it exits without writing an output file | |
1499 | when it issues any error whatsoever. | |
1500 | ||
0a9c1c8e CD |
1501 | @kindex -nostdlib |
1502 | @item -nostdlib | |
1503 | Only search library directories explicitly specified on the | |
1504 | command line. Library directories specified in linker scripts | |
1505 | (including linker scripts specified on the command line) are ignored. | |
1506 | ||
252b5132 | 1507 | @ifclear SingleFormat |
2509a395 SL |
1508 | @kindex --oformat=@var{output-format} |
1509 | @item --oformat=@var{output-format} | |
ff5dcc92 SC |
1510 | @command{ld} may be configured to support more than one kind of object |
1511 | file. If your @command{ld} is configured this way, you can use the | |
252b5132 | 1512 | @samp{--oformat} option to specify the binary format for the output |
ff5dcc92 SC |
1513 | object file. Even when @command{ld} is configured to support alternative |
1514 | object formats, you don't usually need to specify this, as @command{ld} | |
252b5132 RH |
1515 | should be configured to produce as a default output format the most |
1516 | usual format on each machine. @var{output-format} is a text string, the | |
1517 | name of a particular format supported by the BFD libraries. (You can | |
1518 | list the available binary formats with @samp{objdump -i}.) The script | |
1519 | command @code{OUTPUT_FORMAT} can also specify the output format, but | |
1520 | this option overrides it. @xref{BFD}. | |
1521 | @end ifclear | |
1522 | ||
36af4a4e JJ |
1523 | @kindex -pie |
1524 | @kindex --pic-executable | |
1525 | @item -pie | |
1526 | @itemx --pic-executable | |
1527 | @cindex position independent executables | |
1528 | Create a position independent executable. This is currently only supported on | |
1529 | ELF platforms. Position independent executables are similar to shared | |
1530 | libraries in that they are relocated by the dynamic linker to the virtual | |
7e7d5768 | 1531 | address the OS chooses for them (which can vary between invocations). Like |
36af4a4e JJ |
1532 | normal dynamically linked executables they can be executed and symbols |
1533 | defined in the executable cannot be overridden by shared libraries. | |
1534 | ||
252b5132 RH |
1535 | @kindex -qmagic |
1536 | @item -qmagic | |
1537 | This option is ignored for Linux compatibility. | |
1538 | ||
1539 | @kindex -Qy | |
1540 | @item -Qy | |
1541 | This option is ignored for SVR4 compatibility. | |
1542 | ||
1543 | @kindex --relax | |
1544 | @cindex synthesizing linker | |
1545 | @cindex relaxing addressing modes | |
1546 | @item --relax | |
a1ab1d2a | 1547 | An option with machine dependent effects. |
252b5132 RH |
1548 | @ifset GENERIC |
1549 | This option is only supported on a few targets. | |
1550 | @end ifset | |
1551 | @ifset H8300 | |
ff5dcc92 | 1552 | @xref{H8/300,,@command{ld} and the H8/300}. |
252b5132 RH |
1553 | @end ifset |
1554 | @ifset I960 | |
ff5dcc92 | 1555 | @xref{i960,, @command{ld} and the Intel 960 family}. |
252b5132 | 1556 | @end ifset |
e0001a05 NC |
1557 | @ifset XTENSA |
1558 | @xref{Xtensa,, @command{ld} and Xtensa Processors}. | |
1559 | @end ifset | |
93fd0973 SC |
1560 | @ifset M68HC11 |
1561 | @xref{M68HC11/68HC12,,@command{ld} and the 68HC11 and 68HC12}. | |
1562 | @end ifset | |
2a60a7a8 AM |
1563 | @ifset POWERPC |
1564 | @xref{PowerPC ELF32,,@command{ld} and PowerPC 32-bit ELF Support}. | |
1565 | @end ifset | |
252b5132 RH |
1566 | |
1567 | On some platforms, the @samp{--relax} option performs global | |
1568 | optimizations that become possible when the linker resolves addressing | |
1569 | in the program, such as relaxing address modes and synthesizing new | |
1570 | instructions in the output object file. | |
1571 | ||
1572 | On some platforms these link time global optimizations may make symbolic | |
1573 | debugging of the resulting executable impossible. | |
1574 | @ifset GENERIC | |
1575 | This is known to be | |
1576 | the case for the Matsushita MN10200 and MN10300 family of processors. | |
1577 | @end ifset | |
1578 | ||
1579 | @ifset GENERIC | |
1580 | On platforms where this is not supported, @samp{--relax} is accepted, | |
1581 | but ignored. | |
1582 | @end ifset | |
1583 | ||
1584 | @cindex retaining specified symbols | |
1585 | @cindex stripping all but some symbols | |
1586 | @cindex symbols, retaining selectively | |
2509a395 SL |
1587 | @kindex --retain-symbols-file=@var{filename} |
1588 | @item --retain-symbols-file=@var{filename} | |
252b5132 RH |
1589 | Retain @emph{only} the symbols listed in the file @var{filename}, |
1590 | discarding all others. @var{filename} is simply a flat file, with one | |
1591 | symbol name per line. This option is especially useful in environments | |
1592 | @ifset GENERIC | |
1593 | (such as VxWorks) | |
1594 | @end ifset | |
1595 | where a large global symbol table is accumulated gradually, to conserve | |
1596 | run-time memory. | |
1597 | ||
1598 | @samp{--retain-symbols-file} does @emph{not} discard undefined symbols, | |
1599 | or symbols needed for relocations. | |
1600 | ||
1601 | You may only specify @samp{--retain-symbols-file} once in the command | |
1602 | line. It overrides @samp{-s} and @samp{-S}. | |
1603 | ||
1604 | @ifset GENERIC | |
2509a395 | 1605 | @item -rpath=@var{dir} |
252b5132 | 1606 | @cindex runtime library search path |
2509a395 | 1607 | @kindex -rpath=@var{dir} |
252b5132 | 1608 | Add a directory to the runtime library search path. This is used when |
ff5dcc92 | 1609 | linking an ELF executable with shared objects. All @option{-rpath} |
252b5132 | 1610 | arguments are concatenated and passed to the runtime linker, which uses |
ff5dcc92 | 1611 | them to locate shared objects at runtime. The @option{-rpath} option is |
252b5132 RH |
1612 | also used when locating shared objects which are needed by shared |
1613 | objects explicitly included in the link; see the description of the | |
ff5dcc92 | 1614 | @option{-rpath-link} option. If @option{-rpath} is not used when linking an |
252b5132 RH |
1615 | ELF executable, the contents of the environment variable |
1616 | @code{LD_RUN_PATH} will be used if it is defined. | |
1617 | ||
ff5dcc92 | 1618 | The @option{-rpath} option may also be used on SunOS. By default, on |
252b5132 | 1619 | SunOS, the linker will form a runtime search patch out of all the |
ff5dcc92 SC |
1620 | @option{-L} options it is given. If a @option{-rpath} option is used, the |
1621 | runtime search path will be formed exclusively using the @option{-rpath} | |
1622 | options, ignoring the @option{-L} options. This can be useful when using | |
1623 | gcc, which adds many @option{-L} options which may be on NFS mounted | |
b45619c0 | 1624 | file systems. |
252b5132 | 1625 | |
ff5dcc92 | 1626 | For compatibility with other ELF linkers, if the @option{-R} option is |
252b5132 | 1627 | followed by a directory name, rather than a file name, it is treated as |
ff5dcc92 | 1628 | the @option{-rpath} option. |
252b5132 RH |
1629 | @end ifset |
1630 | ||
1631 | @ifset GENERIC | |
1632 | @cindex link-time runtime library search path | |
2509a395 SL |
1633 | @kindex -rpath-link=@var{dir} |
1634 | @item -rpath-link=@var{dir} | |
252b5132 RH |
1635 | When using ELF or SunOS, one shared library may require another. This |
1636 | happens when an @code{ld -shared} link includes a shared library as one | |
1637 | of the input files. | |
1638 | ||
1639 | When the linker encounters such a dependency when doing a non-shared, | |
1640 | non-relocatable link, it will automatically try to locate the required | |
1641 | shared library and include it in the link, if it is not included | |
ff5dcc92 | 1642 | explicitly. In such a case, the @option{-rpath-link} option |
252b5132 | 1643 | specifies the first set of directories to search. The |
ff5dcc92 | 1644 | @option{-rpath-link} option may specify a sequence of directory names |
252b5132 RH |
1645 | either by specifying a list of names separated by colons, or by |
1646 | appearing multiple times. | |
1647 | ||
28c309a2 NC |
1648 | This option should be used with caution as it overrides the search path |
1649 | that may have been hard compiled into a shared library. In such a case it | |
1650 | is possible to use unintentionally a different search path than the | |
1651 | runtime linker would do. | |
1652 | ||
252b5132 | 1653 | The linker uses the following search paths to locate required shared |
ece2d90e | 1654 | libraries: |
252b5132 RH |
1655 | @enumerate |
1656 | @item | |
ff5dcc92 | 1657 | Any directories specified by @option{-rpath-link} options. |
252b5132 | 1658 | @item |
ff5dcc92 SC |
1659 | Any directories specified by @option{-rpath} options. The difference |
1660 | between @option{-rpath} and @option{-rpath-link} is that directories | |
1661 | specified by @option{-rpath} options are included in the executable and | |
1662 | used at runtime, whereas the @option{-rpath-link} option is only effective | |
ece2d90e NC |
1663 | at link time. Searching @option{-rpath} in this way is only supported |
1664 | by native linkers and cross linkers which have been configured with | |
1665 | the @option{--with-sysroot} option. | |
252b5132 | 1666 | @item |
e2a83dd0 NC |
1667 | On an ELF system, for native linkers, if the @option{-rpath} and |
1668 | @option{-rpath-link} options were not used, search the contents of the | |
1669 | environment variable @code{LD_RUN_PATH}. | |
252b5132 | 1670 | @item |
ff5dcc92 SC |
1671 | On SunOS, if the @option{-rpath} option was not used, search any |
1672 | directories specified using @option{-L} options. | |
252b5132 | 1673 | @item |
e2a83dd0 NC |
1674 | For a native linker, the search the contents of the environment |
1675 | variable @code{LD_LIBRARY_PATH}. | |
252b5132 | 1676 | @item |
ec4eb78a L |
1677 | For a native ELF linker, the directories in @code{DT_RUNPATH} or |
1678 | @code{DT_RPATH} of a shared library are searched for shared | |
1679 | libraries needed by it. The @code{DT_RPATH} entries are ignored if | |
1680 | @code{DT_RUNPATH} entries exist. | |
1681 | @item | |
252b5132 RH |
1682 | The default directories, normally @file{/lib} and @file{/usr/lib}. |
1683 | @item | |
1684 | For a native linker on an ELF system, if the file @file{/etc/ld.so.conf} | |
1685 | exists, the list of directories found in that file. | |
1686 | @end enumerate | |
1687 | ||
1688 | If the required shared library is not found, the linker will issue a | |
1689 | warning and continue with the link. | |
1690 | @end ifset | |
1691 | ||
1692 | @kindex -shared | |
1693 | @kindex -Bshareable | |
1694 | @item -shared | |
1695 | @itemx -Bshareable | |
1696 | @cindex shared libraries | |
1697 | Create a shared library. This is currently only supported on ELF, XCOFF | |
1698 | and SunOS platforms. On SunOS, the linker will automatically create a | |
ff5dcc92 | 1699 | shared library if the @option{-e} option is not used and there are |
252b5132 RH |
1700 | undefined symbols in the link. |
1701 | ||
252b5132 | 1702 | @kindex --sort-common |
2509a395 SL |
1703 | @item --sort-common |
1704 | @itemx --sort-common=ascending | |
1705 | @itemx --sort-common=descending | |
de7dd2bd NC |
1706 | This option tells @command{ld} to sort the common symbols by alignment in |
1707 | ascending or descending order when it places them in the appropriate output | |
1708 | sections. The symbol alignments considered are sixteen-byte or larger, | |
1709 | eight-byte, four-byte, two-byte, and one-byte. This is to prevent gaps | |
1710 | between symbols due to alignment constraints. If no sorting order is | |
1711 | specified, then descending order is assumed. | |
252b5132 | 1712 | |
2509a395 SL |
1713 | @kindex --sort-section=name |
1714 | @item --sort-section=name | |
bcaa7b3e L |
1715 | This option will apply @code{SORT_BY_NAME} to all wildcard section |
1716 | patterns in the linker script. | |
1717 | ||
2509a395 SL |
1718 | @kindex --sort-section=alignment |
1719 | @item --sort-section=alignment | |
bcaa7b3e L |
1720 | This option will apply @code{SORT_BY_ALIGNMENT} to all wildcard section |
1721 | patterns in the linker script. | |
1722 | ||
252b5132 | 1723 | @kindex --split-by-file |
2509a395 | 1724 | @item --split-by-file[=@var{size}] |
ff5dcc92 | 1725 | Similar to @option{--split-by-reloc} but creates a new output section for |
a854a4a7 AM |
1726 | each input file when @var{size} is reached. @var{size} defaults to a |
1727 | size of 1 if not given. | |
252b5132 RH |
1728 | |
1729 | @kindex --split-by-reloc | |
2509a395 | 1730 | @item --split-by-reloc[=@var{count}] |
a854a4a7 | 1731 | Tries to creates extra sections in the output file so that no single |
252b5132 | 1732 | output section in the file contains more than @var{count} relocations. |
a854a4a7 | 1733 | This is useful when generating huge relocatable files for downloading into |
252b5132 RH |
1734 | certain real time kernels with the COFF object file format; since COFF |
1735 | cannot represent more than 65535 relocations in a single section. Note | |
1736 | that this will fail to work with object file formats which do not | |
1737 | support arbitrary sections. The linker will not split up individual | |
1738 | input sections for redistribution, so if a single input section contains | |
1739 | more than @var{count} relocations one output section will contain that | |
a854a4a7 | 1740 | many relocations. @var{count} defaults to a value of 32768. |
252b5132 RH |
1741 | |
1742 | @kindex --stats | |
1743 | @item --stats | |
1744 | Compute and display statistics about the operation of the linker, such | |
1745 | as execution time and memory usage. | |
1746 | ||
2509a395 | 1747 | @kindex --sysroot=@var{directory} |
e2243057 RS |
1748 | @item --sysroot=@var{directory} |
1749 | Use @var{directory} as the location of the sysroot, overriding the | |
1750 | configure-time default. This option is only supported by linkers | |
1751 | that were configured using @option{--with-sysroot}. | |
1752 | ||
252b5132 RH |
1753 | @kindex --traditional-format |
1754 | @cindex traditional format | |
1755 | @item --traditional-format | |
ff5dcc92 SC |
1756 | For some targets, the output of @command{ld} is different in some ways from |
1757 | the output of some existing linker. This switch requests @command{ld} to | |
252b5132 RH |
1758 | use the traditional format instead. |
1759 | ||
1760 | @cindex dbx | |
ff5dcc92 | 1761 | For example, on SunOS, @command{ld} combines duplicate entries in the |
252b5132 RH |
1762 | symbol string table. This can reduce the size of an output file with |
1763 | full debugging information by over 30 percent. Unfortunately, the SunOS | |
1764 | @code{dbx} program can not read the resulting program (@code{gdb} has no | |
ff5dcc92 | 1765 | trouble). The @samp{--traditional-format} switch tells @command{ld} to not |
252b5132 RH |
1766 | combine duplicate entries. |
1767 | ||
2509a395 SL |
1768 | @kindex --section-start=@var{sectionname}=@var{org} |
1769 | @item --section-start=@var{sectionname}=@var{org} | |
176355da NC |
1770 | Locate a section in the output file at the absolute |
1771 | address given by @var{org}. You may use this option as many | |
1772 | times as necessary to locate multiple sections in the command | |
1773 | line. | |
1774 | @var{org} must be a single hexadecimal integer; | |
1775 | for compatibility with other linkers, you may omit the leading | |
1776 | @samp{0x} usually associated with hexadecimal values. @emph{Note:} there | |
1777 | should be no white space between @var{sectionname}, the equals | |
1778 | sign (``@key{=}''), and @var{org}. | |
1779 | ||
2509a395 SL |
1780 | @kindex -Tbss=@var{org} |
1781 | @kindex -Tdata=@var{org} | |
1782 | @kindex -Ttext=@var{org} | |
252b5132 | 1783 | @cindex segment origins, cmd line |
2509a395 SL |
1784 | @item -Tbss=@var{org} |
1785 | @itemx -Tdata=@var{org} | |
1786 | @itemx -Ttext=@var{org} | |
1787 | Same as @option{--section-start}, with @code{.bss}, @code{.data} or | |
a6e02871 | 1788 | @code{.text} as the @var{sectionname}. |
252b5132 | 1789 | |
2509a395 SL |
1790 | @kindex -Ttext-segment=@var{org} |
1791 | @item -Ttext-segment=@var{org} | |
258795f5 L |
1792 | @cindex text segment origin, cmd line |
1793 | When creating an ELF executable or shared object, it will set the address | |
1794 | of the first byte of the text segment. | |
1795 | ||
560e09e9 NC |
1796 | @kindex --unresolved-symbols |
1797 | @item --unresolved-symbols=@var{method} | |
1798 | Determine how to handle unresolved symbols. There are four possible | |
1799 | values for @samp{method}: | |
1800 | ||
1801 | @table @samp | |
1802 | @item ignore-all | |
da8bce14 | 1803 | Do not report any unresolved symbols. |
560e09e9 NC |
1804 | |
1805 | @item report-all | |
da8bce14 | 1806 | Report all unresolved symbols. This is the default. |
560e09e9 NC |
1807 | |
1808 | @item ignore-in-object-files | |
1809 | Report unresolved symbols that are contained in shared libraries, but | |
1810 | ignore them if they come from regular object files. | |
1811 | ||
1812 | @item ignore-in-shared-libs | |
1813 | Report unresolved symbols that come from regular object files, but | |
1814 | ignore them if they come from shared libraries. This can be useful | |
1815 | when creating a dynamic binary and it is known that all the shared | |
1816 | libraries that it should be referencing are included on the linker's | |
1817 | command line. | |
1818 | @end table | |
1819 | ||
1820 | The behaviour for shared libraries on their own can also be controlled | |
1821 | by the @option{--[no-]allow-shlib-undefined} option. | |
1822 | ||
1823 | Normally the linker will generate an error message for each reported | |
1824 | unresolved symbol but the option @option{--warn-unresolved-symbols} | |
1825 | can change this to a warning. | |
1826 | ||
252b5132 RH |
1827 | @kindex --verbose |
1828 | @cindex verbose | |
1829 | @item --dll-verbose | |
308b1ffd | 1830 | @itemx --verbose |
ff5dcc92 | 1831 | Display the version number for @command{ld} and list the linker emulations |
252b5132 | 1832 | supported. Display which input files can and cannot be opened. Display |
b9a8de1e | 1833 | the linker script being used by the linker. |
252b5132 RH |
1834 | |
1835 | @kindex --version-script=@var{version-scriptfile} | |
1836 | @cindex version script, symbol versions | |
2509a395 | 1837 | @item --version-script=@var{version-scriptfile} |
252b5132 RH |
1838 | Specify the name of a version script to the linker. This is typically |
1839 | used when creating shared libraries to specify additional information | |
36f63dca | 1840 | about the version hierarchy for the library being created. This option |
09e2aba4 DK |
1841 | is only fully supported on ELF platforms which support shared libraries; |
1842 | see @ref{VERSION}. It is partially supported on PE platforms, which can | |
1843 | use version scripts to filter symbol visibility in auto-export mode: any | |
1844 | symbols marked @samp{local} in the version script will not be exported. | |
1845 | @xref{WIN32}. | |
252b5132 | 1846 | |
7ce691ae | 1847 | @kindex --warn-common |
252b5132 RH |
1848 | @cindex warnings, on combining symbols |
1849 | @cindex combining symbols, warnings on | |
1850 | @item --warn-common | |
1851 | Warn when a common symbol is combined with another common symbol or with | |
560e09e9 | 1852 | a symbol definition. Unix linkers allow this somewhat sloppy practise, |
252b5132 RH |
1853 | but linkers on some other operating systems do not. This option allows |
1854 | you to find potential problems from combining global symbols. | |
560e09e9 | 1855 | Unfortunately, some C libraries use this practise, so you may get some |
252b5132 RH |
1856 | warnings about symbols in the libraries as well as in your programs. |
1857 | ||
1858 | There are three kinds of global symbols, illustrated here by C examples: | |
1859 | ||
1860 | @table @samp | |
1861 | @item int i = 1; | |
1862 | A definition, which goes in the initialized data section of the output | |
1863 | file. | |
1864 | ||
1865 | @item extern int i; | |
1866 | An undefined reference, which does not allocate space. | |
1867 | There must be either a definition or a common symbol for the | |
1868 | variable somewhere. | |
1869 | ||
1870 | @item int i; | |
1871 | A common symbol. If there are only (one or more) common symbols for a | |
1872 | variable, it goes in the uninitialized data area of the output file. | |
1873 | The linker merges multiple common symbols for the same variable into a | |
1874 | single symbol. If they are of different sizes, it picks the largest | |
1875 | size. The linker turns a common symbol into a declaration, if there is | |
1876 | a definition of the same variable. | |
1877 | @end table | |
1878 | ||
1879 | The @samp{--warn-common} option can produce five kinds of warnings. | |
1880 | Each warning consists of a pair of lines: the first describes the symbol | |
1881 | just encountered, and the second describes the previous symbol | |
1882 | encountered with the same name. One or both of the two symbols will be | |
1883 | a common symbol. | |
1884 | ||
1885 | @enumerate | |
1886 | @item | |
1887 | Turning a common symbol into a reference, because there is already a | |
1888 | definition for the symbol. | |
1889 | @smallexample | |
1890 | @var{file}(@var{section}): warning: common of `@var{symbol}' | |
1891 | overridden by definition | |
1892 | @var{file}(@var{section}): warning: defined here | |
1893 | @end smallexample | |
1894 | ||
1895 | @item | |
1896 | Turning a common symbol into a reference, because a later definition for | |
1897 | the symbol is encountered. This is the same as the previous case, | |
1898 | except that the symbols are encountered in a different order. | |
1899 | @smallexample | |
1900 | @var{file}(@var{section}): warning: definition of `@var{symbol}' | |
1901 | overriding common | |
1902 | @var{file}(@var{section}): warning: common is here | |
1903 | @end smallexample | |
1904 | ||
1905 | @item | |
1906 | Merging a common symbol with a previous same-sized common symbol. | |
1907 | @smallexample | |
1908 | @var{file}(@var{section}): warning: multiple common | |
1909 | of `@var{symbol}' | |
1910 | @var{file}(@var{section}): warning: previous common is here | |
1911 | @end smallexample | |
1912 | ||
1913 | @item | |
1914 | Merging a common symbol with a previous larger common symbol. | |
1915 | @smallexample | |
1916 | @var{file}(@var{section}): warning: common of `@var{symbol}' | |
1917 | overridden by larger common | |
1918 | @var{file}(@var{section}): warning: larger common is here | |
1919 | @end smallexample | |
1920 | ||
1921 | @item | |
1922 | Merging a common symbol with a previous smaller common symbol. This is | |
1923 | the same as the previous case, except that the symbols are | |
1924 | encountered in a different order. | |
1925 | @smallexample | |
1926 | @var{file}(@var{section}): warning: common of `@var{symbol}' | |
1927 | overriding smaller common | |
1928 | @var{file}(@var{section}): warning: smaller common is here | |
1929 | @end smallexample | |
1930 | @end enumerate | |
1931 | ||
1932 | @kindex --warn-constructors | |
1933 | @item --warn-constructors | |
1934 | Warn if any global constructors are used. This is only useful for a few | |
1935 | object file formats. For formats like COFF or ELF, the linker can not | |
1936 | detect the use of global constructors. | |
1937 | ||
1938 | @kindex --warn-multiple-gp | |
1939 | @item --warn-multiple-gp | |
1940 | Warn if multiple global pointer values are required in the output file. | |
1941 | This is only meaningful for certain processors, such as the Alpha. | |
1942 | Specifically, some processors put large-valued constants in a special | |
1943 | section. A special register (the global pointer) points into the middle | |
1944 | of this section, so that constants can be loaded efficiently via a | |
1945 | base-register relative addressing mode. Since the offset in | |
1946 | base-register relative mode is fixed and relatively small (e.g., 16 | |
1947 | bits), this limits the maximum size of the constant pool. Thus, in | |
1948 | large programs, it is often necessary to use multiple global pointer | |
1949 | values in order to be able to address all possible constants. This | |
1950 | option causes a warning to be issued whenever this case occurs. | |
1951 | ||
1952 | @kindex --warn-once | |
1953 | @cindex warnings, on undefined symbols | |
1954 | @cindex undefined symbols, warnings on | |
1955 | @item --warn-once | |
1956 | Only warn once for each undefined symbol, rather than once per module | |
1957 | which refers to it. | |
1958 | ||
1959 | @kindex --warn-section-align | |
1960 | @cindex warnings, on section alignment | |
1961 | @cindex section alignment, warnings on | |
1962 | @item --warn-section-align | |
1963 | Warn if the address of an output section is changed because of | |
1964 | alignment. Typically, the alignment will be set by an input section. | |
1965 | The address will only be changed if it not explicitly specified; that | |
1966 | is, if the @code{SECTIONS} command does not specify a start address for | |
1967 | the section (@pxref{SECTIONS}). | |
1968 | ||
8fdd7217 NC |
1969 | @kindex --warn-shared-textrel |
1970 | @item --warn-shared-textrel | |
ece2d90e | 1971 | Warn if the linker adds a DT_TEXTREL to a shared object. |
8fdd7217 | 1972 | |
a0c402a5 L |
1973 | @kindex --warn-alternate-em |
1974 | @item --warn-alternate-em | |
1975 | Warn if an object has alternate ELF machine code. | |
1976 | ||
560e09e9 NC |
1977 | @kindex --warn-unresolved-symbols |
1978 | @item --warn-unresolved-symbols | |
1979 | If the linker is going to report an unresolved symbol (see the option | |
1980 | @option{--unresolved-symbols}) it will normally generate an error. | |
1981 | This option makes it generate a warning instead. | |
1982 | ||
1983 | @kindex --error-unresolved-symbols | |
1984 | @item --error-unresolved-symbols | |
1985 | This restores the linker's default behaviour of generating errors when | |
1986 | it is reporting unresolved symbols. | |
1987 | ||
252b5132 RH |
1988 | @kindex --whole-archive |
1989 | @cindex including an entire archive | |
1990 | @item --whole-archive | |
1991 | For each archive mentioned on the command line after the | |
ff5dcc92 | 1992 | @option{--whole-archive} option, include every object file in the archive |
252b5132 RH |
1993 | in the link, rather than searching the archive for the required object |
1994 | files. This is normally used to turn an archive file into a shared | |
1995 | library, forcing every object to be included in the resulting shared | |
1996 | library. This option may be used more than once. | |
1997 | ||
7ec229ce | 1998 | Two notes when using this option from gcc: First, gcc doesn't know |
ff5dcc92 SC |
1999 | about this option, so you have to use @option{-Wl,-whole-archive}. |
2000 | Second, don't forget to use @option{-Wl,-no-whole-archive} after your | |
7ec229ce DD |
2001 | list of archives, because gcc will add its own list of archives to |
2002 | your link and you may not want this flag to affect those as well. | |
2003 | ||
2509a395 SL |
2004 | @kindex --wrap=@var{symbol} |
2005 | @item --wrap=@var{symbol} | |
252b5132 RH |
2006 | Use a wrapper function for @var{symbol}. Any undefined reference to |
2007 | @var{symbol} will be resolved to @code{__wrap_@var{symbol}}. Any | |
2008 | undefined reference to @code{__real_@var{symbol}} will be resolved to | |
2009 | @var{symbol}. | |
2010 | ||
2011 | This can be used to provide a wrapper for a system function. The | |
2012 | wrapper function should be called @code{__wrap_@var{symbol}}. If it | |
2013 | wishes to call the system function, it should call | |
2014 | @code{__real_@var{symbol}}. | |
2015 | ||
2016 | Here is a trivial example: | |
2017 | ||
2018 | @smallexample | |
2019 | void * | |
cc2f008e | 2020 | __wrap_malloc (size_t c) |
252b5132 | 2021 | @{ |
cc2f008e | 2022 | printf ("malloc called with %zu\n", c); |
252b5132 RH |
2023 | return __real_malloc (c); |
2024 | @} | |
2025 | @end smallexample | |
2026 | ||
ff5dcc92 | 2027 | If you link other code with this file using @option{--wrap malloc}, then |
252b5132 RH |
2028 | all calls to @code{malloc} will call the function @code{__wrap_malloc} |
2029 | instead. The call to @code{__real_malloc} in @code{__wrap_malloc} will | |
2030 | call the real @code{malloc} function. | |
2031 | ||
2032 | You may wish to provide a @code{__real_malloc} function as well, so that | |
ff5dcc92 | 2033 | links without the @option{--wrap} option will succeed. If you do this, |
252b5132 RH |
2034 | you should not put the definition of @code{__real_malloc} in the same |
2035 | file as @code{__wrap_malloc}; if you do, the assembler may resolve the | |
2036 | call before the linker has a chance to wrap it to @code{malloc}. | |
2037 | ||
6aa29e7b JJ |
2038 | @kindex --eh-frame-hdr |
2039 | @item --eh-frame-hdr | |
2040 | Request creation of @code{.eh_frame_hdr} section and ELF | |
2041 | @code{PT_GNU_EH_FRAME} segment header. | |
2042 | ||
6c1439be L |
2043 | @kindex --enable-new-dtags |
2044 | @kindex --disable-new-dtags | |
2045 | @item --enable-new-dtags | |
2046 | @itemx --disable-new-dtags | |
2047 | This linker can create the new dynamic tags in ELF. But the older ELF | |
2048 | systems may not understand them. If you specify | |
ff5dcc92 SC |
2049 | @option{--enable-new-dtags}, the dynamic tags will be created as needed. |
2050 | If you specify @option{--disable-new-dtags}, no new dynamic tags will be | |
6c1439be L |
2051 | created. By default, the new dynamic tags are not created. Note that |
2052 | those options are only available for ELF systems. | |
2053 | ||
2d643429 | 2054 | @kindex --hash-size=@var{number} |
e185dd51 | 2055 | @item --hash-size=@var{number} |
2d643429 NC |
2056 | Set the default size of the linker's hash tables to a prime number |
2057 | close to @var{number}. Increasing this value can reduce the length of | |
2058 | time it takes the linker to perform its tasks, at the expense of | |
2059 | increasing the linker's memory requirements. Similarly reducing this | |
2060 | value can reduce the memory requirements at the expense of speed. | |
2061 | ||
fdc90cb4 JJ |
2062 | @kindex --hash-style=@var{style} |
2063 | @item --hash-style=@var{style} | |
2064 | Set the type of linker's hash table(s). @var{style} can be either | |
2065 | @code{sysv} for classic ELF @code{.hash} section, @code{gnu} for | |
2066 | new style GNU @code{.gnu.hash} section or @code{both} for both | |
2067 | the classic ELF @code{.hash} and new style GNU @code{.gnu.hash} | |
2068 | hash tables. The default is @code{sysv}. | |
2069 | ||
35835446 JR |
2070 | @kindex --reduce-memory-overheads |
2071 | @item --reduce-memory-overheads | |
2072 | This option reduces memory requirements at ld runtime, at the expense of | |
f2a8f148 | 2073 | linking speed. This was introduced to select the old O(n^2) algorithm |
35835446 | 2074 | for link map file generation, rather than the new O(n) algorithm which uses |
2d643429 NC |
2075 | about 40% more memory for symbol storage. |
2076 | ||
4f9c04f7 | 2077 | Another effect of the switch is to set the default hash table size to |
2d643429 | 2078 | 1021, which again saves memory at the cost of lengthening the linker's |
a85785bc | 2079 | run time. This is not done however if the @option{--hash-size} switch |
2d643429 NC |
2080 | has been used. |
2081 | ||
2082 | The @option{--reduce-memory-overheads} switch may be also be used to | |
2083 | enable other tradeoffs in future versions of the linker. | |
35835446 | 2084 | |
c0065db7 RM |
2085 | @kindex --build-id |
2086 | @kindex --build-id=@var{style} | |
2087 | @item --build-id | |
2088 | @itemx --build-id=@var{style} | |
2089 | Request creation of @code{.note.gnu.build-id} ELF note section. | |
2090 | The contents of the note are unique bits identifying this linked | |
2091 | file. @var{style} can be @code{uuid} to use 128 random bits, | |
24382dca RM |
2092 | @code{sha1} to use a 160-bit @sc{SHA1} hash on the normative |
2093 | parts of the output contents, @code{md5} to use a 128-bit | |
2094 | @sc{MD5} hash on the normative parts of the output contents, or | |
2095 | @code{0x@var{hexstring}} to use a chosen bit string specified as | |
2096 | an even number of hexadecimal digits (@code{-} and @code{:} | |
2097 | characters between digit pairs are ignored). If @var{style} is | |
2098 | omitted, @code{sha1} is used. | |
2099 | ||
2100 | The @code{md5} and @code{sha1} styles produces an identifier | |
2101 | that is always the same in an identical output file, but will be | |
2102 | unique among all nonidentical output files. It is not intended | |
2103 | to be compared as a checksum for the file's contents. A linked | |
2104 | file may be changed later by other tools, but the build ID bit | |
2105 | string identifying the original linked file does not change. | |
c0065db7 RM |
2106 | |
2107 | Passing @code{none} for @var{style} disables the setting from any | |
2108 | @code{--build-id} options earlier on the command line. | |
252b5132 RH |
2109 | @end table |
2110 | ||
0285c67d NC |
2111 | @c man end |
2112 | ||
36f63dca | 2113 | @subsection Options Specific to i386 PE Targets |
252b5132 | 2114 | |
0285c67d NC |
2115 | @c man begin OPTIONS |
2116 | ||
ff5dcc92 | 2117 | The i386 PE linker supports the @option{-shared} option, which causes |
252b5132 RH |
2118 | the output to be a dynamically linked library (DLL) instead of a |
2119 | normal executable. You should name the output @code{*.dll} when you | |
2120 | use this option. In addition, the linker fully supports the standard | |
2121 | @code{*.def} files, which may be specified on the linker command line | |
2122 | like an object file (in fact, it should precede archives it exports | |
2123 | symbols from, to ensure that they get linked in, just like a normal | |
2124 | object file). | |
2125 | ||
2126 | In addition to the options common to all targets, the i386 PE linker | |
2127 | support additional command line options that are specific to the i386 | |
2128 | PE target. Options that take values may be separated from their | |
2129 | values by either a space or an equals sign. | |
2130 | ||
ff5dcc92 | 2131 | @table @gcctabopt |
252b5132 RH |
2132 | |
2133 | @kindex --add-stdcall-alias | |
2134 | @item --add-stdcall-alias | |
2135 | If given, symbols with a stdcall suffix (@@@var{nn}) will be exported | |
2136 | as-is and also with the suffix stripped. | |
bb10df36 | 2137 | [This option is specific to the i386 PE targeted port of the linker] |
252b5132 RH |
2138 | |
2139 | @kindex --base-file | |
2140 | @item --base-file @var{file} | |
2141 | Use @var{file} as the name of a file in which to save the base | |
2142 | addresses of all the relocations needed for generating DLLs with | |
2143 | @file{dlltool}. | |
bb10df36 | 2144 | [This is an i386 PE specific option] |
252b5132 RH |
2145 | |
2146 | @kindex --dll | |
2147 | @item --dll | |
2148 | Create a DLL instead of a regular executable. You may also use | |
ff5dcc92 | 2149 | @option{-shared} or specify a @code{LIBRARY} in a given @code{.def} |
252b5132 | 2150 | file. |
bb10df36 | 2151 | [This option is specific to the i386 PE targeted port of the linker] |
252b5132 | 2152 | |
88183869 DK |
2153 | @kindex --enable-long-section-names |
2154 | @kindex --disable-long-section-names | |
2155 | @item --enable-long-section-names | |
2156 | @itemx --disable-long-section-names | |
2157 | The PE variants of the Coff object format add an extension that permits | |
2158 | the use of section names longer than eight characters, the normal limit | |
2159 | for Coff. By default, these names are only allowed in object files, as | |
2160 | fully-linked executable images do not carry the Coff string table required | |
2161 | to support the longer names. As a GNU extension, it is possible to | |
2162 | allow their use in executable images as well, or to (probably pointlessly!) | |
2163 | disallow it in object files, by using these two options. Executable images | |
2164 | generated with these long section names are slightly non-standard, carrying | |
2165 | as they do a string table, and may generate confusing output when examined | |
3efd345c DK |
2166 | with non-GNU PE-aware tools, such as file viewers and dumpers. However, |
2167 | GDB relies on the use of PE long section names to find Dwarf-2 debug | |
2168 | information sections in an executable image at runtime, and so if neither | |
2169 | option is specified on the command-line, @command{ld} will enable long | |
2170 | section names, overriding the default and technically correct behaviour, | |
2171 | when it finds the presence of debug information while linking an executable | |
2172 | image and not stripping symbols. | |
88183869 DK |
2173 | [This option is valid for all PE targeted ports of the linker] |
2174 | ||
252b5132 RH |
2175 | @kindex --enable-stdcall-fixup |
2176 | @kindex --disable-stdcall-fixup | |
2177 | @item --enable-stdcall-fixup | |
2178 | @itemx --disable-stdcall-fixup | |
2179 | If the link finds a symbol that it cannot resolve, it will attempt to | |
36f63dca | 2180 | do ``fuzzy linking'' by looking for another defined symbol that differs |
252b5132 RH |
2181 | only in the format of the symbol name (cdecl vs stdcall) and will |
2182 | resolve that symbol by linking to the match. For example, the | |
2183 | undefined symbol @code{_foo} might be linked to the function | |
2184 | @code{_foo@@12}, or the undefined symbol @code{_bar@@16} might be linked | |
2185 | to the function @code{_bar}. When the linker does this, it prints a | |
2186 | warning, since it normally should have failed to link, but sometimes | |
2187 | import libraries generated from third-party dlls may need this feature | |
ff5dcc92 | 2188 | to be usable. If you specify @option{--enable-stdcall-fixup}, this |
252b5132 | 2189 | feature is fully enabled and warnings are not printed. If you specify |
ff5dcc92 | 2190 | @option{--disable-stdcall-fixup}, this feature is disabled and such |
252b5132 | 2191 | mismatches are considered to be errors. |
bb10df36 | 2192 | [This option is specific to the i386 PE targeted port of the linker] |
