Commit | Line | Data |
---|---|---|
252b5132 RH |
1 | \input texinfo |
2 | @setfilename ld.info | |
219d1afa | 3 | @c Copyright (C) 1991-2018 Free Software Foundation, Inc. |
252b5132 | 4 | @syncodeindex ky cp |
dff70155 | 5 | @c man begin INCLUDE |
252b5132 RH |
6 | @include configdoc.texi |
7 | @c (configdoc.texi is generated by the Makefile) | |
c428fa83 | 8 | @include bfdver.texi |
dff70155 | 9 | @c man end |
252b5132 RH |
10 | |
11 | @c @smallbook | |
12 | ||
ff5dcc92 SC |
13 | @macro gcctabopt{body} |
14 | @code{\body\} | |
15 | @end macro | |
16 | ||
0285c67d NC |
17 | @c man begin NAME |
18 | @ifset man | |
19 | @c Configure for the generation of man pages | |
20 | @set UsesEnvVars | |
21 | @set GENERIC | |
0285c67d | 22 | @set ARM |
ac145307 | 23 | @set C6X |
b8891f8d | 24 | @set CSKY |
49fa1e15 | 25 | @set H8300 |
0285c67d | 26 | @set HPPA |
0285c67d | 27 | @set M68HC11 |
7fb9f789 | 28 | @set M68K |
833794fc | 29 | @set MIPS |
3c3bdf30 | 30 | @set MMIX |
2469cfa2 | 31 | @set MSP430 |
35c08157 | 32 | @set NDS32 |
78058a5e | 33 | @set NIOSII |
2a60a7a8 AM |
34 | @set POWERPC |
35 | @set POWERPC64 | |
49fa1e15 | 36 | @set Renesas |
b4cbbe8f | 37 | @set S/390 |
49fa1e15 AM |
38 | @set SPU |
39 | @set TICOFF | |
2ca22b03 | 40 | @set WIN32 |
e0001a05 | 41 | @set XTENSA |
0285c67d NC |
42 | @end ifset |
43 | @c man end | |
44 | ||
9160ea82 AM |
45 | @ifnottex |
46 | @dircategory Software development | |
47 | @direntry | |
252b5132 | 48 | * Ld: (ld). The GNU linker. |
9160ea82 AM |
49 | @end direntry |
50 | @end ifnottex | |
252b5132 | 51 | |
0e9517a9 | 52 | @copying |
e49e529d JM |
53 | This file documents the @sc{gnu} linker LD |
54 | @ifset VERSION_PACKAGE | |
55 | @value{VERSION_PACKAGE} | |
56 | @end ifset | |
57 | version @value{VERSION}. | |
252b5132 | 58 | |
219d1afa | 59 | Copyright @copyright{} 1991-2018 Free Software Foundation, Inc. |
252b5132 | 60 | |
cf055d54 | 61 | Permission is granted to copy, distribute and/or modify this document |
793c5807 | 62 | under the terms of the GNU Free Documentation License, Version 1.3 |
cf055d54 NC |
63 | or any later version published by the Free Software Foundation; |
64 | with no Invariant Sections, with no Front-Cover Texts, and with no | |
65 | Back-Cover Texts. A copy of the license is included in the | |
36f63dca | 66 | section entitled ``GNU Free Documentation License''. |
0e9517a9 | 67 | @end copying |
252b5132 RH |
68 | @iftex |
69 | @finalout | |
70 | @setchapternewpage odd | |
71ba23f6 | 71 | @settitle The GNU linker |
252b5132 | 72 | @titlepage |
71ba23f6 | 73 | @title The GNU linker |
252b5132 | 74 | @sp 1 |
e49e529d JM |
75 | @subtitle @code{ld} |
76 | @ifset VERSION_PACKAGE | |
77 | @subtitle @value{VERSION_PACKAGE} | |
78 | @end ifset | |
252b5132 RH |
79 | @subtitle Version @value{VERSION} |
80 | @author Steve Chamberlain | |
81 | @author Ian Lance Taylor | |
252b5132 RH |
82 | @page |
83 | ||
84 | @tex | |
85 | {\parskip=0pt | |
704c465c NC |
86 | \hfill Red Hat Inc\par |
87 | \hfill nickc\@credhat.com, doc\@redhat.com\par | |
71ba23f6 | 88 | \hfill {\it The GNU linker}\par |
252b5132 RH |
89 | \hfill Edited by Jeffrey Osier (jeffrey\@cygnus.com)\par |
90 | } | |
91 | \global\parindent=0pt % Steve likes it this way. | |
92 | @end tex | |
93 | ||
94 | @vskip 0pt plus 1filll | |
0285c67d | 95 | @c man begin COPYRIGHT |
219d1afa | 96 | Copyright @copyright{} 1991-2018 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 | 137 | @end ifset |
36f63dca NC |
138 | @ifset ARM |
139 | * ARM:: ld and the ARM family | |
140 | @end ifset | |
93fd0973 SC |
141 | @ifset M68HC11 |
142 | * M68HC11/68HC12:: ld and the Motorola 68HC11 and 68HC12 families | |
143 | @end ifset | |
7ca01ed9 NC |
144 | @ifset HPPA |
145 | * HPPA ELF32:: ld and HPPA 32-bit ELF | |
146 | @end ifset | |
7fb9f789 NC |
147 | @ifset M68K |
148 | * M68K:: ld and Motorola 68K family | |
149 | @end ifset | |
833794fc MR |
150 | @ifset MIPS |
151 | * MIPS:: ld and MIPS 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 | |
b4cbbe8f AK |
159 | @ifset S/390 |
160 | * S/390 ELF:: ld and S/390 ELF Support | |
161 | @end ifset | |
49fa1e15 AM |
162 | @ifset SPU |
163 | * SPU ELF:: ld and SPU ELF Support | |
164 | @end ifset | |
74459f0e TW |
165 | @ifset TICOFF |
166 | * TI COFF:: ld and the TI COFF | |
167 | @end ifset | |
2ca22b03 NC |
168 | @ifset WIN32 |
169 | * Win32:: ld and WIN32 (cygwin/mingw) | |
170 | @end ifset | |
e0001a05 NC |
171 | @ifset XTENSA |
172 | * Xtensa:: ld and Xtensa Processors | |
173 | @end ifset | |
252b5132 RH |
174 | @end ifclear |
175 | @ifclear SingleFormat | |
176 | * BFD:: BFD | |
177 | @end ifclear | |
178 | @c Following blank line required for remaining bug in makeinfo conds/menus | |
179 | ||
180 | * Reporting Bugs:: Reporting Bugs | |
181 | * MRI:: MRI Compatible Script Files | |
704c465c | 182 | * GNU Free Documentation License:: GNU Free Documentation License |
370b66a1 | 183 | * LD Index:: LD Index |
252b5132 | 184 | @end menu |
84ec0e6d | 185 | @end ifnottex |
252b5132 RH |
186 | |
187 | @node Overview | |
188 | @chapter Overview | |
189 | ||
190 | @cindex @sc{gnu} linker | |
191 | @cindex what is this? | |
0285c67d | 192 | |
0879a67a | 193 | @ifset man |
0285c67d | 194 | @c man begin SYNOPSIS |
ff5dcc92 | 195 | ld [@b{options}] @var{objfile} @dots{} |
0285c67d NC |
196 | @c man end |
197 | ||
198 | @c man begin SEEALSO | |
199 | ar(1), nm(1), objcopy(1), objdump(1), readelf(1) and | |
200 | the Info entries for @file{binutils} and | |
201 | @file{ld}. | |
202 | @c man end | |
203 | @end ifset | |
204 | ||
205 | @c man begin DESCRIPTION | |
206 | ||
ff5dcc92 | 207 | @command{ld} combines a number of object and archive files, relocates |
252b5132 | 208 | their data and ties up symbol references. Usually the last step in |
ff5dcc92 | 209 | compiling a program is to run @command{ld}. |
252b5132 | 210 | |
ff5dcc92 | 211 | @command{ld} accepts Linker Command Language files written in |
252b5132 RH |
212 | a superset of AT&T's Link Editor Command Language syntax, |
213 | to provide explicit and total control over the linking process. | |
214 | ||
0285c67d NC |
215 | @ifset man |
216 | @c For the man only | |
ece2d90e | 217 | This man page does not describe the command language; see the |
71ba23f6 NC |
218 | @command{ld} entry in @code{info} for full details on the command |
219 | language and on other aspects of the GNU linker. | |
0285c67d NC |
220 | @end ifset |
221 | ||
252b5132 | 222 | @ifclear SingleFormat |
ff5dcc92 SC |
223 | This version of @command{ld} uses the general purpose BFD libraries |
224 | to operate on object files. This allows @command{ld} to read, combine, and | |
252b5132 RH |
225 | write object files in many different formats---for example, COFF or |
226 | @code{a.out}. Different formats may be linked together to produce any | |
227 | available kind of object file. @xref{BFD}, for more information. | |
228 | @end ifclear | |
229 | ||
230 | Aside from its flexibility, the @sc{gnu} linker is more helpful than other | |
231 | linkers in providing diagnostic information. Many linkers abandon | |
232 | execution immediately upon encountering an error; whenever possible, | |
ff5dcc92 | 233 | @command{ld} continues executing, allowing you to identify other errors |
252b5132 RH |
234 | (or, in some cases, to get an output file in spite of the error). |
235 | ||
0285c67d NC |
236 | @c man end |
237 | ||
252b5132 RH |
238 | @node Invocation |
239 | @chapter Invocation | |
240 | ||
0285c67d NC |
241 | @c man begin DESCRIPTION |
242 | ||
ff5dcc92 | 243 | The @sc{gnu} linker @command{ld} is meant to cover a broad range of situations, |
252b5132 RH |
244 | and to be as compatible as possible with other linkers. As a result, |
245 | you have many choices to control its behavior. | |
246 | ||
0285c67d NC |
247 | @c man end |
248 | ||
252b5132 RH |
249 | @ifset UsesEnvVars |
250 | @menu | |
a05a5b64 | 251 | * Options:: Command-line Options |
252b5132 RH |
252 | * Environment:: Environment Variables |
253 | @end menu | |
254 | ||
255 | @node Options | |
a05a5b64 | 256 | @section Command-line Options |
252b5132 RH |
257 | @end ifset |
258 | ||
259 | @cindex command line | |
260 | @cindex options | |
0285c67d NC |
261 | |
262 | @c man begin OPTIONS | |
263 | ||
252b5132 RH |
264 | The linker supports a plethora of command-line options, but in actual |
265 | practice few of them are used in any particular context. | |
266 | @cindex standard Unix system | |
ff5dcc92 | 267 | For instance, a frequent use of @command{ld} is to link standard Unix |
252b5132 RH |
268 | object files on a standard, supported Unix system. On such a system, to |
269 | link a file @code{hello.o}: | |
270 | ||
271 | @smallexample | |
272 | ld -o @var{output} /lib/crt0.o hello.o -lc | |
273 | @end smallexample | |
274 | ||
ff5dcc92 | 275 | This tells @command{ld} to produce a file called @var{output} as the |
252b5132 RH |
276 | result of linking the file @code{/lib/crt0.o} with @code{hello.o} and |
277 | the library @code{libc.a}, which will come from the standard search | |
278 | directories. (See the discussion of the @samp{-l} option below.) | |
279 | ||
ff5dcc92 | 280 | Some of the command-line options to @command{ld} may be specified at any |
511ab9e9 ILT |
281 | point in the command line. However, options which refer to files, such |
282 | as @samp{-l} or @samp{-T}, cause the file to be read at the point at | |
283 | which the option appears in the command line, relative to the object | |
284 | files and other file options. Repeating non-file options with a | |
285 | different argument will either have no further effect, or override prior | |
252b5132 RH |
286 | occurrences (those further to the left on the command line) of that |
287 | option. Options which may be meaningfully specified more than once are | |
288 | noted in the descriptions below. | |
289 | ||
290 | @cindex object files | |
511ab9e9 ILT |
291 | Non-option arguments are object files or archives which are to be linked |
292 | together. They may follow, precede, or be mixed in with command-line | |
293 | options, except that an object file argument may not be placed between | |
294 | an option and its argument. | |
252b5132 RH |
295 | |
296 | Usually the linker is invoked with at least one object file, but you can | |
297 | specify other forms of binary input files using @samp{-l}, @samp{-R}, | |
298 | and the script command language. If @emph{no} binary input files at all | |
299 | are specified, the linker does not produce any output, and issues the | |
300 | message @samp{No input files}. | |
301 | ||
36f63dca | 302 | If the linker cannot recognize the format of an object file, it will |
252b5132 RH |
303 | assume that it is a linker script. A script specified in this way |
304 | augments the main linker script used for the link (either the default | |
305 | linker script or the one specified by using @samp{-T}). This feature | |
306 | permits the linker to link against a file which appears to be an object | |
307 | or an archive, but actually merely defines some symbol values, or uses | |
53d25da6 AM |
308 | @code{INPUT} or @code{GROUP} to load other objects. Specifying a |
309 | script in this way merely augments the main linker script, with the | |
310 | extra commands placed after the main script; use the @samp{-T} option | |
311 | to replace the default linker script entirely, but note the effect of | |
312 | the @code{INSERT} command. @xref{Scripts}. | |
252b5132 RH |
313 | |
314 | For options whose names are a single letter, | |
315 | option arguments must either follow the option letter without intervening | |
316 | whitespace, or be given as separate arguments immediately following the | |
317 | option that requires them. | |
318 | ||
319 | For options whose names are multiple letters, either one dash or two can | |
e4897a32 | 320 | precede the option name; for example, @samp{-trace-symbol} and |
36f63dca | 321 | @samp{--trace-symbol} are equivalent. Note---there is one exception to |
e4897a32 | 322 | this rule. Multiple letter options that start with a lower case 'o' can |
ba1be17e | 323 | only be preceded by two dashes. This is to reduce confusion with the |
e4897a32 NC |
324 | @samp{-o} option. So for example @samp{-omagic} sets the output file |
325 | name to @samp{magic} whereas @samp{--omagic} sets the NMAGIC flag on the | |
326 | output. | |
327 | ||
328 | Arguments to multiple-letter options must either be separated from the | |
329 | option name by an equals sign, or be given as separate arguments | |
330 | immediately following the option that requires them. For example, | |
331 | @samp{--trace-symbol foo} and @samp{--trace-symbol=foo} are equivalent. | |
332 | Unique abbreviations of the names of multiple-letter options are | |
333 | accepted. | |
252b5132 | 334 | |
36f63dca | 335 | Note---if the linker is being invoked indirectly, via a compiler driver |
a05a5b64 | 336 | (e.g. @samp{gcc}) then all the linker command-line options should be |
fa19fce0 NC |
337 | prefixed by @samp{-Wl,} (or whatever is appropriate for the particular |
338 | compiler driver) like this: | |
4e53152f NC |
339 | |
340 | @smallexample | |
2509a395 | 341 | gcc -Wl,--start-group foo.o bar.o -Wl,--end-group |
4e53152f NC |
342 | @end smallexample |
343 | ||
344 | This is important, because otherwise the compiler driver program may | |
2509a395 SL |
345 | silently drop the linker options, resulting in a bad link. Confusion |
346 | may also arise when passing options that require values through a | |
347 | driver, as the use of a space between option and argument acts as | |
348 | a separator, and causes the driver to pass only the option to the linker | |
349 | and the argument to the compiler. In this case, it is simplest to use | |
350 | the joined forms of both single- and multiple-letter options, such as: | |
351 | ||
352 | @smallexample | |
353 | gcc foo.o bar.o -Wl,-eENTRY -Wl,-Map=a.map | |
354 | @end smallexample | |
4e53152f | 355 | |
a05a5b64 | 356 | Here is a table of the generic command-line switches accepted by the GNU |
4e53152f NC |
357 | linker: |
358 | ||
ff5dcc92 | 359 | @table @gcctabopt |
38fc1cb1 | 360 | @include at-file.texi |
dff70155 | 361 | |
2509a395 SL |
362 | @kindex -a @var{keyword} |
363 | @item -a @var{keyword} | |
252b5132 RH |
364 | This option is supported for HP/UX compatibility. The @var{keyword} |
365 | argument must be one of the strings @samp{archive}, @samp{shared}, or | |
366 | @samp{default}. @samp{-aarchive} is functionally equivalent to | |
367 | @samp{-Bstatic}, and the other two keywords are functionally equivalent | |
368 | to @samp{-Bdynamic}. This option may be used any number of times. | |
369 | ||
7ee314fa AM |
370 | @kindex --audit @var{AUDITLIB} |
371 | @item --audit @var{AUDITLIB} | |
372 | Adds @var{AUDITLIB} to the @code{DT_AUDIT} entry of the dynamic section. | |
373 | @var{AUDITLIB} is not checked for existence, nor will it use the DT_SONAME | |
374 | specified in the library. If specified multiple times @code{DT_AUDIT} | |
375 | will contain a colon separated list of audit interfaces to use. If the linker | |
376 | finds an object with an audit entry while searching for shared libraries, | |
9d5777a3 | 377 | it will add a corresponding @code{DT_DEPAUDIT} entry in the output file. |
7ee314fa | 378 | This option is only meaningful on ELF platforms supporting the rtld-audit |
9d5777a3 | 379 | interface. |
7ee314fa | 380 | |
252b5132 RH |
381 | @ifclear SingleFormat |
382 | @cindex binary input format | |
383 | @kindex -b @var{format} | |
384 | @kindex --format=@var{format} | |
385 | @cindex input format | |
386 | @cindex input format | |
387 | @item -b @var{input-format} | |
388 | @itemx --format=@var{input-format} | |
ff5dcc92 SC |
389 | @command{ld} may be configured to support more than one kind of object |
390 | file. If your @command{ld} is configured this way, you can use the | |
252b5132 | 391 | @samp{-b} option to specify the binary format for input object files |
ff5dcc92 | 392 | that follow this option on the command line. Even when @command{ld} is |
252b5132 | 393 | configured to support alternative object formats, you don't usually need |
ff5dcc92 | 394 | to specify this, as @command{ld} should be configured to expect as a |
252b5132 RH |
395 | default input format the most usual format on each machine. |
396 | @var{input-format} is a text string, the name of a particular format | |
397 | supported by the BFD libraries. (You can list the available binary | |
398 | formats with @samp{objdump -i}.) | |
399 | @xref{BFD}. | |
400 | ||
401 | You may want to use this option if you are linking files with an unusual | |
402 | binary format. You can also use @samp{-b} to switch formats explicitly (when | |
403 | linking object files of different formats), by including | |
404 | @samp{-b @var{input-format}} before each group of object files in a | |
a1ab1d2a | 405 | particular format. |
252b5132 RH |
406 | |
407 | The default format is taken from the environment variable | |
408 | @code{GNUTARGET}. | |
409 | @ifset UsesEnvVars | |
410 | @xref{Environment}. | |
411 | @end ifset | |
412 | You can also define the input format from a script, using the command | |
0285c67d NC |
413 | @code{TARGET}; |
414 | @ifclear man | |
415 | see @ref{Format Commands}. | |
416 | @end ifclear | |
252b5132 RH |
417 | @end ifclear |
418 | ||
419 | @kindex -c @var{MRI-cmdfile} | |
420 | @kindex --mri-script=@var{MRI-cmdfile} | |
421 | @cindex compatibility, MRI | |
422 | @item -c @var{MRI-commandfile} | |
423 | @itemx --mri-script=@var{MRI-commandfile} | |
ff5dcc92 | 424 | For compatibility with linkers produced by MRI, @command{ld} accepts script |
252b5132 | 425 | files written in an alternate, restricted command language, described in |
0285c67d NC |
426 | @ifclear man |
427 | @ref{MRI,,MRI Compatible Script Files}. | |
428 | @end ifclear | |
429 | @ifset man | |
430 | the MRI Compatible Script Files section of GNU ld documentation. | |
431 | @end ifset | |
432 | Introduce MRI script files with | |
252b5132 | 433 | the option @samp{-c}; use the @samp{-T} option to run linker |
ff5dcc92 SC |
434 | scripts written in the general-purpose @command{ld} scripting language. |
435 | If @var{MRI-cmdfile} does not exist, @command{ld} looks for it in the directories | |
252b5132 RH |
436 | specified by any @samp{-L} options. |
437 | ||
438 | @cindex common allocation | |
439 | @kindex -d | |
440 | @kindex -dc | |
441 | @kindex -dp | |
a1ab1d2a | 442 | @item -d |
252b5132 RH |
443 | @itemx -dc |
444 | @itemx -dp | |
445 | These three options are equivalent; multiple forms are supported for | |
446 | compatibility with other linkers. They assign space to common symbols | |
447 | even if a relocatable output file is specified (with @samp{-r}). The | |
448 | script command @code{FORCE_COMMON_ALLOCATION} has the same effect. | |
449 | @xref{Miscellaneous Commands}. | |
450 | ||
7ee314fa AM |
451 | @kindex --depaudit @var{AUDITLIB} |
452 | @kindex -P @var{AUDITLIB} | |
453 | @item --depaudit @var{AUDITLIB} | |
454 | @itemx -P @var{AUDITLIB} | |
455 | Adds @var{AUDITLIB} to the @code{DT_DEPAUDIT} entry of the dynamic section. | |
456 | @var{AUDITLIB} is not checked for existence, nor will it use the DT_SONAME | |
457 | specified in the library. If specified multiple times @code{DT_DEPAUDIT} | |
458 | will contain a colon separated list of audit interfaces to use. This | |
459 | option is only meaningful on ELF platforms supporting the rtld-audit interface. | |
9d5777a3 | 460 | The -P option is provided for Solaris compatibility. |
7ee314fa | 461 | |
252b5132 RH |
462 | @cindex entry point, from command line |
463 | @kindex -e @var{entry} | |
464 | @kindex --entry=@var{entry} | |
a1ab1d2a | 465 | @item -e @var{entry} |
252b5132 RH |
466 | @itemx --entry=@var{entry} |
467 | Use @var{entry} as the explicit symbol for beginning execution of your | |
468 | program, rather than the default entry point. If there is no symbol | |
469 | named @var{entry}, the linker will try to parse @var{entry} as a number, | |
470 | and use that as the entry address (the number will be interpreted in | |
471 | base 10; you may use a leading @samp{0x} for base 16, or a leading | |
472 | @samp{0} for base 8). @xref{Entry Point}, for a discussion of defaults | |
473 | and other ways of specifying the entry point. | |
474 | ||
b58f81ae DJ |
475 | @kindex --exclude-libs |
476 | @item --exclude-libs @var{lib},@var{lib},... | |
477 | Specifies a list of archive libraries from which symbols should not be automatically | |
e1c37eb5 | 478 | exported. The library names may be delimited by commas or colons. Specifying |
b58f81ae DJ |
479 | @code{--exclude-libs ALL} excludes symbols in all archive libraries from |
480 | automatic export. This option is available only for the i386 PE targeted | |
481 | port of the linker and for ELF targeted ports. For i386 PE, symbols | |
482 | explicitly listed in a .def file are still exported, regardless of this | |
483 | option. For ELF targeted ports, symbols affected by this option will | |
484 | be treated as hidden. | |
485 | ||
e1c37eb5 DK |
486 | @kindex --exclude-modules-for-implib |
487 | @item --exclude-modules-for-implib @var{module},@var{module},... | |
488 | Specifies a list of object files or archive members, from which symbols | |
489 | should not be automatically exported, but which should be copied wholesale | |
490 | into the import library being generated during the link. The module names | |
491 | may be delimited by commas or colons, and must match exactly the filenames | |
492 | used by @command{ld} to open the files; for archive members, this is simply | |
493 | the member name, but for object files the name listed must include and | |
494 | match precisely any path used to specify the input file on the linker's | |
495 | command-line. This option is available only for the i386 PE targeted port | |
496 | of the linker. Symbols explicitly listed in a .def file are still exported, | |
497 | regardless of this option. | |
498 | ||
252b5132 RH |
499 | @cindex dynamic symbol table |
500 | @kindex -E | |
501 | @kindex --export-dynamic | |
267e2722 | 502 | @kindex --no-export-dynamic |
252b5132 RH |
503 | @item -E |
504 | @itemx --export-dynamic | |
267e2722 CD |
505 | @itemx --no-export-dynamic |
506 | When creating a dynamically linked executable, using the @option{-E} | |
507 | option or the @option{--export-dynamic} option causes the linker to add | |
508 | all symbols to the dynamic symbol table. The dynamic symbol table is the | |
509 | set of symbols which are visible from dynamic objects at run time. | |
510 | ||
511 | If you do not use either of these options (or use the | |
512 | @option{--no-export-dynamic} option to restore the default behavior), the | |
513 | dynamic symbol table will normally contain only those symbols which are | |
514 | referenced by some dynamic object mentioned in the link. | |
252b5132 RH |
515 | |
516 | If you use @code{dlopen} to load a dynamic object which needs to refer | |
517 | back to the symbols defined by the program, rather than some other | |
518 | dynamic object, then you will probably need to use this option when | |
519 | linking the program itself. | |
520 | ||
55255dae | 521 | You can also use the dynamic list to control what symbols should |
cb840a31 | 522 | be added to the dynamic symbol table if the output format supports it. |
55255dae | 523 | See the description of @samp{--dynamic-list}. |
cb840a31 | 524 | |
8b747e1a DK |
525 | Note that this option is specific to ELF targeted ports. PE targets |
526 | support a similar function to export all symbols from a DLL or EXE; see | |
527 | the description of @samp{--export-all-symbols} below. | |
528 | ||
36f63dca | 529 | @ifclear SingleFormat |
252b5132 RH |
530 | @cindex big-endian objects |
531 | @cindex endianness | |
532 | @kindex -EB | |
533 | @item -EB | |
534 | Link big-endian objects. This affects the default output format. | |
535 | ||
536 | @cindex little-endian objects | |
537 | @kindex -EL | |
538 | @item -EL | |
539 | Link little-endian objects. This affects the default output format. | |
36f63dca | 540 | @end ifclear |
252b5132 | 541 | |
2509a395 SL |
542 | @kindex -f @var{name} |
543 | @kindex --auxiliary=@var{name} | |
544 | @item -f @var{name} | |
545 | @itemx --auxiliary=@var{name} | |
252b5132 RH |
546 | When creating an ELF shared object, set the internal DT_AUXILIARY field |
547 | to the specified name. This tells the dynamic linker that the symbol | |
548 | table of the shared object should be used as an auxiliary filter on the | |
549 | symbol table of the shared object @var{name}. | |
550 | ||
551 | If you later link a program against this filter object, then, when you | |
552 | run the program, the dynamic linker will see the DT_AUXILIARY field. If | |
553 | the dynamic linker resolves any symbols from the filter object, it will | |
554 | first check whether there is a definition in the shared object | |
555 | @var{name}. If there is one, it will be used instead of the definition | |
556 | in the filter object. The shared object @var{name} need not exist. | |
557 | Thus the shared object @var{name} may be used to provide an alternative | |
558 | implementation of certain functions, perhaps for debugging or for | |
559 | machine specific performance. | |
560 | ||
561 | This option may be specified more than once. The DT_AUXILIARY entries | |
562 | will be created in the order in which they appear on the command line. | |
563 | ||
2509a395 SL |
564 | @kindex -F @var{name} |
565 | @kindex --filter=@var{name} | |
252b5132 | 566 | @item -F @var{name} |
2509a395 | 567 | @itemx --filter=@var{name} |
252b5132 RH |
568 | When creating an ELF shared object, set the internal DT_FILTER field to |
569 | the specified name. This tells the dynamic linker that the symbol table | |
570 | of the shared object which is being created should be used as a filter | |
571 | on the symbol table of the shared object @var{name}. | |
572 | ||
573 | If you later link a program against this filter object, then, when you | |
574 | run the program, the dynamic linker will see the DT_FILTER field. The | |
575 | dynamic linker will resolve symbols according to the symbol table of the | |
576 | filter object as usual, but it will actually link to the definitions | |
577 | found in the shared object @var{name}. Thus the filter object can be | |
578 | used to select a subset of the symbols provided by the object | |
579 | @var{name}. | |
580 | ||
ff5dcc92 | 581 | Some older linkers used the @option{-F} option throughout a compilation |
252b5132 | 582 | toolchain for specifying object-file format for both input and output |
36f63dca NC |
583 | object files. |
584 | @ifclear SingleFormat | |
585 | The @sc{gnu} linker uses other mechanisms for this purpose: the | |
ece2d90e | 586 | @option{-b}, @option{--format}, @option{--oformat} options, the |
252b5132 | 587 | @code{TARGET} command in linker scripts, and the @code{GNUTARGET} |
36f63dca NC |
588 | environment variable. |
589 | @end ifclear | |
590 | The @sc{gnu} linker will ignore the @option{-F} option when not | |
591 | creating an ELF shared object. | |
252b5132 | 592 | |
3dbf70a2 | 593 | @cindex finalization function |
2509a395 SL |
594 | @kindex -fini=@var{name} |
595 | @item -fini=@var{name} | |
3dbf70a2 MM |
596 | When creating an ELF executable or shared object, call NAME when the |
597 | executable or shared object is unloaded, by setting DT_FINI to the | |
598 | address of the function. By default, the linker uses @code{_fini} as | |
599 | the function to call. | |
600 | ||
252b5132 RH |
601 | @kindex -g |
602 | @item -g | |
603 | Ignored. Provided for compatibility with other tools. | |
604 | ||
2509a395 SL |
605 | @kindex -G @var{value} |
606 | @kindex --gpsize=@var{value} | |
252b5132 | 607 | @cindex object size |
2509a395 | 608 | @item -G @var{value} |
252b5132 RH |
609 | @itemx --gpsize=@var{value} |
610 | Set the maximum size of objects to be optimized using the GP register to | |
611 | @var{size}. This is only meaningful for object file formats such as | |
e8044f35 | 612 | MIPS ELF that support putting large and small objects into different |
252b5132 RH |
613 | sections. This is ignored for other object file formats. |
614 | ||
615 | @cindex runtime library name | |
2509a395 | 616 | @kindex -h @var{name} |
252b5132 | 617 | @kindex -soname=@var{name} |
2509a395 | 618 | @item -h @var{name} |
252b5132 RH |
619 | @itemx -soname=@var{name} |
620 | When creating an ELF shared object, set the internal DT_SONAME field to | |
621 | the specified name. When an executable is linked with a shared object | |
622 | which has a DT_SONAME field, then when the executable is run the dynamic | |
623 | linker will attempt to load the shared object specified by the DT_SONAME | |
624 | field rather than the using the file name given to the linker. | |
625 | ||
626 | @kindex -i | |
627 | @cindex incremental link | |
628 | @item -i | |
629 | Perform an incremental link (same as option @samp{-r}). | |
630 | ||
3dbf70a2 | 631 | @cindex initialization function |
2509a395 SL |
632 | @kindex -init=@var{name} |
633 | @item -init=@var{name} | |
3dbf70a2 MM |
634 | When creating an ELF executable or shared object, call NAME when the |
635 | executable or shared object is loaded, by setting DT_INIT to the address | |
636 | of the function. By default, the linker uses @code{_init} as the | |
637 | function to call. | |
638 | ||
252b5132 | 639 | @cindex archive files, from cmd line |
2509a395 | 640 | @kindex -l @var{namespec} |
bcb674cf | 641 | @kindex --library=@var{namespec} |
2509a395 | 642 | @item -l @var{namespec} |
bcb674cf RS |
643 | @itemx --library=@var{namespec} |
644 | Add the archive or object file specified by @var{namespec} to the | |
645 | list of files to link. This option may be used any number of times. | |
646 | If @var{namespec} is of the form @file{:@var{filename}}, @command{ld} | |
07d8eb55 | 647 | will search the library path for a file called @var{filename}, otherwise it |
bcb674cf | 648 | will search the library path for a file called @file{lib@var{namespec}.a}. |
252b5132 | 649 | |
ff5dcc92 | 650 | On systems which support shared libraries, @command{ld} may also search for |
bcb674cf RS |
651 | files other than @file{lib@var{namespec}.a}. Specifically, on ELF |
652 | and SunOS systems, @command{ld} will search a directory for a library | |
653 | called @file{lib@var{namespec}.so} before searching for one called | |
654 | @file{lib@var{namespec}.a}. (By convention, a @code{.so} extension | |
655 | indicates a shared library.) Note that this behavior does not apply | |
656 | to @file{:@var{filename}}, which always specifies a file called | |
657 | @var{filename}. | |
252b5132 RH |
658 | |
659 | The linker will search an archive only once, at the location where it is | |
660 | specified on the command line. If the archive defines a symbol which | |
661 | was undefined in some object which appeared before the archive on the | |
662 | command line, the linker will include the appropriate file(s) from the | |
663 | archive. However, an undefined symbol in an object appearing later on | |
664 | the command line will not cause the linker to search the archive again. | |
665 | ||
ff5dcc92 | 666 | See the @option{-(} option for a way to force the linker to search |
252b5132 RH |
667 | archives multiple times. |
668 | ||
669 | You may list the same archive multiple times on the command line. | |
670 | ||
671 | @ifset GENERIC | |
672 | This type of archive searching is standard for Unix linkers. However, | |
ff5dcc92 | 673 | if you are using @command{ld} on AIX, note that it is different from the |
252b5132 RH |
674 | behaviour of the AIX linker. |
675 | @end ifset | |
676 | ||
677 | @cindex search directory, from cmd line | |
2509a395 | 678 | @kindex -L @var{dir} |
252b5132 | 679 | @kindex --library-path=@var{dir} |
2509a395 | 680 | @item -L @var{searchdir} |
252b5132 | 681 | @itemx --library-path=@var{searchdir} |
ff5dcc92 SC |
682 | Add path @var{searchdir} to the list of paths that @command{ld} will search |
683 | for archive libraries and @command{ld} control scripts. You may use this | |
252b5132 RH |
684 | option any number of times. The directories are searched in the order |
685 | in which they are specified on the command line. Directories specified | |
686 | on the command line are searched before the default directories. All | |
ff5dcc92 | 687 | @option{-L} options apply to all @option{-l} options, regardless of the |
7d24f02c KH |
688 | order in which the options appear. @option{-L} options do not affect |
689 | how @command{ld} searches for a linker script unless @option{-T} | |
690 | option is specified. | |
252b5132 | 691 | |
3aa2d05a NC |
692 | If @var{searchdir} begins with @code{=} or @code{$SYSROOT}, then this |
693 | prefix will be replaced by the @dfn{sysroot prefix}, controlled by the | |
694 | @samp{--sysroot} option, or specified when the linker is configured. | |
9c8ebd6a | 695 | |
252b5132 RH |
696 | @ifset UsesEnvVars |
697 | The default set of paths searched (without being specified with | |
ff5dcc92 | 698 | @samp{-L}) depends on which emulation mode @command{ld} is using, and in |
252b5132 RH |
699 | some cases also on how it was configured. @xref{Environment}. |
700 | @end ifset | |
701 | ||
702 | The paths can also be specified in a link script with the | |
703 | @code{SEARCH_DIR} command. Directories specified this way are searched | |
704 | at the point in which the linker script appears in the command line. | |
705 | ||
706 | @cindex emulation | |
707 | @kindex -m @var{emulation} | |
2509a395 | 708 | @item -m @var{emulation} |
252b5132 RH |
709 | Emulate the @var{emulation} linker. You can list the available |
710 | emulations with the @samp{--verbose} or @samp{-V} options. | |
711 | ||
712 | If the @samp{-m} option is not used, the emulation is taken from the | |
713 | @code{LDEMULATION} environment variable, if that is defined. | |
714 | ||
715 | Otherwise, the default emulation depends upon how the linker was | |
716 | configured. | |
717 | ||
718 | @cindex link map | |
719 | @kindex -M | |
720 | @kindex --print-map | |
721 | @item -M | |
722 | @itemx --print-map | |
723 | Print a link map to the standard output. A link map provides | |
724 | information about the link, including the following: | |
725 | ||
726 | @itemize @bullet | |
727 | @item | |
3b83e13a | 728 | Where object files are mapped into memory. |
252b5132 RH |
729 | @item |
730 | How common symbols are allocated. | |
731 | @item | |
732 | All archive members included in the link, with a mention of the symbol | |
733 | which caused the archive member to be brought in. | |
3b83e13a NC |
734 | @item |
735 | The values assigned to symbols. | |
736 | ||
737 | Note - symbols whose values are computed by an expression which | |
738 | involves a reference to a previous value of the same symbol may not | |
739 | have correct result displayed in the link map. This is because the | |
740 | linker discards intermediate results and only retains the final value | |
741 | of an expression. Under such circumstances the linker will display | |
742 | the final value enclosed by square brackets. Thus for example a | |
743 | linker script containing: | |
744 | ||
745 | @smallexample | |
746 | foo = 1 | |
747 | foo = foo * 4 | |
748 | foo = foo + 8 | |
749 | @end smallexample | |
750 | ||
751 | will produce the following output in the link map if the @option{-M} | |
752 | option is used: | |
753 | ||
754 | @smallexample | |
755 | 0x00000001 foo = 0x1 | |
756 | [0x0000000c] foo = (foo * 0x4) | |
757 | [0x0000000c] foo = (foo + 0x8) | |
758 | @end smallexample | |
759 | ||
760 | See @ref{Expressions} for more information about expressions in linker | |
761 | scripts. | |
252b5132 RH |
762 | @end itemize |
763 | ||
764 | @kindex -n | |
765 | @cindex read-only text | |
766 | @cindex NMAGIC | |
767 | @kindex --nmagic | |
768 | @item -n | |
769 | @itemx --nmagic | |
90f5d9d9 JZ |
770 | Turn off page alignment of sections, and disable linking against shared |
771 | libraries. If the output format supports Unix style magic numbers, | |
772 | mark the output as @code{NMAGIC}. | |
252b5132 RH |
773 | |
774 | @kindex -N | |
775 | @kindex --omagic | |
776 | @cindex read/write from cmd line | |
777 | @cindex OMAGIC | |
a1ab1d2a | 778 | @item -N |
252b5132 RH |
779 | @itemx --omagic |
780 | Set the text and data sections to be readable and writable. Also, do | |
63fd3b82 NC |
781 | not page-align the data segment, and disable linking against shared |
782 | libraries. If the output format supports Unix style magic numbers, | |
4d8907ac DS |
783 | mark the output as @code{OMAGIC}. Note: Although a writable text section |
784 | is allowed for PE-COFF targets, it does not conform to the format | |
785 | specification published by Microsoft. | |
63fd3b82 NC |
786 | |
787 | @kindex --no-omagic | |
788 | @cindex OMAGIC | |
789 | @item --no-omagic | |
790 | This option negates most of the effects of the @option{-N} option. It | |
791 | sets the text section to be read-only, and forces the data segment to | |
792 | be page-aligned. Note - this option does not enable linking against | |
793 | shared libraries. Use @option{-Bdynamic} for this. | |
252b5132 RH |
794 | |
795 | @kindex -o @var{output} | |
796 | @kindex --output=@var{output} | |
797 | @cindex naming the output file | |
798 | @item -o @var{output} | |
799 | @itemx --output=@var{output} | |
ff5dcc92 | 800 | Use @var{output} as the name for the program produced by @command{ld}; if this |
252b5132 RH |
801 | option is not specified, the name @file{a.out} is used by default. The |
802 | script command @code{OUTPUT} can also specify the output file name. | |
803 | ||
804 | @kindex -O @var{level} | |
805 | @cindex generating optimized output | |
806 | @item -O @var{level} | |
ff5dcc92 | 807 | If @var{level} is a numeric values greater than zero @command{ld} optimizes |
252b5132 | 808 | the output. This might take significantly longer and therefore probably |
98c503ac NC |
809 | should only be enabled for the final binary. At the moment this |
810 | option only affects ELF shared library generation. Future releases of | |
811 | the linker may make more use of this option. Also currently there is | |
812 | no difference in the linker's behaviour for different non-zero values | |
813 | of this option. Again this may change with future releases. | |
252b5132 | 814 | |
387dd777 DP |
815 | @kindex -plugin @var{name} |
816 | @item -plugin @var{name} | |
817 | Involve a plugin in the linking process. The @var{name} parameter is | |
818 | the absolute filename of the plugin. Usually this parameter is | |
819 | automatically added by the complier, when using link time | |
820 | optimization, but users can also add their own plugins if they so | |
821 | wish. | |
822 | ||
823 | Note that the location of the compiler originated plugins is different | |
824 | from the place where the @command{ar}, @command{nm} and | |
825 | @command{ranlib} programs search for their plugins. In order for | |
826 | those commands to make use of a compiler based plugin it must first be | |
827 | copied into the @file{$@{libdir@}/bfd-plugins} directory. All gcc | |
828 | based linker plugins are backward compatible, so it is sufficient to | |
6c19b93b | 829 | just copy in the newest one. |
387dd777 | 830 | |
26278bb8 UD |
831 | @kindex --push-state |
832 | @cindex push state governing input file handling | |
833 | @item --push-state | |
834 | The @option{--push-state} allows to preserve the current state of the | |
835 | flags which govern the input file handling so that they can all be | |
836 | restored with one corresponding @option{--pop-state} option. | |
837 | ||
838 | The option which are covered are: @option{-Bdynamic}, @option{-Bstatic}, | |
839 | @option{-dn}, @option{-dy}, @option{-call_shared}, @option{-non_shared}, | |
840 | @option{-static}, @option{-N}, @option{-n}, @option{--whole-archive}, | |
841 | @option{--no-whole-archive}, @option{-r}, @option{-Ur}, | |
842 | @option{--copy-dt-needed-entries}, @option{--no-copy-dt-needed-entries}, | |
843 | @option{--as-needed}, @option{--no-as-needed}, and @option{-a}. | |
844 | ||
845 | One target for this option are specifications for @file{pkg-config}. When | |
846 | used with the @option{--libs} option all possibly needed libraries are | |
847 | listed and then possibly linked with all the time. It is better to return | |
848 | something as follows: | |
849 | ||
850 | @smallexample | |
851 | -Wl,--push-state,--as-needed -libone -libtwo -Wl,--pop-state | |
852 | @end smallexample | |
853 | ||
854 | @kindex --pop-state | |
855 | @cindex pop state governing input file handling | |
67cecaba | 856 | @item --pop-state |
26278bb8 UD |
857 | Undoes the effect of --push-state, restores the previous values of the |
858 | flags governing input file handling. | |
859 | ||
a712da20 NC |
860 | @kindex -q |
861 | @kindex --emit-relocs | |
862 | @cindex retain relocations in final executable | |
863 | @item -q | |
864 | @itemx --emit-relocs | |
ba1be17e | 865 | Leave relocation sections and contents in fully linked executables. |
a712da20 NC |
866 | Post link analysis and optimization tools may need this information in |
867 | order to perform correct modifications of executables. This results | |
868 | in larger executables. | |
869 | ||
dbab7a7b NC |
870 | This option is currently only supported on ELF platforms. |
871 | ||
4f471f39 RS |
872 | @kindex --force-dynamic |
873 | @cindex forcing the creation of dynamic sections | |
874 | @item --force-dynamic | |
875 | Force the output file to have dynamic sections. This option is specific | |
876 | to VxWorks targets. | |
877 | ||
252b5132 RH |
878 | @cindex partial link |
879 | @cindex relocatable output | |
880 | @kindex -r | |
1049f94e | 881 | @kindex --relocatable |
252b5132 | 882 | @item -r |
1049f94e | 883 | @itemx --relocatable |
252b5132 | 884 | Generate relocatable output---i.e., generate an output file that can in |
ff5dcc92 | 885 | turn serve as input to @command{ld}. This is often called @dfn{partial |
252b5132 RH |
886 | linking}. As a side effect, in environments that support standard Unix |
887 | magic numbers, this option also sets the output file's magic number to | |
888 | @code{OMAGIC}. | |
ff5dcc92 | 889 | @c ; see @option{-N}. |
252b5132 RH |
890 | If this option is not specified, an absolute file is produced. When |
891 | linking C++ programs, this option @emph{will not} resolve references to | |
892 | constructors; to do that, use @samp{-Ur}. | |
893 | ||
62bf86b4 HPN |
894 | When an input file does not have the same format as the output file, |
895 | partial linking is only supported if that input file does not contain any | |
896 | relocations. Different output formats can have further restrictions; for | |
897 | example some @code{a.out}-based formats do not support partial linking | |
898 | with input files in other formats at all. | |
899 | ||
252b5132 RH |
900 | This option does the same thing as @samp{-i}. |
901 | ||
902 | @kindex -R @var{file} | |
903 | @kindex --just-symbols=@var{file} | |
904 | @cindex symbol-only input | |
905 | @item -R @var{filename} | |
906 | @itemx --just-symbols=@var{filename} | |
907 | Read symbol names and their addresses from @var{filename}, but do not | |
908 | relocate it or include it in the output. This allows your output file | |
909 | to refer symbolically to absolute locations of memory defined in other | |
910 | programs. You may use this option more than once. | |
911 | ||
ff5dcc92 | 912 | For compatibility with other ELF linkers, if the @option{-R} option is |
252b5132 | 913 | followed by a directory name, rather than a file name, it is treated as |
ff5dcc92 | 914 | the @option{-rpath} option. |
252b5132 RH |
915 | |
916 | @kindex -s | |
917 | @kindex --strip-all | |
918 | @cindex strip all symbols | |
a1ab1d2a | 919 | @item -s |
252b5132 RH |
920 | @itemx --strip-all |
921 | Omit all symbol information from the output file. | |
922 | ||
923 | @kindex -S | |
924 | @kindex --strip-debug | |
925 | @cindex strip debugger symbols | |
a1ab1d2a | 926 | @item -S |
252b5132 RH |
927 | @itemx --strip-debug |
928 | Omit debugger symbol information (but not all symbols) from the output file. | |
929 | ||
a70f34c0 NC |
930 | @kindex --strip-discarded |
931 | @kindex --no-strip-discarded | |
932 | @item --strip-discarded | |
933 | @itemx --no-strip-discarded | |
934 | Omit (or do not omit) global symbols defined in discarded sections. | |
935 | Enabled by default. | |
936 | ||
252b5132 RH |
937 | @kindex -t |
938 | @kindex --trace | |
939 | @cindex input files, displaying | |
a1ab1d2a | 940 | @item -t |
252b5132 | 941 | @itemx --trace |
ff5dcc92 | 942 | Print the names of the input files as @command{ld} processes them. |
252b5132 RH |
943 | |
944 | @kindex -T @var{script} | |
945 | @kindex --script=@var{script} | |
946 | @cindex script files | |
947 | @item -T @var{scriptfile} | |
948 | @itemx --script=@var{scriptfile} | |
949 | Use @var{scriptfile} as the linker script. This script replaces | |
ff5dcc92 | 950 | @command{ld}'s default linker script (rather than adding to it), so |
252b5132 | 951 | @var{commandfile} must specify everything necessary to describe the |
114283d8 NC |
952 | output file. @xref{Scripts}. If @var{scriptfile} does not exist in |
953 | the current directory, @code{ld} looks for it in the directories | |
954 | specified by any preceding @samp{-L} options. Multiple @samp{-T} | |
955 | options accumulate. | |
252b5132 | 956 | |
14be8564 L |
957 | @kindex -dT @var{script} |
958 | @kindex --default-script=@var{script} | |
959 | @cindex script files | |
960 | @item -dT @var{scriptfile} | |
961 | @itemx --default-script=@var{scriptfile} | |
962 | Use @var{scriptfile} as the default linker script. @xref{Scripts}. | |
963 | ||
964 | This option is similar to the @option{--script} option except that | |
965 | processing of the script is delayed until after the rest of the | |
966 | command line has been processed. This allows options placed after the | |
967 | @option{--default-script} option on the command line to affect the | |
968 | behaviour of the linker script, which can be important when the linker | |
969 | command line cannot be directly controlled by the user. (eg because | |
970 | the command line is being constructed by another tool, such as | |
971 | @samp{gcc}). | |
972 | ||
252b5132 RH |
973 | @kindex -u @var{symbol} |
974 | @kindex --undefined=@var{symbol} | |
975 | @cindex undefined symbol | |
976 | @item -u @var{symbol} | |
977 | @itemx --undefined=@var{symbol} | |
978 | Force @var{symbol} to be entered in the output file as an undefined | |
979 | symbol. Doing this may, for example, trigger linking of additional | |
980 | modules from standard libraries. @samp{-u} may be repeated with | |
981 | different option arguments to enter additional undefined symbols. This | |
982 | option is equivalent to the @code{EXTERN} linker script command. | |
983 | ||
0a618243 AB |
984 | If this option is being used to force additional modules to be pulled |
985 | into the link, and if it is an error for the symbol to remain | |
986 | undefined, then the option @option{--require-defined} should be used | |
987 | instead. | |
988 | ||
989 | @kindex --require-defined=@var{symbol} | |
990 | @cindex symbols, require defined | |
991 | @cindex defined symbol | |
992 | @item --require-defined=@var{symbol} | |
993 | Require that @var{symbol} is defined in the output file. This option | |
994 | is the same as option @option{--undefined} except that if @var{symbol} | |
995 | is not defined in the output file then the linker will issue an error | |
996 | and exit. The same effect can be achieved in a linker script by using | |
997 | @code{EXTERN}, @code{ASSERT} and @code{DEFINED} together. This option | |
998 | can be used multiple times to require additional symbols. | |
999 | ||
252b5132 RH |
1000 | @kindex -Ur |
1001 | @cindex constructors | |
a1ab1d2a | 1002 | @item -Ur |
252b5132 RH |
1003 | For anything other than C++ programs, this option is equivalent to |
1004 | @samp{-r}: it generates relocatable output---i.e., an output file that can in | |
ff5dcc92 | 1005 | turn serve as input to @command{ld}. When linking C++ programs, @samp{-Ur} |
252b5132 RH |
1006 | @emph{does} resolve references to constructors, unlike @samp{-r}. |
1007 | It does not work to use @samp{-Ur} on files that were themselves linked | |
1008 | with @samp{-Ur}; once the constructor table has been built, it cannot | |
1009 | be added to. Use @samp{-Ur} only for the last partial link, and | |
1010 | @samp{-r} for the others. | |
1011 | ||
c005eb9e AB |
1012 | @kindex --orphan-handling=@var{MODE} |
1013 | @cindex orphan sections | |
1014 | @cindex sections, orphan | |
1015 | @item --orphan-handling=@var{MODE} | |
1016 | Control how orphan sections are handled. An orphan section is one not | |
1017 | specifically mentioned in a linker script. @xref{Orphan Sections}. | |
1018 | ||
1019 | @var{MODE} can have any of the following values: | |
1020 | ||
1021 | @table @code | |
1022 | @item place | |
1023 | Orphan sections are placed into a suitable output section following | |
1024 | the strategy described in @ref{Orphan Sections}. The option | |
d2732b69 | 1025 | @samp{--unique} also affects how sections are placed. |
c005eb9e AB |
1026 | |
1027 | @item discard | |
1028 | All orphan sections are discarded, by placing them in the | |
1029 | @samp{/DISCARD/} section (@pxref{Output Section Discarding}). | |
1030 | ||
1031 | @item warn | |
1032 | The linker will place the orphan section as for @code{place} and also | |
1033 | issue a warning. | |
1034 | ||
1035 | @item error | |
1036 | The linker will exit with an error if any orphan section is found. | |
1037 | @end table | |
1038 | ||
1039 | The default if @samp{--orphan-handling} is not given is @code{place}. | |
1040 | ||
577a0623 AM |
1041 | @kindex --unique[=@var{SECTION}] |
1042 | @item --unique[=@var{SECTION}] | |
1043 | Creates a separate output section for every input section matching | |
1044 | @var{SECTION}, or if the optional wildcard @var{SECTION} argument is | |
1045 | missing, for every orphan input section. An orphan section is one not | |
1046 | specifically mentioned in a linker script. You may use this option | |
1047 | multiple times on the command line; It prevents the normal merging of | |
1048 | input sections with the same name, overriding output section assignments | |
1049 | in a linker script. | |
a854a4a7 | 1050 | |
252b5132 RH |
1051 | @kindex -v |
1052 | @kindex -V | |
1053 | @kindex --version | |
1054 | @cindex version | |
1055 | @item -v | |
1056 | @itemx --version | |
1057 | @itemx -V | |
ff5dcc92 | 1058 | Display the version number for @command{ld}. The @option{-V} option also |
252b5132 RH |
1059 | lists the supported emulations. |
1060 | ||
1061 | @kindex -x | |
1062 | @kindex --discard-all | |
1063 | @cindex deleting local symbols | |
1064 | @item -x | |
1065 | @itemx --discard-all | |
1066 | Delete all local symbols. | |
1067 | ||
1068 | @kindex -X | |
1069 | @kindex --discard-locals | |
1070 | @cindex local symbols, deleting | |
a1ab1d2a | 1071 | @item -X |
252b5132 | 1072 | @itemx --discard-locals |
3c68c38f BW |
1073 | Delete all temporary local symbols. (These symbols start with |
1074 | system-specific local label prefixes, typically @samp{.L} for ELF systems | |
1075 | or @samp{L} for traditional a.out systems.) | |
252b5132 RH |
1076 | |
1077 | @kindex -y @var{symbol} | |
1078 | @kindex --trace-symbol=@var{symbol} | |
1079 | @cindex symbol tracing | |
1080 | @item -y @var{symbol} | |
1081 | @itemx --trace-symbol=@var{symbol} | |
1082 | Print the name of each linked file in which @var{symbol} appears. This | |
1083 | option may be given any number of times. On many systems it is necessary | |
1084 | to prepend an underscore. | |
1085 | ||
1086 | This option is useful when you have an undefined symbol in your link but | |
1087 | don't know where the reference is coming from. | |
1088 | ||
1089 | @kindex -Y @var{path} | |
1090 | @item -Y @var{path} | |
1091 | Add @var{path} to the default library search path. This option exists | |
1092 | for Solaris compatibility. | |
1093 | ||
1094 | @kindex -z @var{keyword} | |
1095 | @item -z @var{keyword} | |
cd6d6c15 NC |
1096 | The recognized keywords are: |
1097 | @table @samp | |
1098 | ||
e6e2dfbd AM |
1099 | @item bndplt |
1100 | Always generate BND prefix in PLT entries. Supported for Linux/x86_64. | |
1101 | ||
1102 | @item call-nop=prefix-addr | |
1103 | @itemx call-nop=suffix-nop | |
1104 | @itemx call-nop=prefix-@var{byte} | |
1105 | @itemx call-nop=suffix-@var{byte} | |
1106 | Specify the 1-byte @code{NOP} padding when transforming indirect call | |
1107 | to a locally defined function, foo, via its GOT slot. | |
1108 | @option{call-nop=prefix-addr} generates @code{0x67 call foo}. | |
1109 | @option{call-nop=suffix-nop} generates @code{call foo 0x90}. | |
1110 | @option{call-nop=prefix-@var{byte}} generates @code{@var{byte} call foo}. | |
1111 | @option{call-nop=suffix-@var{byte}} generates @code{call foo @var{byte}}. | |
1112 | Supported for i386 and x86_64. | |
1113 | ||
cd6d6c15 | 1114 | @item combreloc |
e6e2dfbd AM |
1115 | @itemx nocombreloc |
1116 | Combine multiple dynamic relocation sections and sort to improve | |
1117 | dynamic symbol lookup caching. Do not do this if @samp{nocombreloc}. | |
cd6d6c15 | 1118 | |
b8871f35 | 1119 | @item common |
e6e2dfbd AM |
1120 | @itemx nocommon |
1121 | Generate common symbols with STT_COMMON type during a relocatable | |
1122 | link. Use STT_OBJECT type if @samp{nocommon}. | |
1123 | ||
1124 | @item common-page-size=@var{value} | |
1125 | Set the page size most commonly used to @var{value}. Memory image | |
1126 | layout will be optimized to minimize memory pages if the system is | |
1127 | using pages of this size. | |
b8871f35 | 1128 | |
cd6d6c15 | 1129 | @item defs |
97a232d7 NC |
1130 | Report unresolved symbol references from regular object files. This |
1131 | is done even if the linker is creating a non-symbolic shared library. | |
1132 | This option is the inverse of @samp{-z undefs}. | |
cd6d6c15 | 1133 | |
e6e2dfbd AM |
1134 | @item dynamic-undefined-weak |
1135 | @itemx nodynamic-undefined-weak | |
1136 | Make undefined weak symbols dynamic when building a dynamic object, | |
1137 | if they are referenced from a regular object file and not forced local | |
1138 | by symbol visibility or versioning. Do not make them dynamic if | |
1139 | @samp{nodynamic-undefined-weak}. If neither option is given, a target | |
1140 | may default to either option being in force, or make some other | |
1141 | selection of undefined weak symbols dynamic. Not all targets support | |
1142 | these options. | |
1143 | ||
6aa29e7b JJ |
1144 | @item execstack |
1145 | Marks the object as requiring executable stack. | |
1146 | ||
b039ef04 L |
1147 | @item global |
1148 | This option is only meaningful when building a shared object. It makes | |
1149 | the symbols defined by this shared object available for symbol resolution | |
1150 | of subsequently loaded libraries. | |
1151 | ||
93ab9c0d NC |
1152 | @item globalaudit |
1153 | This option is only meaningful when building a dynamic executable. | |
