Andrea Falcone | 1c4977f | 2020-07-23 10:58:25 -0400 | [diff] [blame] | 1 | This is ld.info, produced by makeinfo version 4.8 from |
| 2 | /Volumes/Android/buildbot/src/android/binutils/toolchain/binutils/binutils-2.27/ld/ld.texinfo. |
| 3 | |
| 4 | INFO-DIR-SECTION Software development |
| 5 | START-INFO-DIR-ENTRY |
| 6 | * Ld: (ld). The GNU linker. |
| 7 | END-INFO-DIR-ENTRY |
| 8 | |
| 9 | This file documents the GNU linker LD (GNU Binutils) version 2.27.0. |
| 10 | |
| 11 | Copyright (C) 1991-2016 Free Software Foundation, Inc. |
| 12 | |
| 13 | Permission is granted to copy, distribute and/or modify this document |
| 14 | under the terms of the GNU Free Documentation License, Version 1.3 or |
| 15 | any later version published by the Free Software Foundation; with no |
| 16 | Invariant Sections, with no Front-Cover Texts, and with no Back-Cover |
| 17 | Texts. A copy of the license is included in the section entitled "GNU |
| 18 | Free Documentation License". |
| 19 | |
| 20 | |
| 21 | File: ld.info, Node: Top, Next: Overview, Up: (dir) |
| 22 | |
| 23 | LD |
| 24 | ** |
| 25 | |
| 26 | This file documents the GNU linker ld (GNU Binutils) version 2.27.0. |
| 27 | |
| 28 | This document is distributed under the terms of the GNU Free |
| 29 | Documentation License version 1.3. A copy of the license is included |
| 30 | in the section entitled "GNU Free Documentation License". |
| 31 | |
| 32 | * Menu: |
| 33 | |
| 34 | * Overview:: Overview |
| 35 | * Invocation:: Invocation |
| 36 | * Scripts:: Linker Scripts |
| 37 | |
| 38 | * Machine Dependent:: Machine Dependent Features |
| 39 | |
| 40 | * BFD:: BFD |
| 41 | |
| 42 | * Reporting Bugs:: Reporting Bugs |
| 43 | * MRI:: MRI Compatible Script Files |
| 44 | * GNU Free Documentation License:: GNU Free Documentation License |
| 45 | * LD Index:: LD Index |
| 46 | |
| 47 | |
| 48 | File: ld.info, Node: Overview, Next: Invocation, Prev: Top, Up: Top |
| 49 | |
| 50 | 1 Overview |
| 51 | ********** |
| 52 | |
| 53 | `ld' combines a number of object and archive files, relocates their |
| 54 | data and ties up symbol references. Usually the last step in compiling |
| 55 | a program is to run `ld'. |
| 56 | |
| 57 | `ld' accepts Linker Command Language files written in a superset of |
| 58 | AT&T's Link Editor Command Language syntax, to provide explicit and |
| 59 | total control over the linking process. |
| 60 | |
| 61 | This version of `ld' uses the general purpose BFD libraries to |
| 62 | operate on object files. This allows `ld' to read, combine, and write |
| 63 | object files in many different formats--for example, COFF or `a.out'. |
| 64 | Different formats may be linked together to produce any available kind |
| 65 | of object file. *Note BFD::, for more information. |
| 66 | |
| 67 | Aside from its flexibility, the GNU linker is more helpful than other |
| 68 | linkers in providing diagnostic information. Many linkers abandon |
| 69 | execution immediately upon encountering an error; whenever possible, |
| 70 | `ld' continues executing, allowing you to identify other errors (or, in |
| 71 | some cases, to get an output file in spite of the error). |
| 72 | |
| 73 | |
| 74 | File: ld.info, Node: Invocation, Next: Scripts, Prev: Overview, Up: Top |
| 75 | |
| 76 | 2 Invocation |
| 77 | ************ |
| 78 | |
| 79 | The GNU linker `ld' is meant to cover a broad range of situations, and |
| 80 | to be as compatible as possible with other linkers. As a result, you |
| 81 | have many choices to control its behavior. |
| 82 | |
| 83 | * Menu: |
| 84 | |
| 85 | * Options:: Command Line Options |
| 86 | * Environment:: Environment Variables |
| 87 | |
| 88 | |
| 89 | File: ld.info, Node: Options, Next: Environment, Up: Invocation |
| 90 | |
| 91 | 2.1 Command Line Options |
| 92 | ======================== |
| 93 | |
| 94 | The linker supports a plethora of command-line options, but in actual |
| 95 | practice few of them are used in any particular context. For instance, |
| 96 | a frequent use of `ld' is to link standard Unix object files on a |
| 97 | standard, supported Unix system. On such a system, to link a file |
| 98 | `hello.o': |
| 99 | |
| 100 | ld -o OUTPUT /lib/crt0.o hello.o -lc |
| 101 | |
| 102 | This tells `ld' to produce a file called OUTPUT as the result of |
| 103 | linking the file `/lib/crt0.o' with `hello.o' and the library `libc.a', |
| 104 | which will come from the standard search directories. (See the |
| 105 | discussion of the `-l' option below.) |
| 106 | |
| 107 | Some of the command-line options to `ld' may be specified at any |
| 108 | point in the command line. However, options which refer to files, such |
| 109 | as `-l' or `-T', cause the file to be read at the point at which the |
| 110 | option appears in the command line, relative to the object files and |
| 111 | other file options. Repeating non-file options with a different |
| 112 | argument will either have no further effect, or override prior |
| 113 | occurrences (those further to the left on the command line) of that |
| 114 | option. Options which may be meaningfully specified more than once are |
| 115 | noted in the descriptions below. |
| 116 | |
| 117 | Non-option arguments are object files or archives which are to be |
| 118 | linked together. They may follow, precede, or be mixed in with |
| 119 | command-line options, except that an object file argument may not be |
| 120 | placed between an option and its argument. |
| 121 | |
| 122 | Usually the linker is invoked with at least one object file, but you |
| 123 | can specify other forms of binary input files using `-l', `-R', and the |
| 124 | script command language. If _no_ binary input files at all are |
| 125 | specified, the linker does not produce any output, and issues the |
| 126 | message `No input files'. |
| 127 | |
| 128 | If the linker cannot recognize the format of an object file, it will |
| 129 | assume that it is a linker script. A script specified in this way |
| 130 | augments the main linker script used for the link (either the default |
| 131 | linker script or the one specified by using `-T'). This feature |
| 132 | permits the linker to link against a file which appears to be an object |
| 133 | or an archive, but actually merely defines some symbol values, or uses |
| 134 | `INPUT' or `GROUP' to load other objects. Specifying a script in this |
| 135 | way merely augments the main linker script, with the extra commands |
| 136 | placed after the main script; use the `-T' option to replace the |
| 137 | default linker script entirely, but note the effect of the `INSERT' |
| 138 | command. *Note Scripts::. |
| 139 | |
| 140 | For options whose names are a single letter, option arguments must |
| 141 | either follow the option letter without intervening whitespace, or be |
| 142 | given as separate arguments immediately following the option that |
| 143 | requires them. |
| 144 | |
| 145 | For options whose names are multiple letters, either one dash or two |
| 146 | can precede the option name; for example, `-trace-symbol' and |
| 147 | `--trace-symbol' are equivalent. Note--there is one exception to this |
| 148 | rule. Multiple letter options that start with a lower case 'o' can |
| 149 | only be preceded by two dashes. This is to reduce confusion with the |
| 150 | `-o' option. So for example `-omagic' sets the output file name to |
| 151 | `magic' whereas `--omagic' sets the NMAGIC flag on the output. |
| 152 | |
| 153 | Arguments to multiple-letter options must either be separated from |
| 154 | the option name by an equals sign, or be given as separate arguments |
| 155 | immediately following the option that requires them. For example, |
| 156 | `--trace-symbol foo' and `--trace-symbol=foo' are equivalent. Unique |
| 157 | abbreviations of the names of multiple-letter options are accepted. |
| 158 | |
| 159 | Note--if the linker is being invoked indirectly, via a compiler |
| 160 | driver (e.g. `gcc') then all the linker command line options should be |
| 161 | prefixed by `-Wl,' (or whatever is appropriate for the particular |
| 162 | compiler driver) like this: |
| 163 | |
| 164 | gcc -Wl,--start-group foo.o bar.o -Wl,--end-group |
| 165 | |
| 166 | This is important, because otherwise the compiler driver program may |
| 167 | silently drop the linker options, resulting in a bad link. Confusion |
| 168 | may also arise when passing options that require values through a |
| 169 | driver, as the use of a space between option and argument acts as a |
| 170 | separator, and causes the driver to pass only the option to the linker |
| 171 | and the argument to the compiler. In this case, it is simplest to use |
| 172 | the joined forms of both single- and multiple-letter options, such as: |
| 173 | |
| 174 | gcc foo.o bar.o -Wl,-eENTRY -Wl,-Map=a.map |
| 175 | |
| 176 | Here is a table of the generic command line switches accepted by the |
| 177 | GNU linker: |
| 178 | |
| 179 | `@FILE' |
| 180 | Read command-line options from FILE. The options read are |
| 181 | inserted in place of the original @FILE option. If FILE does not |
| 182 | exist, or cannot be read, then the option will be treated |
| 183 | literally, and not removed. |
| 184 | |
| 185 | Options in FILE are separated by whitespace. A whitespace |
| 186 | character may be included in an option by surrounding the entire |
| 187 | option in either single or double quotes. Any character |
| 188 | (including a backslash) may be included by prefixing the character |
| 189 | to be included with a backslash. The FILE may itself contain |
| 190 | additional @FILE options; any such options will be processed |
| 191 | recursively. |
| 192 | |
| 193 | `-a KEYWORD' |
| 194 | This option is supported for HP/UX compatibility. The KEYWORD |
| 195 | argument must be one of the strings `archive', `shared', or |
| 196 | `default'. `-aarchive' is functionally equivalent to `-Bstatic', |
| 197 | and the other two keywords are functionally equivalent to |
| 198 | `-Bdynamic'. This option may be used any number of times. |
| 199 | |
| 200 | `--audit AUDITLIB' |
| 201 | Adds AUDITLIB to the `DT_AUDIT' entry of the dynamic section. |
| 202 | AUDITLIB is not checked for existence, nor will it use the |
| 203 | DT_SONAME specified in the library. If specified multiple times |
| 204 | `DT_AUDIT' will contain a colon separated list of audit interfaces |
| 205 | to use. If the linker finds an object with an audit entry while |
| 206 | searching for shared libraries, it will add a corresponding |
| 207 | `DT_DEPAUDIT' entry in the output file. This option is only |
| 208 | meaningful on ELF platforms supporting the rtld-audit interface. |
| 209 | |
| 210 | `-A ARCHITECTURE' |
| 211 | `--architecture=ARCHITECTURE' |
| 212 | In the current release of `ld', this option is useful only for the |
| 213 | Intel 960 family of architectures. In that `ld' configuration, the |
| 214 | ARCHITECTURE argument identifies the particular architecture in |
| 215 | the 960 family, enabling some safeguards and modifying the |
| 216 | archive-library search path. *Note `ld' and the Intel 960 family: |
| 217 | i960, for details. |
| 218 | |
| 219 | Future releases of `ld' may support similar functionality for |
| 220 | other architecture families. |
| 221 | |
| 222 | `-b INPUT-FORMAT' |
| 223 | `--format=INPUT-FORMAT' |
| 224 | `ld' may be configured to support more than one kind of object |
| 225 | file. If your `ld' is configured this way, you can use the `-b' |
| 226 | option to specify the binary format for input object files that |
| 227 | follow this option on the command line. Even when `ld' is |
| 228 | configured to support alternative object formats, you don't |
| 229 | usually need to specify this, as `ld' should be configured to |
| 230 | expect as a default input format the most usual format on each |
| 231 | machine. INPUT-FORMAT is a text string, the name of a particular |
| 232 | format supported by the BFD libraries. (You can list the |
| 233 | available binary formats with `objdump -i'.) *Note BFD::. |
| 234 | |
| 235 | You may want to use this option if you are linking files with an |
| 236 | unusual binary format. You can also use `-b' to switch formats |
| 237 | explicitly (when linking object files of different formats), by |
| 238 | including `-b INPUT-FORMAT' before each group of object files in a |
| 239 | particular format. |
| 240 | |
| 241 | The default format is taken from the environment variable |
| 242 | `GNUTARGET'. *Note Environment::. You can also define the input |
| 243 | format from a script, using the command `TARGET'; see *Note Format |
| 244 | Commands::. |
| 245 | |
| 246 | `-c MRI-COMMANDFILE' |
| 247 | `--mri-script=MRI-COMMANDFILE' |
| 248 | For compatibility with linkers produced by MRI, `ld' accepts script |
| 249 | files written in an alternate, restricted command language, |
| 250 | described in *Note MRI Compatible Script Files: MRI. Introduce |
| 251 | MRI script files with the option `-c'; use the `-T' option to run |
| 252 | linker scripts written in the general-purpose `ld' scripting |
| 253 | language. If MRI-CMDFILE does not exist, `ld' looks for it in the |
| 254 | directories specified by any `-L' options. |
| 255 | |
| 256 | `-d' |
| 257 | `-dc' |
| 258 | `-dp' |
| 259 | These three options are equivalent; multiple forms are supported |
| 260 | for compatibility with other linkers. They assign space to common |
| 261 | symbols even if a relocatable output file is specified (with |
| 262 | `-r'). The script command `FORCE_COMMON_ALLOCATION' has the same |
| 263 | effect. *Note Miscellaneous Commands::. |
| 264 | |
| 265 | `--depaudit AUDITLIB' |
| 266 | `-P AUDITLIB' |
| 267 | Adds AUDITLIB to the `DT_DEPAUDIT' entry of the dynamic section. |
| 268 | AUDITLIB is not checked for existence, nor will it use the |
| 269 | DT_SONAME specified in the library. If specified multiple times |
| 270 | `DT_DEPAUDIT' will contain a colon separated list of audit |
| 271 | interfaces to use. This option is only meaningful on ELF |
| 272 | platforms supporting the rtld-audit interface. The -P option is |
| 273 | provided for Solaris compatibility. |
| 274 | |
| 275 | `-e ENTRY' |
| 276 | `--entry=ENTRY' |
| 277 | Use ENTRY as the explicit symbol for beginning execution of your |
| 278 | program, rather than the default entry point. If there is no |
| 279 | symbol named ENTRY, the linker will try to parse ENTRY as a number, |
| 280 | and use that as the entry address (the number will be interpreted |
| 281 | in base 10; you may use a leading `0x' for base 16, or a leading |
| 282 | `0' for base 8). *Note Entry Point::, for a discussion of defaults |
| 283 | and other ways of specifying the entry point. |
| 284 | |
| 285 | `--exclude-libs LIB,LIB,...' |
| 286 | Specifies a list of archive libraries from which symbols should |
| 287 | not be automatically exported. The library names may be delimited |
| 288 | by commas or colons. Specifying `--exclude-libs ALL' excludes |
| 289 | symbols in all archive libraries from automatic export. This |
| 290 | option is available only for the i386 PE targeted port of the |
| 291 | linker and for ELF targeted ports. For i386 PE, symbols |
| 292 | explicitly listed in a .def file are still exported, regardless of |
| 293 | this option. For ELF targeted ports, symbols affected by this |
| 294 | option will be treated as hidden. |
| 295 | |
| 296 | `--exclude-modules-for-implib MODULE,MODULE,...' |
| 297 | Specifies a list of object files or archive members, from which |
| 298 | symbols should not be automatically exported, but which should be |
| 299 | copied wholesale into the import library being generated during |
| 300 | the link. The module names may be delimited by commas or colons, |
| 301 | and must match exactly the filenames used by `ld' to open the |
| 302 | files; for archive members, this is simply the member name, but |
| 303 | for object files the name listed must include and match precisely |
| 304 | any path used to specify the input file on the linker's |
| 305 | command-line. This option is available only for the i386 PE |
| 306 | targeted port of the linker. Symbols explicitly listed in a .def |
| 307 | file are still exported, regardless of this option. |
| 308 | |
| 309 | `-E' |
| 310 | `--export-dynamic' |
| 311 | `--no-export-dynamic' |
| 312 | When creating a dynamically linked executable, using the `-E' |
| 313 | option or the `--export-dynamic' option causes the linker to add |
| 314 | all symbols to the dynamic symbol table. The dynamic symbol table |
| 315 | is the set of symbols which are visible from dynamic objects at |
| 316 | run time. |
| 317 | |
| 318 | If you do not use either of these options (or use the |
| 319 | `--no-export-dynamic' option to restore the default behavior), the |
| 320 | dynamic symbol table will normally contain only those symbols |
| 321 | which are referenced by some dynamic object mentioned in the link. |
| 322 | |
| 323 | If you use `dlopen' to load a dynamic object which needs to refer |
| 324 | back to the symbols defined by the program, rather than some other |
| 325 | dynamic object, then you will probably need to use this option when |
| 326 | linking the program itself. |
| 327 | |
| 328 | You can also use the dynamic list to control what symbols should |
| 329 | be added to the dynamic symbol table if the output format supports |
| 330 | it. See the description of `--dynamic-list'. |
| 331 | |
| 332 | Note that this option is specific to ELF targeted ports. PE |
| 333 | targets support a similar function to export all symbols from a |
| 334 | DLL or EXE; see the description of `--export-all-symbols' below. |
| 335 | |
| 336 | `-EB' |
| 337 | Link big-endian objects. This affects the default output format. |
| 338 | |
| 339 | `-EL' |
| 340 | Link little-endian objects. This affects the default output |
| 341 | format. |
| 342 | |
| 343 | `-f NAME' |
| 344 | `--auxiliary=NAME' |
| 345 | When creating an ELF shared object, set the internal DT_AUXILIARY |
| 346 | field to the specified name. This tells the dynamic linker that |
| 347 | the symbol table of the shared object should be used as an |
| 348 | auxiliary filter on the symbol table of the shared object NAME. |
| 349 | |
| 350 | If you later link a program against this filter object, then, when |
| 351 | you run the program, the dynamic linker will see the DT_AUXILIARY |
| 352 | field. If the dynamic linker resolves any symbols from the filter |
| 353 | object, it will first check whether there is a definition in the |
| 354 | shared object NAME. If there is one, it will be used instead of |
| 355 | the definition in the filter object. The shared object NAME need |
| 356 | not exist. Thus the shared object NAME may be used to provide an |
| 357 | alternative implementation of certain functions, perhaps for |
| 358 | debugging or for machine specific performance. |
| 359 | |
| 360 | This option may be specified more than once. The DT_AUXILIARY |
| 361 | entries will be created in the order in which they appear on the |
| 362 | command line. |
| 363 | |
| 364 | `-F NAME' |
| 365 | `--filter=NAME' |
| 366 | When creating an ELF shared object, set the internal DT_FILTER |
| 367 | field to the specified name. This tells the dynamic linker that |
| 368 | the symbol table of the shared object which is being created |
| 369 | should be used as a filter on the symbol table of the shared |
| 370 | object NAME. |
| 371 | |
| 372 | If you later link a program against this filter object, then, when |
| 373 | you run the program, the dynamic linker will see the DT_FILTER |