252b5132 RH |
2193 | |
2194 | @cindex DLLs, creating | |
2195 | @kindex --export-all-symbols | |
2196 | @item --export-all-symbols | |
2197 | If given, all global symbols in the objects used to build a DLL will | |
2198 | be exported by the DLL. Note that this is the default if there | |
2199 | otherwise wouldn't be any exported symbols. When symbols are | |
2200 | explicitly exported via DEF files or implicitly exported via function | |
2201 | attributes, the default is to not export anything else unless this | |
2202 | option is given. Note that the symbols @code{DllMain@@12}, | |
ece2d90e | 2203 | @code{DllEntryPoint@@0}, @code{DllMainCRTStartup@@12}, and |
b044cda1 | 2204 | @code{impure_ptr} will not be automatically |
ece2d90e NC |
2205 | exported. Also, symbols imported from other DLLs will not be |
2206 | re-exported, nor will symbols specifying the DLL's internal layout | |
2207 | such as those beginning with @code{_head_} or ending with | |
2208 | @code{_iname}. In addition, no symbols from @code{libgcc}, | |
b044cda1 CW |
2209 | @code{libstd++}, @code{libmingw32}, or @code{crtX.o} will be exported. |
2210 | Symbols whose names begin with @code{__rtti_} or @code{__builtin_} will | |
2211 | not be exported, to help with C++ DLLs. Finally, there is an | |
ece2d90e | 2212 | extensive list of cygwin-private symbols that are not exported |
b044cda1 | 2213 | (obviously, this applies on when building DLLs for cygwin targets). |
ece2d90e | 2214 | These cygwin-excludes are: @code{_cygwin_dll_entry@@12}, |
b044cda1 | 2215 | @code{_cygwin_crt0_common@@8}, @code{_cygwin_noncygwin_dll_entry@@12}, |
ece2d90e | 2216 | @code{_fmode}, @code{_impure_ptr}, @code{cygwin_attach_dll}, |
b044cda1 | 2217 | @code{cygwin_premain0}, @code{cygwin_premain1}, @code{cygwin_premain2}, |
ece2d90e | 2218 | @code{cygwin_premain3}, and @code{environ}. |
bb10df36 | 2219 | [This option is specific to the i386 PE targeted port of the linker] |
252b5132 RH |
2220 | |
2221 | @kindex --exclude-symbols | |
1d0a3c9c | 2222 | @item --exclude-symbols @var{symbol},@var{symbol},... |
252b5132 RH |
2223 | Specifies a list of symbols which should not be automatically |
2224 | exported. The symbol names may be delimited by commas or colons. | |
bb10df36 | 2225 | [This option is specific to the i386 PE targeted port of the linker] |
252b5132 | 2226 | |
2927aaca NC |
2227 | @kindex --exclude-all-symbols |
2228 | @item --exclude-all-symbols | |
2229 | Specifies no symbols should be automatically exported. | |
2230 | [This option is specific to the i386 PE targeted port of the linker] | |
2231 | ||
252b5132 RH |
2232 | @kindex --file-alignment |
2233 | @item --file-alignment | |
2234 | Specify the file alignment. Sections in the file will always begin at | |
2235 | file offsets which are multiples of this number. This defaults to | |
2236 | 512. | |
bb10df36 | 2237 | [This option is specific to the i386 PE targeted port of the linker] |
252b5132 RH |
2238 | |
2239 | @cindex heap size | |
2240 | @kindex --heap | |
2241 | @item --heap @var{reserve} | |
2242 | @itemx --heap @var{reserve},@var{commit} | |
a00b50c5 DS |
2243 | Specify the number of bytes of memory to reserve (and optionally commit) |
2244 | to be used as heap for this program. The default is 1Mb reserved, 4K | |
252b5132 | 2245 | committed. |
bb10df36 | 2246 | [This option is specific to the i386 PE targeted port of the linker] |
252b5132 RH |
2247 | |
2248 | @cindex image base | |
2249 | @kindex --image-base | |
2250 | @item --image-base @var{value} | |
2251 | Use @var{value} as the base address of your program or dll. This is | |
2252 | the lowest memory location that will be used when your program or dll | |
2253 | is loaded. To reduce the need to relocate and improve performance of | |
2254 | your dlls, each should have a unique base address and not overlap any | |
2255 | other dlls. The default is 0x400000 for executables, and 0x10000000 | |
2256 | for dlls. | |
bb10df36 | 2257 | [This option is specific to the i386 PE targeted port of the linker] |
252b5132 RH |
2258 | |
2259 | @kindex --kill-at | |
2260 | @item --kill-at | |
2261 | If given, the stdcall suffixes (@@@var{nn}) will be stripped from | |
2262 | symbols before they are exported. | |
bb10df36 | 2263 | [This option is specific to the i386 PE targeted port of the linker] |
252b5132 | 2264 | |
26d2d8a2 BF |
2265 | @kindex --large-address-aware |
2266 | @item --large-address-aware | |
b45619c0 | 2267 | If given, the appropriate bit in the ``Characteristics'' field of the COFF |
26d2d8a2 | 2268 | header is set to indicate that this executable supports virtual addresses |
b45619c0 | 2269 | greater than 2 gigabytes. This should be used in conjunction with the /3GB |
26d2d8a2 BF |
2270 | or /USERVA=@var{value} megabytes switch in the ``[operating systems]'' |
2271 | section of the BOOT.INI. Otherwise, this bit has no effect. | |
2272 | [This option is specific to PE targeted ports of the linker] | |
2273 | ||
252b5132 RH |
2274 | @kindex --major-image-version |
2275 | @item --major-image-version @var{value} | |
36f63dca | 2276 | Sets the major number of the ``image version''. Defaults to 1. |
bb10df36 | 2277 | [This option is specific to the i386 PE targeted port of the linker] |
252b5132 RH |
2278 | |
2279 | @kindex --major-os-version | |
2280 | @item --major-os-version @var{value} | |
36f63dca | 2281 | Sets the major number of the ``os version''. Defaults to 4. |
bb10df36 | 2282 | [This option is specific to the i386 PE targeted port of the linker] |
252b5132 RH |
2283 | |
2284 | @kindex --major-subsystem-version | |
2285 | @item --major-subsystem-version @var{value} | |
36f63dca | 2286 | Sets the major number of the ``subsystem version''. Defaults to 4. |
bb10df36 | 2287 | [This option is specific to the i386 PE targeted port of the linker] |
252b5132 RH |
2288 | |
2289 | @kindex --minor-image-version | |
2290 | @item --minor-image-version @var{value} | |
36f63dca | 2291 | Sets the minor number of the ``image version''. Defaults to 0. |
bb10df36 | 2292 | [This option is specific to the i386 PE targeted port of the linker] |
252b5132 RH |
2293 | |
2294 | @kindex --minor-os-version | |
2295 | @item --minor-os-version @var{value} | |
36f63dca | 2296 | Sets the minor number of the ``os version''. Defaults to 0. |
bb10df36 | 2297 | [This option is specific to the i386 PE targeted port of the linker] |
252b5132 RH |
2298 | |
2299 | @kindex --minor-subsystem-version | |
2300 | @item --minor-subsystem-version @var{value} | |
36f63dca | 2301 | Sets the minor number of the ``subsystem version''. Defaults to 0. |
bb10df36 | 2302 | [This option is specific to the i386 PE targeted port of the linker] |
252b5132 RH |
2303 | |
2304 | @cindex DEF files, creating | |
2305 | @cindex DLLs, creating | |
2306 | @kindex --output-def | |
2307 | @item --output-def @var{file} | |
2308 | The linker will create the file @var{file} which will contain a DEF | |
2309 | file corresponding to the DLL the linker is generating. This DEF file | |
2310 | (which should be called @code{*.def}) may be used to create an import | |
2311 | library with @code{dlltool} or may be used as a reference to | |
2312 | automatically or implicitly exported symbols. | |
bb10df36 | 2313 | [This option is specific to the i386 PE targeted port of the linker] |
252b5132 | 2314 | |
b044cda1 CW |
2315 | @cindex DLLs, creating |
2316 | @kindex --out-implib | |
2317 | @item --out-implib @var{file} | |
2318 | The linker will create the file @var{file} which will contain an | |
2319 | import lib corresponding to the DLL the linker is generating. This | |
2320 | import lib (which should be called @code{*.dll.a} or @code{*.a} | |
560e09e9 | 2321 | may be used to link clients against the generated DLL; this behaviour |
b044cda1 CW |
2322 | makes it possible to skip a separate @code{dlltool} import library |
2323 | creation step. | |
bb10df36 | 2324 | [This option is specific to the i386 PE targeted port of the linker] |
b044cda1 CW |
2325 | |
2326 | @kindex --enable-auto-image-base | |
2327 | @item --enable-auto-image-base | |
2328 | Automatically choose the image base for DLLs, unless one is specified | |
2329 | using the @code{--image-base} argument. By using a hash generated | |
2330 | from the dllname to create unique image bases for each DLL, in-memory | |
2331 | collisions and relocations which can delay program execution are | |
2332 | avoided. | |
bb10df36 | 2333 | [This option is specific to the i386 PE targeted port of the linker] |
b044cda1 CW |
2334 | |
2335 | @kindex --disable-auto-image-base | |
2336 | @item --disable-auto-image-base | |
2337 | Do not automatically generate a unique image base. If there is no | |
2338 | user-specified image base (@code{--image-base}) then use the platform | |
2339 | default. | |
bb10df36 | 2340 | [This option is specific to the i386 PE targeted port of the linker] |
b044cda1 CW |
2341 | |
2342 | @cindex DLLs, linking to | |
2343 | @kindex --dll-search-prefix | |
2344 | @item --dll-search-prefix @var{string} | |
489d0400 | 2345 | When linking dynamically to a dll without an import library, |
ece2d90e | 2346 | search for @code{<string><basename>.dll} in preference to |
560e09e9 | 2347 | @code{lib<basename>.dll}. This behaviour allows easy distinction |
b044cda1 CW |
2348 | between DLLs built for the various "subplatforms": native, cygwin, |
2349 | uwin, pw, etc. For instance, cygwin DLLs typically use | |
ece2d90e | 2350 | @code{--dll-search-prefix=cyg}. |
bb10df36 | 2351 | [This option is specific to the i386 PE targeted port of the linker] |
b044cda1 CW |
2352 | |
2353 | @kindex --enable-auto-import | |
2354 | @item --enable-auto-import | |
ece2d90e NC |
2355 | Do sophisticated linking of @code{_symbol} to @code{__imp__symbol} for |
2356 | DATA imports from DLLs, and create the necessary thunking symbols when | |
4d8907ac DS |
2357 | building the import libraries with those DATA exports. Note: Use of the |
2358 | 'auto-import' extension will cause the text section of the image file | |
2359 | to be made writable. This does not conform to the PE-COFF format | |
2360 | specification published by Microsoft. | |
2361 | ||
e2a83dd0 NC |
2362 | Note - use of the 'auto-import' extension will also cause read only |
2363 | data which would normally be placed into the .rdata section to be | |
2364 | placed into the .data section instead. This is in order to work | |
2365 | around a problem with consts that is described here: | |
2366 | http://www.cygwin.com/ml/cygwin/2004-09/msg01101.html | |
2367 | ||
4d8907ac DS |
2368 | Using 'auto-import' generally will 'just work' -- but sometimes you may |
2369 | see this message: | |
0d888aac | 2370 | |
ece2d90e | 2371 | "variable '<var>' can't be auto-imported. Please read the |
0d888aac CW |
2372 | documentation for ld's @code{--enable-auto-import} for details." |
2373 | ||
ece2d90e NC |
2374 | This message occurs when some (sub)expression accesses an address |
2375 | ultimately given by the sum of two constants (Win32 import tables only | |
c0065db7 RM |
2376 | allow one). Instances where this may occur include accesses to member |
2377 | fields of struct variables imported from a DLL, as well as using a | |
2378 | constant index into an array variable imported from a DLL. Any | |
2f8d8971 NC |
2379 | multiword variable (arrays, structs, long long, etc) may trigger |
2380 | this error condition. However, regardless of the exact data type | |
2381 | of the offending exported variable, ld will always detect it, issue | |
2382 | the warning, and exit. | |
2383 | ||
2384 | There are several ways to address this difficulty, regardless of the | |
2385 | data type of the exported variable: | |
0d888aac | 2386 | |
2fa9fc65 NC |
2387 | One way is to use --enable-runtime-pseudo-reloc switch. This leaves the task |
2388 | of adjusting references in your client code for runtime environment, so | |
560e09e9 | 2389 | this method works only when runtime environment supports this feature. |
2fa9fc65 | 2390 | |
c0065db7 RM |
2391 | A second solution is to force one of the 'constants' to be a variable -- |
2392 | that is, unknown and un-optimizable at compile time. For arrays, | |
2393 | there are two possibilities: a) make the indexee (the array's address) | |
0d888aac CW |
2394 | a variable, or b) make the 'constant' index a variable. Thus: |
2395 | ||
2396 | @example | |
2397 | extern type extern_array[]; | |
c0065db7 | 2398 | extern_array[1] --> |
0d888aac CW |
2399 | @{ volatile type *t=extern_array; t[1] @} |
2400 | @end example | |
2401 | ||
2402 | or | |
2403 | ||
2404 | @example | |
2405 | extern type extern_array[]; | |
c0065db7 | 2406 | extern_array[1] --> |
0d888aac CW |
2407 | @{ volatile int t=1; extern_array[t] @} |
2408 | @end example | |
2409 | ||
c0065db7 | 2410 | For structs (and most other multiword data types) the only option |
2f8d8971 | 2411 | is to make the struct itself (or the long long, or the ...) variable: |
0d888aac CW |
2412 | |
2413 | @example | |
2414 | extern struct s extern_struct; | |
c0065db7 | 2415 | extern_struct.field --> |
0d888aac CW |
2416 | @{ volatile struct s *t=&extern_struct; t->field @} |
2417 | @end example | |
2418 | ||
c406afaf NC |
2419 | or |
2420 | ||
2421 | @example | |
2422 | extern long long extern_ll; | |
2423 | extern_ll --> | |
2424 | @{ volatile long long * local_ll=&extern_ll; *local_ll @} | |
2425 | @end example | |
2426 | ||
2fa9fc65 | 2427 | A third method of dealing with this difficulty is to abandon |
c0065db7 | 2428 | 'auto-import' for the offending symbol and mark it with |
560e09e9 | 2429 | @code{__declspec(dllimport)}. However, in practise that |
0d888aac | 2430 | requires using compile-time #defines to indicate whether you are |
c0065db7 RM |
2431 | building a DLL, building client code that will link to the DLL, or |
2432 | merely building/linking to a static library. In making the choice | |
2433 | between the various methods of resolving the 'direct address with | |
0d888aac CW |
2434 | constant offset' problem, you should consider typical real-world usage: |
2435 | ||
2436 | Original: | |
2437 | @example | |
2438 | --foo.h | |
2439 | extern int arr[]; | |
2440 | --foo.c | |
2441 | #include "foo.h" | |
2442 | void main(int argc, char **argv)@{ | |
2443 | printf("%d\n",arr[1]); | |
2444 | @} | |
2445 | @end example | |
2446 | ||
2447 | Solution 1: | |
2448 | @example | |
2449 | --foo.h | |
2450 | extern int arr[]; | |
2451 | --foo.c | |
2452 | #include "foo.h" | |
2453 | void main(int argc, char **argv)@{ | |
2454 | /* This workaround is for win32 and cygwin; do not "optimize" */ | |
2455 | volatile int *parr = arr; | |
2456 | printf("%d\n",parr[1]); | |
2457 | @} | |
2458 | @end example | |
2459 | ||
2460 | Solution 2: | |
2461 | @example | |
2462 | --foo.h | |
2463 | /* Note: auto-export is assumed (no __declspec(dllexport)) */ | |
2464 | #if (defined(_WIN32) || defined(__CYGWIN__)) && \ | |
2465 | !(defined(FOO_BUILD_DLL) || defined(FOO_STATIC)) | |
2466 | #define FOO_IMPORT __declspec(dllimport) | |
2467 | #else | |
2468 | #define FOO_IMPORT | |
2469 | #endif | |
2470 | extern FOO_IMPORT int arr[]; | |
2471 | --foo.c | |
2472 | #include "foo.h" | |
2473 | void main(int argc, char **argv)@{ | |
2474 | printf("%d\n",arr[1]); | |
2475 | @} | |
2476 | @end example | |
2477 | ||
c0065db7 | 2478 | A fourth way to avoid this problem is to re-code your |
0d888aac CW |
2479 | library to use a functional interface rather than a data interface |
2480 | for the offending variables (e.g. set_foo() and get_foo() accessor | |
2481 | functions). | |
bb10df36 | 2482 | [This option is specific to the i386 PE targeted port of the linker] |
b044cda1 CW |
2483 | |
2484 | @kindex --disable-auto-import | |
2485 | @item --disable-auto-import | |
c0065db7 | 2486 | Do not attempt to do sophisticated linking of @code{_symbol} to |
b044cda1 | 2487 | @code{__imp__symbol} for DATA imports from DLLs. |
bb10df36 | 2488 | [This option is specific to the i386 PE targeted port of the linker] |
b044cda1 | 2489 | |
2fa9fc65 NC |
2490 | @kindex --enable-runtime-pseudo-reloc |
2491 | @item --enable-runtime-pseudo-reloc | |
2492 | If your code contains expressions described in --enable-auto-import section, | |
2493 | that is, DATA imports from DLL with non-zero offset, this switch will create | |
2494 | a vector of 'runtime pseudo relocations' which can be used by runtime | |
c0065db7 | 2495 | environment to adjust references to such data in your client code. |
bb10df36 | 2496 | [This option is specific to the i386 PE targeted port of the linker] |
2fa9fc65 NC |
2497 | |
2498 | @kindex --disable-runtime-pseudo-reloc | |
2499 | @item --disable-runtime-pseudo-reloc | |
2500 | Do not create pseudo relocations for non-zero offset DATA imports from | |
2501 | DLLs. This is the default. | |
bb10df36 | 2502 | [This option is specific to the i386 PE targeted port of the linker] |
2fa9fc65 | 2503 | |
b044cda1 CW |
2504 | @kindex --enable-extra-pe-debug |
2505 | @item --enable-extra-pe-debug | |
2506 | Show additional debug info related to auto-import symbol thunking. | |
bb10df36 | 2507 | [This option is specific to the i386 PE targeted port of the linker] |
b044cda1 | 2508 | |
252b5132 RH |
2509 | @kindex --section-alignment |
2510 | @item --section-alignment | |
2511 | Sets the section alignment. Sections in memory will always begin at | |
2512 | addresses which are a multiple of this number. Defaults to 0x1000. | |
bb10df36 | 2513 | [This option is specific to the i386 PE targeted port of the linker] |
252b5132 RH |
2514 | |
2515 | @cindex stack size | |
2516 | @kindex --stack | |
2517 | @item --stack @var{reserve} | |
2518 | @itemx --stack @var{reserve},@var{commit} | |
a00b50c5 DS |
2519 | Specify the number of bytes of memory to reserve (and optionally commit) |
2520 | to be used as stack for this program. The default is 2Mb reserved, 4K | |
252b5132 | 2521 | committed. |
bb10df36 | 2522 | [This option is specific to the i386 PE targeted port of the linker] |
252b5132 RH |
2523 | |
2524 | @kindex --subsystem | |
2525 | @item --subsystem @var{which} | |
2526 | @itemx --subsystem @var{which}:@var{major} | |
2527 | @itemx --subsystem @var{which}:@var{major}.@var{minor} | |
2528 | Specifies the subsystem under which your program will execute. The | |
2529 | legal values for @var{which} are @code{native}, @code{windows}, | |
33f362e1 NC |
2530 | @code{console}, @code{posix}, and @code{xbox}. You may optionally set |
2531 | the subsystem version also. Numeric values are also accepted for | |
2532 | @var{which}. | |
bb10df36 | 2533 | [This option is specific to the i386 PE targeted port of the linker] |
252b5132 | 2534 | |
2f563b51 DK |
2535 | The following options set flags in the @code{DllCharacteristics} field |
2536 | of the PE file header: | |
2537 | [These options are specific to PE targeted ports of the linker] | |
2538 | ||
2539 | @kindex --dynamicbase | |
2540 | @item --dynamicbase | |
2541 | The image base address may be relocated using address space layout | |
2542 | randomization (ASLR). This feature was introduced with MS Windows | |
2543 | Vista for i386 PE targets. | |
2544 | ||
2545 | @kindex --forceinteg | |
2546 | @item --forceinteg | |
2547 | Code integrity checks are enforced. | |
2548 | ||
2549 | @kindex --nxcompat | |
2550 | @item --nxcompat | |
2551 | The image is compatible with the Data Execution Prevention. | |
2552 | This feature was introduced with MS Windows XP SP2 for i386 PE targets. | |
2553 | ||
2554 | @kindex --no-isolation | |
2555 | @item --no-isolation | |
2556 | Although the image understands isolation, do not isolate the image. | |
2557 | ||
2558 | @kindex --no-seh | |
2559 | @item --no-seh | |
2560 | The image does not use SEH. No SE handler may be called from | |
2561 | this image. | |
2562 | ||
2563 | @kindex --no-bind | |
2564 | @item --no-bind | |
2565 | Do not bind this image. | |
2566 | ||
2567 | @kindex --wdmdriver | |
2568 | @item --wdmdriver | |
2569 | The driver uses the MS Windows Driver Model. | |
2570 | ||
2571 | @kindex --tsaware | |
2572 | @item --tsaware | |
2573 | The image is Terminal Server aware. | |
2574 | ||
252b5132 RH |
2575 | @end table |
2576 | ||
0285c67d NC |
2577 | @c man end |
2578 | ||
93fd0973 SC |
2579 | @ifset M68HC11 |
2580 | @subsection Options specific to Motorola 68HC11 and 68HC12 targets | |
2581 | ||
2582 | @c man begin OPTIONS | |
2583 | ||
2584 | The 68HC11 and 68HC12 linkers support specific options to control the | |
2585 | memory bank switching mapping and trampoline code generation. | |
2586 | ||
2587 | @table @gcctabopt | |
2588 | ||
2589 | @kindex --no-trampoline | |
2590 | @item --no-trampoline | |
2591 | This option disables the generation of trampoline. By default a trampoline | |
2592 | is generated for each far function which is called using a @code{jsr} | |
2593 | instruction (this happens when a pointer to a far function is taken). | |
2594 | ||
2595 | @kindex --bank-window | |
2596 | @item --bank-window @var{name} | |
2597 | This option indicates to the linker the name of the memory region in | |
2598 | the @samp{MEMORY} specification that describes the memory bank window. | |
2599 | The definition of such region is then used by the linker to compute | |
2600 | paging and addresses within the memory window. | |
2601 | ||
2602 | @end table | |
2603 | ||
2604 | @c man end | |
2605 | @end ifset | |
2606 | ||
7fb9f789 NC |
2607 | @ifset M68K |
2608 | @subsection Options specific to Motorola 68K target | |
2609 | ||
2610 | @c man begin OPTIONS | |
2611 | ||
2612 | The following options are supported to control handling of GOT generation | |
2613 | when linking for 68K targets. | |
2614 | ||
2615 | @table @gcctabopt | |
2616 | ||
2617 | @kindex --got | |
2618 | @item --got=@var{type} | |
2619 | This option tells the linker which GOT generation scheme to use. | |
2620 | @var{type} should be one of @samp{single}, @samp{negative}, | |
2621 | @samp{multigot} or @samp{target}. For more information refer to the | |
2622 | Info entry for @file{ld}. | |
2623 | ||
2624 | @end table | |
2625 | ||
2626 | @c man end | |
2627 | @end ifset | |
2628 | ||
252b5132 RH |
2629 | @ifset UsesEnvVars |
2630 | @node Environment | |
2631 | @section Environment Variables | |
2632 | ||
0285c67d NC |
2633 | @c man begin ENVIRONMENT |
2634 | ||
560e09e9 | 2635 | You can change the behaviour of @command{ld} with the environment variables |
36f63dca NC |
2636 | @ifclear SingleFormat |
2637 | @code{GNUTARGET}, | |
2638 | @end ifclear | |
2639 | @code{LDEMULATION} and @code{COLLECT_NO_DEMANGLE}. | |
252b5132 | 2640 | |
36f63dca | 2641 | @ifclear SingleFormat |
252b5132 RH |
2642 | @kindex GNUTARGET |
2643 | @cindex default input format | |
2644 | @code{GNUTARGET} determines the input-file object format if you don't | |
2645 | use @samp{-b} (or its synonym @samp{--format}). Its value should be one | |
2646 | of the BFD names for an input format (@pxref{BFD}). If there is no | |
ff5dcc92 | 2647 | @code{GNUTARGET} in the environment, @command{ld} uses the natural format |
252b5132 RH |
2648 | of the target. If @code{GNUTARGET} is set to @code{default} then BFD |
2649 | attempts to discover the input format by examining binary input files; | |
2650 | this method often succeeds, but there are potential ambiguities, since | |
2651 | there is no method of ensuring that the magic number used to specify | |
2652 | object-file formats is unique. However, the configuration procedure for | |
2653 | BFD on each system places the conventional format for that system first | |
2654 | in the search-list, so ambiguities are resolved in favor of convention. | |
36f63dca | 2655 | @end ifclear |
252b5132 RH |
2656 | |
2657 | @kindex LDEMULATION | |
2658 | @cindex default emulation | |
2659 | @cindex emulation, default | |
2660 | @code{LDEMULATION} determines the default emulation if you don't use the | |
2661 | @samp{-m} option. The emulation can affect various aspects of linker | |
2662 | behaviour, particularly the default linker script. You can list the | |
2663 | available emulations with the @samp{--verbose} or @samp{-V} options. If | |
2664 | the @samp{-m} option is not used, and the @code{LDEMULATION} environment | |
2665 | variable is not defined, the default emulation depends upon how the | |
2666 | linker was configured. | |
252b5132 RH |
2667 | |
2668 | @kindex COLLECT_NO_DEMANGLE | |
2669 | @cindex demangling, default | |
2670 | Normally, the linker will default to demangling symbols. However, if | |
2671 | @code{COLLECT_NO_DEMANGLE} is set in the environment, then it will | |
2672 | default to not demangling symbols. This environment variable is used in | |
2673 | a similar fashion by the @code{gcc} linker wrapper program. The default | |
2674 | may be overridden by the @samp{--demangle} and @samp{--no-demangle} | |
2675 | options. | |
2676 | ||
0285c67d NC |
2677 | @c man end |
2678 | @end ifset | |
2679 | ||
252b5132 RH |
2680 | @node Scripts |
2681 | @chapter Linker Scripts | |
2682 | ||
2683 | @cindex scripts | |
2684 | @cindex linker scripts | |
2685 | @cindex command files | |
2686 | Every link is controlled by a @dfn{linker script}. This script is | |
2687 | written in the linker command language. | |
2688 | ||
2689 | The main purpose of the linker script is to describe how the sections in | |
2690 | the input files should be mapped into the output file, and to control | |
2691 | the memory layout of the output file. Most linker scripts do nothing | |
2692 | more than this. However, when necessary, the linker script can also | |
2693 | direct the linker to perform many other operations, using the commands | |
2694 | described below. | |
2695 | ||
2696 | The linker always uses a linker script. If you do not supply one | |
2697 | yourself, the linker will use a default script that is compiled into the | |
2698 | linker executable. You can use the @samp{--verbose} command line option | |
2699 | to display the default linker script. Certain command line options, | |
2700 | such as @samp{-r} or @samp{-N}, will affect the default linker script. | |
2701 | ||
2702 | You may supply your own linker script by using the @samp{-T} command | |
2703 | line option. When you do this, your linker script will replace the | |
2704 | default linker script. | |
2705 | ||
2706 | You may also use linker scripts implicitly by naming them as input files | |
2707 | to the linker, as though they were files to be linked. @xref{Implicit | |
2708 | Linker Scripts}. | |
2709 | ||
2710 | @menu | |
2711 | * Basic Script Concepts:: Basic Linker Script Concepts | |
2712 | * Script Format:: Linker Script Format | |
2713 | * Simple Example:: Simple Linker Script Example | |
2714 | * Simple Commands:: Simple Linker Script Commands | |
2715 | * Assignments:: Assigning Values to Symbols | |
2716 | * SECTIONS:: SECTIONS Command | |
2717 | * MEMORY:: MEMORY Command | |
2718 | * PHDRS:: PHDRS Command | |
2719 | * VERSION:: VERSION Command | |
2720 | * Expressions:: Expressions in Linker Scripts | |
2721 | * Implicit Linker Scripts:: Implicit Linker Scripts | |
2722 | @end menu | |
2723 | ||
2724 | @node Basic Script Concepts | |
2725 | @section Basic Linker Script Concepts | |
2726 | @cindex linker script concepts | |
2727 | We need to define some basic concepts and vocabulary in order to | |
2728 | describe the linker script language. | |
2729 | ||
2730 | The linker combines input files into a single output file. The output | |
2731 | file and each input file are in a special data format known as an | |
2732 | @dfn{object file format}. Each file is called an @dfn{object file}. | |
2733 | The output file is often called an @dfn{executable}, but for our | |
2734 | purposes we will also call it an object file. Each object file has, | |
2735 | among other things, a list of @dfn{sections}. We sometimes refer to a | |
2736 | section in an input file as an @dfn{input section}; similarly, a section | |
2737 | in the output file is an @dfn{output section}. | |
2738 | ||
2739 | Each section in an object file has a name and a size. Most sections | |
2740 | also have an associated block of data, known as the @dfn{section | |
2741 | contents}. A section may be marked as @dfn{loadable}, which mean that | |
2742 | the contents should be loaded into memory when the output file is run. | |
2743 | A section with no contents may be @dfn{allocatable}, which means that an | |
2744 | area in memory should be set aside, but nothing in particular should be | |
2745 | loaded there (in some cases this memory must be zeroed out). A section | |
2746 | which is neither loadable nor allocatable typically contains some sort | |
2747 | of debugging information. | |
2748 | ||
2749 | Every loadable or allocatable output section has two addresses. The | |
2750 | first is the @dfn{VMA}, or virtual memory address. This is the address | |
2751 | the section will have when the output file is run. The second is the | |
2752 | @dfn{LMA}, or load memory address. This is the address at which the | |
2753 | section will be loaded. In most cases the two addresses will be the | |
2754 | same. An example of when they might be different is when a data section | |
2755 | is loaded into ROM, and then copied into RAM when the program starts up | |
2756 | (this technique is often used to initialize global variables in a ROM | |
2757 | based system). In this case the ROM address would be the LMA, and the | |
2758 | RAM address would be the VMA. | |
2759 | ||
2760 | You can see the sections in an object file by using the @code{objdump} | |
2761 | program with the @samp{-h} option. | |
2762 | ||
2763 | Every object file also has a list of @dfn{symbols}, known as the | |
2764 | @dfn{symbol table}. A symbol may be defined or undefined. Each symbol | |
2765 | has a name, and each defined symbol has an address, among other | |
2766 | information. If you compile a C or C++ program into an object file, you | |
2767 | will get a defined symbol for every defined function and global or | |
2768 | static variable. Every undefined function or global variable which is | |
2769 | referenced in the input file will become an undefined symbol. | |
2770 | ||
2771 | You can see the symbols in an object file by using the @code{nm} | |
2772 | program, or by using the @code{objdump} program with the @samp{-t} | |
2773 | option. | |
2774 | ||
2775 | @node Script Format | |
2776 | @section Linker Script Format | |
2777 | @cindex linker script format | |
2778 | Linker scripts are text files. | |
2779 | ||
2780 | You write a linker script as a series of commands. Each command is | |
2781 | either a keyword, possibly followed by arguments, or an assignment to a | |
2782 | symbol. You may separate commands using semicolons. Whitespace is | |
2783 | generally ignored. | |
2784 | ||
2785 | Strings such as file or format names can normally be entered directly. | |
2786 | If the file name contains a character such as a comma which would | |
2787 | otherwise serve to separate file names, you may put the file name in | |
2788 | double quotes. There is no way to use a double quote character in a | |
2789 | file name. | |
2790 | ||
2791 | You may include comments in linker scripts just as in C, delimited by | |
2792 | @samp{/*} and @samp{*/}. As in C, comments are syntactically equivalent | |
2793 | to whitespace. | |
2794 | ||
2795 | @node Simple Example | |
2796 | @section Simple Linker Script Example | |
2797 | @cindex linker script example | |
2798 | @cindex example of linker script | |
2799 | Many linker scripts are fairly simple. | |
2800 | ||
2801 | The simplest possible linker script has just one command: | |
2802 | @samp{SECTIONS}. You use the @samp{SECTIONS} command to describe the | |
2803 | memory layout of the output file. | |
2804 | ||
2805 | The @samp{SECTIONS} command is a powerful command. Here we will | |
2806 | describe a simple use of it. Let's assume your program consists only of | |
2807 | code, initialized data, and uninitialized data. These will be in the | |
2808 | @samp{.text}, @samp{.data}, and @samp{.bss} sections, respectively. | |
2809 | Let's assume further that these are the only sections which appear in | |
2810 | your input files. | |
2811 | ||
2812 | For this example, let's say that the code should be loaded at address | |
2813 | 0x10000, and that the data should start at address 0x8000000. Here is a | |
2814 | linker script which will do that: | |
2815 | @smallexample | |
2816 | SECTIONS | |
2817 | @{ | |
2818 | . = 0x10000; | |
2819 | .text : @{ *(.text) @} | |
2820 | . = 0x8000000; | |
2821 | .data : @{ *(.data) @} | |
2822 | .bss : @{ *(.bss) @} | |
2823 | @} | |
2824 | @end smallexample | |
2825 | ||
2826 | You write the @samp{SECTIONS} command as the keyword @samp{SECTIONS}, | |
2827 | followed by a series of symbol assignments and output section | |
2828 | descriptions enclosed in curly braces. | |
2829 | ||
252b5132 RH |
2830 | The first line inside the @samp{SECTIONS} command of the above example |
2831 | sets the value of the special symbol @samp{.}, which is the location | |
2832 | counter. If you do not specify the address of an output section in some | |
2833 | other way (other ways are described later), the address is set from the | |
2834 | current value of the location counter. The location counter is then | |
2835 | incremented by the size of the output section. At the start of the | |
2836 | @samp{SECTIONS} command, the location counter has the value @samp{0}. | |
2837 | ||
2838 | The second line defines an output section, @samp{.text}. The colon is | |
2839 | required syntax which may be ignored for now. Within the curly braces | |
2840 | after the output section name, you list the names of the input sections | |
2841 | which should be placed into this output section. The @samp{*} is a | |
2842 | wildcard which matches any file name. The expression @samp{*(.text)} | |
2843 | means all @samp{.text} input sections in all input files. | |
2844 | ||
2845 | Since the location counter is @samp{0x10000} when the output section | |
2846 | @samp{.text} is defined, the linker will set the address of the | |
2847 | @samp{.text} section in the output file to be @samp{0x10000}. | |
2848 | ||
2849 | The remaining lines define the @samp{.data} and @samp{.bss} sections in | |
2850 | the output file. The linker will place the @samp{.data} output section | |
2851 | at address @samp{0x8000000}. After the linker places the @samp{.data} | |
2852 | output section, the value of the location counter will be | |
2853 | @samp{0x8000000} plus the size of the @samp{.data} output section. The | |
2854 | effect is that the linker will place the @samp{.bss} output section | |
58434bc1 | 2855 | immediately after the @samp{.data} output section in memory. |
252b5132 RH |
2856 | |
2857 | The linker will ensure that each output section has the required | |
2858 | alignment, by increasing the location counter if necessary. In this | |
2859 | example, the specified addresses for the @samp{.text} and @samp{.data} | |
2860 | sections will probably satisfy any alignment constraints, but the linker | |
2861 | may have to create a small gap between the @samp{.data} and @samp{.bss} | |
2862 | sections. | |
2863 | ||
2864 | That's it! That's a simple and complete linker script. | |
2865 | ||
2866 | @node Simple Commands | |
2867 | @section Simple Linker Script Commands | |
2868 | @cindex linker script simple commands | |
2869 | In this section we describe the simple linker script commands. | |
2870 | ||
2871 | @menu | |
2872 | * Entry Point:: Setting the entry point | |
2873 | * File Commands:: Commands dealing with files | |
2874 | @ifclear SingleFormat | |
2875 | * Format Commands:: Commands dealing with object file formats | |
2876 | @end ifclear | |
2877 | ||
4a93e180 | 2878 | * REGION_ALIAS:: Assign alias names to memory regions |
252b5132 RH |
2879 | * Miscellaneous Commands:: Other linker script commands |
2880 | @end menu | |
2881 | ||
2882 | @node Entry Point | |
36f63dca | 2883 | @subsection Setting the Entry Point |
252b5132 RH |
2884 | @kindex ENTRY(@var{symbol}) |
2885 | @cindex start of execution | |
2886 | @cindex first instruction | |
2887 | @cindex entry point | |
2888 | The first instruction to execute in a program is called the @dfn{entry | |
2889 | point}. You can use the @code{ENTRY} linker script command to set the | |
2890 | entry point. The argument is a symbol name: | |
2891 | @smallexample | |
2892 | ENTRY(@var{symbol}) | |
2893 | @end smallexample | |
2894 | ||
2895 | There are several ways to set the entry point. The linker will set the | |
2896 | entry point by trying each of the following methods in order, and | |
2897 | stopping when one of them succeeds: | |
2898 | @itemize @bullet | |
a1ab1d2a | 2899 | @item |
252b5132 | 2900 | the @samp{-e} @var{entry} command-line option; |
a1ab1d2a | 2901 | @item |
252b5132 | 2902 | the @code{ENTRY(@var{symbol})} command in a linker script; |
a1ab1d2a | 2903 | @item |
252b5132 | 2904 | the value of the symbol @code{start}, if defined; |
a1ab1d2a | 2905 | @item |
252b5132 | 2906 | the address of the first byte of the @samp{.text} section, if present; |
a1ab1d2a | 2907 | @item |
252b5132 RH |
2908 | The address @code{0}. |
2909 | @end itemize | |
2910 | ||
2911 | @node File Commands | |
36f63dca | 2912 | @subsection Commands Dealing with Files |
252b5132 RH |
2913 | @cindex linker script file commands |
2914 | Several linker script commands deal with files. | |
2915 | ||
2916 | @table @code | |
2917 | @item INCLUDE @var{filename} | |
2918 | @kindex INCLUDE @var{filename} | |
2919 | @cindex including a linker script | |
2920 | Include the linker script @var{filename} at this point. The file will | |
2921 | be searched for in the current directory, and in any directory specified | |
ff5dcc92 | 2922 | with the @option{-L} option. You can nest calls to @code{INCLUDE} up to |
252b5132 RH |
2923 | 10 levels deep. |
2924 | ||
4006703d NS |
2925 | You can place @code{INCLUDE} directives at the top level, in @code{MEMORY} or |
2926 | @code{SECTIONS} commands, or in output section descriptions. | |
2927 | ||
252b5132 RH |