1154 | This option marks the executable as requiring global auditing by | |
1155 | setting the @code{DF_1_GLOBAUDIT} bit in the @code{DT_FLAGS_1} dynamic | |
1156 | tag. Global auditing requires that any auditing library defined via | |
a05a5b64 | 1157 | the @option{--depaudit} or @option{-P} command-line options be run for |
6c19b93b | 1158 | all dynamic objects loaded by the application. |
93ab9c0d | 1159 | |
e6e2dfbd AM |
1160 | @item ibtplt |
1161 | Generate Intel Indirect Branch Tracking (IBT) enabled PLT entries. | |
1162 | Supported for Linux/i386 and Linux/x86_64. | |
1163 | ||
8df93cb5 | 1164 | @item ibt |
e6e2dfbd AM |
1165 | Generate GNU_PROPERTY_X86_FEATURE_1_IBT in .note.gnu.property section |
1166 | to indicate compatibility with IBT. This also implies @option{ibtplt}. | |
1167 | Supported for Linux/i386 and Linux/x86_64. | |
1168 | ||
cd6d6c15 NC |
1169 | @item initfirst |
1170 | This option is only meaningful when building a shared object. | |
1171 | It marks the object so that its runtime initialization will occur | |
1172 | before the runtime initialization of any other objects brought into | |
1173 | the process at the same time. Similarly the runtime finalization of | |
1174 | the object will occur after the runtime finalization of any other | |
1175 | objects. | |
1176 | ||
1177 | @item interpose | |
e6e2dfbd AM |
1178 | Specify that the dynamic loader should modify its symbol search order |
1179 | so that symbols in this shared library interpose all other shared | |
1180 | libraries not so marked. | |
cd6d6c15 | 1181 | |
5fa222e4 AM |
1182 | @item lazy |
1183 | When generating an executable or shared library, mark it to tell the | |
1184 | dynamic linker to defer function call resolution to the point when | |
1185 | the function is called (lazy binding), rather than at load time. | |
1186 | Lazy binding is the default. | |
1187 | ||
cd6d6c15 | 1188 | @item loadfltr |
e6e2dfbd | 1189 | Specify that the object's filters be processed immediately at runtime. |
cd6d6c15 | 1190 | |
e6e2dfbd AM |
1191 | @item max-page-size=@var{value} |
1192 | Set the maximum memory page size supported to @var{value}. | |
cd6d6c15 | 1193 | |
e6e2dfbd AM |
1194 | @item muldefs |
1195 | Allow multiple definitions. | |
b8871f35 | 1196 | |
cd6d6c15 | 1197 | @item nocopyreloc |
daf220f0 AM |
1198 | Disable linker generated .dynbss variables used in place of variables |
1199 | defined in shared libraries. May result in dynamic text relocations. | |
cd6d6c15 NC |
1200 | |
1201 | @item nodefaultlib | |
e6e2dfbd AM |
1202 | Specify that the dynamic loader search for dependencies of this object |
1203 | should ignore any default library search paths. | |
cd6d6c15 NC |
1204 | |
1205 | @item nodelete | |
e6e2dfbd | 1206 | Specify that the object shouldn't be unloaded at runtime. |
cd6d6c15 NC |
1207 | |
1208 | @item nodlopen | |
e6e2dfbd | 1209 | Specify that the object is not available to @code{dlopen}. |
cd6d6c15 NC |
1210 | |
1211 | @item nodump | |
e6e2dfbd | 1212 | Specify that the object can not be dumped by @code{dldump}. |
cd6d6c15 | 1213 | |
6aa29e7b JJ |
1214 | @item noexecstack |
1215 | Marks the object as not requiring executable stack. | |
1216 | ||
e6e2dfbd AM |
1217 | @item noextern-protected-data |
1218 | Don't treat protected data symbols as external when building a shared | |
1219 | library. This option overrides the linker backend default. It can be | |
1220 | used to work around incorrect relocations against protected data symbols | |
1221 | generated by compiler. Updates on protected data symbols by another | |
1222 | module aren't visible to the resulting shared library. Supported for | |
1223 | i386 and x86-64. | |
8dfef1bd | 1224 | |
e6e2dfbd AM |
1225 | @item noreloc-overflow |
1226 | Disable relocation overflow check. This can be used to disable | |
1227 | relocation overflow check if there will be no dynamic relocation | |
1228 | overflow at run-time. Supported for x86_64. | |
6aa29e7b | 1229 | |
cd6d6c15 NC |
1230 | @item now |
1231 | When generating an executable or shared library, mark it to tell the | |
1232 | dynamic linker to resolve all symbols when the program is started, or | |
e6e2dfbd AM |
1233 | when the shared library is loaded by dlopen, instead of deferring |
1234 | function call resolution to the point when the function is first | |
1235 | called. | |
cd6d6c15 NC |
1236 | |
1237 | @item origin | |
e6e2dfbd | 1238 | Specify that the object requires @samp{$ORIGIN} handling in paths. |
cd6d6c15 | 1239 | |
6aa29e7b | 1240 | @item relro |
e6e2dfbd | 1241 | @itemx norelro |
def5c83c AM |
1242 | Create an ELF @code{PT_GNU_RELRO} segment header in the object. This |
1243 | specifies a memory segment that should be made read-only after | |
1244 | relocation, if supported. Specifying @samp{common-page-size} smaller | |
1245 | than the system page size will render this protection ineffective. | |
e6e2dfbd | 1246 | Don't create an ELF @code{PT_GNU_RELRO} segment if @samp{norelro}. |
6aa29e7b | 1247 | |
47acac12 L |
1248 | @item separate-code |
1249 | @itemx noseparate-code | |
1250 | Create separate code @code{PT_LOAD} segment header in the object. This | |
1251 | specifies a memory segment that should contain only instructions and must | |
1252 | be in wholly disjoint pages from any other data. Don't create separate | |
1253 | code @code{PT_LOAD} segment if @samp{noseparate-code} is used. | |
1254 | ||
e6e2dfbd AM |
1255 | @item shstk |
1256 | Generate GNU_PROPERTY_X86_FEATURE_1_SHSTK in .note.gnu.property section | |
1257 | to indicate compatibility with Intel Shadow Stack. Supported for | |
1258 | Linux/i386 and Linux/x86_64. | |
24718e3b | 1259 | |
04c3a755 | 1260 | @item stack-size=@var{value} |
e6e2dfbd | 1261 | Specify a stack size for an ELF @code{PT_GNU_STACK} segment. |
04c3a755 NS |
1262 | Specifying zero will override any default non-zero sized |
1263 | @code{PT_GNU_STACK} segment creation. | |
1264 | ||
e6e2dfbd AM |
1265 | @item text |
1266 | @itemx notext | |
1267 | @itemx textoff | |
1268 | Report an error if DT_TEXTREL is set, i.e., if the binary has dynamic | |
1269 | relocations in read-only sections. Don't report an error if | |
1270 | @samp{notext} or @samp{textoff}. | |
48580982 | 1271 | |
97a232d7 NC |
1272 | @item undefs |
1273 | Do not report unresolved symbol references from regular object files, | |
1274 | either when creating an executable, or when creating a shared library. | |
1275 | This option is the inverse of @samp{-z defs}. | |
1276 | ||
cd6d6c15 NC |
1277 | @end table |
1278 | ||
ece2d90e | 1279 | Other keywords are ignored for Solaris compatibility. |
252b5132 RH |
1280 | |
1281 | @kindex -( | |
1282 | @cindex groups of archives | |
1283 | @item -( @var{archives} -) | |
1284 | @itemx --start-group @var{archives} --end-group | |
1285 | The @var{archives} should be a list of archive files. They may be | |
1286 | either explicit file names, or @samp{-l} options. | |
1287 | ||
1288 | The specified archives are searched repeatedly until no new undefined | |
1289 | references are created. Normally, an archive is searched only once in | |
1290 | the order that it is specified on the command line. If a symbol in that | |
1291 | archive is needed to resolve an undefined symbol referred to by an | |
1292 | object in an archive that appears later on the command line, the linker | |
1293 | would not be able to resolve that reference. By grouping the archives, | |
1294 | they all be searched repeatedly until all possible references are | |
1295 | resolved. | |
1296 | ||
1297 | Using this option has a significant performance cost. It is best to use | |
1298 | it only when there are unavoidable circular references between two or | |
1299 | more archives. | |
1300 | ||
69da35b5 NC |
1301 | @kindex --accept-unknown-input-arch |
1302 | @kindex --no-accept-unknown-input-arch | |
1303 | @item --accept-unknown-input-arch | |
1304 | @itemx --no-accept-unknown-input-arch | |
1305 | Tells the linker to accept input files whose architecture cannot be | |
2ca22b03 | 1306 | recognised. The assumption is that the user knows what they are doing |
69da35b5 NC |
1307 | and deliberately wants to link in these unknown input files. This was |
1308 | the default behaviour of the linker, before release 2.14. The default | |
1309 | behaviour from release 2.14 onwards is to reject such input files, and | |
1310 | so the @samp{--accept-unknown-input-arch} option has been added to | |
1311 | restore the old behaviour. | |
2ca22b03 | 1312 | |
4a43e768 AM |
1313 | @kindex --as-needed |
1314 | @kindex --no-as-needed | |
1315 | @item --as-needed | |
1316 | @itemx --no-as-needed | |
1317 | This option affects ELF DT_NEEDED tags for dynamic libraries mentioned | |
ddbb8a31 | 1318 | on the command line after the @option{--as-needed} option. Normally |
4a43e768 AM |
1319 | the linker will add a DT_NEEDED tag for each dynamic library mentioned |
1320 | on the command line, regardless of whether the library is actually | |
ddbb8a31 | 1321 | needed or not. @option{--as-needed} causes a DT_NEEDED tag to only be |
ffa9430d AM |
1322 | emitted for a library that @emph{at that point in the link} satisfies a |
1323 | non-weak undefined symbol reference from a regular object file or, if | |
1240be6b AM |
1324 | the library is not found in the DT_NEEDED lists of other needed libraries, a |
1325 | non-weak undefined symbol reference from another needed dynamic library. | |
ffa9430d AM |
1326 | Object files or libraries appearing on the command line @emph{after} |
1327 | the library in question do not affect whether the library is seen as | |
1328 | needed. This is similar to the rules for extraction of object files | |
1329 | from archives. @option{--no-as-needed} restores the default behaviour. | |
4a43e768 | 1330 | |
e56f61be L |
1331 | @kindex --add-needed |
1332 | @kindex --no-add-needed | |
1333 | @item --add-needed | |
1334 | @itemx --no-add-needed | |
ddbb8a31 NC |
1335 | These two options have been deprecated because of the similarity of |
1336 | their names to the @option{--as-needed} and @option{--no-as-needed} | |
1337 | options. They have been replaced by @option{--copy-dt-needed-entries} | |
1338 | and @option{--no-copy-dt-needed-entries}. | |
e56f61be | 1339 | |
252b5132 RH |
1340 | @kindex -assert @var{keyword} |
1341 | @item -assert @var{keyword} | |
1342 | This option is ignored for SunOS compatibility. | |
1343 | ||
1344 | @kindex -Bdynamic | |
1345 | @kindex -dy | |
1346 | @kindex -call_shared | |
1347 | @item -Bdynamic | |
1348 | @itemx -dy | |
1349 | @itemx -call_shared | |
1350 | Link against dynamic libraries. This is only meaningful on platforms | |
1351 | for which shared libraries are supported. This option is normally the | |
1352 | default on such platforms. The different variants of this option are | |
1353 | for compatibility with various systems. You may use this option | |
1354 | multiple times on the command line: it affects library searching for | |
da8bce14 | 1355 | @option{-l} options which follow it. |
252b5132 | 1356 | |
a1ab1d2a UD |
1357 | @kindex -Bgroup |
1358 | @item -Bgroup | |
1359 | Set the @code{DF_1_GROUP} flag in the @code{DT_FLAGS_1} entry in the dynamic | |
1360 | section. This causes the runtime linker to handle lookups in this | |
1361 | object and its dependencies to be performed only inside the group. | |
560e09e9 NC |
1362 | @option{--unresolved-symbols=report-all} is implied. This option is |
1363 | only meaningful on ELF platforms which support shared libraries. | |
a1ab1d2a | 1364 | |
252b5132 RH |
1365 | @kindex -Bstatic |
1366 | @kindex -dn | |
1367 | @kindex -non_shared | |
1368 | @kindex -static | |
a1ab1d2a | 1369 | @item -Bstatic |
252b5132 RH |
1370 | @itemx -dn |
1371 | @itemx -non_shared | |
1372 | @itemx -static | |
1373 | Do not link against shared libraries. This is only meaningful on | |
1374 | platforms for which shared libraries are supported. The different | |
1375 | variants of this option are for compatibility with various systems. You | |
1376 | may use this option multiple times on the command line: it affects | |
560e09e9 | 1377 | library searching for @option{-l} options which follow it. This |
e9156f74 NC |
1378 | option also implies @option{--unresolved-symbols=report-all}. This |
1379 | option can be used with @option{-shared}. Doing so means that a | |
1380 | shared library is being created but that all of the library's external | |
1381 | references must be resolved by pulling in entries from static | |
ece2d90e | 1382 | libraries. |
252b5132 RH |
1383 | |
1384 | @kindex -Bsymbolic | |
1385 | @item -Bsymbolic | |
1386 | When creating a shared library, bind references to global symbols to the | |
1387 | definition within the shared library, if any. Normally, it is possible | |
1388 | for a program linked against a shared library to override the definition | |
e20365c5 L |
1389 | within the shared library. This option can also be used with the |
1390 | @option{--export-dynamic} option, when creating a position independent | |
1391 | executable, to bind references to global symbols to the definition within | |
1392 | the executable. This option is only meaningful on ELF platforms which | |
1393 | support shared libraries and position independent executables. | |
252b5132 | 1394 | |
40b36307 L |
1395 | @kindex -Bsymbolic-functions |
1396 | @item -Bsymbolic-functions | |
1397 | When creating a shared library, bind references to global function | |
c0065db7 | 1398 | symbols to the definition within the shared library, if any. |
e20365c5 L |
1399 | This option can also be used with the @option{--export-dynamic} option, |
1400 | when creating a position independent executable, to bind references | |
1401 | to global function symbols to the definition within the executable. | |
40b36307 | 1402 | This option is only meaningful on ELF platforms which support shared |
e20365c5 | 1403 | libraries and position independent executables. |
40b36307 | 1404 | |
55255dae L |
1405 | @kindex --dynamic-list=@var{dynamic-list-file} |
1406 | @item --dynamic-list=@var{dynamic-list-file} | |
1407 | Specify the name of a dynamic list file to the linker. This is | |
1408 | typically used when creating shared libraries to specify a list of | |
1409 | global symbols whose references shouldn't be bound to the definition | |
1410 | within the shared library, or creating dynamically linked executables | |
1411 | to specify a list of symbols which should be added to the symbol table | |
1412 | in the executable. This option is only meaningful on ELF platforms | |
1413 | which support shared libraries. | |
1414 | ||
1415 | The format of the dynamic list is the same as the version node without | |
1416 | scope and node name. See @ref{VERSION} for more information. | |
1417 | ||
40b36307 L |
1418 | @kindex --dynamic-list-data |
1419 | @item --dynamic-list-data | |
1420 | Include all global data symbols to the dynamic list. | |
1421 | ||
1422 | @kindex --dynamic-list-cpp-new | |
1423 | @item --dynamic-list-cpp-new | |
1424 | Provide the builtin dynamic list for C++ operator new and delete. It | |
1425 | is mainly useful for building shared libstdc++. | |
1426 | ||
0b8a70d9 L |
1427 | @kindex --dynamic-list-cpp-typeinfo |
1428 | @item --dynamic-list-cpp-typeinfo | |
1429 | Provide the builtin dynamic list for C++ runtime type identification. | |
1430 | ||
252b5132 RH |
1431 | @kindex --check-sections |
1432 | @kindex --no-check-sections | |
1433 | @item --check-sections | |
308b1ffd | 1434 | @itemx --no-check-sections |
252b5132 | 1435 | Asks the linker @emph{not} to check section addresses after they have |
7d816a17 | 1436 | been assigned to see if there are any overlaps. Normally the linker will |
252b5132 RH |
1437 | perform this check, and if it finds any overlaps it will produce |
1438 | suitable error messages. The linker does know about, and does make | |
1439 | allowances for sections in overlays. The default behaviour can be | |
a05a5b64 | 1440 | restored by using the command-line switch @option{--check-sections}. |
02b0b1aa NS |
1441 | Section overlap is not usually checked for relocatable links. You can |
1442 | force checking in that case by using the @option{--check-sections} | |
1443 | option. | |
252b5132 | 1444 | |
ddbb8a31 NC |
1445 | @kindex --copy-dt-needed-entries |
1446 | @kindex --no-copy-dt-needed-entries | |
1447 | @item --copy-dt-needed-entries | |
1448 | @itemx --no-copy-dt-needed-entries | |
9d5777a3 | 1449 | This option affects the treatment of dynamic libraries referred to |
ddbb8a31 | 1450 | by DT_NEEDED tags @emph{inside} ELF dynamic libraries mentioned on the |
08efffb8 | 1451 | command line. Normally the linker won't add a DT_NEEDED tag to the |
ddbb8a31 | 1452 | output binary for each library mentioned in a DT_NEEDED tag in an |
08efffb8 | 1453 | input dynamic library. With @option{--copy-dt-needed-entries} |
ddbb8a31 | 1454 | specified on the command line however any dynamic libraries that |
08efffb8 MM |
1455 | follow it will have their DT_NEEDED entries added. The default |
1456 | behaviour can be restored with @option{--no-copy-dt-needed-entries}. | |
ddbb8a31 NC |
1457 | |
1458 | This option also has an effect on the resolution of symbols in dynamic | |
08efffb8 MM |
1459 | libraries. With @option{--copy-dt-needed-entries} dynamic libraries |
1460 | mentioned on the command line will be recursively searched, following | |
1461 | their DT_NEEDED tags to other libraries, in order to resolve symbols | |
1462 | required by the output binary. With the default setting however | |
1463 | the searching of dynamic libraries that follow it will stop with the | |
1464 | dynamic library itself. No DT_NEEDED links will be traversed to resolve | |
ddbb8a31 NC |
1465 | symbols. |
1466 | ||
252b5132 RH |
1467 | @cindex cross reference table |
1468 | @kindex --cref | |
1469 | @item --cref | |
1470 | Output a cross reference table. If a linker map file is being | |
1471 | generated, the cross reference table is printed to the map file. | |
1472 | Otherwise, it is printed on the standard output. | |
1473 | ||
1474 | The format of the table is intentionally simple, so that it may be | |
1475 | easily processed by a script if necessary. The symbols are printed out, | |
1476 | sorted by name. For each symbol, a list of file names is given. If the | |
1477 | symbol is defined, the first file listed is the location of the | |
049c1c8e NC |
1478 | definition. If the symbol is defined as a common value then any files |
1479 | where this happens appear next. Finally any files that reference the | |
1480 | symbol are listed. | |
252b5132 | 1481 | |
4818e05f AM |
1482 | @cindex common allocation |
1483 | @kindex --no-define-common | |
1484 | @item --no-define-common | |
1485 | This option inhibits the assignment of addresses to common symbols. | |
1486 | The script command @code{INHIBIT_COMMON_ALLOCATION} has the same effect. | |
1487 | @xref{Miscellaneous Commands}. | |
1488 | ||
1489 | The @samp{--no-define-common} option allows decoupling | |
1490 | the decision to assign addresses to Common symbols from the choice | |
1491 | of the output file type; otherwise a non-Relocatable output type | |
1492 | forces assigning addresses to Common symbols. | |
1493 | Using @samp{--no-define-common} allows Common symbols that are referenced | |
1494 | from a shared library to be assigned addresses only in the main program. | |
1495 | This eliminates the unused duplicate space in the shared library, | |
1496 | and also prevents any possible confusion over resolving to the wrong | |
1497 | duplicate when there are many dynamic modules with specialized search | |
1498 | paths for runtime symbol resolution. | |
1499 | ||
7bdf4127 AB |
1500 | @cindex group allocation in linker script |
1501 | @cindex section groups | |
1502 | @cindex COMDAT | |
1503 | @kindex --force-group-allocation | |
1504 | @item --force-group-allocation | |
1505 | This option causes the linker to place section group members like | |
1506 | normal input sections, and to delete the section groups. This is the | |
1507 | default behaviour for a final link but this option can be used to | |
1508 | change the behaviour of a relocatable link (@samp{-r}). The script | |
1509 | command @code{FORCE_GROUP_ALLOCATION} has the same | |
1510 | effect. @xref{Miscellaneous Commands}. | |
1511 | ||
252b5132 | 1512 | @cindex symbols, from command line |
2509a395 SL |
1513 | @kindex --defsym=@var{symbol}=@var{exp} |
1514 | @item --defsym=@var{symbol}=@var{expression} | |
252b5132 RH |
1515 | Create a global symbol in the output file, containing the absolute |
1516 | address given by @var{expression}. You may use this option as many | |
1517 | times as necessary to define multiple symbols in the command line. A | |
1518 | limited form of arithmetic is supported for the @var{expression} in this | |
1519 | context: you may give a hexadecimal constant or the name of an existing | |
1520 | symbol, or use @code{+} and @code{-} to add or subtract hexadecimal | |
1521 | constants or symbols. If you need more elaborate expressions, consider | |
66bc8739 AM |
1522 | using the linker command language from a script (@pxref{Assignments}). |
1523 | @emph{Note:} there should be no white space between @var{symbol}, the | |
1524 | equals sign (``@key{=}''), and @var{expression}. | |
252b5132 RH |
1525 | |
1526 | @cindex demangling, from command line | |
28c309a2 | 1527 | @kindex --demangle[=@var{style}] |
252b5132 | 1528 | @kindex --no-demangle |
28c309a2 | 1529 | @item --demangle[=@var{style}] |
252b5132 RH |
1530 | @itemx --no-demangle |
1531 | These options control whether to demangle symbol names in error messages | |
1532 | and other output. When the linker is told to demangle, it tries to | |
1533 | present symbol names in a readable fashion: it strips leading | |
1534 | underscores if they are used by the object file format, and converts C++ | |
a1ab1d2a UD |
1535 | mangled symbol names into user readable names. Different compilers have |
1536 | different mangling styles. The optional demangling style argument can be used | |
1537 | to choose an appropriate demangling style for your compiler. The linker will | |
28c309a2 NC |
1538 | demangle by default unless the environment variable @samp{COLLECT_NO_DEMANGLE} |
1539 | is set. These options may be used to override the default. | |
252b5132 RH |
1540 | |
1541 | @cindex dynamic linker, from command line | |
506eee22 | 1542 | @kindex -I@var{file} |
2509a395 SL |
1543 | @kindex --dynamic-linker=@var{file} |
1544 | @item -I@var{file} | |
1545 | @itemx --dynamic-linker=@var{file} | |
252b5132 RH |
1546 | Set the name of the dynamic linker. This is only meaningful when |
1547 | generating dynamically linked ELF executables. The default dynamic | |
1548 | linker is normally correct; don't use this unless you know what you are | |
1549 | doing. | |
1550 | ||
9b8b325a RF |
1551 | @kindex --no-dynamic-linker |
1552 | @item --no-dynamic-linker | |
1553 | When producing an executable file, omit the request for a dynamic | |
1554 | linker to be used at load-time. This is only meaningful for ELF | |
1555 | executables that contain dynamic relocations, and usually requires | |
1556 | entry point code that is capable of processing these relocations. | |
1557 | ||
a70f34c0 NC |
1558 | @kindex --embedded-relocs |
1559 | @item --embedded-relocs | |
1560 | This option is similar to the @option{--emit-relocs} option except | |
1561 | that the relocs are stored in a target specific section. This option | |
1562 | is only supported by the @samp{BFIN}, @samp{CR16} and @emph{M68K} | |
1563 | targets. | |
1564 | ||
3f0a5f17 ME |
1565 | @kindex --disable-multiple-abs-defs |
1566 | @item --disable-multiple-abs-defs | |
1567 | Do not allow multiple definitions with symbols included | |
1568 | in filename invoked by -R or --just-symbols | |
1569 | ||
7ce691ae | 1570 | @kindex --fatal-warnings |
0fe58ccd | 1571 | @kindex --no-fatal-warnings |
7ce691ae | 1572 | @item --fatal-warnings |
0fe58ccd NC |
1573 | @itemx --no-fatal-warnings |
1574 | Treat all warnings as errors. The default behaviour can be restored | |
1575 | with the option @option{--no-fatal-warnings}. | |
7ce691ae | 1576 | |
252b5132 RH |
1577 | @kindex --force-exe-suffix |
1578 | @item --force-exe-suffix | |
1579 | Make sure that an output file has a .exe suffix. | |
1580 | ||
1581 | If a successfully built fully linked output file does not have a | |
1582 | @code{.exe} or @code{.dll} suffix, this option forces the linker to copy | |
1583 | the output file to one of the same name with a @code{.exe} suffix. This | |
1584 | option is useful when using unmodified Unix makefiles on a Microsoft | |
1585 | Windows host, since some versions of Windows won't run an image unless | |
1586 | it ends in a @code{.exe} suffix. | |
1587 | ||
1588 | @kindex --gc-sections | |
1589 | @kindex --no-gc-sections | |
1590 | @cindex garbage collection | |
c17d87de NC |
1591 | @item --gc-sections |
1592 | @itemx --no-gc-sections | |
252b5132 | 1593 | Enable garbage collection of unused input sections. It is ignored on |
ac69cbc6 | 1594 | targets that do not support this option. The default behaviour (of not |
b3549761 | 1595 | performing this garbage collection) can be restored by specifying |
0f088b2a KT |
1596 | @samp{--no-gc-sections} on the command line. Note that garbage |
1597 | collection for COFF and PE format targets is supported, but the | |
1598 | implementation is currently considered to be experimental. | |
252b5132 | 1599 | |
d5465ba2 AM |
1600 | @samp{--gc-sections} decides which input sections are used by |
1601 | examining symbols and relocations. The section containing the entry | |
1602 | symbol and all sections containing symbols undefined on the | |
1603 | command-line will be kept, as will sections containing symbols | |
1604 | referenced by dynamic objects. Note that when building shared | |
1605 | libraries, the linker must assume that any visible symbol is | |
1606 | referenced. Once this initial set of sections has been determined, | |
1607 | the linker recursively marks as used any section referenced by their | |
1608 | relocations. See @samp{--entry} and @samp{--undefined}. | |
1609 | ||
ac69cbc6 | 1610 | This option can be set when doing a partial link (enabled with option |
9d5777a3 | 1611 | @samp{-r}). In this case the root of symbols kept must be explicitly |
ac69cbc6 TG |
1612 | specified either by an @samp{--entry} or @samp{--undefined} option or by |
1613 | a @code{ENTRY} command in the linker script. | |
1614 | ||
c17d87de NC |
1615 | @kindex --print-gc-sections |
1616 | @kindex --no-print-gc-sections | |
1617 | @cindex garbage collection | |
1618 | @item --print-gc-sections | |
1619 | @itemx --no-print-gc-sections | |
1620 | List all sections removed by garbage collection. The listing is | |
1621 | printed on stderr. This option is only effective if garbage | |
1622 | collection has been enabled via the @samp{--gc-sections}) option. The | |
1623 | default behaviour (of not listing the sections that are removed) can | |
1624 | be restored by specifying @samp{--no-print-gc-sections} on the command | |
1625 | line. | |
1626 | ||
22185505 | 1627 | @kindex --gc-keep-exported |
1628 | @cindex garbage collection | |
1629 | @item --gc-keep-exported | |
1630 | When @samp{--gc-sections} is enabled, this option prevents garbage | |
1631 | collection of unused input sections that contain global symbols having | |
1632 | default or protected visibility. This option is intended to be used for | |
1633 | executables where unreferenced sections would otherwise be garbage | |
1634 | collected regardless of the external visibility of contained symbols. | |
1635 | Note that this option has no effect when linking shared objects since | |
1636 | it is already the default behaviour. This option is only supported for | |
1637 | ELF format targets. | |
1638 | ||
30824704 RM |
1639 | @kindex --print-output-format |
1640 | @cindex output format | |
1641 | @item --print-output-format | |
1642 | Print the name of the default output format (perhaps influenced by | |
1643 | other command-line options). This is the string that would appear | |
1644 | in an @code{OUTPUT_FORMAT} linker script command (@pxref{File Commands}). | |
1645 | ||
3604cb1f TG |
1646 | @kindex --print-memory-usage |
1647 | @cindex memory usage | |
1648 | @item --print-memory-usage | |
1649 | Print used size, total size and used size of memory regions created with | |
1650 | the @ref{MEMORY} command. This is useful on embedded targets to have a | |
1651 | quick view of amount of free memory. The format of the output has one | |
1652 | headline and one line per region. It is both human readable and easily | |
1653 | parsable by tools. Here is an example of an output: | |
1654 | ||
1655 | @smallexample | |
1656 | Memory region Used Size Region Size %age Used | |
1657 | ROM: 256 KB 1 MB 25.00% | |
1658 | RAM: 32 B 2 GB 0.00% | |
1659 | @end smallexample | |
1660 | ||
252b5132 RH |
1661 | @cindex help |
1662 | @cindex usage | |
1663 | @kindex --help | |
1664 | @item --help | |
1665 | Print a summary of the command-line options on the standard output and exit. | |
1666 | ||
ea20a7da CC |
1667 | @kindex --target-help |
1668 | @item --target-help | |
1669 | Print a summary of all target specific options on the standard output and exit. | |
1670 | ||
2509a395 SL |
1671 | @kindex -Map=@var{mapfile} |
1672 | @item -Map=@var{mapfile} | |
252b5132 | 1673 | Print a link map to the file @var{mapfile}. See the description of the |
560e09e9 | 1674 | @option{-M} option, above. |
252b5132 RH |
1675 | |
1676 | @cindex memory usage | |
1677 | @kindex --no-keep-memory | |
1678 | @item --no-keep-memory | |
ff5dcc92 SC |
1679 | @command{ld} normally optimizes for speed over memory usage by caching the |
1680 | symbol tables of input files in memory. This option tells @command{ld} to | |
252b5132 | 1681 | instead optimize for memory usage, by rereading the symbol tables as |
ff5dcc92 | 1682 | necessary. This may be required if @command{ld} runs out of memory space |
252b5132 RH |
1683 | while linking a large executable. |
1684 | ||
1685 | @kindex --no-undefined | |
a1ab1d2a | 1686 | @kindex -z defs |
97a232d7 | 1687 | @kindex -z undefs |
252b5132 | 1688 | @item --no-undefined |
a1ab1d2a | 1689 | @itemx -z defs |
560e09e9 NC |
1690 | Report unresolved symbol references from regular object files. This |
1691 | is done even if the linker is creating a non-symbolic shared library. | |
1692 | The switch @option{--[no-]allow-shlib-undefined} controls the | |
1693 | behaviour for reporting unresolved references found in shared | |
ece2d90e | 1694 | libraries being linked in. |
252b5132 | 1695 | |
97a232d7 NC |
1696 | The effects of this option can be reverted by using @code{-z undefs}. |
1697 | ||
aa713662 L |
1698 | @kindex --allow-multiple-definition |
1699 | @kindex -z muldefs | |
1700 | @item --allow-multiple-definition | |
1701 | @itemx -z muldefs | |
1702 | Normally when a symbol is defined multiple times, the linker will | |
1703 | report a fatal error. These options allow multiple definitions and the | |
1704 | first definition will be used. | |
1705 | ||
b79e8c78 | 1706 | @kindex --allow-shlib-undefined |
ae9a127f | 1707 | @kindex --no-allow-shlib-undefined |
b79e8c78 | 1708 | @item --allow-shlib-undefined |
ae9a127f | 1709 | @itemx --no-allow-shlib-undefined |
903249d7 | 1710 | Allows or disallows undefined symbols in shared libraries. |
560e09e9 NC |
1711 | This switch is similar to @option{--no-undefined} except that it |
1712 | determines the behaviour when the undefined symbols are in a | |
1713 | shared library rather than a regular object file. It does not affect | |
1714 | how undefined symbols in regular object files are handled. | |
1715 | ||
903249d7 NC |
1716 | The default behaviour is to report errors for any undefined symbols |
1717 | referenced in shared libraries if the linker is being used to create | |
1718 | an executable, but to allow them if the linker is being used to create | |
1719 | a shared library. | |
1720 | ||
1721 | The reasons for allowing undefined symbol references in shared | |
1722 | libraries specified at link time are that: | |
1723 | ||
1724 | @itemize @bullet | |
1725 | @item | |
1726 | A shared library specified at link time may not be the same as the one | |
1727 | that is available at load time, so the symbol might actually be | |
1728 | resolvable at load time. | |
1729 | @item | |
1730 | There are some operating systems, eg BeOS and HPPA, where undefined | |
1731 | symbols in shared libraries are normal. | |
1732 | ||
1733 | The BeOS kernel for example patches shared libraries at load time to | |
1734 | select whichever function is most appropriate for the current | |
1735 | architecture. This is used, for example, to dynamically select an | |
1736 | appropriate memset function. | |
1737 | @end itemize | |
b79e8c78 | 1738 | |
31941635 L |
1739 | @kindex --no-undefined-version |
1740 | @item --no-undefined-version | |
1741 | Normally when a symbol has an undefined version, the linker will ignore | |
1742 | it. This option disallows symbols with undefined version and a fatal error | |
1743 | will be issued instead. | |
1744 | ||
3e3b46e5 PB |
1745 | @kindex --default-symver |
1746 | @item --default-symver | |
1747 | Create and use a default symbol version (the soname) for unversioned | |
fc0e6df6 PB |
1748 | exported symbols. |
1749 | ||
1750 | @kindex --default-imported-symver | |
1751 | @item --default-imported-symver | |
1752 | Create and use a default symbol version (the soname) for unversioned | |
1753 | imported symbols. | |
3e3b46e5 | 1754 | |
252b5132 RH |
1755 | @kindex --no-warn-mismatch |
1756 | @item --no-warn-mismatch | |
ff5dcc92 | 1757 | Normally @command{ld} will give an error if you try to link together input |
252b5132 RH |
1758 | files that are mismatched for some reason, perhaps because they have |
1759 | been compiled for different processors or for different endiannesses. | |
ff5dcc92 | 1760 | This option tells @command{ld} that it should silently permit such possible |
252b5132 RH |
1761 | errors. This option should only be used with care, in cases when you |
1762 | have taken some special action that ensures that the linker errors are | |
1763 | inappropriate. | |
1764 | ||
fe7929ce AM |
1765 | @kindex --no-warn-search-mismatch |
1766 | @item --no-warn-search-mismatch | |
1767 | Normally @command{ld} will give a warning if it finds an incompatible | |
1768 | library during a library search. This option silences the warning. | |
1769 | ||
252b5132 RH |
1770 | @kindex --no-whole-archive |
1771 | @item --no-whole-archive | |
ff5dcc92 | 1772 | Turn off the effect of the @option{--whole-archive} option for subsequent |
252b5132 RH |
1773 | archive files. |
1774 | ||
1775 | @cindex output file after errors | |
1776 | @kindex --noinhibit-exec | |
1777 | @item --noinhibit-exec | |
1778 | Retain the executable output file whenever it is still usable. | |
1779 | Normally, the linker will not produce an output file if it encounters | |
1780 | errors during the link process; it exits without writing an output file | |
1781 | when it issues any error whatsoever. | |
1782 | ||
0a9c1c8e CD |
1783 | @kindex -nostdlib |
1784 | @item -nostdlib | |
1785 | Only search library directories explicitly specified on the | |
1786 | command line. Library directories specified in linker scripts | |
1787 | (including linker scripts specified on the command line) are ignored. | |
1788 | ||
252b5132 | 1789 | @ifclear SingleFormat |
2509a395 SL |
1790 | @kindex --oformat=@var{output-format} |
1791 | @item --oformat=@var{output-format} | |
ff5dcc92 SC |
1792 | @command{ld} may be configured to support more than one kind of object |
1793 | file. If your @command{ld} is configured this way, you can use the | |
252b5132 | 1794 | @samp{--oformat} option to specify the binary format for the output |
ff5dcc92 SC |
1795 | object file. Even when @command{ld} is configured to support alternative |
1796 | object formats, you don't usually need to specify this, as @command{ld} | |
252b5132 RH |
1797 | should be configured to produce as a default output format the most |
1798 | usual format on each machine. @var{output-format} is a text string, the | |
1799 | name of a particular format supported by the BFD libraries. (You can | |
1800 | list the available binary formats with @samp{objdump -i}.) The script | |
1801 | command @code{OUTPUT_FORMAT} can also specify the output format, but | |
1802 | this option overrides it. @xref{BFD}. | |
1803 | @end ifclear | |
1804 | ||
76359541 TP |
1805 | @kindex --out-implib |
1806 | @item --out-implib @var{file} | |
1807 | Create an import library in @var{file} corresponding to the executable | |
1808 | the linker is generating (eg. a DLL or ELF program). This import | |
1809 | library (which should be called @code{*.dll.a} or @code{*.a} for DLLs) | |
1810 | may be used to link clients against the generated executable; this | |
1811 | behaviour makes it possible to skip a separate import library creation | |
1812 | step (eg. @code{dlltool} for DLLs). This option is only available for | |
1813 | the i386 PE and ELF targetted ports of the linker. | |
1814 | ||
36af4a4e JJ |
1815 | @kindex -pie |
1816 | @kindex --pic-executable | |
1817 | @item -pie | |
1818 | @itemx --pic-executable | |
1819 | @cindex position independent executables | |
1820 | Create a position independent executable. This is currently only supported on | |
1821 | ELF platforms. Position independent executables are similar to shared | |
1822 | libraries in that they are relocated by the dynamic linker to the virtual | |
7e7d5768 | 1823 | address the OS chooses for them (which can vary between invocations). Like |
36af4a4e JJ |
1824 | normal dynamically linked executables they can be executed and symbols |
1825 | defined in the executable cannot be overridden by shared libraries. | |
1826 | ||
252b5132 RH |
1827 | @kindex -qmagic |
1828 | @item -qmagic | |
1829 | This option is ignored for Linux compatibility. | |
1830 | ||
1831 | @kindex -Qy | |
1832 | @item -Qy | |
1833 | This option is ignored for SVR4 compatibility. | |
1834 | ||
1835 | @kindex --relax | |
1836 | @cindex synthesizing linker | |
1837 | @cindex relaxing addressing modes | |
28d5f677 | 1838 | @cindex --no-relax |
252b5132 | 1839 | @item --relax |
28d5f677 | 1840 | @itemx --no-relax |
a1ab1d2a | 1841 | An option with machine dependent effects. |
252b5132 RH |
1842 | @ifset GENERIC |
1843 | This option is only supported on a few targets. | |
1844 | @end ifset | |
1845 | @ifset H8300 | |
ff5dcc92 | 1846 | @xref{H8/300,,@command{ld} and the H8/300}. |
252b5132 | 1847 | @end ifset |
e0001a05 NC |
1848 | @ifset XTENSA |
1849 | @xref{Xtensa,, @command{ld} and Xtensa Processors}. | |
1850 | @end ifset | |
93fd0973 SC |
1851 | @ifset M68HC11 |
1852 | @xref{M68HC11/68HC12,,@command{ld} and the 68HC11 and 68HC12}. | |
1853 | @end ifset | |
78058a5e SL |
1854 | @ifset NIOSII |
1855 | @xref{Nios II,,@command{ld} and the Altera Nios II}. | |
1856 | @end ifset | |
2a60a7a8 AM |
1857 | @ifset POWERPC |
1858 | @xref{PowerPC ELF32,,@command{ld} and PowerPC 32-bit ELF Support}. | |
1859 | @end ifset | |
252b5132 | 1860 | |
28d5f677 NC |
1861 | On some platforms the @samp{--relax} option performs target specific, |
1862 | global optimizations that become possible when the linker resolves | |
1863 | addressing in the program, such as relaxing address modes, | |
1864 | synthesizing new instructions, selecting shorter version of current | |
11e7fd74 | 1865 | instructions, and combining constant values. |
252b5132 RH |
1866 | |
1867 | On some platforms these link time global optimizations may make symbolic | |
1868 | debugging of the resulting executable impossible. | |
1869 | @ifset GENERIC | |
28d5f677 NC |
1870 | This is known to be the case for the Matsushita MN10200 and MN10300 |
1871 | family of processors. | |
252b5132 RH |
1872 | @end ifset |
1873 | ||
1874 | @ifset GENERIC | |
1875 | On platforms where this is not supported, @samp{--relax} is accepted, | |
1876 | but ignored. | |
1877 | @end ifset | |
1878 | ||
28d5f677 NC |
1879 | On platforms where @samp{--relax} is accepted the option |
1880 | @samp{--no-relax} can be used to disable the feature. | |
1881 | ||
252b5132 RH |
1882 | @cindex retaining specified symbols |
1883 | @cindex stripping all but some symbols | |
1884 | @cindex symbols, retaining selectively | |
2509a395 SL |
1885 | @kindex --retain-symbols-file=@var{filename} |
1886 | @item --retain-symbols-file=@var{filename} | |
252b5132 RH |
1887 | Retain @emph{only} the symbols listed in the file @var{filename}, |
1888 | discarding all others. @var{filename} is simply a flat file, with one | |
1889 | symbol name per line. This option is especially useful in environments | |
1890 | @ifset GENERIC | |
1891 | (such as VxWorks) | |
1892 | @end ifset | |
1893 | where a large global symbol table is accumulated gradually, to conserve | |
1894 | run-time memory. | |
1895 | ||
1896 | @samp{--retain-symbols-file} does @emph{not} discard undefined symbols, | |
1897 | or symbols needed for relocations. | |
1898 | ||
1899 | You may only specify @samp{--retain-symbols-file} once in the command | |
1900 | line. It overrides @samp{-s} and @samp{-S}. | |
1901 | ||
1902 | @ifset GENERIC | |
2509a395 | 1903 | @item -rpath=@var{dir} |
252b5132 | 1904 | @cindex runtime library search path |
2509a395 | 1905 | @kindex -rpath=@var{dir} |
252b5132 | 1906 | Add a directory to the runtime library search path. This is used when |
ff5dcc92 | 1907 | linking an ELF executable with shared objects. All @option{-rpath} |
252b5132 | 1908 | arguments are concatenated and passed to the runtime linker, which uses |
ff5dcc92 | 1909 | them to locate shared objects at runtime. The @option{-rpath} option is |
252b5132 RH |
1910 | also used when locating shared objects which are needed by shared |
1911 | objects explicitly included in the link; see the description of the | |
ff5dcc92 | 1912 | @option{-rpath-link} option. If @option{-rpath} is not used when linking an |
252b5132 RH |
1913 | ELF executable, the contents of the environment variable |
1914 | @code{LD_RUN_PATH} will be used if it is defined. | |
1915 | ||
ff5dcc92 | 1916 | The @option{-rpath} option may also be used on SunOS. By default, on |
d8506323 | 1917 | SunOS, the linker will form a runtime search path out of all the |
ff5dcc92 SC |
1918 | @option{-L} options it is given. If a @option{-rpath} option is used, the |
1919 | runtime search path will be formed exclusively using the @option{-rpath} | |
1920 | options, ignoring the @option{-L} options. This can be useful when using | |
1921 | gcc, which adds many @option{-L} options which may be on NFS mounted | |
b45619c0 | 1922 | file systems. |
252b5132 | 1923 | |
ff5dcc92 | 1924 | For compatibility with other ELF linkers, if the @option{-R} option is |
252b5132 | 1925 | followed by a directory name, rather than a file name, it is treated as |
ff5dcc92 | 1926 | the @option{-rpath} option. |
252b5132 RH |
1927 | @end ifset |
1928 | ||
1929 | @ifset GENERIC | |
1930 | @cindex link-time runtime library search path | |
2509a395 SL |
1931 | @kindex -rpath-link=@var{dir} |
1932 | @item -rpath-link=@var{dir} | |
252b5132 RH |
1933 | When using ELF or SunOS, one shared library may require another. This |
1934 | happens when an @code{ld -shared} link includes a shared library as one | |
1935 | of the input files. | |
1936 | ||
1937 | When the linker encounters such a dependency when doing a non-shared, | |
1938 | non-relocatable link, it will automatically try to locate the required | |
1939 | shared library and include it in the link, if it is not included | |
ff5dcc92 | 1940 | explicitly. In such a case, the @option{-rpath-link} option |
252b5132 | 1941 | specifies the first set of directories to search. The |
ff5dcc92 | 1942 | @option{-rpath-link} option may specify a sequence of directory names |
252b5132 RH |
1943 | either by specifying a list of names separated by colons, or by |
1944 | appearing multiple times. | |
1945 | ||
e680a6b6 NC |
1946 | The tokens @var{$ORIGIN} and @var{$LIB} can appear in these search |
1947 | directories. They will be replaced by the full path to the directory | |
1948 | containing the program or shared object in the case of @var{$ORIGIN} | |
1949 | and either @samp{lib} - for 32-bit binaries - or @samp{lib64} - for | |
1950 | 64-bit binaries - in the case of @var{$LIB}. | |
1951 | ||
1952 | The alternative form of these tokens - @var{$@{ORIGIN@}} and | |
1953 | @var{$@{LIB@}} can also be used. The token @var{$PLATFORM} is not | |
1954 | supported. | |
1955 | ||
28c309a2 NC |
1956 | This option should be used with caution as it overrides the search path |
1957 | that may have been hard compiled into a shared library. In such a case it | |
1958 | is possible to use unintentionally a different search path than the | |
1959 | runtime linker would do. | |
1960 | ||
252b5132 | 1961 | The linker uses the following search paths to locate required shared |
ece2d90e | 1962 | libraries: |
252b5132 RH |
1963 | @enumerate |
1964 | @item | |
ff5dcc92 | 1965 | Any directories specified by @option{-rpath-link} options. |
252b5132 | 1966 | @item |
ff5dcc92 SC |
1967 | Any directories specified by @option{-rpath} options. The difference |
1968 | between @option{-rpath} and @option{-rpath-link} is that directories | |
1969 | specified by @option{-rpath} options are included in the executable and | |
1970 | used at runtime, whereas the @option{-rpath-link} option is only effective | |
ece2d90e NC |
1971 | at link time. Searching @option{-rpath} in this way is only supported |
1972 | by native linkers and cross linkers which have been configured with | |
1973 | the @option{--with-sysroot} option. | |
252b5132 | 1974 | @item |
e2a83dd0 NC |
1975 | On an ELF system, for native linkers, if the @option{-rpath} and |
1976 | @option{-rpath-link} options were not used, search the contents of the | |
1977 | environment variable @code{LD_RUN_PATH}. | |
252b5132 | 1978 | @item |
ff5dcc92 SC |
1979 | On SunOS, if the @option{-rpath} option was not used, search any |
1980 | directories specified using @option{-L} options. | |
252b5132 | 1981 | @item |
a1b8d843 | 1982 | For a native linker, search the contents of the environment |
e2a83dd0 | 1983 | variable @code{LD_LIBRARY_PATH}. |
252b5132 | 1984 | @item |
ec4eb78a L |
1985 | For a native ELF linker, the directories in @code{DT_RUNPATH} or |
1986 | @code{DT_RPATH} of a shared library are searched for shared | |
1987 | libraries needed by it. The @code{DT_RPATH} entries are ignored if | |
1988 | @code{DT_RUNPATH} entries exist. | |
1989 | @item | |
252b5132 RH |
1990 | The default directories, normally @file{/lib} and @file{/usr/lib}. |
1991 | @item | |
1992 | For a native linker on an ELF system, if the file @file{/etc/ld.so.conf} | |
1993 | exists, the list of directories found in that file. | |
1994 | @end enumerate | |
1995 | ||
1996 | If the required shared library is not found, the linker will issue a | |
1997 | warning and continue with the link. | |
1998 | @end ifset | |
1999 | ||
2000 | @kindex -shared | |
2001 | @kindex -Bshareable | |
2002 | @item -shared | |
2003 | @itemx -Bshareable | |
2004 | @cindex shared libraries | |
2005 | Create a shared library. This is currently only supported on ELF, XCOFF | |
2006 | and SunOS platforms. On SunOS, the linker will automatically create a | |
ff5dcc92 | 2007 | shared library if the @option{-e} option is not used and there are |
252b5132 RH |
2008 | undefined symbols in the link. |
2009 | ||
252b5132 | 2010 | @kindex --sort-common |
2509a395 SL |
2011 | @item --sort-common |
2012 | @itemx --sort-common=ascending | |
2013 | @itemx --sort-common=descending | |
de7dd2bd NC |
2014 | This option tells @command{ld} to sort the common symbols by alignment in |
2015 | ascending or descending order when it places them in the appropriate output | |
2016 | sections. The symbol alignments considered are sixteen-byte or larger, | |
2017 | eight-byte, four-byte, two-byte, and one-byte. This is to prevent gaps | |
2018 | between symbols due to alignment constraints. If no sorting order is | |
2019 | specified, then descending order is assumed. | |
252b5132 | 2020 | |
2509a395 SL |
2021 | @kindex --sort-section=name |
2022 | @item --sort-section=name | |
bcaa7b3e L |
2023 | This option will apply @code{SORT_BY_NAME} to all wildcard section |
2024 | patterns in the linker script. | |
2025 | ||
2509a395 SL |
2026 | @kindex --sort-section=alignment |
2027 | @item --sort-section=alignment | |
bcaa7b3e L |
2028 | This option will apply @code{SORT_BY_ALIGNMENT} to all wildcard section |
2029 | patterns in the linker script. | |
2030 | ||
a70f34c0 NC |
2031 | @kindex --spare-dynamic-tags |
2032 | @item --spare-dynamic-tags=@var{count} | |
2033 | This option specifies the number of empty slots to leave in the | |
2034 | .dynamic section of ELF shared objects. Empty slots may be needed by | |
2035 | post processing tools, such as the prelinker. The default is 5. | |
2036 | ||
252b5132 | 2037 | @kindex --split-by-file |
2509a395 | 2038 | @item --split-by-file[=@var{size}] |
ff5dcc92 | 2039 | Similar to @option{--split-by-reloc} but creates a new output section for |
a854a4a7 AM |
2040 | each input file when @var{size} is reached. @var{size} defaults to a |
2041 | size of 1 if not given. | |
252b5132 RH |
2042 | |
2043 | @kindex --split-by-reloc | |
2509a395 | 2044 | @item --split-by-reloc[=@var{count}] |
a854a4a7 | 2045 | Tries to creates extra sections in the output file so that no single |
252b5132 | 2046 | output section in the file contains more than @var{count} relocations. |
a854a4a7 | 2047 | This is useful when generating huge relocatable files for downloading into |
252b5132 RH |
2048 | certain real time kernels with the COFF object file format; since COFF |
2049 | cannot represent more than 65535 relocations in a single section. Note | |
2050 | that this will fail to work with object file formats which do not | |
2051 | support arbitrary sections. The linker will not split up individual | |
2052 | input sections for redistribution, so if a single input section contains | |
2053 | more than @var{count} relocations one output section will contain that | |
a854a4a7 | 2054 | many relocations. @var{count} defaults to a value of 32768. |
252b5132 RH |
2055 | |
2056 | @kindex --stats | |
2057 | @item --stats | |
2058 | Compute and display statistics about the operation of the linker, such | |
2059 | as execution time and memory usage. | |
2060 | ||
2509a395 | 2061 | @kindex --sysroot=@var{directory} |
e2243057 RS |
2062 | @item --sysroot=@var{directory} |
2063 | Use @var{directory} as the location of the sysroot, overriding the | |
2064 | configure-time default. This option is only supported by linkers | |
2065 | that were configured using @option{--with-sysroot}. | |
2066 | ||
a70f34c0 NC |
2067 | @kindex --task-link |
2068 | @item --task-link | |
2069 | This is used by COFF/PE based targets to create a task-linked object | |
2070 | file where all of the global symbols have been converted to statics. | |
2071 | ||
252b5132 RH |
2072 | @kindex --traditional-format |
2073 | @cindex traditional format | |
2074 | @item --traditional-format | |
ff5dcc92 SC |
2075 | For some targets, the output of @command{ld} is different in some ways from |
2076 | the output of some existing linker. This switch requests @command{ld} to | |
252b5132 RH |
2077 | use the traditional format instead. |
2078 | ||
2079 | @cindex dbx | |
ff5dcc92 | 2080 | For example, on SunOS, @command{ld} combines duplicate entries in the |
252b5132 RH |
2081 | symbol string table. This can reduce the size of an output file with |
2082 | full debugging information by over 30 percent. Unfortunately, the SunOS | |
2083 | @code{dbx} program can not read the resulting program (@code{gdb} has no | |
ff5dcc92 | 2084 | trouble). The @samp{--traditional-format} switch tells @command{ld} to not |
252b5132 RH |
2085 | combine duplicate entries. |
2086 | ||
2509a395 SL |
2087 | @kindex --section-start=@var{sectionname}=@var{org} |
2088 | @item --section-start=@var{sectionname}=@var{org} | |
176355da NC |
2089 | Locate a section in the output file at the absolute |
2090 | address given by @var{org}. You may use this option as many | |
2091 | times as necessary to locate multiple sections in the command | |
2092 | line. | |
2093 | @var{org} must be a single hexadecimal integer; | |
2094 | for compatibility with other linkers, you may omit the leading | |
2095 | @samp{0x} usually associated with hexadecimal values. @emph{Note:} there | |
2096 | should be no white space between @var{sectionname}, the equals | |
2097 | sign (``@key{=}''), and @var{org}. | |
2098 | ||
2509a395 SL |
2099 | @kindex -Tbss=@var{org} |
2100 | @kindex -Tdata=@var{org} | |
2101 | @kindex -Ttext=@var{org} | |
252b5132 | 2102 | @cindex segment origins, cmd line |
2509a395 SL |
2103 | @item -Tbss=@var{org} |
2104 | @itemx -Tdata=@var{org} | |
2105 | @itemx -Ttext=@var{org} | |
2106 | Same as @option{--section-start}, with @code{.bss}, @code{.data} or | |
a6e02871 | 2107 | @code{.text} as the @var{sectionname}. |
252b5132 | 2108 | |
2509a395 SL |
2109 | @kindex -Ttext-segment=@var{org} |
2110 | @item -Ttext-segment=@var{org} | |
258795f5 | 2111 | @cindex text segment origin, cmd line |
2b8c06a3 L |
2112 | When creating an ELF executable, it will set the address of the first |
2113 | byte of the text segment. | |
258795f5 | 2114 | |
9d5777a3 RM |
2115 | @kindex -Trodata-segment=@var{org} |
2116 | @item -Trodata-segment=@var{org} | |
2117 | @cindex rodata segment origin, cmd line | |
2118 | When creating an ELF executable or shared object for a target where | |
2119 | the read-only data is in its own segment separate from the executable | |
2120 | text, it will set the address of the first byte of the read-only data segment. | |
2121 | ||
0d705e9f AM |
2122 | @kindex -Tldata-segment=@var{org} |
2123 | @item -Tldata-segment=@var{org} | |
2124 | @cindex ldata segment origin, cmd line | |
2125 | When creating an ELF executable or shared object for x86-64 medium memory | |
2126 | model, it will set the address of the first byte of the ldata segment. | |
2127 | ||
560e09e9 NC |
2128 | @kindex --unresolved-symbols |
2129 | @item --unresolved-symbols=@var{method} | |
2130 | Determine how to handle unresolved symbols. There are four possible | |
2131 | values for @samp{method}: | |
2132 | ||
2133 | @table @samp | |
2134 | @item ignore-all | |
da8bce14 | 2135 | Do not report any unresolved symbols. |
560e09e9 NC |
2136 | |
2137 | @item report-all | |
da8bce14 | 2138 | Report all unresolved symbols. This is the default. |
560e09e9 NC |
2139 | |
2140 | @item ignore-in-object-files | |
2141 | Report unresolved symbols that are contained in shared libraries, but | |
2142 | ignore them if they come from regular object files. | |
2143 | ||
2144 | @item ignore-in-shared-libs | |
2145 | Report unresolved symbols that come from regular object files, but | |
2146 | ignore them if they come from shared libraries. This can be useful | |
2147 | when creating a dynamic binary and it is known that all the shared | |
2148 | libraries that it should be referencing are included on the linker's | |