| 374 | field. The dynamic linker will resolve symbols according to the |
| 375 | symbol table of the filter object as usual, but it will actually |
| 376 | link to the definitions found in the shared object NAME. Thus the |
| 377 | filter object can be used to select a subset of the symbols |
| 378 | provided by the object NAME. |
| 379 | |
| 380 | Some older linkers used the `-F' option throughout a compilation |
| 381 | toolchain for specifying object-file format for both input and |
| 382 | output object files. The GNU linker uses other mechanisms for |
| 383 | this purpose: the `-b', `--format', `--oformat' options, the |
| 384 | `TARGET' command in linker scripts, and the `GNUTARGET' |
| 385 | environment variable. The GNU linker will ignore the `-F' option |
| 386 | when not creating an ELF shared object. |
| 387 | |
| 388 | `-fini=NAME' |
| 389 | When creating an ELF executable or shared object, call NAME when |
| 390 | the executable or shared object is unloaded, by setting DT_FINI to |
| 391 | the address of the function. By default, the linker uses `_fini' |
| 392 | as the function to call. |
| 393 | |
| 394 | `-g' |
| 395 | Ignored. Provided for compatibility with other tools. |
| 396 | |
| 397 | `-G VALUE' |
| 398 | `--gpsize=VALUE' |
| 399 | Set the maximum size of objects to be optimized using the GP |
| 400 | register to SIZE. This is only meaningful for object file formats |
| 401 | such as MIPS ELF that support putting large and small objects into |
| 402 | different sections. This is ignored for other object file formats. |
| 403 | |
| 404 | `-h NAME' |
| 405 | `-soname=NAME' |
| 406 | When creating an ELF shared object, set the internal DT_SONAME |
| 407 | field to the specified name. When an executable is linked with a |
| 408 | shared object which has a DT_SONAME field, then when the |
| 409 | executable is run the dynamic linker will attempt to load the |
| 410 | shared object specified by the DT_SONAME field rather than the |
| 411 | using the file name given to the linker. |
| 412 | |
| 413 | `-i' |
| 414 | Perform an incremental link (same as option `-r'). |
| 415 | |
| 416 | `-init=NAME' |
| 417 | When creating an ELF executable or shared object, call NAME when |
| 418 | the executable or shared object is loaded, by setting DT_INIT to |
| 419 | the address of the function. By default, the linker uses `_init' |
| 420 | as the function to call. |
| 421 | |
| 422 | `-l NAMESPEC' |
| 423 | `--library=NAMESPEC' |
| 424 | Add the archive or object file specified by NAMESPEC to the list |
| 425 | of files to link. This option may be used any number of times. |
| 426 | If NAMESPEC is of the form `:FILENAME', `ld' will search the |
| 427 | library path for a file called FILENAME, otherwise it will search |
| 428 | the library path for a file called `libNAMESPEC.a'. |
| 429 | |
| 430 | On systems which support shared libraries, `ld' may also search for |
| 431 | files other than `libNAMESPEC.a'. Specifically, on ELF and SunOS |
| 432 | systems, `ld' will search a directory for a library called |
| 433 | `libNAMESPEC.so' before searching for one called `libNAMESPEC.a'. |
| 434 | (By convention, a `.so' extension indicates a shared library.) |
| 435 | Note that this behavior does not apply to `:FILENAME', which |
| 436 | always specifies a file called FILENAME. |
| 437 | |
| 438 | The linker will search an archive only once, at the location where |
| 439 | it is specified on the command line. If the archive defines a |
| 440 | symbol which was undefined in some object which appeared before |
| 441 | the archive on the command line, the linker will include the |
| 442 | appropriate file(s) from the archive. However, an undefined |
| 443 | symbol in an object appearing later on the command line will not |
| 444 | cause the linker to search the archive again. |
| 445 | |
| 446 | See the `-(' option for a way to force the linker to search |
| 447 | archives multiple times. |
| 448 | |
| 449 | You may list the same archive multiple times on the command line. |
| 450 | |
| 451 | This type of archive searching is standard for Unix linkers. |
| 452 | However, if you are using `ld' on AIX, note that it is different |
| 453 | from the behaviour of the AIX linker. |
| 454 | |
| 455 | `-L SEARCHDIR' |
| 456 | `--library-path=SEARCHDIR' |
| 457 | Add path SEARCHDIR to the list of paths that `ld' will search for |
| 458 | archive libraries and `ld' control scripts. You may use this |
| 459 | option any number of times. The directories are searched in the |
| 460 | order in which they are specified on the command line. |
| 461 | Directories specified on the command line are searched before the |
| 462 | default directories. All `-L' options apply to all `-l' options, |
| 463 | regardless of the order in which the options appear. `-L' options |
| 464 | do not affect how `ld' searches for a linker script unless `-T' |
| 465 | option is specified. |
| 466 | |
| 467 | If SEARCHDIR begins with `=', then the `=' will be replaced by the |
| 468 | "sysroot prefix", controlled by the `--sysroot' option, or |
| 469 | specified when the linker is configured. |
| 470 | |
| 471 | The default set of paths searched (without being specified with |
| 472 | `-L') depends on which emulation mode `ld' is using, and in some |
| 473 | cases also on how it was configured. *Note Environment::. |
| 474 | |
| 475 | The paths can also be specified in a link script with the |
| 476 | `SEARCH_DIR' command. Directories specified this way are searched |
| 477 | at the point in which the linker script appears in the command |
| 478 | line. |
| 479 | |
| 480 | `-m EMULATION' |
| 481 | Emulate the EMULATION linker. You can list the available |
| 482 | emulations with the `--verbose' or `-V' options. |
| 483 | |
| 484 | If the `-m' option is not used, the emulation is taken from the |
| 485 | `LDEMULATION' environment variable, if that is defined. |
| 486 | |
| 487 | Otherwise, the default emulation depends upon how the linker was |
| 488 | configured. |
| 489 | |
| 490 | `-M' |
| 491 | `--print-map' |
| 492 | Print a link map to the standard output. A link map provides |
| 493 | information about the link, including the following: |
| 494 | |
| 495 | * Where object files are mapped into memory. |
| 496 | |
| 497 | * How common symbols are allocated. |
| 498 | |
| 499 | * All archive members included in the link, with a mention of |
| 500 | the symbol which caused the archive member to be brought in. |
| 501 | |
| 502 | * The values assigned to symbols. |
| 503 | |
| 504 | Note - symbols whose values are computed by an expression |
| 505 | which involves a reference to a previous value of the same |
| 506 | symbol may not have correct result displayed in the link map. |
| 507 | This is because the linker discards intermediate results and |
| 508 | only retains the final value of an expression. Under such |
| 509 | circumstances the linker will display the final value |
| 510 | enclosed by square brackets. Thus for example a linker |
| 511 | script containing: |
| 512 | |
| 513 | foo = 1 |
| 514 | foo = foo * 4 |
| 515 | foo = foo + 8 |
| 516 | |
| 517 | will produce the following output in the link map if the `-M' |
| 518 | option is used: |
| 519 | |
| 520 | 0x00000001 foo = 0x1 |
| 521 | [0x0000000c] foo = (foo * 0x4) |
| 522 | [0x0000000c] foo = (foo + 0x8) |
| 523 | |
| 524 | See *Note Expressions:: for more information about |
| 525 | expressions in linker scripts. |
| 526 | |
| 527 | `-n' |
| 528 | `--nmagic' |
| 529 | Turn off page alignment of sections, and disable linking against |
| 530 | shared libraries. If the output format supports Unix style magic |
| 531 | numbers, mark the output as `NMAGIC'. |
| 532 | |
| 533 | `-N' |
| 534 | `--omagic' |
| 535 | Set the text and data sections to be readable and writable. Also, |
| 536 | do not page-align the data segment, and disable linking against |
| 537 | shared libraries. If the output format supports Unix style magic |
| 538 | numbers, mark the output as `OMAGIC'. Note: Although a writable |
| 539 | text section is allowed for PE-COFF targets, it does not conform |
| 540 | to the format specification published by Microsoft. |
| 541 | |
| 542 | `--no-omagic' |
| 543 | This option negates most of the effects of the `-N' option. It |
| 544 | sets the text section to be read-only, and forces the data segment |
| 545 | to be page-aligned. Note - this option does not enable linking |
| 546 | against shared libraries. Use `-Bdynamic' for this. |
| 547 | |
| 548 | `-o OUTPUT' |
| 549 | `--output=OUTPUT' |
| 550 | Use OUTPUT as the name for the program produced by `ld'; if this |
| 551 | option is not specified, the name `a.out' is used by default. The |
| 552 | script command `OUTPUT' can also specify the output file name. |
| 553 | |
| 554 | `-O LEVEL' |
| 555 | If LEVEL is a numeric values greater than zero `ld' optimizes the |
| 556 | output. This might take significantly longer and therefore |
| 557 | probably should only be enabled for the final binary. At the |
| 558 | moment this option only affects ELF shared library generation. |
| 559 | Future releases of the linker may make more use of this option. |
| 560 | Also currently there is no difference in the linker's behaviour |
| 561 | for different non-zero values of this option. Again this may |
| 562 | change with future releases. |
| 563 | |
| 564 | `--push-state' |
| 565 | The `--push-state' allows to preserve the current state of the |
| 566 | flags which govern the input file handling so that they can all be |
| 567 | restored with one corresponding `--pop-state' option. |
| 568 | |
| 569 | The option which are covered are: `-Bdynamic', `-Bstatic', `-dn', |
| 570 | `-dy', `-call_shared', `-non_shared', `-static', `-N', `-n', |
| 571 | `--whole-archive', `--no-whole-archive', `-r', `-Ur', |
| 572 | `--copy-dt-needed-entries', `--no-copy-dt-needed-entries', |
| 573 | `--as-needed', `--no-as-needed', and `-a'. |
| 574 | |
| 575 | One target for this option are specifications for `pkg-config'. |
| 576 | When used with the `--libs' option all possibly needed libraries |
| 577 | are listed and then possibly linked with all the time. It is |
| 578 | better to return something as follows: |
| 579 | |
| 580 | -Wl,--push-state,--as-needed -libone -libtwo -Wl,--pop-state |
| 581 | |
| 582 | Undoes the effect of -push-state, restores the previous values of |
| 583 | the flags governing input file handling. |
| 584 | |
| 585 | `-q' |
| 586 | `--emit-relocs' |
| 587 | Leave relocation sections and contents in fully linked executables. |
| 588 | Post link analysis and optimization tools may need this |
| 589 | information in order to perform correct modifications of |
| 590 | executables. This results in larger executables. |
| 591 | |
| 592 | This option is currently only supported on ELF platforms. |
| 593 | |
| 594 | `--force-dynamic' |
| 595 | Force the output file to have dynamic sections. This option is |
| 596 | specific to VxWorks targets. |
| 597 | |
| 598 | `-r' |
| 599 | `--relocatable' |
| 600 | Generate relocatable output--i.e., generate an output file that |
| 601 | can in turn serve as input to `ld'. This is often called "partial |
| 602 | linking". As a side effect, in environments that support standard |
| 603 | Unix magic numbers, this option also sets the output file's magic |
| 604 | number to `OMAGIC'. If this option is not specified, an absolute |
| 605 | file is produced. When linking C++ programs, this option _will |
| 606 | not_ resolve references to constructors; to do that, use `-Ur'. |
| 607 | |
| 608 | When an input file does not have the same format as the output |
| 609 | file, partial linking is only supported if that input file does |
| 610 | not contain any relocations. Different output formats can have |
| 611 | further restrictions; for example some `a.out'-based formats do |
| 612 | not support partial linking with input files in other formats at |
| 613 | all. |
| 614 | |
| 615 | This option does the same thing as `-i'. |
| 616 | |
| 617 | `-R FILENAME' |
| 618 | `--just-symbols=FILENAME' |
| 619 | Read symbol names and their addresses from FILENAME, but do not |
| 620 | relocate it or include it in the output. This allows your output |
| 621 | file to refer symbolically to absolute locations of memory defined |
| 622 | in other programs. You may use this option more than once. |
| 623 | |
| 624 | For compatibility with other ELF linkers, if the `-R' option is |
| 625 | followed by a directory name, rather than a file name, it is |
| 626 | treated as the `-rpath' option. |
| 627 | |
| 628 | `-s' |
| 629 | `--strip-all' |
| 630 | Omit all symbol information from the output file. |
| 631 | |
| 632 | `-S' |
| 633 | `--strip-debug' |
| 634 | Omit debugger symbol information (but not all symbols) from the |
| 635 | output file. |
| 636 | |
| 637 | `-t' |
| 638 | `--trace' |
| 639 | Print the names of the input files as `ld' processes them. |
| 640 | |
| 641 | `-T SCRIPTFILE' |
| 642 | `--script=SCRIPTFILE' |
| 643 | Use SCRIPTFILE as the linker script. This script replaces `ld''s |
| 644 | default linker script (rather than adding to it), so COMMANDFILE |
| 645 | must specify everything necessary to describe the output file. |
| 646 | *Note Scripts::. If SCRIPTFILE does not exist in the current |
| 647 | directory, `ld' looks for it in the directories specified by any |
| 648 | preceding `-L' options. Multiple `-T' options accumulate. |
| 649 | |
| 650 | `-dT SCRIPTFILE' |
| 651 | `--default-script=SCRIPTFILE' |
| 652 | Use SCRIPTFILE as the default linker script. *Note Scripts::. |
| 653 | |
| 654 | This option is similar to the `--script' option except that |
| 655 | processing of the script is delayed until after the rest of the |
| 656 | command line has been processed. This allows options placed after |
| 657 | the `--default-script' option on the command line to affect the |
| 658 | behaviour of the linker script, which can be important when the |
| 659 | linker command line cannot be directly controlled by the user. |
| 660 | (eg because the command line is being constructed by another tool, |
| 661 | such as `gcc'). |
| 662 | |
| 663 | `-u SYMBOL' |
| 664 | `--undefined=SYMBOL' |
| 665 | Force SYMBOL to be entered in the output file as an undefined |
| 666 | symbol. Doing this may, for example, trigger linking of additional |
| 667 | modules from standard libraries. `-u' may be repeated with |
| 668 | different option arguments to enter additional undefined symbols. |
| 669 | This option is equivalent to the `EXTERN' linker script command. |
| 670 | |
| 671 | If this option is being used to force additional modules to be |
| 672 | pulled into the link, and if it is an error for the symbol to |
| 673 | remain undefined, then the option `--require-defined' should be |
| 674 | used instead. |
| 675 | |
| 676 | `--require-defined=SYMBOL' |
| 677 | Require that SYMBOL is defined in the output file. This option is |
| 678 | the same as option `--undefined' except that if SYMBOL is not |
| 679 | defined in the output file then the linker will issue an error and |
| 680 | exit. The same effect can be achieved in a linker script by using |
| 681 | `EXTERN', `ASSERT' and `DEFINED' together. This option can be |
| 682 | used multiple times to require additional symbols. |
| 683 | |
| 684 | `-Ur' |
| 685 | For anything other than C++ programs, this option is equivalent to |
| 686 | `-r': it generates relocatable output--i.e., an output file that |
| 687 | can in turn serve as input to `ld'. When linking C++ programs, |
| 688 | `-Ur' _does_ resolve references to constructors, unlike `-r'. It |
| 689 | does not work to use `-Ur' on files that were themselves linked |
| 690 | with `-Ur'; once the constructor table has been built, it cannot |
| 691 | be added to. Use `-Ur' only for the last partial link, and `-r' |
| 692 | for the others. |
| 693 | |
| 694 | `--orphan-handling=MODE' |
| 695 | Control how orphan sections are handled. An orphan section is one |
| 696 | not specifically mentioned in a linker script. *Note Orphan |
| 697 | Sections::. |
| 698 | |
| 699 | MODE can have any of the following values: |
| 700 | |
| 701 | `place' |
| 702 | Orphan sections are placed into a suitable output section |
| 703 | following the strategy described in *Note Orphan Sections::. |
| 704 | The option `--unique' also effects how sections are placed. |
| 705 | |
| 706 | `discard' |
| 707 | All orphan sections are discarded, by placing them in the |
| 708 | `/DISCARD/' section (*note Output Section Discarding::). |
| 709 | |
| 710 | `warn' |
| 711 | The linker will place the orphan section as for `place' and |
| 712 | also issue a warning. |
| 713 | |
| 714 | `error' |
| 715 | The linker will exit with an error if any orphan section is |
| 716 | found. |
| 717 | |
| 718 | The default if `--orphan-handling' is not given is `place'. |
| 719 | |
| 720 | `--unique[=SECTION]' |
| 721 | Creates a separate output section for every input section matching |
| 722 | SECTION, or if the optional wildcard SECTION argument is missing, |
| 723 | for every orphan input section. An orphan section is one not |
| 724 | specifically mentioned in a linker script. You may use this option |
| 725 | multiple times on the command line; It prevents the normal |
| 726 | merging of input sections with the same name, overriding output |
| 727 | section assignments in a linker script. |
| 728 | |
| 729 | `-v' |
| 730 | `--version' |
| 731 | `-V' |
| 732 | Display the version number for `ld'. The `-V' option also lists |
| 733 | the supported emulations. |
| 734 | |
| 735 | `-x' |
| 736 | `--discard-all' |
| 737 | Delete all local symbols. |
| 738 | |
| 739 | `-X' |
| 740 | `--discard-locals' |
| 741 | Delete all temporary local symbols. (These symbols start with |
| 742 | system-specific local label prefixes, typically `.L' for ELF |
| 743 | systems or `L' for traditional a.out systems.) |
| 744 | |
| 745 | `-y SYMBOL' |
| 746 | `--trace-symbol=SYMBOL' |
| 747 | Print the name of each linked file in which SYMBOL appears. This |
| 748 | option may be given any number of times. On many systems it is |
| 749 | necessary to prepend an underscore. |
| 750 | |
| 751 | This option is useful when you have an undefined symbol in your |
| 752 | link but don't know where the reference is coming from. |
| 753 | |
| 754 | `-Y PATH' |
| 755 | Add PATH to the default library search path. This option exists |
| 756 | for Solaris compatibility. |
| 757 | |
| 758 | `-z KEYWORD' |
| 759 | The recognized keywords are: |
| 760 | `combreloc' |
| 761 | Combines multiple reloc sections and sorts them to make |
| 762 | dynamic symbol lookup caching possible. |
| 763 | |
| 764 | `common' |
| 765 | Generate common symbols with the STT_COMMON type druing a |
| 766 | relocatable link. |
| 767 | |
| 768 | `defs' |
| 769 | Disallows undefined symbols in object files. Undefined |
| 770 | symbols in shared libraries are still allowed. |
| 771 | |
| 772 | `execstack' |
| 773 | Marks the object as requiring executable stack. |
| 774 | |
| 775 | `global' |
| 776 | This option is only meaningful when building a shared object. |
| 777 | It makes the symbols defined by this shared object available |
| 778 | for symbol resolution of subsequently loaded libraries. |
| 779 | |
| 780 | `initfirst' |
| 781 | This option is only meaningful when building a shared object. |
| 782 | It marks the object so that its runtime initialization will |
| 783 | occur before the runtime initialization of any other objects |