2928 | @item INPUT(@var{file}, @var{file}, @dots{}) |
2929 | @itemx INPUT(@var{file} @var{file} @dots{}) | |
2930 | @kindex INPUT(@var{files}) | |
2931 | @cindex input files in linker scripts | |
2932 | @cindex input object files in linker scripts | |
2933 | @cindex linker script input object files | |
2934 | The @code{INPUT} command directs the linker to include the named files | |
2935 | in the link, as though they were named on the command line. | |
2936 | ||
2937 | For example, if you always want to include @file{subr.o} any time you do | |
2938 | a link, but you can't be bothered to put it on every link command line, | |
2939 | then you can put @samp{INPUT (subr.o)} in your linker script. | |
2940 | ||
2941 | In fact, if you like, you can list all of your input files in the linker | |
2942 | script, and then invoke the linker with nothing but a @samp{-T} option. | |
2943 | ||
e3f2db7f AO |
2944 | In case a @dfn{sysroot prefix} is configured, and the filename starts |
2945 | with the @samp{/} character, and the script being processed was | |
2946 | located inside the @dfn{sysroot prefix}, the filename will be looked | |
2947 | for in the @dfn{sysroot prefix}. Otherwise, the linker will try to | |
2948 | open the file in the current directory. If it is not found, the | |
2949 | linker will search through the archive library search path. See the | |
2950 | description of @samp{-L} in @ref{Options,,Command Line Options}. | |
252b5132 | 2951 | |
ff5dcc92 | 2952 | If you use @samp{INPUT (-l@var{file})}, @command{ld} will transform the |
252b5132 RH |
2953 | name to @code{lib@var{file}.a}, as with the command line argument |
2954 | @samp{-l}. | |
2955 | ||
2956 | When you use the @code{INPUT} command in an implicit linker script, the | |
2957 | files will be included in the link at the point at which the linker | |
2958 | script file is included. This can affect archive searching. | |
2959 | ||
2960 | @item GROUP(@var{file}, @var{file}, @dots{}) | |
2961 | @itemx GROUP(@var{file} @var{file} @dots{}) | |
2962 | @kindex GROUP(@var{files}) | |
2963 | @cindex grouping input files | |
2964 | The @code{GROUP} command is like @code{INPUT}, except that the named | |
2965 | files should all be archives, and they are searched repeatedly until no | |
2966 | new undefined references are created. See the description of @samp{-(} | |
2967 | in @ref{Options,,Command Line Options}. | |
2968 | ||
b717d30e JJ |
2969 | @item AS_NEEDED(@var{file}, @var{file}, @dots{}) |
2970 | @itemx AS_NEEDED(@var{file} @var{file} @dots{}) | |
2971 | @kindex AS_NEEDED(@var{files}) | |
2972 | This construct can appear only inside of the @code{INPUT} or @code{GROUP} | |
2973 | commands, among other filenames. The files listed will be handled | |
2974 | as if they appear directly in the @code{INPUT} or @code{GROUP} commands, | |
2975 | with the exception of ELF shared libraries, that will be added only | |
2976 | when they are actually needed. This construct essentially enables | |
2977 | @option{--as-needed} option for all the files listed inside of it | |
2978 | and restores previous @option{--as-needed} resp. @option{--no-as-needed} | |
2979 | setting afterwards. | |
2980 | ||
252b5132 RH |
2981 | @item OUTPUT(@var{filename}) |
2982 | @kindex OUTPUT(@var{filename}) | |
b45619c0 | 2983 | @cindex output file name in linker script |
252b5132 RH |
2984 | The @code{OUTPUT} command names the output file. Using |
2985 | @code{OUTPUT(@var{filename})} in the linker script is exactly like using | |
2986 | @samp{-o @var{filename}} on the command line (@pxref{Options,,Command | |
2987 | Line Options}). If both are used, the command line option takes | |
2988 | precedence. | |
2989 | ||
2990 | You can use the @code{OUTPUT} command to define a default name for the | |
2991 | output file other than the usual default of @file{a.out}. | |
2992 | ||
2993 | @item SEARCH_DIR(@var{path}) | |
2994 | @kindex SEARCH_DIR(@var{path}) | |
2995 | @cindex library search path in linker script | |
2996 | @cindex archive search path in linker script | |
2997 | @cindex search path in linker script | |
2998 | The @code{SEARCH_DIR} command adds @var{path} to the list of paths where | |
ff5dcc92 | 2999 | @command{ld} looks for archive libraries. Using |
252b5132 RH |
3000 | @code{SEARCH_DIR(@var{path})} is exactly like using @samp{-L @var{path}} |
3001 | on the command line (@pxref{Options,,Command Line Options}). If both | |
3002 | are used, then the linker will search both paths. Paths specified using | |
3003 | the command line option are searched first. | |
3004 | ||
3005 | @item STARTUP(@var{filename}) | |
3006 | @kindex STARTUP(@var{filename}) | |
3007 | @cindex first input file | |
3008 | The @code{STARTUP} command is just like the @code{INPUT} command, except | |
3009 | that @var{filename} will become the first input file to be linked, as | |
3010 | though it were specified first on the command line. This may be useful | |
3011 | when using a system in which the entry point is always the start of the | |
3012 | first file. | |
3013 | @end table | |
3014 | ||
3015 | @ifclear SingleFormat | |
3016 | @node Format Commands | |
36f63dca | 3017 | @subsection Commands Dealing with Object File Formats |
252b5132 RH |
3018 | A couple of linker script commands deal with object file formats. |
3019 | ||
3020 | @table @code | |
3021 | @item OUTPUT_FORMAT(@var{bfdname}) | |
3022 | @itemx OUTPUT_FORMAT(@var{default}, @var{big}, @var{little}) | |
3023 | @kindex OUTPUT_FORMAT(@var{bfdname}) | |
3024 | @cindex output file format in linker script | |
3025 | The @code{OUTPUT_FORMAT} command names the BFD format to use for the | |
3026 | output file (@pxref{BFD}). Using @code{OUTPUT_FORMAT(@var{bfdname})} is | |
024531e2 | 3027 | exactly like using @samp{--oformat @var{bfdname}} on the command line |
252b5132 RH |
3028 | (@pxref{Options,,Command Line Options}). If both are used, the command |
3029 | line option takes precedence. | |
3030 | ||
3031 | You can use @code{OUTPUT_FORMAT} with three arguments to use different | |
3032 | formats based on the @samp{-EB} and @samp{-EL} command line options. | |
3033 | This permits the linker script to set the output format based on the | |
3034 | desired endianness. | |
3035 | ||
3036 | If neither @samp{-EB} nor @samp{-EL} are used, then the output format | |
3037 | will be the first argument, @var{default}. If @samp{-EB} is used, the | |
3038 | output format will be the second argument, @var{big}. If @samp{-EL} is | |
3039 | used, the output format will be the third argument, @var{little}. | |
3040 | ||
3041 | For example, the default linker script for the MIPS ELF target uses this | |
3042 | command: | |
3043 | @smallexample | |
3044 | OUTPUT_FORMAT(elf32-bigmips, elf32-bigmips, elf32-littlemips) | |
3045 | @end smallexample | |
3046 | This says that the default format for the output file is | |
3047 | @samp{elf32-bigmips}, but if the user uses the @samp{-EL} command line | |
3048 | option, the output file will be created in the @samp{elf32-littlemips} | |
3049 | format. | |
3050 | ||
3051 | @item TARGET(@var{bfdname}) | |
3052 | @kindex TARGET(@var{bfdname}) | |
3053 | @cindex input file format in linker script | |
3054 | The @code{TARGET} command names the BFD format to use when reading input | |
3055 | files. It affects subsequent @code{INPUT} and @code{GROUP} commands. | |
3056 | This command is like using @samp{-b @var{bfdname}} on the command line | |
3057 | (@pxref{Options,,Command Line Options}). If the @code{TARGET} command | |
3058 | is used but @code{OUTPUT_FORMAT} is not, then the last @code{TARGET} | |
3059 | command is also used to set the format for the output file. @xref{BFD}. | |
3060 | @end table | |
3061 | @end ifclear | |
3062 | ||
4a93e180 NC |
3063 | @node REGION_ALIAS |
3064 | @subsection Assign alias names to memory regions | |
3065 | @kindex REGION_ALIAS(@var{alias}, @var{region}) | |
3066 | @cindex region alias | |
3067 | @cindex region names | |
3068 | ||
3069 | Alias names can be added to existing memory regions created with the | |
3070 | @ref{MEMORY} command. Each name corresponds to at most one memory region. | |
3071 | ||
3072 | @smallexample | |
3073 | REGION_ALIAS(@var{alias}, @var{region}) | |
3074 | @end smallexample | |
3075 | ||
3076 | The @code{REGION_ALIAS} function creates an alias name @var{alias} for the | |
3077 | memory region @var{region}. This allows a flexible mapping of output sections | |
3078 | to memory regions. An example follows. | |
3079 | ||
3080 | Suppose we have an application for embedded systems which come with various | |
3081 | memory storage devices. All have a general purpose, volatile memory @code{RAM} | |
3082 | that allows code execution or data storage. Some may have a read-only, | |
3083 | non-volatile memory @code{ROM} that allows code execution and read-only data | |
3084 | access. The last variant is a read-only, non-volatile memory @code{ROM2} with | |
3085 | read-only data access and no code execution capability. We have four output | |
3086 | sections: | |
3087 | ||
3088 | @itemize @bullet | |
3089 | @item | |
3090 | @code{.text} program code; | |
3091 | @item | |
3092 | @code{.rodata} read-only data; | |
3093 | @item | |
3094 | @code{.data} read-write initialized data; | |
3095 | @item | |
3096 | @code{.bss} read-write zero initialized data. | |
3097 | @end itemize | |
3098 | ||
3099 | The goal is to provide a linker command file that contains a system independent | |
3100 | part defining the output sections and a system dependent part mapping the | |
3101 | output sections to the memory regions available on the system. Our embedded | |
3102 | systems come with three different memory setups @code{A}, @code{B} and | |
3103 | @code{C}: | |
3104 | @multitable @columnfractions .25 .25 .25 .25 | |
3105 | @item Section @tab Variant A @tab Variant B @tab Variant C | |
3106 | @item .text @tab RAM @tab ROM @tab ROM | |
3107 | @item .rodata @tab RAM @tab ROM @tab ROM2 | |
3108 | @item .data @tab RAM @tab RAM/ROM @tab RAM/ROM2 | |
3109 | @item .bss @tab RAM @tab RAM @tab RAM | |
3110 | @end multitable | |
3111 | The notation @code{RAM/ROM} or @code{RAM/ROM2} means that this section is | |
3112 | loaded into region @code{ROM} or @code{ROM2} respectively. Please note that | |
3113 | the load address of the @code{.data} section starts in all three variants at | |
3114 | the end of the @code{.rodata} section. | |
3115 | ||
3116 | The base linker script that deals with the output sections follows. It | |
3117 | includes the system dependent @code{linkcmds.memory} file that describes the | |
3118 | memory layout: | |
3119 | @smallexample | |
3120 | INCLUDE linkcmds.memory | |
3121 | ||
3122 | SECTIONS | |
3123 | @{ | |
3124 | .text : | |
3125 | @{ | |
3126 | *(.text) | |
3127 | @} > REGION_TEXT | |
3128 | .rodata : | |
3129 | @{ | |
3130 | *(.rodata) | |
3131 | rodata_end = .; | |
3132 | @} > REGION_RODATA | |
3133 | .data : AT (rodata_end) | |
3134 | @{ | |
3135 | data_start = .; | |
3136 | *(.data) | |
3137 | @} > REGION_DATA | |
3138 | data_size = SIZEOF(.data); | |
3139 | data_load_start = LOADADDR(.data); | |
3140 | .bss : | |
3141 | @{ | |
3142 | *(.bss) | |
3143 | @} > REGION_BSS | |
3144 | @} | |
3145 | @end smallexample | |
3146 | ||
3147 | Now we need three different @code{linkcmds.memory} files to define memory | |
3148 | regions and alias names. The content of @code{linkcmds.memory} for the three | |
3149 | variants @code{A}, @code{B} and @code{C}: | |
3150 | @table @code | |
3151 | @item A | |
3152 | Here everything goes into the @code{RAM}. | |
3153 | @smallexample | |
3154 | MEMORY | |
3155 | @{ | |
3156 | RAM : ORIGIN = 0, LENGTH = 4M | |
3157 | @} | |
3158 | ||
3159 | REGION_ALIAS("REGION_TEXT", RAM); | |
3160 | REGION_ALIAS("REGION_RODATA", RAM); | |
3161 | REGION_ALIAS("REGION_DATA", RAM); | |
3162 | REGION_ALIAS("REGION_BSS", RAM); | |
3163 | @end smallexample | |
3164 | @item B | |
3165 | Program code and read-only data go into the @code{ROM}. Read-write data goes | |
3166 | into the @code{RAM}. An image of the initialized data is loaded into the | |
3167 | @code{ROM} and will be copied during system start into the @code{RAM}. | |
3168 | @smallexample | |
3169 | MEMORY | |
3170 | @{ | |
3171 | ROM : ORIGIN = 0, LENGTH = 3M | |
3172 | RAM : ORIGIN = 0x10000000, LENGTH = 1M | |
3173 | @} | |
3174 | ||
3175 | REGION_ALIAS("REGION_TEXT", ROM); | |
3176 | REGION_ALIAS("REGION_RODATA", ROM); | |
3177 | REGION_ALIAS("REGION_DATA", RAM); | |
3178 | REGION_ALIAS("REGION_BSS", RAM); | |
3179 | @end smallexample | |
3180 | @item C | |
3181 | Program code goes into the @code{ROM}. Read-only data goes into the | |
3182 | @code{ROM2}. Read-write data goes into the @code{RAM}. An image of the | |
3183 | initialized data is loaded into the @code{ROM2} and will be copied during | |
3184 | system start into the @code{RAM}. | |
3185 | @smallexample | |
3186 | MEMORY | |
3187 | @{ | |
3188 | ROM : ORIGIN = 0, LENGTH = 2M | |
3189 | ROM2 : ORIGIN = 0x10000000, LENGTH = 1M | |
3190 | RAM : ORIGIN = 0x20000000, LENGTH = 1M | |
3191 | @} | |
3192 | ||
3193 | REGION_ALIAS("REGION_TEXT", ROM); | |
3194 | REGION_ALIAS("REGION_RODATA", ROM2); | |
3195 | REGION_ALIAS("REGION_DATA", RAM); | |
3196 | REGION_ALIAS("REGION_BSS", RAM); | |
3197 | @end smallexample | |
3198 | @end table | |
3199 | ||
3200 | It is possible to write a common system initialization routine to copy the | |
3201 | @code{.data} section from @code{ROM} or @code{ROM2} into the @code{RAM} if | |
3202 | necessary: | |
3203 | @smallexample | |
3204 | #include <string.h> | |
3205 | ||
3206 | extern char data_start []; | |
3207 | extern char data_size []; | |
3208 | extern char data_load_start []; | |
3209 | ||
3210 | void copy_data(void) | |
3211 | @{ | |
3212 | if (data_start != data_load_start) | |
3213 | @{ | |
3214 | memcpy(data_start, data_load_start, (size_t) data_size); | |
3215 | @} | |
3216 | @} | |
3217 | @end smallexample | |
3218 | ||
252b5132 | 3219 | @node Miscellaneous Commands |
36f63dca | 3220 | @subsection Other Linker Script Commands |
252b5132 RH |
3221 | There are a few other linker scripts commands. |
3222 | ||
3223 | @table @code | |
3224 | @item ASSERT(@var{exp}, @var{message}) | |
3225 | @kindex ASSERT | |
3226 | @cindex assertion in linker script | |
3227 | Ensure that @var{exp} is non-zero. If it is zero, then exit the linker | |
3228 | with an error code, and print @var{message}. | |
3229 | ||
3230 | @item EXTERN(@var{symbol} @var{symbol} @dots{}) | |
3231 | @kindex EXTERN | |
3232 | @cindex undefined symbol in linker script | |
3233 | Force @var{symbol} to be entered in the output file as an undefined | |
3234 | symbol. Doing this may, for example, trigger linking of additional | |
3235 | modules from standard libraries. You may list several @var{symbol}s for | |
3236 | each @code{EXTERN}, and you may use @code{EXTERN} multiple times. This | |
3237 | command has the same effect as the @samp{-u} command-line option. | |
3238 | ||
3239 | @item FORCE_COMMON_ALLOCATION | |
3240 | @kindex FORCE_COMMON_ALLOCATION | |
3241 | @cindex common allocation in linker script | |
3242 | This command has the same effect as the @samp{-d} command-line option: | |
ff5dcc92 | 3243 | to make @command{ld} assign space to common symbols even if a relocatable |
252b5132 RH |
3244 | output file is specified (@samp{-r}). |
3245 | ||
4818e05f AM |
3246 | @item INHIBIT_COMMON_ALLOCATION |
3247 | @kindex INHIBIT_COMMON_ALLOCATION | |
3248 | @cindex common allocation in linker script | |
3249 | This command has the same effect as the @samp{--no-define-common} | |
3250 | command-line option: to make @code{ld} omit the assignment of addresses | |
3251 | to common symbols even for a non-relocatable output file. | |
3252 | ||
53d25da6 AM |
3253 | @item INSERT [ AFTER | BEFORE ] @var{output_section} |
3254 | @kindex INSERT | |
3255 | @cindex insert user script into default script | |
3256 | This command is typically used in a script specified by @samp{-T} to | |
3257 | augment the default @code{SECTIONS} with, for example, overlays. It | |
3258 | inserts all prior linker script statements after (or before) | |
3259 | @var{output_section}, and also causes @samp{-T} to not override the | |
3260 | default linker script. The exact insertion point is as for orphan | |
3261 | sections. @xref{Location Counter}. The insertion happens after the | |
3262 | linker has mapped input sections to output sections. Prior to the | |
3263 | insertion, since @samp{-T} scripts are parsed before the default | |
3264 | linker script, statements in the @samp{-T} script occur before the | |
3265 | default linker script statements in the internal linker representation | |
3266 | of the script. In particular, input section assignments will be made | |
3267 | to @samp{-T} output sections before those in the default script. Here | |
3268 | is an example of how a @samp{-T} script using @code{INSERT} might look: | |
3269 | ||
3270 | @smallexample | |
3271 | SECTIONS | |
3272 | @{ | |
3273 | OVERLAY : | |
3274 | @{ | |
3275 | .ov1 @{ ov1*(.text) @} | |
3276 | .ov2 @{ ov2*(.text) @} | |
3277 | @} | |
3278 | @} | |
3279 | INSERT AFTER .text; | |
3280 | @end smallexample | |
3281 | ||
252b5132 RH |
3282 | @item NOCROSSREFS(@var{section} @var{section} @dots{}) |
3283 | @kindex NOCROSSREFS(@var{sections}) | |
3284 | @cindex cross references | |
ff5dcc92 | 3285 | This command may be used to tell @command{ld} to issue an error about any |
252b5132 RH |
3286 | references among certain output sections. |
3287 | ||
3288 | In certain types of programs, particularly on embedded systems when | |
3289 | using overlays, when one section is loaded into memory, another section | |
3290 | will not be. Any direct references between the two sections would be | |
3291 | errors. For example, it would be an error if code in one section called | |
3292 | a function defined in the other section. | |
3293 | ||
3294 | The @code{NOCROSSREFS} command takes a list of output section names. If | |
ff5dcc92 | 3295 | @command{ld} detects any cross references between the sections, it reports |
252b5132 RH |
3296 | an error and returns a non-zero exit status. Note that the |
3297 | @code{NOCROSSREFS} command uses output section names, not input section | |
3298 | names. | |
3299 | ||
3300 | @ifclear SingleFormat | |
3301 | @item OUTPUT_ARCH(@var{bfdarch}) | |
3302 | @kindex OUTPUT_ARCH(@var{bfdarch}) | |
3303 | @cindex machine architecture | |
3304 | @cindex architecture | |
3305 | Specify a particular output machine architecture. The argument is one | |
3306 | of the names used by the BFD library (@pxref{BFD}). You can see the | |
3307 | architecture of an object file by using the @code{objdump} program with | |
3308 | the @samp{-f} option. | |
3309 | @end ifclear | |
3310 | @end table | |
3311 | ||
3312 | @node Assignments | |
3313 | @section Assigning Values to Symbols | |
3314 | @cindex assignment in scripts | |
3315 | @cindex symbol definition, scripts | |
3316 | @cindex variables, defining | |
3317 | You may assign a value to a symbol in a linker script. This will define | |
73ae6183 | 3318 | the symbol and place it into the symbol table with a global scope. |
252b5132 RH |
3319 | |
3320 | @menu | |
3321 | * Simple Assignments:: Simple Assignments | |
3322 | * PROVIDE:: PROVIDE | |
7af8e998 | 3323 | * PROVIDE_HIDDEN:: PROVIDE_HIDDEN |
73ae6183 | 3324 | * Source Code Reference:: How to use a linker script defined symbol in source code |
252b5132 RH |
3325 | @end menu |
3326 | ||
3327 | @node Simple Assignments | |
3328 | @subsection Simple Assignments | |
3329 | ||
3330 | You may assign to a symbol using any of the C assignment operators: | |
3331 | ||
3332 | @table @code | |
3333 | @item @var{symbol} = @var{expression} ; | |
3334 | @itemx @var{symbol} += @var{expression} ; | |
3335 | @itemx @var{symbol} -= @var{expression} ; | |
3336 | @itemx @var{symbol} *= @var{expression} ; | |
3337 | @itemx @var{symbol} /= @var{expression} ; | |
3338 | @itemx @var{symbol} <<= @var{expression} ; | |
3339 | @itemx @var{symbol} >>= @var{expression} ; | |
3340 | @itemx @var{symbol} &= @var{expression} ; | |
3341 | @itemx @var{symbol} |= @var{expression} ; | |
3342 | @end table | |
3343 | ||
3344 | The first case will define @var{symbol} to the value of | |
3345 | @var{expression}. In the other cases, @var{symbol} must already be | |
3346 | defined, and the value will be adjusted accordingly. | |
3347 | ||
3348 | The special symbol name @samp{.} indicates the location counter. You | |
b5666f2f | 3349 | may only use this within a @code{SECTIONS} command. @xref{Location Counter}. |
252b5132 RH |
3350 | |
3351 | The semicolon after @var{expression} is required. | |
3352 | ||
3353 | Expressions are defined below; see @ref{Expressions}. | |
3354 | ||
3355 | You may write symbol assignments as commands in their own right, or as | |
3356 | statements within a @code{SECTIONS} command, or as part of an output | |
3357 | section description in a @code{SECTIONS} command. | |
3358 | ||
3359 | The section of the symbol will be set from the section of the | |
3360 | expression; for more information, see @ref{Expression Section}. | |
3361 | ||
3362 | Here is an example showing the three different places that symbol | |
3363 | assignments may be used: | |
3364 | ||
3365 | @smallexample | |
3366 | floating_point = 0; | |
3367 | SECTIONS | |
3368 | @{ | |
3369 | .text : | |
3370 | @{ | |
3371 | *(.text) | |
3372 | _etext = .; | |
3373 | @} | |
156e34dd | 3374 | _bdata = (. + 3) & ~ 3; |
252b5132 RH |
3375 | .data : @{ *(.data) @} |
3376 | @} | |
3377 | @end smallexample | |
3378 | @noindent | |
3379 | In this example, the symbol @samp{floating_point} will be defined as | |
3380 | zero. The symbol @samp{_etext} will be defined as the address following | |
3381 | the last @samp{.text} input section. The symbol @samp{_bdata} will be | |
3382 | defined as the address following the @samp{.text} output section aligned | |
3383 | upward to a 4 byte boundary. | |
3384 | ||
3385 | @node PROVIDE | |
3386 | @subsection PROVIDE | |
3387 | @cindex PROVIDE | |
3388 | In some cases, it is desirable for a linker script to define a symbol | |
3389 | only if it is referenced and is not defined by any object included in | |
3390 | the link. For example, traditional linkers defined the symbol | |
3391 | @samp{etext}. However, ANSI C requires that the user be able to use | |
3392 | @samp{etext} as a function name without encountering an error. The | |
3393 | @code{PROVIDE} keyword may be used to define a symbol, such as | |
3394 | @samp{etext}, only if it is referenced but not defined. The syntax is | |
3395 | @code{PROVIDE(@var{symbol} = @var{expression})}. | |
3396 | ||
3397 | Here is an example of using @code{PROVIDE} to define @samp{etext}: | |
3398 | @smallexample | |
3399 | SECTIONS | |
3400 | @{ | |
3401 | .text : | |
3402 | @{ | |
3403 | *(.text) | |
3404 | _etext = .; | |
3405 | PROVIDE(etext = .); | |
3406 | @} | |
3407 | @} | |
3408 | @end smallexample | |
3409 | ||
3410 | In this example, if the program defines @samp{_etext} (with a leading | |
3411 | underscore), the linker will give a multiple definition error. If, on | |
3412 | the other hand, the program defines @samp{etext} (with no leading | |
3413 | underscore), the linker will silently use the definition in the program. | |
3414 | If the program references @samp{etext} but does not define it, the | |
3415 | linker will use the definition in the linker script. | |
3416 | ||
7af8e998 L |
3417 | @node PROVIDE_HIDDEN |
3418 | @subsection PROVIDE_HIDDEN | |
3419 | @cindex PROVIDE_HIDDEN | |
3420 | Similar to @code{PROVIDE}. For ELF targeted ports, the symbol will be | |
3421 | hidden and won't be exported. | |
3422 | ||
73ae6183 NC |
3423 | @node Source Code Reference |
3424 | @subsection Source Code Reference | |
3425 | ||
3426 | Accessing a linker script defined variable from source code is not | |
3427 | intuitive. In particular a linker script symbol is not equivalent to | |
3428 | a variable declaration in a high level language, it is instead a | |
3429 | symbol that does not have a value. | |
3430 | ||
3431 | Before going further, it is important to note that compilers often | |
3432 | transform names in the source code into different names when they are | |
3433 | stored in the symbol table. For example, Fortran compilers commonly | |
3434 | prepend or append an underscore, and C++ performs extensive @samp{name | |
3435 | mangling}. Therefore there might be a discrepancy between the name | |
3436 | of a variable as it is used in source code and the name of the same | |
3437 | variable as it is defined in a linker script. For example in C a | |
3438 | linker script variable might be referred to as: | |
3439 | ||
3440 | @smallexample | |
3441 | extern int foo; | |
3442 | @end smallexample | |
3443 | ||
3444 | But in the linker script it might be defined as: | |
3445 | ||
3446 | @smallexample | |
3447 | _foo = 1000; | |
3448 | @end smallexample | |
3449 | ||
3450 | In the remaining examples however it is assumed that no name | |
3451 | transformation has taken place. | |
3452 | ||
3453 | When a symbol is declared in a high level language such as C, two | |
3454 | things happen. The first is that the compiler reserves enough space | |
3455 | in the program's memory to hold the @emph{value} of the symbol. The | |
3456 | second is that the compiler creates an entry in the program's symbol | |
3457 | table which holds the symbol's @emph{address}. ie the symbol table | |
3458 | contains the address of the block of memory holding the symbol's | |
3459 | value. So for example the following C declaration, at file scope: | |
3460 | ||
3461 | @smallexample | |
3462 | int foo = 1000; | |
3463 | @end smallexample | |
3464 | ||
3465 | creates a entry called @samp{foo} in the symbol table. This entry | |
3466 | holds the address of an @samp{int} sized block of memory where the | |
3467 | number 1000 is initially stored. | |
3468 | ||
3469 | When a program references a symbol the compiler generates code that | |
3470 | first accesses the symbol table to find the address of the symbol's | |
3471 | memory block and then code to read the value from that memory block. | |
3472 | So: | |
3473 | ||
3474 | @smallexample | |
3475 | foo = 1; | |
3476 | @end smallexample | |
3477 | ||
3478 | looks up the symbol @samp{foo} in the symbol table, gets the address | |
3479 | associated with this symbol and then writes the value 1 into that | |
3480 | address. Whereas: | |
3481 | ||
3482 | @smallexample | |
3483 | int * a = & foo; | |
3484 | @end smallexample | |
3485 | ||
3486 | looks up the symbol @samp{foo} in the symbol table, gets it address | |
3487 | and then copies this address into the block of memory associated with | |
3488 | the variable @samp{a}. | |
3489 | ||
3490 | Linker scripts symbol declarations, by contrast, create an entry in | |
3491 | the symbol table but do not assign any memory to them. Thus they are | |
3492 | an address without a value. So for example the linker script definition: | |
3493 | ||
3494 | @smallexample | |
3495 | foo = 1000; | |
3496 | @end smallexample | |
3497 | ||
3498 | creates an entry in the symbol table called @samp{foo} which holds | |
3499 | the address of memory location 1000, but nothing special is stored at | |
3500 | address 1000. This means that you cannot access the @emph{value} of a | |
3501 | linker script defined symbol - it has no value - all you can do is | |
3502 | access the @emph{address} of a linker script defined symbol. | |
3503 | ||
3504 | Hence when you are using a linker script defined symbol in source code | |
3505 | you should always take the address of the symbol, and never attempt to | |
3506 | use its value. For example suppose you want to copy the contents of a | |
3507 | section of memory called .ROM into a section called .FLASH and the | |
3508 | linker script contains these declarations: | |
3509 | ||
3510 | @smallexample | |
3511 | @group | |
3512 | start_of_ROM = .ROM; | |
3513 | end_of_ROM = .ROM + sizeof (.ROM) - 1; | |
3514 | start_of_FLASH = .FLASH; | |
3515 | @end group | |
3516 | @end smallexample | |
3517 | ||
3518 | Then the C source code to perform the copy would be: | |
3519 | ||
3520 | @smallexample | |
3521 | @group | |
3522 | extern char start_of_ROM, end_of_ROM, start_of_FLASH; | |
c0065db7 | 3523 | |
73ae6183 NC |
3524 | memcpy (& start_of_FLASH, & start_of_ROM, & end_of_ROM - & start_of_ROM); |
3525 | @end group | |
3526 | @end smallexample | |
3527 | ||
3528 | Note the use of the @samp{&} operators. These are correct. | |
3529 | ||
252b5132 | 3530 | @node SECTIONS |
36f63dca | 3531 | @section SECTIONS Command |
252b5132 RH |
3532 | @kindex SECTIONS |
3533 | The @code{SECTIONS} command tells the linker how to map input sections | |
3534 | into output sections, and how to place the output sections in memory. | |
3535 | ||
3536 | The format of the @code{SECTIONS} command is: | |
3537 | @smallexample | |
3538 | SECTIONS | |
3539 | @{ | |
3540 | @var{sections-command} | |
3541 | @var{sections-command} | |
3542 | @dots{} | |
3543 | @} | |
3544 | @end smallexample | |
3545 | ||
3546 | Each @var{sections-command} may of be one of the following: | |
3547 | ||
3548 | @itemize @bullet | |
3549 | @item | |
3550 | an @code{ENTRY} command (@pxref{Entry Point,,Entry command}) | |
3551 | @item | |
3552 | a symbol assignment (@pxref{Assignments}) | |
3553 | @item | |
3554 | an output section description | |
3555 | @item | |
3556 | an overlay description | |
3557 | @end itemize | |
3558 | ||
3559 | The @code{ENTRY} command and symbol assignments are permitted inside the | |
3560 | @code{SECTIONS} command for convenience in using the location counter in | |
3561 | those commands. This can also make the linker script easier to | |
3562 | understand because you can use those commands at meaningful points in | |
3563 | the layout of the output file. | |
3564 | ||
3565 | Output section descriptions and overlay descriptions are described | |
3566 | below. | |
3567 | ||
3568 | If you do not use a @code{SECTIONS} command in your linker script, the | |
3569 | linker will place each input section into an identically named output | |
3570 | section in the order that the sections are first encountered in the | |
3571 | input files. If all input sections are present in the first file, for | |
3572 | example, the order of sections in the output file will match the order | |
3573 | in the first input file. The first section will be at address zero. | |
3574 | ||
3575 | @menu | |
3576 | * Output Section Description:: Output section description | |
3577 | * Output Section Name:: Output section name | |
3578 | * Output Section Address:: Output section address | |
3579 | * Input Section:: Input section description | |
3580 | * Output Section Data:: Output section data | |
3581 | * Output Section Keywords:: Output section keywords | |
3582 | * Output Section Discarding:: Output section discarding | |
3583 | * Output Section Attributes:: Output section attributes | |
3584 | * Overlay Description:: Overlay description | |
3585 | @end menu | |
3586 | ||
3587 | @node Output Section Description | |
36f63dca | 3588 | @subsection Output Section Description |
252b5132 RH |
3589 | The full description of an output section looks like this: |
3590 | @smallexample | |
a1ab1d2a | 3591 | @group |
7e7d5768 | 3592 | @var{section} [@var{address}] [(@var{type})] : |
0c71d759 NC |
3593 | [AT(@var{lma})] |
3594 | [ALIGN(@var{section_align})] | |
3595 | [SUBALIGN(@var{subsection_align})] | |
3596 | [@var{constraint}] | |
252b5132 RH |
3597 | @{ |
3598 | @var{output-section-command} | |
3599 | @var{output-section-command} | |
3600 | @dots{} | |
562d3460 | 3601 | @} [>@var{region}] [AT>@var{lma_region}] [:@var{phdr} :@var{phdr} @dots{}] [=@var{fillexp}] |
252b5132 RH |
3602 | @end group |
3603 | @end smallexample | |
3604 | ||
3605 | Most output sections do not use most of the optional section attributes. | |
3606 | ||
3607 | The whitespace around @var{section} is required, so that the section | |
3608 | name is unambiguous. The colon and the curly braces are also required. | |
3609 | The line breaks and other white space are optional. | |
3610 | ||
3611 | Each @var{output-section-command} may be one of the following: | |
3612 | ||
3613 | @itemize @bullet | |
3614 | @item | |
3615 | a symbol assignment (@pxref{Assignments}) | |
3616 | @item | |
3617 | an input section description (@pxref{Input Section}) | |
3618 | @item | |
3619 | data values to include directly (@pxref{Output Section Data}) | |
3620 | @item | |
3621 | a special output section keyword (@pxref{Output Section Keywords}) | |
3622 | @end itemize | |
3623 | ||
3624 | @node Output Section Name | |
36f63dca | 3625 | @subsection Output Section Name |
252b5132 RH |
3626 | @cindex name, section |
3627 | @cindex section name | |
3628 | The name of the output section is @var{section}. @var{section} must | |
3629 | meet the constraints of your output format. In formats which only | |
3630 | support a limited number of sections, such as @code{a.out}, the name | |
3631 | must be one of the names supported by the format (@code{a.out}, for | |
3632 | example, allows only @samp{.text}, @samp{.data} or @samp{.bss}). If the | |
3633 | output format supports any number of sections, but with numbers and not | |
3634 | names (as is the case for Oasys), the name should be supplied as a | |
3635 | quoted numeric string. A section name may consist of any sequence of | |
3636 | characters, but a name which contains any unusual characters such as | |
3637 | commas must be quoted. | |
3638 | ||
3639 | The output section name @samp{/DISCARD/} is special; @ref{Output Section | |
3640 | Discarding}. | |
3641 | ||
3642 | @node Output Section Address | |
2a16d82a | 3643 | @subsection Output Section Address |
252b5132 RH |
3644 | @cindex address, section |
3645 | @cindex section address | |
3646 | The @var{address} is an expression for the VMA (the virtual memory | |
3647 | address) of the output section. If you do not provide @var{address}, | |
3648 | the linker will set it based on @var{region} if present, or otherwise | |
3649 | based on the current value of the location counter. | |
3650 | ||
3651 | If you provide @var{address}, the address of the output section will be | |
3652 | set to precisely that. If you provide neither @var{address} nor | |
3653 | @var{region}, then the address of the output section will be set to the | |
3654 | current value of the location counter aligned to the alignment | |
3655 | requirements of the output section. The alignment requirement of the | |
3656 | output section is the strictest alignment of any input section contained | |
3657 | within the output section. | |
3658 | ||
3659 | For example, | |
3660 | @smallexample | |
3661 | .text . : @{ *(.text) @} | |
3662 | @end smallexample | |
3663 | @noindent | |
3664 | and | |
3665 | @smallexample | |
3666 | .text : @{ *(.text) @} | |
3667 | @end smallexample | |
3668 | @noindent | |
3669 | are subtly different. The first will set the address of the | |
3670 | @samp{.text} output section to the current value of the location | |
3671 | counter. The second will set it to the current value of the location | |
3672 | counter aligned to the strictest alignment of a @samp{.text} input | |
3673 | section. | |
3674 | ||
3675 | The @var{address} may be an arbitrary expression; @ref{Expressions}. | |
3676 | For example, if you want to align the section on a 0x10 byte boundary, | |
3677 | so that the lowest four bits of the section address are zero, you could | |
3678 | do something like this: | |
3679 | @smallexample | |
3680 | .text ALIGN(0x10) : @{ *(.text) @} | |
3681 | @end smallexample | |
3682 | @noindent | |
3683 | This works because @code{ALIGN} returns the current location counter | |
3684 | aligned upward to the specified value. | |
3685 | ||
3686 | Specifying @var{address} for a section will change the value of the | |
6ce340f1 NC |
3687 | location counter, provided that the section is non-empty. (Empty |
3688 | sections are ignored). | |
252b5132 RH |
3689 | |
3690 | @node Input Section | |
36f63dca | 3691 | @subsection Input Section Description |
252b5132 RH |
3692 | @cindex input sections |
3693 | @cindex mapping input sections to output sections | |
3694 | The most common output section command is an input section description. | |
3695 | ||
3696 | The input section description is the most basic linker script operation. | |
3697 | You use output sections to tell the linker how to lay out your program | |
3698 | in memory. You use input section descriptions to tell the linker how to | |
3699 | map the input files into your memory layout. | |
3700 | ||
3701 | @menu | |
3702 | * Input Section Basics:: Input section basics | |
3703 | * Input Section Wildcards:: Input section wildcard patterns | |
3704 | * Input Section Common:: Input section for common symbols | |
3705 | * Input Section Keep:: Input section and garbage collection | |
3706 | * Input Section Example:: Input section example | |
3707 | @end menu | |
3708 | ||
3709 | @node Input Section Basics | |
36f63dca | 3710 | @subsubsection Input Section Basics |
252b5132 RH |
3711 | @cindex input section basics |
3712 | An input section description consists of a file name optionally followed | |
3713 | by a list of section names in parentheses. | |
3714 | ||
3715 | The file name and the section name may be wildcard patterns, which we | |
3716 | describe further below (@pxref{Input Section Wildcards}). | |
3717 | ||
3718 | The most common input section description is to include all input | |
3719 | sections with a particular name in the output section. For example, to | |
3720 | include all input @samp{.text} sections, you would write: | |
3721 | @smallexample | |
3722 | *(.text) | |
3723 | @end smallexample | |
3724 | @noindent | |
18625d54 CM |
3725 | Here the @samp{*} is a wildcard which matches any file name. To exclude a list |
3726 | of files from matching the file name wildcard, EXCLUDE_FILE may be used to | |