2149 | command line. | |
2150 | @end table | |
2151 | ||
2152 | The behaviour for shared libraries on their own can also be controlled | |
2153 | by the @option{--[no-]allow-shlib-undefined} option. | |
2154 | ||
2155 | Normally the linker will generate an error message for each reported | |
2156 | unresolved symbol but the option @option{--warn-unresolved-symbols} | |
2157 | can change this to a warning. | |
2158 | ||
1715a13c L |
2159 | @kindex --verbose[=@var{NUMBER}] |
2160 | @cindex verbose[=@var{NUMBER}] | |
252b5132 | 2161 | @item --dll-verbose |
1715a13c | 2162 | @itemx --verbose[=@var{NUMBER}] |
ff5dcc92 | 2163 | Display the version number for @command{ld} and list the linker emulations |
252b5132 | 2164 | supported. Display which input files can and cannot be opened. Display |
1715a13c L |
2165 | the linker script being used by the linker. If the optional @var{NUMBER} |
2166 | argument > 1, plugin symbol status will also be displayed. | |
252b5132 RH |
2167 | |
2168 | @kindex --version-script=@var{version-scriptfile} | |
2169 | @cindex version script, symbol versions | |
2509a395 | 2170 | @item --version-script=@var{version-scriptfile} |
252b5132 RH |
2171 | Specify the name of a version script to the linker. This is typically |
2172 | used when creating shared libraries to specify additional information | |
36f63dca | 2173 | about the version hierarchy for the library being created. This option |
09e2aba4 DK |
2174 | is only fully supported on ELF platforms which support shared libraries; |
2175 | see @ref{VERSION}. It is partially supported on PE platforms, which can | |
2176 | use version scripts to filter symbol visibility in auto-export mode: any | |
2177 | symbols marked @samp{local} in the version script will not be exported. | |
2178 | @xref{WIN32}. | |
252b5132 | 2179 | |
7ce691ae | 2180 | @kindex --warn-common |
252b5132 RH |
2181 | @cindex warnings, on combining symbols |
2182 | @cindex combining symbols, warnings on | |
2183 | @item --warn-common | |
2184 | Warn when a common symbol is combined with another common symbol or with | |
11e7fd74 | 2185 | a symbol definition. Unix linkers allow this somewhat sloppy practice, |
252b5132 RH |
2186 | but linkers on some other operating systems do not. This option allows |
2187 | you to find potential problems from combining global symbols. | |
11e7fd74 | 2188 | Unfortunately, some C libraries use this practice, so you may get some |
252b5132 RH |
2189 | warnings about symbols in the libraries as well as in your programs. |
2190 | ||
2191 | There are three kinds of global symbols, illustrated here by C examples: | |
2192 | ||
2193 | @table @samp | |
2194 | @item int i = 1; | |
2195 | A definition, which goes in the initialized data section of the output | |
2196 | file. | |
2197 | ||
2198 | @item extern int i; | |
2199 | An undefined reference, which does not allocate space. | |
2200 | There must be either a definition or a common symbol for the | |
2201 | variable somewhere. | |
2202 | ||
2203 | @item int i; | |
2204 | A common symbol. If there are only (one or more) common symbols for a | |
2205 | variable, it goes in the uninitialized data area of the output file. | |
2206 | The linker merges multiple common symbols for the same variable into a | |
2207 | single symbol. If they are of different sizes, it picks the largest | |
2208 | size. The linker turns a common symbol into a declaration, if there is | |
2209 | a definition of the same variable. | |
2210 | @end table | |
2211 | ||
2212 | The @samp{--warn-common} option can produce five kinds of warnings. | |
2213 | Each warning consists of a pair of lines: the first describes the symbol | |
2214 | just encountered, and the second describes the previous symbol | |
2215 | encountered with the same name. One or both of the two symbols will be | |
2216 | a common symbol. | |
2217 | ||
2218 | @enumerate | |
2219 | @item | |
2220 | Turning a common symbol into a reference, because there is already a | |
2221 | definition for the symbol. | |
2222 | @smallexample | |
2223 | @var{file}(@var{section}): warning: common of `@var{symbol}' | |
2224 | overridden by definition | |
2225 | @var{file}(@var{section}): warning: defined here | |
2226 | @end smallexample | |
2227 | ||
2228 | @item | |
2229 | Turning a common symbol into a reference, because a later definition for | |
2230 | the symbol is encountered. This is the same as the previous case, | |
2231 | except that the symbols are encountered in a different order. | |
2232 | @smallexample | |
2233 | @var{file}(@var{section}): warning: definition of `@var{symbol}' | |
2234 | overriding common | |
2235 | @var{file}(@var{section}): warning: common is here | |
2236 | @end smallexample | |
2237 | ||
2238 | @item | |
2239 | Merging a common symbol with a previous same-sized common symbol. | |
2240 | @smallexample | |
2241 | @var{file}(@var{section}): warning: multiple common | |
2242 | of `@var{symbol}' | |
2243 | @var{file}(@var{section}): warning: previous common is here | |
2244 | @end smallexample | |
2245 | ||
2246 | @item | |
2247 | Merging a common symbol with a previous larger common symbol. | |
2248 | @smallexample | |
2249 | @var{file}(@var{section}): warning: common of `@var{symbol}' | |
2250 | overridden by larger common | |
2251 | @var{file}(@var{section}): warning: larger common is here | |
2252 | @end smallexample | |
2253 | ||
2254 | @item | |
2255 | Merging a common symbol with a previous smaller common symbol. This is | |
2256 | the same as the previous case, except that the symbols are | |
2257 | encountered in a different order. | |
2258 | @smallexample | |
2259 | @var{file}(@var{section}): warning: common of `@var{symbol}' | |
2260 | overriding smaller common | |
2261 | @var{file}(@var{section}): warning: smaller common is here | |
2262 | @end smallexample | |
2263 | @end enumerate | |
2264 | ||
2265 | @kindex --warn-constructors | |
2266 | @item --warn-constructors | |
2267 | Warn if any global constructors are used. This is only useful for a few | |
2268 | object file formats. For formats like COFF or ELF, the linker can not | |
2269 | detect the use of global constructors. | |
2270 | ||
2271 | @kindex --warn-multiple-gp | |
2272 | @item --warn-multiple-gp | |
2273 | Warn if multiple global pointer values are required in the output file. | |
2274 | This is only meaningful for certain processors, such as the Alpha. | |
2275 | Specifically, some processors put large-valued constants in a special | |
2276 | section. A special register (the global pointer) points into the middle | |
2277 | of this section, so that constants can be loaded efficiently via a | |
2278 | base-register relative addressing mode. Since the offset in | |
2279 | base-register relative mode is fixed and relatively small (e.g., 16 | |
2280 | bits), this limits the maximum size of the constant pool. Thus, in | |
2281 | large programs, it is often necessary to use multiple global pointer | |
2282 | values in order to be able to address all possible constants. This | |
2283 | option causes a warning to be issued whenever this case occurs. | |
2284 | ||
2285 | @kindex --warn-once | |
2286 | @cindex warnings, on undefined symbols | |
2287 | @cindex undefined symbols, warnings on | |
2288 | @item --warn-once | |
2289 | Only warn once for each undefined symbol, rather than once per module | |
2290 | which refers to it. | |
2291 | ||
2292 | @kindex --warn-section-align | |
2293 | @cindex warnings, on section alignment | |
2294 | @cindex section alignment, warnings on | |
2295 | @item --warn-section-align | |
2296 | Warn if the address of an output section is changed because of | |
2297 | alignment. Typically, the alignment will be set by an input section. | |
2298 | The address will only be changed if it not explicitly specified; that | |
2299 | is, if the @code{SECTIONS} command does not specify a start address for | |
2300 | the section (@pxref{SECTIONS}). | |
2301 | ||
8fdd7217 NC |
2302 | @kindex --warn-shared-textrel |
2303 | @item --warn-shared-textrel | |
ece2d90e | 2304 | Warn if the linker adds a DT_TEXTREL to a shared object. |
8fdd7217 | 2305 | |
a0c402a5 L |
2306 | @kindex --warn-alternate-em |
2307 | @item --warn-alternate-em | |
2308 | Warn if an object has alternate ELF machine code. | |
2309 | ||
560e09e9 NC |
2310 | @kindex --warn-unresolved-symbols |
2311 | @item --warn-unresolved-symbols | |
2312 | If the linker is going to report an unresolved symbol (see the option | |
2313 | @option{--unresolved-symbols}) it will normally generate an error. | |
2314 | This option makes it generate a warning instead. | |
2315 | ||
2316 | @kindex --error-unresolved-symbols | |
2317 | @item --error-unresolved-symbols | |
2318 | This restores the linker's default behaviour of generating errors when | |
2319 | it is reporting unresolved symbols. | |
2320 | ||
252b5132 RH |
2321 | @kindex --whole-archive |
2322 | @cindex including an entire archive | |
2323 | @item --whole-archive | |
2324 | For each archive mentioned on the command line after the | |
ff5dcc92 | 2325 | @option{--whole-archive} option, include every object file in the archive |
252b5132 RH |
2326 | in the link, rather than searching the archive for the required object |
2327 | files. This is normally used to turn an archive file into a shared | |
2328 | library, forcing every object to be included in the resulting shared | |
2329 | library. This option may be used more than once. | |
2330 | ||
7ec229ce | 2331 | Two notes when using this option from gcc: First, gcc doesn't know |
ff5dcc92 SC |
2332 | about this option, so you have to use @option{-Wl,-whole-archive}. |
2333 | Second, don't forget to use @option{-Wl,-no-whole-archive} after your | |
7ec229ce DD |
2334 | list of archives, because gcc will add its own list of archives to |
2335 | your link and you may not want this flag to affect those as well. | |
2336 | ||
2509a395 SL |
2337 | @kindex --wrap=@var{symbol} |
2338 | @item --wrap=@var{symbol} | |
252b5132 RH |
2339 | Use a wrapper function for @var{symbol}. Any undefined reference to |
2340 | @var{symbol} will be resolved to @code{__wrap_@var{symbol}}. Any | |
2341 | undefined reference to @code{__real_@var{symbol}} will be resolved to | |
2342 | @var{symbol}. | |
2343 | ||
2344 | This can be used to provide a wrapper for a system function. The | |
2345 | wrapper function should be called @code{__wrap_@var{symbol}}. If it | |
2346 | wishes to call the system function, it should call | |
2347 | @code{__real_@var{symbol}}. | |
2348 | ||
2349 | Here is a trivial example: | |
2350 | ||
2351 | @smallexample | |
2352 | void * | |
cc2f008e | 2353 | __wrap_malloc (size_t c) |
252b5132 | 2354 | @{ |
cc2f008e | 2355 | printf ("malloc called with %zu\n", c); |
252b5132 RH |
2356 | return __real_malloc (c); |
2357 | @} | |
2358 | @end smallexample | |
2359 | ||
ff5dcc92 | 2360 | If you link other code with this file using @option{--wrap malloc}, then |
252b5132 RH |
2361 | all calls to @code{malloc} will call the function @code{__wrap_malloc} |
2362 | instead. The call to @code{__real_malloc} in @code{__wrap_malloc} will | |
2363 | call the real @code{malloc} function. | |
2364 | ||
2365 | You may wish to provide a @code{__real_malloc} function as well, so that | |
ff5dcc92 | 2366 | links without the @option{--wrap} option will succeed. If you do this, |
252b5132 RH |
2367 | you should not put the definition of @code{__real_malloc} in the same |
2368 | file as @code{__wrap_malloc}; if you do, the assembler may resolve the | |
2369 | call before the linker has a chance to wrap it to @code{malloc}. | |
2370 | ||
6aa29e7b | 2371 | @kindex --eh-frame-hdr |
29063f8b | 2372 | @kindex --no-eh-frame-hdr |
6aa29e7b | 2373 | @item --eh-frame-hdr |
29063f8b NC |
2374 | @itemx --no-eh-frame-hdr |
2375 | Request (@option{--eh-frame-hdr}) or suppress | |
2376 | (@option{--no-eh-frame-hdr}) the creation of @code{.eh_frame_hdr} | |
2377 | section and ELF @code{PT_GNU_EH_FRAME} segment header. | |
6aa29e7b | 2378 | |
e41b3a13 JJ |
2379 | @kindex --ld-generated-unwind-info |
2380 | @item --no-ld-generated-unwind-info | |
2381 | Request creation of @code{.eh_frame} unwind info for linker | |
2382 | generated code sections like PLT. This option is on by default | |
2383 | if linker generated unwind info is supported. | |
2384 | ||
6c1439be L |
2385 | @kindex --enable-new-dtags |
2386 | @kindex --disable-new-dtags | |
2387 | @item --enable-new-dtags | |
2388 | @itemx --disable-new-dtags | |
2389 | This linker can create the new dynamic tags in ELF. But the older ELF | |
2390 | systems may not understand them. If you specify | |
b1b00fcc MF |
2391 | @option{--enable-new-dtags}, the new dynamic tags will be created as needed |
2392 | and older dynamic tags will be omitted. | |
ff5dcc92 | 2393 | If you specify @option{--disable-new-dtags}, no new dynamic tags will be |
6c1439be L |
2394 | created. By default, the new dynamic tags are not created. Note that |
2395 | those options are only available for ELF systems. | |
2396 | ||
2d643429 | 2397 | @kindex --hash-size=@var{number} |
e185dd51 | 2398 | @item --hash-size=@var{number} |
2d643429 NC |
2399 | Set the default size of the linker's hash tables to a prime number |
2400 | close to @var{number}. Increasing this value can reduce the length of | |
2401 | time it takes the linker to perform its tasks, at the expense of | |
2402 | increasing the linker's memory requirements. Similarly reducing this | |
2403 | value can reduce the memory requirements at the expense of speed. | |
2404 | ||
fdc90cb4 JJ |
2405 | @kindex --hash-style=@var{style} |
2406 | @item --hash-style=@var{style} | |
2407 | Set the type of linker's hash table(s). @var{style} can be either | |
2408 | @code{sysv} for classic ELF @code{.hash} section, @code{gnu} for | |
2409 | new style GNU @code{.gnu.hash} section or @code{both} for both | |
2410 | the classic ELF @code{.hash} and new style GNU @code{.gnu.hash} | |
c8455dc9 NC |
2411 | hash tables. The default depends upon how the linker was configured, |
2412 | but for most Linux based systems it will be @code{both}. | |
fdc90cb4 | 2413 | |
0ce398f1 L |
2414 | @kindex --compress-debug-sections=none |
2415 | @kindex --compress-debug-sections=zlib | |
2416 | @kindex --compress-debug-sections=zlib-gnu | |
2417 | @kindex --compress-debug-sections=zlib-gabi | |
2418 | @item --compress-debug-sections=none | |
2419 | @itemx --compress-debug-sections=zlib | |
2420 | @itemx --compress-debug-sections=zlib-gnu | |
2421 | @itemx --compress-debug-sections=zlib-gabi | |
9af89fba NC |
2422 | On ELF platforms, these options control how DWARF debug sections are |
2423 | compressed using zlib. | |
2424 | ||
2425 | @option{--compress-debug-sections=none} doesn't compress DWARF debug | |
2426 | sections. @option{--compress-debug-sections=zlib-gnu} compresses | |
2427 | DWARF debug sections and renames them to begin with @samp{.zdebug} | |
2428 | instead of @samp{.debug}. @option{--compress-debug-sections=zlib-gabi} | |
2429 | also compresses DWARF debug sections, but rather than renaming them it | |
2430 | sets the SHF_COMPRESSED flag in the sections' headers. | |
2431 | ||
2432 | The @option{--compress-debug-sections=zlib} option is an alias for | |
2433 | @option{--compress-debug-sections=zlib-gabi}. | |
2434 | ||
2435 | Note that this option overrides any compression in input debug | |
2436 | sections, so if a binary is linked with @option{--compress-debug-sections=none} | |
2437 | for example, then any compressed debug sections in input files will be | |
2438 | uncompressed before they are copied into the output binary. | |
2439 | ||
2440 | The default compression behaviour varies depending upon the target | |
2441 | involved and the configure options used to build the toolchain. The | |
2442 | default can be determined by examining the output from the linker's | |
2443 | @option{--help} option. | |
0ce398f1 | 2444 | |
35835446 JR |
2445 | @kindex --reduce-memory-overheads |
2446 | @item --reduce-memory-overheads | |
2447 | This option reduces memory requirements at ld runtime, at the expense of | |
f2a8f148 | 2448 | linking speed. This was introduced to select the old O(n^2) algorithm |
35835446 | 2449 | for link map file generation, rather than the new O(n) algorithm which uses |
2d643429 NC |
2450 | about 40% more memory for symbol storage. |
2451 | ||
4f9c04f7 | 2452 | Another effect of the switch is to set the default hash table size to |
2d643429 | 2453 | 1021, which again saves memory at the cost of lengthening the linker's |
a85785bc | 2454 | run time. This is not done however if the @option{--hash-size} switch |
2d643429 NC |
2455 | has been used. |
2456 | ||
2457 | The @option{--reduce-memory-overheads} switch may be also be used to | |
2458 | enable other tradeoffs in future versions of the linker. | |
35835446 | 2459 | |
c0065db7 RM |
2460 | @kindex --build-id |
2461 | @kindex --build-id=@var{style} | |
2462 | @item --build-id | |
2463 | @itemx --build-id=@var{style} | |
61e2488c | 2464 | Request the creation of a @code{.note.gnu.build-id} ELF note section |
6033bf41 | 2465 | or a @code{.buildid} COFF section. The contents of the note are |
61e2488c JT |
2466 | unique bits identifying this linked file. @var{style} can be |
2467 | @code{uuid} to use 128 random bits, @code{sha1} to use a 160-bit | |
2468 | @sc{SHA1} hash on the normative parts of the output contents, | |
2469 | @code{md5} to use a 128-bit @sc{MD5} hash on the normative parts of | |
2470 | the output contents, or @code{0x@var{hexstring}} to use a chosen bit | |
2471 | string specified as an even number of hexadecimal digits (@code{-} and | |
2472 | @code{:} characters between digit pairs are ignored). If @var{style} | |
2473 | is omitted, @code{sha1} is used. | |
24382dca RM |
2474 | |
2475 | The @code{md5} and @code{sha1} styles produces an identifier | |
2476 | that is always the same in an identical output file, but will be | |
2477 | unique among all nonidentical output files. It is not intended | |
2478 | to be compared as a checksum for the file's contents. A linked | |
2479 | file may be changed later by other tools, but the build ID bit | |
2480 | string identifying the original linked file does not change. | |
c0065db7 RM |
2481 | |
2482 | Passing @code{none} for @var{style} disables the setting from any | |
2483 | @code{--build-id} options earlier on the command line. | |
252b5132 RH |
2484 | @end table |
2485 | ||
0285c67d NC |
2486 | @c man end |
2487 | ||
36f63dca | 2488 | @subsection Options Specific to i386 PE Targets |
252b5132 | 2489 | |
0285c67d NC |
2490 | @c man begin OPTIONS |
2491 | ||
ff5dcc92 | 2492 | The i386 PE linker supports the @option{-shared} option, which causes |
252b5132 RH |
2493 | the output to be a dynamically linked library (DLL) instead of a |
2494 | normal executable. You should name the output @code{*.dll} when you | |
2495 | use this option. In addition, the linker fully supports the standard | |
2496 | @code{*.def} files, which may be specified on the linker command line | |
2497 | like an object file (in fact, it should precede archives it exports | |
2498 | symbols from, to ensure that they get linked in, just like a normal | |
2499 | object file). | |
2500 | ||
2501 | In addition to the options common to all targets, the i386 PE linker | |
a05a5b64 | 2502 | support additional command-line options that are specific to the i386 |
252b5132 RH |
2503 | PE target. Options that take values may be separated from their |
2504 | values by either a space or an equals sign. | |
2505 | ||
ff5dcc92 | 2506 | @table @gcctabopt |
252b5132 RH |
2507 | |
2508 | @kindex --add-stdcall-alias | |
2509 | @item --add-stdcall-alias | |
2510 | If given, symbols with a stdcall suffix (@@@var{nn}) will be exported | |
2511 | as-is and also with the suffix stripped. | |
bb10df36 | 2512 | [This option is specific to the i386 PE targeted port of the linker] |
252b5132 RH |
2513 | |
2514 | @kindex --base-file | |
2515 | @item --base-file @var{file} | |
2516 | Use @var{file} as the name of a file in which to save the base | |
2517 | addresses of all the relocations needed for generating DLLs with | |
2518 | @file{dlltool}. | |
bb10df36 | 2519 | [This is an i386 PE specific option] |
252b5132 RH |
2520 | |
2521 | @kindex --dll | |
2522 | @item --dll | |
2523 | Create a DLL instead of a regular executable. You may also use | |
ff5dcc92 | 2524 | @option{-shared} or specify a @code{LIBRARY} in a given @code{.def} |
252b5132 | 2525 | file. |
bb10df36 | 2526 | [This option is specific to the i386 PE targeted port of the linker] |
252b5132 | 2527 | |
88183869 DK |
2528 | @kindex --enable-long-section-names |
2529 | @kindex --disable-long-section-names | |
2530 | @item --enable-long-section-names | |
2531 | @itemx --disable-long-section-names | |
56e6cf80 | 2532 | The PE variants of the COFF object format add an extension that permits |
88183869 | 2533 | the use of section names longer than eight characters, the normal limit |
56e6cf80 NC |
2534 | for COFF. By default, these names are only allowed in object files, as |
2535 | fully-linked executable images do not carry the COFF string table required | |
88183869 DK |
2536 | to support the longer names. As a GNU extension, it is possible to |
2537 | allow their use in executable images as well, or to (probably pointlessly!) | |
2538 | disallow it in object files, by using these two options. Executable images | |
2539 | generated with these long section names are slightly non-standard, carrying | |
2540 | as they do a string table, and may generate confusing output when examined | |
9d5777a3 RM |
2541 | with non-GNU PE-aware tools, such as file viewers and dumpers. However, |
2542 | GDB relies on the use of PE long section names to find Dwarf-2 debug | |
3efd345c DK |
2543 | information sections in an executable image at runtime, and so if neither |
2544 | option is specified on the command-line, @command{ld} will enable long | |
2545 | section names, overriding the default and technically correct behaviour, | |
2546 | when it finds the presence of debug information while linking an executable | |
2547 | image and not stripping symbols. | |
88183869 DK |
2548 | [This option is valid for all PE targeted ports of the linker] |
2549 | ||
252b5132 RH |
2550 | @kindex --enable-stdcall-fixup |
2551 | @kindex --disable-stdcall-fixup | |
2552 | @item --enable-stdcall-fixup | |
2553 | @itemx --disable-stdcall-fixup | |
2554 | If the link finds a symbol that it cannot resolve, it will attempt to | |
36f63dca | 2555 | do ``fuzzy linking'' by looking for another defined symbol that differs |
252b5132 RH |
2556 | only in the format of the symbol name (cdecl vs stdcall) and will |
2557 | resolve that symbol by linking to the match. For example, the | |
2558 | undefined symbol @code{_foo} might be linked to the function | |
2559 | @code{_foo@@12}, or the undefined symbol @code{_bar@@16} might be linked | |
2560 | to the function @code{_bar}. When the linker does this, it prints a | |
2561 | warning, since it normally should have failed to link, but sometimes | |
2562 | import libraries generated from third-party dlls may need this feature | |
ff5dcc92 | 2563 | to be usable. If you specify @option{--enable-stdcall-fixup}, this |
252b5132 | 2564 | feature is fully enabled and warnings are not printed. If you specify |
ff5dcc92 | 2565 | @option{--disable-stdcall-fixup}, this feature is disabled and such |
252b5132 | 2566 | mismatches are considered to be errors. |
bb10df36 | 2567 | [This option is specific to the i386 PE targeted port of the linker] |
252b5132 | 2568 | |
522f09cd KT |
2569 | @kindex --leading-underscore |
2570 | @kindex --no-leading-underscore | |
2571 | @item --leading-underscore | |
2572 | @itemx --no-leading-underscore | |
2573 | For most targets default symbol-prefix is an underscore and is defined | |
2574 | in target's description. By this option it is possible to | |
2575 | disable/enable the default underscore symbol-prefix. | |
2576 | ||
252b5132 RH |
2577 | @cindex DLLs, creating |
2578 | @kindex --export-all-symbols | |
2579 | @item --export-all-symbols | |
2580 | If given, all global symbols in the objects used to build a DLL will | |
2581 | be exported by the DLL. Note that this is the default if there | |
2582 | otherwise wouldn't be any exported symbols. When symbols are | |
2583 | explicitly exported via DEF files or implicitly exported via function | |
2584 | attributes, the default is to not export anything else unless this | |
2585 | option is given. Note that the symbols @code{DllMain@@12}, | |
ece2d90e | 2586 | @code{DllEntryPoint@@0}, @code{DllMainCRTStartup@@12}, and |
b044cda1 | 2587 | @code{impure_ptr} will not be automatically |
ece2d90e NC |
2588 | exported. Also, symbols imported from other DLLs will not be |
2589 | re-exported, nor will symbols specifying the DLL's internal layout | |
2590 | such as those beginning with @code{_head_} or ending with | |
2591 | @code{_iname}. In addition, no symbols from @code{libgcc}, | |
b044cda1 CW |
2592 | @code{libstd++}, @code{libmingw32}, or @code{crtX.o} will be exported. |
2593 | Symbols whose names begin with @code{__rtti_} or @code{__builtin_} will | |
2594 | not be exported, to help with C++ DLLs. Finally, there is an | |
ece2d90e | 2595 | extensive list of cygwin-private symbols that are not exported |
b044cda1 | 2596 | (obviously, this applies on when building DLLs for cygwin targets). |
ece2d90e | 2597 | These cygwin-excludes are: @code{_cygwin_dll_entry@@12}, |
b044cda1 | 2598 | @code{_cygwin_crt0_common@@8}, @code{_cygwin_noncygwin_dll_entry@@12}, |
ece2d90e | 2599 | @code{_fmode}, @code{_impure_ptr}, @code{cygwin_attach_dll}, |
b044cda1 | 2600 | @code{cygwin_premain0}, @code{cygwin_premain1}, @code{cygwin_premain2}, |
ece2d90e | 2601 | @code{cygwin_premain3}, and @code{environ}. |
bb10df36 | 2602 | [This option is specific to the i386 PE targeted port of the linker] |
252b5132 RH |
2603 | |
2604 | @kindex --exclude-symbols | |
1d0a3c9c | 2605 | @item --exclude-symbols @var{symbol},@var{symbol},... |
252b5132 RH |
2606 | Specifies a list of symbols which should not be automatically |
2607 | exported. The symbol names may be delimited by commas or colons. | |
bb10df36 | 2608 | [This option is specific to the i386 PE targeted port of the linker] |
252b5132 | 2609 | |
2927aaca NC |
2610 | @kindex --exclude-all-symbols |
2611 | @item --exclude-all-symbols | |
2612 | Specifies no symbols should be automatically exported. | |
2613 | [This option is specific to the i386 PE targeted port of the linker] | |
2614 | ||
252b5132 RH |
2615 | @kindex --file-alignment |
2616 | @item --file-alignment | |
2617 | Specify the file alignment. Sections in the file will always begin at | |
2618 | file offsets which are multiples of this number. This defaults to | |
2619 | 512. | |
bb10df36 | 2620 | [This option is specific to the i386 PE targeted port of the linker] |
252b5132 RH |
2621 | |
2622 | @cindex heap size | |
2623 | @kindex --heap | |
2624 | @item --heap @var{reserve} | |
2625 | @itemx --heap @var{reserve},@var{commit} | |
a00b50c5 | 2626 | Specify the number of bytes of memory to reserve (and optionally commit) |
fe6d7d6a | 2627 | to be used as heap for this program. The default is 1MB reserved, 4K |
252b5132 | 2628 | committed. |
bb10df36 | 2629 | [This option is specific to the i386 PE targeted port of the linker] |
252b5132 RH |
2630 | |
2631 | @cindex image base | |
2632 | @kindex --image-base | |
2633 | @item --image-base @var{value} | |
2634 | Use @var{value} as the base address of your program or dll. This is | |
2635 | the lowest memory location that will be used when your program or dll | |
2636 | is loaded. To reduce the need to relocate and improve performance of | |
2637 | your dlls, each should have a unique base address and not overlap any | |
2638 | other dlls. The default is 0x400000 for executables, and 0x10000000 | |
2639 | for dlls. | |
bb10df36 | 2640 | [This option is specific to the i386 PE targeted port of the linker] |
252b5132 RH |
2641 | |
2642 | @kindex --kill-at | |
2643 | @item --kill-at | |
2644 | If given, the stdcall suffixes (@@@var{nn}) will be stripped from | |
2645 | symbols before they are exported. | |
bb10df36 | 2646 | [This option is specific to the i386 PE targeted port of the linker] |
252b5132 | 2647 | |
26d2d8a2 BF |
2648 | @kindex --large-address-aware |
2649 | @item --large-address-aware | |
b45619c0 | 2650 | If given, the appropriate bit in the ``Characteristics'' field of the COFF |
26d2d8a2 | 2651 | header is set to indicate that this executable supports virtual addresses |
b45619c0 | 2652 | greater than 2 gigabytes. This should be used in conjunction with the /3GB |
26d2d8a2 BF |
2653 | or /USERVA=@var{value} megabytes switch in the ``[operating systems]'' |
2654 | section of the BOOT.INI. Otherwise, this bit has no effect. | |
2655 | [This option is specific to PE targeted ports of the linker] | |
2656 | ||
f69a2f97 NC |
2657 | @kindex --disable-large-address-aware |
2658 | @item --disable-large-address-aware | |
2659 | Reverts the effect of a previous @samp{--large-address-aware} option. | |
2660 | This is useful if @samp{--large-address-aware} is always set by the compiler | |
2661 | driver (e.g. Cygwin gcc) and the executable does not support virtual | |
2662 | addresses greater than 2 gigabytes. | |
2663 | [This option is specific to PE targeted ports of the linker] | |
2664 | ||
252b5132 RH |
2665 | @kindex --major-image-version |
2666 | @item --major-image-version @var{value} | |
36f63dca | 2667 | Sets the major number of the ``image version''. Defaults to 1. |
bb10df36 | 2668 | [This option is specific to the i386 PE targeted port of the linker] |
252b5132 RH |
2669 | |
2670 | @kindex --major-os-version | |
2671 | @item --major-os-version @var{value} | |
36f63dca | 2672 | Sets the major number of the ``os version''. Defaults to 4. |
bb10df36 | 2673 | [This option is specific to the i386 PE targeted port of the linker] |
252b5132 RH |
2674 | |
2675 | @kindex --major-subsystem-version | |
2676 | @item --major-subsystem-version @var{value} | |
36f63dca | 2677 | Sets the major number of the ``subsystem version''. Defaults to 4. |
bb10df36 | 2678 | [This option is specific to the i386 PE targeted port of the linker] |
252b5132 RH |
2679 | |
2680 | @kindex --minor-image-version | |
2681 | @item --minor-image-version @var{value} | |
36f63dca | 2682 | Sets the minor number of the ``image version''. Defaults to 0. |
bb10df36 | 2683 | [This option is specific to the i386 PE targeted port of the linker] |
252b5132 RH |
2684 | |
2685 | @kindex --minor-os-version | |
2686 | @item --minor-os-version @var{value} | |
36f63dca | 2687 | Sets the minor number of the ``os version''. Defaults to 0. |
bb10df36 | 2688 | [This option is specific to the i386 PE targeted port of the linker] |
252b5132 RH |
2689 | |
2690 | @kindex --minor-subsystem-version | |
2691 | @item --minor-subsystem-version @var{value} | |
36f63dca | 2692 | Sets the minor number of the ``subsystem version''. Defaults to 0. |
bb10df36 | 2693 | [This option is specific to the i386 PE targeted port of the linker] |
252b5132 RH |
2694 | |
2695 | @cindex DEF files, creating | |
2696 | @cindex DLLs, creating | |
2697 | @kindex --output-def | |
2698 | @item --output-def @var{file} | |
2699 | The linker will create the file @var{file} which will contain a DEF | |
2700 | file corresponding to the DLL the linker is generating. This DEF file | |
2701 | (which should be called @code{*.def}) may be used to create an import | |
2702 | library with @code{dlltool} or may be used as a reference to | |
2703 | automatically or implicitly exported symbols. | |
bb10df36 | 2704 | [This option is specific to the i386 PE targeted port of the linker] |
252b5132 | 2705 | |
b044cda1 | 2706 | @cindex DLLs, creating |
b044cda1 CW |
2707 | @kindex --enable-auto-image-base |
2708 | @item --enable-auto-image-base | |
d0e6d77b CF |
2709 | @itemx --enable-auto-image-base=@var{value} |
2710 | Automatically choose the image base for DLLs, optionally starting with base | |
2711 | @var{value}, unless one is specified using the @code{--image-base} argument. | |
2712 | By using a hash generated from the dllname to create unique image bases | |
2713 | for each DLL, in-memory collisions and relocations which can delay program | |
2714 | execution are avoided. | |
bb10df36 | 2715 | [This option is specific to the i386 PE targeted port of the linker] |
b044cda1 CW |
2716 | |
2717 | @kindex --disable-auto-image-base | |
2718 | @item --disable-auto-image-base | |
2719 | Do not automatically generate a unique image base. If there is no | |
2720 | user-specified image base (@code{--image-base}) then use the platform | |
2721 | default. | |
bb10df36 | 2722 | [This option is specific to the i386 PE targeted port of the linker] |
b044cda1 CW |
2723 | |
2724 | @cindex DLLs, linking to | |
2725 | @kindex --dll-search-prefix | |
2726 | @item --dll-search-prefix @var{string} | |
489d0400 | 2727 | When linking dynamically to a dll without an import library, |
ece2d90e | 2728 | search for @code{<string><basename>.dll} in preference to |
560e09e9 | 2729 | @code{lib<basename>.dll}. This behaviour allows easy distinction |
b044cda1 CW |
2730 | between DLLs built for the various "subplatforms": native, cygwin, |
2731 | uwin, pw, etc. For instance, cygwin DLLs typically use | |
ece2d90e | 2732 | @code{--dll-search-prefix=cyg}. |
bb10df36 | 2733 | [This option is specific to the i386 PE targeted port of the linker] |
b044cda1 CW |
2734 | |
2735 | @kindex --enable-auto-import | |
2736 | @item --enable-auto-import | |
ece2d90e | 2737 | Do sophisticated linking of @code{_symbol} to @code{__imp__symbol} for |
317ff008 EB |
2738 | DATA imports from DLLs, thus making it possible to bypass the dllimport |
2739 | mechanism on the user side and to reference unmangled symbol names. | |
2740 | [This option is specific to the i386 PE targeted port of the linker] | |
2741 | ||
2742 | The following remarks pertain to the original implementation of the | |
2743 | feature and are obsolete nowadays for Cygwin and MinGW targets. | |
2744 | ||
2745 | Note: Use of the 'auto-import' extension will cause the text section | |
2746 | of the image file to be made writable. This does not conform to the | |
2747 | PE-COFF format specification published by Microsoft. | |
4d8907ac | 2748 | |
e2a83dd0 NC |
2749 | Note - use of the 'auto-import' extension will also cause read only |
2750 | data which would normally be placed into the .rdata section to be | |
2751 | placed into the .data section instead. This is in order to work | |
2752 | around a problem with consts that is described here: | |
2753 | http://www.cygwin.com/ml/cygwin/2004-09/msg01101.html | |
2754 | ||
4d8907ac DS |
2755 | Using 'auto-import' generally will 'just work' -- but sometimes you may |
2756 | see this message: | |
0d888aac | 2757 | |
ece2d90e | 2758 | "variable '<var>' can't be auto-imported. Please read the |
0d888aac CW |
2759 | documentation for ld's @code{--enable-auto-import} for details." |
2760 | ||
ece2d90e NC |
2761 | This message occurs when some (sub)expression accesses an address |
2762 | ultimately given by the sum of two constants (Win32 import tables only | |
c0065db7 RM |
2763 | allow one). Instances where this may occur include accesses to member |
2764 | fields of struct variables imported from a DLL, as well as using a | |
2765 | constant index into an array variable imported from a DLL. Any | |
2f8d8971 NC |
2766 | multiword variable (arrays, structs, long long, etc) may trigger |
2767 | this error condition. However, regardless of the exact data type | |
2768 | of the offending exported variable, ld will always detect it, issue | |
2769 | the warning, and exit. | |
2770 | ||
2771 | There are several ways to address this difficulty, regardless of the | |
2772 | data type of the exported variable: | |
0d888aac | 2773 | |
2fa9fc65 NC |
2774 | One way is to use --enable-runtime-pseudo-reloc switch. This leaves the task |
2775 | of adjusting references in your client code for runtime environment, so | |
560e09e9 | 2776 | this method works only when runtime environment supports this feature. |
2fa9fc65 | 2777 | |
c0065db7 RM |
2778 | A second solution is to force one of the 'constants' to be a variable -- |
2779 | that is, unknown and un-optimizable at compile time. For arrays, | |
2780 | there are two possibilities: a) make the indexee (the array's address) | |
0d888aac CW |
2781 | a variable, or b) make the 'constant' index a variable. Thus: |
2782 | ||
2783 | @example | |
2784 | extern type extern_array[]; | |
c0065db7 | 2785 | extern_array[1] --> |
0d888aac CW |
2786 | @{ volatile type *t=extern_array; t[1] @} |
2787 | @end example | |
2788 | ||
2789 | or | |
2790 | ||
2791 | @example | |
2792 | extern type extern_array[]; | |
c0065db7 | 2793 | extern_array[1] --> |
0d888aac CW |
2794 | @{ volatile int t=1; extern_array[t] @} |
2795 | @end example | |
2796 | ||
c0065db7 | 2797 | For structs (and most other multiword data types) the only option |
2f8d8971 | 2798 | is to make the struct itself (or the long long, or the ...) variable: |
0d888aac CW |
2799 | |
2800 | @example | |
2801 | extern struct s extern_struct; | |
c0065db7 | 2802 | extern_struct.field --> |
0d888aac CW |
2803 | @{ volatile struct s *t=&extern_struct; t->field @} |
2804 | @end example | |
2805 | ||
c406afaf NC |
2806 | or |
2807 | ||
2808 | @example | |
2809 | extern long long extern_ll; | |
2810 | extern_ll --> | |
2811 | @{ volatile long long * local_ll=&extern_ll; *local_ll @} | |
2812 | @end example | |
2813 | ||
2fa9fc65 | 2814 | A third method of dealing with this difficulty is to abandon |
c0065db7 | 2815 | 'auto-import' for the offending symbol and mark it with |
11e7fd74 | 2816 | @code{__declspec(dllimport)}. However, in practice that |
0d888aac | 2817 | requires using compile-time #defines to indicate whether you are |
c0065db7 RM |
2818 | building a DLL, building client code that will link to the DLL, or |
2819 | merely building/linking to a static library. In making the choice | |
2820 | between the various methods of resolving the 'direct address with | |
0d888aac CW |
2821 | constant offset' problem, you should consider typical real-world usage: |
2822 | ||
2823 | Original: | |
2824 | @example | |
2825 | --foo.h | |
2826 | extern int arr[]; | |
2827 | --foo.c | |
2828 | #include "foo.h" | |
2829 | void main(int argc, char **argv)@{ | |
2830 | printf("%d\n",arr[1]); | |
2831 | @} | |
2832 | @end example | |
2833 | ||
2834 | Solution 1: | |
2835 | @example | |
2836 | --foo.h | |
2837 | extern int arr[]; | |
2838 | --foo.c | |
2839 | #include "foo.h" | |
2840 | void main(int argc, char **argv)@{ | |
2841 | /* This workaround is for win32 and cygwin; do not "optimize" */ | |
2842 | volatile int *parr = arr; | |
2843 | printf("%d\n",parr[1]); | |
2844 | @} | |
2845 | @end example | |
2846 | ||
2847 | Solution 2: | |
2848 | @example | |
2849 | --foo.h | |
2850 | /* Note: auto-export is assumed (no __declspec(dllexport)) */ | |
2851 | #if (defined(_WIN32) || defined(__CYGWIN__)) && \ | |
2852 | !(defined(FOO_BUILD_DLL) || defined(FOO_STATIC)) | |
2853 | #define FOO_IMPORT __declspec(dllimport) | |
2854 | #else | |
2855 | #define FOO_IMPORT | |
2856 | #endif | |
2857 | extern FOO_IMPORT int arr[]; | |
2858 | --foo.c | |
2859 | #include "foo.h" | |
2860 | void main(int argc, char **argv)@{ | |
2861 | printf("%d\n",arr[1]); | |
2862 | @} | |
2863 | @end example | |
2864 | ||
c0065db7 | 2865 | A fourth way to avoid this problem is to re-code your |
0d888aac CW |
2866 | library to use a functional interface rather than a data interface |
2867 | for the offending variables (e.g. set_foo() and get_foo() accessor | |
2868 | functions). | |
b044cda1 CW |
2869 | |
2870 | @kindex --disable-auto-import | |
2871 | @item --disable-auto-import | |
c0065db7 | 2872 | Do not attempt to do sophisticated linking of @code{_symbol} to |
b044cda1 | 2873 | @code{__imp__symbol} for DATA imports from DLLs. |
bb10df36 | 2874 | [This option is specific to the i386 PE targeted port of the linker] |
b044cda1 | 2875 | |
2fa9fc65 NC |
2876 | @kindex --enable-runtime-pseudo-reloc |
2877 | @item --enable-runtime-pseudo-reloc | |
2878 | If your code contains expressions described in --enable-auto-import section, | |
2879 | that is, DATA imports from DLL with non-zero offset, this switch will create | |
2880 | a vector of 'runtime pseudo relocations' which can be used by runtime | |
c0065db7 | 2881 | environment to adjust references to such data in your client code. |
bb10df36 | 2882 | [This option is specific to the i386 PE targeted port of the linker] |
2fa9fc65 NC |
2883 | |
2884 | @kindex --disable-runtime-pseudo-reloc | |
2885 | @item --disable-runtime-pseudo-reloc | |
317ff008 | 2886 | Do not create pseudo relocations for non-zero offset DATA imports from DLLs. |
bb10df36 | 2887 | [This option is specific to the i386 PE targeted port of the linker] |
2fa9fc65 | 2888 | |
b044cda1 CW |
2889 | @kindex --enable-extra-pe-debug |
2890 | @item --enable-extra-pe-debug | |
2891 | Show additional debug info related to auto-import symbol thunking. | |
bb10df36 | 2892 | [This option is specific to the i386 PE targeted port of the linker] |
b044cda1 | 2893 | |
252b5132 RH |
2894 | @kindex --section-alignment |
2895 | @item --section-alignment | |
2896 | Sets the section alignment. Sections in memory will always begin at | |
2897 | addresses which are a multiple of this number. Defaults to 0x1000. | |
bb10df36 | 2898 | [This option is specific to the i386 PE targeted port of the linker] |
252b5132 RH |
2899 | |
2900 | @cindex stack size | |
2901 | @kindex --stack | |
2902 | @item --stack @var{reserve} | |
2903 | @itemx --stack @var{reserve},@var{commit} | |
a00b50c5 | 2904 | Specify the number of bytes of memory to reserve (and optionally commit) |
fe6d7d6a | 2905 | to be used as stack for this program. The default is 2MB reserved, 4K |
252b5132 | 2906 | committed. |
bb10df36 | 2907 | [This option is specific to the i386 PE targeted port of the linker] |
252b5132 RH |
2908 | |
2909 | @kindex --subsystem | |
2910 | @item --subsystem @var{which} | |
2911 | @itemx --subsystem @var{which}:@var{major} | |
2912 | @itemx --subsystem @var{which}:@var{major}.@var{minor} | |
2913 | Specifies the subsystem under which your program will execute. The | |
2914 | legal values for @var{which} are @code{native}, @code{windows}, | |
33f362e1 NC |
2915 | @code{console}, @code{posix}, and @code{xbox}. You may optionally set |
2916 | the subsystem version also. Numeric values are also accepted for | |
2917 | @var{which}. | |
bb10df36 | 2918 | [This option is specific to the i386 PE targeted port of the linker] |
252b5132 | 2919 | |
2f563b51 DK |
2920 | The following options set flags in the @code{DllCharacteristics} field |
2921 | of the PE file header: | |
2922 | [These options are specific to PE targeted ports of the linker] | |
2923 | ||
2d5c3743 NC |
2924 | @kindex --high-entropy-va |
2925 | @item --high-entropy-va | |
2926 | Image is compatible with 64-bit address space layout randomization | |
2927 | (ASLR). | |
2928 | ||
2f563b51 DK |
2929 | @kindex --dynamicbase |
2930 | @item --dynamicbase | |
2931 | The image base address may be relocated using address space layout | |
2932 | randomization (ASLR). This feature was introduced with MS Windows | |
2933 | Vista for i386 PE targets. | |
2934 | ||
2935 | @kindex --forceinteg | |
2936 | @item --forceinteg | |
2937 | Code integrity checks are enforced. | |
2938 | ||
2939 | @kindex --nxcompat | |
2940 | @item --nxcompat | |
2941 | The image is compatible with the Data Execution Prevention. | |
2942 | This feature was introduced with MS Windows XP SP2 for i386 PE targets. | |
2943 | ||
2944 | @kindex --no-isolation | |
2945 | @item --no-isolation | |
2946 | Although the image understands isolation, do not isolate the image. | |
2947 | ||
2948 | @kindex --no-seh | |
2949 | @item --no-seh | |
2950 | The image does not use SEH. No SE handler may be called from | |
2951 | this image. | |
2952 | ||
2953 | @kindex --no-bind | |
2954 | @item --no-bind | |
2955 | Do not bind this image. | |
2956 | ||
2957 | @kindex --wdmdriver | |
2958 | @item --wdmdriver | |
2959 | The driver uses the MS Windows Driver Model. | |
9d5777a3 | 2960 | |
2f563b51 DK |
2961 | @kindex --tsaware |
2962 | @item --tsaware | |
2963 | The image is Terminal Server aware. | |
2964 | ||
0cb112f7 CF |
2965 | @kindex --insert-timestamp |
2966 | @item --insert-timestamp | |
eeb14e5a LZ |
2967 | @itemx --no-insert-timestamp |
2968 | Insert a real timestamp into the image. This is the default behaviour | |
2969 | as it matches legacy code and it means that the image will work with | |
2970 | other, proprietary tools. The problem with this default is that it | |
56e6cf80 | 2971 | will result in slightly different images being produced each time the |
eeb14e5a LZ |
2972 | same sources are linked. The option @option{--no-insert-timestamp} |
2973 | can be used to insert a zero value for the timestamp, this ensuring | |
56e6cf80 | 2974 | that binaries produced from identical sources will compare |
eeb14e5a | 2975 | identically. |
252b5132 RH |
2976 | @end table |
2977 | ||
0285c67d NC |
2978 | @c man end |
2979 | ||
ac145307 BS |
2980 | @ifset C6X |
2981 | @subsection Options specific to C6X uClinux targets | |
2982 | ||
2983 | @c man begin OPTIONS | |
2984 | ||
2985 | The C6X uClinux target uses a binary format called DSBT to support shared | |
2986 | libraries. Each shared library in the system needs to have a unique index; | |
2987 | all executables use an index of 0. | |
2988 | ||
2989 | @table @gcctabopt | |
2990 | ||
2991 | @kindex --dsbt-size | |
2992 | @item --dsbt-size @var{size} | |
56e6cf80 | 2993 | This option sets the number of entries in the DSBT of the current executable |
ac145307 BS |
2994 | or shared library to @var{size}. The default is to create a table with 64 |
2995 | entries. | |
2996 | ||
2997 | @kindex --dsbt-index | |
2998 | @item --dsbt-index @var{index} | |
2999 | This option sets the DSBT index of the current executable or shared library | |
3000 | to @var{index}. The default is 0, which is appropriate for generating | |
3001 | executables. If a shared library is generated with a DSBT index of 0, the | |
3002 | @code{R_C6000_DSBT_INDEX} relocs are copied into the output file. | |
3003 | ||
fbd9ad90 PB |
3004 | @kindex --no-merge-exidx-entries |
3005 | The @samp{--no-merge-exidx-entries} switch disables the merging of adjacent | |
3006 | exidx entries in frame unwind info. | |
3007 | ||
ac145307 BS |
3008 | @end table |
3009 | ||
3010 | @c man end | |
3011 | @end ifset | |
3012 | ||
b8891f8d AJ |
3013 | @ifset CSKY |
3014 | @subsection Options specific to C-SKY targets | |
3015 | ||
3016 | @c man begin OPTIONS | |
3017 | ||
3018 | @table @gcctabopt | |
3019 | ||
3020 | @kindex --branch-stub on C-SKY | |
3021 | @item --branch-stub | |
3022 | This option enables linker branch relaxation by inserting branch stub | |
3023 | sections when needed to extend the range of branches. This option is | |
3024 | usually not required since C-SKY supports branch and call instructions that | |
3025 | can access the full memory range and branch relaxation is normally handled by | |
3026 | the compiler or assembler. | |
3027 | ||
3028 | @kindex --stub-group-size on C-SKY | |
3029 | @item --stub-group-size=@var{N} | |
3030 | This option allows finer control of linker branch stub creation. | |
3031 | It sets the maximum size of a group of input sections that can | |
3032 | be handled by one stub section. A negative value of @var{N} locates | |
3033 | stub sections after their branches, while a positive value allows stub | |
3034 | sections to appear either before or after the branches. Values of | |
3035 | @samp{1} or @samp{-1} indicate that the | |
3036 | linker should choose suitable defaults. | |
3037 | ||
3038 | @end table | |
3039 | ||
3040 | @c man end | |
3041 | @end ifset | |
3042 | ||
93fd0973 SC |
3043 | @ifset M68HC11 |
3044 | @subsection Options specific to Motorola 68HC11 and 68HC12 targets | |
3045 | ||
3046 | @c man begin OPTIONS | |
3047 | ||
3048 | The 68HC11 and 68HC12 linkers support specific options to control the | |
3049 | memory bank switching mapping and trampoline code generation. | |
3050 | ||
3051 | @table @gcctabopt | |
3052 | ||
3053 | @kindex --no-trampoline | |
3054 | @item --no-trampoline | |
3055 | This option disables the generation of trampoline. By default a trampoline | |
3056 | is generated for each far function which is called using a @code{jsr} | |
3057 | instruction (this happens when a pointer to a far function is taken). | |
3058 | ||
3059 | @kindex --bank-window | |
3060 | @item --bank-window @var{name} | |
3061 | This option indicates to the linker the name of the memory region in | |
3062 | the @samp{MEMORY} specification that describes the memory bank window. | |
3063 | The definition of such region is then used by the linker to compute | |
3064 | paging and addresses within the memory window. | |
3065 | ||
3066 | @end table | |
3067 | ||
3068 | @c man end | |
3069 | @end ifset | |
3070 | ||
7fb9f789 NC |
3071 | @ifset M68K |
3072 | @subsection Options specific to Motorola 68K target | |
3073 | ||
3074 | @c man begin OPTIONS | |
3075 | ||
3076 | The following options are supported to control handling of GOT generation | |
3077 | when linking for 68K targets. | |
3078 | ||
3079 | @table @gcctabopt | |
3080 | ||
3081 | @kindex --got | |
3082 | @item --got=@var{type} | |
3083 | This option tells the linker which GOT generation scheme to use. | |
3084 | @var{type} should be one of @samp{single}, @samp{negative}, | |
3085 | @samp{multigot} or @samp{target}. For more information refer to the | |
3086 | Info entry for @file{ld}. | |
3087 | ||
3088 | @end table | |
3089 | ||
3090 | @c man end | |
3091 | @end ifset | |
3092 | ||
833794fc MR |
3093 | @ifset MIPS |
3094 | @subsection Options specific to MIPS targets | |
3095 | ||
3096 | @c man begin OPTIONS | |
3097 | ||
3098 | The following options are supported to control microMIPS instruction | |
8b10b0b3 MR |
3099 | generation and branch relocation checks for ISA mode transitions when |
3100 | linking for MIPS targets. | |
833794fc MR |
3101 | |
3102 | @table @gcctabopt | |
3103 | ||
3104 | @kindex --insn32 | |
3105 | @item --insn32 | |
3106 | @kindex --no-insn32 | |
3107 | @itemx --no-insn32 | |
3108 | These options control the choice of microMIPS instructions used in code | |
3109 | generated by the linker, such as that in the PLT or lazy binding stubs, | |
3110 | or in relaxation. If @samp{--insn32} is used, then the linker only uses | |
3111 | 32-bit instruction encodings. By default or if @samp{--no-insn32} is | |
3112 | used, all instruction encodings are used, including 16-bit ones where | |
3113 | possible. | |
3114 | ||
8b10b0b3 MR |
3115 | @kindex --ignore-branch-isa |
3116 | @item --ignore-branch-isa | |
3117 | @kindex --no-ignore-branch-isa | |
3118 | @itemx --no-ignore-branch-isa | |
3119 | These options control branch relocation checks for invalid ISA mode | |
3120 | transitions. If @samp{--ignore-branch-isa} is used, then the linker | |
3121 | accepts any branch relocations and any ISA mode transition required | |
3122 | is lost in relocation calculation, except for some cases of @code{BAL} | |
3123 | instructions which meet relaxation conditions and are converted to | |
3124 | equivalent @code{JALX} instructions as the associated relocation is | |
3125 | calculated. By default or if @samp{--no-ignore-branch-isa} is used | |
3126 | a check is made causing the loss of an ISA mode transition to produce | |
3127 | an error. | |
3128 | ||
833794fc MR |
3129 | @end table |
3130 | ||
3131 | @c man end | |
3132 | @end ifset | |
3133 | ||
252b5132 RH |
3134 | @ifset UsesEnvVars |
3135 | @node Environment | |
3136 | @section Environment Variables | |
3137 | ||
0285c67d NC |
3138 | @c man begin ENVIRONMENT |
3139 | ||
560e09e9 | 3140 | You can change the behaviour of @command{ld} with the environment variables |
36f63dca NC |
3141 | @ifclear SingleFormat |
3142 | @code{GNUTARGET}, | |
3143 | @end ifclear | |
3144 | @code{LDEMULATION} and @code{COLLECT_NO_DEMANGLE}. | |
252b5132 | 3145 | |
36f63dca | 3146 | @ifclear SingleFormat |
252b5132 RH |
3147 | @kindex GNUTARGET |
3148 | @cindex default input format | |
3149 | @code{GNUTARGET} determines the input-file object format if you don't | |
3150 | use @samp{-b} (or its synonym @samp{--format}). Its value should be one | |
3151 | of the BFD names for an input format (@pxref{BFD}). If there is no | |
ff5dcc92 | 3152 | @code{GNUTARGET} in the environment, @command{ld} uses the natural format |
252b5132 RH |
3153 | of the target. If @code{GNUTARGET} is set to @code{default} then BFD |
3154 | attempts to discover the input format by examining binary input files; | |
3155 | this method often succeeds, but there are potential ambiguities, since | |
3156 | there is no method of ensuring that the magic number used to specify | |
3157 | object-file formats is unique. However, the configuration procedure for | |
3158 | BFD on each system places the conventional format for that system first | |
3159 | in the search-list, so ambiguities are resolved in favor of convention. | |
36f63dca | 3160 | @end ifclear |
252b5132 RH |
3161 | |
3162 | @kindex LDEMULATION | |
3163 | @cindex default emulation | |
3164 | @cindex emulation, default | |
3165 | @code{LDEMULATION} determines the default emulation if you don't use the | |
3166 | @samp{-m} option. The emulation can affect various aspects of linker | |
3167 | behaviour, particularly the default linker script. You can list the | |
3168 | available emulations with the @samp{--verbose} or @samp{-V} options. If | |
3169 | the @samp{-m} option is not used, and the @code{LDEMULATION} environment | |
3170 | variable is not defined, the default emulation depends upon how the | |
3171 | linker was configured. | |
252b5132 RH |
3172 | |
3173 | @kindex COLLECT_NO_DEMANGLE | |
3174 | @cindex demangling, default | |
3175 | Normally, the linker will default to demangling symbols. However, if | |
3176 | @code{COLLECT_NO_DEMANGLE} is set in the environment, then it will | |
3177 | default to not demangling symbols. This environment variable is used in | |
3178 | a similar fashion by the @code{gcc} linker wrapper program. The default | |
3179 | may be overridden by the @samp{--demangle} and @samp{--no-demangle} | |
3180 | options. | |
3181 | ||
0285c67d NC |
3182 | @c man end |
3183 | @end ifset | |
3184 | ||
252b5132 RH |
3185 | @node Scripts |
3186 | @chapter Linker Scripts | |
3187 | ||
3188 | @cindex scripts | |
3189 | @cindex linker scripts | |
3190 | @cindex command files | |
3191 | Every link is controlled by a @dfn{linker script}. This script is | |
3192 | written in the linker command language. | |
3193 | ||
3194 | The main purpose of the linker script is to describe how the sections in | |
3195 | the input files should be mapped into the output file, and to control | |
3196 | the memory layout of the output file. Most linker scripts do nothing | |
3197 | more than this. However, when necessary, the linker script can also | |
3198 | direct the linker to perform many other operations, using the commands | |