| 784 | brought into the process at the same time. Similarly the |
| 785 | runtime finalization of the object will occur after the |
| 786 | runtime finalization of any other objects. |
| 787 | |
| 788 | `interpose' |
| 789 | Marks the object that its symbol table interposes before all |
| 790 | symbols but the primary executable. |
| 791 | |
| 792 | `lazy' |
| 793 | When generating an executable or shared library, mark it to |
| 794 | tell the dynamic linker to defer function call resolution to |
| 795 | the point when the function is called (lazy binding), rather |
| 796 | than at load time. Lazy binding is the default. |
| 797 | |
| 798 | `loadfltr' |
| 799 | Marks the object that its filters be processed immediately at |
| 800 | runtime. |
| 801 | |
| 802 | `muldefs' |
| 803 | Allows multiple definitions. |
| 804 | |
| 805 | `nocombreloc' |
| 806 | Disables multiple reloc sections combining. |
| 807 | |
| 808 | `nocommon' |
| 809 | Generate common symbols with the STT_OBJECT type druing a |
| 810 | relocatable link. |
| 811 | |
| 812 | `nocopyreloc' |
| 813 | Disable linker generated .dynbss variables used in place of |
| 814 | variables defined in shared libraries. May result in dynamic |
| 815 | text relocations. |
| 816 | |
| 817 | `nodefaultlib' |
| 818 | Marks the object that the search for dependencies of this |
| 819 | object will ignore any default library search paths. |
| 820 | |
| 821 | `nodelete' |
| 822 | Marks the object shouldn't be unloaded at runtime. |
| 823 | |
| 824 | `nodlopen' |
| 825 | Marks the object not available to `dlopen'. |
| 826 | |
| 827 | `nodump' |
| 828 | Marks the object can not be dumped by `dldump'. |
| 829 | |
| 830 | `noexecstack' |
| 831 | Marks the object as not requiring executable stack. |
| 832 | |
| 833 | `text' |
| 834 | Treat DT_TEXTREL in shared object as error. |
| 835 | |
| 836 | `notext' |
| 837 | Don't treat DT_TEXTREL in shared object as error. |
| 838 | |
| 839 | `textoff' |
| 840 | Don't treat DT_TEXTREL in shared object as error. |
| 841 | |
| 842 | `norelro' |
| 843 | Don't create an ELF `PT_GNU_RELRO' segment header in the |
| 844 | object. |
| 845 | |
| 846 | `now' |
| 847 | When generating an executable or shared library, mark it to |
| 848 | tell the dynamic linker to resolve all symbols when the |
| 849 | program is started, or when the shared library is linked to |
| 850 | using dlopen, instead of deferring function call resolution |
| 851 | to the point when the function is first called. |
| 852 | |
| 853 | `origin' |
| 854 | Marks the object may contain $ORIGIN. |
| 855 | |
| 856 | `relro' |
| 857 | Create an ELF `PT_GNU_RELRO' segment header in the object. |
| 858 | |
| 859 | `max-page-size=VALUE' |
| 860 | Set the emulation maximum page size to VALUE. |
| 861 | |
| 862 | `common-page-size=VALUE' |
| 863 | Set the emulation common page size to VALUE. |
| 864 | |
| 865 | `stack-size=VALUE' |
| 866 | Specify a stack size for in an ELF `PT_GNU_STACK' segment. |
| 867 | Specifying zero will override any default non-zero sized |
| 868 | `PT_GNU_STACK' segment creation. |
| 869 | |
| 870 | `bndplt' |
| 871 | Always generate BND prefix in PLT entries. Supported for |
| 872 | Linux/x86_64. |
| 873 | |
| 874 | `noextern-protected-data' |
| 875 | Don't treat protected data symbol as external when building |
| 876 | shared library. This option overrides linker backend |
| 877 | default. It can be used to workaround incorrect relocations |
| 878 | against protected data symbols generated by compiler. |
| 879 | Updates on protected data symbols by another module aren't |
| 880 | visible to the resulting shared library. Supported for i386 |
| 881 | and x86-64. |
| 882 | |
| 883 | `nodynamic-undefined-weak' |
| 884 | Don't treat undefined weak symbols as dynamic when building |
| 885 | executable. This option overrides linker backend default. |
| 886 | It can be used to avoid dynamic relocations against undefined |
| 887 | weak symbols in executable. Supported for i386 and x86-64. |
| 888 | |
| 889 | `noreloc-overflow' |
| 890 | Disable relocation overflow check. This can be used to |
| 891 | disable relocation overflow check if there will be no dynamic |
| 892 | relocation overflow at run-time. Supported for x86_64. |
| 893 | |
| 894 | `call-nop=prefix-addr' |
| 895 | `call-nop=prefix-nop' |
| 896 | `call-nop=suffix-nop' |
| 897 | `call-nop=prefix-BYTE' |
| 898 | `call-nop=suffix-BYTE' |
| 899 | Specify the 1-byte `NOP' padding when transforming indirect |
| 900 | call to a locally defined function, foo, via its GOT slot. |
| 901 | `call-nop=prefix-addr' generates `0x67 call foo'. |
| 902 | `call-nop=prefix-nop' generates `0x90 call foo'. |
| 903 | `call-nop=suffix-nop' generates `call foo 0x90'. |
| 904 | `call-nop=prefix-BYTE' generates `BYTE call foo'. |
| 905 | `call-nop=suffix-BYTE' generates `call foo BYTE'. Supported |
| 906 | for i386 and x86_64. |
| 907 | |
| 908 | |
| 909 | Other keywords are ignored for Solaris compatibility. |
| 910 | |
| 911 | `-( ARCHIVES -)' |
| 912 | `--start-group ARCHIVES --end-group' |
| 913 | The ARCHIVES should be a list of archive files. They may be |
| 914 | either explicit file names, or `-l' options. |
| 915 | |
| 916 | The specified archives are searched repeatedly until no new |
| 917 | undefined references are created. Normally, an archive is |
| 918 | searched only once in the order that it is specified on the |
| 919 | command line. If a symbol in that archive is needed to resolve an |
| 920 | undefined symbol referred to by an object in an archive that |
| 921 | appears later on the command line, the linker would not be able to |
| 922 | resolve that reference. By grouping the archives, they all be |
| 923 | searched repeatedly until all possible references are resolved. |
| 924 | |
| 925 | Using this option has a significant performance cost. It is best |
| 926 | to use it only when there are unavoidable circular references |
| 927 | between two or more archives. |
| 928 | |
| 929 | `--accept-unknown-input-arch' |
| 930 | `--no-accept-unknown-input-arch' |
| 931 | Tells the linker to accept input files whose architecture cannot be |
| 932 | recognised. The assumption is that the user knows what they are |
| 933 | doing and deliberately wants to link in these unknown input files. |
| 934 | This was the default behaviour of the linker, before release |
| 935 | 2.14. The default behaviour from release 2.14 onwards is to |
| 936 | reject such input files, and so the `--accept-unknown-input-arch' |
| 937 | option has been added to restore the old behaviour. |
| 938 | |
| 939 | `--as-needed' |
| 940 | `--no-as-needed' |
| 941 | This option affects ELF DT_NEEDED tags for dynamic libraries |
| 942 | mentioned on the command line after the `--as-needed' option. |
| 943 | Normally the linker will add a DT_NEEDED tag for each dynamic |
| 944 | library mentioned on the command line, regardless of whether the |
| 945 | library is actually needed or not. `--as-needed' causes a |
| 946 | DT_NEEDED tag to only be emitted for a library that _at that point |
| 947 | in the link_ satisfies a non-weak undefined symbol reference from |
| 948 | a regular object file or, if the library is not found in the |
| 949 | DT_NEEDED lists of other needed libraries, a non-weak undefined |
| 950 | symbol reference from another needed dynamic library. Object |
| 951 | files or libraries appearing on the command line _after_ the |
| 952 | library in question do not affect whether the library is seen as |
| 953 | needed. This is similar to the rules for extraction of object |
| 954 | files from archives. `--no-as-needed' restores the default |
| 955 | behaviour. |
| 956 | |
| 957 | `--add-needed' |
| 958 | `--no-add-needed' |
| 959 | These two options have been deprecated because of the similarity of |
| 960 | their names to the `--as-needed' and `--no-as-needed' options. |
| 961 | They have been replaced by `--copy-dt-needed-entries' and |
| 962 | `--no-copy-dt-needed-entries'. |
| 963 | |
| 964 | `-assert KEYWORD' |
| 965 | This option is ignored for SunOS compatibility. |
| 966 | |
| 967 | `-Bdynamic' |
| 968 | `-dy' |
| 969 | `-call_shared' |
| 970 | Link against dynamic libraries. This is only meaningful on |
| 971 | platforms for which shared libraries are supported. This option |
| 972 | is normally the default on such platforms. The different variants |
| 973 | of this option are for compatibility with various systems. You |
| 974 | may use this option multiple times on the command line: it affects |
| 975 | library searching for `-l' options which follow it. |
| 976 | |
| 977 | `-Bgroup' |
| 978 | Set the `DF_1_GROUP' flag in the `DT_FLAGS_1' entry in the dynamic |
| 979 | section. This causes the runtime linker to handle lookups in this |
| 980 | object and its dependencies to be performed only inside the group. |
| 981 | `--unresolved-symbols=report-all' is implied. This option is only |
| 982 | meaningful on ELF platforms which support shared libraries. |
| 983 | |
| 984 | `-Bstatic' |
| 985 | `-dn' |
| 986 | `-non_shared' |
| 987 | `-static' |
| 988 | Do not link against shared libraries. This is only meaningful on |
| 989 | platforms for which shared libraries are supported. The different |
| 990 | variants of this option are for compatibility with various |
| 991 | systems. You may use this option multiple times on the command |
| 992 | line: it affects library searching for `-l' options which follow |
| 993 | it. This option also implies `--unresolved-symbols=report-all'. |
| 994 | This option can be used with `-shared'. Doing so means that a |
| 995 | shared library is being created but that all of the library's |
| 996 | external references must be resolved by pulling in entries from |
| 997 | static libraries. |
| 998 | |
| 999 | `-Bsymbolic' |
| 1000 | When creating a shared library, bind references to global symbols |
| 1001 | to the definition within the shared library, if any. Normally, it |
| 1002 | is possible for a program linked against a shared library to |
| 1003 | override the definition within the shared library. This option |
| 1004 | can also be used with the `--export-dynamic' option, when creating |
| 1005 | a position independent executable, to bind references to global |
| 1006 | symbols to the definition within the executable. This option is |
| 1007 | only meaningful on ELF platforms which support shared libraries |
| 1008 | and position independent executables. |
| 1009 | |
| 1010 | `-Bsymbolic-functions' |
| 1011 | When creating a shared library, bind references to global function |
| 1012 | symbols to the definition within the shared library, if any. This |
| 1013 | option can also be used with the `--export-dynamic' option, when |
| 1014 | creating a position independent executable, to bind references to |
| 1015 | global function symbols to the definition within the executable. |
| 1016 | This option is only meaningful on ELF platforms which support |
| 1017 | shared libraries and position independent executables. |
| 1018 | |
| 1019 | `--dynamic-list=DYNAMIC-LIST-FILE' |
| 1020 | Specify the name of a dynamic list file to the linker. This is |
| 1021 | typically used when creating shared libraries to specify a list of |
| 1022 | global symbols whose references shouldn't be bound to the |
| 1023 | definition within the shared library, or creating dynamically |
| 1024 | linked executables to specify a list of symbols which should be |
| 1025 | added to the symbol table in the executable. This option is only |
| 1026 | meaningful on ELF platforms which support shared libraries. |
| 1027 | |
| 1028 | The format of the dynamic list is the same as the version node |
| 1029 | without scope and node name. See *Note VERSION:: for more |
| 1030 | information. |
| 1031 | |
| 1032 | `--dynamic-list-data' |
| 1033 | Include all global data symbols to the dynamic list. |
| 1034 | |
| 1035 | `--dynamic-list-cpp-new' |
| 1036 | Provide the builtin dynamic list for C++ operator new and delete. |
| 1037 | It is mainly useful for building shared libstdc++. |
| 1038 | |
| 1039 | `--dynamic-list-cpp-typeinfo' |
| 1040 | Provide the builtin dynamic list for C++ runtime type |
| 1041 | identification. |
| 1042 | |
| 1043 | `--check-sections' |
| 1044 | `--no-check-sections' |
| 1045 | Asks the linker _not_ to check section addresses after they have |
| 1046 | been assigned to see if there are any overlaps. Normally the |
| 1047 | linker will perform this check, and if it finds any overlaps it |
| 1048 | will produce suitable error messages. The linker does know about, |
| 1049 | and does make allowances for sections in overlays. The default |
| 1050 | behaviour can be restored by using the command line switch |
| 1051 | `--check-sections'. Section overlap is not usually checked for |
| 1052 | relocatable links. You can force checking in that case by using |
| 1053 | the `--check-sections' option. |
| 1054 | |
| 1055 | `--copy-dt-needed-entries' |
| 1056 | `--no-copy-dt-needed-entries' |
| 1057 | This option affects the treatment of dynamic libraries referred to |
| 1058 | by DT_NEEDED tags _inside_ ELF dynamic libraries mentioned on the |
| 1059 | command line. Normally the linker won't add a DT_NEEDED tag to the |
| 1060 | output binary for each library mentioned in a DT_NEEDED tag in an |
| 1061 | input dynamic library. With `--copy-dt-needed-entries' specified |
| 1062 | on the command line however any dynamic libraries that follow it |
| 1063 | will have their DT_NEEDED entries added. The default behaviour |
| 1064 | can be restored with `--no-copy-dt-needed-entries'. |
| 1065 | |
| 1066 | This option also has an effect on the resolution of symbols in |
| 1067 | dynamic libraries. With `--copy-dt-needed-entries' dynamic |
| 1068 | libraries mentioned on the command line will be recursively |
| 1069 | searched, following their DT_NEEDED tags to other libraries, in |
| 1070 | order to resolve symbols required by the output binary. With the |
| 1071 | default setting however the searching of dynamic libraries that |
| 1072 | follow it will stop with the dynamic library itself. No DT_NEEDED |
| 1073 | links will be traversed to resolve symbols. |
| 1074 | |
| 1075 | `--cref' |
| 1076 | Output a cross reference table. If a linker map file is being |
| 1077 | generated, the cross reference table is printed to the map file. |
| 1078 | Otherwise, it is printed on the standard output. |
| 1079 | |
| 1080 | The format of the table is intentionally simple, so that it may be |
| 1081 | easily processed by a script if necessary. The symbols are |
| 1082 | printed out, sorted by name. For each symbol, a list of file |
| 1083 | names is given. If the symbol is defined, the first file listed |
| 1084 | is the location of the definition. If the symbol is defined as a |
| 1085 | common value then any files where this happens appear next. |
| 1086 | Finally any files that reference the symbol are listed. |
| 1087 | |
| 1088 | `--no-define-common' |
| 1089 | This option inhibits the assignment of addresses to common symbols. |
| 1090 | The script command `INHIBIT_COMMON_ALLOCATION' has the same effect. |
| 1091 | *Note Miscellaneous Commands::. |
| 1092 | |
| 1093 | The `--no-define-common' option allows decoupling the decision to |
| 1094 | assign addresses to Common symbols from the choice of the output |
| 1095 | file type; otherwise a non-Relocatable output type forces |
| 1096 | assigning addresses to Common symbols. Using `--no-define-common' |
| 1097 | allows Common symbols that are referenced from a shared library to |
| 1098 | be assigned addresses only in the main program. This eliminates |
| 1099 | the unused duplicate space in the shared library, and also |
| 1100 | prevents any possible confusion over resolving to the wrong |
| 1101 | duplicate when there are many dynamic modules with specialized |
| 1102 | search paths for runtime symbol resolution. |
| 1103 | |
| 1104 | `--defsym=SYMBOL=EXPRESSION' |
| 1105 | Create a global symbol in the output file, containing the absolute |
| 1106 | address given by EXPRESSION. You may use this option as many |
| 1107 | times as necessary to define multiple symbols in the command line. |
| 1108 | A limited form of arithmetic is supported for the EXPRESSION in |
| 1109 | this context: you may give a hexadecimal constant or the name of |
| 1110 | an existing symbol, or use `+' and `-' to add or subtract |
| 1111 | hexadecimal constants or symbols. If you need more elaborate |
| 1112 | expressions, consider using the linker command language from a |
| 1113 | script (*note Assignments::). _Note:_ there should be no white |
| 1114 | space between SYMBOL, the equals sign ("<=>"), and EXPRESSION. |
| 1115 | |
| 1116 | `--demangle[=STYLE]' |
| 1117 | `--no-demangle' |
| 1118 | These options control whether to demangle symbol names in error |
| 1119 | messages and other output. When the linker is told to demangle, |
| 1120 | it tries to present symbol names in a readable fashion: it strips |
| 1121 | leading underscores if they are used by the object file format, |
| 1122 | and converts C++ mangled symbol names into user readable names. |
| 1123 | Different compilers have different mangling styles. The optional |
| 1124 | demangling style argument can be used to choose an appropriate |
| 1125 | demangling style for your compiler. The linker will demangle by |
| 1126 | default unless the environment variable `COLLECT_NO_DEMANGLE' is |
| 1127 | set. These options may be used to override the default. |
| 1128 | |
| 1129 | `-IFILE' |
| 1130 | `--dynamic-linker=FILE' |
| 1131 | Set the name of the dynamic linker. This is only meaningful when |
| 1132 | generating dynamically linked ELF executables. The default dynamic |
| 1133 | linker is normally correct; don't use this unless you know what |
| 1134 | you are doing. |
| 1135 | |
| 1136 | `--no-dynamic-linker' |
| 1137 | When producing an executable file, omit the request for a dynamic |
| 1138 | linker to be used at load-time. This is only meaningful for ELF |
| 1139 | executables that contain dynamic relocations, and usually requires |
| 1140 | entry point code that is capable of processing these relocations. |
| 1141 | |
| 1142 | `--fatal-warnings' |
| 1143 | `--no-fatal-warnings' |
| 1144 | Treat all warnings as errors. The default behaviour can be |
| 1145 | restored with the option `--no-fatal-warnings'. |
| 1146 | |
| 1147 | `--force-exe-suffix' |
| 1148 | Make sure that an output file has a .exe suffix. |
| 1149 | |
| 1150 | If a successfully built fully linked output file does not have a |
| 1151 | `.exe' or `.dll' suffix, this option forces the linker to copy the |
| 1152 | output file to one of the same name with a `.exe' suffix. This |
| 1153 | option is useful when using unmodified Unix makefiles on a |
| 1154 | Microsoft Windows host, since some versions of Windows won't run |
| 1155 | an image unless it ends in a `.exe' suffix. |
| 1156 | |
| 1157 | `--gc-sections' |
| 1158 | `--no-gc-sections' |
| 1159 | Enable garbage collection of unused input sections. It is ignored |
| 1160 | on targets that do not support this option. The default behaviour |
| 1161 | (of not performing this garbage collection) can be restored by |
| 1162 | specifying `--no-gc-sections' on the command line. Note that |
| 1163 | garbage collection for COFF and PE format targets is supported, |
| 1164 | but the implementation is currently considered to be experimental. |
| 1165 | |
| 1166 | `--gc-sections' decides which input sections are used by examining |
| 1167 | symbols and relocations. The section containing the entry symbol |