3727 | match all files except the ones specified in the EXCLUDE_FILE list. For | |
3728 | example: | |
252b5132 | 3729 | @smallexample |
b4346c09 | 3730 | *(EXCLUDE_FILE (*crtend.o *otherfile.o) .ctors) |
252b5132 | 3731 | @end smallexample |
765b7cbe JB |
3732 | will cause all .ctors sections from all files except @file{crtend.o} and |
3733 | @file{otherfile.o} to be included. | |
252b5132 RH |
3734 | |
3735 | There are two ways to include more than one section: | |
3736 | @smallexample | |
3737 | *(.text .rdata) | |
3738 | *(.text) *(.rdata) | |
3739 | @end smallexample | |
3740 | @noindent | |
3741 | The difference between these is the order in which the @samp{.text} and | |
3742 | @samp{.rdata} input sections will appear in the output section. In the | |
b6bf44ba AM |
3743 | first example, they will be intermingled, appearing in the same order as |
3744 | they are found in the linker input. In the second example, all | |
252b5132 RH |
3745 | @samp{.text} input sections will appear first, followed by all |
3746 | @samp{.rdata} input sections. | |
3747 | ||
3748 | You can specify a file name to include sections from a particular file. | |
3749 | You would do this if one or more of your files contain special data that | |
3750 | needs to be at a particular location in memory. For example: | |
3751 | @smallexample | |
3752 | data.o(.data) | |
3753 | @end smallexample | |
3754 | ||
967928e9 AM |
3755 | You can also specify files within archives by writing a pattern |
3756 | matching the archive, a colon, then the pattern matching the file, | |
3757 | with no whitespace around the colon. | |
3758 | ||
3759 | @table @samp | |
3760 | @item archive:file | |
3761 | matches file within archive | |
3762 | @item archive: | |
3763 | matches the whole archive | |
3764 | @item :file | |
3765 | matches file but not one in an archive | |
3766 | @end table | |
3767 | ||
3768 | Either one or both of @samp{archive} and @samp{file} can contain shell | |
3769 | wildcards. On DOS based file systems, the linker will assume that a | |
3770 | single letter followed by a colon is a drive specifier, so | |
3771 | @samp{c:myfile.o} is a simple file specification, not @samp{myfile.o} | |
3772 | within an archive called @samp{c}. @samp{archive:file} filespecs may | |
3773 | also be used within an @code{EXCLUDE_FILE} list, but may not appear in | |
3774 | other linker script contexts. For instance, you cannot extract a file | |
3775 | from an archive by using @samp{archive:file} in an @code{INPUT} | |
3776 | command. | |
3777 | ||
252b5132 RH |
3778 | If you use a file name without a list of sections, then all sections in |
3779 | the input file will be included in the output section. This is not | |
3780 | commonly done, but it may by useful on occasion. For example: | |
3781 | @smallexample | |
3782 | data.o | |
3783 | @end smallexample | |
3784 | ||
967928e9 AM |
3785 | When you use a file name which is not an @samp{archive:file} specifier |
3786 | and does not contain any wild card | |
252b5132 RH |
3787 | characters, the linker will first see if you also specified the file |
3788 | name on the linker command line or in an @code{INPUT} command. If you | |
3789 | did not, the linker will attempt to open the file as an input file, as | |
3790 | though it appeared on the command line. Note that this differs from an | |
3791 | @code{INPUT} command, because the linker will not search for the file in | |
3792 | the archive search path. | |
3793 | ||
3794 | @node Input Section Wildcards | |
36f63dca | 3795 | @subsubsection Input Section Wildcard Patterns |
252b5132 RH |
3796 | @cindex input section wildcards |
3797 | @cindex wildcard file name patterns | |
3798 | @cindex file name wildcard patterns | |
3799 | @cindex section name wildcard patterns | |
3800 | In an input section description, either the file name or the section | |
3801 | name or both may be wildcard patterns. | |
3802 | ||
3803 | The file name of @samp{*} seen in many examples is a simple wildcard | |
3804 | pattern for the file name. | |
3805 | ||
3806 | The wildcard patterns are like those used by the Unix shell. | |
3807 | ||
3808 | @table @samp | |
3809 | @item * | |
3810 | matches any number of characters | |
3811 | @item ? | |
3812 | matches any single character | |
3813 | @item [@var{chars}] | |
3814 | matches a single instance of any of the @var{chars}; the @samp{-} | |
3815 | character may be used to specify a range of characters, as in | |
3816 | @samp{[a-z]} to match any lower case letter | |
3817 | @item \ | |
3818 | quotes the following character | |
3819 | @end table | |
3820 | ||
3821 | When a file name is matched with a wildcard, the wildcard characters | |
3822 | will not match a @samp{/} character (used to separate directory names on | |
3823 | Unix). A pattern consisting of a single @samp{*} character is an | |
3824 | exception; it will always match any file name, whether it contains a | |
3825 | @samp{/} or not. In a section name, the wildcard characters will match | |
3826 | a @samp{/} character. | |
3827 | ||
3828 | File name wildcard patterns only match files which are explicitly | |
3829 | specified on the command line or in an @code{INPUT} command. The linker | |
3830 | does not search directories to expand wildcards. | |
3831 | ||
3832 | If a file name matches more than one wildcard pattern, or if a file name | |
3833 | appears explicitly and is also matched by a wildcard pattern, the linker | |
3834 | will use the first match in the linker script. For example, this | |
3835 | sequence of input section descriptions is probably in error, because the | |
3836 | @file{data.o} rule will not be used: | |
3837 | @smallexample | |
3838 | .data : @{ *(.data) @} | |
3839 | .data1 : @{ data.o(.data) @} | |
3840 | @end smallexample | |
3841 | ||
bcaa7b3e | 3842 | @cindex SORT_BY_NAME |
252b5132 RH |
3843 | Normally, the linker will place files and sections matched by wildcards |
3844 | in the order in which they are seen during the link. You can change | |
bcaa7b3e L |
3845 | this by using the @code{SORT_BY_NAME} keyword, which appears before a wildcard |
3846 | pattern in parentheses (e.g., @code{SORT_BY_NAME(.text*)}). When the | |
3847 | @code{SORT_BY_NAME} keyword is used, the linker will sort the files or sections | |
252b5132 RH |
3848 | into ascending order by name before placing them in the output file. |
3849 | ||
bcaa7b3e L |
3850 | @cindex SORT_BY_ALIGNMENT |
3851 | @code{SORT_BY_ALIGNMENT} is very similar to @code{SORT_BY_NAME}. The | |
3852 | difference is @code{SORT_BY_ALIGNMENT} will sort sections into | |
3853 | ascending order by alignment before placing them in the output file. | |
3854 | ||
3855 | @cindex SORT | |
3856 | @code{SORT} is an alias for @code{SORT_BY_NAME}. | |
3857 | ||
3858 | When there are nested section sorting commands in linker script, there | |
3859 | can be at most 1 level of nesting for section sorting commands. | |
3860 | ||
3861 | @enumerate | |
3862 | @item | |
3863 | @code{SORT_BY_NAME} (@code{SORT_BY_ALIGNMENT} (wildcard section pattern)). | |
3864 | It will sort the input sections by name first, then by alignment if 2 | |
3865 | sections have the same name. | |
3866 | @item | |
3867 | @code{SORT_BY_ALIGNMENT} (@code{SORT_BY_NAME} (wildcard section pattern)). | |
3868 | It will sort the input sections by alignment first, then by name if 2 | |
3869 | sections have the same alignment. | |
3870 | @item | |
c0065db7 | 3871 | @code{SORT_BY_NAME} (@code{SORT_BY_NAME} (wildcard section pattern)) is |
bcaa7b3e L |
3872 | treated the same as @code{SORT_BY_NAME} (wildcard section pattern). |
3873 | @item | |
3874 | @code{SORT_BY_ALIGNMENT} (@code{SORT_BY_ALIGNMENT} (wildcard section pattern)) | |
3875 | is treated the same as @code{SORT_BY_ALIGNMENT} (wildcard section pattern). | |
3876 | @item | |
3877 | All other nested section sorting commands are invalid. | |
3878 | @end enumerate | |
3879 | ||
3880 | When both command line section sorting option and linker script | |
3881 | section sorting command are used, section sorting command always | |
3882 | takes precedence over the command line option. | |
3883 | ||
3884 | If the section sorting command in linker script isn't nested, the | |
3885 | command line option will make the section sorting command to be | |
3886 | treated as nested sorting command. | |
3887 | ||
3888 | @enumerate | |
3889 | @item | |
3890 | @code{SORT_BY_NAME} (wildcard section pattern ) with | |
3891 | @option{--sort-sections alignment} is equivalent to | |
3892 | @code{SORT_BY_NAME} (@code{SORT_BY_ALIGNMENT} (wildcard section pattern)). | |
3893 | @item | |
3894 | @code{SORT_BY_ALIGNMENT} (wildcard section pattern) with | |
3895 | @option{--sort-section name} is equivalent to | |
3896 | @code{SORT_BY_ALIGNMENT} (@code{SORT_BY_NAME} (wildcard section pattern)). | |
3897 | @end enumerate | |
3898 | ||
3899 | If the section sorting command in linker script is nested, the | |
3900 | command line option will be ignored. | |
3901 | ||
252b5132 RH |
3902 | If you ever get confused about where input sections are going, use the |
3903 | @samp{-M} linker option to generate a map file. The map file shows | |
3904 | precisely how input sections are mapped to output sections. | |
3905 | ||
3906 | This example shows how wildcard patterns might be used to partition | |
3907 | files. This linker script directs the linker to place all @samp{.text} | |
3908 | sections in @samp{.text} and all @samp{.bss} sections in @samp{.bss}. | |
3909 | The linker will place the @samp{.data} section from all files beginning | |
3910 | with an upper case character in @samp{.DATA}; for all other files, the | |
3911 | linker will place the @samp{.data} section in @samp{.data}. | |
3912 | @smallexample | |
3913 | @group | |
3914 | SECTIONS @{ | |
3915 | .text : @{ *(.text) @} | |
3916 | .DATA : @{ [A-Z]*(.data) @} | |
3917 | .data : @{ *(.data) @} | |
3918 | .bss : @{ *(.bss) @} | |
3919 | @} | |
3920 | @end group | |
3921 | @end smallexample | |
3922 | ||
3923 | @node Input Section Common | |
36f63dca | 3924 | @subsubsection Input Section for Common Symbols |
252b5132 RH |
3925 | @cindex common symbol placement |
3926 | @cindex uninitialized data placement | |
3927 | A special notation is needed for common symbols, because in many object | |
3928 | file formats common symbols do not have a particular input section. The | |
3929 | linker treats common symbols as though they are in an input section | |
3930 | named @samp{COMMON}. | |
3931 | ||
3932 | You may use file names with the @samp{COMMON} section just as with any | |
3933 | other input sections. You can use this to place common symbols from a | |
3934 | particular input file in one section while common symbols from other | |
3935 | input files are placed in another section. | |
3936 | ||
3937 | In most cases, common symbols in input files will be placed in the | |
3938 | @samp{.bss} section in the output file. For example: | |
3939 | @smallexample | |
3940 | .bss @{ *(.bss) *(COMMON) @} | |
3941 | @end smallexample | |
3942 | ||
3943 | @cindex scommon section | |
3944 | @cindex small common symbols | |
3945 | Some object file formats have more than one type of common symbol. For | |
3946 | example, the MIPS ELF object file format distinguishes standard common | |
3947 | symbols and small common symbols. In this case, the linker will use a | |
3948 | different special section name for other types of common symbols. In | |
3949 | the case of MIPS ELF, the linker uses @samp{COMMON} for standard common | |
3950 | symbols and @samp{.scommon} for small common symbols. This permits you | |
3951 | to map the different types of common symbols into memory at different | |
3952 | locations. | |
3953 | ||
3954 | @cindex [COMMON] | |
3955 | You will sometimes see @samp{[COMMON]} in old linker scripts. This | |
3956 | notation is now considered obsolete. It is equivalent to | |
3957 | @samp{*(COMMON)}. | |
3958 | ||
3959 | @node Input Section Keep | |
36f63dca | 3960 | @subsubsection Input Section and Garbage Collection |
252b5132 RH |
3961 | @cindex KEEP |
3962 | @cindex garbage collection | |
3963 | When link-time garbage collection is in use (@samp{--gc-sections}), | |
a1ab1d2a | 3964 | it is often useful to mark sections that should not be eliminated. |
252b5132 RH |
3965 | This is accomplished by surrounding an input section's wildcard entry |
3966 | with @code{KEEP()}, as in @code{KEEP(*(.init))} or | |
bcaa7b3e | 3967 | @code{KEEP(SORT_BY_NAME(*)(.ctors))}. |
252b5132 RH |
3968 | |
3969 | @node Input Section Example | |
36f63dca | 3970 | @subsubsection Input Section Example |
252b5132 RH |
3971 | The following example is a complete linker script. It tells the linker |
3972 | to read all of the sections from file @file{all.o} and place them at the | |
3973 | start of output section @samp{outputa} which starts at location | |
3974 | @samp{0x10000}. All of section @samp{.input1} from file @file{foo.o} | |
3975 | follows immediately, in the same output section. All of section | |
3976 | @samp{.input2} from @file{foo.o} goes into output section | |
3977 | @samp{outputb}, followed by section @samp{.input1} from @file{foo1.o}. | |
3978 | All of the remaining @samp{.input1} and @samp{.input2} sections from any | |
3979 | files are written to output section @samp{outputc}. | |
3980 | ||
3981 | @smallexample | |
3982 | @group | |
3983 | SECTIONS @{ | |
3984 | outputa 0x10000 : | |
3985 | @{ | |
3986 | all.o | |
3987 | foo.o (.input1) | |
3988 | @} | |
36f63dca NC |
3989 | @end group |
3990 | @group | |
252b5132 RH |
3991 | outputb : |
3992 | @{ | |
3993 | foo.o (.input2) | |
3994 | foo1.o (.input1) | |
3995 | @} | |
36f63dca NC |
3996 | @end group |
3997 | @group | |
252b5132 RH |
3998 | outputc : |
3999 | @{ | |
4000 | *(.input1) | |
4001 | *(.input2) | |
4002 | @} | |
4003 | @} | |
4004 | @end group | |
a1ab1d2a | 4005 | @end smallexample |
252b5132 RH |
4006 | |
4007 | @node Output Section Data | |
36f63dca | 4008 | @subsection Output Section Data |
252b5132 RH |
4009 | @cindex data |
4010 | @cindex section data | |
4011 | @cindex output section data | |
4012 | @kindex BYTE(@var{expression}) | |
4013 | @kindex SHORT(@var{expression}) | |
4014 | @kindex LONG(@var{expression}) | |
4015 | @kindex QUAD(@var{expression}) | |
4016 | @kindex SQUAD(@var{expression}) | |
4017 | You can include explicit bytes of data in an output section by using | |
4018 | @code{BYTE}, @code{SHORT}, @code{LONG}, @code{QUAD}, or @code{SQUAD} as | |
4019 | an output section command. Each keyword is followed by an expression in | |
4020 | parentheses providing the value to store (@pxref{Expressions}). The | |
4021 | value of the expression is stored at the current value of the location | |
4022 | counter. | |
4023 | ||
4024 | The @code{BYTE}, @code{SHORT}, @code{LONG}, and @code{QUAD} commands | |
4025 | store one, two, four, and eight bytes (respectively). After storing the | |
4026 | bytes, the location counter is incremented by the number of bytes | |
4027 | stored. | |
4028 | ||
4029 | For example, this will store the byte 1 followed by the four byte value | |
4030 | of the symbol @samp{addr}: | |
4031 | @smallexample | |
4032 | BYTE(1) | |
4033 | LONG(addr) | |
4034 | @end smallexample | |
4035 | ||
4036 | When using a 64 bit host or target, @code{QUAD} and @code{SQUAD} are the | |
4037 | same; they both store an 8 byte, or 64 bit, value. When both host and | |
4038 | target are 32 bits, an expression is computed as 32 bits. In this case | |
4039 | @code{QUAD} stores a 32 bit value zero extended to 64 bits, and | |
4040 | @code{SQUAD} stores a 32 bit value sign extended to 64 bits. | |
4041 | ||
4042 | If the object file format of the output file has an explicit endianness, | |
4043 | which is the normal case, the value will be stored in that endianness. | |
4044 | When the object file format does not have an explicit endianness, as is | |
4045 | true of, for example, S-records, the value will be stored in the | |
4046 | endianness of the first input object file. | |
4047 | ||
36f63dca | 4048 | Note---these commands only work inside a section description and not |
2b5fc1f5 NC |
4049 | between them, so the following will produce an error from the linker: |
4050 | @smallexample | |
4051 | SECTIONS @{@ .text : @{@ *(.text) @}@ LONG(1) .data : @{@ *(.data) @}@ @}@ | |
4052 | @end smallexample | |
4053 | whereas this will work: | |
4054 | @smallexample | |
4055 | SECTIONS @{@ .text : @{@ *(.text) ; LONG(1) @}@ .data : @{@ *(.data) @}@ @}@ | |
4056 | @end smallexample | |
4057 | ||
252b5132 RH |
4058 | @kindex FILL(@var{expression}) |
4059 | @cindex holes, filling | |
4060 | @cindex unspecified memory | |
4061 | You may use the @code{FILL} command to set the fill pattern for the | |
4062 | current section. It is followed by an expression in parentheses. Any | |
4063 | otherwise unspecified regions of memory within the section (for example, | |
4064 | gaps left due to the required alignment of input sections) are filled | |
a139d329 | 4065 | with the value of the expression, repeated as |
252b5132 RH |
4066 | necessary. A @code{FILL} statement covers memory locations after the |
4067 | point at which it occurs in the section definition; by including more | |
4068 | than one @code{FILL} statement, you can have different fill patterns in | |
4069 | different parts of an output section. | |
4070 | ||
4071 | This example shows how to fill unspecified regions of memory with the | |
563e308f | 4072 | value @samp{0x90}: |
252b5132 | 4073 | @smallexample |
563e308f | 4074 | FILL(0x90909090) |
252b5132 RH |
4075 | @end smallexample |
4076 | ||
4077 | The @code{FILL} command is similar to the @samp{=@var{fillexp}} output | |
9673c93c | 4078 | section attribute, but it only affects the |
252b5132 RH |
4079 | part of the section following the @code{FILL} command, rather than the |
4080 | entire section. If both are used, the @code{FILL} command takes | |
9673c93c | 4081 | precedence. @xref{Output Section Fill}, for details on the fill |
a139d329 | 4082 | expression. |
252b5132 RH |
4083 | |
4084 | @node Output Section Keywords | |
36f63dca | 4085 | @subsection Output Section Keywords |
252b5132 RH |
4086 | There are a couple of keywords which can appear as output section |
4087 | commands. | |
4088 | ||
4089 | @table @code | |
4090 | @kindex CREATE_OBJECT_SYMBOLS | |
4091 | @cindex input filename symbols | |
4092 | @cindex filename symbols | |
4093 | @item CREATE_OBJECT_SYMBOLS | |
4094 | The command tells the linker to create a symbol for each input file. | |
4095 | The name of each symbol will be the name of the corresponding input | |
4096 | file. The section of each symbol will be the output section in which | |
4097 | the @code{CREATE_OBJECT_SYMBOLS} command appears. | |
4098 | ||
4099 | This is conventional for the a.out object file format. It is not | |
4100 | normally used for any other object file format. | |
4101 | ||
4102 | @kindex CONSTRUCTORS | |
4103 | @cindex C++ constructors, arranging in link | |
4104 | @cindex constructors, arranging in link | |
4105 | @item CONSTRUCTORS | |
4106 | When linking using the a.out object file format, the linker uses an | |
4107 | unusual set construct to support C++ global constructors and | |
4108 | destructors. When linking object file formats which do not support | |
4109 | arbitrary sections, such as ECOFF and XCOFF, the linker will | |
4110 | automatically recognize C++ global constructors and destructors by name. | |
4111 | For these object file formats, the @code{CONSTRUCTORS} command tells the | |
4112 | linker to place constructor information in the output section where the | |
4113 | @code{CONSTRUCTORS} command appears. The @code{CONSTRUCTORS} command is | |
4114 | ignored for other object file formats. | |
4115 | ||
4116 | The symbol @w{@code{__CTOR_LIST__}} marks the start of the global | |
7e69709c AM |
4117 | constructors, and the symbol @w{@code{__CTOR_END__}} marks the end. |
4118 | Similarly, @w{@code{__DTOR_LIST__}} and @w{@code{__DTOR_END__}} mark | |
4119 | the start and end of the global destructors. The | |
252b5132 RH |
4120 | first word in the list is the number of entries, followed by the address |
4121 | of each constructor or destructor, followed by a zero word. The | |
4122 | compiler must arrange to actually run the code. For these object file | |
4123 | formats @sc{gnu} C++ normally calls constructors from a subroutine | |
4124 | @code{__main}; a call to @code{__main} is automatically inserted into | |
4125 | the startup code for @code{main}. @sc{gnu} C++ normally runs | |
4126 | destructors either by using @code{atexit}, or directly from the function | |
4127 | @code{exit}. | |
4128 | ||
4129 | For object file formats such as @code{COFF} or @code{ELF} which support | |
4130 | arbitrary section names, @sc{gnu} C++ will normally arrange to put the | |
4131 | addresses of global constructors and destructors into the @code{.ctors} | |
4132 | and @code{.dtors} sections. Placing the following sequence into your | |
4133 | linker script will build the sort of table which the @sc{gnu} C++ | |
4134 | runtime code expects to see. | |
4135 | ||
4136 | @smallexample | |
4137 | __CTOR_LIST__ = .; | |
4138 | LONG((__CTOR_END__ - __CTOR_LIST__) / 4 - 2) | |
4139 | *(.ctors) | |
4140 | LONG(0) | |
4141 | __CTOR_END__ = .; | |
4142 | __DTOR_LIST__ = .; | |
4143 | LONG((__DTOR_END__ - __DTOR_LIST__) / 4 - 2) | |
4144 | *(.dtors) | |
4145 | LONG(0) | |
4146 | __DTOR_END__ = .; | |
4147 | @end smallexample | |
4148 | ||
4149 | If you are using the @sc{gnu} C++ support for initialization priority, | |
4150 | which provides some control over the order in which global constructors | |
4151 | are run, you must sort the constructors at link time to ensure that they | |
4152 | are executed in the correct order. When using the @code{CONSTRUCTORS} | |
bcaa7b3e L |
4153 | command, use @samp{SORT_BY_NAME(CONSTRUCTORS)} instead. When using the |
4154 | @code{.ctors} and @code{.dtors} sections, use @samp{*(SORT_BY_NAME(.ctors))} and | |
4155 | @samp{*(SORT_BY_NAME(.dtors))} instead of just @samp{*(.ctors)} and | |
252b5132 RH |
4156 | @samp{*(.dtors)}. |
4157 | ||
4158 | Normally the compiler and linker will handle these issues automatically, | |
4159 | and you will not need to concern yourself with them. However, you may | |
4160 | need to consider this if you are using C++ and writing your own linker | |
4161 | scripts. | |
4162 | ||
4163 | @end table | |
4164 | ||
4165 | @node Output Section Discarding | |
36f63dca | 4166 | @subsection Output Section Discarding |
252b5132 RH |
4167 | @cindex discarding sections |
4168 | @cindex sections, discarding | |
4169 | @cindex removing sections | |
74541ad4 AM |
4170 | The linker will not create output sections with no contents. This is |
4171 | for convenience when referring to input sections that may or may not | |
4172 | be present in any of the input files. For example: | |
252b5132 | 4173 | @smallexample |
49c13adb | 4174 | .foo : @{ *(.foo) @} |
252b5132 RH |
4175 | @end smallexample |
4176 | @noindent | |
4177 | will only create a @samp{.foo} section in the output file if there is a | |
74541ad4 AM |
4178 | @samp{.foo} section in at least one input file, and if the input |
4179 | sections are not all empty. Other link script directives that allocate | |
4180 | space in an output section will also create the output section. | |
4181 | ||
a0976ea4 | 4182 | The linker will ignore address assignments (@pxref{Output Section Address}) |
74541ad4 AM |
4183 | on discarded output sections, except when the linker script defines |
4184 | symbols in the output section. In that case the linker will obey | |
a0976ea4 AM |
4185 | the address assignments, possibly advancing dot even though the |
4186 | section is discarded. | |
252b5132 RH |
4187 | |
4188 | @cindex /DISCARD/ | |
4189 | The special output section name @samp{/DISCARD/} may be used to discard | |
4190 | input sections. Any input sections which are assigned to an output | |
4191 | section named @samp{/DISCARD/} are not included in the output file. | |
4192 | ||
4193 | @node Output Section Attributes | |
36f63dca | 4194 | @subsection Output Section Attributes |
252b5132 RH |
4195 | @cindex output section attributes |
4196 | We showed above that the full description of an output section looked | |
4197 | like this: | |
0c71d759 | 4198 | |
252b5132 | 4199 | @smallexample |
a1ab1d2a | 4200 | @group |
7e7d5768 | 4201 | @var{section} [@var{address}] [(@var{type})] : |
0c71d759 NC |
4202 | [AT(@var{lma})] |
4203 | [ALIGN(@var{section_align})] | |
4204 | [SUBALIGN(@var{subsection_align})] | |
4205 | [@var{constraint}] | |
252b5132 RH |
4206 | @{ |
4207 | @var{output-section-command} | |
4208 | @var{output-section-command} | |
4209 | @dots{} | |
562d3460 | 4210 | @} [>@var{region}] [AT>@var{lma_region}] [:@var{phdr} :@var{phdr} @dots{}] [=@var{fillexp}] |
252b5132 RH |
4211 | @end group |
4212 | @end smallexample | |
0c71d759 | 4213 | |
252b5132 RH |
4214 | We've already described @var{section}, @var{address}, and |
4215 | @var{output-section-command}. In this section we will describe the | |
4216 | remaining section attributes. | |
4217 | ||
a1ab1d2a | 4218 | @menu |
252b5132 RH |
4219 | * Output Section Type:: Output section type |
4220 | * Output Section LMA:: Output section LMA | |
bbf115d3 | 4221 | * Forced Output Alignment:: Forced Output Alignment |
7e7d5768 | 4222 | * Forced Input Alignment:: Forced Input Alignment |
0c71d759 | 4223 | * Output Section Constraint:: Output section constraint |
252b5132 RH |
4224 | * Output Section Region:: Output section region |
4225 | * Output Section Phdr:: Output section phdr | |
4226 | * Output Section Fill:: Output section fill | |
4227 | @end menu | |
4228 | ||
4229 | @node Output Section Type | |
36f63dca | 4230 | @subsubsection Output Section Type |
252b5132 RH |
4231 | Each output section may have a type. The type is a keyword in |
4232 | parentheses. The following types are defined: | |
4233 | ||
4234 | @table @code | |
4235 | @item NOLOAD | |
4236 | The section should be marked as not loadable, so that it will not be | |
4237 | loaded into memory when the program is run. | |
4238 | @item DSECT | |
4239 | @itemx COPY | |
4240 | @itemx INFO | |
4241 | @itemx OVERLAY | |
4242 | These type names are supported for backward compatibility, and are | |
4243 | rarely used. They all have the same effect: the section should be | |
4244 | marked as not allocatable, so that no memory is allocated for the | |
4245 | section when the program is run. | |
4246 | @end table | |
4247 | ||
4248 | @kindex NOLOAD | |
4249 | @cindex prevent unnecessary loading | |
4250 | @cindex loading, preventing | |
4251 | The linker normally sets the attributes of an output section based on | |
4252 | the input sections which map into it. You can override this by using | |
4253 | the section type. For example, in the script sample below, the | |
4254 | @samp{ROM} section is addressed at memory location @samp{0} and does not | |
4255 | need to be loaded when the program is run. The contents of the | |
4256 | @samp{ROM} section will appear in the linker output file as usual. | |
4257 | @smallexample | |
4258 | @group | |
4259 | SECTIONS @{ | |
4260 | ROM 0 (NOLOAD) : @{ @dots{} @} | |
4261 | @dots{} | |
4262 | @} | |
4263 | @end group | |
4264 | @end smallexample | |
4265 | ||
4266 | @node Output Section LMA | |
36f63dca | 4267 | @subsubsection Output Section LMA |
562d3460 | 4268 | @kindex AT>@var{lma_region} |
252b5132 RH |
4269 | @kindex AT(@var{lma}) |
4270 | @cindex load address | |
4271 | @cindex section load address | |
4272 | Every section has a virtual address (VMA) and a load address (LMA); see | |
4273 | @ref{Basic Script Concepts}. The address expression which may appear in | |
4274 | an output section description sets the VMA (@pxref{Output Section | |
4275 | Address}). | |
4276 | ||
dc0b6aa0 AM |
4277 | The expression @var{lma} that follows the @code{AT} keyword specifies |
4278 | the load address of the section. | |
6bdafbeb NC |
4279 | |
4280 | Alternatively, with @samp{AT>@var{lma_region}} expression, you may | |
4281 | specify a memory region for the section's load address. @xref{MEMORY}. | |
4282 | Note that if the section has not had a VMA assigned to it then the | |
4283 | linker will use the @var{lma_region} as the VMA region as well. | |
dc0b6aa0 AM |
4284 | |
4285 | If neither @code{AT} nor @code{AT>} is specified for an allocatable | |
4286 | section, the linker will set the LMA such that the difference between | |
4287 | VMA and LMA for the section is the same as the preceding output | |
4288 | section in the same region. If there is no preceding output section | |
4289 | or the section is not allocatable, the linker will set the LMA equal | |
4290 | to the VMA. | |
6bdafbeb | 4291 | @xref{Output Section Region}. |
252b5132 RH |
4292 | |
4293 | @cindex ROM initialized data | |
4294 | @cindex initialized data in ROM | |
4295 | This feature is designed to make it easy to build a ROM image. For | |
4296 | example, the following linker script creates three output sections: one | |
4297 | called @samp{.text}, which starts at @code{0x1000}, one called | |
4298 | @samp{.mdata}, which is loaded at the end of the @samp{.text} section | |
4299 | even though its VMA is @code{0x2000}, and one called @samp{.bss} to hold | |
4300 | uninitialized data at address @code{0x3000}. The symbol @code{_data} is | |
4301 | defined with the value @code{0x2000}, which shows that the location | |
4302 | counter holds the VMA value, not the LMA value. | |
4303 | ||
4304 | @smallexample | |
4305 | @group | |
4306 | SECTIONS | |
4307 | @{ | |
4308 | .text 0x1000 : @{ *(.text) _etext = . ; @} | |
a1ab1d2a | 4309 | .mdata 0x2000 : |
252b5132 RH |
4310 | AT ( ADDR (.text) + SIZEOF (.text) ) |
4311 | @{ _data = . ; *(.data); _edata = . ; @} | |
4312 | .bss 0x3000 : | |
4313 | @{ _bstart = . ; *(.bss) *(COMMON) ; _bend = . ;@} | |
4314 | @} | |
4315 | @end group | |
4316 | @end smallexample | |
4317 | ||
4318 | The run-time initialization code for use with a program generated with | |
4319 | this linker script would include something like the following, to copy | |
4320 | the initialized data from the ROM image to its runtime address. Notice | |
4321 | how this code takes advantage of the symbols defined by the linker | |
4322 | script. | |
4323 | ||
4324 | @smallexample | |
4325 | @group | |
4326 | extern char _etext, _data, _edata, _bstart, _bend; | |
4327 | char *src = &_etext; | |
4328 | char *dst = &_data; | |
4329 | ||
4330 | /* ROM has data at end of text; copy it. */ | |
4331 | while (dst < &_edata) @{ | |
4332 | *dst++ = *src++; | |
4333 | @} | |
4334 | ||
4335 | /* Zero bss */ | |
4336 | for (dst = &_bstart; dst< &_bend; dst++) | |
4337 | *dst = 0; | |
4338 | @end group | |
4339 | @end smallexample | |
4340 | ||
bbf115d3 L |
4341 | @node Forced Output Alignment |
4342 | @subsubsection Forced Output Alignment | |
4343 | @kindex ALIGN(@var{section_align}) | |
4344 | @cindex forcing output section alignment | |
4345 | @cindex output section alignment | |
7270c5ed | 4346 | You can increase an output section's alignment by using ALIGN. |
bbf115d3 | 4347 | |
7e7d5768 AM |
4348 | @node Forced Input Alignment |
4349 | @subsubsection Forced Input Alignment | |
4350 | @kindex SUBALIGN(@var{subsection_align}) | |
4351 | @cindex forcing input section alignment | |
4352 | @cindex input section alignment | |
4353 | You can force input section alignment within an output section by using | |
4354 | SUBALIGN. The value specified overrides any alignment given by input | |
4355 | sections, whether larger or smaller. | |
4356 | ||
0c71d759 NC |
4357 | @node Output Section Constraint |
4358 | @subsubsection Output Section Constraint | |
4359 | @kindex ONLY_IF_RO | |
4360 | @kindex ONLY_IF_RW | |
4361 | @cindex constraints on output sections | |
4362 | You can specify that an output section should only be created if all | |
4363 | of its input sections are read-only or all of its input sections are | |
4364 | read-write by using the keyword @code{ONLY_IF_RO} and | |
4365 | @code{ONLY_IF_RW} respectively. | |
4366 | ||
252b5132 | 4367 | @node Output Section Region |
36f63dca | 4368 | @subsubsection Output Section Region |
252b5132 RH |
4369 | @kindex >@var{region} |
4370 | @cindex section, assigning to memory region | |
4371 | @cindex memory regions and sections | |
4372 | You can assign a section to a previously defined region of memory by | |
4373 | using @samp{>@var{region}}. @xref{MEMORY}. | |
4374 | ||
4375 | Here is a simple example: | |
4376 | @smallexample | |
4377 | @group | |
4378 | MEMORY @{ rom : ORIGIN = 0x1000, LENGTH = 0x1000 @} | |
4379 | SECTIONS @{ ROM : @{ *(.text) @} >rom @} | |
4380 | @end group | |
4381 | @end smallexample | |
4382 | ||
4383 | @node Output Section Phdr | |
36f63dca | 4384 | @subsubsection Output Section Phdr |
252b5132 RH |
4385 | @kindex :@var{phdr} |
4386 | @cindex section, assigning to program header | |
4387 | @cindex program headers and sections | |
4388 | You can assign a section to a previously defined program segment by | |
4389 | using @samp{:@var{phdr}}. @xref{PHDRS}. If a section is assigned to | |
4390 | one or more segments, then all subsequent allocated sections will be | |
4391 | assigned to those segments as well, unless they use an explicitly | |
4392 | @code{:@var{phdr}} modifier. You can use @code{:NONE} to tell the | |
4393 | linker to not put the section in any segment at all. | |
4394 | ||
4395 | Here is a simple example: | |
4396 | @smallexample | |
4397 | @group | |
4398 | PHDRS @{ text PT_LOAD ; @} | |
4399 | SECTIONS @{ .text : @{ *(.text) @} :text @} | |
4400 | @end group | |
4401 | @end smallexample | |
4402 | ||
4403 | @node Output Section Fill | |
36f63dca | 4404 | @subsubsection Output Section Fill |
252b5132 RH |
4405 | @kindex =@var{fillexp} |
4406 | @cindex section fill pattern | |
4407 | @cindex fill pattern, entire section | |
4408 | You can set the fill pattern for an entire section by using | |
4409 | @samp{=@var{fillexp}}. @var{fillexp} is an expression | |
4410 | (@pxref{Expressions}). Any otherwise unspecified regions of memory | |
4411 | within the output section (for example, gaps left due to the required | |
a139d329 AM |
4412 | alignment of input sections) will be filled with the value, repeated as |
4413 | necessary. If the fill expression is a simple hex number, ie. a string | |
9673c93c | 4414 | of hex digit starting with @samp{0x} and without a trailing @samp{k} or @samp{M}, then |
a139d329 AM |
4415 | an arbitrarily long sequence of hex digits can be used to specify the |
4416 | fill pattern; Leading zeros become part of the pattern too. For all | |
9673c93c | 4417 | other cases, including extra parentheses or a unary @code{+}, the fill |
a139d329 AM |
4418 | pattern is the four least significant bytes of the value of the |
4419 | expression. In all cases, the number is big-endian. | |
252b5132 RH |
4420 | |
4421 | You can also change the fill value with a @code{FILL} command in the | |
9673c93c | 4422 | output section commands; (@pxref{Output Section Data}). |
252b5132 RH |
4423 | |
4424 | Here is a simple example: | |
4425 | @smallexample | |
4426 | @group | |
563e308f | 4427 | SECTIONS @{ .text : @{ *(.text) @} =0x90909090 @} |
252b5132 RH |
4428 | @end group |
4429 | @end smallexample | |
4430 | ||
4431 | @node Overlay Description | |
36f63dca | 4432 | @subsection Overlay Description |
252b5132 RH |
4433 | @kindex OVERLAY |
4434 | @cindex overlays | |
4435 | An overlay description provides an easy way to describe sections which | |
4436 | are to be loaded as part of a single memory image but are to be run at | |
4437 | the same memory address. At run time, some sort of overlay manager will | |
4438 | copy the overlaid sections in and out of the runtime memory address as | |
4439 | required, perhaps by simply manipulating addressing bits. This approach | |
4440 | can be useful, for example, when a certain region of memory is faster | |
4441 | than another. | |
4442 | ||
4443 | Overlays are described using the @code{OVERLAY} command. The | |
4444 | @code{OVERLAY} command is used within a @code{SECTIONS} command, like an | |
4445 | output section description. The full syntax of the @code{OVERLAY} | |
4446 | command is as follows: | |
4447 | @smallexample | |
4448 | @group | |
4449 | OVERLAY [@var{start}] : [NOCROSSREFS] [AT ( @var{ldaddr} )] | |
4450 | @{ | |
4451 | @var{secname1} | |
4452 | @{ | |
4453 | @var{output-section-command} | |
4454 | @var{output-section-command} | |
4455 | @dots{} | |
4456 | @} [:@var{phdr}@dots{}] [=@var{fill}] | |
4457 | @var{secname2} | |
4458 | @{ | |
4459 | @var{output-section-command} | |
4460 | @var{output-section-command} | |
4461 | @dots{} | |
4462 | @} [:@var{phdr}@dots{}] [=@var{fill}] | |
4463 | @dots{} | |
4464 | @} [>@var{region}] [:@var{phdr}@dots{}] [=@var{fill}] | |
4465 | @end group | |
4466 | @end smallexample | |
4467 | ||
4468 | Everything is optional except @code{OVERLAY} (a keyword), and each | |
4469 | section must have a name (@var{secname1} and @var{secname2} above). The | |
4470 | section definitions within the @code{OVERLAY} construct are identical to | |
4471 | those within the general @code{SECTIONS} contruct (@pxref{SECTIONS}), | |
4472 | except that no addresses and no memory regions may be defined for | |
4473 | sections within an @code{OVERLAY}. | |
4474 | ||
4475 | The sections are all defined with the same starting address. The load | |
4476 | addresses of the sections are arranged such that they are consecutive in | |
4477 | memory starting at the load address used for the @code{OVERLAY} as a | |
4478 | whole (as with normal section definitions, the load address is optional, | |
4479 | and defaults to the start address; the start address is also optional, | |
4480 | and defaults to the current value of the location counter). | |
4481 | ||
4482 | If the @code{NOCROSSREFS} keyword is used, and there any references | |
4483 | among the sections, the linker will report an error. Since the sections | |