3199 | described below. | |
3200 | ||
3201 | The linker always uses a linker script. If you do not supply one | |
3202 | yourself, the linker will use a default script that is compiled into the | |
a05a5b64 TP |
3203 | linker executable. You can use the @samp{--verbose} command-line option |
3204 | to display the default linker script. Certain command-line options, | |
252b5132 RH |
3205 | such as @samp{-r} or @samp{-N}, will affect the default linker script. |
3206 | ||
3207 | You may supply your own linker script by using the @samp{-T} command | |
3208 | line option. When you do this, your linker script will replace the | |
3209 | default linker script. | |
3210 | ||
3211 | You may also use linker scripts implicitly by naming them as input files | |
3212 | to the linker, as though they were files to be linked. @xref{Implicit | |
3213 | Linker Scripts}. | |
3214 | ||
3215 | @menu | |
3216 | * Basic Script Concepts:: Basic Linker Script Concepts | |
3217 | * Script Format:: Linker Script Format | |
3218 | * Simple Example:: Simple Linker Script Example | |
3219 | * Simple Commands:: Simple Linker Script Commands | |
3220 | * Assignments:: Assigning Values to Symbols | |
3221 | * SECTIONS:: SECTIONS Command | |
3222 | * MEMORY:: MEMORY Command | |
3223 | * PHDRS:: PHDRS Command | |
3224 | * VERSION:: VERSION Command | |
3225 | * Expressions:: Expressions in Linker Scripts | |
3226 | * Implicit Linker Scripts:: Implicit Linker Scripts | |
3227 | @end menu | |
3228 | ||
3229 | @node Basic Script Concepts | |
3230 | @section Basic Linker Script Concepts | |
3231 | @cindex linker script concepts | |
3232 | We need to define some basic concepts and vocabulary in order to | |
3233 | describe the linker script language. | |
3234 | ||
3235 | The linker combines input files into a single output file. The output | |
3236 | file and each input file are in a special data format known as an | |
3237 | @dfn{object file format}. Each file is called an @dfn{object file}. | |
3238 | The output file is often called an @dfn{executable}, but for our | |
3239 | purposes we will also call it an object file. Each object file has, | |
3240 | among other things, a list of @dfn{sections}. We sometimes refer to a | |
3241 | section in an input file as an @dfn{input section}; similarly, a section | |
3242 | in the output file is an @dfn{output section}. | |
3243 | ||
3244 | Each section in an object file has a name and a size. Most sections | |
3245 | also have an associated block of data, known as the @dfn{section | |
56dd11f0 | 3246 | contents}. A section may be marked as @dfn{loadable}, which means that |
252b5132 RH |
3247 | the contents should be loaded into memory when the output file is run. |
3248 | A section with no contents may be @dfn{allocatable}, which means that an | |
3249 | area in memory should be set aside, but nothing in particular should be | |
3250 | loaded there (in some cases this memory must be zeroed out). A section | |
3251 | which is neither loadable nor allocatable typically contains some sort | |
3252 | of debugging information. | |
3253 | ||
3254 | Every loadable or allocatable output section has two addresses. The | |
3255 | first is the @dfn{VMA}, or virtual memory address. This is the address | |
3256 | the section will have when the output file is run. The second is the | |
3257 | @dfn{LMA}, or load memory address. This is the address at which the | |
3258 | section will be loaded. In most cases the two addresses will be the | |
3259 | same. An example of when they might be different is when a data section | |
3260 | is loaded into ROM, and then copied into RAM when the program starts up | |
3261 | (this technique is often used to initialize global variables in a ROM | |
3262 | based system). In this case the ROM address would be the LMA, and the | |
3263 | RAM address would be the VMA. | |
3264 | ||
3265 | You can see the sections in an object file by using the @code{objdump} | |
3266 | program with the @samp{-h} option. | |
3267 | ||
3268 | Every object file also has a list of @dfn{symbols}, known as the | |
3269 | @dfn{symbol table}. A symbol may be defined or undefined. Each symbol | |
3270 | has a name, and each defined symbol has an address, among other | |
3271 | information. If you compile a C or C++ program into an object file, you | |
3272 | will get a defined symbol for every defined function and global or | |
3273 | static variable. Every undefined function or global variable which is | |
3274 | referenced in the input file will become an undefined symbol. | |
3275 | ||
3276 | You can see the symbols in an object file by using the @code{nm} | |
3277 | program, or by using the @code{objdump} program with the @samp{-t} | |
3278 | option. | |
3279 | ||
3280 | @node Script Format | |
3281 | @section Linker Script Format | |
3282 | @cindex linker script format | |
3283 | Linker scripts are text files. | |
3284 | ||
3285 | You write a linker script as a series of commands. Each command is | |
3286 | either a keyword, possibly followed by arguments, or an assignment to a | |
3287 | symbol. You may separate commands using semicolons. Whitespace is | |
3288 | generally ignored. | |
3289 | ||
3290 | Strings such as file or format names can normally be entered directly. | |
3291 | If the file name contains a character such as a comma which would | |
3292 | otherwise serve to separate file names, you may put the file name in | |
3293 | double quotes. There is no way to use a double quote character in a | |
3294 | file name. | |
3295 | ||
3296 | You may include comments in linker scripts just as in C, delimited by | |
3297 | @samp{/*} and @samp{*/}. As in C, comments are syntactically equivalent | |
3298 | to whitespace. | |
3299 | ||
3300 | @node Simple Example | |
3301 | @section Simple Linker Script Example | |
3302 | @cindex linker script example | |
3303 | @cindex example of linker script | |
3304 | Many linker scripts are fairly simple. | |
3305 | ||
3306 | The simplest possible linker script has just one command: | |
3307 | @samp{SECTIONS}. You use the @samp{SECTIONS} command to describe the | |
3308 | memory layout of the output file. | |
3309 | ||
3310 | The @samp{SECTIONS} command is a powerful command. Here we will | |
3311 | describe a simple use of it. Let's assume your program consists only of | |
3312 | code, initialized data, and uninitialized data. These will be in the | |
3313 | @samp{.text}, @samp{.data}, and @samp{.bss} sections, respectively. | |
3314 | Let's assume further that these are the only sections which appear in | |
3315 | your input files. | |
3316 | ||
3317 | For this example, let's say that the code should be loaded at address | |
3318 | 0x10000, and that the data should start at address 0x8000000. Here is a | |
3319 | linker script which will do that: | |
3320 | @smallexample | |
3321 | SECTIONS | |
3322 | @{ | |
3323 | . = 0x10000; | |
3324 | .text : @{ *(.text) @} | |
3325 | . = 0x8000000; | |
3326 | .data : @{ *(.data) @} | |
3327 | .bss : @{ *(.bss) @} | |
3328 | @} | |
3329 | @end smallexample | |
3330 | ||
3331 | You write the @samp{SECTIONS} command as the keyword @samp{SECTIONS}, | |
3332 | followed by a series of symbol assignments and output section | |
3333 | descriptions enclosed in curly braces. | |
3334 | ||
252b5132 RH |
3335 | The first line inside the @samp{SECTIONS} command of the above example |
3336 | sets the value of the special symbol @samp{.}, which is the location | |
3337 | counter. If you do not specify the address of an output section in some | |
3338 | other way (other ways are described later), the address is set from the | |
3339 | current value of the location counter. The location counter is then | |
3340 | incremented by the size of the output section. At the start of the | |
3341 | @samp{SECTIONS} command, the location counter has the value @samp{0}. | |
3342 | ||
3343 | The second line defines an output section, @samp{.text}. The colon is | |
3344 | required syntax which may be ignored for now. Within the curly braces | |
3345 | after the output section name, you list the names of the input sections | |
3346 | which should be placed into this output section. The @samp{*} is a | |
3347 | wildcard which matches any file name. The expression @samp{*(.text)} | |
3348 | means all @samp{.text} input sections in all input files. | |
3349 | ||
3350 | Since the location counter is @samp{0x10000} when the output section | |
3351 | @samp{.text} is defined, the linker will set the address of the | |
3352 | @samp{.text} section in the output file to be @samp{0x10000}. | |
3353 | ||
3354 | The remaining lines define the @samp{.data} and @samp{.bss} sections in | |
3355 | the output file. The linker will place the @samp{.data} output section | |
3356 | at address @samp{0x8000000}. After the linker places the @samp{.data} | |
3357 | output section, the value of the location counter will be | |
3358 | @samp{0x8000000} plus the size of the @samp{.data} output section. The | |
3359 | effect is that the linker will place the @samp{.bss} output section | |
58434bc1 | 3360 | immediately after the @samp{.data} output section in memory. |
252b5132 RH |
3361 | |
3362 | The linker will ensure that each output section has the required | |
3363 | alignment, by increasing the location counter if necessary. In this | |
3364 | example, the specified addresses for the @samp{.text} and @samp{.data} | |
3365 | sections will probably satisfy any alignment constraints, but the linker | |
3366 | may have to create a small gap between the @samp{.data} and @samp{.bss} | |
3367 | sections. | |
3368 | ||
3369 | That's it! That's a simple and complete linker script. | |
3370 | ||
3371 | @node Simple Commands | |
3372 | @section Simple Linker Script Commands | |
3373 | @cindex linker script simple commands | |
3374 | In this section we describe the simple linker script commands. | |
3375 | ||
3376 | @menu | |
3377 | * Entry Point:: Setting the entry point | |
3378 | * File Commands:: Commands dealing with files | |
3379 | @ifclear SingleFormat | |
3380 | * Format Commands:: Commands dealing with object file formats | |
3381 | @end ifclear | |
3382 | ||
4a93e180 | 3383 | * REGION_ALIAS:: Assign alias names to memory regions |
252b5132 RH |
3384 | * Miscellaneous Commands:: Other linker script commands |
3385 | @end menu | |
3386 | ||
3387 | @node Entry Point | |
36f63dca | 3388 | @subsection Setting the Entry Point |
252b5132 RH |
3389 | @kindex ENTRY(@var{symbol}) |
3390 | @cindex start of execution | |
3391 | @cindex first instruction | |
3392 | @cindex entry point | |
3393 | The first instruction to execute in a program is called the @dfn{entry | |
3394 | point}. You can use the @code{ENTRY} linker script command to set the | |
3395 | entry point. The argument is a symbol name: | |
3396 | @smallexample | |
3397 | ENTRY(@var{symbol}) | |
3398 | @end smallexample | |
3399 | ||
3400 | There are several ways to set the entry point. The linker will set the | |
3401 | entry point by trying each of the following methods in order, and | |
3402 | stopping when one of them succeeds: | |
3403 | @itemize @bullet | |
a1ab1d2a | 3404 | @item |
252b5132 | 3405 | the @samp{-e} @var{entry} command-line option; |
a1ab1d2a | 3406 | @item |
252b5132 | 3407 | the @code{ENTRY(@var{symbol})} command in a linker script; |
a1ab1d2a | 3408 | @item |
3ab904c4 NC |
3409 | the value of a target specific symbol, if it is defined; For many |
3410 | targets this is @code{start}, but PE and BeOS based systems for example | |
3411 | check a list of possible entry symbols, matching the first one found. | |
a1ab1d2a | 3412 | @item |
252b5132 | 3413 | the address of the first byte of the @samp{.text} section, if present; |
a1ab1d2a | 3414 | @item |
252b5132 RH |
3415 | The address @code{0}. |
3416 | @end itemize | |
3417 | ||
3418 | @node File Commands | |
36f63dca | 3419 | @subsection Commands Dealing with Files |
252b5132 RH |
3420 | @cindex linker script file commands |
3421 | Several linker script commands deal with files. | |
3422 | ||
3423 | @table @code | |
3424 | @item INCLUDE @var{filename} | |
3425 | @kindex INCLUDE @var{filename} | |
3426 | @cindex including a linker script | |
3427 | Include the linker script @var{filename} at this point. The file will | |
3428 | be searched for in the current directory, and in any directory specified | |
ff5dcc92 | 3429 | with the @option{-L} option. You can nest calls to @code{INCLUDE} up to |
252b5132 RH |
3430 | 10 levels deep. |
3431 | ||
4006703d NS |
3432 | You can place @code{INCLUDE} directives at the top level, in @code{MEMORY} or |
3433 | @code{SECTIONS} commands, or in output section descriptions. | |
3434 | ||
252b5132 RH |
3435 | @item INPUT(@var{file}, @var{file}, @dots{}) |
3436 | @itemx INPUT(@var{file} @var{file} @dots{}) | |
3437 | @kindex INPUT(@var{files}) | |
3438 | @cindex input files in linker scripts | |
3439 | @cindex input object files in linker scripts | |
3440 | @cindex linker script input object files | |
3441 | The @code{INPUT} command directs the linker to include the named files | |
3442 | in the link, as though they were named on the command line. | |
3443 | ||
3444 | For example, if you always want to include @file{subr.o} any time you do | |
3445 | a link, but you can't be bothered to put it on every link command line, | |
3446 | then you can put @samp{INPUT (subr.o)} in your linker script. | |
3447 | ||
3448 | In fact, if you like, you can list all of your input files in the linker | |
3449 | script, and then invoke the linker with nothing but a @samp{-T} option. | |
3450 | ||
e3f2db7f AO |
3451 | In case a @dfn{sysroot prefix} is configured, and the filename starts |
3452 | with the @samp{/} character, and the script being processed was | |
3453 | located inside the @dfn{sysroot prefix}, the filename will be looked | |
3454 | for in the @dfn{sysroot prefix}. Otherwise, the linker will try to | |
3455 | open the file in the current directory. If it is not found, the | |
cad60a33 HPN |
3456 | linker will search through the archive library search path. |
3457 | The @dfn{sysroot prefix} can also be forced by specifying @code{=} | |
3aa2d05a NC |
3458 | as the first character in the filename path, or prefixing the filename |
3459 | path with @code{$SYSROOT}. See also the description of @samp{-L} in | |
a05a5b64 | 3460 | @ref{Options,,Command-line Options}. |
252b5132 | 3461 | |
ff5dcc92 | 3462 | If you use @samp{INPUT (-l@var{file})}, @command{ld} will transform the |
a05a5b64 | 3463 | name to @code{lib@var{file}.a}, as with the command-line argument |
252b5132 RH |
3464 | @samp{-l}. |
3465 | ||
3466 | When you use the @code{INPUT} command in an implicit linker script, the | |
3467 | files will be included in the link at the point at which the linker | |
3468 | script file is included. This can affect archive searching. | |
3469 | ||
3470 | @item GROUP(@var{file}, @var{file}, @dots{}) | |
3471 | @itemx GROUP(@var{file} @var{file} @dots{}) | |
3472 | @kindex GROUP(@var{files}) | |
3473 | @cindex grouping input files | |
3474 | The @code{GROUP} command is like @code{INPUT}, except that the named | |
3475 | files should all be archives, and they are searched repeatedly until no | |
3476 | new undefined references are created. See the description of @samp{-(} | |
a05a5b64 | 3477 | in @ref{Options,,Command-line Options}. |
252b5132 | 3478 | |
b717d30e JJ |
3479 | @item AS_NEEDED(@var{file}, @var{file}, @dots{}) |
3480 | @itemx AS_NEEDED(@var{file} @var{file} @dots{}) | |
3481 | @kindex AS_NEEDED(@var{files}) | |
3482 | This construct can appear only inside of the @code{INPUT} or @code{GROUP} | |
3483 | commands, among other filenames. The files listed will be handled | |
3484 | as if they appear directly in the @code{INPUT} or @code{GROUP} commands, | |
3485 | with the exception of ELF shared libraries, that will be added only | |
3486 | when they are actually needed. This construct essentially enables | |
3487 | @option{--as-needed} option for all the files listed inside of it | |
3488 | and restores previous @option{--as-needed} resp. @option{--no-as-needed} | |
3489 | setting afterwards. | |
3490 | ||
252b5132 RH |
3491 | @item OUTPUT(@var{filename}) |
3492 | @kindex OUTPUT(@var{filename}) | |
b45619c0 | 3493 | @cindex output file name in linker script |
252b5132 RH |
3494 | The @code{OUTPUT} command names the output file. Using |
3495 | @code{OUTPUT(@var{filename})} in the linker script is exactly like using | |
3496 | @samp{-o @var{filename}} on the command line (@pxref{Options,,Command | |
a05a5b64 | 3497 | Line Options}). If both are used, the command-line option takes |
252b5132 RH |
3498 | precedence. |
3499 | ||
3500 | You can use the @code{OUTPUT} command to define a default name for the | |
3501 | output file other than the usual default of @file{a.out}. | |
3502 | ||
3503 | @item SEARCH_DIR(@var{path}) | |
3504 | @kindex SEARCH_DIR(@var{path}) | |
3505 | @cindex library search path in linker script | |
3506 | @cindex archive search path in linker script | |
3507 | @cindex search path in linker script | |
3508 | The @code{SEARCH_DIR} command adds @var{path} to the list of paths where | |
ff5dcc92 | 3509 | @command{ld} looks for archive libraries. Using |
252b5132 | 3510 | @code{SEARCH_DIR(@var{path})} is exactly like using @samp{-L @var{path}} |
a05a5b64 | 3511 | on the command line (@pxref{Options,,Command-line Options}). If both |
252b5132 | 3512 | are used, then the linker will search both paths. Paths specified using |
a05a5b64 | 3513 | the command-line option are searched first. |
252b5132 RH |
3514 | |
3515 | @item STARTUP(@var{filename}) | |
3516 | @kindex STARTUP(@var{filename}) | |
3517 | @cindex first input file | |
3518 | The @code{STARTUP} command is just like the @code{INPUT} command, except | |
3519 | that @var{filename} will become the first input file to be linked, as | |
3520 | though it were specified first on the command line. This may be useful | |
3521 | when using a system in which the entry point is always the start of the | |
3522 | first file. | |
3523 | @end table | |
3524 | ||
3525 | @ifclear SingleFormat | |
3526 | @node Format Commands | |
36f63dca | 3527 | @subsection Commands Dealing with Object File Formats |
252b5132 RH |
3528 | A couple of linker script commands deal with object file formats. |
3529 | ||
3530 | @table @code | |
3531 | @item OUTPUT_FORMAT(@var{bfdname}) | |
3532 | @itemx OUTPUT_FORMAT(@var{default}, @var{big}, @var{little}) | |
3533 | @kindex OUTPUT_FORMAT(@var{bfdname}) | |
3534 | @cindex output file format in linker script | |
3535 | The @code{OUTPUT_FORMAT} command names the BFD format to use for the | |
3536 | output file (@pxref{BFD}). Using @code{OUTPUT_FORMAT(@var{bfdname})} is | |
024531e2 | 3537 | exactly like using @samp{--oformat @var{bfdname}} on the command line |
a05a5b64 | 3538 | (@pxref{Options,,Command-line Options}). If both are used, the command |
252b5132 RH |
3539 | line option takes precedence. |
3540 | ||
3541 | You can use @code{OUTPUT_FORMAT} with three arguments to use different | |
a05a5b64 | 3542 | formats based on the @samp{-EB} and @samp{-EL} command-line options. |
252b5132 RH |
3543 | This permits the linker script to set the output format based on the |
3544 | desired endianness. | |
3545 | ||
3546 | If neither @samp{-EB} nor @samp{-EL} are used, then the output format | |
3547 | will be the first argument, @var{default}. If @samp{-EB} is used, the | |
3548 | output format will be the second argument, @var{big}. If @samp{-EL} is | |
3549 | used, the output format will be the third argument, @var{little}. | |
3550 | ||
3551 | For example, the default linker script for the MIPS ELF target uses this | |
3552 | command: | |
3553 | @smallexample | |
3554 | OUTPUT_FORMAT(elf32-bigmips, elf32-bigmips, elf32-littlemips) | |
3555 | @end smallexample | |
3556 | This says that the default format for the output file is | |
a05a5b64 | 3557 | @samp{elf32-bigmips}, but if the user uses the @samp{-EL} command-line |
252b5132 RH |
3558 | option, the output file will be created in the @samp{elf32-littlemips} |
3559 | format. | |
3560 | ||
3561 | @item TARGET(@var{bfdname}) | |
3562 | @kindex TARGET(@var{bfdname}) | |
3563 | @cindex input file format in linker script | |
3564 | The @code{TARGET} command names the BFD format to use when reading input | |
3565 | files. It affects subsequent @code{INPUT} and @code{GROUP} commands. | |
3566 | This command is like using @samp{-b @var{bfdname}} on the command line | |
a05a5b64 | 3567 | (@pxref{Options,,Command-line Options}). If the @code{TARGET} command |
252b5132 RH |
3568 | is used but @code{OUTPUT_FORMAT} is not, then the last @code{TARGET} |
3569 | command is also used to set the format for the output file. @xref{BFD}. | |
3570 | @end table | |
3571 | @end ifclear | |
3572 | ||
4a93e180 NC |
3573 | @node REGION_ALIAS |
3574 | @subsection Assign alias names to memory regions | |
3575 | @kindex REGION_ALIAS(@var{alias}, @var{region}) | |
3576 | @cindex region alias | |
3577 | @cindex region names | |
3578 | ||
3579 | Alias names can be added to existing memory regions created with the | |
3580 | @ref{MEMORY} command. Each name corresponds to at most one memory region. | |
3581 | ||
3582 | @smallexample | |
3583 | REGION_ALIAS(@var{alias}, @var{region}) | |
3584 | @end smallexample | |
3585 | ||
3586 | The @code{REGION_ALIAS} function creates an alias name @var{alias} for the | |
3587 | memory region @var{region}. This allows a flexible mapping of output sections | |
3588 | to memory regions. An example follows. | |
3589 | ||
3590 | Suppose we have an application for embedded systems which come with various | |
3591 | memory storage devices. All have a general purpose, volatile memory @code{RAM} | |
3592 | that allows code execution or data storage. Some may have a read-only, | |
3593 | non-volatile memory @code{ROM} that allows code execution and read-only data | |
3594 | access. The last variant is a read-only, non-volatile memory @code{ROM2} with | |
3595 | read-only data access and no code execution capability. We have four output | |
3596 | sections: | |
3597 | ||
3598 | @itemize @bullet | |
3599 | @item | |
3600 | @code{.text} program code; | |
3601 | @item | |
3602 | @code{.rodata} read-only data; | |
3603 | @item | |
3604 | @code{.data} read-write initialized data; | |
3605 | @item | |
3606 | @code{.bss} read-write zero initialized data. | |
3607 | @end itemize | |
3608 | ||
3609 | The goal is to provide a linker command file that contains a system independent | |
3610 | part defining the output sections and a system dependent part mapping the | |
3611 | output sections to the memory regions available on the system. Our embedded | |
3612 | systems come with three different memory setups @code{A}, @code{B} and | |
3613 | @code{C}: | |
3614 | @multitable @columnfractions .25 .25 .25 .25 | |
3615 | @item Section @tab Variant A @tab Variant B @tab Variant C | |
9d5777a3 | 3616 | @item .text @tab RAM @tab ROM @tab ROM |
4a93e180 NC |
3617 | @item .rodata @tab RAM @tab ROM @tab ROM2 |
3618 | @item .data @tab RAM @tab RAM/ROM @tab RAM/ROM2 | |
3619 | @item .bss @tab RAM @tab RAM @tab RAM | |
3620 | @end multitable | |
3621 | The notation @code{RAM/ROM} or @code{RAM/ROM2} means that this section is | |
3622 | loaded into region @code{ROM} or @code{ROM2} respectively. Please note that | |
3623 | the load address of the @code{.data} section starts in all three variants at | |
3624 | the end of the @code{.rodata} section. | |
3625 | ||
3626 | The base linker script that deals with the output sections follows. It | |
3627 | includes the system dependent @code{linkcmds.memory} file that describes the | |
3628 | memory layout: | |
3629 | @smallexample | |
3630 | INCLUDE linkcmds.memory | |
3631 | ||
3632 | SECTIONS | |
3633 | @{ | |
3634 | .text : | |
3635 | @{ | |
3636 | *(.text) | |
3637 | @} > REGION_TEXT | |
3638 | .rodata : | |
3639 | @{ | |
3640 | *(.rodata) | |
3641 | rodata_end = .; | |
3642 | @} > REGION_RODATA | |
3643 | .data : AT (rodata_end) | |
3644 | @{ | |
3645 | data_start = .; | |
3646 | *(.data) | |
3647 | @} > REGION_DATA | |
3648 | data_size = SIZEOF(.data); | |
3649 | data_load_start = LOADADDR(.data); | |
3650 | .bss : | |
3651 | @{ | |
3652 | *(.bss) | |
3653 | @} > REGION_BSS | |
3654 | @} | |
3655 | @end smallexample | |
3656 | ||
3657 | Now we need three different @code{linkcmds.memory} files to define memory | |
3658 | regions and alias names. The content of @code{linkcmds.memory} for the three | |
3659 | variants @code{A}, @code{B} and @code{C}: | |
3660 | @table @code | |
3661 | @item A | |
3662 | Here everything goes into the @code{RAM}. | |
3663 | @smallexample | |
3664 | MEMORY | |
3665 | @{ | |
3666 | RAM : ORIGIN = 0, LENGTH = 4M | |
3667 | @} | |
3668 | ||
3669 | REGION_ALIAS("REGION_TEXT", RAM); | |
3670 | REGION_ALIAS("REGION_RODATA", RAM); | |
3671 | REGION_ALIAS("REGION_DATA", RAM); | |
3672 | REGION_ALIAS("REGION_BSS", RAM); | |
3673 | @end smallexample | |
3674 | @item B | |
3675 | Program code and read-only data go into the @code{ROM}. Read-write data goes | |
3676 | into the @code{RAM}. An image of the initialized data is loaded into the | |
3677 | @code{ROM} and will be copied during system start into the @code{RAM}. | |
3678 | @smallexample | |
3679 | MEMORY | |
3680 | @{ | |
3681 | ROM : ORIGIN = 0, LENGTH = 3M | |
3682 | RAM : ORIGIN = 0x10000000, LENGTH = 1M | |
3683 | @} | |
3684 | ||
3685 | REGION_ALIAS("REGION_TEXT", ROM); | |
3686 | REGION_ALIAS("REGION_RODATA", ROM); | |
3687 | REGION_ALIAS("REGION_DATA", RAM); | |
3688 | REGION_ALIAS("REGION_BSS", RAM); | |
3689 | @end smallexample | |
3690 | @item C | |
3691 | Program code goes into the @code{ROM}. Read-only data goes into the | |
3692 | @code{ROM2}. Read-write data goes into the @code{RAM}. An image of the | |
3693 | initialized data is loaded into the @code{ROM2} and will be copied during | |
3694 | system start into the @code{RAM}. | |
3695 | @smallexample | |
3696 | MEMORY | |
3697 | @{ | |
3698 | ROM : ORIGIN = 0, LENGTH = 2M | |
3699 | ROM2 : ORIGIN = 0x10000000, LENGTH = 1M | |
3700 | RAM : ORIGIN = 0x20000000, LENGTH = 1M | |
3701 | @} | |
3702 | ||
3703 | REGION_ALIAS("REGION_TEXT", ROM); | |
3704 | REGION_ALIAS("REGION_RODATA", ROM2); | |
3705 | REGION_ALIAS("REGION_DATA", RAM); | |
3706 | REGION_ALIAS("REGION_BSS", RAM); | |
3707 | @end smallexample | |
3708 | @end table | |
3709 | ||
3710 | It is possible to write a common system initialization routine to copy the | |
3711 | @code{.data} section from @code{ROM} or @code{ROM2} into the @code{RAM} if | |
3712 | necessary: | |
3713 | @smallexample | |
3714 | #include <string.h> | |
3715 | ||
3716 | extern char data_start []; | |
3717 | extern char data_size []; | |
3718 | extern char data_load_start []; | |
3719 | ||
3720 | void copy_data(void) | |
3721 | @{ | |
3722 | if (data_start != data_load_start) | |
3723 | @{ | |
3724 | memcpy(data_start, data_load_start, (size_t) data_size); | |
3725 | @} | |
3726 | @} | |
3727 | @end smallexample | |
3728 | ||
252b5132 | 3729 | @node Miscellaneous Commands |
36f63dca | 3730 | @subsection Other Linker Script Commands |
252b5132 RH |
3731 | There are a few other linker scripts commands. |
3732 | ||
3733 | @table @code | |
3734 | @item ASSERT(@var{exp}, @var{message}) | |
3735 | @kindex ASSERT | |
3736 | @cindex assertion in linker script | |
3737 | Ensure that @var{exp} is non-zero. If it is zero, then exit the linker | |
3738 | with an error code, and print @var{message}. | |
3739 | ||
fd1c4238 NC |
3740 | Note that assertions are checked before the final stages of linking |
3741 | take place. This means that expressions involving symbols PROVIDEd | |
3742 | inside section definitions will fail if the user has not set values | |
3743 | for those symbols. The only exception to this rule is PROVIDEd | |
3744 | symbols that just reference dot. Thus an assertion like this: | |
3745 | ||
3746 | @smallexample | |
3747 | .stack : | |
3748 | @{ | |
3749 | PROVIDE (__stack = .); | |
3750 | PROVIDE (__stack_size = 0x100); | |
3751 | ASSERT ((__stack > (_end + __stack_size)), "Error: No room left for the stack"); | |
3752 | @} | |
3753 | @end smallexample | |
3754 | ||
3755 | will fail if @code{__stack_size} is not defined elsewhere. Symbols | |
3756 | PROVIDEd outside of section definitions are evaluated earlier, so they | |
3757 | can be used inside ASSERTions. Thus: | |
3758 | ||
3759 | @smallexample | |
3760 | PROVIDE (__stack_size = 0x100); | |
3761 | .stack : | |
3762 | @{ | |
3763 | PROVIDE (__stack = .); | |
3764 | ASSERT ((__stack > (_end + __stack_size)), "Error: No room left for the stack"); | |
3765 | @} | |
3766 | @end smallexample | |
3767 | ||
3768 | will work. | |
3769 | ||
252b5132 RH |
3770 | @item EXTERN(@var{symbol} @var{symbol} @dots{}) |
3771 | @kindex EXTERN | |
3772 | @cindex undefined symbol in linker script | |
3773 | Force @var{symbol} to be entered in the output file as an undefined | |
3774 | symbol. Doing this may, for example, trigger linking of additional | |
3775 | modules from standard libraries. You may list several @var{symbol}s for | |
3776 | each @code{EXTERN}, and you may use @code{EXTERN} multiple times. This | |
3777 | command has the same effect as the @samp{-u} command-line option. | |
3778 | ||
3779 | @item FORCE_COMMON_ALLOCATION | |
3780 | @kindex FORCE_COMMON_ALLOCATION | |
3781 | @cindex common allocation in linker script | |
3782 | This command has the same effect as the @samp{-d} command-line option: | |
ff5dcc92 | 3783 | to make @command{ld} assign space to common symbols even if a relocatable |
252b5132 RH |
3784 | output file is specified (@samp{-r}). |
3785 | ||
4818e05f AM |
3786 | @item INHIBIT_COMMON_ALLOCATION |
3787 | @kindex INHIBIT_COMMON_ALLOCATION | |
3788 | @cindex common allocation in linker script | |
3789 | This command has the same effect as the @samp{--no-define-common} | |
3790 | command-line option: to make @code{ld} omit the assignment of addresses | |
3791 | to common symbols even for a non-relocatable output file. | |
3792 | ||
7bdf4127 AB |
3793 | @item FORCE_GROUP_ALLOCATION |
3794 | @kindex FORCE_GROUP_ALLOCATION | |
3795 | @cindex group allocation in linker script | |
3796 | @cindex section groups | |
3797 | @cindex COMDAT | |
3798 | This command has the same effect as the | |
3799 | @samp{--force-group-allocation} command-line option: to make | |
3800 | @command{ld} place section group members like normal input sections, | |
3801 | and to delete the section groups even if a relocatable output file is | |
3802 | specified (@samp{-r}). | |
3803 | ||
53d25da6 AM |
3804 | @item INSERT [ AFTER | BEFORE ] @var{output_section} |
3805 | @kindex INSERT | |
3806 | @cindex insert user script into default script | |
3807 | This command is typically used in a script specified by @samp{-T} to | |
3808 | augment the default @code{SECTIONS} with, for example, overlays. It | |
3809 | inserts all prior linker script statements after (or before) | |
3810 | @var{output_section}, and also causes @samp{-T} to not override the | |
3811 | default linker script. The exact insertion point is as for orphan | |
3812 | sections. @xref{Location Counter}. The insertion happens after the | |
3813 | linker has mapped input sections to output sections. Prior to the | |
3814 | insertion, since @samp{-T} scripts are parsed before the default | |
3815 | linker script, statements in the @samp{-T} script occur before the | |
3816 | default linker script statements in the internal linker representation | |
3817 | of the script. In particular, input section assignments will be made | |
3818 | to @samp{-T} output sections before those in the default script. Here | |
3819 | is an example of how a @samp{-T} script using @code{INSERT} might look: | |
3820 | ||
3821 | @smallexample | |
3822 | SECTIONS | |
3823 | @{ | |
3824 | OVERLAY : | |
3825 | @{ | |
3826 | .ov1 @{ ov1*(.text) @} | |
3827 | .ov2 @{ ov2*(.text) @} | |
3828 | @} | |
3829 | @} | |
3830 | INSERT AFTER .text; | |
3831 | @end smallexample | |
3832 | ||
252b5132 RH |
3833 | @item NOCROSSREFS(@var{section} @var{section} @dots{}) |
3834 | @kindex NOCROSSREFS(@var{sections}) | |
3835 | @cindex cross references | |
ff5dcc92 | 3836 | This command may be used to tell @command{ld} to issue an error about any |
252b5132 RH |
3837 | references among certain output sections. |
3838 | ||
3839 | In certain types of programs, particularly on embedded systems when | |
3840 | using overlays, when one section is loaded into memory, another section | |
3841 | will not be. Any direct references between the two sections would be | |
3842 | errors. For example, it would be an error if code in one section called | |
3843 | a function defined in the other section. | |
3844 | ||
3845 | The @code{NOCROSSREFS} command takes a list of output section names. If | |
ff5dcc92 | 3846 | @command{ld} detects any cross references between the sections, it reports |
252b5132 RH |
3847 | an error and returns a non-zero exit status. Note that the |
3848 | @code{NOCROSSREFS} command uses output section names, not input section | |
3849 | names. | |
3850 | ||
cdf96953 MF |
3851 | @item NOCROSSREFS_TO(@var{tosection} @var{fromsection} @dots{}) |
3852 | @kindex NOCROSSREFS_TO(@var{tosection} @var{fromsections}) | |
3853 | @cindex cross references | |
3854 | This command may be used to tell @command{ld} to issue an error about any | |
3855 | references to one section from a list of other sections. | |
3856 | ||
3857 | The @code{NOCROSSREFS} command is useful when ensuring that two or more | |
3858 | output sections are entirely independent but there are situations where | |
3859 | a one-way dependency is needed. For example, in a multi-core application | |
3860 | there may be shared code that can be called from each core but for safety | |
3861 | must never call back. | |
3862 | ||
3863 | The @code{NOCROSSREFS_TO} command takes a list of output section names. | |
3864 | The first section can not be referenced from any of the other sections. | |
3865 | If @command{ld} detects any references to the first section from any of | |
3866 | the other sections, it reports an error and returns a non-zero exit | |
3867 | status. Note that the @code{NOCROSSREFS_TO} command uses output section | |
3868 | names, not input section names. | |
3869 | ||
252b5132 RH |
3870 | @ifclear SingleFormat |
3871 | @item OUTPUT_ARCH(@var{bfdarch}) | |
3872 | @kindex OUTPUT_ARCH(@var{bfdarch}) | |
3873 | @cindex machine architecture | |
3874 | @cindex architecture | |
3875 | Specify a particular output machine architecture. The argument is one | |
3876 | of the names used by the BFD library (@pxref{BFD}). You can see the | |
3877 | architecture of an object file by using the @code{objdump} program with | |
3878 | the @samp{-f} option. | |
3879 | @end ifclear | |
01554a74 AM |
3880 | |
3881 | @item LD_FEATURE(@var{string}) | |
3882 | @kindex LD_FEATURE(@var{string}) | |
3883 | This command may be used to modify @command{ld} behavior. If | |
3884 | @var{string} is @code{"SANE_EXPR"} then absolute symbols and numbers | |
3885 | in a script are simply treated as numbers everywhere. | |
3886 | @xref{Expression Section}. | |
252b5132 RH |
3887 | @end table |
3888 | ||
3889 | @node Assignments | |
3890 | @section Assigning Values to Symbols | |
3891 | @cindex assignment in scripts | |
3892 | @cindex symbol definition, scripts | |
3893 | @cindex variables, defining | |
3894 | You may assign a value to a symbol in a linker script. This will define | |
73ae6183 | 3895 | the symbol and place it into the symbol table with a global scope. |
252b5132 RH |
3896 | |
3897 | @menu | |
3898 | * Simple Assignments:: Simple Assignments | |
eb8476a6 | 3899 | * HIDDEN:: HIDDEN |
252b5132 | 3900 | * PROVIDE:: PROVIDE |
7af8e998 | 3901 | * PROVIDE_HIDDEN:: PROVIDE_HIDDEN |
73ae6183 | 3902 | * Source Code Reference:: How to use a linker script defined symbol in source code |
252b5132 RH |
3903 | @end menu |
3904 | ||
3905 | @node Simple Assignments | |
3906 | @subsection Simple Assignments | |
3907 | ||
3908 | You may assign to a symbol using any of the C assignment operators: | |
3909 | ||
3910 | @table @code | |
3911 | @item @var{symbol} = @var{expression} ; | |
3912 | @itemx @var{symbol} += @var{expression} ; | |
3913 | @itemx @var{symbol} -= @var{expression} ; | |
3914 | @itemx @var{symbol} *= @var{expression} ; | |
3915 | @itemx @var{symbol} /= @var{expression} ; | |
3916 | @itemx @var{symbol} <<= @var{expression} ; | |
3917 | @itemx @var{symbol} >>= @var{expression} ; | |
3918 | @itemx @var{symbol} &= @var{expression} ; | |
3919 | @itemx @var{symbol} |= @var{expression} ; | |
3920 | @end table | |
3921 | ||
3922 | The first case will define @var{symbol} to the value of | |
3923 | @var{expression}. In the other cases, @var{symbol} must already be | |
3924 | defined, and the value will be adjusted accordingly. | |
3925 | ||
3926 | The special symbol name @samp{.} indicates the location counter. You | |
b5666f2f | 3927 | may only use this within a @code{SECTIONS} command. @xref{Location Counter}. |
252b5132 RH |
3928 | |
3929 | The semicolon after @var{expression} is required. | |
3930 | ||
3931 | Expressions are defined below; see @ref{Expressions}. | |
3932 | ||
3933 | You may write symbol assignments as commands in their own right, or as | |
3934 | statements within a @code{SECTIONS} command, or as part of an output | |
3935 | section description in a @code{SECTIONS} command. | |
3936 | ||
3937 | The section of the symbol will be set from the section of the | |
3938 | expression; for more information, see @ref{Expression Section}. | |
3939 | ||
3940 | Here is an example showing the three different places that symbol | |
3941 | assignments may be used: | |
3942 | ||
3943 | @smallexample | |
3944 | floating_point = 0; | |
3945 | SECTIONS | |
3946 | @{ | |
3947 | .text : | |
3948 | @{ | |
3949 | *(.text) | |
3950 | _etext = .; | |
3951 | @} | |
156e34dd | 3952 | _bdata = (. + 3) & ~ 3; |
252b5132 RH |
3953 | .data : @{ *(.data) @} |
3954 | @} | |
3955 | @end smallexample | |
3956 | @noindent | |
3957 | In this example, the symbol @samp{floating_point} will be defined as | |
3958 | zero. The symbol @samp{_etext} will be defined as the address following | |
3959 | the last @samp{.text} input section. The symbol @samp{_bdata} will be | |
3960 | defined as the address following the @samp{.text} output section aligned | |
3961 | upward to a 4 byte boundary. | |
3962 | ||
eb8476a6 MR |
3963 | @node HIDDEN |
3964 | @subsection HIDDEN | |
3965 | @cindex HIDDEN | |
3966 | For ELF targeted ports, define a symbol that will be hidden and won't be | |
3967 | exported. The syntax is @code{HIDDEN(@var{symbol} = @var{expression})}. | |
3968 | ||
3969 | Here is the example from @ref{Simple Assignments}, rewritten to use | |
3970 | @code{HIDDEN}: | |
3971 | ||
3972 | @smallexample | |
3973 | HIDDEN(floating_point = 0); | |
3974 | SECTIONS | |
3975 | @{ | |
3976 | .text : | |
3977 | @{ | |
3978 | *(.text) | |
3979 | HIDDEN(_etext = .); | |
3980 | @} | |
3981 | HIDDEN(_bdata = (. + 3) & ~ 3); | |
3982 | .data : @{ *(.data) @} | |
3983 | @} | |
3984 | @end smallexample | |
3985 | @noindent | |
3986 | In this case none of the three symbols will be visible outside this module. | |
3987 | ||
252b5132 RH |
3988 | @node PROVIDE |
3989 | @subsection PROVIDE | |
3990 | @cindex PROVIDE | |
3991 | In some cases, it is desirable for a linker script to define a symbol | |
3992 | only if it is referenced and is not defined by any object included in | |
3993 | the link. For example, traditional linkers defined the symbol | |
3994 | @samp{etext}. However, ANSI C requires that the user be able to use | |
3995 | @samp{etext} as a function name without encountering an error. The | |
3996 | @code{PROVIDE} keyword may be used to define a symbol, such as | |
3997 | @samp{etext}, only if it is referenced but not defined. The syntax is | |
3998 | @code{PROVIDE(@var{symbol} = @var{expression})}. | |
3999 | ||
4000 | Here is an example of using @code{PROVIDE} to define @samp{etext}: | |
4001 | @smallexample | |
4002 | SECTIONS | |
4003 | @{ | |
4004 | .text : | |
4005 | @{ | |
4006 | *(.text) | |
4007 | _etext = .; | |
4008 | PROVIDE(etext = .); | |
4009 | @} | |
4010 | @} | |
4011 | @end smallexample | |
4012 | ||
4013 | In this example, if the program defines @samp{_etext} (with a leading | |
4014 | underscore), the linker will give a multiple definition error. If, on | |
4015 | the other hand, the program defines @samp{etext} (with no leading | |
4016 | underscore), the linker will silently use the definition in the program. | |
4017 | If the program references @samp{etext} but does not define it, the | |
4018 | linker will use the definition in the linker script. | |
4019 | ||
b0daac83 NC |
4020 | Note - the @code{PROVIDE} directive considers a common symbol to be |
4021 | defined, even though such a symbol could be combined with the symbol | |
4022 | that the @code{PROVIDE} would create. This is particularly important | |
4023 | when considering constructor and destructor list symbols such as | |
4024 | @samp{__CTOR_LIST__} as these are often defined as common symbols. | |
4025 | ||
7af8e998 L |
4026 | @node PROVIDE_HIDDEN |
4027 | @subsection PROVIDE_HIDDEN | |
4028 | @cindex PROVIDE_HIDDEN | |
4029 | Similar to @code{PROVIDE}. For ELF targeted ports, the symbol will be | |
4030 | hidden and won't be exported. | |
4031 | ||
73ae6183 NC |
4032 | @node Source Code Reference |
4033 | @subsection Source Code Reference | |
4034 | ||
4035 | Accessing a linker script defined variable from source code is not | |
4036 | intuitive. In particular a linker script symbol is not equivalent to | |
4037 | a variable declaration in a high level language, it is instead a | |
4038 | symbol that does not have a value. | |
4039 | ||
4040 | Before going further, it is important to note that compilers often | |
4041 | transform names in the source code into different names when they are | |
4042 | stored in the symbol table. For example, Fortran compilers commonly | |
4043 | prepend or append an underscore, and C++ performs extensive @samp{name | |
4044 | mangling}. Therefore there might be a discrepancy between the name | |
4045 | of a variable as it is used in source code and the name of the same | |
4046 | variable as it is defined in a linker script. For example in C a | |
4047 | linker script variable might be referred to as: | |
4048 | ||
4049 | @smallexample | |
4050 | extern int foo; | |
4051 | @end smallexample | |
4052 | ||
4053 | But in the linker script it might be defined as: | |
4054 | ||
4055 | @smallexample | |
4056 | _foo = 1000; | |
4057 | @end smallexample | |
4058 | ||
4059 | In the remaining examples however it is assumed that no name | |
4060 | transformation has taken place. | |
4061 | ||
4062 | When a symbol is declared in a high level language such as C, two | |
4063 | things happen. The first is that the compiler reserves enough space | |
4064 | in the program's memory to hold the @emph{value} of the symbol. The | |
4065 | second is that the compiler creates an entry in the program's symbol | |
4066 | table which holds the symbol's @emph{address}. ie the symbol table | |
4067 | contains the address of the block of memory holding the symbol's | |
4068 | value. So for example the following C declaration, at file scope: | |
4069 | ||
4070 | @smallexample | |
4071 | int foo = 1000; | |
4072 | @end smallexample | |
4073 | ||
10bf6894 | 4074 | creates an entry called @samp{foo} in the symbol table. This entry |
73ae6183 NC |
4075 | holds the address of an @samp{int} sized block of memory where the |
4076 | number 1000 is initially stored. | |
4077 | ||
4078 | When a program references a symbol the compiler generates code that | |
4079 | first accesses the symbol table to find the address of the symbol's | |
4080 | memory block and then code to read the value from that memory block. | |
4081 | So: | |
4082 | ||
4083 | @smallexample | |
4084 | foo = 1; | |
4085 | @end smallexample | |
4086 | ||
4087 | looks up the symbol @samp{foo} in the symbol table, gets the address | |
4088 | associated with this symbol and then writes the value 1 into that | |
4089 | address. Whereas: | |
4090 | ||
4091 | @smallexample | |
4092 | int * a = & foo; | |
4093 | @end smallexample | |
4094 | ||
10bf6894 | 4095 | looks up the symbol @samp{foo} in the symbol table, gets its address |
73ae6183 NC |
4096 | and then copies this address into the block of memory associated with |
4097 | the variable @samp{a}. | |
4098 | ||
4099 | Linker scripts symbol declarations, by contrast, create an entry in | |
4100 | the symbol table but do not assign any memory to them. Thus they are | |
4101 | an address without a value. So for example the linker script definition: | |
4102 | ||
4103 | @smallexample | |
4104 | foo = 1000; | |
4105 | @end smallexample | |
4106 | ||
4107 | creates an entry in the symbol table called @samp{foo} which holds | |
4108 | the address of memory location 1000, but nothing special is stored at | |
4109 | address 1000. This means that you cannot access the @emph{value} of a | |
4110 | linker script defined symbol - it has no value - all you can do is | |
4111 | access the @emph{address} of a linker script defined symbol. | |
4112 | ||
4113 | Hence when you are using a linker script defined symbol in source code | |
4114 | you should always take the address of the symbol, and never attempt to | |
4115 | use its value. For example suppose you want to copy the contents of a | |
4116 | section of memory called .ROM into a section called .FLASH and the | |
4117 | linker script contains these declarations: | |
4118 | ||
4119 | @smallexample | |
4120 | @group | |
4121 | start_of_ROM = .ROM; | |
a5e406b5 | 4122 | end_of_ROM = .ROM + sizeof (.ROM); |
73ae6183 NC |
4123 | start_of_FLASH = .FLASH; |
4124 | @end group | |
4125 | @end smallexample | |
4126 | ||
4127 | Then the C source code to perform the copy would be: | |
4128 | ||
4129 | @smallexample | |
4130 | @group | |
4131 | extern char start_of_ROM, end_of_ROM, start_of_FLASH; | |
c0065db7 | 4132 | |
73ae6183 NC |
4133 | memcpy (& start_of_FLASH, & start_of_ROM, & end_of_ROM - & start_of_ROM); |
4134 | @end group | |
4135 | @end smallexample | |
4136 | ||
4137 | Note the use of the @samp{&} operators. These are correct. | |
5707d2ad NC |
4138 | Alternatively the symbols can be treated as the names of vectors or |
4139 | arrays and then the code will again work as expected: | |
4140 | ||
4141 | @smallexample | |
4142 | @group | |
4143 | extern char start_of_ROM[], end_of_ROM[], start_of_FLASH[]; | |
4144 | ||
4145 | memcpy (start_of_FLASH, start_of_ROM, end_of_ROM - start_of_ROM); | |
4146 | @end group | |
4147 | @end smallexample | |
4148 | ||
4149 | Note how using this method does not require the use of @samp{&} | |
4150 | operators. | |
73ae6183 | 4151 | |
252b5132 | 4152 | @node SECTIONS |
36f63dca | 4153 | @section SECTIONS Command |
252b5132 RH |
4154 | @kindex SECTIONS |
4155 | The @code{SECTIONS} command tells the linker how to map input sections | |
4156 | into output sections, and how to place the output sections in memory. | |
4157 | ||
4158 | The format of the @code{SECTIONS} command is: | |
4159 | @smallexample | |
4160 | SECTIONS | |
4161 | @{ | |
4162 | @var{sections-command} | |
4163 | @var{sections-command} | |
4164 | @dots{} | |
4165 | @} | |
4166 | @end smallexample | |
4167 | ||
4168 | Each @var{sections-command} may of be one of the following: | |
4169 | ||
4170 | @itemize @bullet | |
4171 | @item | |
4172 | an @code{ENTRY} command (@pxref{Entry Point,,Entry command}) | |
4173 | @item | |
4174 | a symbol assignment (@pxref{Assignments}) | |
4175 | @item | |
4176 | an output section description | |
4177 | @item | |
4178 | an overlay description | |
4179 | @end itemize | |
4180 | ||
4181 | The @code{ENTRY} command and symbol assignments are permitted inside the | |
4182 | @code{SECTIONS} command for convenience in using the location counter in | |
4183 | those commands. This can also make the linker script easier to | |
4184 | understand because you can use those commands at meaningful points in | |
4185 | the layout of the output file. | |
4186 | ||
4187 | Output section descriptions and overlay descriptions are described | |
4188 | below. | |
4189 | ||
4190 | If you do not use a @code{SECTIONS} command in your linker script, the | |
4191 | linker will place each input section into an identically named output | |
4192 | section in the order that the sections are first encountered in the | |
4193 | input files. If all input sections are present in the first file, for | |
4194 | example, the order of sections in the output file will match the order | |
4195 | in the first input file. The first section will be at address zero. | |
4196 | ||
4197 | @menu | |
4198 | * Output Section Description:: Output section description | |
4199 | * Output Section Name:: Output section name | |
4200 | * Output Section Address:: Output section address | |
4201 | * Input Section:: Input section description | |
4202 | * Output Section Data:: Output section data | |
4203 | * Output Section Keywords:: Output section keywords | |
4204 | * Output Section Discarding:: Output section discarding | |
4205 | * Output Section Attributes:: Output section attributes | |
4206 | * Overlay Description:: Overlay description | |
4207 | @end menu | |
4208 | ||
4209 | @node Output Section Description | |
36f63dca | 4210 | @subsection Output Section Description |
252b5132 RH |
4211 | The full description of an output section looks like this: |
4212 | @smallexample | |
a1ab1d2a | 4213 | @group |
7e7d5768 | 4214 | @var{section} [@var{address}] [(@var{type})] : |
0c71d759 | 4215 | [AT(@var{lma})] |
1eec346e | 4216 | [ALIGN(@var{section_align}) | ALIGN_WITH_INPUT] |
0c71d759 NC |
4217 | [SUBALIGN(@var{subsection_align})] |
4218 | [@var{constraint}] | |
252b5132 RH |
4219 | @{ |
4220 | @var{output-section-command} | |
4221 | @var{output-section-command} | |
4222 | @dots{} | |
abc9061b | 4223 | @} [>@var{region}] [AT>@var{lma_region}] [:@var{phdr} :@var{phdr} @dots{}] [=@var{fillexp}] [,] |
252b5132 RH |
4224 | @end group |
4225 | @end smallexample | |
4226 | ||
4227 | Most output sections do not use most of the optional section attributes. | |
4228 | ||
4229 | The whitespace around @var{section} is required, so that the section | |
4230 | name is unambiguous. The colon and the curly braces are also required. | |
abc9061b CC |
4231 | The comma at the end may be required if a @var{fillexp} is used and |
4232 | the next @var{sections-command} looks like a continuation of the expression. | |
252b5132 RH |
4233 | The line breaks and other white space are optional. |
4234 | ||
4235 | Each @var{output-section-command} may be one of the following: | |
4236 | ||
4237 | @itemize @bullet | |
4238 | @item | |
4239 | a symbol assignment (@pxref{Assignments}) | |
4240 | @item | |
4241 | an input section description (@pxref{Input Section}) | |
4242 | @item | |
4243 | data values to include directly (@pxref{Output Section Data}) | |
4244 | @item | |
4245 | a special output section keyword (@pxref{Output Section Keywords}) | |
4246 | @end itemize | |
4247 | ||
4248 | @node Output Section Name | |
36f63dca | 4249 | @subsection Output Section Name |
252b5132 RH |
4250 | @cindex name, section |
4251 | @cindex section name | |
4252 | The name of the output section is @var{section}. @var{section} must | |
4253 | meet the constraints of your output format. In formats which only | |
4254 | support a limited number of sections, such as @code{a.out}, the name | |
4255 | must be one of the names supported by the format (@code{a.out}, for | |
4256 | example, allows only @samp{.text}, @samp{.data} or @samp{.bss}). If the | |
4257 | output format supports any number of sections, but with numbers and not | |
4258 | names (as is the case for Oasys), the name should be supplied as a | |
4259 | quoted numeric string. A section name may consist of any sequence of | |
4260 | characters, but a name which contains any unusual characters such as | |
4261 | commas must be quoted. | |
4262 | ||
4263 | The output section name @samp{/DISCARD/} is special; @ref{Output Section | |
4264 | Discarding}. | |
4265 | ||
4266 | @node Output Section Address | |
2a16d82a | 4267 | @subsection Output Section Address |
252b5132 RH |
4268 | @cindex address, section |
4269 | @cindex section address | |
4270 | The @var{address} is an expression for the VMA (the virtual memory | |
ea5cae92 NC |
4271 | address) of the output section. This address is optional, but if it |
4272 | is provided then the output address will be set exactly as specified. | |
4273 | ||
4274 | If the output address is not specified then one will be chosen for the | |
4275 | section, based on the heuristic below. This address will be adjusted | |
4276 | to fit the alignment requirement of the output section. The | |
4277 | alignment requirement is the strictest alignment of any input section | |
4278 | contained within the output section. | |
4279 | ||
4280 | The output section address heuristic is as follows: | |
4281 | ||
4282 | @itemize @bullet | |
4283 | @item | |
4284 | If an output memory @var{region} is set for the section then it | |
4285 | is added to this region and its address will be the next free address | |
4286 | in that region. | |
4287 | ||
4288 | @item | |
4289 | If the MEMORY command has been used to create a list of memory | |
4290 | regions then the first region which has attributes compatible with the | |
4291 | section is selected to contain it. The section's output address will | |
4292 | be the next free address in that region; @ref{MEMORY}. | |
4293 | ||
4294 | @item | |
4295 | If no memory regions were specified, or none match the section then | |
4296 | the output address will be based on the current value of the location | |
4297 | counter. | |
4298 | @end itemize | |
4299 | ||
4300 | @noindent | |
4301 | For example: | |
4302 | ||
252b5132 RH |
4303 | @smallexample |
4304 | .text . : @{ *(.text) @} | |
4305 | @end smallexample | |
ea5cae92 | 4306 | |
252b5132 RH |
4307 | @noindent |
4308 | and | |
ea5cae92 | 4309 | |
252b5132 RH |
4310 | @smallexample |
4311 | .text : @{ *(.text) @} | |
4312 | @end smallexample | |
ea5cae92 | 4313 | |
252b5132 RH |
4314 | @noindent |
4315 | are subtly different. The first will set the address of the | |
4316 | @samp{.text} output section to the current value of the location | |
4317 | counter. The second will set it to the current value of the location | |
ea5cae92 NC |
4318 | counter aligned to the strictest alignment of any of the @samp{.text} |
4319 | input sections. | |
252b5132 RH |
4320 | |
4321 | The @var{address} may be an arbitrary expression; @ref{Expressions}. | |
4322 | For example, if you want to align the section on a 0x10 byte boundary, | |
4323 | so that the lowest four bits of the section address are zero, you could | |
4324 | do something like this: | |
4325 | @smallexample | |
4326 | .text ALIGN(0x10) : @{ *(.text) @} | |
4327 | @end smallexample | |
4328 | @noindent | |
4329 | This works because @code{ALIGN} returns the current location counter | |
4330 | aligned upward to the specified value. | |
4331 | ||
4332 | Specifying @var{address} for a section will change the value of the | |