| 1168 | and all sections containing symbols undefined on the command-line |
| 1169 | will be kept, as will sections containing symbols referenced by |
| 1170 | dynamic objects. Note that when building shared libraries, the |
| 1171 | linker must assume that any visible symbol is referenced. Once |
| 1172 | this initial set of sections has been determined, the linker |
| 1173 | recursively marks as used any section referenced by their |
| 1174 | relocations. See `--entry' and `--undefined'. |
| 1175 | |
| 1176 | This option can be set when doing a partial link (enabled with |
| 1177 | option `-r'). In this case the root of symbols kept must be |
| 1178 | explicitly specified either by an `--entry' or `--undefined' |
| 1179 | option or by a `ENTRY' command in the linker script. |
| 1180 | |
| 1181 | `--print-gc-sections' |
| 1182 | `--no-print-gc-sections' |
| 1183 | List all sections removed by garbage collection. The listing is |
| 1184 | printed on stderr. This option is only effective if garbage |
| 1185 | collection has been enabled via the `--gc-sections') option. The |
| 1186 | default behaviour (of not listing the sections that are removed) |
| 1187 | can be restored by specifying `--no-print-gc-sections' on the |
| 1188 | command line. |
| 1189 | |
| 1190 | `--print-output-format' |
| 1191 | Print the name of the default output format (perhaps influenced by |
| 1192 | other command-line options). This is the string that would appear |
| 1193 | in an `OUTPUT_FORMAT' linker script command (*note File |
| 1194 | Commands::). |
| 1195 | |
| 1196 | `--print-memory-usage' |
| 1197 | Print used size, total size and used size of memory regions |
| 1198 | created with the *Note MEMORY:: command. This is useful on |
| 1199 | embedded targets to have a quick view of amount of free memory. |
| 1200 | The format of the output has one headline and one line per region. |
| 1201 | It is both human readable and easily parsable by tools. Here is |
| 1202 | an example of an output: |
| 1203 | |
| 1204 | Memory region Used Size Region Size %age Used |
| 1205 | ROM: 256 KB 1 MB 25.00% |
| 1206 | RAM: 32 B 2 GB 0.00% |
| 1207 | |
| 1208 | `--help' |
| 1209 | Print a summary of the command-line options on the standard output |
| 1210 | and exit. |
| 1211 | |
| 1212 | `--target-help' |
| 1213 | Print a summary of all target specific options on the standard |
| 1214 | output and exit. |
| 1215 | |
| 1216 | `-Map=MAPFILE' |
| 1217 | Print a link map to the file MAPFILE. See the description of the |
| 1218 | `-M' option, above. |
| 1219 | |
| 1220 | `--no-keep-memory' |
| 1221 | `ld' normally optimizes for speed over memory usage by caching the |
| 1222 | symbol tables of input files in memory. This option tells `ld' to |
| 1223 | instead optimize for memory usage, by rereading the symbol tables |
| 1224 | as necessary. This may be required if `ld' runs out of memory |
| 1225 | space while linking a large executable. |
| 1226 | |
| 1227 | `--no-undefined' |
| 1228 | `-z defs' |
| 1229 | Report unresolved symbol references from regular object files. |
| 1230 | This is done even if the linker is creating a non-symbolic shared |
| 1231 | library. The switch `--[no-]allow-shlib-undefined' controls the |
| 1232 | behaviour for reporting unresolved references found in shared |
| 1233 | libraries being linked in. |
| 1234 | |
| 1235 | `--allow-multiple-definition' |
| 1236 | `-z muldefs' |
| 1237 | Normally when a symbol is defined multiple times, the linker will |
| 1238 | report a fatal error. These options allow multiple definitions and |
| 1239 | the first definition will be used. |
| 1240 | |
| 1241 | `--allow-shlib-undefined' |
| 1242 | `--no-allow-shlib-undefined' |
| 1243 | Allows or disallows undefined symbols in shared libraries. This |
| 1244 | switch is similar to `--no-undefined' except that it determines |
| 1245 | the behaviour when the undefined symbols are in a shared library |
| 1246 | rather than a regular object file. It does not affect how |
| 1247 | undefined symbols in regular object files are handled. |
| 1248 | |
| 1249 | The default behaviour is to report errors for any undefined symbols |
| 1250 | referenced in shared libraries if the linker is being used to |
| 1251 | create an executable, but to allow them if the linker is being |
| 1252 | used to create a shared library. |
| 1253 | |
| 1254 | The reasons for allowing undefined symbol references in shared |
| 1255 | libraries specified at link time are that: |
| 1256 | |
| 1257 | * A shared library specified at link time may not be the same |
| 1258 | as the one that is available at load time, so the symbol |
| 1259 | might actually be resolvable at load time. |
| 1260 | |
| 1261 | * There are some operating systems, eg BeOS and HPPA, where |
| 1262 | undefined symbols in shared libraries are normal. |
| 1263 | |
| 1264 | The BeOS kernel for example patches shared libraries at load |
| 1265 | time to select whichever function is most appropriate for the |
| 1266 | current architecture. This is used, for example, to |
| 1267 | dynamically select an appropriate memset function. |
| 1268 | |
| 1269 | `--no-undefined-version' |
| 1270 | Normally when a symbol has an undefined version, the linker will |
| 1271 | ignore it. This option disallows symbols with undefined version |
| 1272 | and a fatal error will be issued instead. |
| 1273 | |
| 1274 | `--default-symver' |
| 1275 | Create and use a default symbol version (the soname) for |
| 1276 | unversioned exported symbols. |
| 1277 | |
| 1278 | `--default-imported-symver' |
| 1279 | Create and use a default symbol version (the soname) for |
| 1280 | unversioned imported symbols. |
| 1281 | |
| 1282 | `--no-warn-mismatch' |
| 1283 | Normally `ld' will give an error if you try to link together input |
| 1284 | files that are mismatched for some reason, perhaps because they |
| 1285 | have been compiled for different processors or for different |
| 1286 | endiannesses. This option tells `ld' that it should silently |
| 1287 | permit such possible errors. This option should only be used with |
| 1288 | care, in cases when you have taken some special action that |
| 1289 | ensures that the linker errors are inappropriate. |
| 1290 | |
| 1291 | `--no-warn-search-mismatch' |
| 1292 | Normally `ld' will give a warning if it finds an incompatible |
| 1293 | library during a library search. This option silences the warning. |
| 1294 | |
| 1295 | `--no-whole-archive' |
| 1296 | Turn off the effect of the `--whole-archive' option for subsequent |
| 1297 | archive files. |
| 1298 | |
| 1299 | `--noinhibit-exec' |
| 1300 | Retain the executable output file whenever it is still usable. |
| 1301 | Normally, the linker will not produce an output file if it |
| 1302 | encounters errors during the link process; it exits without |
| 1303 | writing an output file when it issues any error whatsoever. |
| 1304 | |
| 1305 | `-nostdlib' |
| 1306 | Only search library directories explicitly specified on the |
| 1307 | command line. Library directories specified in linker scripts |
| 1308 | (including linker scripts specified on the command line) are |
| 1309 | ignored. |
| 1310 | |
| 1311 | `--oformat=OUTPUT-FORMAT' |
| 1312 | `ld' may be configured to support more than one kind of object |
| 1313 | file. If your `ld' is configured this way, you can use the |
| 1314 | `--oformat' option to specify the binary format for the output |
| 1315 | object file. Even when `ld' is configured to support alternative |
| 1316 | object formats, you don't usually need to specify this, as `ld' |
| 1317 | should be configured to produce as a default output format the most |
| 1318 | usual format on each machine. OUTPUT-FORMAT is a text string, the |
| 1319 | name of a particular format supported by the BFD libraries. (You |
| 1320 | can list the available binary formats with `objdump -i'.) The |
| 1321 | script command `OUTPUT_FORMAT' can also specify the output format, |
| 1322 | but this option overrides it. *Note BFD::. |
| 1323 | |
| 1324 | `-pie' |
| 1325 | `--pic-executable' |
| 1326 | Create a position independent executable. This is currently only |
| 1327 | supported on ELF platforms. Position independent executables are |
| 1328 | similar to shared libraries in that they are relocated by the |
| 1329 | dynamic linker to the virtual address the OS chooses for them |
| 1330 | (which can vary between invocations). Like normal dynamically |
| 1331 | linked executables they can be executed and symbols defined in the |
| 1332 | executable cannot be overridden by shared libraries. |
| 1333 | |
| 1334 | `-qmagic' |
| 1335 | This option is ignored for Linux compatibility. |
| 1336 | |
| 1337 | `-Qy' |
| 1338 | This option is ignored for SVR4 compatibility. |
| 1339 | |
| 1340 | `--relax' |
| 1341 | `--no-relax' |
| 1342 | An option with machine dependent effects. This option is only |
| 1343 | supported on a few targets. *Note `ld' and the H8/300: H8/300. |
| 1344 | *Note `ld' and the Intel 960 family: i960. *Note `ld' and Xtensa |
| 1345 | Processors: Xtensa. *Note `ld' and the 68HC11 and 68HC12: |
| 1346 | M68HC11/68HC12. *Note `ld' and the Altera Nios II: Nios II. |
| 1347 | *Note `ld' and PowerPC 32-bit ELF Support: PowerPC ELF32. |
| 1348 | |
| 1349 | On some platforms the `--relax' option performs target specific, |
| 1350 | global optimizations that become possible when the linker resolves |
| 1351 | addressing in the program, such as relaxing address modes, |
| 1352 | synthesizing new instructions, selecting shorter version of current |
| 1353 | instructions, and combining constant values. |
| 1354 | |
| 1355 | On some platforms these link time global optimizations may make |
| 1356 | symbolic debugging of the resulting executable impossible. This |
| 1357 | is known to be the case for the Matsushita MN10200 and MN10300 |
| 1358 | family of processors. |
| 1359 | |
| 1360 | On platforms where this is not supported, `--relax' is accepted, |
| 1361 | but ignored. |
| 1362 | |
| 1363 | On platforms where `--relax' is accepted the option `--no-relax' |
| 1364 | can be used to disable the feature. |
| 1365 | |
| 1366 | `--retain-symbols-file=FILENAME' |
| 1367 | Retain _only_ the symbols listed in the file FILENAME, discarding |
| 1368 | all others. FILENAME is simply a flat file, with one symbol name |
| 1369 | per line. This option is especially useful in environments (such |
| 1370 | as VxWorks) where a large global symbol table is accumulated |
| 1371 | gradually, to conserve run-time memory. |
| 1372 | |
| 1373 | `--retain-symbols-file' does _not_ discard undefined symbols, or |
| 1374 | symbols needed for relocations. |
| 1375 | |
| 1376 | You may only specify `--retain-symbols-file' once in the command |
| 1377 | line. It overrides `-s' and `-S'. |
| 1378 | |
| 1379 | `-rpath=DIR' |
| 1380 | Add a directory to the runtime library search path. This is used |
| 1381 | when linking an ELF executable with shared objects. All `-rpath' |
| 1382 | arguments are concatenated and passed to the runtime linker, which |
| 1383 | uses them to locate shared objects at runtime. The `-rpath' |
| 1384 | option is also used when locating shared objects which are needed |
| 1385 | by shared objects explicitly included in the link; see the |
| 1386 | description of the `-rpath-link' option. If `-rpath' is not used |
| 1387 | when linking an ELF executable, the contents of the environment |
| 1388 | variable `LD_RUN_PATH' will be used if it is defined. |
| 1389 | |
| 1390 | The `-rpath' option may also be used on SunOS. By default, on |
| 1391 | SunOS, the linker will form a runtime search path out of all the |
| 1392 | `-L' options it is given. If a `-rpath' option is used, the |
| 1393 | runtime search path will be formed exclusively using the `-rpath' |
| 1394 | options, ignoring the `-L' options. This can be useful when using |
| 1395 | gcc, which adds many `-L' options which may be on NFS mounted file |
| 1396 | systems. |
| 1397 | |
| 1398 | For compatibility with other ELF linkers, if the `-R' option is |
| 1399 | followed by a directory name, rather than a file name, it is |
| 1400 | treated as the `-rpath' option. |
| 1401 | |
| 1402 | `-rpath-link=DIR' |
| 1403 | When using ELF or SunOS, one shared library may require another. |
| 1404 | This happens when an `ld -shared' link includes a shared library |
| 1405 | as one of the input files. |
| 1406 | |
| 1407 | When the linker encounters such a dependency when doing a |
| 1408 | non-shared, non-relocatable link, it will automatically try to |
| 1409 | locate the required shared library and include it in the link, if |
| 1410 | it is not included explicitly. In such a case, the `-rpath-link' |
| 1411 | option specifies the first set of directories to search. The |
| 1412 | `-rpath-link' option may specify a sequence of directory names |
| 1413 | either by specifying a list of names separated by colons, or by |
| 1414 | appearing multiple times. |
| 1415 | |
| 1416 | This option should be used with caution as it overrides the search |
| 1417 | path that may have been hard compiled into a shared library. In |
| 1418 | such a case it is possible to use unintentionally a different |
| 1419 | search path than the runtime linker would do. |
| 1420 | |
| 1421 | The linker uses the following search paths to locate required |
| 1422 | shared libraries: |
| 1423 | 1. Any directories specified by `-rpath-link' options. |
| 1424 | |
| 1425 | 2. Any directories specified by `-rpath' options. The difference |
| 1426 | between `-rpath' and `-rpath-link' is that directories |
| 1427 | specified by `-rpath' options are included in the executable |
| 1428 | and used at runtime, whereas the `-rpath-link' option is only |
| 1429 | effective at link time. Searching `-rpath' in this way is |
| 1430 | only supported by native linkers and cross linkers which have |
| 1431 | been configured with the `--with-sysroot' option. |
| 1432 | |
| 1433 | 3. On an ELF system, for native linkers, if the `-rpath' and |
| 1434 | `-rpath-link' options were not used, search the contents of |
| 1435 | the environment variable `LD_RUN_PATH'. |
| 1436 | |
| 1437 | 4. On SunOS, if the `-rpath' option was not used, search any |
| 1438 | directories specified using `-L' options. |
| 1439 | |
| 1440 | 5. For a native linker, search the contents of the environment |
| 1441 | variable `LD_LIBRARY_PATH'. |
| 1442 | |
| 1443 | 6. For a native ELF linker, the directories in `DT_RUNPATH' or |
| 1444 | `DT_RPATH' of a shared library are searched for shared |
| 1445 | libraries needed by it. The `DT_RPATH' entries are ignored if |
| 1446 | `DT_RUNPATH' entries exist. |
| 1447 | |
| 1448 | 7. The default directories, normally `/lib' and `/usr/lib'. |
| 1449 | |
| 1450 | 8. For a native linker on an ELF system, if the file |
| 1451 | `/etc/ld.so.conf' exists, the list of directories found in |
| 1452 | that file. |
| 1453 | |
| 1454 | If the required shared library is not found, the linker will issue |
| 1455 | a warning and continue with the link. |
| 1456 | |
| 1457 | `-shared' |
| 1458 | `-Bshareable' |
| 1459 | Create a shared library. This is currently only supported on ELF, |
| 1460 | XCOFF and SunOS platforms. On SunOS, the linker will |
| 1461 | automatically create a shared library if the `-e' option is not |
| 1462 | used and there are undefined symbols in the link. |
| 1463 | |
| 1464 | `--sort-common' |
| 1465 | `--sort-common=ascending' |
| 1466 | `--sort-common=descending' |
| 1467 | This option tells `ld' to sort the common symbols by alignment in |
| 1468 | ascending or descending order when it places them in the |
| 1469 | appropriate output sections. The symbol alignments considered are |
| 1470 | sixteen-byte or larger, eight-byte, four-byte, two-byte, and |
| 1471 | one-byte. This is to prevent gaps between symbols due to alignment |
| 1472 | constraints. If no sorting order is specified, then descending |
| 1473 | order is assumed. |
| 1474 | |
| 1475 | `--sort-section=name' |
| 1476 | This option will apply `SORT_BY_NAME' to all wildcard section |
| 1477 | patterns in the linker script. |
| 1478 | |
| 1479 | `--sort-section=alignment' |
| 1480 | This option will apply `SORT_BY_ALIGNMENT' to all wildcard section |
| 1481 | patterns in the linker script. |
| 1482 | |
| 1483 | `--split-by-file[=SIZE]' |
| 1484 | Similar to `--split-by-reloc' but creates a new output section for |
| 1485 | each input file when SIZE is reached. SIZE defaults to a size of |
| 1486 | 1 if not given. |
| 1487 | |
| 1488 | `--split-by-reloc[=COUNT]' |
| 1489 | Tries to creates extra sections in the output file so that no |
| 1490 | single output section in the file contains more than COUNT |
| 1491 | relocations. This is useful when generating huge relocatable |
| 1492 | files for downloading into certain real time kernels with the COFF |
| 1493 | object file format; since COFF cannot represent more than 65535 |
| 1494 | relocations in a single section. Note that this will fail to work |
| 1495 | with object file formats which do not support arbitrary sections. |
| 1496 | The linker will not split up individual input sections for |
| 1497 | redistribution, so if a single input section contains more than |
| 1498 | COUNT relocations one output section will contain that many |
| 1499 | relocations. COUNT defaults to a value of 32768. |
| 1500 | |
| 1501 | `--stats' |
| 1502 | Compute and display statistics about the operation of the linker, |
| 1503 | such as execution time and memory usage. |
| 1504 | |
| 1505 | `--sysroot=DIRECTORY' |
| 1506 | Use DIRECTORY as the location of the sysroot, overriding the |
| 1507 | configure-time default. This option is only supported by linkers |
| 1508 | that were configured using `--with-sysroot'. |
| 1509 | |
| 1510 | `--traditional-format' |
| 1511 | For some targets, the output of `ld' is different in some ways from |
| 1512 | the output of some existing linker. This switch requests `ld' to |
| 1513 | use the traditional format instead. |
| 1514 | |
| 1515 | For example, on SunOS, `ld' combines duplicate entries in the |
| 1516 | symbol string table. This can reduce the size of an output file |
| 1517 | with full debugging information by over 30 percent. |
| 1518 | Unfortunately, the SunOS `dbx' program can not read the resulting |
| 1519 | program (`gdb' has no trouble). The `--traditional-format' switch |
| 1520 | tells `ld' to not combine duplicate entries. |
| 1521 | |
| 1522 | `--section-start=SECTIONNAME=ORG' |
| 1523 | Locate a section in the output file at the absolute address given |
| 1524 | by ORG. You may use this option as many times as necessary to |
| 1525 | locate multiple sections in the command line. ORG must be a |
| 1526 | single hexadecimal integer; for compatibility with other linkers, |
| 1527 | you may omit the leading `0x' usually associated with hexadecimal |
| 1528 | values. _Note:_ there should be no white space between |
| 1529 | SECTIONNAME, the equals sign ("<=>"), and ORG. |
| 1530 | |
| 1531 | `-Tbss=ORG' |
| 1532 | `-Tdata=ORG' |
| 1533 | `-Ttext=ORG' |
| 1534 | Same as `--section-start', with `.bss', `.data' or `.text' as the |
| 1535 | SECTIONNAME. |
| 1536 | |
| 1537 | `-Ttext-segment=ORG' |
| 1538 | When creating an ELF executable, it will set the address of the |
| 1539 | first byte of the text segment. |
| 1540 | |
| 1541 | `-Trodata-segment=ORG' |
| 1542 | When creating an ELF executable or shared object for a target where |
| 1543 | the read-only data is in its own segment separate from the |
| 1544 | executable text, it will set the address of the first byte of the |