4484 | all run at the same address, it normally does not make sense for one | |
4485 | section to refer directly to another. @xref{Miscellaneous Commands, | |
4486 | NOCROSSREFS}. | |
4487 | ||
4488 | For each section within the @code{OVERLAY}, the linker automatically | |
34711ca3 | 4489 | provides two symbols. The symbol @code{__load_start_@var{secname}} is |
252b5132 RH |
4490 | defined as the starting load address of the section. The symbol |
4491 | @code{__load_stop_@var{secname}} is defined as the final load address of | |
4492 | the section. Any characters within @var{secname} which are not legal | |
4493 | within C identifiers are removed. C (or assembler) code may use these | |
4494 | symbols to move the overlaid sections around as necessary. | |
4495 | ||
4496 | At the end of the overlay, the value of the location counter is set to | |
4497 | the start address of the overlay plus the size of the largest section. | |
4498 | ||
4499 | Here is an example. Remember that this would appear inside a | |
4500 | @code{SECTIONS} construct. | |
4501 | @smallexample | |
4502 | @group | |
4503 | OVERLAY 0x1000 : AT (0x4000) | |
4504 | @{ | |
4505 | .text0 @{ o1/*.o(.text) @} | |
4506 | .text1 @{ o2/*.o(.text) @} | |
4507 | @} | |
4508 | @end group | |
4509 | @end smallexample | |
4510 | @noindent | |
4511 | This will define both @samp{.text0} and @samp{.text1} to start at | |
4512 | address 0x1000. @samp{.text0} will be loaded at address 0x4000, and | |
4513 | @samp{.text1} will be loaded immediately after @samp{.text0}. The | |
34711ca3 | 4514 | following symbols will be defined if referenced: @code{__load_start_text0}, |
252b5132 RH |
4515 | @code{__load_stop_text0}, @code{__load_start_text1}, |
4516 | @code{__load_stop_text1}. | |
4517 | ||
4518 | C code to copy overlay @code{.text1} into the overlay area might look | |
4519 | like the following. | |
4520 | ||
4521 | @smallexample | |
4522 | @group | |
4523 | extern char __load_start_text1, __load_stop_text1; | |
4524 | memcpy ((char *) 0x1000, &__load_start_text1, | |
4525 | &__load_stop_text1 - &__load_start_text1); | |
4526 | @end group | |
4527 | @end smallexample | |
4528 | ||
4529 | Note that the @code{OVERLAY} command is just syntactic sugar, since | |
4530 | everything it does can be done using the more basic commands. The above | |
4531 | example could have been written identically as follows. | |
4532 | ||
4533 | @smallexample | |
4534 | @group | |
4535 | .text0 0x1000 : AT (0x4000) @{ o1/*.o(.text) @} | |
34711ca3 AM |
4536 | PROVIDE (__load_start_text0 = LOADADDR (.text0)); |
4537 | PROVIDE (__load_stop_text0 = LOADADDR (.text0) + SIZEOF (.text0)); | |
252b5132 | 4538 | .text1 0x1000 : AT (0x4000 + SIZEOF (.text0)) @{ o2/*.o(.text) @} |
34711ca3 AM |
4539 | PROVIDE (__load_start_text1 = LOADADDR (.text1)); |
4540 | PROVIDE (__load_stop_text1 = LOADADDR (.text1) + SIZEOF (.text1)); | |
252b5132 RH |
4541 | . = 0x1000 + MAX (SIZEOF (.text0), SIZEOF (.text1)); |
4542 | @end group | |
4543 | @end smallexample | |
4544 | ||
4545 | @node MEMORY | |
36f63dca | 4546 | @section MEMORY Command |
252b5132 RH |
4547 | @kindex MEMORY |
4548 | @cindex memory regions | |
4549 | @cindex regions of memory | |
4550 | @cindex allocating memory | |
4551 | @cindex discontinuous memory | |
4552 | The linker's default configuration permits allocation of all available | |
4553 | memory. You can override this by using the @code{MEMORY} command. | |
4554 | ||
4555 | The @code{MEMORY} command describes the location and size of blocks of | |
4556 | memory in the target. You can use it to describe which memory regions | |
4557 | may be used by the linker, and which memory regions it must avoid. You | |
4558 | can then assign sections to particular memory regions. The linker will | |
4559 | set section addresses based on the memory regions, and will warn about | |
4560 | regions that become too full. The linker will not shuffle sections | |
4561 | around to fit into the available regions. | |
4562 | ||
4563 | A linker script may contain at most one use of the @code{MEMORY} | |
4564 | command. However, you can define as many blocks of memory within it as | |
4565 | you wish. The syntax is: | |
4566 | @smallexample | |
4567 | @group | |
a1ab1d2a | 4568 | MEMORY |
252b5132 RH |
4569 | @{ |
4570 | @var{name} [(@var{attr})] : ORIGIN = @var{origin}, LENGTH = @var{len} | |
4571 | @dots{} | |
4572 | @} | |
4573 | @end group | |
4574 | @end smallexample | |
4575 | ||
4576 | The @var{name} is a name used in the linker script to refer to the | |
4577 | region. The region name has no meaning outside of the linker script. | |
4578 | Region names are stored in a separate name space, and will not conflict | |
4579 | with symbol names, file names, or section names. Each memory region | |
4a93e180 NC |
4580 | must have a distinct name within the @code{MEMORY} command. However you can |
4581 | add later alias names to existing memory regions with the @ref{REGION_ALIAS} | |
4582 | command. | |
252b5132 RH |
4583 | |
4584 | @cindex memory region attributes | |
4585 | The @var{attr} string is an optional list of attributes that specify | |
4586 | whether to use a particular memory region for an input section which is | |
4587 | not explicitly mapped in the linker script. As described in | |
4588 | @ref{SECTIONS}, if you do not specify an output section for some input | |
4589 | section, the linker will create an output section with the same name as | |
4590 | the input section. If you define region attributes, the linker will use | |
4591 | them to select the memory region for the output section that it creates. | |
4592 | ||
4593 | The @var{attr} string must consist only of the following characters: | |
4594 | @table @samp | |
4595 | @item R | |
4596 | Read-only section | |
4597 | @item W | |
4598 | Read/write section | |
4599 | @item X | |
4600 | Executable section | |
4601 | @item A | |
4602 | Allocatable section | |
4603 | @item I | |
4604 | Initialized section | |
4605 | @item L | |
4606 | Same as @samp{I} | |
4607 | @item ! | |
4608 | Invert the sense of any of the preceding attributes | |
4609 | @end table | |
4610 | ||
4611 | If a unmapped section matches any of the listed attributes other than | |
4612 | @samp{!}, it will be placed in the memory region. The @samp{!} | |
4613 | attribute reverses this test, so that an unmapped section will be placed | |
4614 | in the memory region only if it does not match any of the listed | |
4615 | attributes. | |
4616 | ||
4617 | @kindex ORIGIN = | |
4618 | @kindex o = | |
4619 | @kindex org = | |
9cd6d51a NC |
4620 | The @var{origin} is an numerical expression for the start address of |
4621 | the memory region. The expression must evaluate to a constant and it | |
4622 | cannot involve any symbols. The keyword @code{ORIGIN} may be | |
4623 | abbreviated to @code{org} or @code{o} (but not, for example, | |
4624 | @code{ORG}). | |
252b5132 RH |
4625 | |
4626 | @kindex LENGTH = | |
4627 | @kindex len = | |
4628 | @kindex l = | |
4629 | The @var{len} is an expression for the size in bytes of the memory | |
4630 | region. As with the @var{origin} expression, the expression must | |
9cd6d51a NC |
4631 | be numerical only and must evaluate to a constant. The keyword |
4632 | @code{LENGTH} may be abbreviated to @code{len} or @code{l}. | |
252b5132 RH |
4633 | |
4634 | In the following example, we specify that there are two memory regions | |
4635 | available for allocation: one starting at @samp{0} for 256 kilobytes, | |
4636 | and the other starting at @samp{0x40000000} for four megabytes. The | |
4637 | linker will place into the @samp{rom} memory region every section which | |
4638 | is not explicitly mapped into a memory region, and is either read-only | |
4639 | or executable. The linker will place other sections which are not | |
4640 | explicitly mapped into a memory region into the @samp{ram} memory | |
4641 | region. | |
4642 | ||
4643 | @smallexample | |
4644 | @group | |
a1ab1d2a | 4645 | MEMORY |
252b5132 RH |
4646 | @{ |
4647 | rom (rx) : ORIGIN = 0, LENGTH = 256K | |
4648 | ram (!rx) : org = 0x40000000, l = 4M | |
4649 | @} | |
4650 | @end group | |
4651 | @end smallexample | |
4652 | ||
4653 | Once you define a memory region, you can direct the linker to place | |
4654 | specific output sections into that memory region by using the | |
4655 | @samp{>@var{region}} output section attribute. For example, if you have | |
4656 | a memory region named @samp{mem}, you would use @samp{>mem} in the | |
4657 | output section definition. @xref{Output Section Region}. If no address | |
4658 | was specified for the output section, the linker will set the address to | |
4659 | the next available address within the memory region. If the combined | |
4660 | output sections directed to a memory region are too large for the | |
4661 | region, the linker will issue an error message. | |
4662 | ||
3ec57632 | 4663 | It is possible to access the origin and length of a memory in an |
c0065db7 | 4664 | expression via the @code{ORIGIN(@var{memory})} and |
3ec57632 NC |
4665 | @code{LENGTH(@var{memory})} functions: |
4666 | ||
4667 | @smallexample | |
4668 | @group | |
c0065db7 | 4669 | _fstack = ORIGIN(ram) + LENGTH(ram) - 4; |
3ec57632 NC |
4670 | @end group |
4671 | @end smallexample | |
4672 | ||
252b5132 RH |
4673 | @node PHDRS |
4674 | @section PHDRS Command | |
4675 | @kindex PHDRS | |
4676 | @cindex program headers | |
4677 | @cindex ELF program headers | |
4678 | @cindex program segments | |
4679 | @cindex segments, ELF | |
4680 | The ELF object file format uses @dfn{program headers}, also knows as | |
4681 | @dfn{segments}. The program headers describe how the program should be | |
4682 | loaded into memory. You can print them out by using the @code{objdump} | |
4683 | program with the @samp{-p} option. | |
4684 | ||
4685 | When you run an ELF program on a native ELF system, the system loader | |
4686 | reads the program headers in order to figure out how to load the | |
4687 | program. This will only work if the program headers are set correctly. | |
4688 | This manual does not describe the details of how the system loader | |
4689 | interprets program headers; for more information, see the ELF ABI. | |
4690 | ||
4691 | The linker will create reasonable program headers by default. However, | |
4692 | in some cases, you may need to specify the program headers more | |
4693 | precisely. You may use the @code{PHDRS} command for this purpose. When | |
4694 | the linker sees the @code{PHDRS} command in the linker script, it will | |
4695 | not create any program headers other than the ones specified. | |
4696 | ||
4697 | The linker only pays attention to the @code{PHDRS} command when | |
4698 | generating an ELF output file. In other cases, the linker will simply | |
4699 | ignore @code{PHDRS}. | |
4700 | ||
4701 | This is the syntax of the @code{PHDRS} command. The words @code{PHDRS}, | |
4702 | @code{FILEHDR}, @code{AT}, and @code{FLAGS} are keywords. | |
4703 | ||
4704 | @smallexample | |
4705 | @group | |
4706 | PHDRS | |
4707 | @{ | |
4708 | @var{name} @var{type} [ FILEHDR ] [ PHDRS ] [ AT ( @var{address} ) ] | |
4709 | [ FLAGS ( @var{flags} ) ] ; | |
4710 | @} | |
4711 | @end group | |
4712 | @end smallexample | |
4713 | ||
4714 | The @var{name} is used only for reference in the @code{SECTIONS} command | |
4715 | of the linker script. It is not put into the output file. Program | |
4716 | header names are stored in a separate name space, and will not conflict | |
4717 | with symbol names, file names, or section names. Each program header | |
5c1a3f0f NS |
4718 | must have a distinct name. The headers are processed in order and it |
4719 | is usual for them to map to sections in ascending load address order. | |
252b5132 RH |
4720 | |
4721 | Certain program header types describe segments of memory which the | |
4722 | system loader will load from the file. In the linker script, you | |
4723 | specify the contents of these segments by placing allocatable output | |
4724 | sections in the segments. You use the @samp{:@var{phdr}} output section | |
4725 | attribute to place a section in a particular segment. @xref{Output | |
4726 | Section Phdr}. | |
4727 | ||
4728 | It is normal to put certain sections in more than one segment. This | |
4729 | merely implies that one segment of memory contains another. You may | |
4730 | repeat @samp{:@var{phdr}}, using it once for each segment which should | |
4731 | contain the section. | |
4732 | ||
4733 | If you place a section in one or more segments using @samp{:@var{phdr}}, | |
4734 | then the linker will place all subsequent allocatable sections which do | |
4735 | not specify @samp{:@var{phdr}} in the same segments. This is for | |
4736 | convenience, since generally a whole set of contiguous sections will be | |
4737 | placed in a single segment. You can use @code{:NONE} to override the | |
4738 | default segment and tell the linker to not put the section in any | |
4739 | segment at all. | |
4740 | ||
4741 | @kindex FILEHDR | |
4742 | @kindex PHDRS | |
5c1a3f0f | 4743 | You may use the @code{FILEHDR} and @code{PHDRS} keywords after |
252b5132 RH |
4744 | the program header type to further describe the contents of the segment. |
4745 | The @code{FILEHDR} keyword means that the segment should include the ELF | |
4746 | file header. The @code{PHDRS} keyword means that the segment should | |
5c1a3f0f | 4747 | include the ELF program headers themselves. If applied to a loadable |
4100cea3 AM |
4748 | segment (@code{PT_LOAD}), all prior loadable segments must have one of |
4749 | these keywords. | |
252b5132 RH |
4750 | |
4751 | The @var{type} may be one of the following. The numbers indicate the | |
4752 | value of the keyword. | |
4753 | ||
4754 | @table @asis | |
4755 | @item @code{PT_NULL} (0) | |
4756 | Indicates an unused program header. | |
4757 | ||
4758 | @item @code{PT_LOAD} (1) | |
4759 | Indicates that this program header describes a segment to be loaded from | |
4760 | the file. | |
4761 | ||
4762 | @item @code{PT_DYNAMIC} (2) | |
4763 | Indicates a segment where dynamic linking information can be found. | |
4764 | ||
4765 | @item @code{PT_INTERP} (3) | |
4766 | Indicates a segment where the name of the program interpreter may be | |
4767 | found. | |
4768 | ||
4769 | @item @code{PT_NOTE} (4) | |
4770 | Indicates a segment holding note information. | |
4771 | ||
4772 | @item @code{PT_SHLIB} (5) | |
4773 | A reserved program header type, defined but not specified by the ELF | |
4774 | ABI. | |
4775 | ||
4776 | @item @code{PT_PHDR} (6) | |
4777 | Indicates a segment where the program headers may be found. | |
4778 | ||
4779 | @item @var{expression} | |
4780 | An expression giving the numeric type of the program header. This may | |
4781 | be used for types not defined above. | |
4782 | @end table | |
4783 | ||
4784 | You can specify that a segment should be loaded at a particular address | |
4785 | in memory by using an @code{AT} expression. This is identical to the | |
4786 | @code{AT} command used as an output section attribute (@pxref{Output | |
4787 | Section LMA}). The @code{AT} command for a program header overrides the | |
4788 | output section attribute. | |
4789 | ||
4790 | The linker will normally set the segment flags based on the sections | |
4791 | which comprise the segment. You may use the @code{FLAGS} keyword to | |
4792 | explicitly specify the segment flags. The value of @var{flags} must be | |
4793 | an integer. It is used to set the @code{p_flags} field of the program | |
4794 | header. | |
4795 | ||
4796 | Here is an example of @code{PHDRS}. This shows a typical set of program | |
4797 | headers used on a native ELF system. | |
4798 | ||
4799 | @example | |
4800 | @group | |
4801 | PHDRS | |
4802 | @{ | |
4803 | headers PT_PHDR PHDRS ; | |
4804 | interp PT_INTERP ; | |
4805 | text PT_LOAD FILEHDR PHDRS ; | |
4806 | data PT_LOAD ; | |
4807 | dynamic PT_DYNAMIC ; | |
4808 | @} | |
4809 | ||
4810 | SECTIONS | |
4811 | @{ | |
4812 | . = SIZEOF_HEADERS; | |
4813 | .interp : @{ *(.interp) @} :text :interp | |
4814 | .text : @{ *(.text) @} :text | |
4815 | .rodata : @{ *(.rodata) @} /* defaults to :text */ | |
4816 | @dots{} | |
4817 | . = . + 0x1000; /* move to a new page in memory */ | |
4818 | .data : @{ *(.data) @} :data | |
4819 | .dynamic : @{ *(.dynamic) @} :data :dynamic | |
4820 | @dots{} | |
4821 | @} | |
4822 | @end group | |
4823 | @end example | |
4824 | ||
4825 | @node VERSION | |
4826 | @section VERSION Command | |
4827 | @kindex VERSION @{script text@} | |
4828 | @cindex symbol versions | |
4829 | @cindex version script | |
4830 | @cindex versions of symbols | |
4831 | The linker supports symbol versions when using ELF. Symbol versions are | |
4832 | only useful when using shared libraries. The dynamic linker can use | |
4833 | symbol versions to select a specific version of a function when it runs | |
4834 | a program that may have been linked against an earlier version of the | |
4835 | shared library. | |
4836 | ||
4837 | You can include a version script directly in the main linker script, or | |
4838 | you can supply the version script as an implicit linker script. You can | |
4839 | also use the @samp{--version-script} linker option. | |
4840 | ||
4841 | The syntax of the @code{VERSION} command is simply | |
4842 | @smallexample | |
4843 | VERSION @{ version-script-commands @} | |
4844 | @end smallexample | |
4845 | ||
4846 | The format of the version script commands is identical to that used by | |
4847 | Sun's linker in Solaris 2.5. The version script defines a tree of | |
4848 | version nodes. You specify the node names and interdependencies in the | |
4849 | version script. You can specify which symbols are bound to which | |
4850 | version nodes, and you can reduce a specified set of symbols to local | |
4851 | scope so that they are not globally visible outside of the shared | |
4852 | library. | |
4853 | ||
4854 | The easiest way to demonstrate the version script language is with a few | |
4855 | examples. | |
4856 | ||
4857 | @smallexample | |
4858 | VERS_1.1 @{ | |
4859 | global: | |
4860 | foo1; | |
4861 | local: | |
a1ab1d2a UD |
4862 | old*; |
4863 | original*; | |
4864 | new*; | |
252b5132 RH |
4865 | @}; |
4866 | ||
4867 | VERS_1.2 @{ | |
4868 | foo2; | |
4869 | @} VERS_1.1; | |
4870 | ||
4871 | VERS_2.0 @{ | |
4872 | bar1; bar2; | |
c0065db7 | 4873 | extern "C++" @{ |
86043bbb MM |
4874 | ns::*; |
4875 | "int f(int, double)"; | |
c0065db7 | 4876 | @} |
252b5132 RH |
4877 | @} VERS_1.2; |
4878 | @end smallexample | |
4879 | ||
4880 | This example version script defines three version nodes. The first | |
4881 | version node defined is @samp{VERS_1.1}; it has no other dependencies. | |
4882 | The script binds the symbol @samp{foo1} to @samp{VERS_1.1}. It reduces | |
4883 | a number of symbols to local scope so that they are not visible outside | |
313e35ee AM |
4884 | of the shared library; this is done using wildcard patterns, so that any |
4885 | symbol whose name begins with @samp{old}, @samp{original}, or @samp{new} | |
4886 | is matched. The wildcard patterns available are the same as those used | |
4887 | in the shell when matching filenames (also known as ``globbing''). | |
86043bbb MM |
4888 | However, if you specify the symbol name inside double quotes, then the |
4889 | name is treated as literal, rather than as a glob pattern. | |
252b5132 RH |
4890 | |
4891 | Next, the version script defines node @samp{VERS_1.2}. This node | |
4892 | depends upon @samp{VERS_1.1}. The script binds the symbol @samp{foo2} | |
4893 | to the version node @samp{VERS_1.2}. | |
4894 | ||
4895 | Finally, the version script defines node @samp{VERS_2.0}. This node | |
4896 | depends upon @samp{VERS_1.2}. The scripts binds the symbols @samp{bar1} | |
4897 | and @samp{bar2} are bound to the version node @samp{VERS_2.0}. | |
4898 | ||
4899 | When the linker finds a symbol defined in a library which is not | |
4900 | specifically bound to a version node, it will effectively bind it to an | |
4901 | unspecified base version of the library. You can bind all otherwise | |
a981ed6f | 4902 | unspecified symbols to a given version node by using @samp{global: *;} |
ae5a3597 AM |
4903 | somewhere in the version script. Note that it's slightly crazy to use |
4904 | wildcards in a global spec except on the last version node. Global | |
4905 | wildcards elsewhere run the risk of accidentally adding symbols to the | |
4906 | set exported for an old version. That's wrong since older versions | |
4907 | ought to have a fixed set of symbols. | |
252b5132 RH |
4908 | |
4909 | The names of the version nodes have no specific meaning other than what | |
4910 | they might suggest to the person reading them. The @samp{2.0} version | |
4911 | could just as well have appeared in between @samp{1.1} and @samp{1.2}. | |
4912 | However, this would be a confusing way to write a version script. | |
4913 | ||
0f6bf451 | 4914 | Node name can be omitted, provided it is the only version node |
6b9b879a JJ |
4915 | in the version script. Such version script doesn't assign any versions to |
4916 | symbols, only selects which symbols will be globally visible out and which | |
4917 | won't. | |
4918 | ||
4919 | @smallexample | |
7c9c73be | 4920 | @{ global: foo; bar; local: *; @}; |
9d201f2f | 4921 | @end smallexample |
6b9b879a | 4922 | |
252b5132 RH |
4923 | When you link an application against a shared library that has versioned |
4924 | symbols, the application itself knows which version of each symbol it | |
4925 | requires, and it also knows which version nodes it needs from each | |
4926 | shared library it is linked against. Thus at runtime, the dynamic | |
4927 | loader can make a quick check to make sure that the libraries you have | |
4928 | linked against do in fact supply all of the version nodes that the | |
4929 | application will need to resolve all of the dynamic symbols. In this | |
4930 | way it is possible for the dynamic linker to know with certainty that | |
4931 | all external symbols that it needs will be resolvable without having to | |
4932 | search for each symbol reference. | |
4933 | ||
4934 | The symbol versioning is in effect a much more sophisticated way of | |
4935 | doing minor version checking that SunOS does. The fundamental problem | |
4936 | that is being addressed here is that typically references to external | |
4937 | functions are bound on an as-needed basis, and are not all bound when | |
4938 | the application starts up. If a shared library is out of date, a | |
4939 | required interface may be missing; when the application tries to use | |
4940 | that interface, it may suddenly and unexpectedly fail. With symbol | |
4941 | versioning, the user will get a warning when they start their program if | |
4942 | the libraries being used with the application are too old. | |
4943 | ||
4944 | There are several GNU extensions to Sun's versioning approach. The | |
4945 | first of these is the ability to bind a symbol to a version node in the | |
4946 | source file where the symbol is defined instead of in the versioning | |
4947 | script. This was done mainly to reduce the burden on the library | |
4948 | maintainer. You can do this by putting something like: | |
4949 | @smallexample | |
4950 | __asm__(".symver original_foo,foo@@VERS_1.1"); | |
4951 | @end smallexample | |
4952 | @noindent | |
4953 | in the C source file. This renames the function @samp{original_foo} to | |
4954 | be an alias for @samp{foo} bound to the version node @samp{VERS_1.1}. | |
4955 | The @samp{local:} directive can be used to prevent the symbol | |
96a94295 L |
4956 | @samp{original_foo} from being exported. A @samp{.symver} directive |
4957 | takes precedence over a version script. | |
252b5132 RH |
4958 | |
4959 | The second GNU extension is to allow multiple versions of the same | |
4960 | function to appear in a given shared library. In this way you can make | |
4961 | an incompatible change to an interface without increasing the major | |
4962 | version number of the shared library, while still allowing applications | |
4963 | linked against the old interface to continue to function. | |
4964 | ||
4965 | To do this, you must use multiple @samp{.symver} directives in the | |
4966 | source file. Here is an example: | |
4967 | ||
4968 | @smallexample | |
4969 | __asm__(".symver original_foo,foo@@"); | |
4970 | __asm__(".symver old_foo,foo@@VERS_1.1"); | |
4971 | __asm__(".symver old_foo1,foo@@VERS_1.2"); | |
4972 | __asm__(".symver new_foo,foo@@@@VERS_2.0"); | |
4973 | @end smallexample | |
4974 | ||
4975 | In this example, @samp{foo@@} represents the symbol @samp{foo} bound to the | |
4976 | unspecified base version of the symbol. The source file that contains this | |
4977 | example would define 4 C functions: @samp{original_foo}, @samp{old_foo}, | |
4978 | @samp{old_foo1}, and @samp{new_foo}. | |
4979 | ||
4980 | When you have multiple definitions of a given symbol, there needs to be | |
4981 | some way to specify a default version to which external references to | |
4982 | this symbol will be bound. You can do this with the | |
4983 | @samp{foo@@@@VERS_2.0} type of @samp{.symver} directive. You can only | |
4984 | declare one version of a symbol as the default in this manner; otherwise | |
4985 | you would effectively have multiple definitions of the same symbol. | |
4986 | ||
4987 | If you wish to bind a reference to a specific version of the symbol | |
4988 | within the shared library, you can use the aliases of convenience | |
36f63dca | 4989 | (i.e., @samp{old_foo}), or you can use the @samp{.symver} directive to |
252b5132 RH |
4990 | specifically bind to an external version of the function in question. |
4991 | ||
cb840a31 L |
4992 | You can also specify the language in the version script: |
4993 | ||
4994 | @smallexample | |
4995 | VERSION extern "lang" @{ version-script-commands @} | |
4996 | @end smallexample | |
4997 | ||
c0065db7 | 4998 | The supported @samp{lang}s are @samp{C}, @samp{C++}, and @samp{Java}. |
cb840a31 L |
4999 | The linker will iterate over the list of symbols at the link time and |
5000 | demangle them according to @samp{lang} before matching them to the | |
5001 | patterns specified in @samp{version-script-commands}. | |
5002 | ||
86043bbb MM |
5003 | Demangled names may contains spaces and other special characters. As |
5004 | described above, you can use a glob pattern to match demangled names, | |
5005 | or you can use a double-quoted string to match the string exactly. In | |
5006 | the latter case, be aware that minor differences (such as differing | |
5007 | whitespace) between the version script and the demangler output will | |
5008 | cause a mismatch. As the exact string generated by the demangler | |
5009 | might change in the future, even if the mangled name does not, you | |
5010 | should check that all of your version directives are behaving as you | |
5011 | expect when you upgrade. | |
5012 | ||
252b5132 RH |
5013 | @node Expressions |
5014 | @section Expressions in Linker Scripts | |
5015 | @cindex expressions | |
5016 | @cindex arithmetic | |
5017 | The syntax for expressions in the linker script language is identical to | |
5018 | that of C expressions. All expressions are evaluated as integers. All | |
5019 | expressions are evaluated in the same size, which is 32 bits if both the | |
5020 | host and target are 32 bits, and is otherwise 64 bits. | |
5021 | ||
5022 | You can use and set symbol values in expressions. | |
5023 | ||
5024 | The linker defines several special purpose builtin functions for use in | |
5025 | expressions. | |
5026 | ||
5027 | @menu | |
5028 | * Constants:: Constants | |
0c71d759 | 5029 | * Symbolic Constants:: Symbolic constants |
252b5132 | 5030 | * Symbols:: Symbol Names |
ecca9871 | 5031 | * Orphan Sections:: Orphan Sections |
252b5132 RH |
5032 | * Location Counter:: The Location Counter |
5033 | * Operators:: Operators | |
5034 | * Evaluation:: Evaluation | |
5035 | * Expression Section:: The Section of an Expression | |
5036 | * Builtin Functions:: Builtin Functions | |
5037 | @end menu | |
5038 | ||
5039 | @node Constants | |
5040 | @subsection Constants | |
5041 | @cindex integer notation | |
5042 | @cindex constants in linker scripts | |
5043 | All constants are integers. | |
5044 | ||
5045 | As in C, the linker considers an integer beginning with @samp{0} to be | |
5046 | octal, and an integer beginning with @samp{0x} or @samp{0X} to be | |
8a308ae8 NC |
5047 | hexadecimal. Alternatively the linker accepts suffixes of @samp{h} or |
5048 | @samp{H} for hexadeciaml, @samp{o} or @samp{O} for octal, @samp{b} or | |
5049 | @samp{B} for binary and @samp{d} or @samp{D} for decimal. Any integer | |
5050 | value without a prefix or a suffix is considered to be decimal. | |
252b5132 RH |
5051 | |
5052 | @cindex scaled integers | |
5053 | @cindex K and M integer suffixes | |
5054 | @cindex M and K integer suffixes | |
5055 | @cindex suffixes for integers | |
5056 | @cindex integer suffixes | |
5057 | In addition, you can use the suffixes @code{K} and @code{M} to scale a | |
5058 | constant by | |
5059 | @c TEXI2ROFF-KILL | |
36f63dca | 5060 | @ifnottex |
252b5132 RH |
5061 | @c END TEXI2ROFF-KILL |
5062 | @code{1024} or @code{1024*1024} | |
5063 | @c TEXI2ROFF-KILL | |
36f63dca | 5064 | @end ifnottex |
252b5132 RH |
5065 | @tex |
5066 | ${\rm 1024}$ or ${\rm 1024}^2$ | |
5067 | @end tex | |
5068 | @c END TEXI2ROFF-KILL | |
8a308ae8 NC |
5069 | respectively. For example, the following |
5070 | all refer to the same quantity: | |
5071 | ||
252b5132 | 5072 | @smallexample |
36f63dca NC |
5073 | _fourk_1 = 4K; |
5074 | _fourk_2 = 4096; | |
5075 | _fourk_3 = 0x1000; | |
8a308ae8 | 5076 | _fourk_4 = 10000o; |
252b5132 RH |
5077 | @end smallexample |
5078 | ||
8a308ae8 NC |
5079 | Note - the @code{K} and @code{M} suffixes cannot be used in |
5080 | conjunction with the base suffixes mentioned above. | |
5081 | ||
0c71d759 NC |
5082 | @node Symbolic Constants |
5083 | @subsection Symbolic Constants | |
5084 | @cindex symbolic constants | |
5085 | @kindex CONSTANT | |
5086 | It is possible to refer to target specific constants via the use of | |
5087 | the @code{CONSTANT(@var{name})} operator, where @var{name} is one of: | |
5088 | ||
5089 | @table @code | |
5090 | @item MAXPAGESIZE | |
5091 | @kindex MAXPAGESIZE | |
5092 | The target's maximum page size. | |
5093 | ||
5094 | @item COMMONPAGESIZE | |
5095 | @kindex COMMONPAGESIZE | |
5096 | The target's default page size. | |
5097 | @end table | |
5098 | ||
5099 | So for example: | |
5100 | ||
5101 | @smallexample | |
5102 | .text ALIGN (CONSTANT (MAXPAGESIZE)) : @{ *(.text) @} | |
5103 | @end smallexample | |
5104 | ||
5105 | will create a text section aligned to the largest page boundary | |
5106 | supported by the target. | |
5107 | ||
252b5132 RH |
5108 | @node Symbols |
5109 | @subsection Symbol Names | |
5110 | @cindex symbol names | |
5111 | @cindex names | |
5112 | @cindex quoted symbol names | |
5113 | @kindex " | |
5114 | Unless quoted, symbol names start with a letter, underscore, or period | |
5115 | and may include letters, digits, underscores, periods, and hyphens. | |
5116 | Unquoted symbol names must not conflict with any keywords. You can | |
5117 | specify a symbol which contains odd characters or has the same name as a | |
5118 | keyword by surrounding the symbol name in double quotes: | |
5119 | @smallexample | |
36f63dca NC |
5120 | "SECTION" = 9; |
5121 | "with a space" = "also with a space" + 10; | |
252b5132 RH |
5122 | @end smallexample |
5123 | ||
5124 | Since symbols can contain many non-alphabetic characters, it is safest | |
5125 | to delimit symbols with spaces. For example, @samp{A-B} is one symbol, | |
5126 | whereas @samp{A - B} is an expression involving subtraction. | |
5127 | ||
ecca9871 L |
5128 | @node Orphan Sections |
5129 | @subsection Orphan Sections | |
5130 | @cindex orphan | |
5131 | Orphan sections are sections present in the input files which | |
5132 | are not explicitly placed into the output file by the linker | |
5133 | script. The linker will still copy these sections into the | |
5134 | output file, but it has to guess as to where they should be | |
5135 | placed. The linker uses a simple heuristic to do this. It | |
5136 | attempts to place orphan sections after non-orphan sections of the | |
5137 | same attribute, such as code vs data, loadable vs non-loadable, etc. | |
5138 | If there is not enough room to do this then it places | |
5139 | at the end of the file. | |
5140 | ||
5141 | For ELF targets, the attribute of the section includes section type as | |
5142 | well as section flag. | |
5143 | ||
41911f68 | 5144 | If an orphaned section's name is representable as a C identifier then |
a61ca861 | 5145 | the linker will automatically @pxref{PROVIDE} two symbols: |
41911f68 NC |
5146 | __start_SECNAME and __end_SECNAME, where SECNAME is the name of the |
5147 | section. These indicate the start address and end address of the | |
5148 | orphaned section respectively. Note: most section names are not | |
5149 | representable as C identifiers because they contain a @samp{.} | |
5150 | character. | |
5151 | ||
252b5132 RH |
5152 | @node Location Counter |
5153 | @subsection The Location Counter | |
5154 | @kindex . | |
5155 | @cindex dot | |
5156 | @cindex location counter | |
5157 | @cindex current output location | |
5158 | The special linker variable @dfn{dot} @samp{.} always contains the | |
5159 | current output location counter. Since the @code{.} always refers to a | |
5160 | location in an output section, it may only appear in an expression | |
5161 | within a @code{SECTIONS} command. The @code{.} symbol may appear | |
5162 | anywhere that an ordinary symbol is allowed in an expression. | |
5163 | ||
5164 | @cindex holes | |
5165 | Assigning a value to @code{.} will cause the location counter to be | |
5166 | moved. This may be used to create holes in the output section. The | |
dc0b6aa0 AM |
5167 | location counter may not be moved backwards inside an output section, |
5168 | and may not be moved backwards outside of an output section if so | |
5169 | doing creates areas with overlapping LMAs. | |
252b5132 RH |
5170 | |
5171 | @smallexample | |
5172 | SECTIONS | |
5173 | @{ | |
5174 | output : | |
5175 | @{ | |
5176 | file1(.text) | |
5177 | . = . + 1000; | |
5178 | file2(.text) | |
5179 | . += 1000; | |
5180 | file3(.text) | |
563e308f | 5181 | @} = 0x12345678; |
252b5132 RH |
5182 | @} |
5183 | @end smallexample | |
5184 | @noindent | |
5185 | In the previous example, the @samp{.text} section from @file{file1} is | |
5186 | located at the beginning of the output section @samp{output}. It is | |
5187 | followed by a 1000 byte gap. Then the @samp{.text} section from | |
5188 | @file{file2} appears, also with a 1000 byte gap following before the | |
563e308f | 5189 | @samp{.text} section from @file{file3}. The notation @samp{= 0x12345678} |
252b5132 RH |
5190 | specifies what data to write in the gaps (@pxref{Output Section Fill}). |
5191 | ||
5c6bbab8 NC |
5192 | @cindex dot inside sections |
5193 | Note: @code{.} actually refers to the byte offset from the start of the | |
5194 | current containing object. Normally this is the @code{SECTIONS} | |
69da35b5 | 5195 | statement, whose start address is 0, hence @code{.} can be used as an |
5c6bbab8 NC |
5196 | absolute address. If @code{.} is used inside a section description |
5197 | however, it refers to the byte offset from the start of that section, | |
5198 | not an absolute address. Thus in a script like this: | |
5199 | ||
5200 | @smallexample | |
5201 | SECTIONS | |
5202 | @{ | |
5203 | . = 0x100 | |
5204 | .text: @{ | |
5205 | *(.text) | |
5206 | . = 0x200 | |
5207 | @} | |
5208 | . = 0x500 | |
5209 | .data: @{ | |
5210 | *(.data) | |
5211 | . += 0x600 | |
5212 | @} | |
5213 | @} | |
5214 | @end smallexample | |
5215 | ||
5216 | The @samp{.text} section will be assigned a starting address of 0x100 | |
5217 | and a size of exactly 0x200 bytes, even if there is not enough data in | |
5218 | the @samp{.text} input sections to fill this area. (If there is too | |
5219 | much data, an error will be produced because this would be an attempt to | |
5220 | move @code{.} backwards). The @samp{.data} section will start at 0x500 | |
5221 | and it will have an extra 0x600 bytes worth of space after the end of | |
5222 | the values from the @samp{.data} input sections and before the end of | |
5223 | the @samp{.data} output section itself. | |
5224 | ||
b5666f2f AM |
5225 | @cindex dot outside sections |
5226 | Setting symbols to the value of the location counter outside of an | |
5227 | output section statement can result in unexpected values if the linker | |
5228 | needs to place orphan sections. For example, given the following: | |
5229 | ||
5230 | @smallexample | |
5231 | SECTIONS | |
5232 | @{ | |
5233 | start_of_text = . ; | |
5234 | .text: @{ *(.text) @} | |
5235 | end_of_text = . ; | |
5236 | ||
5237 | start_of_data = . ; | |
5238 | .data: @{ *(.data) @} | |
5239 | end_of_data = . ; | |
5240 | @} | |
5241 | @end smallexample | |
5242 | ||
5243 | If the linker needs to place some input section, e.g. @code{.rodata}, | |
5244 | not mentioned in the script, it might choose to place that section | |
5245 | between @code{.text} and @code{.data}. You might think the linker | |
5246 | should place @code{.rodata} on the blank line in the above script, but | |
5247 | blank lines are of no particular significance to the linker. As well, | |
5248 | the linker doesn't associate the above symbol names with their | |
5249 | sections. Instead, it assumes that all assignments or other | |
5250 | statements belong to the previous output section, except for the | |
5251 | special case of an assignment to @code{.}. I.e., the linker will | |
5252 | place the orphan @code{.rodata} section as if the script was written | |
5253 | as follows: | |
5254 | ||
5255 | @smallexample | |
5256 | SECTIONS | |
5257 | @{ | |
5258 | start_of_text = . ; | |
5259 | .text: @{ *(.text) @} | |