6ce340f1 NC |
4333 | location counter, provided that the section is non-empty. (Empty |
4334 | sections are ignored). | |
252b5132 RH |
4335 | |
4336 | @node Input Section | |
36f63dca | 4337 | @subsection Input Section Description |
252b5132 RH |
4338 | @cindex input sections |
4339 | @cindex mapping input sections to output sections | |
4340 | The most common output section command is an input section description. | |
4341 | ||
4342 | The input section description is the most basic linker script operation. | |
4343 | You use output sections to tell the linker how to lay out your program | |
4344 | in memory. You use input section descriptions to tell the linker how to | |
4345 | map the input files into your memory layout. | |
4346 | ||
4347 | @menu | |
4348 | * Input Section Basics:: Input section basics | |
4349 | * Input Section Wildcards:: Input section wildcard patterns | |
4350 | * Input Section Common:: Input section for common symbols | |
4351 | * Input Section Keep:: Input section and garbage collection | |
4352 | * Input Section Example:: Input section example | |
4353 | @end menu | |
4354 | ||
4355 | @node Input Section Basics | |
36f63dca | 4356 | @subsubsection Input Section Basics |
252b5132 RH |
4357 | @cindex input section basics |
4358 | An input section description consists of a file name optionally followed | |
4359 | by a list of section names in parentheses. | |
4360 | ||
4361 | The file name and the section name may be wildcard patterns, which we | |
4362 | describe further below (@pxref{Input Section Wildcards}). | |
4363 | ||
4364 | The most common input section description is to include all input | |
4365 | sections with a particular name in the output section. For example, to | |
4366 | include all input @samp{.text} sections, you would write: | |
4367 | @smallexample | |
4368 | *(.text) | |
4369 | @end smallexample | |
4370 | @noindent | |
18625d54 | 4371 | Here the @samp{*} is a wildcard which matches any file name. To exclude a list |
8f1732fc | 4372 | @cindex EXCLUDE_FILE |
18625d54 CM |
4373 | of files from matching the file name wildcard, EXCLUDE_FILE may be used to |
4374 | match all files except the ones specified in the EXCLUDE_FILE list. For | |
4375 | example: | |
252b5132 | 4376 | @smallexample |
8f1732fc AB |
4377 | EXCLUDE_FILE (*crtend.o *otherfile.o) *(.ctors) |
4378 | @end smallexample | |
4379 | @noindent | |
4380 | will cause all .ctors sections from all files except @file{crtend.o} | |
4381 | and @file{otherfile.o} to be included. The EXCLUDE_FILE can also be | |
4382 | placed inside the section list, for example: | |
4383 | @smallexample | |
b4346c09 | 4384 | *(EXCLUDE_FILE (*crtend.o *otherfile.o) .ctors) |
252b5132 | 4385 | @end smallexample |
8f1732fc AB |
4386 | @noindent |
4387 | The result of this is identically to the previous example. Supporting | |
4388 | two syntaxes for EXCLUDE_FILE is useful if the section list contains | |
4389 | more than one section, as described below. | |
252b5132 RH |
4390 | |
4391 | There are two ways to include more than one section: | |
4392 | @smallexample | |
4393 | *(.text .rdata) | |
4394 | *(.text) *(.rdata) | |
4395 | @end smallexample | |
4396 | @noindent | |
4397 | The difference between these is the order in which the @samp{.text} and | |
4398 | @samp{.rdata} input sections will appear in the output section. In the | |
b6bf44ba AM |
4399 | first example, they will be intermingled, appearing in the same order as |
4400 | they are found in the linker input. In the second example, all | |
252b5132 RH |
4401 | @samp{.text} input sections will appear first, followed by all |
4402 | @samp{.rdata} input sections. | |
4403 | ||
8f1732fc AB |
4404 | When using EXCLUDE_FILE with more than one section, if the exclusion |
4405 | is within the section list then the exclusion only applies to the | |
4406 | immediately following section, for example: | |
a5bf7d4f AB |
4407 | @smallexample |
4408 | *(EXCLUDE_FILE (*somefile.o) .text .rdata) | |
4409 | @end smallexample | |
4410 | @noindent | |
4411 | will cause all @samp{.text} sections from all files except | |
4412 | @file{somefile.o} to be included, while all @samp{.rdata} sections | |
4413 | from all files, including @file{somefile.o}, will be included. To | |
4414 | exclude the @samp{.rdata} sections from @file{somefile.o} the example | |
8f1732fc | 4415 | could be modified to: |
a5bf7d4f AB |
4416 | @smallexample |
4417 | *(EXCLUDE_FILE (*somefile.o) .text EXCLUDE_FILE (*somefile.o) .rdata) | |
4418 | @end smallexample | |
8f1732fc AB |
4419 | @noindent |
4420 | Alternatively, placing the EXCLUDE_FILE outside of the section list, | |
4421 | before the input file selection, will cause the exclusion to apply for | |
4422 | all sections. Thus the previous example can be rewritten as: | |
4423 | @smallexample | |
4424 | EXCLUDE_FILE (*somefile.o) *(.text .rdata) | |
4425 | @end smallexample | |
a5bf7d4f | 4426 | |
252b5132 RH |
4427 | You can specify a file name to include sections from a particular file. |
4428 | You would do this if one or more of your files contain special data that | |
4429 | needs to be at a particular location in memory. For example: | |
4430 | @smallexample | |
4431 | data.o(.data) | |
4432 | @end smallexample | |
4433 | ||
ae17ab41 CM |
4434 | To refine the sections that are included based on the section flags |
4435 | of an input section, INPUT_SECTION_FLAGS may be used. | |
4436 | ||
4437 | Here is a simple example for using Section header flags for ELF sections: | |
4438 | ||
4439 | @smallexample | |
4440 | @group | |
4441 | SECTIONS @{ | |
4442 | .text : @{ INPUT_SECTION_FLAGS (SHF_MERGE & SHF_STRINGS) *(.text) @} | |
4443 | .text2 : @{ INPUT_SECTION_FLAGS (!SHF_WRITE) *(.text) @} | |
4444 | @} | |
4445 | @end group | |
4446 | @end smallexample | |
4447 | ||
4448 | In this example, the output section @samp{.text} will be comprised of any | |
4449 | input section matching the name *(.text) whose section header flags | |
4450 | @code{SHF_MERGE} and @code{SHF_STRINGS} are set. The output section | |
4451 | @samp{.text2} will be comprised of any input section matching the name *(.text) | |
4452 | whose section header flag @code{SHF_WRITE} is clear. | |
4453 | ||
967928e9 AM |
4454 | You can also specify files within archives by writing a pattern |
4455 | matching the archive, a colon, then the pattern matching the file, | |
4456 | with no whitespace around the colon. | |
4457 | ||
4458 | @table @samp | |
4459 | @item archive:file | |
4460 | matches file within archive | |
4461 | @item archive: | |
4462 | matches the whole archive | |
4463 | @item :file | |
4464 | matches file but not one in an archive | |
4465 | @end table | |
4466 | ||
4467 | Either one or both of @samp{archive} and @samp{file} can contain shell | |
4468 | wildcards. On DOS based file systems, the linker will assume that a | |
4469 | single letter followed by a colon is a drive specifier, so | |
4470 | @samp{c:myfile.o} is a simple file specification, not @samp{myfile.o} | |
4471 | within an archive called @samp{c}. @samp{archive:file} filespecs may | |
4472 | also be used within an @code{EXCLUDE_FILE} list, but may not appear in | |
4473 | other linker script contexts. For instance, you cannot extract a file | |
4474 | from an archive by using @samp{archive:file} in an @code{INPUT} | |
4475 | command. | |
4476 | ||
252b5132 RH |
4477 | If you use a file name without a list of sections, then all sections in |
4478 | the input file will be included in the output section. This is not | |
4479 | commonly done, but it may by useful on occasion. For example: | |
4480 | @smallexample | |
4481 | data.o | |
4482 | @end smallexample | |
4483 | ||
967928e9 AM |
4484 | When you use a file name which is not an @samp{archive:file} specifier |
4485 | and does not contain any wild card | |
252b5132 RH |
4486 | characters, the linker will first see if you also specified the file |
4487 | name on the linker command line or in an @code{INPUT} command. If you | |
4488 | did not, the linker will attempt to open the file as an input file, as | |
4489 | though it appeared on the command line. Note that this differs from an | |
4490 | @code{INPUT} command, because the linker will not search for the file in | |
4491 | the archive search path. | |
4492 | ||
4493 | @node Input Section Wildcards | |
36f63dca | 4494 | @subsubsection Input Section Wildcard Patterns |
252b5132 RH |
4495 | @cindex input section wildcards |
4496 | @cindex wildcard file name patterns | |
4497 | @cindex file name wildcard patterns | |
4498 | @cindex section name wildcard patterns | |
4499 | In an input section description, either the file name or the section | |
4500 | name or both may be wildcard patterns. | |
4501 | ||
4502 | The file name of @samp{*} seen in many examples is a simple wildcard | |
4503 | pattern for the file name. | |
4504 | ||
4505 | The wildcard patterns are like those used by the Unix shell. | |
4506 | ||
4507 | @table @samp | |
4508 | @item * | |
4509 | matches any number of characters | |
4510 | @item ? | |
4511 | matches any single character | |
4512 | @item [@var{chars}] | |
4513 | matches a single instance of any of the @var{chars}; the @samp{-} | |
4514 | character may be used to specify a range of characters, as in | |
4515 | @samp{[a-z]} to match any lower case letter | |
4516 | @item \ | |
4517 | quotes the following character | |
4518 | @end table | |
4519 | ||
4520 | When a file name is matched with a wildcard, the wildcard characters | |
4521 | will not match a @samp{/} character (used to separate directory names on | |
4522 | Unix). A pattern consisting of a single @samp{*} character is an | |
4523 | exception; it will always match any file name, whether it contains a | |
4524 | @samp{/} or not. In a section name, the wildcard characters will match | |
4525 | a @samp{/} character. | |
4526 | ||
4527 | File name wildcard patterns only match files which are explicitly | |
4528 | specified on the command line or in an @code{INPUT} command. The linker | |
4529 | does not search directories to expand wildcards. | |
4530 | ||
4531 | If a file name matches more than one wildcard pattern, or if a file name | |
4532 | appears explicitly and is also matched by a wildcard pattern, the linker | |
4533 | will use the first match in the linker script. For example, this | |
4534 | sequence of input section descriptions is probably in error, because the | |
4535 | @file{data.o} rule will not be used: | |
4536 | @smallexample | |
4537 | .data : @{ *(.data) @} | |
4538 | .data1 : @{ data.o(.data) @} | |
4539 | @end smallexample | |
4540 | ||
bcaa7b3e | 4541 | @cindex SORT_BY_NAME |
252b5132 RH |
4542 | Normally, the linker will place files and sections matched by wildcards |
4543 | in the order in which they are seen during the link. You can change | |
bcaa7b3e L |
4544 | this by using the @code{SORT_BY_NAME} keyword, which appears before a wildcard |
4545 | pattern in parentheses (e.g., @code{SORT_BY_NAME(.text*)}). When the | |
4546 | @code{SORT_BY_NAME} keyword is used, the linker will sort the files or sections | |
252b5132 RH |
4547 | into ascending order by name before placing them in the output file. |
4548 | ||
bcaa7b3e L |
4549 | @cindex SORT_BY_ALIGNMENT |
4550 | @code{SORT_BY_ALIGNMENT} is very similar to @code{SORT_BY_NAME}. The | |
4551 | difference is @code{SORT_BY_ALIGNMENT} will sort sections into | |
1ae5c3ae | 4552 | descending order by alignment before placing them in the output file. |
ee83b8a6 NC |
4553 | Larger alignments are placed before smaller alignments in order to |
4554 | reduce the amount of padding necessary. | |
bcaa7b3e | 4555 | |
02ecc8e9 L |
4556 | @cindex SORT_BY_INIT_PRIORITY |
4557 | @code{SORT_BY_INIT_PRIORITY} is very similar to @code{SORT_BY_NAME}. The | |
4558 | difference is @code{SORT_BY_INIT_PRIORITY} will sort sections into | |
4559 | ascending order by numerical value of the GCC init_priority attribute | |
4560 | encoded in the section name before placing them in the output file. | |
4561 | ||
bcaa7b3e L |
4562 | @cindex SORT |
4563 | @code{SORT} is an alias for @code{SORT_BY_NAME}. | |
4564 | ||
4565 | When there are nested section sorting commands in linker script, there | |
4566 | can be at most 1 level of nesting for section sorting commands. | |
4567 | ||
4568 | @enumerate | |
4569 | @item | |
4570 | @code{SORT_BY_NAME} (@code{SORT_BY_ALIGNMENT} (wildcard section pattern)). | |
1ae5c3ae | 4571 | It will sort the input sections by name first, then by alignment if two |
bcaa7b3e L |
4572 | sections have the same name. |
4573 | @item | |
4574 | @code{SORT_BY_ALIGNMENT} (@code{SORT_BY_NAME} (wildcard section pattern)). | |
1ae5c3ae | 4575 | It will sort the input sections by alignment first, then by name if two |
bcaa7b3e L |
4576 | sections have the same alignment. |
4577 | @item | |
c0065db7 | 4578 | @code{SORT_BY_NAME} (@code{SORT_BY_NAME} (wildcard section pattern)) is |
bcaa7b3e L |
4579 | treated the same as @code{SORT_BY_NAME} (wildcard section pattern). |
4580 | @item | |
4581 | @code{SORT_BY_ALIGNMENT} (@code{SORT_BY_ALIGNMENT} (wildcard section pattern)) | |
4582 | is treated the same as @code{SORT_BY_ALIGNMENT} (wildcard section pattern). | |
4583 | @item | |
4584 | All other nested section sorting commands are invalid. | |
4585 | @end enumerate | |
4586 | ||
a05a5b64 | 4587 | When both command-line section sorting option and linker script |
bcaa7b3e | 4588 | section sorting command are used, section sorting command always |
a05a5b64 | 4589 | takes precedence over the command-line option. |
bcaa7b3e L |
4590 | |
4591 | If the section sorting command in linker script isn't nested, the | |
a05a5b64 | 4592 | command-line option will make the section sorting command to be |
bcaa7b3e L |
4593 | treated as nested sorting command. |
4594 | ||
4595 | @enumerate | |
4596 | @item | |
4597 | @code{SORT_BY_NAME} (wildcard section pattern ) with | |
4598 | @option{--sort-sections alignment} is equivalent to | |
4599 | @code{SORT_BY_NAME} (@code{SORT_BY_ALIGNMENT} (wildcard section pattern)). | |
4600 | @item | |
4601 | @code{SORT_BY_ALIGNMENT} (wildcard section pattern) with | |
4602 | @option{--sort-section name} is equivalent to | |
4603 | @code{SORT_BY_ALIGNMENT} (@code{SORT_BY_NAME} (wildcard section pattern)). | |
4604 | @end enumerate | |
4605 | ||
4606 | If the section sorting command in linker script is nested, the | |
a05a5b64 | 4607 | command-line option will be ignored. |
bcaa7b3e | 4608 | |
eda680f8 | 4609 | @cindex SORT_NONE |
a05a5b64 | 4610 | @code{SORT_NONE} disables section sorting by ignoring the command-line |
eda680f8 L |
4611 | section sorting option. |
4612 | ||
252b5132 RH |
4613 | If you ever get confused about where input sections are going, use the |
4614 | @samp{-M} linker option to generate a map file. The map file shows | |
4615 | precisely how input sections are mapped to output sections. | |
4616 | ||
4617 | This example shows how wildcard patterns might be used to partition | |
4618 | files. This linker script directs the linker to place all @samp{.text} | |
4619 | sections in @samp{.text} and all @samp{.bss} sections in @samp{.bss}. | |
4620 | The linker will place the @samp{.data} section from all files beginning | |
4621 | with an upper case character in @samp{.DATA}; for all other files, the | |
4622 | linker will place the @samp{.data} section in @samp{.data}. | |
4623 | @smallexample | |
4624 | @group | |
4625 | SECTIONS @{ | |
4626 | .text : @{ *(.text) @} | |
4627 | .DATA : @{ [A-Z]*(.data) @} | |
4628 | .data : @{ *(.data) @} | |
4629 | .bss : @{ *(.bss) @} | |
4630 | @} | |
4631 | @end group | |
4632 | @end smallexample | |
4633 | ||
4634 | @node Input Section Common | |
36f63dca | 4635 | @subsubsection Input Section for Common Symbols |
252b5132 RH |
4636 | @cindex common symbol placement |
4637 | @cindex uninitialized data placement | |
4638 | A special notation is needed for common symbols, because in many object | |
4639 | file formats common symbols do not have a particular input section. The | |
4640 | linker treats common symbols as though they are in an input section | |
4641 | named @samp{COMMON}. | |
4642 | ||
4643 | You may use file names with the @samp{COMMON} section just as with any | |
4644 | other input sections. You can use this to place common symbols from a | |
4645 | particular input file in one section while common symbols from other | |
4646 | input files are placed in another section. | |
4647 | ||
4648 | In most cases, common symbols in input files will be placed in the | |
4649 | @samp{.bss} section in the output file. For example: | |
4650 | @smallexample | |
4651 | .bss @{ *(.bss) *(COMMON) @} | |
4652 | @end smallexample | |
4653 | ||
4654 | @cindex scommon section | |
4655 | @cindex small common symbols | |
4656 | Some object file formats have more than one type of common symbol. For | |
4657 | example, the MIPS ELF object file format distinguishes standard common | |
4658 | symbols and small common symbols. In this case, the linker will use a | |
4659 | different special section name for other types of common symbols. In | |
4660 | the case of MIPS ELF, the linker uses @samp{COMMON} for standard common | |
4661 | symbols and @samp{.scommon} for small common symbols. This permits you | |
4662 | to map the different types of common symbols into memory at different | |
4663 | locations. | |
4664 | ||
4665 | @cindex [COMMON] | |
4666 | You will sometimes see @samp{[COMMON]} in old linker scripts. This | |
4667 | notation is now considered obsolete. It is equivalent to | |
4668 | @samp{*(COMMON)}. | |
4669 | ||
4670 | @node Input Section Keep | |
36f63dca | 4671 | @subsubsection Input Section and Garbage Collection |
252b5132 RH |
4672 | @cindex KEEP |
4673 | @cindex garbage collection | |
4674 | When link-time garbage collection is in use (@samp{--gc-sections}), | |
a1ab1d2a | 4675 | it is often useful to mark sections that should not be eliminated. |
252b5132 RH |
4676 | This is accomplished by surrounding an input section's wildcard entry |
4677 | with @code{KEEP()}, as in @code{KEEP(*(.init))} or | |
bcaa7b3e | 4678 | @code{KEEP(SORT_BY_NAME(*)(.ctors))}. |
252b5132 RH |
4679 | |
4680 | @node Input Section Example | |
36f63dca | 4681 | @subsubsection Input Section Example |
252b5132 RH |
4682 | The following example is a complete linker script. It tells the linker |
4683 | to read all of the sections from file @file{all.o} and place them at the | |
4684 | start of output section @samp{outputa} which starts at location | |
4685 | @samp{0x10000}. All of section @samp{.input1} from file @file{foo.o} | |
4686 | follows immediately, in the same output section. All of section | |
4687 | @samp{.input2} from @file{foo.o} goes into output section | |
4688 | @samp{outputb}, followed by section @samp{.input1} from @file{foo1.o}. | |
4689 | All of the remaining @samp{.input1} and @samp{.input2} sections from any | |
4690 | files are written to output section @samp{outputc}. | |
4691 | ||
4692 | @smallexample | |
4693 | @group | |
4694 | SECTIONS @{ | |
4695 | outputa 0x10000 : | |
4696 | @{ | |
4697 | all.o | |
4698 | foo.o (.input1) | |
4699 | @} | |
36f63dca NC |
4700 | @end group |
4701 | @group | |
252b5132 RH |
4702 | outputb : |
4703 | @{ | |
4704 | foo.o (.input2) | |
4705 | foo1.o (.input1) | |
4706 | @} | |
36f63dca NC |
4707 | @end group |
4708 | @group | |
252b5132 RH |
4709 | outputc : |
4710 | @{ | |
4711 | *(.input1) | |
4712 | *(.input2) | |
4713 | @} | |
4714 | @} | |
4715 | @end group | |
a1ab1d2a | 4716 | @end smallexample |
252b5132 | 4717 | |
cbd0eecf L |
4718 | If an output section's name is the same as the input section's name |
4719 | and is representable as a C identifier, then the linker will | |
4720 | automatically @pxref{PROVIDE} two symbols: __start_SECNAME and | |
4721 | __stop_SECNAME, where SECNAME is the name of the section. These | |
4722 | indicate the start address and end address of the output section | |
4723 | respectively. Note: most section names are not representable as | |
4724 | C identifiers because they contain a @samp{.} character. | |
4725 | ||
252b5132 | 4726 | @node Output Section Data |
36f63dca | 4727 | @subsection Output Section Data |
252b5132 RH |
4728 | @cindex data |
4729 | @cindex section data | |
4730 | @cindex output section data | |
4731 | @kindex BYTE(@var{expression}) | |
4732 | @kindex SHORT(@var{expression}) | |
4733 | @kindex LONG(@var{expression}) | |
4734 | @kindex QUAD(@var{expression}) | |
4735 | @kindex SQUAD(@var{expression}) | |
4736 | You can include explicit bytes of data in an output section by using | |
4737 | @code{BYTE}, @code{SHORT}, @code{LONG}, @code{QUAD}, or @code{SQUAD} as | |
4738 | an output section command. Each keyword is followed by an expression in | |
4739 | parentheses providing the value to store (@pxref{Expressions}). The | |
4740 | value of the expression is stored at the current value of the location | |
4741 | counter. | |
4742 | ||
4743 | The @code{BYTE}, @code{SHORT}, @code{LONG}, and @code{QUAD} commands | |
4744 | store one, two, four, and eight bytes (respectively). After storing the | |
4745 | bytes, the location counter is incremented by the number of bytes | |
4746 | stored. | |
4747 | ||
4748 | For example, this will store the byte 1 followed by the four byte value | |
4749 | of the symbol @samp{addr}: | |
4750 | @smallexample | |
4751 | BYTE(1) | |
4752 | LONG(addr) | |
4753 | @end smallexample | |
4754 | ||
4755 | When using a 64 bit host or target, @code{QUAD} and @code{SQUAD} are the | |
4756 | same; they both store an 8 byte, or 64 bit, value. When both host and | |
4757 | target are 32 bits, an expression is computed as 32 bits. In this case | |
4758 | @code{QUAD} stores a 32 bit value zero extended to 64 bits, and | |
4759 | @code{SQUAD} stores a 32 bit value sign extended to 64 bits. | |
4760 | ||
4761 | If the object file format of the output file has an explicit endianness, | |
4762 | which is the normal case, the value will be stored in that endianness. | |
4763 | When the object file format does not have an explicit endianness, as is | |
4764 | true of, for example, S-records, the value will be stored in the | |
4765 | endianness of the first input object file. | |
4766 | ||
36f63dca | 4767 | Note---these commands only work inside a section description and not |
2b5fc1f5 NC |
4768 | between them, so the following will produce an error from the linker: |
4769 | @smallexample | |
4770 | SECTIONS @{@ .text : @{@ *(.text) @}@ LONG(1) .data : @{@ *(.data) @}@ @}@ | |
4771 | @end smallexample | |
4772 | whereas this will work: | |
4773 | @smallexample | |
4774 | SECTIONS @{@ .text : @{@ *(.text) ; LONG(1) @}@ .data : @{@ *(.data) @}@ @}@ | |
4775 | @end smallexample | |
4776 | ||
252b5132 RH |
4777 | @kindex FILL(@var{expression}) |
4778 | @cindex holes, filling | |
4779 | @cindex unspecified memory | |
4780 | You may use the @code{FILL} command to set the fill pattern for the | |
4781 | current section. It is followed by an expression in parentheses. Any | |
4782 | otherwise unspecified regions of memory within the section (for example, | |
4783 | gaps left due to the required alignment of input sections) are filled | |
a139d329 | 4784 | with the value of the expression, repeated as |
252b5132 RH |
4785 | necessary. A @code{FILL} statement covers memory locations after the |
4786 | point at which it occurs in the section definition; by including more | |
4787 | than one @code{FILL} statement, you can have different fill patterns in | |
4788 | different parts of an output section. | |
4789 | ||
4790 | This example shows how to fill unspecified regions of memory with the | |
563e308f | 4791 | value @samp{0x90}: |
252b5132 | 4792 | @smallexample |
563e308f | 4793 | FILL(0x90909090) |
252b5132 RH |
4794 | @end smallexample |
4795 | ||
4796 | The @code{FILL} command is similar to the @samp{=@var{fillexp}} output | |
9673c93c | 4797 | section attribute, but it only affects the |
252b5132 RH |
4798 | part of the section following the @code{FILL} command, rather than the |
4799 | entire section. If both are used, the @code{FILL} command takes | |
9673c93c | 4800 | precedence. @xref{Output Section Fill}, for details on the fill |
a139d329 | 4801 | expression. |
252b5132 RH |
4802 | |
4803 | @node Output Section Keywords | |
36f63dca | 4804 | @subsection Output Section Keywords |
252b5132 RH |
4805 | There are a couple of keywords which can appear as output section |
4806 | commands. | |
4807 | ||
4808 | @table @code | |
4809 | @kindex CREATE_OBJECT_SYMBOLS | |
4810 | @cindex input filename symbols | |
4811 | @cindex filename symbols | |
4812 | @item CREATE_OBJECT_SYMBOLS | |
4813 | The command tells the linker to create a symbol for each input file. | |
4814 | The name of each symbol will be the name of the corresponding input | |
4815 | file. The section of each symbol will be the output section in which | |
4816 | the @code{CREATE_OBJECT_SYMBOLS} command appears. | |
4817 | ||
4818 | This is conventional for the a.out object file format. It is not | |
4819 | normally used for any other object file format. | |
4820 | ||
4821 | @kindex CONSTRUCTORS | |
4822 | @cindex C++ constructors, arranging in link | |
4823 | @cindex constructors, arranging in link | |
4824 | @item CONSTRUCTORS | |
4825 | When linking using the a.out object file format, the linker uses an | |
4826 | unusual set construct to support C++ global constructors and | |
4827 | destructors. When linking object file formats which do not support | |
4828 | arbitrary sections, such as ECOFF and XCOFF, the linker will | |
4829 | automatically recognize C++ global constructors and destructors by name. | |
4830 | For these object file formats, the @code{CONSTRUCTORS} command tells the | |
4831 | linker to place constructor information in the output section where the | |
4832 | @code{CONSTRUCTORS} command appears. The @code{CONSTRUCTORS} command is | |
4833 | ignored for other object file formats. | |
4834 | ||
4835 | The symbol @w{@code{__CTOR_LIST__}} marks the start of the global | |
7e69709c AM |
4836 | constructors, and the symbol @w{@code{__CTOR_END__}} marks the end. |
4837 | Similarly, @w{@code{__DTOR_LIST__}} and @w{@code{__DTOR_END__}} mark | |
4838 | the start and end of the global destructors. The | |
252b5132 RH |
4839 | first word in the list is the number of entries, followed by the address |
4840 | of each constructor or destructor, followed by a zero word. The | |
4841 | compiler must arrange to actually run the code. For these object file | |
4842 | formats @sc{gnu} C++ normally calls constructors from a subroutine | |
4843 | @code{__main}; a call to @code{__main} is automatically inserted into | |
4844 | the startup code for @code{main}. @sc{gnu} C++ normally runs | |
4845 | destructors either by using @code{atexit}, or directly from the function | |
4846 | @code{exit}. | |
4847 | ||
4848 | For object file formats such as @code{COFF} or @code{ELF} which support | |
4849 | arbitrary section names, @sc{gnu} C++ will normally arrange to put the | |
4850 | addresses of global constructors and destructors into the @code{.ctors} | |
4851 | and @code{.dtors} sections. Placing the following sequence into your | |
4852 | linker script will build the sort of table which the @sc{gnu} C++ | |
4853 | runtime code expects to see. | |
4854 | ||
4855 | @smallexample | |
4856 | __CTOR_LIST__ = .; | |
4857 | LONG((__CTOR_END__ - __CTOR_LIST__) / 4 - 2) | |
4858 | *(.ctors) | |
4859 | LONG(0) | |
4860 | __CTOR_END__ = .; | |
4861 | __DTOR_LIST__ = .; | |
4862 | LONG((__DTOR_END__ - __DTOR_LIST__) / 4 - 2) | |
4863 | *(.dtors) | |
4864 | LONG(0) | |
4865 | __DTOR_END__ = .; | |
4866 | @end smallexample | |
4867 | ||
4868 | If you are using the @sc{gnu} C++ support for initialization priority, | |
4869 | which provides some control over the order in which global constructors | |
4870 | are run, you must sort the constructors at link time to ensure that they | |
4871 | are executed in the correct order. When using the @code{CONSTRUCTORS} | |
bcaa7b3e L |
4872 | command, use @samp{SORT_BY_NAME(CONSTRUCTORS)} instead. When using the |
4873 | @code{.ctors} and @code{.dtors} sections, use @samp{*(SORT_BY_NAME(.ctors))} and | |
4874 | @samp{*(SORT_BY_NAME(.dtors))} instead of just @samp{*(.ctors)} and | |
252b5132 RH |
4875 | @samp{*(.dtors)}. |
4876 | ||
4877 | Normally the compiler and linker will handle these issues automatically, | |
4878 | and you will not need to concern yourself with them. However, you may | |
4879 | need to consider this if you are using C++ and writing your own linker | |
4880 | scripts. | |
4881 | ||
4882 | @end table | |
4883 | ||
4884 | @node Output Section Discarding | |
36f63dca | 4885 | @subsection Output Section Discarding |
252b5132 RH |
4886 | @cindex discarding sections |
4887 | @cindex sections, discarding | |
4888 | @cindex removing sections | |
2edab91c AM |
4889 | The linker will not normally create output sections with no contents. |
4890 | This is for convenience when referring to input sections that may or | |
4891 | may not be present in any of the input files. For example: | |
252b5132 | 4892 | @smallexample |
49c13adb | 4893 | .foo : @{ *(.foo) @} |
252b5132 RH |
4894 | @end smallexample |
4895 | @noindent | |
4896 | will only create a @samp{.foo} section in the output file if there is a | |
74541ad4 AM |
4897 | @samp{.foo} section in at least one input file, and if the input |
4898 | sections are not all empty. Other link script directives that allocate | |
2edab91c AM |
4899 | space in an output section will also create the output section. So |
4900 | too will assignments to dot even if the assignment does not create | |
4901 | space, except for @samp{. = 0}, @samp{. = . + 0}, @samp{. = sym}, | |
4902 | @samp{. = . + sym} and @samp{. = ALIGN (. != 0, expr, 1)} when | |
4903 | @samp{sym} is an absolute symbol of value 0 defined in the script. | |
4904 | This allows you to force output of an empty section with @samp{. = .}. | |
74541ad4 | 4905 | |
a0976ea4 | 4906 | The linker will ignore address assignments (@pxref{Output Section Address}) |
74541ad4 AM |
4907 | on discarded output sections, except when the linker script defines |
4908 | symbols in the output section. In that case the linker will obey | |
a0976ea4 AM |
4909 | the address assignments, possibly advancing dot even though the |
4910 | section is discarded. | |
252b5132 RH |
4911 | |
4912 | @cindex /DISCARD/ | |
4913 | The special output section name @samp{/DISCARD/} may be used to discard | |
4914 | input sections. Any input sections which are assigned to an output | |
4915 | section named @samp{/DISCARD/} are not included in the output file. | |
4916 | ||
4917 | @node Output Section Attributes | |
36f63dca | 4918 | @subsection Output Section Attributes |
252b5132 RH |
4919 | @cindex output section attributes |
4920 | We showed above that the full description of an output section looked | |
4921 | like this: | |
0c71d759 | 4922 | |
252b5132 | 4923 | @smallexample |
a1ab1d2a | 4924 | @group |
7e7d5768 | 4925 | @var{section} [@var{address}] [(@var{type})] : |
0c71d759 NC |
4926 | [AT(@var{lma})] |
4927 | [ALIGN(@var{section_align})] | |
4928 | [SUBALIGN(@var{subsection_align})] | |
4929 | [@var{constraint}] | |
252b5132 RH |
4930 | @{ |
4931 | @var{output-section-command} | |
4932 | @var{output-section-command} | |
4933 | @dots{} | |
562d3460 | 4934 | @} [>@var{region}] [AT>@var{lma_region}] [:@var{phdr} :@var{phdr} @dots{}] [=@var{fillexp}] |
252b5132 RH |
4935 | @end group |
4936 | @end smallexample | |
0c71d759 | 4937 | |
252b5132 RH |
4938 | We've already described @var{section}, @var{address}, and |
4939 | @var{output-section-command}. In this section we will describe the | |
4940 | remaining section attributes. | |
4941 | ||
a1ab1d2a | 4942 | @menu |
252b5132 RH |
4943 | * Output Section Type:: Output section type |
4944 | * Output Section LMA:: Output section LMA | |
bbf115d3 | 4945 | * Forced Output Alignment:: Forced Output Alignment |
7e7d5768 | 4946 | * Forced Input Alignment:: Forced Input Alignment |
0c71d759 | 4947 | * Output Section Constraint:: Output section constraint |
252b5132 RH |
4948 | * Output Section Region:: Output section region |
4949 | * Output Section Phdr:: Output section phdr | |
4950 | * Output Section Fill:: Output section fill | |
4951 | @end menu | |
4952 | ||
4953 | @node Output Section Type | |
36f63dca | 4954 | @subsubsection Output Section Type |
252b5132 RH |
4955 | Each output section may have a type. The type is a keyword in |
4956 | parentheses. The following types are defined: | |
4957 | ||
4958 | @table @code | |
4959 | @item NOLOAD | |
4960 | The section should be marked as not loadable, so that it will not be | |
4961 | loaded into memory when the program is run. | |
4962 | @item DSECT | |
4963 | @itemx COPY | |
4964 | @itemx INFO | |
4965 | @itemx OVERLAY | |
4966 | These type names are supported for backward compatibility, and are | |
4967 | rarely used. They all have the same effect: the section should be | |
4968 | marked as not allocatable, so that no memory is allocated for the | |
4969 | section when the program is run. | |
4970 | @end table | |
4971 | ||
4972 | @kindex NOLOAD | |
4973 | @cindex prevent unnecessary loading | |
4974 | @cindex loading, preventing | |
4975 | The linker normally sets the attributes of an output section based on | |
4976 | the input sections which map into it. You can override this by using | |
4977 | the section type. For example, in the script sample below, the | |
4978 | @samp{ROM} section is addressed at memory location @samp{0} and does not | |
2e76e85a | 4979 | need to be loaded when the program is run. |
252b5132 RH |
4980 | @smallexample |
4981 | @group | |
4982 | SECTIONS @{ | |
4983 | ROM 0 (NOLOAD) : @{ @dots{} @} | |
4984 | @dots{} | |
4985 | @} | |
4986 | @end group | |
4987 | @end smallexample | |
4988 | ||
4989 | @node Output Section LMA | |
36f63dca | 4990 | @subsubsection Output Section LMA |
562d3460 | 4991 | @kindex AT>@var{lma_region} |
252b5132 RH |
4992 | @kindex AT(@var{lma}) |
4993 | @cindex load address | |
4994 | @cindex section load address | |
4995 | Every section has a virtual address (VMA) and a load address (LMA); see | |
ea5cae92 NC |
4996 | @ref{Basic Script Concepts}. The virtual address is specified by the |
4997 | @pxref{Output Section Address} described earlier. The load address is | |
4998 | specified by the @code{AT} or @code{AT>} keywords. Specifying a load | |
4999 | address is optional. | |
6bdafbeb | 5000 | |
ea5cae92 NC |
5001 | The @code{AT} keyword takes an expression as an argument. This |
5002 | specifies the exact load address of the section. The @code{AT>} keyword | |
5003 | takes the name of a memory region as an argument. @xref{MEMORY}. The | |
5004 | load address of the section is set to the next free address in the | |
5005 | region, aligned to the section's alignment requirements. | |
dc0b6aa0 AM |
5006 | |
5007 | If neither @code{AT} nor @code{AT>} is specified for an allocatable | |
ea5cae92 NC |
5008 | section, the linker will use the following heuristic to determine the |
5009 | load address: | |
5010 | ||
5011 | @itemize @bullet | |
5012 | @item | |
5013 | If the section has a specific VMA address, then this is used as | |
5014 | the LMA address as well. | |
5015 | ||
5016 | @item | |
5017 | If the section is not allocatable then its LMA is set to its VMA. | |
5018 | ||
5019 | @item | |
5020 | Otherwise if a memory region can be found that is compatible | |
5021 | with the current section, and this region contains at least one | |
5022 | section, then the LMA is set so the difference between the | |
5023 | VMA and LMA is the same as the difference between the VMA and LMA of | |
5024 | the last section in the located region. | |
5025 | ||
5026 | @item | |
5027 | If no memory regions have been declared then a default region | |
5028 | that covers the entire address space is used in the previous step. | |
5029 | ||
5030 | @item | |
5031 | If no suitable region could be found, or there was no previous | |
5032 | section then the LMA is set equal to the VMA. | |
5033 | @end itemize | |
252b5132 RH |
5034 | |
5035 | @cindex ROM initialized data | |
5036 | @cindex initialized data in ROM | |
5037 | This feature is designed to make it easy to build a ROM image. For | |
5038 | example, the following linker script creates three output sections: one | |
5039 | called @samp{.text}, which starts at @code{0x1000}, one called | |
5040 | @samp{.mdata}, which is loaded at the end of the @samp{.text} section | |
5041 | even though its VMA is @code{0x2000}, and one called @samp{.bss} to hold | |
5042 | uninitialized data at address @code{0x3000}. The symbol @code{_data} is | |
5043 | defined with the value @code{0x2000}, which shows that the location | |
5044 | counter holds the VMA value, not the LMA value. | |
5045 | ||
5046 | @smallexample | |
5047 | @group | |
5048 | SECTIONS | |
5049 | @{ | |
5050 | .text 0x1000 : @{ *(.text) _etext = . ; @} | |
a1ab1d2a | 5051 | .mdata 0x2000 : |
252b5132 RH |
5052 | AT ( ADDR (.text) + SIZEOF (.text) ) |
5053 | @{ _data = . ; *(.data); _edata = . ; @} | |
5054 | .bss 0x3000 : | |
5055 | @{ _bstart = . ; *(.bss) *(COMMON) ; _bend = . ;@} | |
5056 | @} | |
5057 | @end group | |
5058 | @end smallexample | |
5059 | ||
5060 | The run-time initialization code for use with a program generated with | |
5061 | this linker script would include something like the following, to copy | |
5062 | the initialized data from the ROM image to its runtime address. Notice | |
5063 | how this code takes advantage of the symbols defined by the linker | |
5064 | script. | |
5065 | ||
5066 | @smallexample | |
5067 | @group | |
5068 | extern char _etext, _data, _edata, _bstart, _bend; | |
5069 | char *src = &_etext; | |
5070 | char *dst = &_data; | |
5071 | ||
ea5cae92 NC |
5072 | /* ROM has data at end of text; copy it. */ |
5073 | while (dst < &_edata) | |
252b5132 | 5074 | *dst++ = *src++; |
252b5132 | 5075 | |
ea5cae92 | 5076 | /* Zero bss. */ |
252b5132 RH |
5077 | for (dst = &_bstart; dst< &_bend; dst++) |
5078 | *dst = 0; | |
5079 | @end group | |
5080 | @end smallexample | |
5081 | ||
bbf115d3 L |
5082 | @node Forced Output Alignment |
5083 | @subsubsection Forced Output Alignment | |
5084 | @kindex ALIGN(@var{section_align}) | |
5085 | @cindex forcing output section alignment | |
5086 | @cindex output section alignment | |
1eec346e | 5087 | You can increase an output section's alignment by using ALIGN. As an |
13075d04 SH |
5088 | alternative you can enforce that the difference between the VMA and LMA remains |
5089 | intact throughout this output section with the ALIGN_WITH_INPUT attribute. | |
bbf115d3 | 5090 | |
7e7d5768 AM |
5091 | @node Forced Input Alignment |
5092 | @subsubsection Forced Input Alignment | |
5093 | @kindex SUBALIGN(@var{subsection_align}) | |
5094 | @cindex forcing input section alignment | |
5095 | @cindex input section alignment | |
5096 | You can force input section alignment within an output section by using | |
5097 | SUBALIGN. The value specified overrides any alignment given by input | |
5098 | sections, whether larger or smaller. | |
5099 | ||
0c71d759 NC |
5100 | @node Output Section Constraint |
5101 | @subsubsection Output Section Constraint | |
5102 | @kindex ONLY_IF_RO | |
5103 | @kindex ONLY_IF_RW | |
5104 | @cindex constraints on output sections | |
5105 | You can specify that an output section should only be created if all | |
5106 | of its input sections are read-only or all of its input sections are | |
5107 | read-write by using the keyword @code{ONLY_IF_RO} and | |
5108 | @code{ONLY_IF_RW} respectively. | |
5109 | ||
252b5132 | 5110 | @node Output Section Region |
36f63dca | 5111 | @subsubsection Output Section Region |
252b5132 RH |
5112 | @kindex >@var{region} |
5113 | @cindex section, assigning to memory region | |
5114 | @cindex memory regions and sections | |
5115 | You can assign a section to a previously defined region of memory by | |
5116 | using @samp{>@var{region}}. @xref{MEMORY}. | |
5117 | ||
5118 | Here is a simple example: | |
5119 | @smallexample | |
5120 | @group | |
5121 | MEMORY @{ rom : ORIGIN = 0x1000, LENGTH = 0x1000 @} | |
5122 | SECTIONS @{ ROM : @{ *(.text) @} >rom @} | |
5123 | @end group | |
5124 | @end smallexample | |
5125 | ||
5126 | @node Output Section Phdr | |
36f63dca | 5127 | @subsubsection Output Section Phdr |
252b5132 RH |
5128 | @kindex :@var{phdr} |
5129 | @cindex section, assigning to program header | |
5130 | @cindex program headers and sections | |
5131 | You can assign a section to a previously defined program segment by | |
5132 | using @samp{:@var{phdr}}. @xref{PHDRS}. If a section is assigned to | |
5133 | one or more segments, then all subsequent allocated sections will be | |
5134 | assigned to those segments as well, unless they use an explicitly | |
5135 | @code{:@var{phdr}} modifier. You can use @code{:NONE} to tell the | |
5136 | linker to not put the section in any segment at all. | |
5137 | ||
5138 | Here is a simple example: | |
5139 | @smallexample | |
5140 | @group | |
5141 | PHDRS @{ text PT_LOAD ; @} | |
5142 | SECTIONS @{ .text : @{ *(.text) @} :text @} | |
5143 | @end group | |
5144 | @end smallexample | |
5145 | ||
5146 | @node Output Section Fill | |
36f63dca | 5147 | @subsubsection Output Section Fill |
252b5132 RH |
5148 | @kindex =@var{fillexp} |
5149 | @cindex section fill pattern | |
5150 | @cindex fill pattern, entire section | |
5151 | You can set the fill pattern for an entire section by using | |
5152 | @samp{=@var{fillexp}}. @var{fillexp} is an expression | |
5153 | (@pxref{Expressions}). Any otherwise unspecified regions of memory | |
5154 | within the output section (for example, gaps left due to the required | |
a139d329 AM |
5155 | alignment of input sections) will be filled with the value, repeated as |
5156 | necessary. If the fill expression is a simple hex number, ie. a string | |
9673c93c | 5157 | of hex digit starting with @samp{0x} and without a trailing @samp{k} or @samp{M}, then |
a139d329 AM |
5158 | an arbitrarily long sequence of hex digits can be used to specify the |
5159 | fill pattern; Leading zeros become part of the pattern too. For all | |
9673c93c | 5160 | other cases, including extra parentheses or a unary @code{+}, the fill |
a139d329 AM |
5161 | pattern is the four least significant bytes of the value of the |
5162 | expression. In all cases, the number is big-endian. | |
252b5132 RH |
5163 | |
5164 | You can also change the fill value with a @code{FILL} command in the | |
9673c93c | 5165 | output section commands; (@pxref{Output Section Data}). |
252b5132 RH |
5166 | |
5167 | Here is a simple example: | |
5168 | @smallexample | |
5169 | @group | |
563e308f | 5170 | SECTIONS @{ .text : @{ *(.text) @} =0x90909090 @} |
252b5132 RH |
5171 | @end group |
5172 | @end smallexample | |
5173 | ||
5174 | @node Overlay Description | |
36f63dca | 5175 | @subsection Overlay Description |
252b5132 RH |
5176 | @kindex OVERLAY |
5177 | @cindex overlays | |
5178 | An overlay description provides an easy way to describe sections which | |
5179 | are to be loaded as part of a single memory image but are to be run at | |
5180 | the same memory address. At run time, some sort of overlay manager will | |
5181 | copy the overlaid sections in and out of the runtime memory address as | |
5182 | required, perhaps by simply manipulating addressing bits. This approach | |
5183 | can be useful, for example, when a certain region of memory is faster | |
5184 | than another. | |
5185 | ||
5186 | Overlays are described using the @code{OVERLAY} command. The | |
5187 | @code{OVERLAY} command is used within a @code{SECTIONS} command, like an | |
5188 | output section description. The full syntax of the @code{OVERLAY} | |
5189 | command is as follows: | |
5190 | @smallexample | |
5191 | @group | |
5192 | OVERLAY [@var{start}] : [NOCROSSREFS] [AT ( @var{ldaddr} )] | |
5193 | @{ | |
5194 | @var{secname1} | |
5195 | @{ | |
5196 | @var{output-section-command} | |
5197 | @var{output-section-command} | |
5198 | @dots{} | |
5199 | @} [:@var{phdr}@dots{}] [=@var{fill}] | |
5200 | @var{secname2} | |
5201 | @{ | |
5202 | @var{output-section-command} | |
5203 | @var{output-section-command} | |
5204 | @dots{} | |
5205 | @} [:@var{phdr}@dots{}] [=@var{fill}] | |
5206 | @dots{} | |
abc9061b | 5207 | @} [>@var{region}] [:@var{phdr}@dots{}] [=@var{fill}] [,] |
252b5132 RH |
5208 | @end group |
5209 | @end smallexample | |
5210 | ||
5211 | Everything is optional except @code{OVERLAY} (a keyword), and each | |
5212 | section must have a name (@var{secname1} and @var{secname2} above). The | |
5213 | section definitions within the @code{OVERLAY} construct are identical to | |
11e7fd74 | 5214 | those within the general @code{SECTIONS} construct (@pxref{SECTIONS}), |
252b5132 RH |
5215 | except that no addresses and no memory regions may be defined for |
5216 | sections within an @code{OVERLAY}. | |
5217 | ||
abc9061b CC |
5218 | The comma at the end may be required if a @var{fill} is used and |
5219 | the next @var{sections-command} looks like a continuation of the expression. | |
5220 | ||
252b5132 RH |
5221 | The sections are all defined with the same starting address. The load |
5222 | addresses of the sections are arranged such that they are consecutive in | |
5223 | memory starting at the load address used for the @code{OVERLAY} as a | |
5224 | whole (as with normal section definitions, the load address is optional, | |
5225 | and defaults to the start address; the start address is also optional, | |
5226 | and defaults to the current value of the location counter). | |
5227 | ||
56dd11f0 NC |
5228 | If the @code{NOCROSSREFS} keyword is used, and there are any |
5229 | references among the sections, the linker will report an error. Since | |
5230 | the sections all run at the same address, it normally does not make | |
5231 | sense for one section to refer directly to another. | |
5232 | @xref{Miscellaneous Commands, NOCROSSREFS}. | |
252b5132 RH |
5233 | |
5234 | For each section within the @code{OVERLAY}, the linker automatically | |
34711ca3 | 5235 | provides two symbols. The symbol @code{__load_start_@var{secname}} is |
252b5132 RH |
5236 | defined as the starting load address of the section. The symbol |
5237 | @code{__load_stop_@var{secname}} is defined as the final load address of | |
5238 | the section. Any characters within @var{secname} which are not legal | |
5239 | within C identifiers are removed. C (or assembler) code may use these | |
5240 | symbols to move the overlaid sections around as necessary. | |
5241 | ||
5242 | At the end of the overlay, the value of the location counter is set to | |
5243 | the start address of the overlay plus the size of the largest section. | |
5244 | ||
5245 | Here is an example. Remember that this would appear inside a | |
5246 | @code{SECTIONS} construct. | |
5247 | @smallexample | |
5248 | @group | |
5249 | OVERLAY 0x1000 : AT (0x4000) | |
5250 | @{ | |
5251 | .text0 @{ o1/*.o(.text) @} | |
5252 | .text1 @{ o2/*.o(.text) @} | |
5253 | @} | |
5254 | @end group | |
5255 | @end smallexample | |
5256 | @noindent | |
5257 | This will define both @samp{.text0} and @samp{.text1} to start at | |
5258 | address 0x1000. @samp{.text0} will be loaded at address 0x4000, and | |
5259 | @samp{.text1} will be loaded immediately after @samp{.text0}. The | |
34711ca3 | 5260 | following symbols will be defined if referenced: @code{__load_start_text0}, |
252b5132 RH |
5261 | @code{__load_stop_text0}, @code{__load_start_text1}, |
5262 | @code{__load_stop_text1}. | |
5263 | ||
5264 | C code to copy overlay @code{.text1} into the overlay area might look | |
5265 | like the following. | |
5266 | ||
5267 | @smallexample | |
5268 | @group | |
5269 | extern char __load_start_text1, __load_stop_text1; | |
5270 | memcpy ((char *) 0x1000, &__load_start_text1, | |
5271 | &__load_stop_text1 - &__load_start_text1); | |
5272 | @end group | |
5273 | @end smallexample | |
5274 | ||
5275 | Note that the @code{OVERLAY} command is just syntactic sugar, since | |
5276 | everything it does can be done using the more basic commands. The above | |
5277 | example could have been written identically as follows. | |
5278 | ||
5279 | @smallexample | |
5280 | @group | |
5281 | .text0 0x1000 : AT (0x4000) @{ o1/*.o(.text) @} | |
34711ca3 AM |
5282 | PROVIDE (__load_start_text0 = LOADADDR (.text0)); |
5283 | PROVIDE (__load_stop_text0 = LOADADDR (.text0) + SIZEOF (.text0)); | |
252b5132 | 5284 | .text1 0x1000 : AT (0x4000 + SIZEOF (.text0)) @{ o2/*.o(.text) @} |
34711ca3 AM |
5285 | PROVIDE (__load_start_text1 = LOADADDR (.text1)); |
5286 | PROVIDE (__load_stop_text1 = LOADADDR (.text1) + SIZEOF (.text1)); | |
252b5132 RH |
5287 | . = 0x1000 + MAX (SIZEOF (.text0), SIZEOF (.text1)); |
5288 | @end group | |
5289 | @end smallexample | |
5290 | ||
5291 | @node MEMORY | |
36f63dca | 5292 | @section MEMORY Command |
252b5132 RH |
5293 | @kindex MEMORY |
5294 | @cindex memory regions | |
5295 | @cindex regions of memory | |
5296 | @cindex allocating memory | |
5297 | @cindex discontinuous memory | |
5298 | The linker's default configuration permits allocation of all available | |
5299 | memory. You can override this by using the @code{MEMORY} command. | |
5300 | ||
5301 | The @code{MEMORY} command describes the location and size of blocks of | |
5302 | memory in the target. You can use it to describe which memory regions | |
5303 | may be used by the linker, and which memory regions it must avoid. You | |
5304 | can then assign sections to particular memory regions. The linker will | |
5305 | set section addresses based on the memory regions, and will warn about | |
5306 | regions that become too full. The linker will not shuffle sections | |
5307 | around to fit into the available regions. | |
5308 | ||
127fcdff AB |
5309 | A linker script may contain many uses of the @code{MEMORY} command, |
5310 | however, all memory blocks defined are treated as if they were | |
5311 | specified inside a single @code{MEMORY} command. The syntax for | |
5312 | @code{MEMORY} is: | |
252b5132 RH |
5313 | @smallexample |
5314 | @group | |
a1ab1d2a | 5315 | MEMORY |
252b5132 RH |
5316 | @{ |
5317 | @var{name} [(@var{attr})] : ORIGIN = @var{origin}, LENGTH = @var{len} | |
5318 | @dots{} | |
5319 | @} | |
5320 | @end group | |
5321 | @end smallexample | |
5322 | ||
5323 | The @var{name} is a name used in the linker script to refer to the | |
5324 | region. The region name has no meaning outside of the linker script. | |
5325 | Region names are stored in a separate name space, and will not conflict | |
5326 | with symbol names, file names, or section names. Each memory region | |
4a93e180 NC |
5327 | must have a distinct name within the @code{MEMORY} command. However you can |
5328 | add later alias names to existing memory regions with the @ref{REGION_ALIAS} | |
9d5777a3 | 5329 | command. |
252b5132 RH |
5330 | |
5331 | @cindex memory region attributes | |
5332 | The @var{attr} string is an optional list of attributes that specify | |
5333 | whether to use a particular memory region for an input section which is | |
5334 | not explicitly mapped in the linker script. As described in | |
5335 | @ref{SECTIONS}, if you do not specify an output section for some input | |
5336 | section, the linker will create an output section with the same name as | |
5337 | the input section. If you define region attributes, the linker will use | |
5338 | them to select the memory region for the output section that it creates. | |
5339 | ||
5340 | The @var{attr} string must consist only of the following characters: | |
5341 | @table @samp | |
5342 | @item R | |
5343 | Read-only section | |
5344 | @item W | |
5345 | Read/write section | |
5346 | @item X | |
5347 | Executable section | |
5348 | @item A | |
5349 | Allocatable section | |
5350 | @item I | |
5351 | Initialized section | |
5352 | @item L | |
5353 | Same as @samp{I} | |
5354 | @item ! | |
c09e9a8c | 5355 | Invert the sense of any of the attributes that follow |
252b5132 RH |
5356 | @end table |
5357 | ||
5358 | If a unmapped section matches any of the listed attributes other than | |
5359 | @samp{!}, it will be placed in the memory region. The @samp{!} | |
5360 | attribute reverses this test, so that an unmapped section will be placed | |
5361 | in the memory region only if it does not match any of the listed | |
5362 | attributes. | |
5363 | ||
5364 | @kindex ORIGIN = | |
5365 | @kindex o = | |
5366 | @kindex org = | |
9cd6d51a NC |
5367 | The @var{origin} is an numerical expression for the start address of |
5368 | the memory region. The expression must evaluate to a constant and it | |
5369 | cannot involve any symbols. The keyword @code{ORIGIN} may be | |
5370 | abbreviated to @code{org} or @code{o} (but not, for example, | |
5371 | @code{ORG}). | |
252b5132 RH |
5372 | |
5373 | @kindex LENGTH = | |
5374 | @kindex len = | |
5375 | @kindex l = | |
5376 | The @var{len} is an expression for the size in bytes of the memory | |
5377 | region. As with the @var{origin} expression, the expression must | |
9cd6d51a NC |
5378 | be numerical only and must evaluate to a constant. The keyword |
5379 | @code{LENGTH} may be abbreviated to @code{len} or @code{l}. | |
252b5132 RH |
5380 | |
5381 | In the following example, we specify that there are two memory regions | |
5382 | available for allocation: one starting at @samp{0} for 256 kilobytes, | |
5383 | and the other starting at @samp{0x40000000} for four megabytes. The | |
5384 | linker will place into the @samp{rom} memory region every section which | |
5385 | is not explicitly mapped into a memory region, and is either read-only | |
5386 | or executable. The linker will place other sections which are not | |
5387 | explicitly mapped into a memory region into the @samp{ram} memory | |
5388 | region. | |
5389 | ||
5390 | @smallexample | |
5391 | @group | |
a1ab1d2a | 5392 | MEMORY |
252b5132 RH |
5393 | @{ |
5394 | rom (rx) : ORIGIN = 0, LENGTH = 256K | |
5395 | ram (!rx) : org = 0x40000000, l = 4M | |
5396 | @} | |
5397 | @end group | |
5398 | @end smallexample | |
5399 | ||
5400 | Once you define a memory region, you can direct the linker to place | |
5401 | specific output sections into that memory region by using the | |
5402 | @samp{>@var{region}} output section attribute. For example, if you have | |
5403 | a memory region named @samp{mem}, you would use @samp{>mem} in the | |
5404 | output section definition. @xref{Output Section Region}. If no address | |
5405 | was specified for the output section, the linker will set the address to | |
5406 | the next available address within the memory region. If the combined | |
5407 | output sections directed to a memory region are too large for the | |
5408 | region, the linker will issue an error message. | |
5409 | ||
3ec57632 | 5410 | It is possible to access the origin and length of a memory in an |
c0065db7 | 5411 | expression via the @code{ORIGIN(@var{memory})} and |
3ec57632 NC |
5412 | @code{LENGTH(@var{memory})} functions: |
5413 | ||
5414 | @smallexample | |
5415 | @group | |
c0065db7 | 5416 | _fstack = ORIGIN(ram) + LENGTH(ram) - 4; |
3ec57632 NC |
5417 | @end group |