| 1545 | read-only data segment. |
| 1546 | |
| 1547 | `-Tldata-segment=ORG' |
| 1548 | When creating an ELF executable or shared object for x86-64 medium |
| 1549 | memory model, it will set the address of the first byte of the |
| 1550 | ldata segment. |
| 1551 | |
| 1552 | `--unresolved-symbols=METHOD' |
| 1553 | Determine how to handle unresolved symbols. There are four |
| 1554 | possible values for `method': |
| 1555 | |
| 1556 | `ignore-all' |
| 1557 | Do not report any unresolved symbols. |
| 1558 | |
| 1559 | `report-all' |
| 1560 | Report all unresolved symbols. This is the default. |
| 1561 | |
| 1562 | `ignore-in-object-files' |
| 1563 | Report unresolved symbols that are contained in shared |
| 1564 | libraries, but ignore them if they come from regular object |
| 1565 | files. |
| 1566 | |
| 1567 | `ignore-in-shared-libs' |
| 1568 | Report unresolved symbols that come from regular object |
| 1569 | files, but ignore them if they come from shared libraries. |
| 1570 | This can be useful when creating a dynamic binary and it is |
| 1571 | known that all the shared libraries that it should be |
| 1572 | referencing are included on the linker's command line. |
| 1573 | |
| 1574 | The behaviour for shared libraries on their own can also be |
| 1575 | controlled by the `--[no-]allow-shlib-undefined' option. |
| 1576 | |
| 1577 | Normally the linker will generate an error message for each |
| 1578 | reported unresolved symbol but the option |
| 1579 | `--warn-unresolved-symbols' can change this to a warning. |
| 1580 | |
| 1581 | `--dll-verbose' |
| 1582 | `--verbose[=NUMBER]' |
| 1583 | Display the version number for `ld' and list the linker emulations |
| 1584 | supported. Display which input files can and cannot be opened. |
| 1585 | Display the linker script being used by the linker. If the |
| 1586 | optional NUMBER argument > 1, plugin symbol status will also be |
| 1587 | displayed. |
| 1588 | |
| 1589 | `--version-script=VERSION-SCRIPTFILE' |
| 1590 | Specify the name of a version script to the linker. This is |
| 1591 | typically used when creating shared libraries to specify |
| 1592 | additional information about the version hierarchy for the library |
| 1593 | being created. This option is only fully supported on ELF |
| 1594 | platforms which support shared libraries; see *Note VERSION::. It |
| 1595 | is partially supported on PE platforms, which can use version |
| 1596 | scripts to filter symbol visibility in auto-export mode: any |
| 1597 | symbols marked `local' in the version script will not be exported. |
| 1598 | *Note WIN32::. |
| 1599 | |
| 1600 | `--warn-common' |
| 1601 | Warn when a common symbol is combined with another common symbol |
| 1602 | or with a symbol definition. Unix linkers allow this somewhat |
| 1603 | sloppy practice, but linkers on some other operating systems do |
| 1604 | not. This option allows you to find potential problems from |
| 1605 | combining global symbols. Unfortunately, some C libraries use |
| 1606 | this practice, so you may get some warnings about symbols in the |
| 1607 | libraries as well as in your programs. |
| 1608 | |
| 1609 | There are three kinds of global symbols, illustrated here by C |
| 1610 | examples: |
| 1611 | |
| 1612 | `int i = 1;' |
| 1613 | A definition, which goes in the initialized data section of |
| 1614 | the output file. |
| 1615 | |
| 1616 | `extern int i;' |
| 1617 | An undefined reference, which does not allocate space. There |
| 1618 | must be either a definition or a common symbol for the |
| 1619 | variable somewhere. |
| 1620 | |
| 1621 | `int i;' |
| 1622 | A common symbol. If there are only (one or more) common |
| 1623 | symbols for a variable, it goes in the uninitialized data |
| 1624 | area of the output file. The linker merges multiple common |
| 1625 | symbols for the same variable into a single symbol. If they |
| 1626 | are of different sizes, it picks the largest size. The |
| 1627 | linker turns a common symbol into a declaration, if there is |
| 1628 | a definition of the same variable. |
| 1629 | |
| 1630 | The `--warn-common' option can produce five kinds of warnings. |
| 1631 | Each warning consists of a pair of lines: the first describes the |
| 1632 | symbol just encountered, and the second describes the previous |
| 1633 | symbol encountered with the same name. One or both of the two |
| 1634 | symbols will be a common symbol. |
| 1635 | |
| 1636 | 1. Turning a common symbol into a reference, because there is |
| 1637 | already a definition for the symbol. |
| 1638 | FILE(SECTION): warning: common of `SYMBOL' |
| 1639 | overridden by definition |
| 1640 | FILE(SECTION): warning: defined here |
| 1641 | |
| 1642 | 2. Turning a common symbol into a reference, because a later |
| 1643 | definition for the symbol is encountered. This is the same |
| 1644 | as the previous case, except that the symbols are encountered |
| 1645 | in a different order. |
| 1646 | FILE(SECTION): warning: definition of `SYMBOL' |
| 1647 | overriding common |
| 1648 | FILE(SECTION): warning: common is here |
| 1649 | |
| 1650 | 3. Merging a common symbol with a previous same-sized common |
| 1651 | symbol. |
| 1652 | FILE(SECTION): warning: multiple common |
| 1653 | of `SYMBOL' |
| 1654 | FILE(SECTION): warning: previous common is here |
| 1655 | |
| 1656 | 4. Merging a common symbol with a previous larger common symbol. |
| 1657 | FILE(SECTION): warning: common of `SYMBOL' |
| 1658 | overridden by larger common |
| 1659 | FILE(SECTION): warning: larger common is here |
| 1660 | |
| 1661 | 5. Merging a common symbol with a previous smaller common |
| 1662 | symbol. This is the same as the previous case, except that |
| 1663 | the symbols are encountered in a different order. |
| 1664 | FILE(SECTION): warning: common of `SYMBOL' |
| 1665 | overriding smaller common |
| 1666 | FILE(SECTION): warning: smaller common is here |
| 1667 | |
| 1668 | `--warn-constructors' |
| 1669 | Warn if any global constructors are used. This is only useful for |
| 1670 | a few object file formats. For formats like COFF or ELF, the |
| 1671 | linker can not detect the use of global constructors. |
| 1672 | |
| 1673 | `--warn-multiple-gp' |
| 1674 | Warn if multiple global pointer values are required in the output |
| 1675 | file. This is only meaningful for certain processors, such as the |
| 1676 | Alpha. Specifically, some processors put large-valued constants |
| 1677 | in a special section. A special register (the global pointer) |
| 1678 | points into the middle of this section, so that constants can be |
| 1679 | loaded efficiently via a base-register relative addressing mode. |
| 1680 | Since the offset in base-register relative mode is fixed and |
| 1681 | relatively small (e.g., 16 bits), this limits the maximum size of |
| 1682 | the constant pool. Thus, in large programs, it is often necessary |
| 1683 | to use multiple global pointer values in order to be able to |
| 1684 | address all possible constants. This option causes a warning to |
| 1685 | be issued whenever this case occurs. |
| 1686 | |
| 1687 | `--warn-once' |
| 1688 | Only warn once for each undefined symbol, rather than once per |
| 1689 | module which refers to it. |
| 1690 | |
| 1691 | `--warn-section-align' |
| 1692 | Warn if the address of an output section is changed because of |
| 1693 | alignment. Typically, the alignment will be set by an input |
| 1694 | section. The address will only be changed if it not explicitly |
| 1695 | specified; that is, if the `SECTIONS' command does not specify a |
| 1696 | start address for the section (*note SECTIONS::). |
| 1697 | |
| 1698 | `--warn-shared-textrel' |
| 1699 | Warn if the linker adds a DT_TEXTREL to a shared object. |
| 1700 | |
| 1701 | `--warn-alternate-em' |
| 1702 | Warn if an object has alternate ELF machine code. |
| 1703 | |
| 1704 | `--warn-unresolved-symbols' |
| 1705 | If the linker is going to report an unresolved symbol (see the |
| 1706 | option `--unresolved-symbols') it will normally generate an error. |
| 1707 | This option makes it generate a warning instead. |
| 1708 | |
| 1709 | `--error-unresolved-symbols' |
| 1710 | This restores the linker's default behaviour of generating errors |
| 1711 | when it is reporting unresolved symbols. |
| 1712 | |
| 1713 | `--whole-archive' |
| 1714 | For each archive mentioned on the command line after the |
| 1715 | `--whole-archive' option, include every object file in the archive |
| 1716 | in the link, rather than searching the archive for the required |
| 1717 | object files. This is normally used to turn an archive file into |
| 1718 | a shared library, forcing every object to be included in the |
| 1719 | resulting shared library. This option may be used more than once. |
| 1720 | |
| 1721 | Two notes when using this option from gcc: First, gcc doesn't know |
| 1722 | about this option, so you have to use `-Wl,-whole-archive'. |
| 1723 | Second, don't forget to use `-Wl,-no-whole-archive' after your |
| 1724 | list of archives, because gcc will add its own list of archives to |
| 1725 | your link and you may not want this flag to affect those as well. |
| 1726 | |
| 1727 | `--wrap=SYMBOL' |
| 1728 | Use a wrapper function for SYMBOL. Any undefined reference to |
| 1729 | SYMBOL will be resolved to `__wrap_SYMBOL'. Any undefined |
| 1730 | reference to `__real_SYMBOL' will be resolved to SYMBOL. |
| 1731 | |
| 1732 | This can be used to provide a wrapper for a system function. The |
| 1733 | wrapper function should be called `__wrap_SYMBOL'. If it wishes |
| 1734 | to call the system function, it should call `__real_SYMBOL'. |
| 1735 | |
| 1736 | Here is a trivial example: |
| 1737 | |
| 1738 | void * |
| 1739 | __wrap_malloc (size_t c) |
| 1740 | { |
| 1741 | printf ("malloc called with %zu\n", c); |
| 1742 | return __real_malloc (c); |
| 1743 | } |
| 1744 | |
| 1745 | If you link other code with this file using `--wrap malloc', then |
| 1746 | all calls to `malloc' will call the function `__wrap_malloc' |
| 1747 | instead. The call to `__real_malloc' in `__wrap_malloc' will call |
| 1748 | the real `malloc' function. |
| 1749 | |
| 1750 | You may wish to provide a `__real_malloc' function as well, so that |
| 1751 | links without the `--wrap' option will succeed. If you do this, |
| 1752 | you should not put the definition of `__real_malloc' in the same |
| 1753 | file as `__wrap_malloc'; if you do, the assembler may resolve the |
| 1754 | call before the linker has a chance to wrap it to `malloc'. |
| 1755 | |
| 1756 | `--eh-frame-hdr' |
| 1757 | Request creation of `.eh_frame_hdr' section and ELF |
| 1758 | `PT_GNU_EH_FRAME' segment header. |
| 1759 | |
| 1760 | `--no-ld-generated-unwind-info' |
| 1761 | Request creation of `.eh_frame' unwind info for linker generated |
| 1762 | code sections like PLT. This option is on by default if linker |
| 1763 | generated unwind info is supported. |
| 1764 | |
| 1765 | `--enable-new-dtags' |
| 1766 | `--disable-new-dtags' |
| 1767 | This linker can create the new dynamic tags in ELF. But the older |
| 1768 | ELF systems may not understand them. If you specify |
| 1769 | `--enable-new-dtags', the new dynamic tags will be created as |
| 1770 | needed and older dynamic tags will be omitted. If you specify |
| 1771 | `--disable-new-dtags', no new dynamic tags will be created. By |
| 1772 | default, the new dynamic tags are not created. Note that those |
| 1773 | options are only available for ELF systems. |
| 1774 | |
| 1775 | `--hash-size=NUMBER' |
| 1776 | Set the default size of the linker's hash tables to a prime number |
| 1777 | close to NUMBER. Increasing this value can reduce the length of |
| 1778 | time it takes the linker to perform its tasks, at the expense of |
| 1779 | increasing the linker's memory requirements. Similarly reducing |
| 1780 | this value can reduce the memory requirements at the expense of |
| 1781 | speed. |
| 1782 | |
| 1783 | `--hash-style=STYLE' |
| 1784 | Set the type of linker's hash table(s). STYLE can be either |
| 1785 | `sysv' for classic ELF `.hash' section, `gnu' for new style GNU |
| 1786 | `.gnu.hash' section or `both' for both the classic ELF `.hash' and |
| 1787 | new style GNU `.gnu.hash' hash tables. The default is `sysv'. |
| 1788 | |
| 1789 | `--compress-debug-sections=none' |
| 1790 | `--compress-debug-sections=zlib' |
| 1791 | `--compress-debug-sections=zlib-gnu' |
| 1792 | `--compress-debug-sections=zlib-gabi' |
| 1793 | On ELF platforms , these options control how DWARF debug sections |
| 1794 | are compressed using zlib. `--compress-debug-sections=none' |
| 1795 | doesn't compress DWARF debug sections. |
| 1796 | `--compress-debug-sections=zlib-gnu' compresses DWARF debug |
| 1797 | sections and rename debug section names to begin with `.zdebug' |
| 1798 | instead of `.debug'. `--compress-debug-sections=zlib' and |
| 1799 | `--compress-debug-sections=zlib-gabi' compress DWARF debug |
| 1800 | sections with SHF_COMPRESSED from the ELF ABI. The default |
| 1801 | behaviour varies depending upon the target involved and the |
| 1802 | configure options used to build the toolchain. The default can be |
| 1803 | determined by examing the output from the linker's `--help' option. |
| 1804 | |
| 1805 | `--reduce-memory-overheads' |
| 1806 | This option reduces memory requirements at ld runtime, at the |
| 1807 | expense of linking speed. This was introduced to select the old |
| 1808 | O(n^2) algorithm for link map file generation, rather than the new |
| 1809 | O(n) algorithm which uses about 40% more memory for symbol storage. |
| 1810 | |
| 1811 | Another effect of the switch is to set the default hash table size |
| 1812 | to 1021, which again saves memory at the cost of lengthening the |
| 1813 | linker's run time. This is not done however if the `--hash-size' |
| 1814 | switch has been used. |
| 1815 | |
| 1816 | The `--reduce-memory-overheads' switch may be also be used to |
| 1817 | enable other tradeoffs in future versions of the linker. |
| 1818 | |
| 1819 | `--build-id' |
| 1820 | `--build-id=STYLE' |
| 1821 | Request the creation of a `.note.gnu.build-id' ELF note section or |
| 1822 | a `.buildid' COFF section. The contents of the note are unique |
| 1823 | bits identifying this linked file. STYLE can be `uuid' to use 128 |
| 1824 | random bits, `sha1' to use a 160-bit SHA1 hash on the normative |
| 1825 | parts of the output contents, `md5' to use a 128-bit MD5 hash on |
| 1826 | the normative parts of the output contents, or `0xHEXSTRING' to |
| 1827 | use a chosen bit string specified as an even number of hexadecimal |
| 1828 | digits (`-' and `:' characters between digit pairs are ignored). |
| 1829 | If STYLE is omitted, `sha1' is used. |
| 1830 | |
| 1831 | The `md5' and `sha1' styles produces an identifier that is always |
| 1832 | the same in an identical output file, but will be unique among all |
| 1833 | nonidentical output files. It is not intended to be compared as a |
| 1834 | checksum for the file's contents. A linked file may be changed |
| 1835 | later by other tools, but the build ID bit string identifying the |
| 1836 | original linked file does not change. |
| 1837 | |
| 1838 | Passing `none' for STYLE disables the setting from any |
| 1839 | `--build-id' options earlier on the command line. |
| 1840 | |
| 1841 | `--warn-poison-system-directories' |
| 1842 | Warn for `-L' options using system directories such as `/usr/lib' |
| 1843 | when cross linking. This option is intended for use in |
| 1844 | environments that want to detect and reject incorrect link |
| 1845 | settings. |
| 1846 | |
| 1847 | `--no-warn-poison-system-directories' |
| 1848 | Do not warn for `-L' options using system directories such as |
| 1849 | `/usr/lib' when cross linking. This option is intended for use in |
| 1850 | chroot environments when such directories contain the correct |
| 1851 | libraries for the target system rather than the host. |
| 1852 | |
| 1853 | `--error-poison-system-directories' |
| 1854 | Give an error instead of a warning for `-L' options using system |
| 1855 | directories when cross linking. |
| 1856 | |
| 1857 | 2.1.1 Options Specific to i386 PE Targets |
| 1858 | ----------------------------------------- |
| 1859 | |
| 1860 | The i386 PE linker supports the `-shared' option, which causes the |
| 1861 | output to be a dynamically linked library (DLL) instead of a normal |
| 1862 | executable. You should name the output `*.dll' when you use this |
| 1863 | option. In addition, the linker fully supports the standard `*.def' |
| 1864 | files, which may be specified on the linker command line like an object |
| 1865 | file (in fact, it should precede archives it exports symbols from, to |
| 1866 | ensure that they get linked in, just like a normal object file). |
| 1867 | |
| 1868 | In addition to the options common to all targets, the i386 PE linker |
| 1869 | support additional command line options that are specific to the i386 |
| 1870 | PE target. Options that take values may be separated from their values |
| 1871 | by either a space or an equals sign. |
| 1872 | |
| 1873 | `--add-stdcall-alias' |
| 1874 | If given, symbols with a stdcall suffix (@NN) will be exported |
| 1875 | as-is and also with the suffix stripped. [This option is specific |
| 1876 | to the i386 PE targeted port of the linker] |
| 1877 | |
| 1878 | `--base-file FILE' |
| 1879 | Use FILE as the name of a file in which to save the base addresses |
| 1880 | of all the relocations needed for generating DLLs with `dlltool'. |
| 1881 | [This is an i386 PE specific option] |
| 1882 | |
| 1883 | `--dll' |
| 1884 | Create a DLL instead of a regular executable. You may also use |
| 1885 | `-shared' or specify a `LIBRARY' in a given `.def' file. [This |
| 1886 | option is specific to the i386 PE targeted port of the linker] |
| 1887 | |
| 1888 | `--enable-long-section-names' |
| 1889 | `--disable-long-section-names' |
| 1890 | The PE variants of the COFF object format add an extension that |
| 1891 | permits the use of section names longer than eight characters, the |
| 1892 | normal limit for COFF. By default, these names are only allowed |
| 1893 | in object files, as fully-linked executable images do not carry |
| 1894 | the COFF string table required to support the longer names. As a |
| 1895 | GNU extension, it is possible to allow their use in executable |
| 1896 | images as well, or to (probably pointlessly!) disallow it in |
| 1897 | object files, by using these two options. Executable images |
| 1898 | generated with these long section names are slightly non-standard, |
| 1899 | carrying as they do a string table, and may generate confusing |
| 1900 | output when examined with non-GNU PE-aware tools, such as file |
| 1901 | viewers and dumpers. However, GDB relies on the use of PE long |
| 1902 | section names to find Dwarf-2 debug information sections in an |
| 1903 | executable image at runtime, and so if neither option is specified |
| 1904 | on the command-line, `ld' will enable long section names, |
| 1905 | overriding the default and technically correct behaviour, when it |
| 1906 | finds the presence of debug information while linking an executable |
| 1907 | image and not stripping symbols. [This option is valid for all PE |
| 1908 | targeted ports of the linker] |
| 1909 | |
| 1910 | `--enable-stdcall-fixup' |
| 1911 | `--disable-stdcall-fixup' |
| 1912 | If the link finds a symbol that it cannot resolve, it will attempt |
| 1913 | to do "fuzzy linking" by looking for another defined symbol that |