5260 | end_of_text = . ; | |
5261 | ||
5262 | start_of_data = . ; | |
5263 | .rodata: @{ *(.rodata) @} | |
5264 | .data: @{ *(.data) @} | |
5265 | end_of_data = . ; | |
5266 | @} | |
5267 | @end smallexample | |
5268 | ||
5269 | This may or may not be the script author's intention for the value of | |
5270 | @code{start_of_data}. One way to influence the orphan section | |
5271 | placement is to assign the location counter to itself, as the linker | |
5272 | assumes that an assignment to @code{.} is setting the start address of | |
5273 | a following output section and thus should be grouped with that | |
5274 | section. So you could write: | |
5275 | ||
5276 | @smallexample | |
5277 | SECTIONS | |
5278 | @{ | |
5279 | start_of_text = . ; | |
5280 | .text: @{ *(.text) @} | |
5281 | end_of_text = . ; | |
5282 | ||
5283 | . = . ; | |
5284 | start_of_data = . ; | |
5285 | .data: @{ *(.data) @} | |
5286 | end_of_data = . ; | |
5287 | @} | |
5288 | @end smallexample | |
5289 | ||
5290 | Now, the orphan @code{.rodata} section will be placed between | |
5291 | @code{end_of_text} and @code{start_of_data}. | |
5292 | ||
252b5132 RH |
5293 | @need 2000 |
5294 | @node Operators | |
5295 | @subsection Operators | |
5296 | @cindex operators for arithmetic | |
5297 | @cindex arithmetic operators | |
5298 | @cindex precedence in expressions | |
5299 | The linker recognizes the standard C set of arithmetic operators, with | |
5300 | the standard bindings and precedence levels: | |
5301 | @c TEXI2ROFF-KILL | |
36f63dca | 5302 | @ifnottex |
252b5132 RH |
5303 | @c END TEXI2ROFF-KILL |
5304 | @smallexample | |
5305 | precedence associativity Operators Notes | |
5306 | (highest) | |
5307 | 1 left ! - ~ (1) | |
5308 | 2 left * / % | |
5309 | 3 left + - | |
5310 | 4 left >> << | |
5311 | 5 left == != > < <= >= | |
5312 | 6 left & | |
5313 | 7 left | | |
5314 | 8 left && | |
5315 | 9 left || | |
5316 | 10 right ? : | |
5317 | 11 right &= += -= *= /= (2) | |
5318 | (lowest) | |
5319 | @end smallexample | |
5320 | Notes: | |
a1ab1d2a | 5321 | (1) Prefix operators |
252b5132 RH |
5322 | (2) @xref{Assignments}. |
5323 | @c TEXI2ROFF-KILL | |
36f63dca | 5324 | @end ifnottex |
252b5132 RH |
5325 | @tex |
5326 | \vskip \baselineskip | |
5327 | %"lispnarrowing" is the extra indent used generally for smallexample | |
5328 | \hskip\lispnarrowing\vbox{\offinterlineskip | |
5329 | \hrule | |
5330 | \halign | |
5331 | {\vrule#&\strut\hfil\ #\ \hfil&\vrule#&\strut\hfil\ #\ \hfil&\vrule#&\strut\hfil\ {\tt #}\ \hfil&\vrule#\cr | |
5332 | height2pt&\omit&&\omit&&\omit&\cr | |
5333 | &Precedence&& Associativity &&{\rm Operators}&\cr | |
5334 | height2pt&\omit&&\omit&&\omit&\cr | |
5335 | \noalign{\hrule} | |
5336 | height2pt&\omit&&\omit&&\omit&\cr | |
5337 | &highest&&&&&\cr | |
5338 | % '176 is tilde, '~' in tt font | |
a1ab1d2a | 5339 | &1&&left&&\qquad- \char'176\ !\qquad\dag&\cr |
252b5132 RH |
5340 | &2&&left&&* / \%&\cr |
5341 | &3&&left&&+ -&\cr | |
5342 | &4&&left&&>> <<&\cr | |
5343 | &5&&left&&== != > < <= >=&\cr | |
5344 | &6&&left&&\&&\cr | |
5345 | &7&&left&&|&\cr | |
5346 | &8&&left&&{\&\&}&\cr | |
5347 | &9&&left&&||&\cr | |
5348 | &10&&right&&? :&\cr | |
5349 | &11&&right&&\qquad\&= += -= *= /=\qquad\ddag&\cr | |
5350 | &lowest&&&&&\cr | |
5351 | height2pt&\omit&&\omit&&\omit&\cr} | |
5352 | \hrule} | |
5353 | @end tex | |
5354 | @iftex | |
5355 | { | |
5356 | @obeylines@parskip=0pt@parindent=0pt | |
5357 | @dag@quad Prefix operators. | |
5358 | @ddag@quad @xref{Assignments}. | |
5359 | } | |
5360 | @end iftex | |
5361 | @c END TEXI2ROFF-KILL | |
5362 | ||
5363 | @node Evaluation | |
5364 | @subsection Evaluation | |
5365 | @cindex lazy evaluation | |
5366 | @cindex expression evaluation order | |
5367 | The linker evaluates expressions lazily. It only computes the value of | |
5368 | an expression when absolutely necessary. | |
5369 | ||
5370 | The linker needs some information, such as the value of the start | |
5371 | address of the first section, and the origins and lengths of memory | |
5372 | regions, in order to do any linking at all. These values are computed | |
5373 | as soon as possible when the linker reads in the linker script. | |
5374 | ||
5375 | However, other values (such as symbol values) are not known or needed | |
5376 | until after storage allocation. Such values are evaluated later, when | |
5377 | other information (such as the sizes of output sections) is available | |
5378 | for use in the symbol assignment expression. | |
5379 | ||
5380 | The sizes of sections cannot be known until after allocation, so | |
5381 | assignments dependent upon these are not performed until after | |
5382 | allocation. | |
5383 | ||
5384 | Some expressions, such as those depending upon the location counter | |
5385 | @samp{.}, must be evaluated during section allocation. | |
5386 | ||
5387 | If the result of an expression is required, but the value is not | |
5388 | available, then an error results. For example, a script like the | |
5389 | following | |
5390 | @smallexample | |
5391 | @group | |
5392 | SECTIONS | |
5393 | @{ | |
a1ab1d2a | 5394 | .text 9+this_isnt_constant : |
252b5132 RH |
5395 | @{ *(.text) @} |
5396 | @} | |
5397 | @end group | |
5398 | @end smallexample | |
5399 | @noindent | |
5400 | will cause the error message @samp{non constant expression for initial | |
5401 | address}. | |
5402 | ||
5403 | @node Expression Section | |
5404 | @subsection The Section of an Expression | |
5405 | @cindex expression sections | |
5406 | @cindex absolute expressions | |
5407 | @cindex relative expressions | |
5408 | @cindex absolute and relocatable symbols | |
5409 | @cindex relocatable and absolute symbols | |
5410 | @cindex symbols, relocatable and absolute | |
5411 | When the linker evaluates an expression, the result is either absolute | |
5412 | or relative to some section. A relative expression is expressed as a | |
5413 | fixed offset from the base of a section. | |
5414 | ||
5415 | The position of the expression within the linker script determines | |
5416 | whether it is absolute or relative. An expression which appears within | |
5417 | an output section definition is relative to the base of the output | |
5418 | section. An expression which appears elsewhere will be absolute. | |
5419 | ||
5420 | A symbol set to a relative expression will be relocatable if you request | |
5421 | relocatable output using the @samp{-r} option. That means that a | |
5422 | further link operation may change the value of the symbol. The symbol's | |
5423 | section will be the section of the relative expression. | |
5424 | ||
5425 | A symbol set to an absolute expression will retain the same value | |
5426 | through any further link operation. The symbol will be absolute, and | |
5427 | will not have any particular associated section. | |
5428 | ||
5429 | You can use the builtin function @code{ABSOLUTE} to force an expression | |
5430 | to be absolute when it would otherwise be relative. For example, to | |
5431 | create an absolute symbol set to the address of the end of the output | |
5432 | section @samp{.data}: | |
5433 | @smallexample | |
5434 | SECTIONS | |
5435 | @{ | |
5436 | .data : @{ *(.data) _edata = ABSOLUTE(.); @} | |
5437 | @} | |
5438 | @end smallexample | |
5439 | @noindent | |
5440 | If @samp{ABSOLUTE} were not used, @samp{_edata} would be relative to the | |
5441 | @samp{.data} section. | |
5442 | ||
5443 | @node Builtin Functions | |
5444 | @subsection Builtin Functions | |
5445 | @cindex functions in expressions | |
5446 | The linker script language includes a number of builtin functions for | |
5447 | use in linker script expressions. | |
5448 | ||
5449 | @table @code | |
5450 | @item ABSOLUTE(@var{exp}) | |
5451 | @kindex ABSOLUTE(@var{exp}) | |
5452 | @cindex expression, absolute | |
5453 | Return the absolute (non-relocatable, as opposed to non-negative) value | |
5454 | of the expression @var{exp}. Primarily useful to assign an absolute | |
5455 | value to a symbol within a section definition, where symbol values are | |
5456 | normally section relative. @xref{Expression Section}. | |
5457 | ||
5458 | @item ADDR(@var{section}) | |
5459 | @kindex ADDR(@var{section}) | |
5460 | @cindex section address in expression | |
5461 | Return the absolute address (the VMA) of the named @var{section}. Your | |
5462 | script must previously have defined the location of that section. In | |
5463 | the following example, @code{symbol_1} and @code{symbol_2} are assigned | |
5464 | identical values: | |
5465 | @smallexample | |
5466 | @group | |
5467 | SECTIONS @{ @dots{} | |
5468 | .output1 : | |
a1ab1d2a | 5469 | @{ |
252b5132 RH |
5470 | start_of_output_1 = ABSOLUTE(.); |
5471 | @dots{} | |
5472 | @} | |
5473 | .output : | |
5474 | @{ | |
5475 | symbol_1 = ADDR(.output1); | |
5476 | symbol_2 = start_of_output_1; | |
5477 | @} | |
5478 | @dots{} @} | |
5479 | @end group | |
5480 | @end smallexample | |
5481 | ||
876f4090 NS |
5482 | @item ALIGN(@var{align}) |
5483 | @itemx ALIGN(@var{exp},@var{align}) | |
5484 | @kindex ALIGN(@var{align}) | |
5485 | @kindex ALIGN(@var{exp},@var{align}) | |
252b5132 RH |
5486 | @cindex round up location counter |
5487 | @cindex align location counter | |
876f4090 NS |
5488 | @cindex round up expression |
5489 | @cindex align expression | |
5490 | Return the location counter (@code{.}) or arbitrary expression aligned | |
5491 | to the next @var{align} boundary. The single operand @code{ALIGN} | |
5492 | doesn't change the value of the location counter---it just does | |
5493 | arithmetic on it. The two operand @code{ALIGN} allows an arbitrary | |
5494 | expression to be aligned upwards (@code{ALIGN(@var{align})} is | |
5495 | equivalent to @code{ALIGN(., @var{align})}). | |
5496 | ||
5497 | Here is an example which aligns the output @code{.data} section to the | |
5498 | next @code{0x2000} byte boundary after the preceding section and sets a | |
5499 | variable within the section to the next @code{0x8000} boundary after the | |
5500 | input sections: | |
252b5132 RH |
5501 | @smallexample |
5502 | @group | |
5503 | SECTIONS @{ @dots{} | |
5504 | .data ALIGN(0x2000): @{ | |
5505 | *(.data) | |
5506 | variable = ALIGN(0x8000); | |
5507 | @} | |
5508 | @dots{} @} | |
5509 | @end group | |
5510 | @end smallexample | |
5511 | @noindent | |
5512 | The first use of @code{ALIGN} in this example specifies the location of | |
5513 | a section because it is used as the optional @var{address} attribute of | |
5514 | a section definition (@pxref{Output Section Address}). The second use | |
5515 | of @code{ALIGN} is used to defines the value of a symbol. | |
5516 | ||
5517 | The builtin function @code{NEXT} is closely related to @code{ALIGN}. | |
5518 | ||
362c1d1a NS |
5519 | @item ALIGNOF(@var{section}) |
5520 | @kindex ALIGNOF(@var{section}) | |
5521 | @cindex section alignment | |
5522 | Return the alignment in bytes of the named @var{section}, if that section has | |
5523 | been allocated. If the section has not been allocated when this is | |
5524 | evaluated, the linker will report an error. In the following example, | |
5525 | the alignment of the @code{.output} section is stored as the first | |
5526 | value in that section. | |
5527 | @smallexample | |
5528 | @group | |
5529 | SECTIONS@{ @dots{} | |
5530 | .output @{ | |
5531 | LONG (ALIGNOF (.output)) | |
5532 | @dots{} | |
5533 | @} | |
5534 | @dots{} @} | |
5535 | @end group | |
5536 | @end smallexample | |
5537 | ||
252b5132 RH |
5538 | @item BLOCK(@var{exp}) |
5539 | @kindex BLOCK(@var{exp}) | |
5540 | This is a synonym for @code{ALIGN}, for compatibility with older linker | |
5541 | scripts. It is most often seen when setting the address of an output | |
5542 | section. | |
5543 | ||
2d20f7bf JJ |
5544 | @item DATA_SEGMENT_ALIGN(@var{maxpagesize}, @var{commonpagesize}) |
5545 | @kindex DATA_SEGMENT_ALIGN(@var{maxpagesize}, @var{commonpagesize}) | |
5546 | This is equivalent to either | |
5547 | @smallexample | |
5548 | (ALIGN(@var{maxpagesize}) + (. & (@var{maxpagesize} - 1))) | |
5549 | @end smallexample | |
5550 | or | |
5551 | @smallexample | |
5552 | (ALIGN(@var{maxpagesize}) + (. & (@var{maxpagesize} - @var{commonpagesize}))) | |
5553 | @end smallexample | |
5554 | @noindent | |
5555 | depending on whether the latter uses fewer @var{commonpagesize} sized pages | |
5556 | for the data segment (area between the result of this expression and | |
5557 | @code{DATA_SEGMENT_END}) than the former or not. | |
5558 | If the latter form is used, it means @var{commonpagesize} bytes of runtime | |
5559 | memory will be saved at the expense of up to @var{commonpagesize} wasted | |
5560 | bytes in the on-disk file. | |
5561 | ||
5562 | This expression can only be used directly in @code{SECTIONS} commands, not in | |
5563 | any output section descriptions and only once in the linker script. | |
5564 | @var{commonpagesize} should be less or equal to @var{maxpagesize} and should | |
5565 | be the system page size the object wants to be optimized for (while still | |
5566 | working on system page sizes up to @var{maxpagesize}). | |
5567 | ||
5568 | @noindent | |
5569 | Example: | |
5570 | @smallexample | |
5571 | . = DATA_SEGMENT_ALIGN(0x10000, 0x2000); | |
5572 | @end smallexample | |
5573 | ||
5574 | @item DATA_SEGMENT_END(@var{exp}) | |
5575 | @kindex DATA_SEGMENT_END(@var{exp}) | |
5576 | This defines the end of data segment for @code{DATA_SEGMENT_ALIGN} | |
5577 | evaluation purposes. | |
5578 | ||
5579 | @smallexample | |
5580 | . = DATA_SEGMENT_END(.); | |
5581 | @end smallexample | |
5582 | ||
a4f5ad88 JJ |
5583 | @item DATA_SEGMENT_RELRO_END(@var{offset}, @var{exp}) |
5584 | @kindex DATA_SEGMENT_RELRO_END(@var{offset}, @var{exp}) | |
5585 | This defines the end of the @code{PT_GNU_RELRO} segment when | |
5586 | @samp{-z relro} option is used. Second argument is returned. | |
5587 | When @samp{-z relro} option is not present, @code{DATA_SEGMENT_RELRO_END} | |
5588 | does nothing, otherwise @code{DATA_SEGMENT_ALIGN} is padded so that | |
5589 | @var{exp} + @var{offset} is aligned to the most commonly used page | |
5590 | boundary for particular target. If present in the linker script, | |
5591 | it must always come in between @code{DATA_SEGMENT_ALIGN} and | |
5592 | @code{DATA_SEGMENT_END}. | |
5593 | ||
5594 | @smallexample | |
5595 | . = DATA_SEGMENT_RELRO_END(24, .); | |
5596 | @end smallexample | |
5597 | ||
252b5132 RH |
5598 | @item DEFINED(@var{symbol}) |
5599 | @kindex DEFINED(@var{symbol}) | |
5600 | @cindex symbol defaults | |
5601 | Return 1 if @var{symbol} is in the linker global symbol table and is | |
420e579c HPN |
5602 | defined before the statement using DEFINED in the script, otherwise |
5603 | return 0. You can use this function to provide | |
252b5132 RH |
5604 | default values for symbols. For example, the following script fragment |
5605 | shows how to set a global symbol @samp{begin} to the first location in | |
5606 | the @samp{.text} section---but if a symbol called @samp{begin} already | |
5607 | existed, its value is preserved: | |
5608 | ||
5609 | @smallexample | |
5610 | @group | |
5611 | SECTIONS @{ @dots{} | |
5612 | .text : @{ | |
5613 | begin = DEFINED(begin) ? begin : . ; | |
5614 | @dots{} | |
5615 | @} | |
5616 | @dots{} | |
5617 | @} | |
5618 | @end group | |
5619 | @end smallexample | |
5620 | ||
3ec57632 NC |
5621 | @item LENGTH(@var{memory}) |
5622 | @kindex LENGTH(@var{memory}) | |
5623 | Return the length of the memory region named @var{memory}. | |
5624 | ||
252b5132 RH |
5625 | @item LOADADDR(@var{section}) |
5626 | @kindex LOADADDR(@var{section}) | |
5627 | @cindex section load address in expression | |
5628 | Return the absolute LMA of the named @var{section}. This is normally | |
5629 | the same as @code{ADDR}, but it may be different if the @code{AT} | |
5630 | attribute is used in the output section definition (@pxref{Output | |
5631 | Section LMA}). | |
5632 | ||
5633 | @kindex MAX | |
5634 | @item MAX(@var{exp1}, @var{exp2}) | |
5635 | Returns the maximum of @var{exp1} and @var{exp2}. | |
5636 | ||
5637 | @kindex MIN | |
5638 | @item MIN(@var{exp1}, @var{exp2}) | |
5639 | Returns the minimum of @var{exp1} and @var{exp2}. | |
5640 | ||
5641 | @item NEXT(@var{exp}) | |
5642 | @kindex NEXT(@var{exp}) | |
5643 | @cindex unallocated address, next | |
5644 | Return the next unallocated address that is a multiple of @var{exp}. | |
5645 | This function is closely related to @code{ALIGN(@var{exp})}; unless you | |
5646 | use the @code{MEMORY} command to define discontinuous memory for the | |
5647 | output file, the two functions are equivalent. | |
5648 | ||
3ec57632 NC |
5649 | @item ORIGIN(@var{memory}) |
5650 | @kindex ORIGIN(@var{memory}) | |
5651 | Return the origin of the memory region named @var{memory}. | |
5652 | ||
ba916c8a MM |
5653 | @item SEGMENT_START(@var{segment}, @var{default}) |
5654 | @kindex SEGMENT_START(@var{segment}, @var{default}) | |
5655 | Return the base address of the named @var{segment}. If an explicit | |
5656 | value has been given for this segment (with a command-line @samp{-T} | |
5657 | option) that value will be returned; otherwise the value will be | |
5658 | @var{default}. At present, the @samp{-T} command-line option can only | |
5659 | be used to set the base address for the ``text'', ``data'', and | |
5660 | ``bss'' sections, but you use @code{SEGMENT_START} with any segment | |
5661 | name. | |
5662 | ||
252b5132 RH |
5663 | @item SIZEOF(@var{section}) |
5664 | @kindex SIZEOF(@var{section}) | |
5665 | @cindex section size | |
5666 | Return the size in bytes of the named @var{section}, if that section has | |
5667 | been allocated. If the section has not been allocated when this is | |
5668 | evaluated, the linker will report an error. In the following example, | |
5669 | @code{symbol_1} and @code{symbol_2} are assigned identical values: | |
5670 | @smallexample | |
5671 | @group | |
5672 | SECTIONS@{ @dots{} | |
5673 | .output @{ | |
5674 | .start = . ; | |
5675 | @dots{} | |
5676 | .end = . ; | |
5677 | @} | |
5678 | symbol_1 = .end - .start ; | |
5679 | symbol_2 = SIZEOF(.output); | |
5680 | @dots{} @} | |
5681 | @end group | |
5682 | @end smallexample | |
5683 | ||
5684 | @item SIZEOF_HEADERS | |
5685 | @itemx sizeof_headers | |
5686 | @kindex SIZEOF_HEADERS | |
5687 | @cindex header size | |
5688 | Return the size in bytes of the output file's headers. This is | |
5689 | information which appears at the start of the output file. You can use | |
5690 | this number when setting the start address of the first section, if you | |
5691 | choose, to facilitate paging. | |
5692 | ||
5693 | @cindex not enough room for program headers | |
5694 | @cindex program headers, not enough room | |
5695 | When producing an ELF output file, if the linker script uses the | |
5696 | @code{SIZEOF_HEADERS} builtin function, the linker must compute the | |
5697 | number of program headers before it has determined all the section | |
5698 | addresses and sizes. If the linker later discovers that it needs | |
5699 | additional program headers, it will report an error @samp{not enough | |
5700 | room for program headers}. To avoid this error, you must avoid using | |
5701 | the @code{SIZEOF_HEADERS} function, or you must rework your linker | |
5702 | script to avoid forcing the linker to use additional program headers, or | |
5703 | you must define the program headers yourself using the @code{PHDRS} | |
5704 | command (@pxref{PHDRS}). | |
5705 | @end table | |
5706 | ||
5707 | @node Implicit Linker Scripts | |
5708 | @section Implicit Linker Scripts | |
5709 | @cindex implicit linker scripts | |
5710 | If you specify a linker input file which the linker can not recognize as | |
5711 | an object file or an archive file, it will try to read the file as a | |
5712 | linker script. If the file can not be parsed as a linker script, the | |
5713 | linker will report an error. | |
5714 | ||
5715 | An implicit linker script will not replace the default linker script. | |
5716 | ||
5717 | Typically an implicit linker script would contain only symbol | |
5718 | assignments, or the @code{INPUT}, @code{GROUP}, or @code{VERSION} | |
5719 | commands. | |
5720 | ||
5721 | Any input files read because of an implicit linker script will be read | |
5722 | at the position in the command line where the implicit linker script was | |
5723 | read. This can affect archive searching. | |
5724 | ||
5725 | @ifset GENERIC | |
5726 | @node Machine Dependent | |
5727 | @chapter Machine Dependent Features | |
5728 | ||
5729 | @cindex machine dependencies | |
ff5dcc92 SC |
5730 | @command{ld} has additional features on some platforms; the following |
5731 | sections describe them. Machines where @command{ld} has no additional | |
252b5132 RH |
5732 | functionality are not listed. |
5733 | ||
5734 | @menu | |
36f63dca NC |
5735 | @ifset H8300 |
5736 | * H8/300:: @command{ld} and the H8/300 | |
5737 | @end ifset | |
5738 | @ifset I960 | |
5739 | * i960:: @command{ld} and the Intel 960 family | |
5740 | @end ifset | |
5741 | @ifset ARM | |
5742 | * ARM:: @command{ld} and the ARM family | |
5743 | @end ifset | |
5744 | @ifset HPPA | |
5745 | * HPPA ELF32:: @command{ld} and HPPA 32-bit ELF | |
5746 | @end ifset | |
7fb9f789 NC |
5747 | @ifset M68K |
5748 | * M68K:: @command{ld} and the Motorola 68K family | |
5749 | @end ifset | |
3c3bdf30 | 5750 | @ifset MMIX |
36f63dca | 5751 | * MMIX:: @command{ld} and MMIX |
3c3bdf30 | 5752 | @end ifset |
2469cfa2 | 5753 | @ifset MSP430 |
36f63dca | 5754 | * MSP430:: @command{ld} and MSP430 |
2469cfa2 | 5755 | @end ifset |
93fd0973 SC |
5756 | @ifset M68HC11 |
5757 | * M68HC11/68HC12:: @code{ld} and the Motorola 68HC11 and 68HC12 families | |
5758 | @end ifset | |
2a60a7a8 AM |
5759 | @ifset POWERPC |
5760 | * PowerPC ELF32:: @command{ld} and PowerPC 32-bit ELF Support | |
5761 | @end ifset | |
5762 | @ifset POWERPC64 | |
5763 | * PowerPC64 ELF64:: @command{ld} and PowerPC64 64-bit ELF Support | |
5764 | @end ifset | |
49fa1e15 AM |
5765 | @ifset SPU |
5766 | * SPU ELF:: @command{ld} and SPU ELF Support | |
5767 | @end ifset | |
74459f0e | 5768 | @ifset TICOFF |
ff5dcc92 | 5769 | * TI COFF:: @command{ld} and TI COFF |
74459f0e | 5770 | @end ifset |
2ca22b03 NC |
5771 | @ifset WIN32 |
5772 | * WIN32:: @command{ld} and WIN32 (cygwin/mingw) | |
5773 | @end ifset | |
e0001a05 NC |
5774 | @ifset XTENSA |
5775 | * Xtensa:: @command{ld} and Xtensa Processors | |
5776 | @end ifset | |
252b5132 RH |
5777 | @end menu |
5778 | @end ifset | |
5779 | ||
252b5132 RH |
5780 | @ifset H8300 |
5781 | @ifclear GENERIC | |
5782 | @raisesections | |
5783 | @end ifclear | |
5784 | ||
5785 | @node H8/300 | |
ff5dcc92 | 5786 | @section @command{ld} and the H8/300 |
252b5132 RH |
5787 | |
5788 | @cindex H8/300 support | |
ff5dcc92 | 5789 | For the H8/300, @command{ld} can perform these global optimizations when |
252b5132 RH |
5790 | you specify the @samp{--relax} command-line option. |
5791 | ||
5792 | @table @emph | |
5793 | @cindex relaxing on H8/300 | |
5794 | @item relaxing address modes | |
ff5dcc92 | 5795 | @command{ld} finds all @code{jsr} and @code{jmp} instructions whose |
252b5132 RH |
5796 | targets are within eight bits, and turns them into eight-bit |
5797 | program-counter relative @code{bsr} and @code{bra} instructions, | |
5798 | respectively. | |
5799 | ||
5800 | @cindex synthesizing on H8/300 | |
5801 | @item synthesizing instructions | |
5802 | @c FIXME: specifically mov.b, or any mov instructions really? | |
ff5dcc92 | 5803 | @command{ld} finds all @code{mov.b} instructions which use the |
252b5132 RH |
5804 | sixteen-bit absolute address form, but refer to the top |
5805 | page of memory, and changes them to use the eight-bit address form. | |
5806 | (That is: the linker turns @samp{mov.b @code{@@}@var{aa}:16} into | |
5807 | @samp{mov.b @code{@@}@var{aa}:8} whenever the address @var{aa} is in the | |
5808 | top page of memory). | |
1502569c NC |
5809 | |
5810 | @item bit manipulation instructions | |
c0065db7 | 5811 | @command{ld} finds all bit manipulation instructions like @code{band, bclr, |
1502569c | 5812 | biand, bild, bior, bist, bixor, bld, bnot, bor, bset, bst, btst, bxor} |
c0065db7 | 5813 | which use 32 bit and 16 bit absolute address form, but refer to the top |
1502569c NC |
5814 | page of memory, and changes them to use the 8 bit address form. |
5815 | (That is: the linker turns @samp{bset #xx:3,@code{@@}@var{aa}:32} into | |
c0065db7 | 5816 | @samp{bset #xx:3,@code{@@}@var{aa}:8} whenever the address @var{aa} is in |
1502569c NC |
5817 | the top page of memory). |
5818 | ||
5819 | @item system control instructions | |
c0065db7 RM |
5820 | @command{ld} finds all @code{ldc.w, stc.w} instructions which use the |
5821 | 32 bit absolute address form, but refer to the top page of memory, and | |
1502569c NC |
5822 | changes them to use 16 bit address form. |
5823 | (That is: the linker turns @samp{ldc.w @code{@@}@var{aa}:32,ccr} into | |
c0065db7 | 5824 | @samp{ldc.w @code{@@}@var{aa}:16,ccr} whenever the address @var{aa} is in |
1502569c | 5825 | the top page of memory). |
252b5132 RH |
5826 | @end table |
5827 | ||
5828 | @ifclear GENERIC | |
5829 | @lowersections | |
5830 | @end ifclear | |
5831 | @end ifset | |
5832 | ||
36f63dca | 5833 | @ifclear GENERIC |
c2dcd04e | 5834 | @ifset Renesas |
36f63dca | 5835 | @c This stuff is pointless to say unless you're especially concerned |
c2dcd04e NC |
5836 | @c with Renesas chips; don't enable it for generic case, please. |
5837 | @node Renesas | |
5838 | @chapter @command{ld} and Other Renesas Chips | |
36f63dca | 5839 | |
c2dcd04e NC |
5840 | @command{ld} also supports the Renesas (formerly Hitachi) H8/300H, |
5841 | H8/500, and SH chips. No special features, commands, or command-line | |
5842 | options are required for these chips. | |
36f63dca NC |
5843 | @end ifset |
5844 | @end ifclear | |
5845 | ||
5846 | @ifset I960 | |
5847 | @ifclear GENERIC | |
5848 | @raisesections | |
5849 | @end ifclear | |
5850 | ||
5851 | @node i960 | |
5852 | @section @command{ld} and the Intel 960 Family | |
5853 | ||
5854 | @cindex i960 support | |
5855 | ||
5856 | You can use the @samp{-A@var{architecture}} command line option to | |
5857 | specify one of the two-letter names identifying members of the 960 | |
5858 | family; the option specifies the desired output target, and warns of any | |
5859 | incompatible instructions in the input files. It also modifies the | |
5860 | linker's search strategy for archive libraries, to support the use of | |
5861 | libraries specific to each particular architecture, by including in the | |
5862 | search loop names suffixed with the string identifying the architecture. | |
5863 | ||
5864 | For example, if your @command{ld} command line included @w{@samp{-ACA}} as | |
5865 | well as @w{@samp{-ltry}}, the linker would look (in its built-in search | |
5866 | paths, and in any paths you specify with @samp{-L}) for a library with | |
5867 | the names | |
5868 | ||
5869 | @smallexample | |
5870 | @group | |
5871 | try | |
5872 | libtry.a | |
5873 | tryca | |
5874 | libtryca.a | |
5875 | @end group | |
5876 | @end smallexample | |
5877 | ||
5878 | @noindent | |
5879 | The first two possibilities would be considered in any event; the last | |
5880 | two are due to the use of @w{@samp{-ACA}}. | |
5881 | ||
5882 | You can meaningfully use @samp{-A} more than once on a command line, since | |
5883 | the 960 architecture family allows combination of target architectures; each | |
5884 | use will add another pair of name variants to search for when @w{@samp{-l}} | |
5885 | specifies a library. | |
5886 | ||
5887 | @cindex @option{--relax} on i960 | |
5888 | @cindex relaxing on i960 | |
5889 | @command{ld} supports the @samp{--relax} option for the i960 family. If | |
5890 | you specify @samp{--relax}, @command{ld} finds all @code{balx} and | |
5891 | @code{calx} instructions whose targets are within 24 bits, and turns | |
5892 | them into 24-bit program-counter relative @code{bal} and @code{cal} | |
5893 | instructions, respectively. @command{ld} also turns @code{cal} | |
5894 | instructions into @code{bal} instructions when it determines that the | |
5895 | target subroutine is a leaf routine (that is, the target subroutine does | |
5896 | not itself call any subroutines). | |
5897 | ||
48229727 JB |
5898 | @cindex Cortex-A8 erratum workaround |
5899 | @kindex --fix-cortex-a8 | |
5900 | @kindex --no-fix-cortex-a8 | |
5901 | 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}. | |
5902 | ||
5903 | The erratum only affects Thumb-2 code. Please contact ARM for further details. | |
5904 | ||
36f63dca NC |
5905 | @ifclear GENERIC |
5906 | @lowersections | |
5907 | @end ifclear | |
5908 | @end ifset | |
5909 | ||
5910 | @ifset ARM | |
5911 | @ifclear GENERIC | |
5912 | @raisesections | |
5913 | @end ifclear | |
5914 | ||
93fd0973 SC |
5915 | @ifset M68HC11 |
5916 | @ifclear GENERIC | |
5917 | @raisesections | |
5918 | @end ifclear | |
5919 | ||
5920 | @node M68HC11/68HC12 | |
5921 | @section @command{ld} and the Motorola 68HC11 and 68HC12 families | |
5922 | ||
5923 | @cindex M68HC11 and 68HC12 support | |
5924 | ||
5925 | @subsection Linker Relaxation | |
5926 | ||
5927 | For the Motorola 68HC11, @command{ld} can perform these global | |
5928 | optimizations when you specify the @samp{--relax} command-line option. | |
5929 | ||
5930 | @table @emph | |
5931 | @cindex relaxing on M68HC11 | |
5932 | @item relaxing address modes | |
5933 | @command{ld} finds all @code{jsr} and @code{jmp} instructions whose | |
5934 | targets are within eight bits, and turns them into eight-bit | |
5935 | program-counter relative @code{bsr} and @code{bra} instructions, | |
5936 | respectively. | |
5937 | ||
5938 | @command{ld} also looks at all 16-bit extended addressing modes and | |
5939 | transforms them in a direct addressing mode when the address is in | |
5940 | page 0 (between 0 and 0x0ff). | |
5941 | ||
5942 | @item relaxing gcc instruction group | |
5943 | When @command{gcc} is called with @option{-mrelax}, it can emit group | |
5944 | of instructions that the linker can optimize to use a 68HC11 direct | |
5945 | addressing mode. These instructions consists of @code{bclr} or | |
5946 | @code{bset} instructions. | |
5947 | ||
5948 | @end table | |
5949 | ||
5950 | @subsection Trampoline Generation | |
5951 | ||
5952 | @cindex trampoline generation on M68HC11 | |
5953 | @cindex trampoline generation on M68HC12 | |
5954 | For 68HC11 and 68HC12, @command{ld} can generate trampoline code to | |
5955 | call a far function using a normal @code{jsr} instruction. The linker | |
c0065db7 | 5956 | will also change the relocation to some far function to use the |
93fd0973 SC |
5957 | trampoline address instead of the function address. This is typically the |
5958 | case when a pointer to a function is taken. The pointer will in fact | |
5959 | point to the function trampoline. | |
5960 | ||
5961 | @ifclear GENERIC | |
5962 | @lowersections | |
5963 | @end ifclear | |
5964 | @end ifset | |
5965 | ||
36f63dca | 5966 | @node ARM |
3674e28a | 5967 | @section @command{ld} and the ARM family |
36f63dca NC |
5968 | |
5969 | @cindex ARM interworking support | |
5970 | @kindex --support-old-code | |
5971 | For the ARM, @command{ld} will generate code stubs to allow functions calls | |
b45619c0 | 5972 | between ARM and Thumb code. These stubs only work with code that has |
36f63dca NC |
5973 | been compiled and assembled with the @samp{-mthumb-interwork} command |
5974 | line option. If it is necessary to link with old ARM object files or | |
5975 | libraries, which have not been compiled with the -mthumb-interwork | |
5976 | option then the @samp{--support-old-code} command line switch should be | |
5977 | given to the linker. This will make it generate larger stub functions | |
5978 | which will work with non-interworking aware ARM code. Note, however, | |
5979 | the linker does not support generating stubs for function calls to | |
5980 | non-interworking aware Thumb code. | |
5981 | ||
5982 | @cindex thumb entry point | |
5983 | @cindex entry point, thumb | |
5984 | @kindex --thumb-entry=@var{entry} | |
5985 | The @samp{--thumb-entry} switch is a duplicate of the generic | |
5986 | @samp{--entry} switch, in that it sets the program's starting address. | |
5987 | But it also sets the bottom bit of the address, so that it can be | |
5988 | branched to using a BX instruction, and the program will start | |
5989 | executing in Thumb mode straight away. | |
5990 | ||
ce11ba6c KT |
5991 | @cindex PE import table prefixing |
5992 | @kindex --use-nul-prefixed-import-tables | |
5993 | The @samp{--use-nul-prefixed-import-tables} switch is specifying, that | |
5994 | the import tables idata4 and idata5 have to be generated with a zero | |
5995 | elememt prefix for import libraries. This is the old style to generate | |
5996 | import tables. By default this option is turned off. | |
5997 | ||
e489d0ae PB |
5998 | @cindex BE8 |
5999 | @kindex --be8 | |
6000 | The @samp{--be8} switch instructs @command{ld} to generate BE8 format | |
6001 | executables. This option is only valid when linking big-endian objects. | |
6002 | The resulting image will contain big-endian data and little-endian code. | |
6003 | ||
3674e28a PB |
6004 | @cindex TARGET1 |
6005 | @kindex --target1-rel | |
6006 | @kindex --target1-abs | |
6007 | The @samp{R_ARM_TARGET1} relocation is typically used for entries in the | |
6008 | @samp{.init_array} section. It is interpreted as either @samp{R_ARM_REL32} | |
6009 | or @samp{R_ARM_ABS32}, depending on the target. The @samp{--target1-rel} | |
6010 | and @samp{--target1-abs} switches override the default. | |
6011 | ||
6012 | @cindex TARGET2 | |
6013 | @kindex --target2=@var{type} | |
6014 | The @samp{--target2=type} switch overrides the default definition of the | |
6015 | @samp{R_ARM_TARGET2} relocation. Valid values for @samp{type}, their | |
6016 | meanings, and target defaults are as follows: | |
6017 | @table @samp | |
6018 | @item rel | |
eeac373a PB |
6019 | @samp{R_ARM_REL32} (arm*-*-elf, arm*-*-eabi) |
6020 | @item abs | |
6021 | @samp{R_ARM_ABS32} (arm*-*-symbianelf) | |
3674e28a PB |
6022 | @item got-rel |
6023 | @samp{R_ARM_GOT_PREL} (arm*-*-linux, arm*-*-*bsd) | |
6024 | @end table | |
6025 | ||
319850b4 JB |
6026 | @cindex FIX_V4BX |
6027 | @kindex --fix-v4bx | |
6028 | The @samp{R_ARM_V4BX} relocation (defined by the ARM AAELF | |
6029 | specification) enables objects compiled for the ARMv4 architecture to be | |
6030 | interworking-safe when linked with other objects compiled for ARMv4t, but | |
6031 | also allows pure ARMv4 binaries to be built from the same ARMv4 objects. | |
6032 | ||
6033 | In the latter case, the switch @option{--fix-v4bx} must be passed to the | |
6034 | linker, which causes v4t @code{BX rM} instructions to be rewritten as | |
6035 | @code{MOV PC,rM}, since v4 processors do not have a @code{BX} instruction. | |
6036 | ||
6037 | In the former case, the switch should not be used, and @samp{R_ARM_V4BX} | |
6038 | relocations are ignored. | |
6039 | ||
845b51d6 PB |
6040 | @cindex FIX_V4BX_INTERWORKING |
6041 | @kindex --fix-v4bx-interworking | |
6042 | Replace @code{BX rM} instructions identified by @samp{R_ARM_V4BX} | |
6043 | relocations with a branch to the following veneer: | |
6044 | ||
6045 | @smallexample | |
6046 | TST rM, #1 | |
6047 | MOVEQ PC, rM | |
6048 | BX Rn | |
6049 | @end smallexample | |
6050 | ||
6051 | This allows generation of libraries/applications that work on ARMv4 cores | |
6052 | and are still interworking safe. Note that the above veneer clobbers the | |
6053 | condition flags, so may cause incorrect progrm behavior in rare cases. | |
6054 | ||
33bfe774 JB |
6055 | @cindex USE_BLX |
6056 | @kindex --use-blx | |
6057 | The @samp{--use-blx} switch enables the linker to use ARM/Thumb | |
6058 | BLX instructions (available on ARMv5t and above) in various | |
6059 | situations. Currently it is used to perform calls via the PLT from Thumb | |
6060 | code using BLX rather than using BX and a mode-switching stub before | |
6061 | each PLT entry. This should lead to such calls executing slightly faster. | |
6062 | ||
6063 | This option is enabled implicitly for SymbianOS, so there is no need to | |
6064 | specify it if you are using that target. | |
6065 | ||
c6dd86c6 JB |
6066 | @cindex VFP11_DENORM_FIX |
6067 | @kindex --vfp11-denorm-fix | |
6068 | The @samp{--vfp11-denorm-fix} switch enables a link-time workaround for a | |
6069 | bug in certain VFP11 coprocessor hardware, which sometimes allows | |
6070 | instructions with denorm operands (which must be handled by support code) | |
6071 | to have those operands overwritten by subsequent instructions before | |
6072 | the support code can read the intended values. | |
6073 | ||
6074 | The bug may be avoided in scalar mode if you allow at least one | |
6075 | intervening instruction between a VFP11 instruction which uses a register | |
6076 | and another instruction which writes to the same register, or at least two | |
6077 | intervening instructions if vector mode is in use. The bug only affects | |
6078 | full-compliance floating-point mode: you do not need this workaround if | |
6079 | you are using "runfast" mode. Please contact ARM for further details. | |
6080 | ||
6081 | If you know you are using buggy VFP11 hardware, you can | |
6082 | enable this workaround by specifying the linker option | |
6083 | @samp{--vfp-denorm-fix=scalar} if you are using the VFP11 scalar | |
6084 | mode only, or @samp{--vfp-denorm-fix=vector} if you are using | |
6085 | vector mode (the latter also works for scalar code). The default is | |
6086 | @samp{--vfp-denorm-fix=none}. | |
6087 | ||
6088 | If the workaround is enabled, instructions are scanned for | |
6089 | potentially-troublesome sequences, and a veneer is created for each | |
6090 | such sequence which may trigger the erratum. The veneer consists of the | |
6091 | first instruction of the sequence and a branch back to the subsequent | |