5418 | @end smallexample | |
5419 | ||
252b5132 RH |
5420 | @node PHDRS |
5421 | @section PHDRS Command | |
5422 | @kindex PHDRS | |
5423 | @cindex program headers | |
5424 | @cindex ELF program headers | |
5425 | @cindex program segments | |
5426 | @cindex segments, ELF | |
5427 | The ELF object file format uses @dfn{program headers}, also knows as | |
5428 | @dfn{segments}. The program headers describe how the program should be | |
5429 | loaded into memory. You can print them out by using the @code{objdump} | |
5430 | program with the @samp{-p} option. | |
5431 | ||
5432 | When you run an ELF program on a native ELF system, the system loader | |
5433 | reads the program headers in order to figure out how to load the | |
5434 | program. This will only work if the program headers are set correctly. | |
5435 | This manual does not describe the details of how the system loader | |
5436 | interprets program headers; for more information, see the ELF ABI. | |
5437 | ||
5438 | The linker will create reasonable program headers by default. However, | |
5439 | in some cases, you may need to specify the program headers more | |
5440 | precisely. You may use the @code{PHDRS} command for this purpose. When | |
5441 | the linker sees the @code{PHDRS} command in the linker script, it will | |
5442 | not create any program headers other than the ones specified. | |
5443 | ||
5444 | The linker only pays attention to the @code{PHDRS} command when | |
5445 | generating an ELF output file. In other cases, the linker will simply | |
5446 | ignore @code{PHDRS}. | |
5447 | ||
5448 | This is the syntax of the @code{PHDRS} command. The words @code{PHDRS}, | |
5449 | @code{FILEHDR}, @code{AT}, and @code{FLAGS} are keywords. | |
5450 | ||
5451 | @smallexample | |
5452 | @group | |
5453 | PHDRS | |
5454 | @{ | |
5455 | @var{name} @var{type} [ FILEHDR ] [ PHDRS ] [ AT ( @var{address} ) ] | |
5456 | [ FLAGS ( @var{flags} ) ] ; | |
5457 | @} | |
5458 | @end group | |
5459 | @end smallexample | |
5460 | ||
5461 | The @var{name} is used only for reference in the @code{SECTIONS} command | |
5462 | of the linker script. It is not put into the output file. Program | |
5463 | header names are stored in a separate name space, and will not conflict | |
5464 | with symbol names, file names, or section names. Each program header | |
5c1a3f0f NS |
5465 | must have a distinct name. The headers are processed in order and it |
5466 | is usual for them to map to sections in ascending load address order. | |
252b5132 RH |
5467 | |
5468 | Certain program header types describe segments of memory which the | |
5469 | system loader will load from the file. In the linker script, you | |
5470 | specify the contents of these segments by placing allocatable output | |
5471 | sections in the segments. You use the @samp{:@var{phdr}} output section | |
5472 | attribute to place a section in a particular segment. @xref{Output | |
5473 | Section Phdr}. | |
5474 | ||
5475 | It is normal to put certain sections in more than one segment. This | |
5476 | merely implies that one segment of memory contains another. You may | |
5477 | repeat @samp{:@var{phdr}}, using it once for each segment which should | |
5478 | contain the section. | |
5479 | ||
5480 | If you place a section in one or more segments using @samp{:@var{phdr}}, | |
5481 | then the linker will place all subsequent allocatable sections which do | |
5482 | not specify @samp{:@var{phdr}} in the same segments. This is for | |
5483 | convenience, since generally a whole set of contiguous sections will be | |
5484 | placed in a single segment. You can use @code{:NONE} to override the | |
5485 | default segment and tell the linker to not put the section in any | |
5486 | segment at all. | |
5487 | ||
5488 | @kindex FILEHDR | |
5489 | @kindex PHDRS | |
5c1a3f0f | 5490 | You may use the @code{FILEHDR} and @code{PHDRS} keywords after |
252b5132 RH |
5491 | the program header type to further describe the contents of the segment. |
5492 | The @code{FILEHDR} keyword means that the segment should include the ELF | |
5493 | file header. The @code{PHDRS} keyword means that the segment should | |
5c1a3f0f | 5494 | include the ELF program headers themselves. If applied to a loadable |
4100cea3 AM |
5495 | segment (@code{PT_LOAD}), all prior loadable segments must have one of |
5496 | these keywords. | |
252b5132 RH |
5497 | |
5498 | The @var{type} may be one of the following. The numbers indicate the | |
5499 | value of the keyword. | |
5500 | ||
5501 | @table @asis | |
5502 | @item @code{PT_NULL} (0) | |
5503 | Indicates an unused program header. | |
5504 | ||
5505 | @item @code{PT_LOAD} (1) | |
5506 | Indicates that this program header describes a segment to be loaded from | |
5507 | the file. | |
5508 | ||
5509 | @item @code{PT_DYNAMIC} (2) | |
5510 | Indicates a segment where dynamic linking information can be found. | |
5511 | ||
5512 | @item @code{PT_INTERP} (3) | |
5513 | Indicates a segment where the name of the program interpreter may be | |
5514 | found. | |
5515 | ||
5516 | @item @code{PT_NOTE} (4) | |
5517 | Indicates a segment holding note information. | |
5518 | ||
5519 | @item @code{PT_SHLIB} (5) | |
5520 | A reserved program header type, defined but not specified by the ELF | |
5521 | ABI. | |
5522 | ||
5523 | @item @code{PT_PHDR} (6) | |
5524 | Indicates a segment where the program headers may be found. | |
5525 | ||
1a9ccd70 NC |
5526 | @item @code{PT_TLS} (7) |
5527 | Indicates a segment containing thread local storage. | |
5528 | ||
252b5132 RH |
5529 | @item @var{expression} |
5530 | An expression giving the numeric type of the program header. This may | |
5531 | be used for types not defined above. | |
5532 | @end table | |
5533 | ||
5534 | You can specify that a segment should be loaded at a particular address | |
5535 | in memory by using an @code{AT} expression. This is identical to the | |
5536 | @code{AT} command used as an output section attribute (@pxref{Output | |
5537 | Section LMA}). The @code{AT} command for a program header overrides the | |
5538 | output section attribute. | |
5539 | ||
5540 | The linker will normally set the segment flags based on the sections | |
5541 | which comprise the segment. You may use the @code{FLAGS} keyword to | |
5542 | explicitly specify the segment flags. The value of @var{flags} must be | |
5543 | an integer. It is used to set the @code{p_flags} field of the program | |
5544 | header. | |
5545 | ||
5546 | Here is an example of @code{PHDRS}. This shows a typical set of program | |
5547 | headers used on a native ELF system. | |
5548 | ||
5549 | @example | |
5550 | @group | |
5551 | PHDRS | |
5552 | @{ | |
5553 | headers PT_PHDR PHDRS ; | |
5554 | interp PT_INTERP ; | |
5555 | text PT_LOAD FILEHDR PHDRS ; | |
5556 | data PT_LOAD ; | |
5557 | dynamic PT_DYNAMIC ; | |
5558 | @} | |
5559 | ||
5560 | SECTIONS | |
5561 | @{ | |
5562 | . = SIZEOF_HEADERS; | |
5563 | .interp : @{ *(.interp) @} :text :interp | |
5564 | .text : @{ *(.text) @} :text | |
5565 | .rodata : @{ *(.rodata) @} /* defaults to :text */ | |
5566 | @dots{} | |
5567 | . = . + 0x1000; /* move to a new page in memory */ | |
5568 | .data : @{ *(.data) @} :data | |
5569 | .dynamic : @{ *(.dynamic) @} :data :dynamic | |
5570 | @dots{} | |
5571 | @} | |
5572 | @end group | |
5573 | @end example | |
5574 | ||
5575 | @node VERSION | |
5576 | @section VERSION Command | |
5577 | @kindex VERSION @{script text@} | |
5578 | @cindex symbol versions | |
5579 | @cindex version script | |
5580 | @cindex versions of symbols | |
5581 | The linker supports symbol versions when using ELF. Symbol versions are | |
5582 | only useful when using shared libraries. The dynamic linker can use | |
5583 | symbol versions to select a specific version of a function when it runs | |
5584 | a program that may have been linked against an earlier version of the | |
5585 | shared library. | |
5586 | ||
5587 | You can include a version script directly in the main linker script, or | |
5588 | you can supply the version script as an implicit linker script. You can | |
5589 | also use the @samp{--version-script} linker option. | |
5590 | ||
5591 | The syntax of the @code{VERSION} command is simply | |
5592 | @smallexample | |
5593 | VERSION @{ version-script-commands @} | |
5594 | @end smallexample | |
5595 | ||
5596 | The format of the version script commands is identical to that used by | |
5597 | Sun's linker in Solaris 2.5. The version script defines a tree of | |
5598 | version nodes. You specify the node names and interdependencies in the | |
5599 | version script. You can specify which symbols are bound to which | |
5600 | version nodes, and you can reduce a specified set of symbols to local | |
5601 | scope so that they are not globally visible outside of the shared | |
5602 | library. | |
5603 | ||
5604 | The easiest way to demonstrate the version script language is with a few | |
5605 | examples. | |
5606 | ||
5607 | @smallexample | |
5608 | VERS_1.1 @{ | |
5609 | global: | |
5610 | foo1; | |
5611 | local: | |
a1ab1d2a UD |
5612 | old*; |
5613 | original*; | |
5614 | new*; | |
252b5132 RH |
5615 | @}; |
5616 | ||
5617 | VERS_1.2 @{ | |
5618 | foo2; | |
5619 | @} VERS_1.1; | |
5620 | ||
5621 | VERS_2.0 @{ | |
5622 | bar1; bar2; | |
c0065db7 | 5623 | extern "C++" @{ |
86043bbb | 5624 | ns::*; |
bb1515f2 MF |
5625 | "f(int, double)"; |
5626 | @}; | |
252b5132 RH |
5627 | @} VERS_1.2; |
5628 | @end smallexample | |
5629 | ||
5630 | This example version script defines three version nodes. The first | |
5631 | version node defined is @samp{VERS_1.1}; it has no other dependencies. | |
5632 | The script binds the symbol @samp{foo1} to @samp{VERS_1.1}. It reduces | |
5633 | a number of symbols to local scope so that they are not visible outside | |
313e35ee AM |
5634 | of the shared library; this is done using wildcard patterns, so that any |
5635 | symbol whose name begins with @samp{old}, @samp{original}, or @samp{new} | |
5636 | is matched. The wildcard patterns available are the same as those used | |
5637 | in the shell when matching filenames (also known as ``globbing''). | |
86043bbb MM |
5638 | However, if you specify the symbol name inside double quotes, then the |
5639 | name is treated as literal, rather than as a glob pattern. | |
252b5132 RH |
5640 | |
5641 | Next, the version script defines node @samp{VERS_1.2}. This node | |
5642 | depends upon @samp{VERS_1.1}. The script binds the symbol @samp{foo2} | |
5643 | to the version node @samp{VERS_1.2}. | |
5644 | ||
5645 | Finally, the version script defines node @samp{VERS_2.0}. This node | |
5646 | depends upon @samp{VERS_1.2}. The scripts binds the symbols @samp{bar1} | |
5647 | and @samp{bar2} are bound to the version node @samp{VERS_2.0}. | |
5648 | ||
5649 | When the linker finds a symbol defined in a library which is not | |
5650 | specifically bound to a version node, it will effectively bind it to an | |
5651 | unspecified base version of the library. You can bind all otherwise | |
a981ed6f | 5652 | unspecified symbols to a given version node by using @samp{global: *;} |
ae5a3597 AM |
5653 | somewhere in the version script. Note that it's slightly crazy to use |
5654 | wildcards in a global spec except on the last version node. Global | |
5655 | wildcards elsewhere run the risk of accidentally adding symbols to the | |
5656 | set exported for an old version. That's wrong since older versions | |
5657 | ought to have a fixed set of symbols. | |
252b5132 RH |
5658 | |
5659 | The names of the version nodes have no specific meaning other than what | |
5660 | they might suggest to the person reading them. The @samp{2.0} version | |
5661 | could just as well have appeared in between @samp{1.1} and @samp{1.2}. | |
5662 | However, this would be a confusing way to write a version script. | |
5663 | ||
0f6bf451 | 5664 | Node name can be omitted, provided it is the only version node |
6b9b879a JJ |
5665 | in the version script. Such version script doesn't assign any versions to |
5666 | symbols, only selects which symbols will be globally visible out and which | |
5667 | won't. | |
5668 | ||
5669 | @smallexample | |
7c9c73be | 5670 | @{ global: foo; bar; local: *; @}; |
9d201f2f | 5671 | @end smallexample |
6b9b879a | 5672 | |
252b5132 RH |
5673 | When you link an application against a shared library that has versioned |
5674 | symbols, the application itself knows which version of each symbol it | |
5675 | requires, and it also knows which version nodes it needs from each | |
5676 | shared library it is linked against. Thus at runtime, the dynamic | |
5677 | loader can make a quick check to make sure that the libraries you have | |
5678 | linked against do in fact supply all of the version nodes that the | |
5679 | application will need to resolve all of the dynamic symbols. In this | |
5680 | way it is possible for the dynamic linker to know with certainty that | |
5681 | all external symbols that it needs will be resolvable without having to | |
5682 | search for each symbol reference. | |
5683 | ||
5684 | The symbol versioning is in effect a much more sophisticated way of | |
5685 | doing minor version checking that SunOS does. The fundamental problem | |
5686 | that is being addressed here is that typically references to external | |
5687 | functions are bound on an as-needed basis, and are not all bound when | |
5688 | the application starts up. If a shared library is out of date, a | |
5689 | required interface may be missing; when the application tries to use | |
5690 | that interface, it may suddenly and unexpectedly fail. With symbol | |
5691 | versioning, the user will get a warning when they start their program if | |
5692 | the libraries being used with the application are too old. | |
5693 | ||
5694 | There are several GNU extensions to Sun's versioning approach. The | |
5695 | first of these is the ability to bind a symbol to a version node in the | |
5696 | source file where the symbol is defined instead of in the versioning | |
5697 | script. This was done mainly to reduce the burden on the library | |
5698 | maintainer. You can do this by putting something like: | |
5699 | @smallexample | |
5700 | __asm__(".symver original_foo,foo@@VERS_1.1"); | |
5701 | @end smallexample | |
5702 | @noindent | |
5703 | in the C source file. This renames the function @samp{original_foo} to | |
5704 | be an alias for @samp{foo} bound to the version node @samp{VERS_1.1}. | |
5705 | The @samp{local:} directive can be used to prevent the symbol | |
96a94295 L |
5706 | @samp{original_foo} from being exported. A @samp{.symver} directive |
5707 | takes precedence over a version script. | |
252b5132 RH |
5708 | |
5709 | The second GNU extension is to allow multiple versions of the same | |
5710 | function to appear in a given shared library. In this way you can make | |
5711 | an incompatible change to an interface without increasing the major | |
5712 | version number of the shared library, while still allowing applications | |
5713 | linked against the old interface to continue to function. | |
5714 | ||
5715 | To do this, you must use multiple @samp{.symver} directives in the | |
5716 | source file. Here is an example: | |
5717 | ||
5718 | @smallexample | |
5719 | __asm__(".symver original_foo,foo@@"); | |
5720 | __asm__(".symver old_foo,foo@@VERS_1.1"); | |
5721 | __asm__(".symver old_foo1,foo@@VERS_1.2"); | |
5722 | __asm__(".symver new_foo,foo@@@@VERS_2.0"); | |
5723 | @end smallexample | |
5724 | ||
5725 | In this example, @samp{foo@@} represents the symbol @samp{foo} bound to the | |
5726 | unspecified base version of the symbol. The source file that contains this | |
5727 | example would define 4 C functions: @samp{original_foo}, @samp{old_foo}, | |
5728 | @samp{old_foo1}, and @samp{new_foo}. | |
5729 | ||
5730 | When you have multiple definitions of a given symbol, there needs to be | |
5731 | some way to specify a default version to which external references to | |
5732 | this symbol will be bound. You can do this with the | |
5733 | @samp{foo@@@@VERS_2.0} type of @samp{.symver} directive. You can only | |
5734 | declare one version of a symbol as the default in this manner; otherwise | |
5735 | you would effectively have multiple definitions of the same symbol. | |
5736 | ||
5737 | If you wish to bind a reference to a specific version of the symbol | |
5738 | within the shared library, you can use the aliases of convenience | |
36f63dca | 5739 | (i.e., @samp{old_foo}), or you can use the @samp{.symver} directive to |
252b5132 RH |
5740 | specifically bind to an external version of the function in question. |
5741 | ||
cb840a31 L |
5742 | You can also specify the language in the version script: |
5743 | ||
5744 | @smallexample | |
5745 | VERSION extern "lang" @{ version-script-commands @} | |
5746 | @end smallexample | |
5747 | ||
c0065db7 | 5748 | The supported @samp{lang}s are @samp{C}, @samp{C++}, and @samp{Java}. |
cb840a31 L |
5749 | The linker will iterate over the list of symbols at the link time and |
5750 | demangle them according to @samp{lang} before matching them to the | |
bb1515f2 MF |
5751 | patterns specified in @samp{version-script-commands}. The default |
5752 | @samp{lang} is @samp{C}. | |
cb840a31 | 5753 | |
86043bbb MM |
5754 | Demangled names may contains spaces and other special characters. As |
5755 | described above, you can use a glob pattern to match demangled names, | |
5756 | or you can use a double-quoted string to match the string exactly. In | |
5757 | the latter case, be aware that minor differences (such as differing | |
5758 | whitespace) between the version script and the demangler output will | |
5759 | cause a mismatch. As the exact string generated by the demangler | |
5760 | might change in the future, even if the mangled name does not, you | |
5761 | should check that all of your version directives are behaving as you | |
5762 | expect when you upgrade. | |
5763 | ||
252b5132 RH |
5764 | @node Expressions |
5765 | @section Expressions in Linker Scripts | |
5766 | @cindex expressions | |
5767 | @cindex arithmetic | |
5768 | The syntax for expressions in the linker script language is identical to | |
5769 | that of C expressions. All expressions are evaluated as integers. All | |
5770 | expressions are evaluated in the same size, which is 32 bits if both the | |
5771 | host and target are 32 bits, and is otherwise 64 bits. | |
5772 | ||
5773 | You can use and set symbol values in expressions. | |
5774 | ||
5775 | The linker defines several special purpose builtin functions for use in | |
5776 | expressions. | |
5777 | ||
5778 | @menu | |
5779 | * Constants:: Constants | |
0c71d759 | 5780 | * Symbolic Constants:: Symbolic constants |
252b5132 | 5781 | * Symbols:: Symbol Names |
ecca9871 | 5782 | * Orphan Sections:: Orphan Sections |
252b5132 RH |
5783 | * Location Counter:: The Location Counter |
5784 | * Operators:: Operators | |
5785 | * Evaluation:: Evaluation | |
5786 | * Expression Section:: The Section of an Expression | |
5787 | * Builtin Functions:: Builtin Functions | |
5788 | @end menu | |
5789 | ||
5790 | @node Constants | |
5791 | @subsection Constants | |
5792 | @cindex integer notation | |
5793 | @cindex constants in linker scripts | |
5794 | All constants are integers. | |
5795 | ||
5796 | As in C, the linker considers an integer beginning with @samp{0} to be | |
5797 | octal, and an integer beginning with @samp{0x} or @samp{0X} to be | |
8a308ae8 | 5798 | hexadecimal. Alternatively the linker accepts suffixes of @samp{h} or |
11e7fd74 | 5799 | @samp{H} for hexadecimal, @samp{o} or @samp{O} for octal, @samp{b} or |
8a308ae8 NC |
5800 | @samp{B} for binary and @samp{d} or @samp{D} for decimal. Any integer |
5801 | value without a prefix or a suffix is considered to be decimal. | |
252b5132 RH |
5802 | |
5803 | @cindex scaled integers | |
5804 | @cindex K and M integer suffixes | |
5805 | @cindex M and K integer suffixes | |
5806 | @cindex suffixes for integers | |
5807 | @cindex integer suffixes | |
5808 | In addition, you can use the suffixes @code{K} and @code{M} to scale a | |
5809 | constant by | |
5810 | @c TEXI2ROFF-KILL | |
36f63dca | 5811 | @ifnottex |
252b5132 RH |
5812 | @c END TEXI2ROFF-KILL |
5813 | @code{1024} or @code{1024*1024} | |
5814 | @c TEXI2ROFF-KILL | |
36f63dca | 5815 | @end ifnottex |
252b5132 RH |
5816 | @tex |
5817 | ${\rm 1024}$ or ${\rm 1024}^2$ | |
5818 | @end tex | |
5819 | @c END TEXI2ROFF-KILL | |
8a308ae8 NC |
5820 | respectively. For example, the following |
5821 | all refer to the same quantity: | |
5822 | ||
252b5132 | 5823 | @smallexample |
36f63dca NC |
5824 | _fourk_1 = 4K; |
5825 | _fourk_2 = 4096; | |
5826 | _fourk_3 = 0x1000; | |
8a308ae8 | 5827 | _fourk_4 = 10000o; |
252b5132 RH |
5828 | @end smallexample |
5829 | ||
8a308ae8 NC |
5830 | Note - the @code{K} and @code{M} suffixes cannot be used in |
5831 | conjunction with the base suffixes mentioned above. | |
5832 | ||
0c71d759 NC |
5833 | @node Symbolic Constants |
5834 | @subsection Symbolic Constants | |
5835 | @cindex symbolic constants | |
5836 | @kindex CONSTANT | |
5837 | It is possible to refer to target specific constants via the use of | |
5838 | the @code{CONSTANT(@var{name})} operator, where @var{name} is one of: | |
5839 | ||
5840 | @table @code | |
5841 | @item MAXPAGESIZE | |
5842 | @kindex MAXPAGESIZE | |
5843 | The target's maximum page size. | |
5844 | ||
5845 | @item COMMONPAGESIZE | |
5846 | @kindex COMMONPAGESIZE | |
5847 | The target's default page size. | |
5848 | @end table | |
5849 | ||
5850 | So for example: | |
5851 | ||
5852 | @smallexample | |
9d5777a3 | 5853 | .text ALIGN (CONSTANT (MAXPAGESIZE)) : @{ *(.text) @} |
0c71d759 NC |
5854 | @end smallexample |
5855 | ||
5856 | will create a text section aligned to the largest page boundary | |
5857 | supported by the target. | |
5858 | ||
252b5132 RH |
5859 | @node Symbols |
5860 | @subsection Symbol Names | |
5861 | @cindex symbol names | |
5862 | @cindex names | |
5863 | @cindex quoted symbol names | |
5864 | @kindex " | |
5865 | Unless quoted, symbol names start with a letter, underscore, or period | |
5866 | and may include letters, digits, underscores, periods, and hyphens. | |
5867 | Unquoted symbol names must not conflict with any keywords. You can | |
5868 | specify a symbol which contains odd characters or has the same name as a | |
5869 | keyword by surrounding the symbol name in double quotes: | |
5870 | @smallexample | |
36f63dca NC |
5871 | "SECTION" = 9; |
5872 | "with a space" = "also with a space" + 10; | |
252b5132 RH |
5873 | @end smallexample |
5874 | ||
5875 | Since symbols can contain many non-alphabetic characters, it is safest | |
5876 | to delimit symbols with spaces. For example, @samp{A-B} is one symbol, | |
5877 | whereas @samp{A - B} is an expression involving subtraction. | |
5878 | ||
ecca9871 L |
5879 | @node Orphan Sections |
5880 | @subsection Orphan Sections | |
5881 | @cindex orphan | |
5882 | Orphan sections are sections present in the input files which | |
5883 | are not explicitly placed into the output file by the linker | |
5884 | script. The linker will still copy these sections into the | |
a87ded7b AB |
5885 | output file by either finding, or creating a suitable output section |
5886 | in which to place the orphaned input section. | |
5887 | ||
5888 | If the name of an orphaned input section exactly matches the name of | |
5889 | an existing output section, then the orphaned input section will be | |
5890 | placed at the end of that output section. | |
5891 | ||
5892 | If there is no output section with a matching name then new output | |
5893 | sections will be created. Each new output section will have the same | |
5894 | name as the orphan section placed within it. If there are multiple | |
5895 | orphan sections with the same name, these will all be combined into | |
5896 | one new output section. | |
5897 | ||
5898 | If new output sections are created to hold orphaned input sections, | |
5899 | then the linker must decide where to place these new output sections | |
e299b355 AM |
5900 | in relation to existing output sections. On most modern targets, the |
5901 | linker attempts to place orphan sections after sections of the same | |
5902 | attribute, such as code vs data, loadable vs non-loadable, etc. If no | |
5903 | sections with matching attributes are found, or your target lacks this | |
5904 | support, the orphan section is placed at the end of the file. | |
ecca9871 | 5905 | |
a05a5b64 TP |
5906 | The command-line options @samp{--orphan-handling} and @samp{--unique} |
5907 | (@pxref{Options,,Command-line Options}) can be used to control which | |
c005eb9e AB |
5908 | output sections an orphan is placed in. |
5909 | ||
252b5132 RH |
5910 | @node Location Counter |
5911 | @subsection The Location Counter | |
5912 | @kindex . | |
5913 | @cindex dot | |
5914 | @cindex location counter | |
5915 | @cindex current output location | |
5916 | The special linker variable @dfn{dot} @samp{.} always contains the | |
5917 | current output location counter. Since the @code{.} always refers to a | |
5918 | location in an output section, it may only appear in an expression | |
5919 | within a @code{SECTIONS} command. The @code{.} symbol may appear | |
5920 | anywhere that an ordinary symbol is allowed in an expression. | |
5921 | ||
5922 | @cindex holes | |
5923 | Assigning a value to @code{.} will cause the location counter to be | |
5924 | moved. This may be used to create holes in the output section. The | |
dc0b6aa0 AM |
5925 | location counter may not be moved backwards inside an output section, |
5926 | and may not be moved backwards outside of an output section if so | |
5927 | doing creates areas with overlapping LMAs. | |
252b5132 RH |
5928 | |
5929 | @smallexample | |
5930 | SECTIONS | |
5931 | @{ | |
5932 | output : | |
5933 | @{ | |
5934 | file1(.text) | |
5935 | . = . + 1000; | |
5936 | file2(.text) | |
5937 | . += 1000; | |
5938 | file3(.text) | |
563e308f | 5939 | @} = 0x12345678; |
252b5132 RH |
5940 | @} |
5941 | @end smallexample | |
5942 | @noindent | |
5943 | In the previous example, the @samp{.text} section from @file{file1} is | |
5944 | located at the beginning of the output section @samp{output}. It is | |
5945 | followed by a 1000 byte gap. Then the @samp{.text} section from | |
5946 | @file{file2} appears, also with a 1000 byte gap following before the | |
563e308f | 5947 | @samp{.text} section from @file{file3}. The notation @samp{= 0x12345678} |
252b5132 RH |
5948 | specifies what data to write in the gaps (@pxref{Output Section Fill}). |
5949 | ||
5c6bbab8 NC |
5950 | @cindex dot inside sections |
5951 | Note: @code{.} actually refers to the byte offset from the start of the | |
5952 | current containing object. Normally this is the @code{SECTIONS} | |
69da35b5 | 5953 | statement, whose start address is 0, hence @code{.} can be used as an |
5c6bbab8 NC |
5954 | absolute address. If @code{.} is used inside a section description |
5955 | however, it refers to the byte offset from the start of that section, | |
5956 | not an absolute address. Thus in a script like this: | |
5957 | ||
5958 | @smallexample | |
5959 | SECTIONS | |
5960 | @{ | |
5961 | . = 0x100 | |
5962 | .text: @{ | |
5963 | *(.text) | |
5964 | . = 0x200 | |
5965 | @} | |
5966 | . = 0x500 | |
5967 | .data: @{ | |
5968 | *(.data) | |
5969 | . += 0x600 | |
5970 | @} | |
5971 | @} | |
5972 | @end smallexample | |
5973 | ||
5974 | The @samp{.text} section will be assigned a starting address of 0x100 | |
5975 | and a size of exactly 0x200 bytes, even if there is not enough data in | |
5976 | the @samp{.text} input sections to fill this area. (If there is too | |
5977 | much data, an error will be produced because this would be an attempt to | |
5978 | move @code{.} backwards). The @samp{.data} section will start at 0x500 | |
5979 | and it will have an extra 0x600 bytes worth of space after the end of | |
5980 | the values from the @samp{.data} input sections and before the end of | |
5981 | the @samp{.data} output section itself. | |
5982 | ||
b5666f2f AM |
5983 | @cindex dot outside sections |
5984 | Setting symbols to the value of the location counter outside of an | |
5985 | output section statement can result in unexpected values if the linker | |
5986 | needs to place orphan sections. For example, given the following: | |
5987 | ||
5988 | @smallexample | |
5989 | SECTIONS | |
5990 | @{ | |
5991 | start_of_text = . ; | |
5992 | .text: @{ *(.text) @} | |
5993 | end_of_text = . ; | |
5994 | ||
5995 | start_of_data = . ; | |
5996 | .data: @{ *(.data) @} | |
5997 | end_of_data = . ; | |
5998 | @} | |
5999 | @end smallexample | |
6000 | ||
6001 | If the linker needs to place some input section, e.g. @code{.rodata}, | |
6002 | not mentioned in the script, it might choose to place that section | |
6003 | between @code{.text} and @code{.data}. You might think the linker | |
6004 | should place @code{.rodata} on the blank line in the above script, but | |
6005 | blank lines are of no particular significance to the linker. As well, | |
6006 | the linker doesn't associate the above symbol names with their | |
6007 | sections. Instead, it assumes that all assignments or other | |
6008 | statements belong to the previous output section, except for the | |
6009 | special case of an assignment to @code{.}. I.e., the linker will | |
6010 | place the orphan @code{.rodata} section as if the script was written | |
6011 | as follows: | |
6012 | ||
6013 | @smallexample | |
6014 | SECTIONS | |
6015 | @{ | |
6016 | start_of_text = . ; | |
6017 | .text: @{ *(.text) @} | |
6018 | end_of_text = . ; | |
6019 | ||
6020 | start_of_data = . ; | |
6021 | .rodata: @{ *(.rodata) @} | |
6022 | .data: @{ *(.data) @} | |
6023 | end_of_data = . ; | |
6024 | @} | |
6025 | @end smallexample | |
6026 | ||
6027 | This may or may not be the script author's intention for the value of | |
6028 | @code{start_of_data}. One way to influence the orphan section | |
6029 | placement is to assign the location counter to itself, as the linker | |
6030 | assumes that an assignment to @code{.} is setting the start address of | |
6031 | a following output section and thus should be grouped with that | |
6032 | section. So you could write: | |
6033 | ||
6034 | @smallexample | |
6035 | SECTIONS | |
6036 | @{ | |
6037 | start_of_text = . ; | |
6038 | .text: @{ *(.text) @} | |
6039 | end_of_text = . ; | |
6040 | ||
6041 | . = . ; | |
6042 | start_of_data = . ; | |
6043 | .data: @{ *(.data) @} | |
6044 | end_of_data = . ; | |
6045 | @} | |
6046 | @end smallexample | |
6047 | ||
6048 | Now, the orphan @code{.rodata} section will be placed between | |
6049 | @code{end_of_text} and @code{start_of_data}. | |
6050 | ||
252b5132 RH |
6051 | @need 2000 |
6052 | @node Operators | |
6053 | @subsection Operators | |
6054 | @cindex operators for arithmetic | |
6055 | @cindex arithmetic operators | |
6056 | @cindex precedence in expressions | |
6057 | The linker recognizes the standard C set of arithmetic operators, with | |
6058 | the standard bindings and precedence levels: | |
6059 | @c TEXI2ROFF-KILL | |
36f63dca | 6060 | @ifnottex |
252b5132 RH |
6061 | @c END TEXI2ROFF-KILL |
6062 | @smallexample | |
6063 | precedence associativity Operators Notes | |
6064 | (highest) | |
6065 | 1 left ! - ~ (1) | |
6066 | 2 left * / % | |
6067 | 3 left + - | |
6068 | 4 left >> << | |
6069 | 5 left == != > < <= >= | |
6070 | 6 left & | |
6071 | 7 left | | |
6072 | 8 left && | |
6073 | 9 left || | |
6074 | 10 right ? : | |
6075 | 11 right &= += -= *= /= (2) | |
6076 | (lowest) | |
6077 | @end smallexample | |
6078 | Notes: | |
a1ab1d2a | 6079 | (1) Prefix operators |
252b5132 RH |
6080 | (2) @xref{Assignments}. |
6081 | @c TEXI2ROFF-KILL | |
36f63dca | 6082 | @end ifnottex |
252b5132 RH |
6083 | @tex |
6084 | \vskip \baselineskip | |
6085 | %"lispnarrowing" is the extra indent used generally for smallexample | |
6086 | \hskip\lispnarrowing\vbox{\offinterlineskip | |
6087 | \hrule | |
6088 | \halign | |
6089 | {\vrule#&\strut\hfil\ #\ \hfil&\vrule#&\strut\hfil\ #\ \hfil&\vrule#&\strut\hfil\ {\tt #}\ \hfil&\vrule#\cr | |
6090 | height2pt&\omit&&\omit&&\omit&\cr | |
6091 | &Precedence&& Associativity &&{\rm Operators}&\cr | |
6092 | height2pt&\omit&&\omit&&\omit&\cr | |
6093 | \noalign{\hrule} | |
6094 | height2pt&\omit&&\omit&&\omit&\cr | |
6095 | &highest&&&&&\cr | |
6096 | % '176 is tilde, '~' in tt font | |
a1ab1d2a | 6097 | &1&&left&&\qquad- \char'176\ !\qquad\dag&\cr |
252b5132 RH |
6098 | &2&&left&&* / \%&\cr |
6099 | &3&&left&&+ -&\cr | |
6100 | &4&&left&&>> <<&\cr | |
6101 | &5&&left&&== != > < <= >=&\cr | |
6102 | &6&&left&&\&&\cr | |
6103 | &7&&left&&|&\cr | |
6104 | &8&&left&&{\&\&}&\cr | |
6105 | &9&&left&&||&\cr | |
6106 | &10&&right&&? :&\cr | |
6107 | &11&&right&&\qquad\&= += -= *= /=\qquad\ddag&\cr | |
6108 | &lowest&&&&&\cr | |
6109 | height2pt&\omit&&\omit&&\omit&\cr} | |
6110 | \hrule} | |
6111 | @end tex | |
6112 | @iftex | |
6113 | { | |
6114 | @obeylines@parskip=0pt@parindent=0pt | |
6115 | @dag@quad Prefix operators. | |
6116 | @ddag@quad @xref{Assignments}. | |
6117 | } | |
6118 | @end iftex | |
6119 | @c END TEXI2ROFF-KILL | |
6120 | ||
6121 | @node Evaluation | |
6122 | @subsection Evaluation | |
6123 | @cindex lazy evaluation | |
6124 | @cindex expression evaluation order | |
6125 | The linker evaluates expressions lazily. It only computes the value of | |
6126 | an expression when absolutely necessary. | |
6127 | ||
6128 | The linker needs some information, such as the value of the start | |
6129 | address of the first section, and the origins and lengths of memory | |
6130 | regions, in order to do any linking at all. These values are computed | |
6131 | as soon as possible when the linker reads in the linker script. | |
6132 | ||
6133 | However, other values (such as symbol values) are not known or needed | |
6134 | until after storage allocation. Such values are evaluated later, when | |
6135 | other information (such as the sizes of output sections) is available | |
6136 | for use in the symbol assignment expression. | |
6137 | ||
6138 | The sizes of sections cannot be known until after allocation, so | |
6139 | assignments dependent upon these are not performed until after | |
6140 | allocation. | |
6141 | ||
6142 | Some expressions, such as those depending upon the location counter | |
6143 | @samp{.}, must be evaluated during section allocation. | |
6144 | ||
6145 | If the result of an expression is required, but the value is not | |
6146 | available, then an error results. For example, a script like the | |
6147 | following | |
6148 | @smallexample | |
6149 | @group | |
6150 | SECTIONS | |
6151 | @{ | |
a1ab1d2a | 6152 | .text 9+this_isnt_constant : |
252b5132 RH |
6153 | @{ *(.text) @} |
6154 | @} | |
6155 | @end group | |
6156 | @end smallexample | |
6157 | @noindent | |
6158 | will cause the error message @samp{non constant expression for initial | |
6159 | address}. | |
6160 | ||
6161 | @node Expression Section | |
6162 | @subsection The Section of an Expression | |
6163 | @cindex expression sections | |
6164 | @cindex absolute expressions | |
6165 | @cindex relative expressions | |
6166 | @cindex absolute and relocatable symbols | |
6167 | @cindex relocatable and absolute symbols | |
6168 | @cindex symbols, relocatable and absolute | |
7542af2a AM |
6169 | Addresses and symbols may be section relative, or absolute. A section |
6170 | relative symbol is relocatable. If you request relocatable output | |
6171 | using the @samp{-r} option, a further link operation may change the | |
6172 | value of a section relative symbol. On the other hand, an absolute | |
6173 | symbol will retain the same value throughout any further link | |
6174 | operations. | |
6175 | ||
abf4be64 AM |
6176 | Some terms in linker expressions are addresses. This is true of |
6177 | section relative symbols and for builtin functions that return an | |
6178 | address, such as @code{ADDR}, @code{LOADADDR}, @code{ORIGIN} and | |
6179 | @code{SEGMENT_START}. Other terms are simply numbers, or are builtin | |
6180 | functions that return a non-address value, such as @code{LENGTH}. | |
01554a74 AM |
6181 | One complication is that unless you set @code{LD_FEATURE ("SANE_EXPR")} |
6182 | (@pxref{Miscellaneous Commands}), numbers and absolute symbols are treated | |
5c3049d2 AM |
6183 | differently depending on their location, for compatibility with older |
6184 | versions of @code{ld}. Expressions appearing outside an output | |
6185 | section definition treat all numbers as absolute addresses. | |
6186 | Expressions appearing inside an output section definition treat | |
01554a74 AM |
6187 | absolute symbols as numbers. If @code{LD_FEATURE ("SANE_EXPR")} is |
6188 | given, then absolute symbols and numbers are simply treated as numbers | |
6189 | everywhere. | |
5c3049d2 AM |
6190 | |
6191 | In the following simple example, | |
252b5132 | 6192 | |
7542af2a AM |
6193 | @smallexample |
6194 | @group | |
6195 | SECTIONS | |
6196 | @{ | |
6197 | . = 0x100; | |
6198 | __executable_start = 0x100; | |
6199 | .data : | |
6200 | @{ | |
6201 | . = 0x10; | |
6202 | __data_start = 0x10; | |
6203 | *(.data) | |
6204 | @} | |
6205 | @dots{} | |
6206 | @} | |
6207 | @end group | |
6208 | @end smallexample | |
252b5132 | 6209 | |
7542af2a AM |
6210 | both @code{.} and @code{__executable_start} are set to the absolute |
6211 | address 0x100 in the first two assignments, then both @code{.} and | |
6212 | @code{__data_start} are set to 0x10 relative to the @code{.data} | |
6213 | section in the second two assignments. | |
252b5132 | 6214 | |
5c3049d2 AM |
6215 | For expressions involving numbers, relative addresses and absolute |
6216 | addresses, ld follows these rules to evaluate terms: | |
7542af2a AM |
6217 | |
6218 | @itemize @bullet | |
6219 | @item | |
c05f749e AM |
6220 | Unary operations on an absolute address or number, and binary |
6221 | operations on two absolute addresses or two numbers, or between one | |
6222 | absolute address and a number, apply the operator to the value(s). | |
6223 | @item | |
7542af2a AM |
6224 | Unary operations on a relative address, and binary operations on two |
6225 | relative addresses in the same section or between one relative address | |
6226 | and a number, apply the operator to the offset part of the address(es). | |
6227 | @item | |
c05f749e AM |
6228 | Other binary operations, that is, between two relative addresses not |
6229 | in the same section, or between a relative address and an absolute | |
6230 | address, first convert any non-absolute term to an absolute address | |
6231 | before applying the operator. | |
7542af2a AM |
6232 | @end itemize |
6233 | ||
6234 | The result section of each sub-expression is as follows: | |
6235 | ||
6236 | @itemize @bullet | |
6237 | @item | |
6238 | An operation involving only numbers results in a number. | |
6239 | @item | |
6240 | The result of comparisons, @samp{&&} and @samp{||} is also a number. | |
6241 | @item | |
9bc8bb33 | 6242 | The result of other binary arithmetic and logical operations on two |
11e7fd74 | 6243 | relative addresses in the same section or two absolute addresses |
94b41882 AM |
6244 | (after above conversions) is also a number when |
6245 | @code{LD_FEATURE ("SANE_EXPR")} or inside an output section definition | |
6246 | but an absolute address otherwise. | |
9bc8bb33 AM |
6247 | @item |
6248 | The result of other operations on relative addresses or one | |
6249 | relative address and a number, is a relative address in the same | |
6250 | section as the relative operand(s). | |
7542af2a AM |
6251 | @item |
6252 | The result of other operations on absolute addresses (after above | |
6253 | conversions) is an absolute address. | |
6254 | @end itemize | |
252b5132 RH |
6255 | |
6256 | You can use the builtin function @code{ABSOLUTE} to force an expression | |
6257 | to be absolute when it would otherwise be relative. For example, to | |
6258 | create an absolute symbol set to the address of the end of the output | |
6259 | section @samp{.data}: | |
6260 | @smallexample | |
6261 | SECTIONS | |
6262 | @{ | |
6263 | .data : @{ *(.data) _edata = ABSOLUTE(.); @} | |
6264 | @} | |
6265 | @end smallexample | |
6266 | @noindent | |
6267 | If @samp{ABSOLUTE} were not used, @samp{_edata} would be relative to the | |
6268 | @samp{.data} section. | |
6269 | ||
7542af2a AM |
6270 | Using @code{LOADADDR} also forces an expression absolute, since this |
6271 | particular builtin function returns an absolute address. | |
6272 | ||
252b5132 RH |
6273 | @node Builtin Functions |
6274 | @subsection Builtin Functions | |
6275 | @cindex functions in expressions | |
6276 | The linker script language includes a number of builtin functions for | |
6277 | use in linker script expressions. | |
6278 | ||
6279 | @table @code | |
6280 | @item ABSOLUTE(@var{exp}) | |
6281 | @kindex ABSOLUTE(@var{exp}) | |
6282 | @cindex expression, absolute | |
6283 | Return the absolute (non-relocatable, as opposed to non-negative) value | |
6284 | of the expression @var{exp}. Primarily useful to assign an absolute | |
6285 | value to a symbol within a section definition, where symbol values are | |
6286 | normally section relative. @xref{Expression Section}. | |
6287 | ||
6288 | @item ADDR(@var{section}) | |
6289 | @kindex ADDR(@var{section}) | |
6290 | @cindex section address in expression | |
7542af2a | 6291 | Return the address (VMA) of the named @var{section}. Your |
252b5132 | 6292 | script must previously have defined the location of that section. In |
7542af2a AM |
6293 | the following example, @code{start_of_output_1}, @code{symbol_1} and |
6294 | @code{symbol_2} are assigned equivalent values, except that | |
6295 | @code{symbol_1} will be relative to the @code{.output1} section while | |
6296 | the other two will be absolute: | |
252b5132 RH |
6297 | @smallexample |
6298 | @group | |
6299 | SECTIONS @{ @dots{} | |
6300 | .output1 : | |
a1ab1d2a | 6301 | @{ |
252b5132 RH |
6302 | start_of_output_1 = ABSOLUTE(.); |
6303 | @dots{} | |
6304 | @} | |
6305 | .output : | |
6306 | @{ | |
6307 | symbol_1 = ADDR(.output1); | |
6308 | symbol_2 = start_of_output_1; | |
6309 | @} | |
6310 | @dots{} @} | |
6311 | @end group | |
6312 | @end smallexample | |
6313 | ||
876f4090 NS |
6314 | @item ALIGN(@var{align}) |
6315 | @itemx ALIGN(@var{exp},@var{align}) | |
6316 | @kindex ALIGN(@var{align}) | |
6317 | @kindex ALIGN(@var{exp},@var{align}) | |
252b5132 RH |
6318 | @cindex round up location counter |
6319 | @cindex align location counter | |
876f4090 NS |
6320 | @cindex round up expression |
6321 | @cindex align expression | |
6322 | Return the location counter (@code{.}) or arbitrary expression aligned | |
6323 | to the next @var{align} boundary. The single operand @code{ALIGN} | |
6324 | doesn't change the value of the location counter---it just does | |
6325 | arithmetic on it. The two operand @code{ALIGN} allows an arbitrary | |
6326 | expression to be aligned upwards (@code{ALIGN(@var{align})} is | |
e0a3af22 | 6327 | equivalent to @code{ALIGN(ABSOLUTE(.), @var{align})}). |
876f4090 NS |
6328 | |
6329 | Here is an example which aligns the output @code{.data} section to the | |
6330 | next @code{0x2000} byte boundary after the preceding section and sets a | |
6331 | variable within the section to the next @code{0x8000} boundary after the | |
6332 | input sections: | |
252b5132 RH |
6333 | @smallexample |
6334 | @group | |
6335 | SECTIONS @{ @dots{} | |
6336 | .data ALIGN(0x2000): @{ | |
6337 | *(.data) | |
6338 | variable = ALIGN(0x8000); | |
6339 | @} | |
6340 | @dots{} @} | |
6341 | @end group | |
6342 | @end smallexample | |
6343 | @noindent | |
6344 | The first use of @code{ALIGN} in this example specifies the location of | |
6345 | a section because it is used as the optional @var{address} attribute of | |
6346 | a section definition (@pxref{Output Section Address}). The second use | |
6347 | of @code{ALIGN} is used to defines the value of a symbol. | |
6348 | ||
6349 | The builtin function @code{NEXT} is closely related to @code{ALIGN}. | |
6350 | ||
362c1d1a NS |
6351 | @item ALIGNOF(@var{section}) |
6352 | @kindex ALIGNOF(@var{section}) | |
6353 | @cindex section alignment | |
6354 | Return the alignment in bytes of the named @var{section}, if that section has | |
6355 | been allocated. If the section has not been allocated when this is | |
6356 | evaluated, the linker will report an error. In the following example, | |
6357 | the alignment of the @code{.output} section is stored as the first | |
6358 | value in that section. | |
6359 | @smallexample | |
6360 | @group | |
6361 | SECTIONS@{ @dots{} | |
6362 | .output @{ | |
6363 | LONG (ALIGNOF (.output)) | |
6364 | @dots{} | |
6365 | @} | |
6366 | @dots{} @} | |
6367 | @end group | |
6368 | @end smallexample | |
6369 | ||
252b5132 RH |
6370 | @item BLOCK(@var{exp}) |
6371 | @kindex BLOCK(@var{exp}) | |
6372 | This is a synonym for @code{ALIGN}, for compatibility with older linker | |
6373 | scripts. It is most often seen when setting the address of an output | |
6374 | section. | |
6375 | ||
2d20f7bf JJ |
6376 | @item DATA_SEGMENT_ALIGN(@var{maxpagesize}, @var{commonpagesize}) |
6377 | @kindex DATA_SEGMENT_ALIGN(@var{maxpagesize}, @var{commonpagesize}) | |
6378 | This is equivalent to either | |
6379 | @smallexample | |
6380 | (ALIGN(@var{maxpagesize}) + (. & (@var{maxpagesize} - 1))) | |
6381 | @end smallexample | |
6382 | or | |
6383 | @smallexample | |
fe6052e1 AM |
6384 | (ALIGN(@var{maxpagesize}) |
6385 | + ((. + @var{commonpagesize} - 1) & (@var{maxpagesize} - @var{commonpagesize}))) | |
2d20f7bf JJ |
6386 | @end smallexample |
6387 | @noindent | |
6388 | depending on whether the latter uses fewer @var{commonpagesize} sized pages | |
6389 | for the data segment (area between the result of this expression and | |
6390 | @code{DATA_SEGMENT_END}) than the former or not. | |
6391 | If the latter form is used, it means @var{commonpagesize} bytes of runtime | |
6392 | memory will be saved at the expense of up to @var{commonpagesize} wasted | |
6393 | bytes in the on-disk file. | |
6394 | ||
6395 | This expression can only be used directly in @code{SECTIONS} commands, not in | |
6396 | any output section descriptions and only once in the linker script. | |
6397 | @var{commonpagesize} should be less or equal to @var{maxpagesize} and should | |
def5c83c AM |
6398 | be the system page size the object wants to be optimized for while still |
6399 | running on system page sizes up to @var{maxpagesize}. Note however | |
6400 | that @samp{-z relro} protection will not be effective if the system | |
6401 | page size is larger than @var{commonpagesize}. | |
2d20f7bf JJ |
6402 | |
6403 | @noindent | |
6404 | Example: | |
6405 | @smallexample | |
6406 | . = DATA_SEGMENT_ALIGN(0x10000, 0x2000); | |
6407 | @end smallexample | |
6408 | ||
6409 | @item DATA_SEGMENT_END(@var{exp}) | |
6410 | @kindex DATA_SEGMENT_END(@var{exp}) | |
6411 | This defines the end of data segment for @code{DATA_SEGMENT_ALIGN} | |
6412 | evaluation purposes. | |
6413 | ||
6414 | @smallexample | |
6415 | . = DATA_SEGMENT_END(.); | |
6416 | @end smallexample | |
6417 | ||
a4f5ad88 JJ |
6418 | @item DATA_SEGMENT_RELRO_END(@var{offset}, @var{exp}) |
6419 | @kindex DATA_SEGMENT_RELRO_END(@var{offset}, @var{exp}) | |
6420 | This defines the end of the @code{PT_GNU_RELRO} segment when | |
eec2f3ed | 6421 | @samp{-z relro} option is used. |
a4f5ad88 JJ |
6422 | When @samp{-z relro} option is not present, @code{DATA_SEGMENT_RELRO_END} |
6423 | does nothing, otherwise @code{DATA_SEGMENT_ALIGN} is padded so that | |
def5c83c AM |
6424 | @var{exp} + @var{offset} is aligned to the @var{commonpagesize} |
6425 | argument given to @code{DATA_SEGMENT_ALIGN}. If present in the linker | |
6426 | script, it must be placed between @code{DATA_SEGMENT_ALIGN} and | |
eec2f3ed AM |
6427 | @code{DATA_SEGMENT_END}. Evaluates to the second argument plus any |
6428 | padding needed at the end of the @code{PT_GNU_RELRO} segment due to | |
6429 | section alignment. | |
a4f5ad88 JJ |
6430 | |
6431 | @smallexample | |
6432 | . = DATA_SEGMENT_RELRO_END(24, .); | |
6433 | @end smallexample | |
6434 | ||
252b5132 RH |
6435 | @item DEFINED(@var{symbol}) |
6436 | @kindex DEFINED(@var{symbol}) | |
6437 | @cindex symbol defaults | |
6438 | Return 1 if @var{symbol} is in the linker global symbol table and is | |
420e579c HPN |
6439 | defined before the statement using DEFINED in the script, otherwise |
6440 | return 0. You can use this function to provide | |
252b5132 RH |
6441 | default values for symbols. For example, the following script fragment |
6442 | shows how to set a global symbol @samp{begin} to the first location in | |
6443 | the @samp{.text} section---but if a symbol called @samp{begin} already | |
6444 | existed, its value is preserved: | |
6445 | ||
6446 | @smallexample | |
6447 | @group | |
6448 | SECTIONS @{ @dots{} | |
6449 | .text : @{ | |
6450 | begin = DEFINED(begin) ? begin : . ; | |
6451 | @dots{} | |
6452 | @} | |
6453 | @dots{} | |
6454 | @} | |
6455 | @end group | |
6456 | @end smallexample | |
6457 | ||
3ec57632 NC |
6458 | @item LENGTH(@var{memory}) |
6459 | @kindex LENGTH(@var{memory}) | |
6460 | Return the length of the memory region named @var{memory}. | |
6461 | ||
252b5132 RH |
6462 | @item LOADADDR(@var{section}) |
6463 | @kindex LOADADDR(@var{section}) | |
6464 | @cindex section load address in expression | |
7542af2a | 6465 | Return the absolute LMA of the named @var{section}. (@pxref{Output |
252b5132 RH |
6466 | Section LMA}). |
6467 | ||
2e53f7d6 NC |
6468 | @item LOG2CEIL(@var{exp}) |
6469 | @kindex LOG2CEIL(@var{exp}) | |
6470 | Return the binary logarithm of @var{exp} rounded towards infinity. | |
6471 | @code{LOG2CEIL(0)} returns 0. | |
6472 | ||
252b5132 RH |
6473 | @kindex MAX |
6474 | @item MAX(@var{exp1}, @var{exp2}) | |
6475 | Returns the maximum of @var{exp1} and @var{exp2}. | |
6476 | ||
6477 | @kindex MIN | |
6478 | @item MIN(@var{exp1}, @var{exp2}) | |
6479 | Returns the minimum of @var{exp1} and @var{exp2}. | |
6480 | ||
6481 | @item NEXT(@var{exp}) | |
6482 | @kindex NEXT(@var{exp}) | |
6483 | @cindex unallocated address, next | |
6484 | Return the next unallocated address that is a multiple of @var{exp}. | |
6485 | This function is closely related to @code{ALIGN(@var{exp})}; unless you | |
6486 | use the @code{MEMORY} command to define discontinuous memory for the | |
6487 | output file, the two functions are equivalent. | |
6488 | ||
3ec57632 NC |
6489 | @item ORIGIN(@var{memory}) |
6490 | @kindex ORIGIN(@var{memory}) | |
6491 | Return the origin of the memory region named @var{memory}. | |
6492 | ||
ba916c8a MM |
6493 | @item SEGMENT_START(@var{segment}, @var{default}) |
6494 | @kindex SEGMENT_START(@var{segment}, @var{default}) | |
6495 | Return the base address of the named @var{segment}. If an explicit | |
c5da8c7d NC |
6496 | value has already been given for this segment (with a command-line |
6497 | @samp{-T} option) then that value will be returned otherwise the value | |
6498 | will be @var{default}. At present, the @samp{-T} command-line option | |
6499 | can only be used to set the base address for the ``text'', ``data'', and | |
7542af2a | 6500 | ``bss'' sections, but you can use @code{SEGMENT_START} with any segment |
ba916c8a MM |
6501 | name. |
6502 | ||
252b5132 RH |
6503 | @item SIZEOF(@var{section}) |
6504 | @kindex SIZEOF(@var{section}) | |
6505 | @cindex section size | |
6506 | Return the size in bytes of the named @var{section}, if that section has | |
6507 | been allocated. If the section has not been allocated when this is | |
6508 | evaluated, the linker will report an error. In the following example, | |
6509 | @code{symbol_1} and @code{symbol_2} are assigned identical values: | |
6510 | @smallexample | |
6511 | @group | |
6512 | SECTIONS@{ @dots{} | |
6513 | .output @{ | |
6514 | .start = . ; | |
6515 | @dots{} | |
6516 | .end = . ; | |
6517 | @} | |
6518 | symbol_1 = .end - .start ; | |
6519 | symbol_2 = SIZEOF(.output); | |
6520 | @dots{} @} | |
6521 | @end group | |
6522 | @end smallexample | |
6523 | ||
6524 | @item SIZEOF_HEADERS | |
6525 | @itemx sizeof_headers | |
6526 | @kindex SIZEOF_HEADERS | |
6527 | @cindex header size | |
6528 | Return the size in bytes of the output file's headers. This is | |
6529 | information which appears at the start of the output file. You can use | |
6530 | this number when setting the start address of the first section, if you | |
6531 | choose, to facilitate paging. | |
6532 | ||
6533 | @cindex not enough room for program headers | |
6534 | @cindex program headers, not enough room | |
6535 | When producing an ELF output file, if the linker script uses the | |
6536 | @code{SIZEOF_HEADERS} builtin function, the linker must compute the | |
6537 | number of program headers before it has determined all the section | |
6538 | addresses and sizes. If the linker later discovers that it needs | |
6539 | additional program headers, it will report an error @samp{not enough | |
6540 | room for program headers}. To avoid this error, you must avoid using | |
6541 | the @code{SIZEOF_HEADERS} function, or you must rework your linker | |
6542 | script to avoid forcing the linker to use additional program headers, or | |
6543 | you must define the program headers yourself using the @code{PHDRS} | |
6544 | command (@pxref{PHDRS}). | |
6545 | @end table | |
6546 | ||
6547 | @node Implicit Linker Scripts | |
6548 | @section Implicit Linker Scripts | |
6549 | @cindex implicit linker scripts | |
6550 | If you specify a linker input file which the linker can not recognize as | |
6551 | an object file or an archive file, it will try to read the file as a | |
6552 | linker script. If the file can not be parsed as a linker script, the | |
6553 | linker will report an error. | |
6554 | ||
6555 | An implicit linker script will not replace the default linker script. | |
6556 | ||
6557 | Typically an implicit linker script would contain only symbol | |
6558 | assignments, or the @code{INPUT}, @code{GROUP}, or @code{VERSION} | |
6559 | commands. | |
6560 | ||
6561 | Any input files read because of an implicit linker script will be read | |
6562 | at the position in the command line where the implicit linker script was | |
6563 | read. This can affect archive searching. | |
6564 | ||
6565 | @ifset GENERIC | |
6566 | @node Machine Dependent | |
6567 | @chapter Machine Dependent Features | |
6568 | ||
6569 | @cindex machine dependencies | |
ff5dcc92 SC |
6570 | @command{ld} has additional features on some platforms; the following |
6571 | sections describe them. Machines where @command{ld} has no additional | |
252b5132 RH |
6572 | functionality are not listed. |
6573 | ||
6574 | @menu | |
36f63dca NC |
6575 | @ifset H8300 |
6576 | * H8/300:: @command{ld} and the H8/300 | |
6577 | @end ifset | |
7ca01ed9 NC |
6578 | @ifset M68HC11 |
6579 | * M68HC11/68HC12:: @code{ld} and the Motorola 68HC11 and 68HC12 families | |
6580 | @end ifset | |
36f63dca NC |
6581 | @ifset ARM |
6582 | * ARM:: @command{ld} and the ARM family | |
6583 | @end ifset | |
6584 | @ifset HPPA | |
6585 | * HPPA ELF32:: @command{ld} and HPPA 32-bit ELF | |
6586 | @end ifset | |
7fb9f789 NC |
6587 | @ifset M68K |