| 1914 | differs only in the format of the symbol name (cdecl vs stdcall) |
| 1915 | and will resolve that symbol by linking to the match. For |
| 1916 | example, the undefined symbol `_foo' might be linked to the |
| 1917 | function `_foo@12', or the undefined symbol `_bar@16' might be |
| 1918 | linked to the function `_bar'. When the linker does this, it |
| 1919 | prints a warning, since it normally should have failed to link, |
| 1920 | but sometimes import libraries generated from third-party dlls may |
| 1921 | need this feature to be usable. If you specify |
| 1922 | `--enable-stdcall-fixup', this feature is fully enabled and |
| 1923 | warnings are not printed. If you specify |
| 1924 | `--disable-stdcall-fixup', this feature is disabled and such |
| 1925 | mismatches are considered to be errors. [This option is specific |
| 1926 | to the i386 PE targeted port of the linker] |
| 1927 | |
| 1928 | `--leading-underscore' |
| 1929 | `--no-leading-underscore' |
| 1930 | For most targets default symbol-prefix is an underscore and is |
| 1931 | defined in target's description. By this option it is possible to |
| 1932 | disable/enable the default underscore symbol-prefix. |
| 1933 | |
| 1934 | `--export-all-symbols' |
| 1935 | If given, all global symbols in the objects used to build a DLL |
| 1936 | will be exported by the DLL. Note that this is the default if |
| 1937 | there otherwise wouldn't be any exported symbols. When symbols are |
| 1938 | explicitly exported via DEF files or implicitly exported via |
| 1939 | function attributes, the default is to not export anything else |
| 1940 | unless this option is given. Note that the symbols `DllMain@12', |
| 1941 | `DllEntryPoint@0', `DllMainCRTStartup@12', and `impure_ptr' will |
| 1942 | not be automatically exported. Also, symbols imported from other |
| 1943 | DLLs will not be re-exported, nor will symbols specifying the |
| 1944 | DLL's internal layout such as those beginning with `_head_' or |
| 1945 | ending with `_iname'. In addition, no symbols from `libgcc', |
| 1946 | `libstd++', `libmingw32', or `crtX.o' will be exported. Symbols |
| 1947 | whose names begin with `__rtti_' or `__builtin_' will not be |
| 1948 | exported, to help with C++ DLLs. Finally, there is an extensive |
| 1949 | list of cygwin-private symbols that are not exported (obviously, |
| 1950 | this applies on when building DLLs for cygwin targets). These |
| 1951 | cygwin-excludes are: `_cygwin_dll_entry@12', |
| 1952 | `_cygwin_crt0_common@8', `_cygwin_noncygwin_dll_entry@12', |
| 1953 | `_fmode', `_impure_ptr', `cygwin_attach_dll', `cygwin_premain0', |
| 1954 | `cygwin_premain1', `cygwin_premain2', `cygwin_premain3', and |
| 1955 | `environ'. [This option is specific to the i386 PE targeted port |
| 1956 | of the linker] |
| 1957 | |
| 1958 | `--exclude-symbols SYMBOL,SYMBOL,...' |
| 1959 | Specifies a list of symbols which should not be automatically |
| 1960 | exported. The symbol names may be delimited by commas or colons. |
| 1961 | [This option is specific to the i386 PE targeted port of the |
| 1962 | linker] |
| 1963 | |
| 1964 | `--exclude-all-symbols' |
| 1965 | Specifies no symbols should be automatically exported. [This |
| 1966 | option is specific to the i386 PE targeted port of the linker] |
| 1967 | |
| 1968 | `--file-alignment' |
| 1969 | Specify the file alignment. Sections in the file will always |
| 1970 | begin at file offsets which are multiples of this number. This |
| 1971 | defaults to 512. [This option is specific to the i386 PE targeted |
| 1972 | port of the linker] |
| 1973 | |
| 1974 | `--heap RESERVE' |
| 1975 | `--heap RESERVE,COMMIT' |
| 1976 | Specify the number of bytes of memory to reserve (and optionally |
| 1977 | commit) to be used as heap for this program. The default is 1MB |
| 1978 | reserved, 4K committed. [This option is specific to the i386 PE |
| 1979 | targeted port of the linker] |
| 1980 | |
| 1981 | `--image-base VALUE' |
| 1982 | Use VALUE as the base address of your program or dll. This is the |
| 1983 | lowest memory location that will be used when your program or dll |
| 1984 | is loaded. To reduce the need to relocate and improve performance |
| 1985 | of your dlls, each should have a unique base address and not |
| 1986 | overlap any other dlls. The default is 0x400000 for executables, |
| 1987 | and 0x10000000 for dlls. [This option is specific to the i386 PE |
| 1988 | targeted port of the linker] |
| 1989 | |
| 1990 | `--kill-at' |
| 1991 | If given, the stdcall suffixes (@NN) will be stripped from symbols |
| 1992 | before they are exported. [This option is specific to the i386 PE |
| 1993 | targeted port of the linker] |
| 1994 | |
| 1995 | `--large-address-aware' |
| 1996 | If given, the appropriate bit in the "Characteristics" field of |
| 1997 | the COFF header is set to indicate that this executable supports |
| 1998 | virtual addresses greater than 2 gigabytes. This should be used |
| 1999 | in conjunction with the /3GB or /USERVA=VALUE megabytes switch in |
| 2000 | the "[operating systems]" section of the BOOT.INI. Otherwise, |
| 2001 | this bit has no effect. [This option is specific to PE targeted |
| 2002 | ports of the linker] |
| 2003 | |
| 2004 | `--disable-large-address-aware' |
| 2005 | Reverts the effect of a previous `--large-address-aware' option. |
| 2006 | This is useful if `--large-address-aware' is always set by the |
| 2007 | compiler driver (e.g. Cygwin gcc) and the executable does not |
| 2008 | support virtual addresses greater than 2 gigabytes. [This option |
| 2009 | is specific to PE targeted ports of the linker] |
| 2010 | |
| 2011 | `--major-image-version VALUE' |
| 2012 | Sets the major number of the "image version". Defaults to 1. |
| 2013 | [This option is specific to the i386 PE targeted port of the |
| 2014 | linker] |
| 2015 | |
| 2016 | `--major-os-version VALUE' |
| 2017 | Sets the major number of the "os version". Defaults to 4. [This |
| 2018 | option is specific to the i386 PE targeted port of the linker] |
| 2019 | |
| 2020 | `--major-subsystem-version VALUE' |
| 2021 | Sets the major number of the "subsystem version". Defaults to 4. |
| 2022 | [This option is specific to the i386 PE targeted port of the |
| 2023 | linker] |
| 2024 | |
| 2025 | `--minor-image-version VALUE' |
| 2026 | Sets the minor number of the "image version". Defaults to 0. |
| 2027 | [This option is specific to the i386 PE targeted port of the |
| 2028 | linker] |
| 2029 | |
| 2030 | `--minor-os-version VALUE' |
| 2031 | Sets the minor number of the "os version". Defaults to 0. [This |
| 2032 | option is specific to the i386 PE targeted port of the linker] |
| 2033 | |
| 2034 | `--minor-subsystem-version VALUE' |
| 2035 | Sets the minor number of the "subsystem version". Defaults to 0. |
| 2036 | [This option is specific to the i386 PE targeted port of the |
| 2037 | linker] |
| 2038 | |
| 2039 | `--output-def FILE' |
| 2040 | The linker will create the file FILE which will contain a DEF file |
| 2041 | corresponding to the DLL the linker is generating. This DEF file |
| 2042 | (which should be called `*.def') may be used to create an import |
| 2043 | library with `dlltool' or may be used as a reference to |
| 2044 | automatically or implicitly exported symbols. [This option is |
| 2045 | specific to the i386 PE targeted port of the linker] |
| 2046 | |
| 2047 | `--out-implib FILE' |
| 2048 | The linker will create the file FILE which will contain an import |
| 2049 | lib corresponding to the DLL the linker is generating. This import |
| 2050 | lib (which should be called `*.dll.a' or `*.a' may be used to link |
| 2051 | clients against the generated DLL; this behaviour makes it |
| 2052 | possible to skip a separate `dlltool' import library creation step. |
| 2053 | [This option is specific to the i386 PE targeted port of the |
| 2054 | linker] |
| 2055 | |
| 2056 | `--enable-auto-image-base' |
| 2057 | `--enable-auto-image-base=VALUE' |
| 2058 | Automatically choose the image base for DLLs, optionally starting |
| 2059 | with base VALUE, unless one is specified using the `--image-base' |
| 2060 | argument. By using a hash generated from the dllname to create |
| 2061 | unique image bases for each DLL, in-memory collisions and |
| 2062 | relocations which can delay program execution are avoided. [This |
| 2063 | option is specific to the i386 PE targeted port of the linker] |
| 2064 | |
| 2065 | `--disable-auto-image-base' |
| 2066 | Do not automatically generate a unique image base. If there is no |
| 2067 | user-specified image base (`--image-base') then use the platform |
| 2068 | default. [This option is specific to the i386 PE targeted port of |
| 2069 | the linker] |
| 2070 | |
| 2071 | `--dll-search-prefix STRING' |
| 2072 | When linking dynamically to a dll without an import library, |
| 2073 | search for `<string><basename>.dll' in preference to |
| 2074 | `lib<basename>.dll'. This behaviour allows easy distinction |
| 2075 | between DLLs built for the various "subplatforms": native, cygwin, |
| 2076 | uwin, pw, etc. For instance, cygwin DLLs typically use |
| 2077 | `--dll-search-prefix=cyg'. [This option is specific to the i386 |
| 2078 | PE targeted port of the linker] |
| 2079 | |
| 2080 | `--enable-auto-import' |
| 2081 | Do sophisticated linking of `_symbol' to `__imp__symbol' for DATA |
| 2082 | imports from DLLs, and create the necessary thunking symbols when |
| 2083 | building the import libraries with those DATA exports. Note: Use |
| 2084 | of the 'auto-import' extension will cause the text section of the |
| 2085 | image file to be made writable. This does not conform to the |
| 2086 | PE-COFF format specification published by Microsoft. |
| 2087 | |
| 2088 | Note - use of the 'auto-import' extension will also cause read only |
| 2089 | data which would normally be placed into the .rdata section to be |
| 2090 | placed into the .data section instead. This is in order to work |
| 2091 | around a problem with consts that is described here: |
| 2092 | http://www.cygwin.com/ml/cygwin/2004-09/msg01101.html |
| 2093 | |
| 2094 | Using 'auto-import' generally will 'just work' - but sometimes you |
| 2095 | may see this message: |
| 2096 | |
| 2097 | "variable '<var>' can't be auto-imported. Please read the |
| 2098 | documentation for ld's `--enable-auto-import' for details." |
| 2099 | |
| 2100 | This message occurs when some (sub)expression accesses an address |
| 2101 | ultimately given by the sum of two constants (Win32 import tables |
| 2102 | only allow one). Instances where this may occur include accesses |
| 2103 | to member fields of struct variables imported from a DLL, as well |
| 2104 | as using a constant index into an array variable imported from a |
| 2105 | DLL. Any multiword variable (arrays, structs, long long, etc) may |
| 2106 | trigger this error condition. However, regardless of the exact |
| 2107 | data type of the offending exported variable, ld will always |
| 2108 | detect it, issue the warning, and exit. |
| 2109 | |
| 2110 | There are several ways to address this difficulty, regardless of |
| 2111 | the data type of the exported variable: |
| 2112 | |
| 2113 | One way is to use -enable-runtime-pseudo-reloc switch. This leaves |
| 2114 | the task of adjusting references in your client code for runtime |
| 2115 | environment, so this method works only when runtime environment |
| 2116 | supports this feature. |
| 2117 | |
| 2118 | A second solution is to force one of the 'constants' to be a |
| 2119 | variable - that is, unknown and un-optimizable at compile time. |
| 2120 | For arrays, there are two possibilities: a) make the indexee (the |
| 2121 | array's address) a variable, or b) make the 'constant' index a |
| 2122 | variable. Thus: |
| 2123 | |
| 2124 | extern type extern_array[]; |
| 2125 | extern_array[1] --> |
| 2126 | { volatile type *t=extern_array; t[1] } |
| 2127 | |
| 2128 | or |
| 2129 | |
| 2130 | extern type extern_array[]; |
| 2131 | extern_array[1] --> |
| 2132 | { volatile int t=1; extern_array[t] } |
| 2133 | |
| 2134 | For structs (and most other multiword data types) the only option |
| 2135 | is to make the struct itself (or the long long, or the ...) |
| 2136 | variable: |
| 2137 | |
| 2138 | extern struct s extern_struct; |
| 2139 | extern_struct.field --> |
| 2140 | { volatile struct s *t=&extern_struct; t->field } |
| 2141 | |
| 2142 | or |
| 2143 | |
| 2144 | extern long long extern_ll; |
| 2145 | extern_ll --> |
| 2146 | { volatile long long * local_ll=&extern_ll; *local_ll } |
| 2147 | |
| 2148 | A third method of dealing with this difficulty is to abandon |
| 2149 | 'auto-import' for the offending symbol and mark it with |
| 2150 | `__declspec(dllimport)'. However, in practice that requires using |
| 2151 | compile-time #defines to indicate whether you are building a DLL, |
| 2152 | building client code that will link to the DLL, or merely |
| 2153 | building/linking to a static library. In making the choice |
| 2154 | between the various methods of resolving the 'direct address with |
| 2155 | constant offset' problem, you should consider typical real-world |
| 2156 | usage: |
| 2157 | |
| 2158 | Original: |
| 2159 | --foo.h |
| 2160 | extern int arr[]; |
| 2161 | --foo.c |
| 2162 | #include "foo.h" |
| 2163 | void main(int argc, char **argv){ |
| 2164 | printf("%d\n",arr[1]); |
| 2165 | } |
| 2166 | |
| 2167 | Solution 1: |
| 2168 | --foo.h |
| 2169 | extern int arr[]; |
| 2170 | --foo.c |
| 2171 | #include "foo.h" |
| 2172 | void main(int argc, char **argv){ |
| 2173 | /* This workaround is for win32 and cygwin; do not "optimize" */ |
| 2174 | volatile int *parr = arr; |
| 2175 | printf("%d\n",parr[1]); |
| 2176 | } |
| 2177 | |
| 2178 | Solution 2: |
| 2179 | --foo.h |
| 2180 | /* Note: auto-export is assumed (no __declspec(dllexport)) */ |
| 2181 | #if (defined(_WIN32) || defined(__CYGWIN__)) && \ |
| 2182 | !(defined(FOO_BUILD_DLL) || defined(FOO_STATIC)) |
| 2183 | #define FOO_IMPORT __declspec(dllimport) |
| 2184 | #else |
| 2185 | #define FOO_IMPORT |
| 2186 | #endif |
| 2187 | extern FOO_IMPORT int arr[]; |
| 2188 | --foo.c |
| 2189 | #include "foo.h" |
| 2190 | void main(int argc, char **argv){ |
| 2191 | printf("%d\n",arr[1]); |
| 2192 | } |
| 2193 | |
| 2194 | A fourth way to avoid this problem is to re-code your library to |
| 2195 | use a functional interface rather than a data interface for the |
| 2196 | offending variables (e.g. set_foo() and get_foo() accessor |
| 2197 | functions). [This option is specific to the i386 PE targeted port |
| 2198 | of the linker] |
| 2199 | |
| 2200 | `--disable-auto-import' |
| 2201 | Do not attempt to do sophisticated linking of `_symbol' to |
| 2202 | `__imp__symbol' for DATA imports from DLLs. [This option is |
| 2203 | specific to the i386 PE targeted port of the linker] |
| 2204 | |
| 2205 | `--enable-runtime-pseudo-reloc' |
| 2206 | If your code contains expressions described in -enable-auto-import |
| 2207 | section, that is, DATA imports from DLL with non-zero offset, this |
| 2208 | switch will create a vector of 'runtime pseudo relocations' which |
| 2209 | can be used by runtime environment to adjust references to such |
| 2210 | data in your client code. [This option is specific to the i386 PE |
| 2211 | targeted port of the linker] |
| 2212 | |
| 2213 | `--disable-runtime-pseudo-reloc' |
| 2214 | Do not create pseudo relocations for non-zero offset DATA imports |
| 2215 | from DLLs. [This option is specific to the i386 PE targeted port |
| 2216 | of the linker] |
| 2217 | |
| 2218 | `--enable-extra-pe-debug' |
| 2219 | Show additional debug info related to auto-import symbol thunking. |
| 2220 | [This option is specific to the i386 PE targeted port of the |
| 2221 | linker] |
| 2222 | |
| 2223 | `--section-alignment' |
| 2224 | Sets the section alignment. Sections in memory will always begin |
| 2225 | at addresses which are a multiple of this number. Defaults to |
| 2226 | 0x1000. [This option is specific to the i386 PE targeted port of |
| 2227 | the linker] |
| 2228 | |
| 2229 | `--stack RESERVE' |
| 2230 | `--stack RESERVE,COMMIT' |
| 2231 | Specify the number of bytes of memory to reserve (and optionally |
| 2232 | commit) to be used as stack for this program. The default is 2MB |
| 2233 | reserved, 4K committed. [This option is specific to the i386 PE |
| 2234 | targeted port of the linker] |
| 2235 | |
| 2236 | `--subsystem WHICH' |
| 2237 | `--subsystem WHICH:MAJOR' |
| 2238 | `--subsystem WHICH:MAJOR.MINOR' |
| 2239 | Specifies the subsystem under which your program will execute. The |
| 2240 | legal values for WHICH are `native', `windows', `console', |
| 2241 | `posix', and `xbox'. You may optionally set the subsystem version |
| 2242 | also. Numeric values are also accepted for WHICH. [This option |
| 2243 | is specific to the i386 PE targeted port of the linker] |
| 2244 | |
| 2245 | The following options set flags in the `DllCharacteristics' field |
| 2246 | of the PE file header: [These options are specific to PE targeted |
| 2247 | ports of the linker] |
| 2248 | |
| 2249 | `--high-entropy-va' |
| 2250 | Image is compatible with 64-bit address space layout randomization |
| 2251 | (ASLR). |
| 2252 | |
| 2253 | `--dynamicbase' |
| 2254 | The image base address may be relocated using address space layout |
| 2255 | randomization (ASLR). This feature was introduced with MS Windows |
| 2256 | Vista for i386 PE targets. |
| 2257 | |
| 2258 | `--forceinteg' |
| 2259 | Code integrity checks are enforced. |
| 2260 | |
| 2261 | `--nxcompat' |
| 2262 | The image is compatible with the Data Execution Prevention. This |
| 2263 | feature was introduced with MS Windows XP SP2 for i386 PE targets. |
| 2264 | |
| 2265 | `--no-isolation' |
| 2266 | Although the image understands isolation, do not isolate the image. |
| 2267 | |
| 2268 | `--no-seh' |
| 2269 | The image does not use SEH. No SE handler may be called from this |
| 2270 | image. |
| 2271 | |
| 2272 | `--no-bind' |
| 2273 | Do not bind this image. |
| 2274 | |
| 2275 | `--wdmdriver' |
| 2276 | The driver uses the MS Windows Driver Model. |
| 2277 | |
| 2278 | `--tsaware' |
| 2279 | The image is Terminal Server aware. |
| 2280 | |
| 2281 | `--insert-timestamp' |
| 2282 | `--no-insert-timestamp' |
| 2283 | Insert a real timestamp into the image. This is the default |
| 2284 | behaviour as it matches legacy code and it means that the image |
| 2285 | will work with other, proprietary tools. The problem with this |
| 2286 | default is that it will result in slightly different images being |
| 2287 | produced each time the same sources are linked. The option |
| 2288 | `--no-insert-timestamp' can be used to insert a zero value for the |
| 2289 | timestamp, this ensuring that binaries produced from identical |
| 2290 | sources will compare identically. |
| 2291 | |
| 2292 | 2.1.2 Options specific to C6X uClinux targets |
| 2293 | --------------------------------------------- |
| 2294 | |
| 2295 | The C6X uClinux target uses a binary format called DSBT to support |
| 2296 | shared libraries. Each shared library in the system needs to have a |
| 2297 | unique index; all executables use an index of 0. |
| 2298 | |
| 2299 | `--dsbt-size SIZE' |
| 2300 | This option sets the number of entries in the DSBT of the current |
| 2301 | executable or shared library to SIZE. The default is to create a |
| 2302 | table with 64 entries. |
| 2303 | |
| 2304 | `--dsbt-index INDEX' |
| 2305 | This option sets the DSBT index of the current executable or |