6092 | instruction. The original instruction is then replaced with a branch to | |
6093 | the veneer. The extra cycles required to call and return from the veneer | |
6094 | are sufficient to avoid the erratum in both the scalar and vector cases. | |
6095 | ||
bf21ed78 MS |
6096 | @cindex NO_ENUM_SIZE_WARNING |
6097 | @kindex --no-enum-size-warning | |
726150b7 | 6098 | The @option{--no-enum-size-warning} switch prevents the linker from |
bf21ed78 MS |
6099 | warning when linking object files that specify incompatible EABI |
6100 | enumeration size attributes. For example, with this switch enabled, | |
6101 | linking of an object file using 32-bit enumeration values with another | |
6102 | using enumeration values fitted into the smallest possible space will | |
6103 | not be diagnosed. | |
a9dc9481 JM |
6104 | |
6105 | @cindex NO_WCHAR_SIZE_WARNING | |
6106 | @kindex --no-wchar-size-warning | |
6107 | The @option{--no-wchar-size-warning} switch prevents the linker from | |
6108 | warning when linking object files that specify incompatible EABI | |
6109 | @code{wchar_t} size attributes. For example, with this switch enabled, | |
6110 | linking of an object file using 32-bit @code{wchar_t} values with another | |
6111 | using 16-bit @code{wchar_t} values will not be diagnosed. | |
bf21ed78 | 6112 | |
726150b7 NC |
6113 | @cindex PIC_VENEER |
6114 | @kindex --pic-veneer | |
6115 | The @samp{--pic-veneer} switch makes the linker use PIC sequences for | |
6116 | ARM/Thumb interworking veneers, even if the rest of the binary | |
6117 | is not PIC. This avoids problems on uClinux targets where | |
6118 | @samp{--emit-relocs} is used to generate relocatable binaries. | |
6119 | ||
6120 | @cindex STUB_GROUP_SIZE | |
6121 | @kindex --stub-group-size=@var{N} | |
6122 | The linker will automatically generate and insert small sequences of | |
6123 | code into a linked ARM ELF executable whenever an attempt is made to | |
6124 | perform a function call to a symbol that is too far away. The | |
6125 | placement of these sequences of instructions - called stubs - is | |
6126 | controlled by the command line option @option{--stub-group-size=N}. | |
6127 | The placement is important because a poor choice can create a need for | |
6128 | duplicate stubs, increasing the code sizw. The linker will try to | |
6129 | group stubs together in order to reduce interruptions to the flow of | |
6130 | code, but it needs guidance as to how big these groups should be and | |
6131 | where they should be placed. | |
6132 | ||
6133 | The value of @samp{N}, the parameter to the | |
6134 | @option{--stub-group-size=} option controls where the stub groups are | |
07d72278 | 6135 | placed. If it is negative then all stubs are placed after the first |
726150b7 NC |
6136 | branch that needs them. If it is positive then the stubs can be |
6137 | placed either before or after the branches that need them. If the | |
6138 | value of @samp{N} is 1 (either +1 or -1) then the linker will choose | |
6139 | exactly where to place groups of stubs, using its built in heuristics. | |
6140 | A value of @samp{N} greater than 1 (or smaller than -1) tells the | |
6141 | linker that a single group of stubs can service at most @samp{N} bytes | |
6142 | from the input sections. | |
6143 | ||
6144 | The default, if @option{--stub-group-size=} is not specified, is | |
6145 | @samp{N = +1}. | |
6146 | ||
1a51c1a4 NC |
6147 | Farcalls stubs insertion is fully supported for the ARM-EABI target |
6148 | only, because it relies on object files properties not present | |
6149 | otherwise. | |
6150 | ||
36f63dca NC |
6151 | @ifclear GENERIC |
6152 | @lowersections | |
6153 | @end ifclear | |
6154 | @end ifset | |
6155 | ||
6156 | @ifset HPPA | |
6157 | @ifclear GENERIC | |
6158 | @raisesections | |
6159 | @end ifclear | |
6160 | ||
6161 | @node HPPA ELF32 | |
6162 | @section @command{ld} and HPPA 32-bit ELF Support | |
6163 | @cindex HPPA multiple sub-space stubs | |
6164 | @kindex --multi-subspace | |
6165 | When generating a shared library, @command{ld} will by default generate | |
6166 | import stubs suitable for use with a single sub-space application. | |
6167 | The @samp{--multi-subspace} switch causes @command{ld} to generate export | |
6168 | stubs, and different (larger) import stubs suitable for use with | |
6169 | multiple sub-spaces. | |
6170 | ||
6171 | @cindex HPPA stub grouping | |
6172 | @kindex --stub-group-size=@var{N} | |
6173 | Long branch stubs and import/export stubs are placed by @command{ld} in | |
6174 | stub sections located between groups of input sections. | |
6175 | @samp{--stub-group-size} specifies the maximum size of a group of input | |
6176 | sections handled by one stub section. Since branch offsets are signed, | |
6177 | a stub section may serve two groups of input sections, one group before | |
6178 | the stub section, and one group after it. However, when using | |
6179 | conditional branches that require stubs, it may be better (for branch | |
6180 | prediction) that stub sections only serve one group of input sections. | |
6181 | A negative value for @samp{N} chooses this scheme, ensuring that | |
6182 | branches to stubs always use a negative offset. Two special values of | |
6183 | @samp{N} are recognized, @samp{1} and @samp{-1}. These both instruct | |
6184 | @command{ld} to automatically size input section groups for the branch types | |
6185 | detected, with the same behaviour regarding stub placement as other | |
6186 | positive or negative values of @samp{N} respectively. | |
6187 | ||
6188 | Note that @samp{--stub-group-size} does not split input sections. A | |
6189 | single input section larger than the group size specified will of course | |
6190 | create a larger group (of one section). If input sections are too | |
6191 | large, it may not be possible for a branch to reach its stub. | |
6192 | ||
6193 | @ifclear GENERIC | |
6194 | @lowersections | |
6195 | @end ifclear | |
6196 | @end ifset | |
6197 | ||
7fb9f789 NC |
6198 | @ifset M68K |
6199 | @ifclear GENERIC | |
6200 | @raisesections | |
6201 | @end ifclear | |
6202 | ||
6203 | @node M68K | |
6204 | @section @command{ld} and the Motorola 68K family | |
6205 | ||
6206 | @cindex Motorola 68K GOT generation | |
6207 | @kindex --got=@var{type} | |
6208 | The @samp{--got=@var{type}} option lets you choose the GOT generation scheme. | |
6209 | The choices are @samp{single}, @samp{negative}, @samp{multigot} and | |
6210 | @samp{target}. When @samp{target} is selected the linker chooses | |
6211 | the default GOT generation scheme for the current target. | |
6212 | @samp{single} tells the linker to generate a single GOT with | |
6213 | entries only at non-negative offsets. | |
6214 | @samp{negative} instructs the linker to generate a single GOT with | |
6215 | entries at both negative and positive offsets. Not all environments | |
6216 | support such GOTs. | |
6217 | @samp{multigot} allows the linker to generate several GOTs in the | |
6218 | output file. All GOT references from a single input object | |
6219 | file access the same GOT, but references from different input object | |
6220 | files might access different GOTs. Not all environments support such GOTs. | |
6221 | ||
6222 | @ifclear GENERIC | |
6223 | @lowersections | |
6224 | @end ifclear | |
6225 | @end ifset | |
6226 | ||
36f63dca NC |
6227 | @ifset MMIX |
6228 | @ifclear GENERIC | |
6229 | @raisesections | |
6230 | @end ifclear | |
6231 | ||
6232 | @node MMIX | |
6233 | @section @code{ld} and MMIX | |
6234 | For MMIX, there is a choice of generating @code{ELF} object files or | |
6235 | @code{mmo} object files when linking. The simulator @code{mmix} | |
6236 | understands the @code{mmo} format. The binutils @code{objcopy} utility | |
6237 | can translate between the two formats. | |
6238 | ||
6239 | There is one special section, the @samp{.MMIX.reg_contents} section. | |
6240 | Contents in this section is assumed to correspond to that of global | |
6241 | registers, and symbols referring to it are translated to special symbols, | |
6242 | equal to registers. In a final link, the start address of the | |
6243 | @samp{.MMIX.reg_contents} section corresponds to the first allocated | |
6244 | global register multiplied by 8. Register @code{$255} is not included in | |
6245 | this section; it is always set to the program entry, which is at the | |
6246 | symbol @code{Main} for @code{mmo} files. | |
6247 | ||
7a2de473 HPN |
6248 | Global symbols with the prefix @code{__.MMIX.start.}, for example |
6249 | @code{__.MMIX.start..text} and @code{__.MMIX.start..data} are special. | |
6250 | The default linker script uses these to set the default start address | |
6251 | of a section. | |
36f63dca NC |
6252 | |
6253 | Initial and trailing multiples of zero-valued 32-bit words in a section, | |
6254 | are left out from an mmo file. | |
6255 | ||
6256 | @ifclear GENERIC | |
6257 | @lowersections | |
6258 | @end ifclear | |
6259 | @end ifset | |
6260 | ||
6261 | @ifset MSP430 | |
6262 | @ifclear GENERIC | |
6263 | @raisesections | |
6264 | @end ifclear | |
6265 | ||
6266 | @node MSP430 | |
6267 | @section @code{ld} and MSP430 | |
6268 | For the MSP430 it is possible to select the MPU architecture. The flag @samp{-m [mpu type]} | |
6269 | will select an appropriate linker script for selected MPU type. (To get a list of known MPUs | |
6270 | just pass @samp{-m help} option to the linker). | |
6271 | ||
6272 | @cindex MSP430 extra sections | |
6273 | The linker will recognize some extra sections which are MSP430 specific: | |
6274 | ||
6275 | @table @code | |
6276 | @item @samp{.vectors} | |
6277 | Defines a portion of ROM where interrupt vectors located. | |
6278 | ||
6279 | @item @samp{.bootloader} | |
6280 | Defines the bootloader portion of the ROM (if applicable). Any code | |
6281 | in this section will be uploaded to the MPU. | |
6282 | ||
6283 | @item @samp{.infomem} | |
6284 | Defines an information memory section (if applicable). Any code in | |
6285 | this section will be uploaded to the MPU. | |
6286 | ||
c0065db7 | 6287 | @item @samp{.infomemnobits} |
36f63dca NC |
6288 | This is the same as the @samp{.infomem} section except that any code |
6289 | in this section will not be uploaded to the MPU. | |
6290 | ||
6291 | @item @samp{.noinit} | |
6292 | Denotes a portion of RAM located above @samp{.bss} section. | |
6293 | ||
c0065db7 | 6294 | The last two sections are used by gcc. |
36f63dca NC |
6295 | @end table |
6296 | ||
6297 | @ifclear GENERIC | |
6298 | @lowersections | |
6299 | @end ifclear | |
6300 | @end ifset | |
6301 | ||
2a60a7a8 AM |
6302 | @ifset POWERPC |
6303 | @ifclear GENERIC | |
6304 | @raisesections | |
6305 | @end ifclear | |
6306 | ||
6307 | @node PowerPC ELF32 | |
6308 | @section @command{ld} and PowerPC 32-bit ELF Support | |
6309 | @cindex PowerPC long branches | |
6310 | @kindex --relax on PowerPC | |
6311 | Branches on PowerPC processors are limited to a signed 26-bit | |
6312 | displacement, which may result in @command{ld} giving | |
6313 | @samp{relocation truncated to fit} errors with very large programs. | |
6314 | @samp{--relax} enables the generation of trampolines that can access | |
6315 | the entire 32-bit address space. These trampolines are inserted at | |
6316 | section boundaries, so may not themselves be reachable if an input | |
c8a1f254 NS |
6317 | section exceeds 33M in size. You may combine @samp{-r} and |
6318 | @samp{--relax} to add trampolines in a partial link. In that case | |
6319 | both branches to undefined symbols and inter-section branches are also | |
6320 | considered potentially out of range, and trampolines inserted. | |
2a60a7a8 AM |
6321 | |
6322 | @cindex PowerPC ELF32 options | |
6323 | @table @option | |
6324 | @cindex PowerPC PLT | |
6325 | @kindex --bss-plt | |
6326 | @item --bss-plt | |
6327 | Current PowerPC GCC accepts a @samp{-msecure-plt} option that | |
6328 | generates code capable of using a newer PLT and GOT layout that has | |
6329 | the security advantage of no executable section ever needing to be | |
6330 | writable and no writable section ever being executable. PowerPC | |
6331 | @command{ld} will generate this layout, including stubs to access the | |
6332 | PLT, if all input files (including startup and static libraries) were | |
6333 | compiled with @samp{-msecure-plt}. @samp{--bss-plt} forces the old | |
6334 | BSS PLT (and GOT layout) which can give slightly better performance. | |
6335 | ||
016687f8 AM |
6336 | @kindex --secure-plt |
6337 | @item --secure-plt | |
6338 | @command{ld} will use the new PLT and GOT layout if it is linking new | |
6339 | @samp{-fpic} or @samp{-fPIC} code, but does not do so automatically | |
6340 | when linking non-PIC code. This option requests the new PLT and GOT | |
6341 | layout. A warning will be given if some object file requires the old | |
6342 | style BSS PLT. | |
6343 | ||
2a60a7a8 AM |
6344 | @cindex PowerPC GOT |
6345 | @kindex --sdata-got | |
6346 | @item --sdata-got | |
6347 | The new secure PLT and GOT are placed differently relative to other | |
6348 | sections compared to older BSS PLT and GOT placement. The location of | |
6349 | @code{.plt} must change because the new secure PLT is an initialized | |
6350 | section while the old PLT is uninitialized. The reason for the | |
6351 | @code{.got} change is more subtle: The new placement allows | |
6352 | @code{.got} to be read-only in applications linked with | |
6353 | @samp{-z relro -z now}. However, this placement means that | |
6354 | @code{.sdata} cannot always be used in shared libraries, because the | |
6355 | PowerPC ABI accesses @code{.sdata} in shared libraries from the GOT | |
6356 | pointer. @samp{--sdata-got} forces the old GOT placement. PowerPC | |
6357 | GCC doesn't use @code{.sdata} in shared libraries, so this option is | |
6358 | really only useful for other compilers that may do so. | |
6359 | ||
6360 | @cindex PowerPC stub symbols | |
6361 | @kindex --emit-stub-syms | |
6362 | @item --emit-stub-syms | |
6363 | This option causes @command{ld} to label linker stubs with a local | |
6364 | symbol that encodes the stub type and destination. | |
6365 | ||
6366 | @cindex PowerPC TLS optimization | |
6367 | @kindex --no-tls-optimize | |
6368 | @item --no-tls-optimize | |
6369 | PowerPC @command{ld} normally performs some optimization of code | |
6370 | sequences used to access Thread-Local Storage. Use this option to | |
6371 | disable the optimization. | |
6372 | @end table | |
6373 | ||
6374 | @ifclear GENERIC | |
6375 | @lowersections | |
6376 | @end ifclear | |
6377 | @end ifset | |
6378 | ||
6379 | @ifset POWERPC64 | |
6380 | @ifclear GENERIC | |
6381 | @raisesections | |
6382 | @end ifclear | |
6383 | ||
6384 | @node PowerPC64 ELF64 | |
6385 | @section @command{ld} and PowerPC64 64-bit ELF Support | |
6386 | ||
6387 | @cindex PowerPC64 ELF64 options | |
6388 | @table @option | |
6389 | @cindex PowerPC64 stub grouping | |
6390 | @kindex --stub-group-size | |
6391 | @item --stub-group-size | |
6392 | Long branch stubs, PLT call stubs and TOC adjusting stubs are placed | |
6393 | by @command{ld} in stub sections located between groups of input sections. | |
6394 | @samp{--stub-group-size} specifies the maximum size of a group of input | |
6395 | sections handled by one stub section. Since branch offsets are signed, | |
6396 | a stub section may serve two groups of input sections, one group before | |
6397 | the stub section, and one group after it. However, when using | |
6398 | conditional branches that require stubs, it may be better (for branch | |
6399 | prediction) that stub sections only serve one group of input sections. | |
6400 | A negative value for @samp{N} chooses this scheme, ensuring that | |
6401 | branches to stubs always use a negative offset. Two special values of | |
6402 | @samp{N} are recognized, @samp{1} and @samp{-1}. These both instruct | |
6403 | @command{ld} to automatically size input section groups for the branch types | |
6404 | detected, with the same behaviour regarding stub placement as other | |
6405 | positive or negative values of @samp{N} respectively. | |
6406 | ||
6407 | Note that @samp{--stub-group-size} does not split input sections. A | |
6408 | single input section larger than the group size specified will of course | |
6409 | create a larger group (of one section). If input sections are too | |
6410 | large, it may not be possible for a branch to reach its stub. | |
6411 | ||
6412 | @cindex PowerPC64 stub symbols | |
6413 | @kindex --emit-stub-syms | |
6414 | @item --emit-stub-syms | |
6415 | This option causes @command{ld} to label linker stubs with a local | |
6416 | symbol that encodes the stub type and destination. | |
6417 | ||
6418 | @cindex PowerPC64 dot symbols | |
6419 | @kindex --dotsyms | |
6420 | @kindex --no-dotsyms | |
6421 | @item --dotsyms, --no-dotsyms | |
6422 | These two options control how @command{ld} interprets version patterns | |
6423 | in a version script. Older PowerPC64 compilers emitted both a | |
6424 | function descriptor symbol with the same name as the function, and a | |
6425 | code entry symbol with the name prefixed by a dot (@samp{.}). To | |
6426 | properly version a function @samp{foo}, the version script thus needs | |
6427 | to control both @samp{foo} and @samp{.foo}. The option | |
6428 | @samp{--dotsyms}, on by default, automatically adds the required | |
6429 | dot-prefixed patterns. Use @samp{--no-dotsyms} to disable this | |
6430 | feature. | |
6431 | ||
6432 | @cindex PowerPC64 TLS optimization | |
6433 | @kindex --no-tls-optimize | |
6434 | @item --no-tls-optimize | |
6435 | PowerPC64 @command{ld} normally performs some optimization of code | |
6436 | sequences used to access Thread-Local Storage. Use this option to | |
6437 | disable the optimization. | |
6438 | ||
6439 | @cindex PowerPC64 OPD optimization | |
6440 | @kindex --no-opd-optimize | |
6441 | @item --no-opd-optimize | |
6442 | PowerPC64 @command{ld} normally removes @code{.opd} section entries | |
6443 | corresponding to deleted link-once functions, or functions removed by | |
e7fc76dd | 6444 | the action of @samp{--gc-sections} or linker script @code{/DISCARD/}. |
2a60a7a8 AM |
6445 | Use this option to disable @code{.opd} optimization. |
6446 | ||
6447 | @cindex PowerPC64 OPD spacing | |
6448 | @kindex --non-overlapping-opd | |
6449 | @item --non-overlapping-opd | |
6450 | Some PowerPC64 compilers have an option to generate compressed | |
6451 | @code{.opd} entries spaced 16 bytes apart, overlapping the third word, | |
6452 | the static chain pointer (unused in C) with the first word of the next | |
6453 | entry. This option expands such entries to the full 24 bytes. | |
6454 | ||
6455 | @cindex PowerPC64 TOC optimization | |
6456 | @kindex --no-toc-optimize | |
6457 | @item --no-toc-optimize | |
6458 | PowerPC64 @command{ld} normally removes unused @code{.toc} section | |
6459 | entries. Such entries are detected by examining relocations that | |
6460 | reference the TOC in code sections. A reloc in a deleted code section | |
6461 | marks a TOC word as unneeded, while a reloc in a kept code section | |
6462 | marks a TOC word as needed. Since the TOC may reference itself, TOC | |
6463 | relocs are also examined. TOC words marked as both needed and | |
6464 | unneeded will of course be kept. TOC words without any referencing | |
6465 | reloc are assumed to be part of a multi-word entry, and are kept or | |
6466 | discarded as per the nearest marked preceding word. This works | |
6467 | reliably for compiler generated code, but may be incorrect if assembly | |
6468 | code is used to insert TOC entries. Use this option to disable the | |
6469 | optimization. | |
6470 | ||
6471 | @cindex PowerPC64 multi-TOC | |
6472 | @kindex --no-multi-toc | |
6473 | @item --no-multi-toc | |
6474 | By default, PowerPC64 GCC generates code for a TOC model where TOC | |
6475 | entries are accessed with a 16-bit offset from r2. This limits the | |
6476 | total TOC size to 64K. PowerPC64 @command{ld} extends this limit by | |
6477 | grouping code sections such that each group uses less than 64K for its | |
6478 | TOC entries, then inserts r2 adjusting stubs between inter-group | |
6479 | calls. @command{ld} does not split apart input sections, so cannot | |
6480 | help if a single input file has a @code{.toc} section that exceeds | |
6481 | 64K, most likely from linking multiple files with @command{ld -r}. | |
6482 | Use this option to turn off this feature. | |
6483 | @end table | |
6484 | ||
6485 | @ifclear GENERIC | |
6486 | @lowersections | |
6487 | @end ifclear | |
6488 | @end ifset | |
6489 | ||
49fa1e15 AM |
6490 | @ifset SPU |
6491 | @ifclear GENERIC | |
6492 | @raisesections | |
6493 | @end ifclear | |
6494 | ||
6495 | @node SPU ELF | |
6496 | @section @command{ld} and SPU ELF Support | |
6497 | ||
6498 | @cindex SPU ELF options | |
6499 | @table @option | |
6500 | ||
6501 | @cindex SPU plugins | |
6502 | @kindex --plugin | |
6503 | @item --plugin | |
6504 | This option marks an executable as a PIC plugin module. | |
6505 | ||
6506 | @cindex SPU overlays | |
6507 | @kindex --no-overlays | |
6508 | @item --no-overlays | |
6509 | Normally, @command{ld} recognizes calls to functions within overlay | |
6510 | regions, and redirects such calls to an overlay manager via a stub. | |
6511 | @command{ld} also provides a built-in overlay manager. This option | |
6512 | turns off all this special overlay handling. | |
6513 | ||
6514 | @cindex SPU overlay stub symbols | |
6515 | @kindex --emit-stub-syms | |
6516 | @item --emit-stub-syms | |
6517 | This option causes @command{ld} to label overlay stubs with a local | |
6518 | symbol that encodes the stub type and destination. | |
6519 | ||
6520 | @cindex SPU extra overlay stubs | |
6521 | @kindex --extra-overlay-stubs | |
6522 | @item --extra-overlay-stubs | |
6523 | This option causes @command{ld} to add overlay call stubs on all | |
6524 | function calls out of overlay regions. Normally stubs are not added | |
6525 | on calls to non-overlay regions. | |
6526 | ||
6527 | @cindex SPU local store size | |
6528 | @kindex --local-store=lo:hi | |
6529 | @item --local-store=lo:hi | |
6530 | @command{ld} usually checks that a final executable for SPU fits in | |
6531 | the address range 0 to 256k. This option may be used to change the | |
6532 | range. Disable the check entirely with @option{--local-store=0:0}. | |
6533 | ||
c0065db7 | 6534 | @cindex SPU |
49fa1e15 AM |
6535 | @kindex --stack-analysis |
6536 | @item --stack-analysis | |
6537 | SPU local store space is limited. Over-allocation of stack space | |
6538 | unnecessarily limits space available for code and data, while | |
6539 | under-allocation results in runtime failures. If given this option, | |
6540 | @command{ld} will provide an estimate of maximum stack usage. | |
6541 | @command{ld} does this by examining symbols in code sections to | |
6542 | determine the extents of functions, and looking at function prologues | |
6543 | for stack adjusting instructions. A call-graph is created by looking | |
6544 | for relocations on branch instructions. The graph is then searched | |
6545 | for the maximum stack usage path. Note that this analysis does not | |
6546 | find calls made via function pointers, and does not handle recursion | |
6547 | and other cycles in the call graph. Stack usage may be | |
6548 | under-estimated if your code makes such calls. Also, stack usage for | |
6549 | dynamic allocation, e.g. alloca, will not be detected. If a link map | |
6550 | is requested, detailed information about each function's stack usage | |
6551 | and calls will be given. | |
6552 | ||
c0065db7 | 6553 | @cindex SPU |
49fa1e15 AM |
6554 | @kindex --emit-stack-syms |
6555 | @item --emit-stack-syms | |
6556 | This option, if given along with @option{--stack-analysis} will result | |
6557 | in @command{ld} emitting stack sizing symbols for each function. | |
6558 | These take the form @code{__stack_<function_name>} for global | |
6559 | functions, and @code{__stack_<number>_<function_name>} for static | |
6560 | functions. @code{<number>} is the section id in hex. The value of | |
6561 | such symbols is the stack requirement for the corresponding function. | |
6562 | The symbol size will be zero, type @code{STT_NOTYPE}, binding | |
c0065db7 | 6563 | @code{STB_LOCAL}, and section @code{SHN_ABS}. |
49fa1e15 AM |
6564 | @end table |
6565 | ||
6566 | @ifclear GENERIC | |
6567 | @lowersections | |
6568 | @end ifclear | |
6569 | @end ifset | |
6570 | ||
36f63dca NC |
6571 | @ifset TICOFF |
6572 | @ifclear GENERIC | |
6573 | @raisesections | |
6574 | @end ifclear | |
6575 | ||
6576 | @node TI COFF | |
6577 | @section @command{ld}'s Support for Various TI COFF Versions | |
6578 | @cindex TI COFF versions | |
6579 | @kindex --format=@var{version} | |
6580 | The @samp{--format} switch allows selection of one of the various | |
6581 | TI COFF versions. The latest of this writing is 2; versions 0 and 1 are | |
6582 | also supported. The TI COFF versions also vary in header byte-order | |
6583 | format; @command{ld} will read any version or byte order, but the output | |
6584 | header format depends on the default specified by the specific target. | |
6585 | ||
6586 | @ifclear GENERIC | |
6587 | @lowersections | |
6588 | @end ifclear | |
6589 | @end ifset | |
6590 | ||
2ca22b03 NC |
6591 | @ifset WIN32 |
6592 | @ifclear GENERIC | |
6593 | @raisesections | |
6594 | @end ifclear | |
6595 | ||
6596 | @node WIN32 | |
6597 | @section @command{ld} and WIN32 (cygwin/mingw) | |
6598 | ||
c0065db7 | 6599 | This section describes some of the win32 specific @command{ld} issues. |
b45619c0 | 6600 | See @ref{Options,,Command Line Options} for detailed description of the |
dc8465bf | 6601 | command line options mentioned here. |
2ca22b03 NC |
6602 | |
6603 | @table @emph | |
c0065db7 RM |
6604 | @cindex import libraries |
6605 | @item import libraries | |
69da35b5 | 6606 | The standard Windows linker creates and uses so-called import |
2ca22b03 | 6607 | libraries, which contains information for linking to dll's. They are |
69da35b5 NC |
6608 | regular static archives and are handled as any other static |
6609 | archive. The cygwin and mingw ports of @command{ld} have specific | |
2ca22b03 NC |
6610 | support for creating such libraries provided with the |
6611 | @samp{--out-implib} command line option. | |
6612 | ||
c0065db7 RM |
6613 | @item exporting DLL symbols |
6614 | @cindex exporting DLL symbols | |
dc8465bf NC |
6615 | The cygwin/mingw @command{ld} has several ways to export symbols for dll's. |
6616 | ||
6617 | @table @emph | |
6618 | @item using auto-export functionality | |
6619 | @cindex using auto-export functionality | |
6620 | By default @command{ld} exports symbols with the auto-export functionality, | |
6621 | which is controlled by the following command line options: | |
6622 | ||
0a5d968e NC |
6623 | @itemize |
6624 | @item --export-all-symbols [This is the default] | |
6625 | @item --exclude-symbols | |
6626 | @item --exclude-libs | |
e1c37eb5 | 6627 | @item --exclude-modules-for-implib |
09e2aba4 | 6628 | @item --version-script |
0a5d968e NC |
6629 | @end itemize |
6630 | ||
09e2aba4 DK |
6631 | When auto-export is in operation, @command{ld} will export all the non-local |
6632 | (global and common) symbols it finds in a DLL, with the exception of a few | |
6633 | symbols known to belong to the system's runtime and libraries. As it will | |
6634 | often not be desirable to export all of a DLL's symbols, which may include | |
6635 | private functions that are not part of any public interface, the command-line | |
6636 | options listed above may be used to filter symbols out from the list for | |
6637 | exporting. The @samp{--output-def} option can be used in order to see the | |
6638 | final list of exported symbols with all exclusions taken into effect. | |
6639 | ||
6640 | If @samp{--export-all-symbols} is not given explicitly on the | |
0a5d968e NC |
6641 | command line, then the default auto-export behavior will be @emph{disabled} |
6642 | if either of the following are true: | |
6643 | ||
6644 | @itemize | |
6645 | @item A DEF file is used. | |
6646 | @item Any symbol in any object file was marked with the __declspec(dllexport) attribute. | |
6647 | @end itemize | |
dc8465bf | 6648 | |
c0065db7 RM |
6649 | @item using a DEF file |
6650 | @cindex using a DEF file | |
dc8465bf NC |
6651 | Another way of exporting symbols is using a DEF file. A DEF file is |
6652 | an ASCII file containing definitions of symbols which should be | |
6653 | exported when a dll is created. Usually it is named @samp{<dll | |
6654 | name>.def} and is added as any other object file to the linker's | |
0a5d968e | 6655 | command line. The file's name must end in @samp{.def} or @samp{.DEF}. |
dc8465bf NC |
6656 | |
6657 | @example | |
6658 | gcc -o <output> <objectfiles> <dll name>.def | |
6659 | @end example | |
6660 | ||
0a5d968e NC |
6661 | Using a DEF file turns off the normal auto-export behavior, unless the |
6662 | @samp{--export-all-symbols} option is also used. | |
6663 | ||
dc8465bf NC |
6664 | Here is an example of a DEF file for a shared library called @samp{xyz.dll}: |
6665 | ||
6666 | @example | |
4b5bd4e7 | 6667 | LIBRARY "xyz.dll" BASE=0x20000000 |
dc8465bf NC |
6668 | |
6669 | EXPORTS | |
6670 | foo | |
6671 | bar | |
6672 | _bar = bar | |
4b5bd4e7 DS |
6673 | another_foo = abc.dll.afoo |
6674 | var1 DATA | |
c0065db7 | 6675 | @end example |
dc8465bf | 6676 | |
4b5bd4e7 DS |
6677 | This example defines a DLL with a non-default base address and five |
6678 | symbols in the export table. The third exported symbol @code{_bar} is an | |
6679 | alias for the second. The fourth symbol, @code{another_foo} is resolved | |
6680 | by "forwarding" to another module and treating it as an alias for | |
6681 | @code{afoo} exported from the DLL @samp{abc.dll}. The final symbol | |
6682 | @code{var1} is declared to be a data object. | |
6683 | ||
6b31ad16 DS |
6684 | The optional @code{LIBRARY <name>} command indicates the @emph{internal} |
6685 | name of the output DLL. If @samp{<name>} does not include a suffix, | |
6686 | the default library suffix, @samp{.DLL} is appended. | |
6687 | ||
b45619c0 NC |
6688 | When the .DEF file is used to build an application, rather than a |
6689 | library, the @code{NAME <name>} command should be used instead of | |
6b31ad16 | 6690 | @code{LIBRARY}. If @samp{<name>} does not include a suffix, the default |
c0065db7 | 6691 | executable suffix, @samp{.EXE} is appended. |
6b31ad16 DS |
6692 | |
6693 | With either @code{LIBRARY <name>} or @code{NAME <name>} the optional | |
6694 | specification @code{BASE = <number>} may be used to specify a | |
c0065db7 | 6695 | non-default base address for the image. |
6b31ad16 DS |
6696 | |
6697 | If neither @code{LIBRARY <name>} nor @code{NAME <name>} is specified, | |
a2877985 DS |
6698 | or they specify an empty string, the internal name is the same as the |
6699 | filename specified on the command line. | |
6b31ad16 | 6700 | |
4b5bd4e7 DS |
6701 | The complete specification of an export symbol is: |
6702 | ||
6703 | @example | |
6704 | EXPORTS | |
6705 | ( ( ( <name1> [ = <name2> ] ) | |
6706 | | ( <name1> = <module-name> . <external-name>)) | |
6707 | [ @@ <integer> ] [NONAME] [DATA] [CONSTANT] [PRIVATE] ) * | |
c0065db7 | 6708 | @end example |
4b5bd4e7 DS |
6709 | |
6710 | Declares @samp{<name1>} as an exported symbol from the DLL, or declares | |
6711 | @samp{<name1>} as an exported alias for @samp{<name2>}; or declares | |
6712 | @samp{<name1>} as a "forward" alias for the symbol | |
6713 | @samp{<external-name>} in the DLL @samp{<module-name>}. | |
6714 | Optionally, the symbol may be exported by the specified ordinal | |
6715 | @samp{<integer>} alias. | |
6716 | ||
6717 | The optional keywords that follow the declaration indicate: | |
6718 | ||
6719 | @code{NONAME}: Do not put the symbol name in the DLL's export table. It | |
6720 | will still be exported by its ordinal alias (either the value specified | |
6721 | by the .def specification or, otherwise, the value assigned by the | |
6722 | linker). The symbol name, however, does remain visible in the import | |
6723 | library (if any), unless @code{PRIVATE} is also specified. | |
6724 | ||
6725 | @code{DATA}: The symbol is a variable or object, rather than a function. | |
6726 | The import lib will export only an indirect reference to @code{foo} as | |
6727 | the symbol @code{_imp__foo} (ie, @code{foo} must be resolved as | |
6728 | @code{*_imp__foo}). | |
6729 | ||
6730 | @code{CONSTANT}: Like @code{DATA}, but put the undecorated @code{foo} as | |
6731 | well as @code{_imp__foo} into the import library. Both refer to the | |
6732 | read-only import address table's pointer to the variable, not to the | |
6733 | variable itself. This can be dangerous. If the user code fails to add | |
6734 | the @code{dllimport} attribute and also fails to explicitly add the | |
6735 | extra indirection that the use of the attribute enforces, the | |
6736 | application will behave unexpectedly. | |
6737 | ||
6738 | @code{PRIVATE}: Put the symbol in the DLL's export table, but do not put | |
6739 | it into the static import library used to resolve imports at link time. The | |
6740 | symbol can still be imported using the @code{LoadLibrary/GetProcAddress} | |
6741 | API at runtime or by by using the GNU ld extension of linking directly to | |
6742 | the DLL without an import library. | |
c0065db7 | 6743 | |
4b5bd4e7 DS |
6744 | See ld/deffilep.y in the binutils sources for the full specification of |
6745 | other DEF file statements | |
dc8465bf NC |
6746 | |
6747 | @cindex creating a DEF file | |
6748 | While linking a shared dll, @command{ld} is able to create a DEF file | |
6749 | with the @samp{--output-def <file>} command line option. | |
0a5d968e NC |
6750 | |
6751 | @item Using decorations | |
6752 | @cindex Using decorations | |
6753 | Another way of marking symbols for export is to modify the source code | |
6754 | itself, so that when building the DLL each symbol to be exported is | |
6755 | declared as: | |
6756 | ||
6757 | @example | |
6758 | __declspec(dllexport) int a_variable | |
6759 | __declspec(dllexport) void a_function(int with_args) | |
6760 | @end example | |
6761 | ||
6762 | All such symbols will be exported from the DLL. If, however, | |
6763 | any of the object files in the DLL contain symbols decorated in | |
6764 | this way, then the normal auto-export behavior is disabled, unless | |
6765 | the @samp{--export-all-symbols} option is also used. | |
6766 | ||
6767 | Note that object files that wish to access these symbols must @emph{not} | |
c0065db7 | 6768 | decorate them with dllexport. Instead, they should use dllimport, |
0a5d968e NC |
6769 | instead: |
6770 | ||
6771 | @example | |
6772 | __declspec(dllimport) int a_variable | |
6773 | __declspec(dllimport) void a_function(int with_args) | |
6774 | @end example | |
6775 | ||
c0065db7 RM |
6776 | This complicates the structure of library header files, because |
6777 | when included by the library itself the header must declare the | |
0a5d968e NC |
6778 | variables and functions as dllexport, but when included by client |
6779 | code the header must declare them as dllimport. There are a number | |
c0065db7 | 6780 | of idioms that are typically used to do this; often client code can |
0a5d968e NC |
6781 | omit the __declspec() declaration completely. See |
6782 | @samp{--enable-auto-import} and @samp{automatic data imports} for more | |
b45619c0 | 6783 | information. |
c0065db7 | 6784 | @end table |
dc8465bf | 6785 | |
2ca22b03 NC |
6786 | @cindex automatic data imports |
6787 | @item automatic data imports | |
6788 | The standard Windows dll format supports data imports from dlls only | |
69da35b5 | 6789 | by adding special decorations (dllimport/dllexport), which let the |
2ca22b03 | 6790 | compiler produce specific assembler instructions to deal with this |
c0065db7 | 6791 | issue. This increases the effort necessary to port existing Un*x |
69da35b5 | 6792 | code to these platforms, especially for large |
2ca22b03 | 6793 | c++ libraries and applications. The auto-import feature, which was |
c0065db7 | 6794 | initially provided by Paul Sokolovsky, allows one to omit the |
b45619c0 | 6795 | decorations to achieve a behavior that conforms to that on POSIX/Un*x |
c0065db7 | 6796 | platforms. This feature is enabled with the @samp{--enable-auto-import} |
69da35b5 NC |
6797 | command-line option, although it is enabled by default on cygwin/mingw. |
6798 | The @samp{--enable-auto-import} option itself now serves mainly to | |
6799 | suppress any warnings that are ordinarily emitted when linked objects | |
6800 | trigger the feature's use. | |
6801 | ||
c0065db7 | 6802 | auto-import of variables does not always work flawlessly without |
69da35b5 NC |
6803 | additional assistance. Sometimes, you will see this message |
6804 | ||
c0065db7 | 6805 | "variable '<var>' can't be auto-imported. Please read the |
69da35b5 NC |
6806 | documentation for ld's @code{--enable-auto-import} for details." |
6807 | ||
c0065db7 RM |
6808 | The @samp{--enable-auto-import} documentation explains why this error |
6809 | occurs, and several methods that can be used to overcome this difficulty. | |
6810 | One of these methods is the @emph{runtime pseudo-relocs} feature, described | |
69da35b5 NC |
6811 | below. |
6812 | ||
6813 | @cindex runtime pseudo-relocation | |
c0065db7 RM |
6814 | For complex variables imported from DLLs (such as structs or classes), |
6815 | object files typically contain a base address for the variable and an | |
6816 | offset (@emph{addend}) within the variable--to specify a particular | |
6817 | field or public member, for instance. Unfortunately, the runtime loader used | |
6818 | in win32 environments is incapable of fixing these references at runtime | |
69da35b5 | 6819 | without the additional information supplied by dllimport/dllexport decorations. |