6588 | * M68K:: @command{ld} and the Motorola 68K family | |
6589 | @end ifset | |
833794fc MR |
6590 | @ifset MIPS |
6591 | * MIPS:: @command{ld} and the MIPS family | |
6592 | @end ifset | |
3c3bdf30 | 6593 | @ifset MMIX |
36f63dca | 6594 | * MMIX:: @command{ld} and MMIX |
3c3bdf30 | 6595 | @end ifset |
2469cfa2 | 6596 | @ifset MSP430 |
36f63dca | 6597 | * MSP430:: @command{ld} and MSP430 |
2469cfa2 | 6598 | @end ifset |
35c08157 KLC |
6599 | @ifset NDS32 |
6600 | * NDS32:: @command{ld} and NDS32 | |
6601 | @end ifset | |
78058a5e SL |
6602 | @ifset NIOSII |
6603 | * Nios II:: @command{ld} and the Altera Nios II | |
6604 | @end ifset | |
2a60a7a8 AM |
6605 | @ifset POWERPC |
6606 | * PowerPC ELF32:: @command{ld} and PowerPC 32-bit ELF Support | |
6607 | @end ifset | |
6608 | @ifset POWERPC64 | |
6609 | * PowerPC64 ELF64:: @command{ld} and PowerPC64 64-bit ELF Support | |
6610 | @end ifset | |
b4cbbe8f AK |
6611 | @ifset S/390 |
6612 | * S/390 ELF:: @command{ld} and S/390 ELF Support | |
6613 | @end ifset | |
49fa1e15 AM |
6614 | @ifset SPU |
6615 | * SPU ELF:: @command{ld} and SPU ELF Support | |
6616 | @end ifset | |
74459f0e | 6617 | @ifset TICOFF |
ff5dcc92 | 6618 | * TI COFF:: @command{ld} and TI COFF |
74459f0e | 6619 | @end ifset |
2ca22b03 NC |
6620 | @ifset WIN32 |
6621 | * WIN32:: @command{ld} and WIN32 (cygwin/mingw) | |
6622 | @end ifset | |
e0001a05 NC |
6623 | @ifset XTENSA |
6624 | * Xtensa:: @command{ld} and Xtensa Processors | |
6625 | @end ifset | |
252b5132 RH |
6626 | @end menu |
6627 | @end ifset | |
6628 | ||
252b5132 RH |
6629 | @ifset H8300 |
6630 | @ifclear GENERIC | |
6631 | @raisesections | |
6632 | @end ifclear | |
6633 | ||
6634 | @node H8/300 | |
ff5dcc92 | 6635 | @section @command{ld} and the H8/300 |
252b5132 RH |
6636 | |
6637 | @cindex H8/300 support | |
ff5dcc92 | 6638 | For the H8/300, @command{ld} can perform these global optimizations when |
252b5132 RH |
6639 | you specify the @samp{--relax} command-line option. |
6640 | ||
6641 | @table @emph | |
6642 | @cindex relaxing on H8/300 | |
6643 | @item relaxing address modes | |
ff5dcc92 | 6644 | @command{ld} finds all @code{jsr} and @code{jmp} instructions whose |
252b5132 RH |
6645 | targets are within eight bits, and turns them into eight-bit |
6646 | program-counter relative @code{bsr} and @code{bra} instructions, | |
6647 | respectively. | |
6648 | ||
6649 | @cindex synthesizing on H8/300 | |
6650 | @item synthesizing instructions | |
81f5558e | 6651 | @c FIXME: specifically mov.b, or any mov instructions really? -> mov.b only, at least on H8, H8H, H8S |
ff5dcc92 | 6652 | @command{ld} finds all @code{mov.b} instructions which use the |
252b5132 RH |
6653 | sixteen-bit absolute address form, but refer to the top |
6654 | page of memory, and changes them to use the eight-bit address form. | |
6655 | (That is: the linker turns @samp{mov.b @code{@@}@var{aa}:16} into | |
6656 | @samp{mov.b @code{@@}@var{aa}:8} whenever the address @var{aa} is in the | |
6657 | top page of memory). | |
1502569c | 6658 | |
81f5558e NC |
6659 | @command{ld} finds all @code{mov} instructions which use the register |
6660 | indirect with 32-bit displacement addressing mode, but use a small | |
6661 | displacement inside 16-bit displacement range, and changes them to use | |
6662 | the 16-bit displacement form. (That is: the linker turns @samp{mov.b | |
6663 | @code{@@}@var{d}:32,ERx} into @samp{mov.b @code{@@}@var{d}:16,ERx} | |
6664 | whenever the displacement @var{d} is in the 16 bit signed integer | |
6665 | range. Only implemented in ELF-format ld). | |
6666 | ||
1502569c | 6667 | @item bit manipulation instructions |
c0065db7 | 6668 | @command{ld} finds all bit manipulation instructions like @code{band, bclr, |
1502569c | 6669 | biand, bild, bior, bist, bixor, bld, bnot, bor, bset, bst, btst, bxor} |
c0065db7 | 6670 | which use 32 bit and 16 bit absolute address form, but refer to the top |
1502569c NC |
6671 | page of memory, and changes them to use the 8 bit address form. |
6672 | (That is: the linker turns @samp{bset #xx:3,@code{@@}@var{aa}:32} into | |
c0065db7 | 6673 | @samp{bset #xx:3,@code{@@}@var{aa}:8} whenever the address @var{aa} is in |
1502569c NC |
6674 | the top page of memory). |
6675 | ||
6676 | @item system control instructions | |
c0065db7 RM |
6677 | @command{ld} finds all @code{ldc.w, stc.w} instructions which use the |
6678 | 32 bit absolute address form, but refer to the top page of memory, and | |
1502569c NC |
6679 | changes them to use 16 bit address form. |
6680 | (That is: the linker turns @samp{ldc.w @code{@@}@var{aa}:32,ccr} into | |
c0065db7 | 6681 | @samp{ldc.w @code{@@}@var{aa}:16,ccr} whenever the address @var{aa} is in |
1502569c | 6682 | the top page of memory). |
252b5132 RH |
6683 | @end table |
6684 | ||
6685 | @ifclear GENERIC | |
6686 | @lowersections | |
6687 | @end ifclear | |
6688 | @end ifset | |
6689 | ||
36f63dca | 6690 | @ifclear GENERIC |
c2dcd04e | 6691 | @ifset Renesas |
36f63dca | 6692 | @c This stuff is pointless to say unless you're especially concerned |
c2dcd04e NC |
6693 | @c with Renesas chips; don't enable it for generic case, please. |
6694 | @node Renesas | |
6695 | @chapter @command{ld} and Other Renesas Chips | |
36f63dca | 6696 | |
c2dcd04e NC |
6697 | @command{ld} also supports the Renesas (formerly Hitachi) H8/300H, |
6698 | H8/500, and SH chips. No special features, commands, or command-line | |
6699 | options are required for these chips. | |
36f63dca NC |
6700 | @end ifset |
6701 | @end ifclear | |
6702 | ||
36f63dca NC |
6703 | @ifset ARM |
6704 | @ifclear GENERIC | |
6705 | @raisesections | |
6706 | @end ifclear | |
6707 | ||
93fd0973 SC |
6708 | @ifset M68HC11 |
6709 | @ifclear GENERIC | |
6710 | @raisesections | |
6711 | @end ifclear | |
6712 | ||
6713 | @node M68HC11/68HC12 | |
6714 | @section @command{ld} and the Motorola 68HC11 and 68HC12 families | |
6715 | ||
6716 | @cindex M68HC11 and 68HC12 support | |
6717 | ||
6718 | @subsection Linker Relaxation | |
6719 | ||
6720 | For the Motorola 68HC11, @command{ld} can perform these global | |
6721 | optimizations when you specify the @samp{--relax} command-line option. | |
6722 | ||
6723 | @table @emph | |
6724 | @cindex relaxing on M68HC11 | |
6725 | @item relaxing address modes | |
6726 | @command{ld} finds all @code{jsr} and @code{jmp} instructions whose | |
6727 | targets are within eight bits, and turns them into eight-bit | |
6728 | program-counter relative @code{bsr} and @code{bra} instructions, | |
6729 | respectively. | |
6730 | ||
6731 | @command{ld} also looks at all 16-bit extended addressing modes and | |
6732 | transforms them in a direct addressing mode when the address is in | |
6733 | page 0 (between 0 and 0x0ff). | |
6734 | ||
6735 | @item relaxing gcc instruction group | |
6736 | When @command{gcc} is called with @option{-mrelax}, it can emit group | |
6737 | of instructions that the linker can optimize to use a 68HC11 direct | |
6738 | addressing mode. These instructions consists of @code{bclr} or | |
6739 | @code{bset} instructions. | |
6740 | ||
6741 | @end table | |
6742 | ||
6743 | @subsection Trampoline Generation | |
6744 | ||
6745 | @cindex trampoline generation on M68HC11 | |
6746 | @cindex trampoline generation on M68HC12 | |
6747 | For 68HC11 and 68HC12, @command{ld} can generate trampoline code to | |
6748 | call a far function using a normal @code{jsr} instruction. The linker | |
c0065db7 | 6749 | will also change the relocation to some far function to use the |
93fd0973 SC |
6750 | trampoline address instead of the function address. This is typically the |
6751 | case when a pointer to a function is taken. The pointer will in fact | |
6752 | point to the function trampoline. | |
6753 | ||
6754 | @ifclear GENERIC | |
6755 | @lowersections | |
6756 | @end ifclear | |
6757 | @end ifset | |
6758 | ||
36f63dca | 6759 | @node ARM |
3674e28a | 6760 | @section @command{ld} and the ARM family |
36f63dca NC |
6761 | |
6762 | @cindex ARM interworking support | |
6763 | @kindex --support-old-code | |
6764 | For the ARM, @command{ld} will generate code stubs to allow functions calls | |
b45619c0 | 6765 | between ARM and Thumb code. These stubs only work with code that has |
36f63dca NC |
6766 | been compiled and assembled with the @samp{-mthumb-interwork} command |
6767 | line option. If it is necessary to link with old ARM object files or | |
6768 | libraries, which have not been compiled with the -mthumb-interwork | |
a05a5b64 | 6769 | option then the @samp{--support-old-code} command-line switch should be |
36f63dca NC |
6770 | given to the linker. This will make it generate larger stub functions |
6771 | which will work with non-interworking aware ARM code. Note, however, | |
6772 | the linker does not support generating stubs for function calls to | |
6773 | non-interworking aware Thumb code. | |
6774 | ||
6775 | @cindex thumb entry point | |
6776 | @cindex entry point, thumb | |
6777 | @kindex --thumb-entry=@var{entry} | |
6778 | The @samp{--thumb-entry} switch is a duplicate of the generic | |
6779 | @samp{--entry} switch, in that it sets the program's starting address. | |
6780 | But it also sets the bottom bit of the address, so that it can be | |
6781 | branched to using a BX instruction, and the program will start | |
6782 | executing in Thumb mode straight away. | |
6783 | ||
ce11ba6c KT |
6784 | @cindex PE import table prefixing |
6785 | @kindex --use-nul-prefixed-import-tables | |
6786 | The @samp{--use-nul-prefixed-import-tables} switch is specifying, that | |
6787 | the import tables idata4 and idata5 have to be generated with a zero | |
11e7fd74 | 6788 | element prefix for import libraries. This is the old style to generate |
ce11ba6c KT |
6789 | import tables. By default this option is turned off. |
6790 | ||
e489d0ae PB |
6791 | @cindex BE8 |
6792 | @kindex --be8 | |
6793 | The @samp{--be8} switch instructs @command{ld} to generate BE8 format | |
080bb7bb NC |
6794 | executables. This option is only valid when linking big-endian |
6795 | objects - ie ones which have been assembled with the @option{-EB} | |
6796 | option. The resulting image will contain big-endian data and | |
6797 | little-endian code. | |
e489d0ae | 6798 | |
3674e28a PB |
6799 | @cindex TARGET1 |
6800 | @kindex --target1-rel | |
6801 | @kindex --target1-abs | |
6802 | The @samp{R_ARM_TARGET1} relocation is typically used for entries in the | |
6803 | @samp{.init_array} section. It is interpreted as either @samp{R_ARM_REL32} | |
6804 | or @samp{R_ARM_ABS32}, depending on the target. The @samp{--target1-rel} | |
6805 | and @samp{--target1-abs} switches override the default. | |
6806 | ||
6807 | @cindex TARGET2 | |
6808 | @kindex --target2=@var{type} | |
6809 | The @samp{--target2=type} switch overrides the default definition of the | |
6810 | @samp{R_ARM_TARGET2} relocation. Valid values for @samp{type}, their | |
6811 | meanings, and target defaults are as follows: | |
6812 | @table @samp | |
6813 | @item rel | |
eeac373a PB |
6814 | @samp{R_ARM_REL32} (arm*-*-elf, arm*-*-eabi) |
6815 | @item abs | |
6816 | @samp{R_ARM_ABS32} (arm*-*-symbianelf) | |
3674e28a PB |
6817 | @item got-rel |
6818 | @samp{R_ARM_GOT_PREL} (arm*-*-linux, arm*-*-*bsd) | |
6819 | @end table | |
6820 | ||
319850b4 JB |
6821 | @cindex FIX_V4BX |
6822 | @kindex --fix-v4bx | |
6823 | The @samp{R_ARM_V4BX} relocation (defined by the ARM AAELF | |
6824 | specification) enables objects compiled for the ARMv4 architecture to be | |
6825 | interworking-safe when linked with other objects compiled for ARMv4t, but | |
6826 | also allows pure ARMv4 binaries to be built from the same ARMv4 objects. | |
6827 | ||
6828 | In the latter case, the switch @option{--fix-v4bx} must be passed to the | |
6829 | linker, which causes v4t @code{BX rM} instructions to be rewritten as | |
6830 | @code{MOV PC,rM}, since v4 processors do not have a @code{BX} instruction. | |
6831 | ||
6832 | In the former case, the switch should not be used, and @samp{R_ARM_V4BX} | |
6833 | relocations are ignored. | |
6834 | ||
845b51d6 PB |
6835 | @cindex FIX_V4BX_INTERWORKING |
6836 | @kindex --fix-v4bx-interworking | |
6837 | Replace @code{BX rM} instructions identified by @samp{R_ARM_V4BX} | |
6838 | relocations with a branch to the following veneer: | |
6839 | ||
6840 | @smallexample | |
6841 | TST rM, #1 | |
6842 | MOVEQ PC, rM | |
6843 | BX Rn | |
6844 | @end smallexample | |
6845 | ||
6846 | This allows generation of libraries/applications that work on ARMv4 cores | |
6847 | and are still interworking safe. Note that the above veneer clobbers the | |
11e7fd74 | 6848 | condition flags, so may cause incorrect program behavior in rare cases. |
845b51d6 | 6849 | |
33bfe774 JB |
6850 | @cindex USE_BLX |
6851 | @kindex --use-blx | |
6852 | The @samp{--use-blx} switch enables the linker to use ARM/Thumb | |
6853 | BLX instructions (available on ARMv5t and above) in various | |
6854 | situations. Currently it is used to perform calls via the PLT from Thumb | |
6855 | code using BLX rather than using BX and a mode-switching stub before | |
6856 | each PLT entry. This should lead to such calls executing slightly faster. | |
6857 | ||
6858 | This option is enabled implicitly for SymbianOS, so there is no need to | |
6859 | specify it if you are using that target. | |
6860 | ||
c6dd86c6 JB |
6861 | @cindex VFP11_DENORM_FIX |
6862 | @kindex --vfp11-denorm-fix | |
6863 | The @samp{--vfp11-denorm-fix} switch enables a link-time workaround for a | |
6864 | bug in certain VFP11 coprocessor hardware, which sometimes allows | |
6865 | instructions with denorm operands (which must be handled by support code) | |
6866 | to have those operands overwritten by subsequent instructions before | |
6867 | the support code can read the intended values. | |
6868 | ||
6869 | The bug may be avoided in scalar mode if you allow at least one | |
6870 | intervening instruction between a VFP11 instruction which uses a register | |
6871 | and another instruction which writes to the same register, or at least two | |
6872 | intervening instructions if vector mode is in use. The bug only affects | |
6873 | full-compliance floating-point mode: you do not need this workaround if | |
6874 | you are using "runfast" mode. Please contact ARM for further details. | |
6875 | ||
6876 | If you know you are using buggy VFP11 hardware, you can | |
6877 | enable this workaround by specifying the linker option | |
6878 | @samp{--vfp-denorm-fix=scalar} if you are using the VFP11 scalar | |
6879 | mode only, or @samp{--vfp-denorm-fix=vector} if you are using | |
6880 | vector mode (the latter also works for scalar code). The default is | |
6881 | @samp{--vfp-denorm-fix=none}. | |
6882 | ||
6883 | If the workaround is enabled, instructions are scanned for | |
6884 | potentially-troublesome sequences, and a veneer is created for each | |
6885 | such sequence which may trigger the erratum. The veneer consists of the | |
6886 | first instruction of the sequence and a branch back to the subsequent | |
6887 | instruction. The original instruction is then replaced with a branch to | |
6888 | the veneer. The extra cycles required to call and return from the veneer | |
6889 | are sufficient to avoid the erratum in both the scalar and vector cases. | |
6890 | ||
2de70689 MGD |
6891 | @cindex ARM1176 erratum workaround |
6892 | @kindex --fix-arm1176 | |
6893 | @kindex --no-fix-arm1176 | |
9d5777a3 RM |
6894 | The @samp{--fix-arm1176} switch enables a link-time workaround for an erratum |
6895 | in certain ARM1176 processors. The workaround is enabled by default if you | |
6896 | are targeting ARM v6 (excluding ARM v6T2) or earlier. It can be disabled | |
2de70689 MGD |
6897 | unconditionally by specifying @samp{--no-fix-arm1176}. |
6898 | ||
9d5777a3 | 6899 | Further information is available in the ``ARM1176JZ-S and ARM1176JZF-S |
11e7fd74 | 6900 | Programmer Advice Notice'' available on the ARM documentation website at: |
2de70689 MGD |
6901 | http://infocenter.arm.com/. |
6902 | ||
a504d23a LA |
6903 | @cindex STM32L4xx erratum workaround |
6904 | @kindex --fix-stm32l4xx-629360 | |
6905 | ||
6906 | The @samp{--fix-stm32l4xx-629360} switch enables a link-time | |
6907 | workaround for a bug in the bus matrix / memory controller for some of | |
6908 | the STM32 Cortex-M4 based products (STM32L4xx). When accessing | |
6909 | off-chip memory via the affected bus for bus reads of 9 words or more, | |
6910 | the bus can generate corrupt data and/or abort. These are only | |
6911 | core-initiated accesses (not DMA), and might affect any access: | |
6912 | integer loads such as LDM, POP and floating-point loads such as VLDM, | |
6913 | VPOP. Stores are not affected. | |
6914 | ||
6915 | The bug can be avoided by splitting memory accesses into the | |
6916 | necessary chunks to keep bus reads below 8 words. | |
6917 | ||
6918 | The workaround is not enabled by default, this is equivalent to use | |
6919 | @samp{--fix-stm32l4xx-629360=none}. If you know you are using buggy | |
6920 | STM32L4xx hardware, you can enable the workaround by specifying the | |
6921 | linker option @samp{--fix-stm32l4xx-629360}, or the equivalent | |
6922 | @samp{--fix-stm32l4xx-629360=default}. | |
6923 | ||
6924 | If the workaround is enabled, instructions are scanned for | |
6925 | potentially-troublesome sequences, and a veneer is created for each | |
6926 | such sequence which may trigger the erratum. The veneer consists in a | |
6927 | replacement sequence emulating the behaviour of the original one and a | |
6928 | branch back to the subsequent instruction. The original instruction is | |
6929 | then replaced with a branch to the veneer. | |
6930 | ||
6931 | The workaround does not always preserve the memory access order for | |
6932 | the LDMDB instruction, when the instruction loads the PC. | |
6933 | ||
6934 | The workaround is not able to handle problematic instructions when | |
6935 | they are in the middle of an IT block, since a branch is not allowed | |
6936 | there. In that case, the linker reports a warning and no replacement | |
6937 | occurs. | |
6938 | ||
6939 | The workaround is not able to replace problematic instructions with a | |
6940 | PC-relative branch instruction if the @samp{.text} section is too | |
6941 | large. In that case, when the branch that replaces the original code | |
6942 | cannot be encoded, the linker reports a warning and no replacement | |
6943 | occurs. | |
6944 | ||
bf21ed78 MS |
6945 | @cindex NO_ENUM_SIZE_WARNING |
6946 | @kindex --no-enum-size-warning | |
726150b7 | 6947 | The @option{--no-enum-size-warning} switch prevents the linker from |
bf21ed78 MS |
6948 | warning when linking object files that specify incompatible EABI |
6949 | enumeration size attributes. For example, with this switch enabled, | |
6950 | linking of an object file using 32-bit enumeration values with another | |
6951 | using enumeration values fitted into the smallest possible space will | |
6952 | not be diagnosed. | |
a9dc9481 JM |
6953 | |
6954 | @cindex NO_WCHAR_SIZE_WARNING | |
6955 | @kindex --no-wchar-size-warning | |
6956 | The @option{--no-wchar-size-warning} switch prevents the linker from | |
6957 | warning when linking object files that specify incompatible EABI | |
6958 | @code{wchar_t} size attributes. For example, with this switch enabled, | |
6959 | linking of an object file using 32-bit @code{wchar_t} values with another | |
6960 | using 16-bit @code{wchar_t} values will not be diagnosed. | |
bf21ed78 | 6961 | |
726150b7 NC |
6962 | @cindex PIC_VENEER |
6963 | @kindex --pic-veneer | |
6964 | The @samp{--pic-veneer} switch makes the linker use PIC sequences for | |
6965 | ARM/Thumb interworking veneers, even if the rest of the binary | |
6966 | is not PIC. This avoids problems on uClinux targets where | |
6967 | @samp{--emit-relocs} is used to generate relocatable binaries. | |
6968 | ||
6969 | @cindex STUB_GROUP_SIZE | |
6970 | @kindex --stub-group-size=@var{N} | |
6971 | The linker will automatically generate and insert small sequences of | |
6972 | code into a linked ARM ELF executable whenever an attempt is made to | |
6973 | perform a function call to a symbol that is too far away. The | |
6974 | placement of these sequences of instructions - called stubs - is | |
a05a5b64 | 6975 | controlled by the command-line option @option{--stub-group-size=N}. |
726150b7 | 6976 | The placement is important because a poor choice can create a need for |
11e7fd74 | 6977 | duplicate stubs, increasing the code size. The linker will try to |
726150b7 NC |
6978 | group stubs together in order to reduce interruptions to the flow of |
6979 | code, but it needs guidance as to how big these groups should be and | |
6980 | where they should be placed. | |
6981 | ||
6982 | The value of @samp{N}, the parameter to the | |
6983 | @option{--stub-group-size=} option controls where the stub groups are | |
07d72278 | 6984 | placed. If it is negative then all stubs are placed after the first |
726150b7 NC |
6985 | branch that needs them. If it is positive then the stubs can be |
6986 | placed either before or after the branches that need them. If the | |
6987 | value of @samp{N} is 1 (either +1 or -1) then the linker will choose | |
6988 | exactly where to place groups of stubs, using its built in heuristics. | |
6989 | A value of @samp{N} greater than 1 (or smaller than -1) tells the | |
6990 | linker that a single group of stubs can service at most @samp{N} bytes | |
6991 | from the input sections. | |
6992 | ||
6993 | The default, if @option{--stub-group-size=} is not specified, is | |
6994 | @samp{N = +1}. | |
6995 | ||
1a51c1a4 NC |
6996 | Farcalls stubs insertion is fully supported for the ARM-EABI target |
6997 | only, because it relies on object files properties not present | |
6998 | otherwise. | |
6999 | ||
1db37fe6 YG |
7000 | @cindex Cortex-A8 erratum workaround |
7001 | @kindex --fix-cortex-a8 | |
7002 | @kindex --no-fix-cortex-a8 | |
7003 | 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}. | |
7004 | ||
7005 | The erratum only affects Thumb-2 code. Please contact ARM for further details. | |
7006 | ||
68fcca92 JW |
7007 | @cindex Cortex-A53 erratum 835769 workaround |
7008 | @kindex --fix-cortex-a53-835769 | |
7009 | @kindex --no-fix-cortex-a53-835769 | |
7010 | The @samp{--fix-cortex-a53-835769} switch enables a link-time workaround for erratum 835769 present on certain early revisions of Cortex-A53 processors. The workaround is disabled by default. It can be enabled by specifying @samp{--fix-cortex-a53-835769}, or disabled unconditionally by specifying @samp{--no-fix-cortex-a53-835769}. | |
7011 | ||
7012 | Please contact ARM for further details. | |
7013 | ||
1db37fe6 YG |
7014 | @kindex --merge-exidx-entries |
7015 | @kindex --no-merge-exidx-entries | |
7016 | @cindex Merging exidx entries | |
7017 | The @samp{--no-merge-exidx-entries} switch disables the merging of adjacent exidx entries in debuginfo. | |
7018 | ||
7019 | @kindex --long-plt | |
7020 | @cindex 32-bit PLT entries | |
7021 | The @samp{--long-plt} option enables the use of 16 byte PLT entries | |
7022 | which support up to 4Gb of code. The default is to use 12 byte PLT | |
7023 | entries which only support 512Mb of code. | |
7024 | ||
1f56df9d JW |
7025 | @kindex --no-apply-dynamic-relocs |
7026 | @cindex AArch64 rela addend | |
7027 | The @samp{--no-apply-dynamic-relocs} option makes AArch64 linker do not apply | |
7028 | link-time values for dynamic relocations. | |
7029 | ||
4ba2ef8f TP |
7030 | @cindex Placement of SG veneers |
7031 | All SG veneers are placed in the special output section @code{.gnu.sgstubs}. | |
a05a5b64 | 7032 | Its start address must be set, either with the command-line option |
4ba2ef8f TP |
7033 | @samp{--section-start} or in a linker script, to indicate where to place these |
7034 | veneers in memory. | |
7035 | ||
54ddd295 TP |
7036 | @kindex --cmse-implib |
7037 | @cindex Secure gateway import library | |
7038 | The @samp{--cmse-implib} option requests that the import libraries | |
7039 | specified by the @samp{--out-implib} and @samp{--in-implib} options are | |
7040 | secure gateway import libraries, suitable for linking a non-secure | |
7041 | executable against secure code as per ARMv8-M Security Extensions. | |
7042 | ||
0955507f TP |
7043 | @kindex --in-implib=@var{file} |
7044 | @cindex Input import library | |
7045 | The @samp{--in-implib=file} specifies an input import library whose symbols | |
7046 | must keep the same address in the executable being produced. A warning is | |
7047 | given if no @samp{--out-implib} is given but new symbols have been introduced | |
7048 | in the executable that should be listed in its import library. Otherwise, if | |
7049 | @samp{--out-implib} is specified, the symbols are added to the output import | |
7050 | library. A warning is also given if some symbols present in the input import | |
7051 | library have disappeared from the executable. This option is only effective | |
7052 | for Secure Gateway import libraries, ie. when @samp{--cmse-implib} is | |
7053 | specified. | |
7054 | ||
36f63dca NC |
7055 | @ifclear GENERIC |
7056 | @lowersections | |
7057 | @end ifclear | |
7058 | @end ifset | |
7059 | ||
7060 | @ifset HPPA | |
7061 | @ifclear GENERIC | |
7062 | @raisesections | |
7063 | @end ifclear | |
7064 | ||
7065 | @node HPPA ELF32 | |
7066 | @section @command{ld} and HPPA 32-bit ELF Support | |
7067 | @cindex HPPA multiple sub-space stubs | |
7068 | @kindex --multi-subspace | |
7069 | When generating a shared library, @command{ld} will by default generate | |
7070 | import stubs suitable for use with a single sub-space application. | |
7071 | The @samp{--multi-subspace} switch causes @command{ld} to generate export | |
7072 | stubs, and different (larger) import stubs suitable for use with | |
7073 | multiple sub-spaces. | |
7074 | ||
7075 | @cindex HPPA stub grouping | |
7076 | @kindex --stub-group-size=@var{N} | |
7077 | Long branch stubs and import/export stubs are placed by @command{ld} in | |
7078 | stub sections located between groups of input sections. | |
7079 | @samp{--stub-group-size} specifies the maximum size of a group of input | |
7080 | sections handled by one stub section. Since branch offsets are signed, | |
7081 | a stub section may serve two groups of input sections, one group before | |
7082 | the stub section, and one group after it. However, when using | |
7083 | conditional branches that require stubs, it may be better (for branch | |
7084 | prediction) that stub sections only serve one group of input sections. | |
7085 | A negative value for @samp{N} chooses this scheme, ensuring that | |
7086 | branches to stubs always use a negative offset. Two special values of | |
7087 | @samp{N} are recognized, @samp{1} and @samp{-1}. These both instruct | |
7088 | @command{ld} to automatically size input section groups for the branch types | |
7089 | detected, with the same behaviour regarding stub placement as other | |
7090 | positive or negative values of @samp{N} respectively. | |
7091 | ||
7092 | Note that @samp{--stub-group-size} does not split input sections. A | |
7093 | single input section larger than the group size specified will of course | |
7094 | create a larger group (of one section). If input sections are too | |
7095 | large, it may not be possible for a branch to reach its stub. | |
7096 | ||
7097 | @ifclear GENERIC | |
7098 | @lowersections | |
7099 | @end ifclear | |
7100 | @end ifset | |
7101 | ||
7fb9f789 NC |
7102 | @ifset M68K |
7103 | @ifclear GENERIC | |
7104 | @raisesections | |
7105 | @end ifclear | |
7106 | ||
7107 | @node M68K | |
7108 | @section @command{ld} and the Motorola 68K family | |
7109 | ||
7110 | @cindex Motorola 68K GOT generation | |
7111 | @kindex --got=@var{type} | |
7112 | The @samp{--got=@var{type}} option lets you choose the GOT generation scheme. | |
7113 | The choices are @samp{single}, @samp{negative}, @samp{multigot} and | |
7114 | @samp{target}. When @samp{target} is selected the linker chooses | |
7115 | the default GOT generation scheme for the current target. | |
7116 | @samp{single} tells the linker to generate a single GOT with | |
7117 | entries only at non-negative offsets. | |
7118 | @samp{negative} instructs the linker to generate a single GOT with | |
7119 | entries at both negative and positive offsets. Not all environments | |
7120 | support such GOTs. | |
7121 | @samp{multigot} allows the linker to generate several GOTs in the | |
7122 | output file. All GOT references from a single input object | |
7123 | file access the same GOT, but references from different input object | |
7124 | files might access different GOTs. Not all environments support such GOTs. | |
7125 | ||
7126 | @ifclear GENERIC | |
7127 | @lowersections | |
7128 | @end ifclear | |
7129 | @end ifset | |
7130 | ||
833794fc MR |
7131 | @ifset MIPS |
7132 | @ifclear GENERIC | |
7133 | @raisesections | |
7134 | @end ifclear | |
7135 | ||
7136 | @node MIPS | |
7137 | @section @command{ld} and the MIPS family | |
7138 | ||
7139 | @cindex MIPS microMIPS instruction choice selection | |
7140 | @kindex --insn32 | |
7141 | @kindex --no-insn32 | |
7142 | The @samp{--insn32} and @samp{--no-insn32} options control the choice of | |
7143 | microMIPS instructions used in code generated by the linker, such as that | |
7144 | in the PLT or lazy binding stubs, or in relaxation. If @samp{--insn32} is | |
7145 | used, then the linker only uses 32-bit instruction encodings. By default | |
7146 | or if @samp{--no-insn32} is used, all instruction encodings are used, | |
7147 | including 16-bit ones where possible. | |
7148 | ||
8b10b0b3 MR |
7149 | @cindex MIPS branch relocation check control |
7150 | @kindex --ignore-branch-isa | |
7151 | @kindex --no-ignore-branch-isa | |
7152 | The @samp{--ignore-branch-isa} and @samp{--no-ignore-branch-isa} options | |
7153 | control branch relocation checks for invalid ISA mode transitions. If | |
7154 | @samp{--ignore-branch-isa} is used, then the linker accepts any branch | |
7155 | relocations and any ISA mode transition required is lost in relocation | |
7156 | calculation, except for some cases of @code{BAL} instructions which meet | |
7157 | relaxation conditions and are converted to equivalent @code{JALX} | |
7158 | instructions as the associated relocation is calculated. By default | |
7159 | or if @samp{--no-ignore-branch-isa} is used a check is made causing | |
7160 | the loss of an ISA mode transition to produce an error. | |
7161 | ||
833794fc MR |
7162 | @ifclear GENERIC |
7163 | @lowersections | |
7164 | @end ifclear | |
7165 | @end ifset | |
7166 | ||
36f63dca NC |
7167 | @ifset MMIX |
7168 | @ifclear GENERIC | |
7169 | @raisesections | |
7170 | @end ifclear | |
7171 | ||
7172 | @node MMIX | |
7173 | @section @code{ld} and MMIX | |
7174 | For MMIX, there is a choice of generating @code{ELF} object files or | |
7175 | @code{mmo} object files when linking. The simulator @code{mmix} | |
7176 | understands the @code{mmo} format. The binutils @code{objcopy} utility | |
7177 | can translate between the two formats. | |
7178 | ||
7179 | There is one special section, the @samp{.MMIX.reg_contents} section. | |
7180 | Contents in this section is assumed to correspond to that of global | |
7181 | registers, and symbols referring to it are translated to special symbols, | |
7182 | equal to registers. In a final link, the start address of the | |
7183 | @samp{.MMIX.reg_contents} section corresponds to the first allocated | |
7184 | global register multiplied by 8. Register @code{$255} is not included in | |
7185 | this section; it is always set to the program entry, which is at the | |
7186 | symbol @code{Main} for @code{mmo} files. | |
7187 | ||
7a2de473 HPN |
7188 | Global symbols with the prefix @code{__.MMIX.start.}, for example |
7189 | @code{__.MMIX.start..text} and @code{__.MMIX.start..data} are special. | |
7190 | The default linker script uses these to set the default start address | |
7191 | of a section. | |
36f63dca NC |
7192 | |
7193 | Initial and trailing multiples of zero-valued 32-bit words in a section, | |
7194 | are left out from an mmo file. | |
7195 | ||
7196 | @ifclear GENERIC | |
7197 | @lowersections | |
7198 | @end ifclear | |
7199 | @end ifset | |
7200 | ||
7201 | @ifset MSP430 | |
7202 | @ifclear GENERIC | |
7203 | @raisesections | |
7204 | @end ifclear | |
7205 | ||
7206 | @node MSP430 | |
7207 | @section @code{ld} and MSP430 | |
7208 | For the MSP430 it is possible to select the MPU architecture. The flag @samp{-m [mpu type]} | |
7209 | will select an appropriate linker script for selected MPU type. (To get a list of known MPUs | |
7210 | just pass @samp{-m help} option to the linker). | |
7211 | ||
7212 | @cindex MSP430 extra sections | |
7213 | The linker will recognize some extra sections which are MSP430 specific: | |
7214 | ||
7215 | @table @code | |
7216 | @item @samp{.vectors} | |
7217 | Defines a portion of ROM where interrupt vectors located. | |
7218 | ||
7219 | @item @samp{.bootloader} | |
7220 | Defines the bootloader portion of the ROM (if applicable). Any code | |
7221 | in this section will be uploaded to the MPU. | |
7222 | ||
7223 | @item @samp{.infomem} | |
7224 | Defines an information memory section (if applicable). Any code in | |
7225 | this section will be uploaded to the MPU. | |
7226 | ||
c0065db7 | 7227 | @item @samp{.infomemnobits} |
36f63dca NC |
7228 | This is the same as the @samp{.infomem} section except that any code |
7229 | in this section will not be uploaded to the MPU. | |
7230 | ||
7231 | @item @samp{.noinit} | |
7232 | Denotes a portion of RAM located above @samp{.bss} section. | |
7233 | ||
c0065db7 | 7234 | The last two sections are used by gcc. |
36f63dca NC |
7235 | @end table |
7236 | ||
7ef3addb JL |
7237 | @table @option |
7238 | @cindex MSP430 Options | |
7239 | @kindex --code-region | |
7240 | @item --code-region=[either,lower,upper,none] | |
7241 | This will transform .text* sections to [either,lower,upper].text* sections. The | |
7242 | argument passed to GCC for -mcode-region is propagated to the linker | |
7243 | using this option. | |
7244 | ||
7245 | @kindex --data-region | |
7246 | @item --data-region=[either,lower,upper,none] | |
7247 | This will transform .data*, .bss* and .rodata* sections to | |
7248 | [either,lower,upper].[data,bss,rodata]* sections. The argument passed to GCC | |
7249 | for -mdata-region is propagated to the linker using this option. | |
7250 | ||
7251 | @kindex --disable-sec-transformation | |
7252 | @item --disable-sec-transformation | |
7253 | Prevent the transformation of sections as specified by the @code{--code-region} | |
7254 | and @code{--data-region} options. | |
7255 | This is useful if you are compiling and linking using a single call to the GCC | |
7256 | wrapper, and want to compile the source files using -m[code,data]-region but | |
7257 | not transform the sections for prebuilt libraries and objects. | |
7258 | @end table | |
7259 | ||
36f63dca NC |
7260 | @ifclear GENERIC |
7261 | @lowersections | |
7262 | @end ifclear | |
7263 | @end ifset | |
7264 | ||
35c08157 KLC |
7265 | @ifset NDS32 |
7266 | @ifclear GENERIC | |
7267 | @raisesections | |
7268 | @end ifclear | |
7269 | ||
7270 | @node NDS32 | |
7271 | @section @code{ld} and NDS32 | |
7272 | @kindex relaxing on NDS32 | |
7273 | For NDS32, there are some options to select relaxation behavior. The linker | |
7274 | relaxes objects according to these options. | |
7275 | ||
7276 | @table @code | |
7277 | @item @samp{--m[no-]fp-as-gp} | |
7278 | Disable/enable fp-as-gp relaxation. | |
7279 | ||
7280 | @item @samp{--mexport-symbols=FILE} | |
7281 | Exporting symbols and their address into FILE as linker script. | |
7282 | ||
7283 | @item @samp{--m[no-]ex9} | |
7284 | Disable/enable link-time EX9 relaxation. | |
7285 | ||
7286 | @item @samp{--mexport-ex9=FILE} | |
7287 | Export the EX9 table after linking. | |
7288 | ||
7289 | @item @samp{--mimport-ex9=FILE} | |
7290 | Import the Ex9 table for EX9 relaxation. | |
7291 | ||
7292 | @item @samp{--mupdate-ex9} | |
7293 | Update the existing EX9 table. | |
7294 | ||
7295 | @item @samp{--mex9-limit=NUM} | |
7296 | Maximum number of entries in the ex9 table. | |
7297 | ||
7298 | @item @samp{--mex9-loop-aware} | |
7299 | Avoid generating the EX9 instruction inside the loop. | |
7300 | ||
7301 | @item @samp{--m[no-]ifc} | |
7302 | Disable/enable the link-time IFC optimization. | |
7303 | ||
7304 | @item @samp{--mifc-loop-aware} | |
7305 | Avoid generating the IFC instruction inside the loop. | |
7306 | @end table | |
7307 | ||
7308 | @ifclear GENERIC | |
7309 | @lowersections | |
7310 | @end ifclear | |
7311 | @end ifset | |
7312 | ||
78058a5e SL |
7313 | @ifset NIOSII |
7314 | @ifclear GENERIC | |
7315 | @raisesections | |
7316 | @end ifclear | |
7317 | ||
7318 | @node Nios II | |
7319 | @section @command{ld} and the Altera Nios II | |
7320 | @cindex Nios II call relaxation | |
7321 | @kindex --relax on Nios II | |
7322 | ||
7323 | Call and immediate jump instructions on Nios II processors are limited to | |
7324 | transferring control to addresses in the same 256MB memory segment, | |
7325 | which may result in @command{ld} giving | |
7326 | @samp{relocation truncated to fit} errors with very large programs. | |
7327 | The command-line option @option{--relax} enables the generation of | |
7328 | trampolines that can access the entire 32-bit address space for calls | |
7329 | outside the normal @code{call} and @code{jmpi} address range. These | |
7330 | trampolines are inserted at section boundaries, so may not themselves | |
7331 | be reachable if an input section and its associated call trampolines are | |
7332 | larger than 256MB. | |
7333 | ||
7334 | The @option{--relax} option is enabled by default unless @option{-r} | |
7335 | is also specified. You can disable trampoline generation by using the | |
7336 | @option{--no-relax} linker option. You can also disable this optimization | |
7337 | locally by using the @samp{set .noat} directive in assembly-language | |
7338 | source files, as the linker-inserted trampolines use the @code{at} | |
7339 | register as a temporary. | |
7340 | ||
7341 | Note that the linker @option{--relax} option is independent of assembler | |
7342 | relaxation options, and that using the GNU assembler's @option{-relax-all} | |
7343 | option interferes with the linker's more selective call instruction relaxation. | |
7344 | ||
7345 | @ifclear GENERIC | |
7346 | @lowersections | |
7347 | @end ifclear | |
7348 | @end ifset | |
7349 | ||
2a60a7a8 AM |
7350 | @ifset POWERPC |
7351 | @ifclear GENERIC | |
7352 | @raisesections | |
7353 | @end ifclear | |
7354 | ||
7355 | @node PowerPC ELF32 | |
7356 | @section @command{ld} and PowerPC 32-bit ELF Support | |
7357 | @cindex PowerPC long branches | |
7358 | @kindex --relax on PowerPC | |
7359 | Branches on PowerPC processors are limited to a signed 26-bit | |
7360 | displacement, which may result in @command{ld} giving | |
7361 | @samp{relocation truncated to fit} errors with very large programs. | |
7362 | @samp{--relax} enables the generation of trampolines that can access | |
7363 | the entire 32-bit address space. These trampolines are inserted at | |
7364 | section boundaries, so may not themselves be reachable if an input | |
c8a1f254 NS |
7365 | section exceeds 33M in size. You may combine @samp{-r} and |
7366 | @samp{--relax} to add trampolines in a partial link. In that case | |
7367 | both branches to undefined symbols and inter-section branches are also | |
7368 | considered potentially out of range, and trampolines inserted. | |
2a60a7a8 AM |
7369 | |
7370 | @cindex PowerPC ELF32 options | |
7371 | @table @option | |
7372 | @cindex PowerPC PLT | |
7373 | @kindex --bss-plt | |
7374 | @item --bss-plt | |
7375 | Current PowerPC GCC accepts a @samp{-msecure-plt} option that | |
7376 | generates code capable of using a newer PLT and GOT layout that has | |
7377 | the security advantage of no executable section ever needing to be | |
7378 | writable and no writable section ever being executable. PowerPC | |
7379 | @command{ld} will generate this layout, including stubs to access the | |
7380 | PLT, if all input files (including startup and static libraries) were | |
7381 | compiled with @samp{-msecure-plt}. @samp{--bss-plt} forces the old | |
7382 | BSS PLT (and GOT layout) which can give slightly better performance. | |
7383 | ||
016687f8 AM |
7384 | @kindex --secure-plt |
7385 | @item --secure-plt | |
7386 | @command{ld} will use the new PLT and GOT layout if it is linking new | |
7387 | @samp{-fpic} or @samp{-fPIC} code, but does not do so automatically | |
7388 | when linking non-PIC code. This option requests the new PLT and GOT | |
7389 | layout. A warning will be given if some object file requires the old | |
7390 | style BSS PLT. | |
7391 | ||
2a60a7a8 AM |
7392 | @cindex PowerPC GOT |
7393 | @kindex --sdata-got | |
7394 | @item --sdata-got | |
7395 | The new secure PLT and GOT are placed differently relative to other | |
7396 | sections compared to older BSS PLT and GOT placement. The location of | |
7397 | @code{.plt} must change because the new secure PLT is an initialized | |
7398 | section while the old PLT is uninitialized. The reason for the | |
7399 | @code{.got} change is more subtle: The new placement allows | |
7400 | @code{.got} to be read-only in applications linked with | |
7401 | @samp{-z relro -z now}. However, this placement means that | |
7402 | @code{.sdata} cannot always be used in shared libraries, because the | |
7403 | PowerPC ABI accesses @code{.sdata} in shared libraries from the GOT | |
7404 | pointer. @samp{--sdata-got} forces the old GOT placement. PowerPC | |
7405 | GCC doesn't use @code{.sdata} in shared libraries, so this option is | |
7406 | really only useful for other compilers that may do so. | |
7407 | ||
7408 | @cindex PowerPC stub symbols | |
7409 | @kindex --emit-stub-syms | |
7410 | @item --emit-stub-syms | |
7411 | This option causes @command{ld} to label linker stubs with a local | |
7412 | symbol that encodes the stub type and destination. | |
7413 | ||
7414 | @cindex PowerPC TLS optimization | |
7415 | @kindex --no-tls-optimize | |
7416 | @item --no-tls-optimize | |
7417 | PowerPC @command{ld} normally performs some optimization of code | |
7418 | sequences used to access Thread-Local Storage. Use this option to | |
7419 | disable the optimization. | |
7420 | @end table | |
7421 | ||
7422 | @ifclear GENERIC | |
7423 | @lowersections | |
7424 | @end ifclear | |
7425 | @end ifset | |
7426 | ||
7427 | @ifset POWERPC64 | |
7428 | @ifclear GENERIC | |
7429 | @raisesections | |
7430 | @end ifclear | |
7431 | ||
7432 | @node PowerPC64 ELF64 | |
7433 | @section @command{ld} and PowerPC64 64-bit ELF Support | |
7434 | ||
7435 | @cindex PowerPC64 ELF64 options | |
7436 | @table @option | |
7437 | @cindex PowerPC64 stub grouping | |
7438 | @kindex --stub-group-size | |
7439 | @item --stub-group-size | |
7440 | Long branch stubs, PLT call stubs and TOC adjusting stubs are placed | |
7441 | by @command{ld} in stub sections located between groups of input sections. | |
7442 | @samp{--stub-group-size} specifies the maximum size of a group of input | |
7443 | sections handled by one stub section. Since branch offsets are signed, | |
7444 | a stub section may serve two groups of input sections, one group before | |
7445 | the stub section, and one group after it. However, when using | |
7446 | conditional branches that require stubs, it may be better (for branch | |
7447 | prediction) that stub sections only serve one group of input sections. | |
7448 | A negative value for @samp{N} chooses this scheme, ensuring that | |
7449 | branches to stubs always use a negative offset. Two special values of | |
7450 | @samp{N} are recognized, @samp{1} and @samp{-1}. These both instruct | |
7451 | @command{ld} to automatically size input section groups for the branch types | |
7452 | detected, with the same behaviour regarding stub placement as other | |
7453 | positive or negative values of @samp{N} respectively. | |
7454 | ||
7455 | Note that @samp{--stub-group-size} does not split input sections. A | |
7456 | single input section larger than the group size specified will of course | |
7457 | create a larger group (of one section). If input sections are too | |
7458 | large, it may not be possible for a branch to reach its stub. | |
7459 | ||
7460 | @cindex PowerPC64 stub symbols | |
7461 | @kindex --emit-stub-syms | |
7462 | @item --emit-stub-syms | |
7463 | This option causes @command{ld} to label linker stubs with a local | |
7464 | symbol that encodes the stub type and destination. | |
7465 | ||
7466 | @cindex PowerPC64 dot symbols | |
7467 | @kindex --dotsyms | |
7468 | @kindex --no-dotsyms | |
95421fb9 AM |
7469 | @item --dotsyms |
7470 | @itemx --no-dotsyms | |
2a60a7a8 AM |
7471 | These two options control how @command{ld} interprets version patterns |
7472 | in a version script. Older PowerPC64 compilers emitted both a | |
7473 | function descriptor symbol with the same name as the function, and a | |
7474 | code entry symbol with the name prefixed by a dot (@samp{.}). To | |
7475 | properly version a function @samp{foo}, the version script thus needs | |
7476 | to control both @samp{foo} and @samp{.foo}. The option | |
7477 | @samp{--dotsyms}, on by default, automatically adds the required | |
7478 | dot-prefixed patterns. Use @samp{--no-dotsyms} to disable this | |
7479 | feature. | |
7480 | ||
7ae4ea7d AM |
7481 | @cindex PowerPC64 register save/restore functions |
7482 | @kindex --save-restore-funcs | |
7483 | @kindex --no-save-restore-funcs | |
95421fb9 AM |
7484 | @item --save-restore-funcs |
7485 | @itemx --no-save-restore-funcs | |
7ae4ea7d AM |
7486 | These two options control whether PowerPC64 @command{ld} automatically |
7487 | provides out-of-line register save and restore functions used by | |
7488 | @samp{-Os} code. The default is to provide any such referenced | |
7489 | function for a normal final link, and to not do so for a relocatable | |
7490 | link. | |
7491 | ||
2a60a7a8 AM |
7492 | @cindex PowerPC64 TLS optimization |
7493 | @kindex --no-tls-optimize | |
7494 | @item --no-tls-optimize | |
7495 | PowerPC64 @command{ld} normally performs some optimization of code | |
7496 | sequences used to access Thread-Local Storage. Use this option to | |
7497 | disable the optimization. | |
7498 | ||
7c9cf415 AM |
7499 | @cindex PowerPC64 __tls_get_addr optimization |
7500 | @kindex --tls-get-addr-optimize | |
7501 | @kindex --no-tls-get-addr-optimize | |
95421fb9 AM |
7502 | @item --tls-get-addr-optimize |
7503 | @itemx --no-tls-get-addr-optimize | |
7c9cf415 AM |
7504 | These options control whether PowerPC64 @command{ld} uses a special |
7505 | stub to call __tls_get_addr. PowerPC64 glibc 2.22 and later support | |
7506 | an optimization that allows the second and subsequent calls to | |
7507 | @code{__tls_get_addr} for a given symbol to be resolved by the special | |
7508 | stub without calling in to glibc. By default the linker enables this | |
7509 | option when glibc advertises the availability of __tls_get_addr_opt. | |
7510 | Forcing this option on when using an older glibc won't do much besides | |
7511 | slow down your applications, but may be useful if linking an | |
7512 | application against an older glibc with the expectation that it will | |
7513 | normally be used on systems having a newer glibc. | |
7514 | ||
2a60a7a8 AM |
7515 | @cindex PowerPC64 OPD optimization |
7516 | @kindex --no-opd-optimize | |
7517 | @item --no-opd-optimize | |
7518 | PowerPC64 @command{ld} normally removes @code{.opd} section entries | |
7519 | corresponding to deleted link-once functions, or functions removed by | |
e7fc76dd | 7520 | the action of @samp{--gc-sections} or linker script @code{/DISCARD/}. |
2a60a7a8 AM |
7521 | Use this option to disable @code{.opd} optimization. |
7522 | ||
7523 | @cindex PowerPC64 OPD spacing | |
7524 | @kindex --non-overlapping-opd | |
7525 | @item --non-overlapping-opd | |
7526 | Some PowerPC64 compilers have an option to generate compressed | |
7527 | @code{.opd} entries spaced 16 bytes apart, overlapping the third word, | |
7528 | the static chain pointer (unused in C) with the first word of the next | |
7529 | entry. This option expands such entries to the full 24 bytes. | |
7530 | ||
7531 | @cindex PowerPC64 TOC optimization | |
7532 | @kindex --no-toc-optimize | |
7533 | @item --no-toc-optimize | |
7534 | PowerPC64 @command{ld} normally removes unused @code{.toc} section | |
7535 | entries. Such entries are detected by examining relocations that | |
7536 | reference the TOC in code sections. A reloc in a deleted code section | |
7537 | marks a TOC word as unneeded, while a reloc in a kept code section | |
7538 | marks a TOC word as needed. Since the TOC may reference itself, TOC | |
7539 | relocs are also examined. TOC words marked as both needed and | |
7540 | unneeded will of course be kept. TOC words without any referencing | |
7541 | reloc are assumed to be part of a multi-word entry, and are kept or | |
7542 | discarded as per the nearest marked preceding word. This works | |
7543 | reliably for compiler generated code, but may be incorrect if assembly | |
7544 | code is used to insert TOC entries. Use this option to disable the | |
7545 | optimization. | |
7546 | ||
7547 | @cindex PowerPC64 multi-TOC | |
7548 | @kindex --no-multi-toc | |
7549 | @item --no-multi-toc | |
794e51c0 AM |
7550 | If given any toc option besides @code{-mcmodel=medium} or |
7551 | @code{-mcmodel=large}, PowerPC64 GCC generates code for a TOC model | |
7552 | where TOC | |
2a60a7a8 AM |
7553 | entries are accessed with a 16-bit offset from r2. This limits the |
7554 | total TOC size to 64K. PowerPC64 @command{ld} extends this limit by | |
7555 | grouping code sections such that each group uses less than 64K for its | |
7556 | TOC entries, then inserts r2 adjusting stubs between inter-group | |
7557 | calls. @command{ld} does not split apart input sections, so cannot | |
7558 | help if a single input file has a @code{.toc} section that exceeds | |
7559 | 64K, most likely from linking multiple files with @command{ld -r}. | |
7560 | Use this option to turn off this feature. | |
794e51c0 AM |
7561 | |
7562 | @cindex PowerPC64 TOC sorting | |
7563 | @kindex --no-toc-sort | |
7564 | @item --no-toc-sort | |
7565 | By default, @command{ld} sorts TOC sections so that those whose file | |
7566 | happens to have a section called @code{.init} or @code{.fini} are | |
7567 | placed first, followed by TOC sections referenced by code generated | |
7568 | with PowerPC64 gcc's @code{-mcmodel=small}, and lastly TOC sections | |
7569 | referenced only by code generated with PowerPC64 gcc's | |
7570 | @code{-mcmodel=medium} or @code{-mcmodel=large} options. Doing this | |
7571 | results in better TOC grouping for multi-TOC. Use this option to turn | |
7572 | off this feature. | |
7573 | ||
7574 | @cindex PowerPC64 PLT stub alignment | |
7575 | @kindex --plt-align | |
7576 | @kindex --no-plt-align | |
7577 | @item --plt-align | |
7578 | @itemx --no-plt-align | |
7579 | Use these options to control whether individual PLT call stubs are | |
2420fff6 AM |
7580 | aligned to a 32-byte boundary, or to the specified power of two |
7581 | boundary when using @code{--plt-align=}. A negative value may be | |
7582 | specified to pad PLT call stubs so that they do not cross the | |
7583 | specified power of two boundary (or the minimum number of boundaries | |
7584 | if a PLT stub is so large that it must cross a boundary). By default | |
7585 | PLT call stubs are aligned to 32-byte boundaries. | |
794e51c0 AM |
7586 | |
7587 | @cindex PowerPC64 PLT call stub static chain | |
7588 | @kindex --plt-static-chain | |
7589 | @kindex --no-plt-static-chain | |
7590 | @item --plt-static-chain | |
7591 | @itemx --no-plt-static-chain | |
7592 | Use these options to control whether PLT call stubs load the static | |
7593 | chain pointer (r11). @code{ld} defaults to not loading the static | |
7594 | chain since there is never any need to do so on a PLT call. | |
7595 | ||
7596 | @cindex PowerPC64 PLT call stub thread safety | |
7597 | @kindex --plt-thread-safe | |
7598 | @kindex --no-plt-thread-safe | |
7599 | @item --plt-thread-safe | |
1be5d8d3 | 7600 | @itemx --no-plt-thread-safe |
794e51c0 AM |
7601 | With power7's weakly ordered memory model, it is possible when using |
7602 | lazy binding for ld.so to update a plt entry in one thread and have | |
7603 | another thread see the individual plt entry words update in the wrong | |
7604 | order, despite ld.so carefully writing in the correct order and using | |
7605 | memory write barriers. To avoid this we need some sort of read | |
7606 | barrier in the call stub, or use LD_BIND_NOW=1. By default, @code{ld} | |
7607 | looks for calls to commonly used functions that create threads, and if | |
7608 | seen, adds the necessary barriers. Use these options to change the | |
7609 | default behaviour. | |
8b5f1ed8 AM |
7610 | |
7611 | @cindex PowerPC64 ELFv2 PLT localentry optimization | |
7612 | @kindex --plt-localentry | |
7613 | @kindex --no-plt-localentry | |
7614 | @item --plt-localentry | |
7615 | @itemx --no-localentry | |
7616 | ELFv2 functions with localentry:0 are those with a single entry point, | |
7617 | ie. global entry == local entry, and that have no requirement on r2 | |
7618 | (the TOC/GOT pointer) or r12, and guarantee r2 is unchanged on return. | |
7619 | Such an external function can be called via the PLT without saving r2 | |
7620 | or restoring it on return, avoiding a common load-hit-store for small | |
7621 | functions. The optimization is attractive, with up to 40% reduction | |
7622 | in execution time for a small function, but can result in symbol | |
d44c746a AM |
7623 | interposition failures. Also, minor changes in a shared library, |
7624 | including system libraries, can cause a function that was localentry:0 | |
7625 | to become localentry:8. This will result in a dynamic loader | |
7626 | complaint and failure to run. The option is experimental, use with | |
7627 | care. @option{--no-plt-localentry} is the default. | |
2a60a7a8 AM |
7628 | @end table |
7629 | ||
7630 | @ifclear GENERIC | |
7631 | @lowersections | |
7632 | @end ifclear | |
7633 | @end ifset | |
7634 | ||
b4cbbe8f AK |
7635 | @ifset S/390 |
7636 | @ifclear GENERIC | |
7637 | @raisesections | |
7638 | @end ifclear | |
7639 | ||
7640 | @node S/390 ELF | |
7641 | @section @command{ld} and S/390 ELF Support | |
7642 | ||
7643 | @cindex S/390 ELF options | |
7644 | @table @option | |
7645 | ||
7646 | @cindex S/390 | |
7647 | @kindex --s390-pgste | |
7648 | @item --s390-pgste | |
7649 | This option marks the result file with a @code{PT_S390_PGSTE} | |
7650 | segment. The Linux kernel is supposed to allocate 4k page tables for | |
7651 | binaries marked that way. | |
7652 | @end table | |
7653 | ||
7654 | @ifclear GENERIC | |
7655 | @lowersections | |
7656 | @end ifclear | |
7657 | @end ifset | |
7658 | ||
49fa1e15 AM |
7659 | @ifset SPU |
7660 | @ifclear GENERIC | |
7661 | @raisesections | |
7662 | @end ifclear | |
7663 | ||
7664 | @node SPU ELF | |
7665 | @section @command{ld} and SPU ELF Support | |
7666 | ||
7667 | @cindex SPU ELF options | |