| 2306 | shared library to INDEX. The default is 0, which is appropriate |
| 2307 | for generating executables. If a shared library is generated with |
| 2308 | a DSBT index of 0, the `R_C6000_DSBT_INDEX' relocs are copied into |
| 2309 | the output file. |
| 2310 | |
| 2311 | The `--no-merge-exidx-entries' switch disables the merging of |
| 2312 | adjacent exidx entries in frame unwind info. |
| 2313 | |
| 2314 | |
| 2315 | 2.1.3 Options specific to Motorola 68HC11 and 68HC12 targets |
| 2316 | ------------------------------------------------------------ |
| 2317 | |
| 2318 | The 68HC11 and 68HC12 linkers support specific options to control the |
| 2319 | memory bank switching mapping and trampoline code generation. |
| 2320 | |
| 2321 | `--no-trampoline' |
| 2322 | This option disables the generation of trampoline. By default a |
| 2323 | trampoline is generated for each far function which is called |
| 2324 | using a `jsr' instruction (this happens when a pointer to a far |
| 2325 | function is taken). |
| 2326 | |
| 2327 | `--bank-window NAME' |
| 2328 | This option indicates to the linker the name of the memory region |
| 2329 | in the `MEMORY' specification that describes the memory bank |
| 2330 | window. The definition of such region is then used by the linker |
| 2331 | to compute paging and addresses within the memory window. |
| 2332 | |
| 2333 | |
| 2334 | 2.1.4 Options specific to Motorola 68K target |
| 2335 | --------------------------------------------- |
| 2336 | |
| 2337 | The following options are supported to control handling of GOT |
| 2338 | generation when linking for 68K targets. |
| 2339 | |
| 2340 | `--got=TYPE' |
| 2341 | This option tells the linker which GOT generation scheme to use. |
| 2342 | TYPE should be one of `single', `negative', `multigot' or |
| 2343 | `target'. For more information refer to the Info entry for `ld'. |
| 2344 | |
| 2345 | |
| 2346 | 2.1.5 Options specific to MIPS targets |
| 2347 | -------------------------------------- |
| 2348 | |
| 2349 | The following options are supported to control microMIPS instruction |
| 2350 | generation when linking for MIPS targets. |
| 2351 | |
| 2352 | `--insn32' |
| 2353 | `--no-insn32' |
| 2354 | These options control the choice of microMIPS instructions used in |
| 2355 | code generated by the linker, such as that in the PLT or lazy |
| 2356 | binding stubs, or in relaxation. If `--insn32' is used, then the |
| 2357 | linker only uses 32-bit instruction encodings. By default or if |
| 2358 | `--no-insn32' is used, all instruction encodings are used, |
| 2359 | including 16-bit ones where possible. |
| 2360 | |
| 2361 | |
| 2362 | |
| 2363 | File: ld.info, Node: Environment, Prev: Options, Up: Invocation |
| 2364 | |
| 2365 | 2.2 Environment Variables |
| 2366 | ========================= |
| 2367 | |
| 2368 | You can change the behaviour of `ld' with the environment variables |
| 2369 | `GNUTARGET', `LDEMULATION' and `COLLECT_NO_DEMANGLE'. |
| 2370 | |
| 2371 | `GNUTARGET' determines the input-file object format if you don't use |
| 2372 | `-b' (or its synonym `--format'). Its value should be one of the BFD |
| 2373 | names for an input format (*note BFD::). If there is no `GNUTARGET' in |
| 2374 | the environment, `ld' uses the natural format of the target. If |
| 2375 | `GNUTARGET' is set to `default' then BFD attempts to discover the input |
| 2376 | format by examining binary input files; this method often succeeds, but |
| 2377 | there are potential ambiguities, since there is no method of ensuring |
| 2378 | that the magic number used to specify object-file formats is unique. |
| 2379 | However, the configuration procedure for BFD on each system places the |
| 2380 | conventional format for that system first in the search-list, so |
| 2381 | ambiguities are resolved in favor of convention. |
| 2382 | |
| 2383 | `LDEMULATION' determines the default emulation if you don't use the |
| 2384 | `-m' option. The emulation can affect various aspects of linker |
| 2385 | behaviour, particularly the default linker script. You can list the |
| 2386 | available emulations with the `--verbose' or `-V' options. If the `-m' |
| 2387 | option is not used, and the `LDEMULATION' environment variable is not |
| 2388 | defined, the default emulation depends upon how the linker was |
| 2389 | configured. |
| 2390 | |
| 2391 | Normally, the linker will default to demangling symbols. However, if |
| 2392 | `COLLECT_NO_DEMANGLE' is set in the environment, then it will default |
| 2393 | to not demangling symbols. This environment variable is used in a |
| 2394 | similar fashion by the `gcc' linker wrapper program. The default may |
| 2395 | be overridden by the `--demangle' and `--no-demangle' options. |
| 2396 | |
| 2397 | |
| 2398 | File: ld.info, Node: Scripts, Next: Machine Dependent, Prev: Invocation, Up: Top |
| 2399 | |
| 2400 | 3 Linker Scripts |
| 2401 | **************** |
| 2402 | |
| 2403 | Every link is controlled by a "linker script". This script is written |
| 2404 | in the linker command language. |
| 2405 | |
| 2406 | The main purpose of the linker script is to describe how the |
| 2407 | sections in the input files should be mapped into the output file, and |
| 2408 | to control the memory layout of the output file. Most linker scripts |
| 2409 | do nothing more than this. However, when necessary, the linker script |
| 2410 | can also direct the linker to perform many other operations, using the |
| 2411 | commands described below. |
| 2412 | |
| 2413 | The linker always uses a linker script. If you do not supply one |
| 2414 | yourself, the linker will use a default script that is compiled into the |
| 2415 | linker executable. You can use the `--verbose' command line option to |
| 2416 | display the default linker script. Certain command line options, such |
| 2417 | as `-r' or `-N', will affect the default linker script. |
| 2418 | |
| 2419 | You may supply your own linker script by using the `-T' command line |
| 2420 | option. When you do this, your linker script will replace the default |
| 2421 | linker script. |
| 2422 | |
| 2423 | You may also use linker scripts implicitly by naming them as input |
| 2424 | files to the linker, as though they were files to be linked. *Note |
| 2425 | Implicit Linker Scripts::. |
| 2426 | |
| 2427 | * Menu: |
| 2428 | |
| 2429 | * Basic Script Concepts:: Basic Linker Script Concepts |
| 2430 | * Script Format:: Linker Script Format |
| 2431 | * Simple Example:: Simple Linker Script Example |
| 2432 | * Simple Commands:: Simple Linker Script Commands |
| 2433 | * Assignments:: Assigning Values to Symbols |
| 2434 | * SECTIONS:: SECTIONS Command |
| 2435 | * MEMORY:: MEMORY Command |
| 2436 | * PHDRS:: PHDRS Command |
| 2437 | * VERSION:: VERSION Command |
| 2438 | * Expressions:: Expressions in Linker Scripts |
| 2439 | * Implicit Linker Scripts:: Implicit Linker Scripts |
| 2440 | |
| 2441 | |
| 2442 | File: ld.info, Node: Basic Script Concepts, Next: Script Format, Up: Scripts |
| 2443 | |
| 2444 | 3.1 Basic Linker Script Concepts |
| 2445 | ================================ |
| 2446 | |
| 2447 | We need to define some basic concepts and vocabulary in order to |
| 2448 | describe the linker script language. |
| 2449 | |
| 2450 | The linker combines input files into a single output file. The |
| 2451 | output file and each input file are in a special data format known as an |
| 2452 | "object file format". Each file is called an "object file". The |
| 2453 | output file is often called an "executable", but for our purposes we |
| 2454 | will also call it an object file. Each object file has, among other |
| 2455 | things, a list of "sections". We sometimes refer to a section in an |
| 2456 | input file as an "input section"; similarly, a section in the output |
| 2457 | file is an "output section". |
| 2458 | |
| 2459 | Each section in an object file has a name and a size. Most sections |
| 2460 | also have an associated block of data, known as the "section contents". |
| 2461 | A section may be marked as "loadable", which means that the contents |
| 2462 | should be loaded into memory when the output file is run. A section |
| 2463 | with no contents may be "allocatable", which means that an area in |
| 2464 | memory should be set aside, but nothing in particular should be loaded |
| 2465 | there (in some cases this memory must be zeroed out). A section which |
| 2466 | is neither loadable nor allocatable typically contains some sort of |
| 2467 | debugging information. |
| 2468 | |
| 2469 | Every loadable or allocatable output section has two addresses. The |
| 2470 | first is the "VMA", or virtual memory address. This is the address the |
| 2471 | section will have when the output file is run. The second is the |
| 2472 | "LMA", or load memory address. This is the address at which the |
| 2473 | section will be loaded. In most cases the two addresses will be the |
| 2474 | same. An example of when they might be different is when a data section |
| 2475 | is loaded into ROM, and then copied into RAM when the program starts up |
| 2476 | (this technique is often used to initialize global variables in a ROM |
| 2477 | based system). In this case the ROM address would be the LMA, and the |
| 2478 | RAM address would be the VMA. |
| 2479 | |
| 2480 | You can see the sections in an object file by using the `objdump' |
| 2481 | program with the `-h' option. |
| 2482 | |
| 2483 | Every object file also has a list of "symbols", known as the "symbol |
| 2484 | table". A symbol may be defined or undefined. Each symbol has a name, |
| 2485 | and each defined symbol has an address, among other information. If |
| 2486 | you compile a C or C++ program into an object file, you will get a |
| 2487 | defined symbol for every defined function and global or static |
| 2488 | variable. Every undefined function or global variable which is |
| 2489 | referenced in the input file will become an undefined symbol. |
| 2490 | |
| 2491 | You can see the symbols in an object file by using the `nm' program, |
| 2492 | or by using the `objdump' program with the `-t' option. |
| 2493 | |
| 2494 | |
| 2495 | File: ld.info, Node: Script Format, Next: Simple Example, Prev: Basic Script Concepts, Up: Scripts |
| 2496 | |
| 2497 | 3.2 Linker Script Format |
| 2498 | ======================== |
| 2499 | |
| 2500 | Linker scripts are text files. |
| 2501 | |
| 2502 | You write a linker script as a series of commands. Each command is |
| 2503 | either a keyword, possibly followed by arguments, or an assignment to a |
| 2504 | symbol. You may separate commands using semicolons. Whitespace is |
| 2505 | generally ignored. |
| 2506 | |
| 2507 | Strings such as file or format names can normally be entered |
| 2508 | directly. If the file name contains a character such as a comma which |
| 2509 | would otherwise serve to separate file names, you may put the file name |
| 2510 | in double quotes. There is no way to use a double quote character in a |
| 2511 | file name. |
| 2512 | |
| 2513 | You may include comments in linker scripts just as in C, delimited by |
| 2514 | `/*' and `*/'. As in C, comments are syntactically equivalent to |
| 2515 | whitespace. |
| 2516 | |
| 2517 | |
| 2518 | File: ld.info, Node: Simple Example, Next: Simple Commands, Prev: Script Format, Up: Scripts |
| 2519 | |
| 2520 | 3.3 Simple Linker Script Example |
| 2521 | ================================ |
| 2522 | |
| 2523 | Many linker scripts are fairly simple. |
| 2524 | |
| 2525 | The simplest possible linker script has just one command: |
| 2526 | `SECTIONS'. You use the `SECTIONS' command to describe the memory |
| 2527 | layout of the output file. |
| 2528 | |
| 2529 | The `SECTIONS' command is a powerful command. Here we will describe |
| 2530 | a simple use of it. Let's assume your program consists only of code, |
| 2531 | initialized data, and uninitialized data. These will be in the |
| 2532 | `.text', `.data', and `.bss' sections, respectively. Let's assume |
| 2533 | further that these are the only sections which appear in your input |
| 2534 | files. |
| 2535 | |
| 2536 | For this example, let's say that the code should be loaded at address |
| 2537 | 0x10000, and that the data should start at address 0x8000000. Here is a |
| 2538 | linker script which will do that: |
| 2539 | SECTIONS |
| 2540 | { |
| 2541 | . = 0x10000; |
| 2542 | .text : { *(.text) } |
| 2543 | . = 0x8000000; |
| 2544 | .data : { *(.data) } |
| 2545 | .bss : { *(.bss) } |
| 2546 | } |
| 2547 | |
| 2548 | You write the `SECTIONS' command as the keyword `SECTIONS', followed |
| 2549 | by a series of symbol assignments and output section descriptions |
| 2550 | enclosed in curly braces. |
| 2551 | |
| 2552 | The first line inside the `SECTIONS' command of the above example |
| 2553 | sets the value of the special symbol `.', which is the location |
| 2554 | counter. If you do not specify the address of an output section in some |
| 2555 | other way (other ways are described later), the address is set from the |
| 2556 | current value of the location counter. The location counter is then |
| 2557 | incremented by the size of the output section. At the start of the |
| 2558 | `SECTIONS' command, the location counter has the value `0'. |
| 2559 | |
| 2560 | The second line defines an output section, `.text'. The colon is |
| 2561 | required syntax which may be ignored for now. Within the curly braces |
| 2562 | after the output section name, you list the names of the input sections |
| 2563 | which should be placed into this output section. The `*' is a wildcard |
| 2564 | which matches any file name. The expression `*(.text)' means all |
| 2565 | `.text' input sections in all input files. |
| 2566 | |
| 2567 | Since the location counter is `0x10000' when the output section |
| 2568 | `.text' is defined, the linker will set the address of the `.text' |
| 2569 | section in the output file to be `0x10000'. |
| 2570 | |
| 2571 | The remaining lines define the `.data' and `.bss' sections in the |
| 2572 | output file. The linker will place the `.data' output section at |
| 2573 | address `0x8000000'. After the linker places the `.data' output |
| 2574 | section, the value of the location counter will be `0x8000000' plus the |
| 2575 | size of the `.data' output section. The effect is that the linker will |
| 2576 | place the `.bss' output section immediately after the `.data' output |
| 2577 | section in memory. |
| 2578 | |
| 2579 | The linker will ensure that each output section has the required |
| 2580 | alignment, by increasing the location counter if necessary. In this |
| 2581 | example, the specified addresses for the `.text' and `.data' sections |
| 2582 | will probably satisfy any alignment constraints, but the linker may |
| 2583 | have to create a small gap between the `.data' and `.bss' sections. |
| 2584 | |
| 2585 | That's it! That's a simple and complete linker script. |
| 2586 | |
| 2587 | |
| 2588 | File: ld.info, Node: Simple Commands, Next: Assignments, Prev: Simple Example, Up: Scripts |
| 2589 | |
| 2590 | 3.4 Simple Linker Script Commands |
| 2591 | ================================= |
| 2592 | |
| 2593 | In this section we describe the simple linker script commands. |
| 2594 | |
| 2595 | * Menu: |
| 2596 | |
| 2597 | * Entry Point:: Setting the entry point |
| 2598 | * File Commands:: Commands dealing with files |
| 2599 | |
| 2600 | * Format Commands:: Commands dealing with object file formats |
| 2601 | |
| 2602 | * REGION_ALIAS:: Assign alias names to memory regions |
| 2603 | * Miscellaneous Commands:: Other linker script commands |
| 2604 | |
| 2605 | |
| 2606 | File: ld.info, Node: Entry Point, Next: File Commands, Up: Simple Commands |
| 2607 | |
| 2608 | 3.4.1 Setting the Entry Point |
| 2609 | ----------------------------- |
| 2610 | |
| 2611 | The first instruction to execute in a program is called the "entry |
| 2612 | point". You can use the `ENTRY' linker script command to set the entry |
| 2613 | point. The argument is a symbol name: |
| 2614 | ENTRY(SYMBOL) |
| 2615 | |
| 2616 | There are several ways to set the entry point. The linker will set |
| 2617 | the entry point by trying each of the following methods in order, and |
| 2618 | stopping when one of them succeeds: |
| 2619 | * the `-e' ENTRY command-line option; |
| 2620 | |
| 2621 | * the `ENTRY(SYMBOL)' command in a linker script; |
| 2622 | |
| 2623 | * the value of a target specific symbol, if it is defined; For many |
| 2624 | targets this is `start', but PE and BeOS based systems for example |
| 2625 | check a list of possible entry symbols, matching the first one |
| 2626 | found. |
| 2627 | |
| 2628 | * the address of the first byte of the `.text' section, if present; |
| 2629 | |
| 2630 | * The address `0'. |
| 2631 | |
| 2632 | |
| 2633 | File: ld.info, Node: File Commands, Next: Format Commands, Prev: Entry Point, Up: Simple Commands |
| 2634 | |
| 2635 | 3.4.2 Commands Dealing with Files |
| 2636 | --------------------------------- |
| 2637 | |
| 2638 | Several linker script commands deal with files. |
| 2639 | |
| 2640 | `INCLUDE FILENAME' |
| 2641 | Include the linker script FILENAME at this point. The file will |
| 2642 | be searched for in the current directory, and in any directory |
| 2643 | specified with the `-L' option. You can nest calls to `INCLUDE' |
| 2644 | up to 10 levels deep. |
| 2645 | |
| 2646 | You can place `INCLUDE' directives at the top level, in `MEMORY' or |
| 2647 | `SECTIONS' commands, or in output section descriptions. |
| 2648 | |
| 2649 | `INPUT(FILE, FILE, ...)' |
| 2650 | `INPUT(FILE FILE ...)' |
| 2651 | The `INPUT' command directs the linker to include the named files |
| 2652 | in the link, as though they were named on the command line. |
| 2653 | |
| 2654 | For example, if you always want to include `subr.o' any time you do |
| 2655 | a link, but you can't be bothered to put it on every link command |
| 2656 | line, then you can put `INPUT (subr.o)' in your linker script. |
| 2657 | |
| 2658 | In fact, if you like, you can list all of your input files in the |
| 2659 | linker script, and then invoke the linker with nothing but a `-T' |
| 2660 | option. |
| 2661 | |
| 2662 | In case a "sysroot prefix" is configured, and the filename starts |
| 2663 | with the `/' character, and the script being processed was located |
| 2664 | inside the "sysroot prefix", the filename will be looked for in |
| 2665 | the "sysroot prefix". Otherwise, the linker will try to open the |
| 2666 | file in the current directory. If it is not found, the linker |
| 2667 | will search through the archive library search path. The "sysroot |
| 2668 | prefix" can also be forced by specifying `=' as the first |
| 2669 | character in the filename path. See also the description of `-L' |
| 2670 | in *Note Command Line Options: Options. |
| 2671 | |
| 2672 | If you use `INPUT (-lFILE)', `ld' will transform the name to |
| 2673 | `libFILE.a', as with the command line argument `-l'. |
| 2674 | |
| 2675 | When you use the `INPUT' command in an implicit linker script, the |
| 2676 | files will be included in the link at the point at which the linker |
| 2677 | script file is included. This can affect archive searching. |
| 2678 | |
| 2679 | `GROUP(FILE, FILE, ...)' |
| 2680 | `GROUP(FILE FILE ...)' |
| 2681 | The `GROUP' command is like `INPUT', except that the named files |
| 2682 | should all be archives, and they are searched repeatedly until no |
| 2683 | new undefined references are created. See the description of `-(' |