c0065db7 | 6820 | The standard auto-import feature described above is unable to resolve these |
69da35b5 NC |
6821 | references. |
6822 | ||
c0065db7 RM |
6823 | The @samp{--enable-runtime-pseudo-relocs} switch allows these references to |
6824 | be resolved without error, while leaving the task of adjusting the references | |
6825 | themselves (with their non-zero addends) to specialized code provided by the | |
6826 | runtime environment. Recent versions of the cygwin and mingw environments and | |
6827 | compilers provide this runtime support; older versions do not. However, the | |
6828 | support is only necessary on the developer's platform; the compiled result will | |
69da35b5 NC |
6829 | run without error on an older system. |
6830 | ||
c0065db7 RM |
6831 | @samp{--enable-runtime-pseudo-relocs} is not the default; it must be explicitly |
6832 | enabled as needed. | |
2ca22b03 NC |
6833 | |
6834 | @cindex direct linking to a dll | |
6835 | @item direct linking to a dll | |
6836 | The cygwin/mingw ports of @command{ld} support the direct linking, | |
6837 | including data symbols, to a dll without the usage of any import | |
69da35b5 | 6838 | libraries. This is much faster and uses much less memory than does the |
b45619c0 | 6839 | traditional import library method, especially when linking large |
c0065db7 RM |
6840 | libraries or applications. When @command{ld} creates an import lib, each |
6841 | function or variable exported from the dll is stored in its own bfd, even | |
6842 | though a single bfd could contain many exports. The overhead involved in | |
69da35b5 | 6843 | storing, loading, and processing so many bfd's is quite large, and explains the |
c0065db7 | 6844 | tremendous time, memory, and storage needed to link against particularly |
69da35b5 NC |
6845 | large or complex libraries when using import libs. |
6846 | ||
c0065db7 | 6847 | Linking directly to a dll uses no extra command-line switches other than |
69da35b5 | 6848 | @samp{-L} and @samp{-l}, because @command{ld} already searches for a number |
c0065db7 | 6849 | of names to match each library. All that is needed from the developer's |
69da35b5 NC |
6850 | perspective is an understanding of this search, in order to force ld to |
6851 | select the dll instead of an import library. | |
6852 | ||
2ca22b03 | 6853 | |
69da35b5 NC |
6854 | For instance, when ld is called with the argument @samp{-lxxx} it will attempt |
6855 | to find, in the first directory of its search path, | |
2ca22b03 NC |
6856 | |
6857 | @example | |
45e948fe NC |
6858 | libxxx.dll.a |
6859 | xxx.dll.a | |
6860 | libxxx.a | |
6861 | xxx.lib | |
69da35b5 | 6862 | cygxxx.dll (*) |
45e948fe NC |
6863 | libxxx.dll |
6864 | xxx.dll | |
2ca22b03 NC |
6865 | @end example |
6866 | ||
69da35b5 NC |
6867 | before moving on to the next directory in the search path. |
6868 | ||
c0065db7 RM |
6869 | (*) Actually, this is not @samp{cygxxx.dll} but in fact is @samp{<prefix>xxx.dll}, |
6870 | where @samp{<prefix>} is set by the @command{ld} option | |
6871 | @samp{--dll-search-prefix=<prefix>}. In the case of cygwin, the standard gcc spec | |
6872 | file includes @samp{--dll-search-prefix=cyg}, so in effect we actually search for | |
69da35b5 NC |
6873 | @samp{cygxxx.dll}. |
6874 | ||
c0065db7 RM |
6875 | Other win32-based unix environments, such as mingw or pw32, may use other |
6876 | @samp{<prefix>}es, although at present only cygwin makes use of this feature. It | |
69da35b5 NC |
6877 | was originally intended to help avoid name conflicts among dll's built for the |
6878 | various win32/un*x environments, so that (for example) two versions of a zlib dll | |
6879 | could coexist on the same machine. | |
6880 | ||
2ca22b03 NC |
6881 | The generic cygwin/mingw path layout uses a @samp{bin} directory for |
6882 | applications and dll's and a @samp{lib} directory for the import | |
69da35b5 | 6883 | libraries (using cygwin nomenclature): |
2ca22b03 NC |
6884 | |
6885 | @example | |
6886 | bin/ | |
6887 | cygxxx.dll | |
6888 | lib/ | |
6889 | libxxx.dll.a (in case of dll's) | |
c0065db7 | 6890 | libxxx.a (in case of static archive) |
2ca22b03 NC |
6891 | @end example |
6892 | ||
c0065db7 RM |
6893 | Linking directly to a dll without using the import library can be |
6894 | done two ways: | |
2ca22b03 NC |
6895 | |
6896 | 1. Use the dll directly by adding the @samp{bin} path to the link line | |
6897 | @example | |
6898 | gcc -Wl,-verbose -o a.exe -L../bin/ -lxxx | |
c0065db7 | 6899 | @end example |
2ca22b03 | 6900 | |
69da35b5 NC |
6901 | However, as the dll's often have version numbers appended to their names |
6902 | (@samp{cygncurses-5.dll}) this will often fail, unless one specifies | |
6903 | @samp{-L../bin -lncurses-5} to include the version. Import libs are generally | |
6904 | not versioned, and do not have this difficulty. | |
6905 | ||
2ca22b03 NC |
6906 | 2. Create a symbolic link from the dll to a file in the @samp{lib} |
6907 | directory according to the above mentioned search pattern. This | |
6908 | should be used to avoid unwanted changes in the tools needed for | |
6909 | making the app/dll. | |
6910 | ||
6911 | @example | |
6912 | ln -s bin/cygxxx.dll lib/[cyg|lib|]xxx.dll[.a] | |
c0065db7 | 6913 | @end example |
2ca22b03 NC |
6914 | |
6915 | Then you can link without any make environment changes. | |
6916 | ||
6917 | @example | |
6918 | gcc -Wl,-verbose -o a.exe -L../lib/ -lxxx | |
c0065db7 | 6919 | @end example |
69da35b5 NC |
6920 | |
6921 | This technique also avoids the version number problems, because the following is | |
6922 | perfectly legal | |
6923 | ||
6924 | @example | |
6925 | bin/ | |
6926 | cygxxx-5.dll | |
6927 | lib/ | |
c0065db7 | 6928 | libxxx.dll.a -> ../bin/cygxxx-5.dll |
69da35b5 NC |
6929 | @end example |
6930 | ||
dc8465bf | 6931 | Linking directly to a dll without using an import lib will work |
69da35b5 NC |
6932 | even when auto-import features are exercised, and even when |
6933 | @samp{--enable-runtime-pseudo-relocs} is used. | |
6934 | ||
6935 | Given the improvements in speed and memory usage, one might justifiably | |
45e948fe | 6936 | wonder why import libraries are used at all. There are three reasons: |
69da35b5 NC |
6937 | |
6938 | 1. Until recently, the link-directly-to-dll functionality did @emph{not} | |
6939 | work with auto-imported data. | |
6940 | ||
dc8465bf NC |
6941 | 2. Sometimes it is necessary to include pure static objects within the |
6942 | import library (which otherwise contains only bfd's for indirection | |
6943 | symbols that point to the exports of a dll). Again, the import lib | |
6944 | for the cygwin kernel makes use of this ability, and it is not | |
6945 | possible to do this without an import lib. | |
69da35b5 | 6946 | |
45e948fe NC |
6947 | 3. Symbol aliases can only be resolved using an import lib. This is |
6948 | critical when linking against OS-supplied dll's (eg, the win32 API) | |
6949 | in which symbols are usually exported as undecorated aliases of their | |
6950 | stdcall-decorated assembly names. | |
6951 | ||
69da35b5 | 6952 | So, import libs are not going away. But the ability to replace |
c0065db7 RM |
6953 | true import libs with a simple symbolic link to (or a copy of) |
6954 | a dll, in many cases, is a useful addition to the suite of tools | |
6955 | binutils makes available to the win32 developer. Given the | |
69da35b5 NC |
6956 | massive improvements in memory requirements during linking, storage |
6957 | requirements, and linking speed, we expect that many developers | |
6958 | will soon begin to use this feature whenever possible. | |
dc8465bf | 6959 | |
c0065db7 | 6960 | @item symbol aliasing |
dc8465bf | 6961 | @table @emph |
c0065db7 RM |
6962 | @item adding additional names |
6963 | Sometimes, it is useful to export symbols with additional names. | |
dc8465bf NC |
6964 | A symbol @samp{foo} will be exported as @samp{foo}, but it can also be |
6965 | exported as @samp{_foo} by using special directives in the DEF file | |
6966 | when creating the dll. This will affect also the optional created | |
c0065db7 | 6967 | import library. Consider the following DEF file: |
dc8465bf | 6968 | |
c0065db7 | 6969 | @example |
dc8465bf NC |
6970 | LIBRARY "xyz.dll" BASE=0x61000000 |
6971 | ||
6972 | EXPORTS | |
c0065db7 | 6973 | foo |
dc8465bf | 6974 | _foo = foo |
c0065db7 | 6975 | @end example |
dc8465bf NC |
6976 | |
6977 | The line @samp{_foo = foo} maps the symbol @samp{foo} to @samp{_foo}. | |
6978 | ||
6979 | Another method for creating a symbol alias is to create it in the | |
6980 | source code using the "weak" attribute: | |
6981 | ||
c0065db7 RM |
6982 | @example |
6983 | void foo () @{ /* Do something. */; @} | |
dc8465bf | 6984 | void _foo () __attribute__ ((weak, alias ("foo"))); |
c0065db7 | 6985 | @end example |
dc8465bf NC |
6986 | |
6987 | See the gcc manual for more information about attributes and weak | |
6988 | symbols. | |
6989 | ||
6990 | @item renaming symbols | |
6991 | Sometimes it is useful to rename exports. For instance, the cygwin | |
c0065db7 | 6992 | kernel does this regularly. A symbol @samp{_foo} can be exported as |
dc8465bf NC |
6993 | @samp{foo} but not as @samp{_foo} by using special directives in the |
6994 | DEF file. (This will also affect the import library, if it is | |
c0065db7 | 6995 | created). In the following example: |
dc8465bf | 6996 | |
c0065db7 | 6997 | @example |
dc8465bf NC |
6998 | LIBRARY "xyz.dll" BASE=0x61000000 |
6999 | ||
7000 | EXPORTS | |
7001 | _foo = foo | |
c0065db7 | 7002 | @end example |
dc8465bf NC |
7003 | |
7004 | The line @samp{_foo = foo} maps the exported symbol @samp{foo} to | |
7005 | @samp{_foo}. | |
c0065db7 | 7006 | @end table |
dc8465bf | 7007 | |
0a5d968e | 7008 | Note: using a DEF file disables the default auto-export behavior, |
c0065db7 | 7009 | unless the @samp{--export-all-symbols} command line option is used. |
0a5d968e | 7010 | If, however, you are trying to rename symbols, then you should list |
c0065db7 RM |
7011 | @emph{all} desired exports in the DEF file, including the symbols |
7012 | that are not being renamed, and do @emph{not} use the | |
7013 | @samp{--export-all-symbols} option. If you list only the | |
7014 | renamed symbols in the DEF file, and use @samp{--export-all-symbols} | |
7015 | to handle the other symbols, then the both the new names @emph{and} | |
7016 | the original names for the renamed symbols will be exported. | |
7017 | In effect, you'd be aliasing those symbols, not renaming them, | |
0a5d968e | 7018 | which is probably not what you wanted. |
c87db184 CF |
7019 | |
7020 | @cindex weak externals | |
7021 | @item weak externals | |
7022 | The Windows object format, PE, specifies a form of weak symbols called | |
7023 | weak externals. When a weak symbol is linked and the symbol is not | |
7024 | defined, the weak symbol becomes an alias for some other symbol. There | |
7025 | are three variants of weak externals: | |
7026 | @itemize | |
7027 | @item Definition is searched for in objects and libraries, historically | |
7028 | called lazy externals. | |
7029 | @item Definition is searched for only in other objects, not in libraries. | |
7030 | This form is not presently implemented. | |
7031 | @item No search; the symbol is an alias. This form is not presently | |
7032 | implemented. | |
7033 | @end itemize | |
7034 | As a GNU extension, weak symbols that do not specify an alternate symbol | |
7035 | are supported. If the symbol is undefined when linking, the symbol | |
7036 | uses a default value. | |
c1711530 DK |
7037 | |
7038 | @cindex aligned common symbols | |
7039 | @item aligned common symbols | |
7040 | As a GNU extension to the PE file format, it is possible to specify the | |
7041 | desired alignment for a common symbol. This information is conveyed from | |
7042 | the assembler or compiler to the linker by means of GNU-specific commands | |
7043 | carried in the object file's @samp{.drectve} section, which are recognized | |
7044 | by @command{ld} and respected when laying out the common symbols. Native | |
7045 | tools will be able to process object files employing this GNU extension, | |
7046 | but will fail to respect the alignment instructions, and may issue noisy | |
7047 | warnings about unknown linker directives. | |
2ca22b03 NC |
7048 | @end table |
7049 | ||
7050 | @ifclear GENERIC | |
7051 | @lowersections | |
7052 | @end ifclear | |
7053 | @end ifset | |
7054 | ||
e0001a05 NC |
7055 | @ifset XTENSA |
7056 | @ifclear GENERIC | |
7057 | @raisesections | |
7058 | @end ifclear | |
7059 | ||
7060 | @node Xtensa | |
7061 | @section @code{ld} and Xtensa Processors | |
7062 | ||
7063 | @cindex Xtensa processors | |
7064 | The default @command{ld} behavior for Xtensa processors is to interpret | |
7065 | @code{SECTIONS} commands so that lists of explicitly named sections in a | |
7066 | specification with a wildcard file will be interleaved when necessary to | |
7067 | keep literal pools within the range of PC-relative load offsets. For | |
7068 | example, with the command: | |
7069 | ||
7070 | @smallexample | |
7071 | SECTIONS | |
7072 | @{ | |
7073 | .text : @{ | |
7074 | *(.literal .text) | |
7075 | @} | |
7076 | @} | |
7077 | @end smallexample | |
7078 | ||
7079 | @noindent | |
7080 | @command{ld} may interleave some of the @code{.literal} | |
7081 | and @code{.text} sections from different object files to ensure that the | |
7082 | literal pools are within the range of PC-relative load offsets. A valid | |
7083 | interleaving might place the @code{.literal} sections from an initial | |
7084 | group of files followed by the @code{.text} sections of that group of | |
7085 | files. Then, the @code{.literal} sections from the rest of the files | |
7086 | and the @code{.text} sections from the rest of the files would follow. | |
e0001a05 | 7087 | |
43cd72b9 | 7088 | @cindex @option{--relax} on Xtensa |
e0001a05 | 7089 | @cindex relaxing on Xtensa |
43cd72b9 BW |
7090 | Relaxation is enabled by default for the Xtensa version of @command{ld} and |
7091 | provides two important link-time optimizations. The first optimization | |
7092 | is to combine identical literal values to reduce code size. A redundant | |
7093 | literal will be removed and all the @code{L32R} instructions that use it | |
7094 | will be changed to reference an identical literal, as long as the | |
7095 | location of the replacement literal is within the offset range of all | |
7096 | the @code{L32R} instructions. The second optimization is to remove | |
7097 | unnecessary overhead from assembler-generated ``longcall'' sequences of | |
7098 | @code{L32R}/@code{CALLX@var{n}} when the target functions are within | |
7099 | range of direct @code{CALL@var{n}} instructions. | |
7100 | ||
7101 | For each of these cases where an indirect call sequence can be optimized | |
7102 | to a direct call, the linker will change the @code{CALLX@var{n}} | |
7103 | instruction to a @code{CALL@var{n}} instruction, remove the @code{L32R} | |
7104 | instruction, and remove the literal referenced by the @code{L32R} | |
7105 | instruction if it is not used for anything else. Removing the | |
7106 | @code{L32R} instruction always reduces code size but can potentially | |
7107 | hurt performance by changing the alignment of subsequent branch targets. | |
7108 | By default, the linker will always preserve alignments, either by | |
7109 | switching some instructions between 24-bit encodings and the equivalent | |
7110 | density instructions or by inserting a no-op in place of the @code{L32R} | |
7111 | instruction that was removed. If code size is more important than | |
7112 | performance, the @option{--size-opt} option can be used to prevent the | |
7113 | linker from widening density instructions or inserting no-ops, except in | |
7114 | a few cases where no-ops are required for correctness. | |
7115 | ||
7116 | The following Xtensa-specific command-line options can be used to | |
7117 | control the linker: | |
7118 | ||
7119 | @cindex Xtensa options | |
7120 | @table @option | |
e0001a05 | 7121 | @kindex --no-relax |
43cd72b9 BW |
7122 | @item --no-relax |
7123 | Since the Xtensa version of @code{ld} enables the @option{--relax} option | |
7124 | by default, the @option{--no-relax} option is provided to disable | |
7125 | relaxation. | |
7126 | ||
7127 | @item --size-opt | |
7128 | When optimizing indirect calls to direct calls, optimize for code size | |
7129 | more than performance. With this option, the linker will not insert | |
7130 | no-ops or widen density instructions to preserve branch target | |
7131 | alignment. There may still be some cases where no-ops are required to | |
7132 | preserve the correctness of the code. | |
7133 | @end table | |
e0001a05 NC |
7134 | |
7135 | @ifclear GENERIC | |
7136 | @lowersections | |
7137 | @end ifclear | |
7138 | @end ifset | |
7139 | ||
252b5132 RH |
7140 | @ifclear SingleFormat |
7141 | @node BFD | |
7142 | @chapter BFD | |
7143 | ||
7144 | @cindex back end | |
7145 | @cindex object file management | |
7146 | @cindex object formats available | |
7147 | @kindex objdump -i | |
7148 | The linker accesses object and archive files using the BFD libraries. | |
7149 | These libraries allow the linker to use the same routines to operate on | |
7150 | object files whatever the object file format. A different object file | |
7151 | format can be supported simply by creating a new BFD back end and adding | |
7152 | it to the library. To conserve runtime memory, however, the linker and | |
7153 | associated tools are usually configured to support only a subset of the | |
7154 | object file formats available. You can use @code{objdump -i} | |
7155 | (@pxref{objdump,,objdump,binutils.info,The GNU Binary Utilities}) to | |
7156 | list all the formats available for your configuration. | |
7157 | ||
7158 | @cindex BFD requirements | |
7159 | @cindex requirements for BFD | |
7160 | As with most implementations, BFD is a compromise between | |
7161 | several conflicting requirements. The major factor influencing | |
7162 | BFD design was efficiency: any time used converting between | |
7163 | formats is time which would not have been spent had BFD not | |
7164 | been involved. This is partly offset by abstraction payback; since | |
7165 | BFD simplifies applications and back ends, more time and care | |
7166 | may be spent optimizing algorithms for a greater speed. | |
7167 | ||
7168 | One minor artifact of the BFD solution which you should bear in | |
7169 | mind is the potential for information loss. There are two places where | |
7170 | useful information can be lost using the BFD mechanism: during | |
7171 | conversion and during output. @xref{BFD information loss}. | |
7172 | ||
7173 | @menu | |
7174 | * BFD outline:: How it works: an outline of BFD | |
7175 | @end menu | |
7176 | ||
7177 | @node BFD outline | |
36f63dca | 7178 | @section How It Works: An Outline of BFD |
252b5132 RH |
7179 | @cindex opening object files |
7180 | @include bfdsumm.texi | |
7181 | @end ifclear | |
7182 | ||
7183 | @node Reporting Bugs | |
7184 | @chapter Reporting Bugs | |
ff5dcc92 SC |
7185 | @cindex bugs in @command{ld} |
7186 | @cindex reporting bugs in @command{ld} | |
252b5132 | 7187 | |
ff5dcc92 | 7188 | Your bug reports play an essential role in making @command{ld} reliable. |
252b5132 RH |
7189 | |
7190 | Reporting a bug may help you by bringing a solution to your problem, or | |
7191 | it may not. But in any case the principal function of a bug report is | |
ff5dcc92 | 7192 | to help the entire community by making the next version of @command{ld} |
252b5132 | 7193 | work better. Bug reports are your contribution to the maintenance of |
ff5dcc92 | 7194 | @command{ld}. |
252b5132 RH |
7195 | |
7196 | In order for a bug report to serve its purpose, you must include the | |
7197 | information that enables us to fix the bug. | |
7198 | ||
7199 | @menu | |
7200 | * Bug Criteria:: Have you found a bug? | |
7201 | * Bug Reporting:: How to report bugs | |
7202 | @end menu | |
7203 | ||
7204 | @node Bug Criteria | |
36f63dca | 7205 | @section Have You Found a Bug? |
252b5132 RH |
7206 | @cindex bug criteria |
7207 | ||
7208 | If you are not sure whether you have found a bug, here are some guidelines: | |
7209 | ||
7210 | @itemize @bullet | |
7211 | @cindex fatal signal | |
7212 | @cindex linker crash | |
7213 | @cindex crash of linker | |
7214 | @item | |
7215 | If the linker gets a fatal signal, for any input whatever, that is a | |
ff5dcc92 | 7216 | @command{ld} bug. Reliable linkers never crash. |
252b5132 RH |
7217 | |
7218 | @cindex error on valid input | |
7219 | @item | |
ff5dcc92 | 7220 | If @command{ld} produces an error message for valid input, that is a bug. |
252b5132 RH |
7221 | |
7222 | @cindex invalid input | |
7223 | @item | |
ff5dcc92 | 7224 | If @command{ld} does not produce an error message for invalid input, that |
252b5132 RH |
7225 | may be a bug. In the general case, the linker can not verify that |
7226 | object files are correct. | |
7227 | ||
7228 | @item | |
7229 | If you are an experienced user of linkers, your suggestions for | |
ff5dcc92 | 7230 | improvement of @command{ld} are welcome in any case. |
252b5132 RH |
7231 | @end itemize |
7232 | ||
7233 | @node Bug Reporting | |
36f63dca | 7234 | @section How to Report Bugs |
252b5132 | 7235 | @cindex bug reports |
ff5dcc92 | 7236 | @cindex @command{ld} bugs, reporting |
252b5132 RH |
7237 | |
7238 | A number of companies and individuals offer support for @sc{gnu} | |
ff5dcc92 | 7239 | products. If you obtained @command{ld} from a support organization, we |
252b5132 RH |
7240 | recommend you contact that organization first. |
7241 | ||
7242 | You can find contact information for many support companies and | |
7243 | individuals in the file @file{etc/SERVICE} in the @sc{gnu} Emacs | |
7244 | distribution. | |
7245 | ||
ad22bfe8 | 7246 | @ifset BUGURL |
ff5dcc92 | 7247 | Otherwise, send bug reports for @command{ld} to |
ad22bfe8 JM |
7248 | @value{BUGURL}. |
7249 | @end ifset | |
252b5132 RH |
7250 | |
7251 | The fundamental principle of reporting bugs usefully is this: | |
7252 | @strong{report all the facts}. If you are not sure whether to state a | |
7253 | fact or leave it out, state it! | |
7254 | ||
7255 | Often people omit facts because they think they know what causes the | |
7256 | problem and assume that some details do not matter. Thus, you might | |
b553b183 NC |
7257 | assume that the name of a symbol you use in an example does not |
7258 | matter. Well, probably it does not, but one cannot be sure. Perhaps | |
7259 | the bug is a stray memory reference which happens to fetch from the | |
7260 | location where that name is stored in memory; perhaps, if the name | |
7261 | were different, the contents of that location would fool the linker | |
7262 | into doing the right thing despite the bug. Play it safe and give a | |
7263 | specific, complete example. That is the easiest thing for you to do, | |
c0065db7 | 7264 | and the most helpful. |
b553b183 NC |
7265 | |
7266 | Keep in mind that the purpose of a bug report is to enable us to fix | |
7267 | the bug if it is new to us. Therefore, always write your bug reports | |
7268 | on the assumption that the bug has not been reported previously. | |
252b5132 RH |
7269 | |
7270 | Sometimes people give a few sketchy facts and ask, ``Does this ring a | |
36f63dca NC |
7271 | bell?'' This cannot help us fix a bug, so it is basically useless. We |
7272 | respond by asking for enough details to enable us to investigate. | |
7273 | You might as well expedite matters by sending them to begin with. | |
252b5132 RH |
7274 | |
7275 | To enable us to fix the bug, you should include all these things: | |
7276 | ||
7277 | @itemize @bullet | |
7278 | @item | |
ff5dcc92 | 7279 | The version of @command{ld}. @command{ld} announces it if you start it with |
252b5132 RH |
7280 | the @samp{--version} argument. |
7281 | ||
7282 | Without this, we will not know whether there is any point in looking for | |
ff5dcc92 | 7283 | the bug in the current version of @command{ld}. |
252b5132 RH |
7284 | |
7285 | @item | |
ff5dcc92 | 7286 | Any patches you may have applied to the @command{ld} source, including any |
252b5132 RH |
7287 | patches made to the @code{BFD} library. |
7288 | ||
7289 | @item | |
7290 | The type of machine you are using, and the operating system name and | |
7291 | version number. | |
7292 | ||
7293 | @item | |
ff5dcc92 | 7294 | What compiler (and its version) was used to compile @command{ld}---e.g. |
252b5132 RH |
7295 | ``@code{gcc-2.7}''. |
7296 | ||
7297 | @item | |
7298 | The command arguments you gave the linker to link your example and | |
7299 | observe the bug. To guarantee you will not omit something important, | |
7300 | list them all. A copy of the Makefile (or the output from make) is | |
7301 | sufficient. | |
7302 | ||
7303 | If we were to try to guess the arguments, we would probably guess wrong | |
7304 | and then we might not encounter the bug. | |
7305 | ||
7306 | @item | |
7307 | A complete input file, or set of input files, that will reproduce the | |
b553b183 NC |
7308 | bug. It is generally most helpful to send the actual object files |
7309 | provided that they are reasonably small. Say no more than 10K. For | |
7310 | bigger files you can either make them available by FTP or HTTP or else | |
7311 | state that you are willing to send the object file(s) to whomever | |
7312 | requests them. (Note - your email will be going to a mailing list, so | |
7313 | we do not want to clog it up with large attachments). But small | |
7314 | attachments are best. | |
252b5132 RH |
7315 | |
7316 | If the source files were assembled using @code{gas} or compiled using | |
7317 | @code{gcc}, then it may be OK to send the source files rather than the | |
7318 | object files. In this case, be sure to say exactly what version of | |
7319 | @code{gas} or @code{gcc} was used to produce the object files. Also say | |
7320 | how @code{gas} or @code{gcc} were configured. | |
7321 | ||
7322 | @item | |
7323 | A description of what behavior you observe that you believe is | |
7324 | incorrect. For example, ``It gets a fatal signal.'' | |
7325 | ||
ff5dcc92 | 7326 | Of course, if the bug is that @command{ld} gets a fatal signal, then we |
252b5132 RH |
7327 | will certainly notice it. But if the bug is incorrect output, we might |
7328 | not notice unless it is glaringly wrong. You might as well not give us | |
7329 | a chance to make a mistake. | |
7330 | ||
7331 | Even if the problem you experience is a fatal signal, you should still | |
7332 | say so explicitly. Suppose something strange is going on, such as, your | |
b45619c0 | 7333 | copy of @command{ld} is out of sync, or you have encountered a bug in the |
252b5132 RH |
7334 | C library on your system. (This has happened!) Your copy might crash |
7335 | and ours would not. If you told us to expect a crash, then when ours | |
7336 | fails to crash, we would know that the bug was not happening for us. If | |
7337 | you had not told us to expect a crash, then we would not be able to draw | |
7338 | any conclusion from our observations. | |
7339 | ||
7340 | @item | |
ff5dcc92 | 7341 | If you wish to suggest changes to the @command{ld} source, send us context |
252b5132 RH |
7342 | diffs, as generated by @code{diff} with the @samp{-u}, @samp{-c}, or |
7343 | @samp{-p} option. Always send diffs from the old file to the new file. | |
ff5dcc92 | 7344 | If you even discuss something in the @command{ld} source, refer to it by |
252b5132 RH |
7345 | context, not by line number. |
7346 | ||
7347 | The line numbers in our development sources will not match those in your | |
7348 | sources. Your line numbers would convey no useful information to us. | |
7349 | @end itemize | |
7350 | ||
7351 | Here are some things that are not necessary: | |
7352 | ||
7353 | @itemize @bullet | |
7354 | @item | |
7355 | A description of the envelope of the bug. | |
7356 | ||
7357 | Often people who encounter a bug spend a lot of time investigating | |
7358 | which changes to the input file will make the bug go away and which | |
7359 | changes will not affect it. | |
7360 | ||
7361 | This is often time consuming and not very useful, because the way we | |
7362 | will find the bug is by running a single example under the debugger | |
7363 | with breakpoints, not by pure deduction from a series of examples. | |
7364 | We recommend that you save your time for something else. | |
7365 | ||
7366 | Of course, if you can find a simpler example to report @emph{instead} | |
7367 | of the original one, that is a convenience for us. Errors in the | |
7368 | output will be easier to spot, running under the debugger will take | |
7369 | less time, and so on. | |
7370 | ||
7371 | However, simplification is not vital; if you do not want to do this, | |
7372 | report the bug anyway and send us the entire test case you used. | |
7373 | ||
7374 | @item | |
7375 | A patch for the bug. | |
7376 | ||
7377 | A patch for the bug does help us if it is a good one. But do not omit | |
7378 | the necessary information, such as the test case, on the assumption that | |
7379 | a patch is all we need. We might see problems with your patch and decide | |
7380 | to fix the problem another way, or we might not understand it at all. | |
7381 | ||
ff5dcc92 | 7382 | Sometimes with a program as complicated as @command{ld} it is very hard to |
252b5132 RH |
7383 | construct an example that will make the program follow a certain path |
7384 | through the code. If you do not send us the example, we will not be | |
7385 | able to construct one, so we will not be able to verify that the bug is | |
7386 | fixed. | |
7387 | ||
7388 | And if we cannot understand what bug you are trying to fix, or why your | |
7389 | patch should be an improvement, we will not install it. A test case will | |
7390 | help us to understand. | |
7391 | ||
7392 | @item | |
7393 | A guess about what the bug is or what it depends on. | |
7394 | ||
7395 | Such guesses are usually wrong. Even we cannot guess right about such | |
7396 | things without first using the debugger to find the facts. | |
7397 | @end itemize | |
7398 | ||
7399 | @node MRI | |
7400 | @appendix MRI Compatible Script Files | |
7401 | @cindex MRI compatibility | |
ff5dcc92 SC |
7402 | To aid users making the transition to @sc{gnu} @command{ld} from the MRI |
7403 | linker, @command{ld} can use MRI compatible linker scripts as an | |
252b5132 RH |
7404 | alternative to the more general-purpose linker scripting language |
7405 | described in @ref{Scripts}. MRI compatible linker scripts have a much | |
7406 | simpler command set than the scripting language otherwise used with | |
ff5dcc92 | 7407 | @command{ld}. @sc{gnu} @command{ld} supports the most commonly used MRI |
252b5132 RH |
7408 | linker commands; these commands are described here. |
7409 | ||
7410 | In general, MRI scripts aren't of much use with the @code{a.out} object | |
7411 | file format, since it only has three sections and MRI scripts lack some | |
7412 | features to make use of them. | |
7413 | ||
7414 | You can specify a file containing an MRI-compatible script using the | |
7415 | @samp{-c} command-line option. | |
7416 | ||
7417 | Each command in an MRI-compatible script occupies its own line; each | |
7418 | command line starts with the keyword that identifies the command (though | |
7419 | blank lines are also allowed for punctuation). If a line of an | |
ff5dcc92 | 7420 | MRI-compatible script begins with an unrecognized keyword, @command{ld} |
252b5132 RH |
7421 | issues a warning message, but continues processing the script. |
7422 | ||
7423 | Lines beginning with @samp{*} are comments. | |
7424 | ||
7425 | You can write these commands using all upper-case letters, or all | |
7426 | lower case; for example, @samp{chip} is the same as @samp{CHIP}. | |
7427 | The following list shows only the upper-case form of each command. | |
7428 | ||
7429 | @table @code | |
7430 | @cindex @code{ABSOLUTE} (MRI) | |
7431 | @item ABSOLUTE @var{secname} | |
7432 | @itemx ABSOLUTE @var{secname}, @var{secname}, @dots{} @var{secname} | |
ff5dcc92 | 7433 | Normally, @command{ld} includes in the output file all sections from all |
252b5132 RH |
7434 | the input files. However, in an MRI-compatible script, you can use the |
7435 | @code{ABSOLUTE} command to restrict the sections that will be present in | |
7436 | your output program. If the @code{ABSOLUTE} command is used at all in a | |
7437 | script, then only the sections named explicitly in @code{ABSOLUTE} | |
7438 | commands will appear in the linker output. You can still use other | |
7439 | input sections (whatever you select on the command line, or using | |
7440 | @code{LOAD}) to resolve addresses in the output file. | |
7441 | ||
7442 | @cindex @code{ALIAS} (MRI) | |
7443 | @item ALIAS @var{out-secname}, @var{in-secname} | |
7444 | Use this command to place the data from input section @var{in-secname} | |
7445 | in a section called @var{out-secname} in the linker output file. | |
7446 | ||
7447 | @var{in-secname} may be an integer. | |
7448 | ||
7449 | @cindex @code{ALIGN} (MRI) | |
7450 | @item ALIGN @var{secname} = @var{expression} | |
7451 | Align the section called @var{secname} to @var{expression}. The | |
7452 | @var{expression} should be a power of two. | |
7453 | ||
7454 | @cindex @code{BASE} (MRI) | |
7455 | @item BASE @var{expression} | |
7456 | Use the value of @var{expression} as the lowest address (other than | |
7457 | absolute addresses) in the output file. | |
7458 | ||
7459 | @cindex @code{CHIP} (MRI) | |
7460 | @item CHIP @var{expression} | |
7461 | @itemx CHIP @var{expression}, @var{expression} | |
7462 | This command does nothing; it is accepted only for compatibility. | |
7463 | ||
7464 | @cindex @code{END} (MRI) | |
7465 | @item END | |
7466 | This command does nothing whatever; it's only accepted for compatibility. | |
7467 | ||
7468 | @cindex @code{FORMAT} (MRI) | |
7469 | @item FORMAT @var{output-format} | |
7470 | Similar to the @code{OUTPUT_FORMAT} command in the more general linker | |
a1ab1d2a | 7471 | language, but restricted to one of these output formats: |
252b5132 RH |
7472 | |
7473 | @enumerate | |
a1ab1d2a | 7474 | @item |
252b5132 RH |
7475 | S-records, if @var{output-format} is @samp{S} |
7476 | ||
7477 | @item | |
7478 | IEEE, if @var{output-format} is @samp{IEEE} | |
7479 | ||
7480 | @item | |
7481 | COFF (the @samp{coff-m68k} variant in BFD), if @var{output-format} is | |
7482 | @samp{COFF} | |
7483 | @end enumerate | |
7484 | ||
7485 | @cindex @code{LIST} (MRI) | |
7486 | @item LIST @var{anything}@dots{} | |
7487 | Print (to the standard output file) a link map, as produced by the | |
ff5dcc92 | 7488 | @command{ld} command-line option @samp{-M}. |
252b5132 RH |
7489 | |
7490 | The keyword @code{LIST} may be followed by anything on the | |
7491 | same line, with no change in its effect. | |
7492 | ||
7493 | @cindex @code{LOAD} (MRI) | |
7494 | @item LOAD @var{filename} | |
7495 | @itemx LOAD @var{filename}, @var{filename}, @dots{} @var{filename} | |
7496 | Include one or more object file @var{filename} in the link; this has the | |
ff5dcc92 | 7497 | same effect as specifying @var{filename} directly on the @command{ld} |
252b5132 RH |
7498 | command line. |
7499 | ||
7500 | @cindex @code{NAME} (MRI) | |
7501 | @item NAME @var{output-name} | |
ff5dcc92 | 7502 | @var{output-name} is the name for the program produced by @command{ld}; the |
252b5132 RH |
7503 | MRI-compatible command @code{NAME} is equivalent to the command-line |
7504 | option @samp{-o} or the general script language command @code{OUTPUT}. | |
7505 | ||
7506 | @cindex @code{ORDER} (MRI) | |
7507 | @item ORDER @var{secname}, @var{secname}, @dots{} @var{secname} | |
7508 | @itemx ORDER @var{secname} @var{secname} @var{secname} | |
ff5dcc92 | 7509 | Normally, @command{ld} orders the sections in its output file in the |
252b5132 RH |
7510 | order in which they first appear in the input files. In an MRI-compatible |
7511 | script, you can override this ordering with the @code{ORDER} command. The | |
7512 | sections you list with @code{ORDER} will appear first in your output | |
7513 | file, in the order specified. | |
7514 | ||
7515 | @cindex @code{PUBLIC} (MRI) | |
7516 | @item PUBLIC @var{name}=@var{expression} | |
7517 | @itemx PUBLIC @var{name},@var{expression} | |
7518 | @itemx PUBLIC @var{name} @var{expression} | |
7519 | Supply a value (@var{expression}) for external symbol | |
7520 | @var{name} used in the linker input files. | |
7521 | ||
7522 | @cindex @code{SECT} (MRI) | |
7523 | @item SECT @var{secname}, @var{expression} | |
7524 | @itemx SECT @var{secname}=@var{expression} | |
7525 | @itemx SECT @var{secname} @var{expression} | |
7526 | You can use any of these three forms of the @code{SECT} command to | |
7527 | specify the start address (@var{expression}) for section @var{secname}. | |
7528 | If you have more than one @code{SECT} statement for the same | |
7529 | @var{secname}, only the @emph{first} sets the start address. | |
7530 | @end table | |
7531 | ||
793c5807 NC |
7532 | @node GNU Free Documentation License |
7533 | @appendix GNU Free Documentation License | |
36f63dca | 7534 | @include fdl.texi |
704c465c | 7535 | |
370b66a1 CD |
7536 | @node LD Index |
7537 | @unnumbered LD Index | |
252b5132 RH |
7538 | |
7539 | @printindex cp | |
7540 | ||
7541 | @tex | |
7542 | % I think something like @colophon should be in texinfo. In the | |
7543 | % meantime: | |
7544 | \long\def\colophon{\hbox to0pt{}\vfill | |
7545 | \centerline{The body of this manual is set in} | |
7546 | \centerline{\fontname\tenrm,} | |
7547 | \centerline{with headings in {\bf\fontname\tenbf}} | |
7548 | \centerline{and examples in {\tt\fontname\tentt}.} | |
7549 | \centerline{{\it\fontname\tenit\/} and} | |
7550 | \centerline{{\sl\fontname\tensl\/}} | |
7551 | \centerline{are used for emphasis.}\vfill} | |
7552 | \page\colophon | |
7553 | % Blame: doc@cygnus.com, 28mar91. | |
7554 | @end tex | |
7555 | ||
252b5132 | 7556 | @bye |