7668 | @table @option | |
7669 | ||
7670 | @cindex SPU plugins | |
7671 | @kindex --plugin | |
7672 | @item --plugin | |
7673 | This option marks an executable as a PIC plugin module. | |
7674 | ||
7675 | @cindex SPU overlays | |
7676 | @kindex --no-overlays | |
7677 | @item --no-overlays | |
7678 | Normally, @command{ld} recognizes calls to functions within overlay | |
7679 | regions, and redirects such calls to an overlay manager via a stub. | |
7680 | @command{ld} also provides a built-in overlay manager. This option | |
7681 | turns off all this special overlay handling. | |
7682 | ||
7683 | @cindex SPU overlay stub symbols | |
7684 | @kindex --emit-stub-syms | |
7685 | @item --emit-stub-syms | |
7686 | This option causes @command{ld} to label overlay stubs with a local | |
7687 | symbol that encodes the stub type and destination. | |
7688 | ||
7689 | @cindex SPU extra overlay stubs | |
7690 | @kindex --extra-overlay-stubs | |
7691 | @item --extra-overlay-stubs | |
7692 | This option causes @command{ld} to add overlay call stubs on all | |
7693 | function calls out of overlay regions. Normally stubs are not added | |
7694 | on calls to non-overlay regions. | |
7695 | ||
7696 | @cindex SPU local store size | |
7697 | @kindex --local-store=lo:hi | |
7698 | @item --local-store=lo:hi | |
7699 | @command{ld} usually checks that a final executable for SPU fits in | |
7700 | the address range 0 to 256k. This option may be used to change the | |
7701 | range. Disable the check entirely with @option{--local-store=0:0}. | |
7702 | ||
c0065db7 | 7703 | @cindex SPU |
49fa1e15 AM |
7704 | @kindex --stack-analysis |
7705 | @item --stack-analysis | |
7706 | SPU local store space is limited. Over-allocation of stack space | |
7707 | unnecessarily limits space available for code and data, while | |
7708 | under-allocation results in runtime failures. If given this option, | |
7709 | @command{ld} will provide an estimate of maximum stack usage. | |
7710 | @command{ld} does this by examining symbols in code sections to | |
7711 | determine the extents of functions, and looking at function prologues | |
7712 | for stack adjusting instructions. A call-graph is created by looking | |
7713 | for relocations on branch instructions. The graph is then searched | |
7714 | for the maximum stack usage path. Note that this analysis does not | |
7715 | find calls made via function pointers, and does not handle recursion | |
7716 | and other cycles in the call graph. Stack usage may be | |
7717 | under-estimated if your code makes such calls. Also, stack usage for | |
7718 | dynamic allocation, e.g. alloca, will not be detected. If a link map | |
7719 | is requested, detailed information about each function's stack usage | |
7720 | and calls will be given. | |
7721 | ||
c0065db7 | 7722 | @cindex SPU |
49fa1e15 AM |
7723 | @kindex --emit-stack-syms |
7724 | @item --emit-stack-syms | |
7725 | This option, if given along with @option{--stack-analysis} will result | |
7726 | in @command{ld} emitting stack sizing symbols for each function. | |
7727 | These take the form @code{__stack_<function_name>} for global | |
7728 | functions, and @code{__stack_<number>_<function_name>} for static | |
7729 | functions. @code{<number>} is the section id in hex. The value of | |
7730 | such symbols is the stack requirement for the corresponding function. | |
7731 | The symbol size will be zero, type @code{STT_NOTYPE}, binding | |
c0065db7 | 7732 | @code{STB_LOCAL}, and section @code{SHN_ABS}. |
49fa1e15 AM |
7733 | @end table |
7734 | ||
7735 | @ifclear GENERIC | |
7736 | @lowersections | |
7737 | @end ifclear | |
7738 | @end ifset | |
7739 | ||
36f63dca NC |
7740 | @ifset TICOFF |
7741 | @ifclear GENERIC | |
7742 | @raisesections | |
7743 | @end ifclear | |
7744 | ||
7745 | @node TI COFF | |
7746 | @section @command{ld}'s Support for Various TI COFF Versions | |
7747 | @cindex TI COFF versions | |
7748 | @kindex --format=@var{version} | |
7749 | The @samp{--format} switch allows selection of one of the various | |
7750 | TI COFF versions. The latest of this writing is 2; versions 0 and 1 are | |
7751 | also supported. The TI COFF versions also vary in header byte-order | |
7752 | format; @command{ld} will read any version or byte order, but the output | |
7753 | header format depends on the default specified by the specific target. | |
7754 | ||
7755 | @ifclear GENERIC | |
7756 | @lowersections | |
7757 | @end ifclear | |
7758 | @end ifset | |
7759 | ||
2ca22b03 NC |
7760 | @ifset WIN32 |
7761 | @ifclear GENERIC | |
7762 | @raisesections | |
7763 | @end ifclear | |
7764 | ||
7765 | @node WIN32 | |
7766 | @section @command{ld} and WIN32 (cygwin/mingw) | |
7767 | ||
c0065db7 | 7768 | This section describes some of the win32 specific @command{ld} issues. |
a05a5b64 TP |
7769 | See @ref{Options,,Command-line Options} for detailed description of the |
7770 | command-line options mentioned here. | |
2ca22b03 NC |
7771 | |
7772 | @table @emph | |
c0065db7 RM |
7773 | @cindex import libraries |
7774 | @item import libraries | |
69da35b5 | 7775 | The standard Windows linker creates and uses so-called import |
2ca22b03 | 7776 | libraries, which contains information for linking to dll's. They are |
69da35b5 NC |
7777 | regular static archives and are handled as any other static |
7778 | archive. The cygwin and mingw ports of @command{ld} have specific | |
2ca22b03 | 7779 | support for creating such libraries provided with the |
a05a5b64 | 7780 | @samp{--out-implib} command-line option. |
2ca22b03 | 7781 | |
c0065db7 RM |
7782 | @item exporting DLL symbols |
7783 | @cindex exporting DLL symbols | |
dc8465bf NC |
7784 | The cygwin/mingw @command{ld} has several ways to export symbols for dll's. |
7785 | ||
7786 | @table @emph | |
7787 | @item using auto-export functionality | |
7788 | @cindex using auto-export functionality | |
7789 | By default @command{ld} exports symbols with the auto-export functionality, | |
a05a5b64 | 7790 | which is controlled by the following command-line options: |
dc8465bf | 7791 | |
0a5d968e NC |
7792 | @itemize |
7793 | @item --export-all-symbols [This is the default] | |
7794 | @item --exclude-symbols | |
7795 | @item --exclude-libs | |
e1c37eb5 | 7796 | @item --exclude-modules-for-implib |
09e2aba4 | 7797 | @item --version-script |
0a5d968e NC |
7798 | @end itemize |
7799 | ||
09e2aba4 DK |
7800 | When auto-export is in operation, @command{ld} will export all the non-local |
7801 | (global and common) symbols it finds in a DLL, with the exception of a few | |
7802 | symbols known to belong to the system's runtime and libraries. As it will | |
7803 | often not be desirable to export all of a DLL's symbols, which may include | |
7804 | private functions that are not part of any public interface, the command-line | |
9d5777a3 | 7805 | options listed above may be used to filter symbols out from the list for |
09e2aba4 DK |
7806 | exporting. The @samp{--output-def} option can be used in order to see the |
7807 | final list of exported symbols with all exclusions taken into effect. | |
7808 | ||
7809 | If @samp{--export-all-symbols} is not given explicitly on the | |
0a5d968e NC |
7810 | command line, then the default auto-export behavior will be @emph{disabled} |
7811 | if either of the following are true: | |
7812 | ||
7813 | @itemize | |
7814 | @item A DEF file is used. | |
7815 | @item Any symbol in any object file was marked with the __declspec(dllexport) attribute. | |
7816 | @end itemize | |
dc8465bf | 7817 | |
c0065db7 RM |
7818 | @item using a DEF file |
7819 | @cindex using a DEF file | |
dc8465bf NC |
7820 | Another way of exporting symbols is using a DEF file. A DEF file is |
7821 | an ASCII file containing definitions of symbols which should be | |
7822 | exported when a dll is created. Usually it is named @samp{<dll | |
7823 | name>.def} and is added as any other object file to the linker's | |
0a5d968e | 7824 | command line. The file's name must end in @samp{.def} or @samp{.DEF}. |
dc8465bf NC |
7825 | |
7826 | @example | |
7827 | gcc -o <output> <objectfiles> <dll name>.def | |
7828 | @end example | |
7829 | ||
0a5d968e NC |
7830 | Using a DEF file turns off the normal auto-export behavior, unless the |
7831 | @samp{--export-all-symbols} option is also used. | |
7832 | ||
dc8465bf NC |
7833 | Here is an example of a DEF file for a shared library called @samp{xyz.dll}: |
7834 | ||
7835 | @example | |
4b5bd4e7 | 7836 | LIBRARY "xyz.dll" BASE=0x20000000 |
dc8465bf NC |
7837 | |
7838 | EXPORTS | |
7839 | foo | |
7840 | bar | |
7841 | _bar = bar | |
4b5bd4e7 DS |
7842 | another_foo = abc.dll.afoo |
7843 | var1 DATA | |
7fcab871 KT |
7844 | doo = foo == foo2 |
7845 | eoo DATA == var1 | |
c0065db7 | 7846 | @end example |
dc8465bf | 7847 | |
7fcab871 | 7848 | This example defines a DLL with a non-default base address and seven |
4b5bd4e7 DS |
7849 | symbols in the export table. The third exported symbol @code{_bar} is an |
7850 | alias for the second. The fourth symbol, @code{another_foo} is resolved | |
7851 | by "forwarding" to another module and treating it as an alias for | |
7852 | @code{afoo} exported from the DLL @samp{abc.dll}. The final symbol | |
7fcab871 KT |
7853 | @code{var1} is declared to be a data object. The @samp{doo} symbol in |
7854 | export library is an alias of @samp{foo}, which gets the string name | |
7855 | in export table @samp{foo2}. The @samp{eoo} symbol is an data export | |
7856 | symbol, which gets in export table the name @samp{var1}. | |
4b5bd4e7 | 7857 | |
6b31ad16 DS |
7858 | The optional @code{LIBRARY <name>} command indicates the @emph{internal} |
7859 | name of the output DLL. If @samp{<name>} does not include a suffix, | |
7860 | the default library suffix, @samp{.DLL} is appended. | |
7861 | ||
b45619c0 NC |
7862 | When the .DEF file is used to build an application, rather than a |
7863 | library, the @code{NAME <name>} command should be used instead of | |
6b31ad16 | 7864 | @code{LIBRARY}. If @samp{<name>} does not include a suffix, the default |
c0065db7 | 7865 | executable suffix, @samp{.EXE} is appended. |
6b31ad16 DS |
7866 | |
7867 | With either @code{LIBRARY <name>} or @code{NAME <name>} the optional | |
7868 | specification @code{BASE = <number>} may be used to specify a | |
c0065db7 | 7869 | non-default base address for the image. |
6b31ad16 DS |
7870 | |
7871 | If neither @code{LIBRARY <name>} nor @code{NAME <name>} is specified, | |
a2877985 DS |
7872 | or they specify an empty string, the internal name is the same as the |
7873 | filename specified on the command line. | |
6b31ad16 | 7874 | |
4b5bd4e7 DS |
7875 | The complete specification of an export symbol is: |
7876 | ||
7877 | @example | |
7878 | EXPORTS | |
7879 | ( ( ( <name1> [ = <name2> ] ) | |
7880 | | ( <name1> = <module-name> . <external-name>)) | |
7fcab871 | 7881 | [ @@ <integer> ] [NONAME] [DATA] [CONSTANT] [PRIVATE] [== <name3>] ) * |
c0065db7 | 7882 | @end example |
4b5bd4e7 DS |
7883 | |
7884 | Declares @samp{<name1>} as an exported symbol from the DLL, or declares | |
7885 | @samp{<name1>} as an exported alias for @samp{<name2>}; or declares | |
7886 | @samp{<name1>} as a "forward" alias for the symbol | |
7887 | @samp{<external-name>} in the DLL @samp{<module-name>}. | |
7888 | Optionally, the symbol may be exported by the specified ordinal | |
7fcab871 KT |
7889 | @samp{<integer>} alias. The optional @samp{<name3>} is the to be used |
7890 | string in import/export table for the symbol. | |
4b5bd4e7 DS |
7891 | |
7892 | The optional keywords that follow the declaration indicate: | |
7893 | ||
7894 | @code{NONAME}: Do not put the symbol name in the DLL's export table. It | |
7895 | will still be exported by its ordinal alias (either the value specified | |
7896 | by the .def specification or, otherwise, the value assigned by the | |
7897 | linker). The symbol name, however, does remain visible in the import | |
7898 | library (if any), unless @code{PRIVATE} is also specified. | |
7899 | ||
7900 | @code{DATA}: The symbol is a variable or object, rather than a function. | |
7901 | The import lib will export only an indirect reference to @code{foo} as | |
7902 | the symbol @code{_imp__foo} (ie, @code{foo} must be resolved as | |
7903 | @code{*_imp__foo}). | |
7904 | ||
7905 | @code{CONSTANT}: Like @code{DATA}, but put the undecorated @code{foo} as | |
7906 | well as @code{_imp__foo} into the import library. Both refer to the | |
7907 | read-only import address table's pointer to the variable, not to the | |
7908 | variable itself. This can be dangerous. If the user code fails to add | |
7909 | the @code{dllimport} attribute and also fails to explicitly add the | |
7910 | extra indirection that the use of the attribute enforces, the | |
7911 | application will behave unexpectedly. | |
7912 | ||
7913 | @code{PRIVATE}: Put the symbol in the DLL's export table, but do not put | |
7914 | it into the static import library used to resolve imports at link time. The | |
7915 | symbol can still be imported using the @code{LoadLibrary/GetProcAddress} | |
de194d85 | 7916 | API at runtime or by using the GNU ld extension of linking directly to |
4b5bd4e7 | 7917 | the DLL without an import library. |
c0065db7 | 7918 | |
4b5bd4e7 DS |
7919 | See ld/deffilep.y in the binutils sources for the full specification of |
7920 | other DEF file statements | |
dc8465bf NC |
7921 | |
7922 | @cindex creating a DEF file | |
7923 | While linking a shared dll, @command{ld} is able to create a DEF file | |
a05a5b64 | 7924 | with the @samp{--output-def <file>} command-line option. |
0a5d968e NC |
7925 | |
7926 | @item Using decorations | |
7927 | @cindex Using decorations | |
7928 | Another way of marking symbols for export is to modify the source code | |
7929 | itself, so that when building the DLL each symbol to be exported is | |
7930 | declared as: | |
7931 | ||
7932 | @example | |
7933 | __declspec(dllexport) int a_variable | |
7934 | __declspec(dllexport) void a_function(int with_args) | |
7935 | @end example | |
7936 | ||
7937 | All such symbols will be exported from the DLL. If, however, | |
7938 | any of the object files in the DLL contain symbols decorated in | |
7939 | this way, then the normal auto-export behavior is disabled, unless | |
7940 | the @samp{--export-all-symbols} option is also used. | |
7941 | ||
7942 | Note that object files that wish to access these symbols must @emph{not} | |
c0065db7 | 7943 | decorate them with dllexport. Instead, they should use dllimport, |
0a5d968e NC |
7944 | instead: |
7945 | ||
7946 | @example | |
7947 | __declspec(dllimport) int a_variable | |
7948 | __declspec(dllimport) void a_function(int with_args) | |
7949 | @end example | |
7950 | ||
c0065db7 RM |
7951 | This complicates the structure of library header files, because |
7952 | when included by the library itself the header must declare the | |
0a5d968e NC |
7953 | variables and functions as dllexport, but when included by client |
7954 | code the header must declare them as dllimport. There are a number | |
c0065db7 | 7955 | of idioms that are typically used to do this; often client code can |
0a5d968e NC |
7956 | omit the __declspec() declaration completely. See |
7957 | @samp{--enable-auto-import} and @samp{automatic data imports} for more | |
b45619c0 | 7958 | information. |
c0065db7 | 7959 | @end table |
dc8465bf | 7960 | |
2ca22b03 NC |
7961 | @cindex automatic data imports |
7962 | @item automatic data imports | |
7963 | The standard Windows dll format supports data imports from dlls only | |
69da35b5 | 7964 | by adding special decorations (dllimport/dllexport), which let the |
2ca22b03 | 7965 | compiler produce specific assembler instructions to deal with this |
c0065db7 | 7966 | issue. This increases the effort necessary to port existing Un*x |
69da35b5 | 7967 | code to these platforms, especially for large |
2ca22b03 | 7968 | c++ libraries and applications. The auto-import feature, which was |
c0065db7 | 7969 | initially provided by Paul Sokolovsky, allows one to omit the |
b45619c0 | 7970 | decorations to achieve a behavior that conforms to that on POSIX/Un*x |
c0065db7 | 7971 | platforms. This feature is enabled with the @samp{--enable-auto-import} |
69da35b5 NC |
7972 | command-line option, although it is enabled by default on cygwin/mingw. |
7973 | The @samp{--enable-auto-import} option itself now serves mainly to | |
7974 | suppress any warnings that are ordinarily emitted when linked objects | |
7975 | trigger the feature's use. | |
7976 | ||
c0065db7 | 7977 | auto-import of variables does not always work flawlessly without |
69da35b5 NC |
7978 | additional assistance. Sometimes, you will see this message |
7979 | ||
c0065db7 | 7980 | "variable '<var>' can't be auto-imported. Please read the |
69da35b5 NC |
7981 | documentation for ld's @code{--enable-auto-import} for details." |
7982 | ||
c0065db7 RM |
7983 | The @samp{--enable-auto-import} documentation explains why this error |
7984 | occurs, and several methods that can be used to overcome this difficulty. | |
7985 | One of these methods is the @emph{runtime pseudo-relocs} feature, described | |
69da35b5 NC |
7986 | below. |
7987 | ||
7988 | @cindex runtime pseudo-relocation | |
c0065db7 RM |
7989 | For complex variables imported from DLLs (such as structs or classes), |
7990 | object files typically contain a base address for the variable and an | |
7991 | offset (@emph{addend}) within the variable--to specify a particular | |
7992 | field or public member, for instance. Unfortunately, the runtime loader used | |
7993 | in win32 environments is incapable of fixing these references at runtime | |
69da35b5 | 7994 | without the additional information supplied by dllimport/dllexport decorations. |
c0065db7 | 7995 | The standard auto-import feature described above is unable to resolve these |
69da35b5 NC |
7996 | references. |
7997 | ||
c0065db7 RM |
7998 | The @samp{--enable-runtime-pseudo-relocs} switch allows these references to |
7999 | be resolved without error, while leaving the task of adjusting the references | |
8000 | themselves (with their non-zero addends) to specialized code provided by the | |
8001 | runtime environment. Recent versions of the cygwin and mingw environments and | |
8002 | compilers provide this runtime support; older versions do not. However, the | |
8003 | support is only necessary on the developer's platform; the compiled result will | |
69da35b5 NC |
8004 | run without error on an older system. |
8005 | ||
c0065db7 RM |
8006 | @samp{--enable-runtime-pseudo-relocs} is not the default; it must be explicitly |
8007 | enabled as needed. | |
2ca22b03 NC |
8008 | |
8009 | @cindex direct linking to a dll | |
8010 | @item direct linking to a dll | |
8011 | The cygwin/mingw ports of @command{ld} support the direct linking, | |
8012 | including data symbols, to a dll without the usage of any import | |
69da35b5 | 8013 | libraries. This is much faster and uses much less memory than does the |
b45619c0 | 8014 | traditional import library method, especially when linking large |
c0065db7 RM |
8015 | libraries or applications. When @command{ld} creates an import lib, each |
8016 | function or variable exported from the dll is stored in its own bfd, even | |
8017 | though a single bfd could contain many exports. The overhead involved in | |
69da35b5 | 8018 | storing, loading, and processing so many bfd's is quite large, and explains the |
c0065db7 | 8019 | tremendous time, memory, and storage needed to link against particularly |
69da35b5 NC |
8020 | large or complex libraries when using import libs. |
8021 | ||
c0065db7 | 8022 | Linking directly to a dll uses no extra command-line switches other than |
69da35b5 | 8023 | @samp{-L} and @samp{-l}, because @command{ld} already searches for a number |
c0065db7 | 8024 | of names to match each library. All that is needed from the developer's |
69da35b5 NC |
8025 | perspective is an understanding of this search, in order to force ld to |
8026 | select the dll instead of an import library. | |
8027 | ||
2ca22b03 | 8028 | |
69da35b5 NC |
8029 | For instance, when ld is called with the argument @samp{-lxxx} it will attempt |
8030 | to find, in the first directory of its search path, | |
2ca22b03 NC |
8031 | |
8032 | @example | |
45e948fe NC |
8033 | libxxx.dll.a |
8034 | xxx.dll.a | |
8035 | libxxx.a | |
8036 | xxx.lib | |
f6c4d4b1 | 8037 | libxxx.lib |
69da35b5 | 8038 | cygxxx.dll (*) |
45e948fe NC |
8039 | libxxx.dll |
8040 | xxx.dll | |
2ca22b03 NC |
8041 | @end example |
8042 | ||
69da35b5 NC |
8043 | before moving on to the next directory in the search path. |
8044 | ||
c0065db7 RM |
8045 | (*) Actually, this is not @samp{cygxxx.dll} but in fact is @samp{<prefix>xxx.dll}, |
8046 | where @samp{<prefix>} is set by the @command{ld} option | |
8047 | @samp{--dll-search-prefix=<prefix>}. In the case of cygwin, the standard gcc spec | |
8048 | file includes @samp{--dll-search-prefix=cyg}, so in effect we actually search for | |
69da35b5 NC |
8049 | @samp{cygxxx.dll}. |
8050 | ||
c0065db7 RM |
8051 | Other win32-based unix environments, such as mingw or pw32, may use other |
8052 | @samp{<prefix>}es, although at present only cygwin makes use of this feature. It | |
69da35b5 NC |
8053 | was originally intended to help avoid name conflicts among dll's built for the |
8054 | various win32/un*x environments, so that (for example) two versions of a zlib dll | |
8055 | could coexist on the same machine. | |
8056 | ||
2ca22b03 NC |
8057 | The generic cygwin/mingw path layout uses a @samp{bin} directory for |
8058 | applications and dll's and a @samp{lib} directory for the import | |
69da35b5 | 8059 | libraries (using cygwin nomenclature): |
2ca22b03 NC |
8060 | |
8061 | @example | |
8062 | bin/ | |
8063 | cygxxx.dll | |
8064 | lib/ | |
8065 | libxxx.dll.a (in case of dll's) | |
c0065db7 | 8066 | libxxx.a (in case of static archive) |
2ca22b03 NC |
8067 | @end example |
8068 | ||
c0065db7 RM |
8069 | Linking directly to a dll without using the import library can be |
8070 | done two ways: | |
2ca22b03 NC |
8071 | |
8072 | 1. Use the dll directly by adding the @samp{bin} path to the link line | |
8073 | @example | |
8074 | gcc -Wl,-verbose -o a.exe -L../bin/ -lxxx | |
c0065db7 | 8075 | @end example |
2ca22b03 | 8076 | |
69da35b5 NC |
8077 | However, as the dll's often have version numbers appended to their names |
8078 | (@samp{cygncurses-5.dll}) this will often fail, unless one specifies | |
8079 | @samp{-L../bin -lncurses-5} to include the version. Import libs are generally | |
8080 | not versioned, and do not have this difficulty. | |
8081 | ||
2ca22b03 NC |
8082 | 2. Create a symbolic link from the dll to a file in the @samp{lib} |
8083 | directory according to the above mentioned search pattern. This | |
8084 | should be used to avoid unwanted changes in the tools needed for | |
8085 | making the app/dll. | |
8086 | ||
8087 | @example | |
8088 | ln -s bin/cygxxx.dll lib/[cyg|lib|]xxx.dll[.a] | |
c0065db7 | 8089 | @end example |
2ca22b03 NC |
8090 | |
8091 | Then you can link without any make environment changes. | |
8092 | ||
8093 | @example | |
8094 | gcc -Wl,-verbose -o a.exe -L../lib/ -lxxx | |
c0065db7 | 8095 | @end example |
69da35b5 NC |
8096 | |
8097 | This technique also avoids the version number problems, because the following is | |
8098 | perfectly legal | |
8099 | ||
8100 | @example | |
8101 | bin/ | |
8102 | cygxxx-5.dll | |
8103 | lib/ | |
c0065db7 | 8104 | libxxx.dll.a -> ../bin/cygxxx-5.dll |
69da35b5 NC |
8105 | @end example |
8106 | ||
dc8465bf | 8107 | Linking directly to a dll without using an import lib will work |
69da35b5 NC |
8108 | even when auto-import features are exercised, and even when |
8109 | @samp{--enable-runtime-pseudo-relocs} is used. | |
8110 | ||
8111 | Given the improvements in speed and memory usage, one might justifiably | |
45e948fe | 8112 | wonder why import libraries are used at all. There are three reasons: |
69da35b5 NC |
8113 | |
8114 | 1. Until recently, the link-directly-to-dll functionality did @emph{not} | |
8115 | work with auto-imported data. | |
8116 | ||
dc8465bf NC |
8117 | 2. Sometimes it is necessary to include pure static objects within the |
8118 | import library (which otherwise contains only bfd's for indirection | |
8119 | symbols that point to the exports of a dll). Again, the import lib | |
8120 | for the cygwin kernel makes use of this ability, and it is not | |
8121 | possible to do this without an import lib. | |
69da35b5 | 8122 | |
45e948fe NC |
8123 | 3. Symbol aliases can only be resolved using an import lib. This is |
8124 | critical when linking against OS-supplied dll's (eg, the win32 API) | |
8125 | in which symbols are usually exported as undecorated aliases of their | |
8126 | stdcall-decorated assembly names. | |
8127 | ||
69da35b5 | 8128 | So, import libs are not going away. But the ability to replace |
c0065db7 RM |
8129 | true import libs with a simple symbolic link to (or a copy of) |
8130 | a dll, in many cases, is a useful addition to the suite of tools | |
8131 | binutils makes available to the win32 developer. Given the | |
69da35b5 NC |
8132 | massive improvements in memory requirements during linking, storage |
8133 | requirements, and linking speed, we expect that many developers | |
8134 | will soon begin to use this feature whenever possible. | |
dc8465bf | 8135 | |
c0065db7 | 8136 | @item symbol aliasing |
dc8465bf | 8137 | @table @emph |
c0065db7 RM |
8138 | @item adding additional names |
8139 | Sometimes, it is useful to export symbols with additional names. | |
dc8465bf NC |
8140 | A symbol @samp{foo} will be exported as @samp{foo}, but it can also be |
8141 | exported as @samp{_foo} by using special directives in the DEF file | |
8142 | when creating the dll. This will affect also the optional created | |
c0065db7 | 8143 | import library. Consider the following DEF file: |
dc8465bf | 8144 | |
c0065db7 | 8145 | @example |
dc8465bf NC |
8146 | LIBRARY "xyz.dll" BASE=0x61000000 |
8147 | ||
8148 | EXPORTS | |
c0065db7 | 8149 | foo |
dc8465bf | 8150 | _foo = foo |
c0065db7 | 8151 | @end example |
dc8465bf NC |
8152 | |
8153 | The line @samp{_foo = foo} maps the symbol @samp{foo} to @samp{_foo}. | |
8154 | ||
8155 | Another method for creating a symbol alias is to create it in the | |
8156 | source code using the "weak" attribute: | |
8157 | ||
c0065db7 RM |
8158 | @example |
8159 | void foo () @{ /* Do something. */; @} | |
dc8465bf | 8160 | void _foo () __attribute__ ((weak, alias ("foo"))); |
c0065db7 | 8161 | @end example |
dc8465bf NC |
8162 | |
8163 | See the gcc manual for more information about attributes and weak | |
8164 | symbols. | |
8165 | ||
8166 | @item renaming symbols | |
8167 | Sometimes it is useful to rename exports. For instance, the cygwin | |
c0065db7 | 8168 | kernel does this regularly. A symbol @samp{_foo} can be exported as |
dc8465bf NC |
8169 | @samp{foo} but not as @samp{_foo} by using special directives in the |
8170 | DEF file. (This will also affect the import library, if it is | |
c0065db7 | 8171 | created). In the following example: |
dc8465bf | 8172 | |
c0065db7 | 8173 | @example |
dc8465bf NC |
8174 | LIBRARY "xyz.dll" BASE=0x61000000 |
8175 | ||
8176 | EXPORTS | |
8177 | _foo = foo | |
c0065db7 | 8178 | @end example |
dc8465bf NC |
8179 | |
8180 | The line @samp{_foo = foo} maps the exported symbol @samp{foo} to | |
8181 | @samp{_foo}. | |
c0065db7 | 8182 | @end table |
dc8465bf | 8183 | |
0a5d968e | 8184 | Note: using a DEF file disables the default auto-export behavior, |
a05a5b64 | 8185 | unless the @samp{--export-all-symbols} command-line option is used. |
0a5d968e | 8186 | If, however, you are trying to rename symbols, then you should list |
c0065db7 RM |
8187 | @emph{all} desired exports in the DEF file, including the symbols |
8188 | that are not being renamed, and do @emph{not} use the | |
8189 | @samp{--export-all-symbols} option. If you list only the | |
8190 | renamed symbols in the DEF file, and use @samp{--export-all-symbols} | |
8191 | to handle the other symbols, then the both the new names @emph{and} | |
8192 | the original names for the renamed symbols will be exported. | |
8193 | In effect, you'd be aliasing those symbols, not renaming them, | |
0a5d968e | 8194 | which is probably not what you wanted. |
c87db184 CF |
8195 | |
8196 | @cindex weak externals | |
8197 | @item weak externals | |
8198 | The Windows object format, PE, specifies a form of weak symbols called | |
8199 | weak externals. When a weak symbol is linked and the symbol is not | |
8200 | defined, the weak symbol becomes an alias for some other symbol. There | |
8201 | are three variants of weak externals: | |
8202 | @itemize | |
8203 | @item Definition is searched for in objects and libraries, historically | |
8204 | called lazy externals. | |
8205 | @item Definition is searched for only in other objects, not in libraries. | |
8206 | This form is not presently implemented. | |
8207 | @item No search; the symbol is an alias. This form is not presently | |
8208 | implemented. | |
8209 | @end itemize | |
8210 | As a GNU extension, weak symbols that do not specify an alternate symbol | |
8211 | are supported. If the symbol is undefined when linking, the symbol | |
8212 | uses a default value. | |
c1711530 DK |
8213 | |
8214 | @cindex aligned common symbols | |
8215 | @item aligned common symbols | |
8216 | As a GNU extension to the PE file format, it is possible to specify the | |
8217 | desired alignment for a common symbol. This information is conveyed from | |
8218 | the assembler or compiler to the linker by means of GNU-specific commands | |
8219 | carried in the object file's @samp{.drectve} section, which are recognized | |
8220 | by @command{ld} and respected when laying out the common symbols. Native | |
8221 | tools will be able to process object files employing this GNU extension, | |
8222 | but will fail to respect the alignment instructions, and may issue noisy | |
8223 | warnings about unknown linker directives. | |
5063daf7 | 8224 | |
2ca22b03 NC |
8225 | @end table |
8226 | ||
8227 | @ifclear GENERIC | |
8228 | @lowersections | |
8229 | @end ifclear | |
8230 | @end ifset | |
8231 | ||
e0001a05 NC |
8232 | @ifset XTENSA |
8233 | @ifclear GENERIC | |
8234 | @raisesections | |
8235 | @end ifclear | |
8236 | ||
8237 | @node Xtensa | |
8238 | @section @code{ld} and Xtensa Processors | |
8239 | ||
8240 | @cindex Xtensa processors | |
8241 | The default @command{ld} behavior for Xtensa processors is to interpret | |
8242 | @code{SECTIONS} commands so that lists of explicitly named sections in a | |
8243 | specification with a wildcard file will be interleaved when necessary to | |
8244 | keep literal pools within the range of PC-relative load offsets. For | |
8245 | example, with the command: | |
8246 | ||
8247 | @smallexample | |
8248 | SECTIONS | |
8249 | @{ | |
8250 | .text : @{ | |
8251 | *(.literal .text) | |
8252 | @} | |
8253 | @} | |
8254 | @end smallexample | |
8255 | ||
8256 | @noindent | |
8257 | @command{ld} may interleave some of the @code{.literal} | |
8258 | and @code{.text} sections from different object files to ensure that the | |
8259 | literal pools are within the range of PC-relative load offsets. A valid | |
8260 | interleaving might place the @code{.literal} sections from an initial | |
8261 | group of files followed by the @code{.text} sections of that group of | |
8262 | files. Then, the @code{.literal} sections from the rest of the files | |
8263 | and the @code{.text} sections from the rest of the files would follow. | |
e0001a05 | 8264 | |
43cd72b9 | 8265 | @cindex @option{--relax} on Xtensa |
e0001a05 | 8266 | @cindex relaxing on Xtensa |
43cd72b9 BW |
8267 | Relaxation is enabled by default for the Xtensa version of @command{ld} and |
8268 | provides two important link-time optimizations. The first optimization | |
8269 | is to combine identical literal values to reduce code size. A redundant | |
8270 | literal will be removed and all the @code{L32R} instructions that use it | |
8271 | will be changed to reference an identical literal, as long as the | |
8272 | location of the replacement literal is within the offset range of all | |
8273 | the @code{L32R} instructions. The second optimization is to remove | |
8274 | unnecessary overhead from assembler-generated ``longcall'' sequences of | |
8275 | @code{L32R}/@code{CALLX@var{n}} when the target functions are within | |
8276 | range of direct @code{CALL@var{n}} instructions. | |
8277 | ||
8278 | For each of these cases where an indirect call sequence can be optimized | |
8279 | to a direct call, the linker will change the @code{CALLX@var{n}} | |
8280 | instruction to a @code{CALL@var{n}} instruction, remove the @code{L32R} | |
8281 | instruction, and remove the literal referenced by the @code{L32R} | |
8282 | instruction if it is not used for anything else. Removing the | |
8283 | @code{L32R} instruction always reduces code size but can potentially | |
8284 | hurt performance by changing the alignment of subsequent branch targets. | |
8285 | By default, the linker will always preserve alignments, either by | |
8286 | switching some instructions between 24-bit encodings and the equivalent | |
8287 | density instructions or by inserting a no-op in place of the @code{L32R} | |
8288 | instruction that was removed. If code size is more important than | |
8289 | performance, the @option{--size-opt} option can be used to prevent the | |
8290 | linker from widening density instructions or inserting no-ops, except in | |
8291 | a few cases where no-ops are required for correctness. | |
8292 | ||
8293 | The following Xtensa-specific command-line options can be used to | |
8294 | control the linker: | |
8295 | ||
8296 | @cindex Xtensa options | |
8297 | @table @option | |
43cd72b9 BW |
8298 | @item --size-opt |
8299 | When optimizing indirect calls to direct calls, optimize for code size | |
8300 | more than performance. With this option, the linker will not insert | |
8301 | no-ops or widen density instructions to preserve branch target | |
8302 | alignment. There may still be some cases where no-ops are required to | |
8303 | preserve the correctness of the code. | |
8304 | @end table | |
e0001a05 NC |
8305 | |
8306 | @ifclear GENERIC | |
8307 | @lowersections | |
8308 | @end ifclear | |
8309 | @end ifset | |
8310 | ||
252b5132 RH |
8311 | @ifclear SingleFormat |
8312 | @node BFD | |
8313 | @chapter BFD | |
8314 | ||
8315 | @cindex back end | |
8316 | @cindex object file management | |
8317 | @cindex object formats available | |
8318 | @kindex objdump -i | |
8319 | The linker accesses object and archive files using the BFD libraries. | |
8320 | These libraries allow the linker to use the same routines to operate on | |
8321 | object files whatever the object file format. A different object file | |
8322 | format can be supported simply by creating a new BFD back end and adding | |
8323 | it to the library. To conserve runtime memory, however, the linker and | |
8324 | associated tools are usually configured to support only a subset of the | |
8325 | object file formats available. You can use @code{objdump -i} | |
8326 | (@pxref{objdump,,objdump,binutils.info,The GNU Binary Utilities}) to | |
8327 | list all the formats available for your configuration. | |
8328 | ||
8329 | @cindex BFD requirements | |
8330 | @cindex requirements for BFD | |
8331 | As with most implementations, BFD is a compromise between | |
8332 | several conflicting requirements. The major factor influencing | |
8333 | BFD design was efficiency: any time used converting between | |
8334 | formats is time which would not have been spent had BFD not | |
8335 | been involved. This is partly offset by abstraction payback; since | |
8336 | BFD simplifies applications and back ends, more time and care | |
8337 | may be spent optimizing algorithms for a greater speed. | |
8338 | ||
8339 | One minor artifact of the BFD solution which you should bear in | |
8340 | mind is the potential for information loss. There are two places where | |
8341 | useful information can be lost using the BFD mechanism: during | |
8342 | conversion and during output. @xref{BFD information loss}. | |
8343 | ||
8344 | @menu | |
8345 | * BFD outline:: How it works: an outline of BFD | |
8346 | @end menu | |
8347 | ||
8348 | @node BFD outline | |
36f63dca | 8349 | @section How It Works: An Outline of BFD |
252b5132 RH |
8350 | @cindex opening object files |
8351 | @include bfdsumm.texi | |
8352 | @end ifclear | |
8353 | ||
8354 | @node Reporting Bugs | |
8355 | @chapter Reporting Bugs | |
ff5dcc92 SC |
8356 | @cindex bugs in @command{ld} |
8357 | @cindex reporting bugs in @command{ld} | |
252b5132 | 8358 | |
ff5dcc92 | 8359 | Your bug reports play an essential role in making @command{ld} reliable. |
252b5132 RH |
8360 | |
8361 | Reporting a bug may help you by bringing a solution to your problem, or | |
8362 | it may not. But in any case the principal function of a bug report is | |
ff5dcc92 | 8363 | to help the entire community by making the next version of @command{ld} |
252b5132 | 8364 | work better. Bug reports are your contribution to the maintenance of |
ff5dcc92 | 8365 | @command{ld}. |
252b5132 RH |
8366 | |
8367 | In order for a bug report to serve its purpose, you must include the | |
8368 | information that enables us to fix the bug. | |
8369 | ||
8370 | @menu | |
8371 | * Bug Criteria:: Have you found a bug? | |
8372 | * Bug Reporting:: How to report bugs | |
8373 | @end menu | |
8374 | ||
8375 | @node Bug Criteria | |
36f63dca | 8376 | @section Have You Found a Bug? |
252b5132 RH |
8377 | @cindex bug criteria |
8378 | ||
8379 | If you are not sure whether you have found a bug, here are some guidelines: | |
8380 | ||
8381 | @itemize @bullet | |
8382 | @cindex fatal signal | |
8383 | @cindex linker crash | |
8384 | @cindex crash of linker | |
8385 | @item | |
8386 | If the linker gets a fatal signal, for any input whatever, that is a | |
ff5dcc92 | 8387 | @command{ld} bug. Reliable linkers never crash. |
252b5132 RH |
8388 | |
8389 | @cindex error on valid input | |
8390 | @item | |
ff5dcc92 | 8391 | If @command{ld} produces an error message for valid input, that is a bug. |
252b5132 RH |
8392 | |
8393 | @cindex invalid input | |
8394 | @item | |
ff5dcc92 | 8395 | If @command{ld} does not produce an error message for invalid input, that |
252b5132 RH |
8396 | may be a bug. In the general case, the linker can not verify that |
8397 | object files are correct. | |
8398 | ||
8399 | @item | |
8400 | If you are an experienced user of linkers, your suggestions for | |
ff5dcc92 | 8401 | improvement of @command{ld} are welcome in any case. |
252b5132 RH |
8402 | @end itemize |
8403 | ||
8404 | @node Bug Reporting | |
36f63dca | 8405 | @section How to Report Bugs |
252b5132 | 8406 | @cindex bug reports |
ff5dcc92 | 8407 | @cindex @command{ld} bugs, reporting |
252b5132 RH |
8408 | |
8409 | A number of companies and individuals offer support for @sc{gnu} | |
ff5dcc92 | 8410 | products. If you obtained @command{ld} from a support organization, we |
252b5132 RH |
8411 | recommend you contact that organization first. |
8412 | ||
8413 | You can find contact information for many support companies and | |
8414 | individuals in the file @file{etc/SERVICE} in the @sc{gnu} Emacs | |
8415 | distribution. | |
8416 | ||
ad22bfe8 | 8417 | @ifset BUGURL |
ff5dcc92 | 8418 | Otherwise, send bug reports for @command{ld} to |
ad22bfe8 JM |
8419 | @value{BUGURL}. |
8420 | @end ifset | |
252b5132 RH |
8421 | |
8422 | The fundamental principle of reporting bugs usefully is this: | |
8423 | @strong{report all the facts}. If you are not sure whether to state a | |
8424 | fact or leave it out, state it! | |
8425 | ||
8426 | Often people omit facts because they think they know what causes the | |
8427 | problem and assume that some details do not matter. Thus, you might | |
b553b183 NC |
8428 | assume that the name of a symbol you use in an example does not |
8429 | matter. Well, probably it does not, but one cannot be sure. Perhaps | |
8430 | the bug is a stray memory reference which happens to fetch from the | |
8431 | location where that name is stored in memory; perhaps, if the name | |
8432 | were different, the contents of that location would fool the linker | |
8433 | into doing the right thing despite the bug. Play it safe and give a | |
8434 | specific, complete example. That is the easiest thing for you to do, | |
c0065db7 | 8435 | and the most helpful. |
b553b183 NC |
8436 | |
8437 | Keep in mind that the purpose of a bug report is to enable us to fix | |
8438 | the bug if it is new to us. Therefore, always write your bug reports | |
8439 | on the assumption that the bug has not been reported previously. | |
252b5132 RH |
8440 | |
8441 | Sometimes people give a few sketchy facts and ask, ``Does this ring a | |
36f63dca NC |
8442 | bell?'' This cannot help us fix a bug, so it is basically useless. We |
8443 | respond by asking for enough details to enable us to investigate. | |
8444 | You might as well expedite matters by sending them to begin with. | |
252b5132 RH |
8445 | |
8446 | To enable us to fix the bug, you should include all these things: | |
8447 | ||
8448 | @itemize @bullet | |
8449 | @item | |
ff5dcc92 | 8450 | The version of @command{ld}. @command{ld} announces it if you start it with |
252b5132 RH |
8451 | the @samp{--version} argument. |
8452 | ||
8453 | Without this, we will not know whether there is any point in looking for | |
ff5dcc92 | 8454 | the bug in the current version of @command{ld}. |
252b5132 RH |
8455 | |
8456 | @item | |
ff5dcc92 | 8457 | Any patches you may have applied to the @command{ld} source, including any |
252b5132 RH |
8458 | patches made to the @code{BFD} library. |
8459 | ||
8460 | @item | |
8461 | The type of machine you are using, and the operating system name and | |
8462 | version number. | |
8463 | ||
8464 | @item | |
ff5dcc92 | 8465 | What compiler (and its version) was used to compile @command{ld}---e.g. |
252b5132 RH |
8466 | ``@code{gcc-2.7}''. |
8467 | ||
8468 | @item | |
8469 | The command arguments you gave the linker to link your example and | |
8470 | observe the bug. To guarantee you will not omit something important, | |
8471 | list them all. A copy of the Makefile (or the output from make) is | |
8472 | sufficient. | |
8473 | ||
8474 | If we were to try to guess the arguments, we would probably guess wrong | |
8475 | and then we might not encounter the bug. | |
8476 | ||
8477 | @item | |
8478 | A complete input file, or set of input files, that will reproduce the | |
b553b183 NC |
8479 | bug. It is generally most helpful to send the actual object files |
8480 | provided that they are reasonably small. Say no more than 10K. For | |
8481 | bigger files you can either make them available by FTP or HTTP or else | |
8482 | state that you are willing to send the object file(s) to whomever | |
8483 | requests them. (Note - your email will be going to a mailing list, so | |
8484 | we do not want to clog it up with large attachments). But small | |
8485 | attachments are best. | |
252b5132 RH |
8486 | |
8487 | If the source files were assembled using @code{gas} or compiled using | |
8488 | @code{gcc}, then it may be OK to send the source files rather than the | |
8489 | object files. In this case, be sure to say exactly what version of | |
8490 | @code{gas} or @code{gcc} was used to produce the object files. Also say | |
8491 | how @code{gas} or @code{gcc} were configured. | |
8492 | ||
8493 | @item | |
8494 | A description of what behavior you observe that you believe is | |
8495 | incorrect. For example, ``It gets a fatal signal.'' | |
8496 | ||
ff5dcc92 | 8497 | Of course, if the bug is that @command{ld} gets a fatal signal, then we |
252b5132 RH |
8498 | will certainly notice it. But if the bug is incorrect output, we might |
8499 | not notice unless it is glaringly wrong. You might as well not give us | |
8500 | a chance to make a mistake. | |
8501 | ||
8502 | Even if the problem you experience is a fatal signal, you should still | |
8503 | say so explicitly. Suppose something strange is going on, such as, your | |
b45619c0 | 8504 | copy of @command{ld} is out of sync, or you have encountered a bug in the |
252b5132 RH |
8505 | C library on your system. (This has happened!) Your copy might crash |
8506 | and ours would not. If you told us to expect a crash, then when ours | |
8507 | fails to crash, we would know that the bug was not happening for us. If | |
8508 | you had not told us to expect a crash, then we would not be able to draw | |
8509 | any conclusion from our observations. | |
8510 | ||
8511 | @item | |
ff5dcc92 | 8512 | If you wish to suggest changes to the @command{ld} source, send us context |
252b5132 RH |
8513 | diffs, as generated by @code{diff} with the @samp{-u}, @samp{-c}, or |
8514 | @samp{-p} option. Always send diffs from the old file to the new file. | |
ff5dcc92 | 8515 | If you even discuss something in the @command{ld} source, refer to it by |
252b5132 RH |
8516 | context, not by line number. |
8517 | ||
8518 | The line numbers in our development sources will not match those in your | |
8519 | sources. Your line numbers would convey no useful information to us. | |
8520 | @end itemize | |
8521 | ||
8522 | Here are some things that are not necessary: | |
8523 | ||
8524 | @itemize @bullet | |
8525 | @item | |
8526 | A description of the envelope of the bug. | |
8527 | ||
8528 | Often people who encounter a bug spend a lot of time investigating | |
8529 | which changes to the input file will make the bug go away and which | |
8530 | changes will not affect it. | |
8531 | ||
8532 | This is often time consuming and not very useful, because the way we | |
8533 | will find the bug is by running a single example under the debugger | |
8534 | with breakpoints, not by pure deduction from a series of examples. | |
8535 | We recommend that you save your time for something else. | |
8536 | ||
8537 | Of course, if you can find a simpler example to report @emph{instead} | |
8538 | of the original one, that is a convenience for us. Errors in the | |
8539 | output will be easier to spot, running under the debugger will take | |
8540 | less time, and so on. | |
8541 | ||
8542 | However, simplification is not vital; if you do not want to do this, | |
8543 | report the bug anyway and send us the entire test case you used. | |
8544 | ||
8545 | @item | |
8546 | A patch for the bug. | |
8547 | ||
8548 | A patch for the bug does help us if it is a good one. But do not omit | |
8549 | the necessary information, such as the test case, on the assumption that | |
8550 | a patch is all we need. We might see problems with your patch and decide | |
8551 | to fix the problem another way, or we might not understand it at all. | |
8552 | ||
ff5dcc92 | 8553 | Sometimes with a program as complicated as @command{ld} it is very hard to |
252b5132 RH |
8554 | construct an example that will make the program follow a certain path |
8555 | through the code. If you do not send us the example, we will not be | |
8556 | able to construct one, so we will not be able to verify that the bug is | |
8557 | fixed. | |
8558 | ||
8559 | And if we cannot understand what bug you are trying to fix, or why your | |
8560 | patch should be an improvement, we will not install it. A test case will | |
8561 | help us to understand. | |
8562 | ||
8563 | @item | |
8564 | A guess about what the bug is or what it depends on. | |
8565 | ||
8566 | Such guesses are usually wrong. Even we cannot guess right about such | |
8567 | things without first using the debugger to find the facts. | |
8568 | @end itemize | |
8569 | ||
8570 | @node MRI | |
8571 | @appendix MRI Compatible Script Files | |
8572 | @cindex MRI compatibility | |
ff5dcc92 SC |
8573 | To aid users making the transition to @sc{gnu} @command{ld} from the MRI |
8574 | linker, @command{ld} can use MRI compatible linker scripts as an | |
252b5132 RH |
8575 | alternative to the more general-purpose linker scripting language |
8576 | described in @ref{Scripts}. MRI compatible linker scripts have a much | |
8577 | simpler command set than the scripting language otherwise used with | |
ff5dcc92 | 8578 | @command{ld}. @sc{gnu} @command{ld} supports the most commonly used MRI |
252b5132 RH |
8579 | linker commands; these commands are described here. |
8580 | ||
8581 | In general, MRI scripts aren't of much use with the @code{a.out} object | |
8582 | file format, since it only has three sections and MRI scripts lack some | |
8583 | features to make use of them. | |
8584 | ||
8585 | You can specify a file containing an MRI-compatible script using the | |
8586 | @samp{-c} command-line option. | |
8587 | ||
8588 | Each command in an MRI-compatible script occupies its own line; each | |
8589 | command line starts with the keyword that identifies the command (though | |
8590 | blank lines are also allowed for punctuation). If a line of an | |
ff5dcc92 | 8591 | MRI-compatible script begins with an unrecognized keyword, @command{ld} |
252b5132 RH |
8592 | issues a warning message, but continues processing the script. |
8593 | ||
8594 | Lines beginning with @samp{*} are comments. | |
8595 | ||
8596 | You can write these commands using all upper-case letters, or all | |
8597 | lower case; for example, @samp{chip} is the same as @samp{CHIP}. | |
8598 | The following list shows only the upper-case form of each command. | |
8599 | ||
8600 | @table @code | |
8601 | @cindex @code{ABSOLUTE} (MRI) | |
8602 | @item ABSOLUTE @var{secname} | |
8603 | @itemx ABSOLUTE @var{secname}, @var{secname}, @dots{} @var{secname} | |
ff5dcc92 | 8604 | Normally, @command{ld} includes in the output file all sections from all |
252b5132 RH |
8605 | the input files. However, in an MRI-compatible script, you can use the |
8606 | @code{ABSOLUTE} command to restrict the sections that will be present in | |
8607 | your output program. If the @code{ABSOLUTE} command is used at all in a | |
8608 | script, then only the sections named explicitly in @code{ABSOLUTE} | |
8609 | commands will appear in the linker output. You can still use other | |
8610 | input sections (whatever you select on the command line, or using | |
8611 | @code{LOAD}) to resolve addresses in the output file. | |
8612 | ||
8613 | @cindex @code{ALIAS} (MRI) | |
8614 | @item ALIAS @var{out-secname}, @var{in-secname} | |
8615 | Use this command to place the data from input section @var{in-secname} | |
8616 | in a section called @var{out-secname} in the linker output file. | |
8617 | ||
8618 | @var{in-secname} may be an integer. | |
8619 | ||
8620 | @cindex @code{ALIGN} (MRI) | |
8621 | @item ALIGN @var{secname} = @var{expression} | |
8622 | Align the section called @var{secname} to @var{expression}. The | |
8623 | @var{expression} should be a power of two. | |
8624 | ||
8625 | @cindex @code{BASE} (MRI) | |
8626 | @item BASE @var{expression} | |
8627 | Use the value of @var{expression} as the lowest address (other than | |
8628 | absolute addresses) in the output file. | |
8629 | ||
8630 | @cindex @code{CHIP} (MRI) | |
8631 | @item CHIP @var{expression} | |
8632 | @itemx CHIP @var{expression}, @var{expression} | |
8633 | This command does nothing; it is accepted only for compatibility. | |
8634 | ||
8635 | @cindex @code{END} (MRI) | |
8636 | @item END | |
8637 | This command does nothing whatever; it's only accepted for compatibility. | |
8638 | ||
8639 | @cindex @code{FORMAT} (MRI) | |
8640 | @item FORMAT @var{output-format} | |
8641 | Similar to the @code{OUTPUT_FORMAT} command in the more general linker | |
dc12032b | 8642 | language, but restricted to S-records, if @var{output-format} is @samp{S} |
252b5132 RH |
8643 | |
8644 | @cindex @code{LIST} (MRI) | |
8645 | @item LIST @var{anything}@dots{} | |
8646 | Print (to the standard output file) a link map, as produced by the | |
ff5dcc92 | 8647 | @command{ld} command-line option @samp{-M}. |
252b5132 RH |
8648 | |
8649 | The keyword @code{LIST} may be followed by anything on the | |
8650 | same line, with no change in its effect. | |
8651 | ||
8652 | @cindex @code{LOAD} (MRI) | |
8653 | @item LOAD @var{filename} | |
8654 | @itemx LOAD @var{filename}, @var{filename}, @dots{} @var{filename} | |
8655 | Include one or more object file @var{filename} in the link; this has the | |
ff5dcc92 | 8656 | same effect as specifying @var{filename} directly on the @command{ld} |
252b5132 RH |
8657 | command line. |
8658 | ||
8659 | @cindex @code{NAME} (MRI) | |
8660 | @item NAME @var{output-name} | |
ff5dcc92 | 8661 | @var{output-name} is the name for the program produced by @command{ld}; the |
252b5132 RH |
8662 | MRI-compatible command @code{NAME} is equivalent to the command-line |
8663 | option @samp{-o} or the general script language command @code{OUTPUT}. | |
8664 | ||
8665 | @cindex @code{ORDER} (MRI) | |
8666 | @item ORDER @var{secname}, @var{secname}, @dots{} @var{secname} | |
8667 | @itemx ORDER @var{secname} @var{secname} @var{secname} | |
ff5dcc92 | 8668 | Normally, @command{ld} orders the sections in its output file in the |
252b5132 RH |
8669 | order in which they first appear in the input files. In an MRI-compatible |
8670 | script, you can override this ordering with the @code{ORDER} command. The | |
8671 | sections you list with @code{ORDER} will appear first in your output | |
8672 | file, in the order specified. | |
8673 | ||
8674 | @cindex @code{PUBLIC} (MRI) | |
8675 | @item PUBLIC @var{name}=@var{expression} | |
8676 | @itemx PUBLIC @var{name},@var{expression} | |
8677 | @itemx PUBLIC @var{name} @var{expression} | |
8678 | Supply a value (@var{expression}) for external symbol | |
8679 | @var{name} used in the linker input files. | |
8680 | ||
8681 | @cindex @code{SECT} (MRI) | |
8682 | @item SECT @var{secname}, @var{expression} | |
8683 | @itemx SECT @var{secname}=@var{expression} | |
8684 | @itemx SECT @var{secname} @var{expression} | |
8685 | You can use any of these three forms of the @code{SECT} command to | |
8686 | specify the start address (@var{expression}) for section @var{secname}. | |
8687 | If you have more than one @code{SECT} statement for the same | |
8688 | @var{secname}, only the @emph{first} sets the start address. | |
8689 | @end table | |
8690 | ||
793c5807 NC |
8691 | @node GNU Free Documentation License |
8692 | @appendix GNU Free Documentation License | |
36f63dca | 8693 | @include fdl.texi |
704c465c | 8694 | |
370b66a1 CD |
8695 | @node LD Index |
8696 | @unnumbered LD Index | |
252b5132 RH |
8697 | |
8698 | @printindex cp | |
8699 | ||
8700 | @tex | |
7ca01ed9 | 8701 | % I think something like @@colophon should be in texinfo. In the |
252b5132 RH |
8702 | % meantime: |
8703 | \long\def\colophon{\hbox to0pt{}\vfill | |
8704 | \centerline{The body of this manual is set in} | |
8705 | \centerline{\fontname\tenrm,} | |
8706 | \centerline{with headings in {\bf\fontname\tenbf}} | |
8707 | \centerline{and examples in {\tt\fontname\tentt}.} | |
8708 | \centerline{{\it\fontname\tenit\/} and} | |
8709 | \centerline{{\sl\fontname\tensl\/}} | |
8710 | \centerline{are used for emphasis.}\vfill} | |
8711 | \page\colophon | |
7ca01ed9 | 8712 | % Blame: doc@@cygnus.com, 28mar91. |
252b5132 RH |
8713 | @end tex |
8714 | ||
252b5132 | 8715 | @bye |