| 2684 | in *Note Command Line Options: Options. |
| 2685 | |
| 2686 | `AS_NEEDED(FILE, FILE, ...)' |
| 2687 | `AS_NEEDED(FILE FILE ...)' |
| 2688 | This construct can appear only inside of the `INPUT' or `GROUP' |
| 2689 | commands, among other filenames. The files listed will be handled |
| 2690 | as if they appear directly in the `INPUT' or `GROUP' commands, |
| 2691 | with the exception of ELF shared libraries, that will be added only |
| 2692 | when they are actually needed. This construct essentially enables |
| 2693 | `--as-needed' option for all the files listed inside of it and |
| 2694 | restores previous `--as-needed' resp. `--no-as-needed' setting |
| 2695 | afterwards. |
| 2696 | |
| 2697 | `OUTPUT(FILENAME)' |
| 2698 | The `OUTPUT' command names the output file. Using |
| 2699 | `OUTPUT(FILENAME)' in the linker script is exactly like using `-o |
| 2700 | FILENAME' on the command line (*note Command Line Options: |
| 2701 | Options.). If both are used, the command line option takes |
| 2702 | precedence. |
| 2703 | |
| 2704 | You can use the `OUTPUT' command to define a default name for the |
| 2705 | output file other than the usual default of `a.out'. |
| 2706 | |
| 2707 | `SEARCH_DIR(PATH)' |
| 2708 | The `SEARCH_DIR' command adds PATH to the list of paths where `ld' |
| 2709 | looks for archive libraries. Using `SEARCH_DIR(PATH)' is exactly |
| 2710 | like using `-L PATH' on the command line (*note Command Line |
| 2711 | Options: Options.). If both are used, then the linker will search |
| 2712 | both paths. Paths specified using the command line option are |
| 2713 | searched first. |
| 2714 | |
| 2715 | `STARTUP(FILENAME)' |
| 2716 | The `STARTUP' command is just like the `INPUT' command, except |
| 2717 | that FILENAME will become the first input file to be linked, as |
| 2718 | though it were specified first on the command line. This may be |
| 2719 | useful when using a system in which the entry point is always the |
| 2720 | start of the first file. |
| 2721 | |
| 2722 | |
| 2723 | File: ld.info, Node: Format Commands, Next: REGION_ALIAS, Prev: File Commands, Up: Simple Commands |
| 2724 | |
| 2725 | 3.4.3 Commands Dealing with Object File Formats |
| 2726 | ----------------------------------------------- |
| 2727 | |
| 2728 | A couple of linker script commands deal with object file formats. |
| 2729 | |
| 2730 | `OUTPUT_FORMAT(BFDNAME)' |
| 2731 | `OUTPUT_FORMAT(DEFAULT, BIG, LITTLE)' |
| 2732 | The `OUTPUT_FORMAT' command names the BFD format to use for the |
| 2733 | output file (*note BFD::). Using `OUTPUT_FORMAT(BFDNAME)' is |
| 2734 | exactly like using `--oformat BFDNAME' on the command line (*note |
| 2735 | Command Line Options: Options.). If both are used, the command |
| 2736 | line option takes precedence. |
| 2737 | |
| 2738 | You can use `OUTPUT_FORMAT' with three arguments to use different |
| 2739 | formats based on the `-EB' and `-EL' command line options. This |
| 2740 | permits the linker script to set the output format based on the |
| 2741 | desired endianness. |
| 2742 | |
| 2743 | If neither `-EB' nor `-EL' are used, then the output format will |
| 2744 | be the first argument, DEFAULT. If `-EB' is used, the output |
| 2745 | format will be the second argument, BIG. If `-EL' is used, the |
| 2746 | output format will be the third argument, LITTLE. |
| 2747 | |
| 2748 | For example, the default linker script for the MIPS ELF target |
| 2749 | uses this command: |
| 2750 | OUTPUT_FORMAT(elf32-bigmips, elf32-bigmips, elf32-littlemips) |
| 2751 | This says that the default format for the output file is |
| 2752 | `elf32-bigmips', but if the user uses the `-EL' command line |
| 2753 | option, the output file will be created in the `elf32-littlemips' |
| 2754 | format. |
| 2755 | |
| 2756 | `TARGET(BFDNAME)' |
| 2757 | The `TARGET' command names the BFD format to use when reading input |
| 2758 | files. It affects subsequent `INPUT' and `GROUP' commands. This |
| 2759 | command is like using `-b BFDNAME' on the command line (*note |
| 2760 | Command Line Options: Options.). If the `TARGET' command is used |
| 2761 | but `OUTPUT_FORMAT' is not, then the last `TARGET' command is also |
| 2762 | used to set the format for the output file. *Note BFD::. |
| 2763 | |
| 2764 | |
| 2765 | File: ld.info, Node: REGION_ALIAS, Next: Miscellaneous Commands, Prev: Format Commands, Up: Simple Commands |
| 2766 | |
| 2767 | 3.4.4 Assign alias names to memory regions |
| 2768 | ------------------------------------------ |
| 2769 | |
| 2770 | Alias names can be added to existing memory regions created with the |
| 2771 | *Note MEMORY:: command. Each name corresponds to at most one memory |
| 2772 | region. |
| 2773 | |
| 2774 | REGION_ALIAS(ALIAS, REGION) |
| 2775 | |
| 2776 | The `REGION_ALIAS' function creates an alias name ALIAS for the |
| 2777 | memory region REGION. This allows a flexible mapping of output sections |
| 2778 | to memory regions. An example follows. |
| 2779 | |
| 2780 | Suppose we have an application for embedded systems which come with |
| 2781 | various memory storage devices. All have a general purpose, volatile |
| 2782 | memory `RAM' that allows code execution or data storage. Some may have |
| 2783 | a read-only, non-volatile memory `ROM' that allows code execution and |
| 2784 | read-only data access. The last variant is a read-only, non-volatile |
| 2785 | memory `ROM2' with read-only data access and no code execution |
| 2786 | capability. We have four output sections: |
| 2787 | |
| 2788 | * `.text' program code; |
| 2789 | |
| 2790 | * `.rodata' read-only data; |
| 2791 | |
| 2792 | * `.data' read-write initialized data; |
| 2793 | |
| 2794 | * `.bss' read-write zero initialized data. |
| 2795 | |
| 2796 | The goal is to provide a linker command file that contains a system |
| 2797 | independent part defining the output sections and a system dependent |
| 2798 | part mapping the output sections to the memory regions available on the |
| 2799 | system. Our embedded systems come with three different memory setups |
| 2800 | `A', `B' and `C': |
| 2801 | Section Variant A Variant B Variant C |
| 2802 | .text RAM ROM ROM |
| 2803 | .rodata RAM ROM ROM2 |
| 2804 | .data RAM RAM/ROM RAM/ROM2 |
| 2805 | .bss RAM RAM RAM |
| 2806 | The notation `RAM/ROM' or `RAM/ROM2' means that this section is |
| 2807 | loaded into region `ROM' or `ROM2' respectively. Please note that the |
| 2808 | load address of the `.data' section starts in all three variants at the |
| 2809 | end of the `.rodata' section. |
| 2810 | |
| 2811 | The base linker script that deals with the output sections follows. |
| 2812 | It includes the system dependent `linkcmds.memory' file that describes |
| 2813 | the memory layout: |
| 2814 | INCLUDE linkcmds.memory |
| 2815 | |
| 2816 | SECTIONS |
| 2817 | { |
| 2818 | .text : |
| 2819 | { |
| 2820 | *(.text) |
| 2821 | } > REGION_TEXT |
| 2822 | .rodata : |
| 2823 | { |
| 2824 | *(.rodata) |
| 2825 | rodata_end = .; |
| 2826 | } > REGION_RODATA |
| 2827 | .data : AT (rodata_end) |
| 2828 | { |
| 2829 | data_start = .; |
| 2830 | *(.data) |
| 2831 | } > REGION_DATA |
| 2832 | data_size = SIZEOF(.data); |
| 2833 | data_load_start = LOADADDR(.data); |
| 2834 | .bss : |
| 2835 | { |
| 2836 | *(.bss) |
| 2837 | } > REGION_BSS |
| 2838 | } |
| 2839 | |
| 2840 | Now we need three different `linkcmds.memory' files to define memory |
| 2841 | regions and alias names. The content of `linkcmds.memory' for the three |
| 2842 | variants `A', `B' and `C': |
| 2843 | `A' |
| 2844 | Here everything goes into the `RAM'. |
| 2845 | MEMORY |
| 2846 | { |
| 2847 | RAM : ORIGIN = 0, LENGTH = 4M |
| 2848 | } |
| 2849 | |
| 2850 | REGION_ALIAS("REGION_TEXT", RAM); |
| 2851 | REGION_ALIAS("REGION_RODATA", RAM); |
| 2852 | REGION_ALIAS("REGION_DATA", RAM); |
| 2853 | REGION_ALIAS("REGION_BSS", RAM); |
| 2854 | |
| 2855 | `B' |
| 2856 | Program code and read-only data go into the `ROM'. Read-write |
| 2857 | data goes into the `RAM'. An image of the initialized data is |
| 2858 | loaded into the `ROM' and will be copied during system start into |
| 2859 | the `RAM'. |
| 2860 | MEMORY |
| 2861 | { |
| 2862 | ROM : ORIGIN = 0, LENGTH = 3M |
| 2863 | RAM : ORIGIN = 0x10000000, LENGTH = 1M |
| 2864 | } |
| 2865 | |
| 2866 | REGION_ALIAS("REGION_TEXT", ROM); |
| 2867 | REGION_ALIAS("REGION_RODATA", ROM); |
| 2868 | REGION_ALIAS("REGION_DATA", RAM); |
| 2869 | REGION_ALIAS("REGION_BSS", RAM); |
| 2870 | |
| 2871 | `C' |
| 2872 | Program code goes into the `ROM'. Read-only data goes into the |
| 2873 | `ROM2'. Read-write data goes into the `RAM'. An image of the |
| 2874 | initialized data is loaded into the `ROM2' and will be copied |
| 2875 | during system start into the `RAM'. |
| 2876 | MEMORY |
| 2877 | { |
| 2878 | ROM : ORIGIN = 0, LENGTH = 2M |
| 2879 | ROM2 : ORIGIN = 0x10000000, LENGTH = 1M |
| 2880 | RAM : ORIGIN = 0x20000000, LENGTH = 1M |
| 2881 | } |
| 2882 | |
| 2883 | REGION_ALIAS("REGION_TEXT", ROM); |
| 2884 | REGION_ALIAS("REGION_RODATA", ROM2); |
| 2885 | REGION_ALIAS("REGION_DATA", RAM); |
| 2886 | REGION_ALIAS("REGION_BSS", RAM); |
| 2887 | |
| 2888 | It is possible to write a common system initialization routine to |
| 2889 | copy the `.data' section from `ROM' or `ROM2' into the `RAM' if |
| 2890 | necessary: |
| 2891 | #include <string.h> |
| 2892 | |
| 2893 | extern char data_start []; |
| 2894 | extern char data_size []; |
| 2895 | extern char data_load_start []; |
| 2896 | |
| 2897 | void copy_data(void) |
| 2898 | { |
| 2899 | if (data_start != data_load_start) |
| 2900 | { |
| 2901 | memcpy(data_start, data_load_start, (size_t) data_size); |
| 2902 | } |
| 2903 | } |
| 2904 | |
| 2905 | |
| 2906 | File: ld.info, Node: Miscellaneous Commands, Prev: REGION_ALIAS, Up: Simple Commands |
| 2907 | |
| 2908 | 3.4.5 Other Linker Script Commands |
| 2909 | ---------------------------------- |
| 2910 | |
| 2911 | There are a few other linker scripts commands. |
| 2912 | |
| 2913 | `ASSERT(EXP, MESSAGE)' |
| 2914 | Ensure that EXP is non-zero. If it is zero, then exit the linker |
| 2915 | with an error code, and print MESSAGE. |
| 2916 | |
| 2917 | Note that assertions are checked before the final stages of linking |
| 2918 | take place. This means that expressions involving symbols PROVIDEd |
| 2919 | inside section definitions will fail if the user has not set values |
| 2920 | for those symbols. The only exception to this rule is PROVIDEd |
| 2921 | symbols that just reference dot. Thus an assertion like this: |
| 2922 | |
| 2923 | .stack : |
| 2924 | { |
| 2925 | PROVIDE (__stack = .); |
| 2926 | PROVIDE (__stack_size = 0x100); |
| 2927 | ASSERT ((__stack > (_end + __stack_size)), "Error: No room left for the stack"); |
| 2928 | } |
| 2929 | |
| 2930 | will fail if `__stack_size' is not defined elsewhere. Symbols |
| 2931 | PROVIDEd outside of section definitions are evaluated earlier, so |
| 2932 | they can be used inside ASSERTions. Thus: |
| 2933 | |
| 2934 | PROVIDE (__stack_size = 0x100); |
| 2935 | .stack : |
| 2936 | { |
| 2937 | PROVIDE (__stack = .); |
| 2938 | ASSERT ((__stack > (_end + __stack_size)), "Error: No room left for the stack"); |
| 2939 | } |
| 2940 | |
| 2941 | will work. |
| 2942 | |
| 2943 | `EXTERN(SYMBOL SYMBOL ...)' |
| 2944 | Force SYMBOL to be entered in the output file as an undefined |
| 2945 | symbol. Doing this may, for example, trigger linking of additional |
| 2946 | modules from standard libraries. You may list several SYMBOLs for |
| 2947 | each `EXTERN', and you may use `EXTERN' multiple times. This |
| 2948 | command has the same effect as the `-u' command-line option. |
| 2949 | |
| 2950 | `FORCE_COMMON_ALLOCATION' |
| 2951 | This command has the same effect as the `-d' command-line option: |
| 2952 | to make `ld' assign space to common symbols even if a relocatable |
| 2953 | output file is specified (`-r'). |
| 2954 | |
| 2955 | `INHIBIT_COMMON_ALLOCATION' |
| 2956 | This command has the same effect as the `--no-define-common' |
| 2957 | command-line option: to make `ld' omit the assignment of addresses |
| 2958 | to common symbols even for a non-relocatable output file. |
| 2959 | |
| 2960 | `INSERT [ AFTER | BEFORE ] OUTPUT_SECTION' |
| 2961 | This command is typically used in a script specified by `-T' to |
| 2962 | augment the default `SECTIONS' with, for example, overlays. It |
| 2963 | inserts all prior linker script statements after (or before) |
| 2964 | OUTPUT_SECTION, and also causes `-T' to not override the default |
| 2965 | linker script. The exact insertion point is as for orphan |
| 2966 | sections. *Note Location Counter::. The insertion happens after |
| 2967 | the linker has mapped input sections to output sections. Prior to |
| 2968 | the insertion, since `-T' scripts are parsed before the default |
| 2969 | linker script, statements in the `-T' script occur before the |
| 2970 | default linker script statements in the internal linker |
| 2971 | representation of the script. In particular, input section |
| 2972 | assignments will be made to `-T' output sections before those in |
| 2973 | the default script. Here is an example of how a `-T' script using |
| 2974 | `INSERT' might look: |
| 2975 | |
| 2976 | SECTIONS |
| 2977 | { |
| 2978 | OVERLAY : |
| 2979 | { |
| 2980 | .ov1 { ov1*(.text) } |
| 2981 | .ov2 { ov2*(.text) } |
| 2982 | } |
| 2983 | } |
| 2984 | INSERT AFTER .text; |
| 2985 | |
| 2986 | `NOCROSSREFS(SECTION SECTION ...)' |
| 2987 | This command may be used to tell `ld' to issue an error about any |
| 2988 | references among certain output sections. |
| 2989 | |
| 2990 | In certain types of programs, particularly on embedded systems when |
| 2991 | using overlays, when one section is loaded into memory, another |
| 2992 | section will not be. Any direct references between the two |
| 2993 | sections would be errors. For example, it would be an error if |
| 2994 | code in one section called a function defined in the other section. |
| 2995 | |
| 2996 | The `NOCROSSREFS' command takes a list of output section names. If |
| 2997 | `ld' detects any cross references between the sections, it reports |
| 2998 | an error and returns a non-zero exit status. Note that the |
| 2999 | `NOCROSSREFS' command uses output section names, not input section |
| 3000 | names. |
| 3001 | |
| 3002 | `NOCROSSREFS_TO(TOSECTION FROMSECTION ...)' |
| 3003 | This command may be used to tell `ld' to issue an error about any |
| 3004 | references to one section from a list of other sections. |
| 3005 | |
| 3006 | The `NOCROSSREFS' command is useful when ensuring that two or more |
| 3007 | output sections are entirely independent but there are situations |
| 3008 | where a one-way dependency is needed. For example, in a multi-core |
| 3009 | application there may be shared code that can be called from each |
| 3010 | core but for safety must never call back. |
| 3011 | |
| 3012 | The `NOCROSSREFS_TO' command takes a list of output section names. |
| 3013 | The first section can not be referenced from any of the other |
| 3014 | sections. If `ld' detects any references to the first section |
| 3015 | from any of the other sections, it reports an error and returns a |
| 3016 | non-zero exit status. Note that the `NOCROSSREFS_TO' command uses |
| 3017 | output section names, not input section names. |
| 3018 | |
| 3019 | `OUTPUT_ARCH(BFDARCH)' |
| 3020 | Specify a particular output machine architecture. The argument is |
| 3021 | one of the names used by the BFD library (*note BFD::). You can |
| 3022 | see the architecture of an object file by using the `objdump' |
| 3023 | program with the `-f' option. |
| 3024 | |
| 3025 | `LD_FEATURE(STRING)' |
| 3026 | This command may be used to modify `ld' behavior. If STRING is |
| 3027 | `"SANE_EXPR"' then absolute symbols and numbers in a script are |
| 3028 | simply treated as numbers everywhere. *Note Expression Section::. |
| 3029 | |
| 3030 | |
| 3031 | File: ld.info, Node: Assignments, Next: SECTIONS, Prev: Simple Commands, Up: Scripts |
| 3032 | |
| 3033 | 3.5 Assigning Values to Symbols |
| 3034 | =============================== |
| 3035 | |
| 3036 | You may assign a value to a symbol in a linker script. This will define |
| 3037 | the symbol and place it into the symbol table with a global scope. |
| 3038 | |
| 3039 | * Menu: |
| 3040 | |
| 3041 | * Simple Assignments:: Simple Assignments |
| 3042 | * HIDDEN:: HIDDEN |
| 3043 | * PROVIDE:: PROVIDE |
| 3044 | * PROVIDE_HIDDEN:: PROVIDE_HIDDEN |
| 3045 | * Source Code Reference:: How to use a linker script defined symbol in source code |
| 3046 | |
| 3047 | |
| 3048 | File: ld.info, Node: Simple Assignments, Next: HIDDEN, Up: Assignments |
| 3049 | |
| 3050 | 3.5.1 Simple Assignments |
| 3051 | ------------------------ |
| 3052 | |
| 3053 | You may assign to a symbol using any of the C assignment operators: |
| 3054 | |
| 3055 | `SYMBOL = EXPRESSION ;' |
| 3056 | `SYMBOL += EXPRESSION ;' |
| 3057 | `SYMBOL -= EXPRESSION ;' |
| 3058 | `SYMBOL *= EXPRESSION ;' |
| 3059 | `SYMBOL /= EXPRESSION ;' |
| 3060 | `SYMBOL <<= EXPRESSION ;' |
| 3061 | `SYMBOL >>= EXPRESSION ;' |
| 3062 | `SYMBOL &= EXPRESSION ;' |
| 3063 | `SYMBOL |= EXPRESSION ;' |
| 3064 | |
| 3065 | The first case will define SYMBOL to the value of EXPRESSION. In |
| 3066 | the other cases, SYMBOL must already be defined, and the value will be |
| 3067 | adjusted accordingly. |
| 3068 | |
| 3069 | The special symbol name `.' indicates the location counter. You may |
| 3070 | only use this within a `SECTIONS' command. *Note Location Counter::. |
| 3071 | |
| 3072 | The semicolon after EXPRESSION is required. |
| 3073 | |
| 3074 | Expressions are defined below; see *Note Expressions::. |
| 3075 | |
| 3076 | You may write symbol assignments as commands in their own right, or |
| 3077 | as statements within a `SECTIONS' command, or as part of an output |
| 3078 | section description in a `SECTIONS' command. |
| 3079 | |
| 3080 | The section of the symbol will be set from the section of the |
| 3081 | expression; for more information, see *Note Expression Section::. |
| 3082 | |
| 3083 | Here is an example showing the three different places that symbol |
| 3084 | assignments may be used: |
| 3085 | |
| 3086 | floating_point = 0; |
| 3087 | SECTIONS |
| 3088 | { |
| 3089 | .text : |
| 3090 | { |
| 3091 | *(.text) |
| 3092 | _etext = .; |
| 3093 | } |
| 3094 | _bdata = (. + 3) & ~ 3; |
| 3095 | .data : { *(.data) } |
| 3096 | } |
| 3097 | In this example, the symbol `floating_point' will be defined as |
| 3098 | zero. The symbol `_etext' will be defined as the address following the |
| 3099 | last `.text' input section. The symbol `_bdata' will be defined as the |
| 3100 | address following the `.text' output section aligned upward to a 4 byte |
| 3101 | boundary. |
| 3102 | |
| 3103 | |
| 3104 | File: ld.info, Node: HIDDEN, Next: PROVIDE, Prev: Simple Assignments, Up: Assignments |
| 3105 | |
| 3106 | 3.5.2 HIDDEN |
| 3107 | ------------ |
| 3108 | |
| 3109 | For ELF targeted ports, define a symbol that will be hidden and won't be |
| 3110 | exported. The syntax is `HIDDEN(SYMBOL = EXPRESSION)'. |
| 3111 | |
| 3112 | Here is the example from *Note Simple Assignments::, rewritten to use |
| 3113 | `HIDDEN': |
| 3114 | |
| 3115 | HIDDEN(floating_point = 0); |
| 3116 | SECTIONS |
| 3117 | { |
| 3118 | .text : |
| 3119 | { |
| 3120 | *(.text) |
| 3121 | HIDDEN(_etext = .); |
| 3122 | } |
| 3123 | HIDDEN(_bdata = (. + 3) & ~ 3); |
| 3124 | .data : { *(.data) } |
| 3125 | } |
| 3126 | In this case none of the three symbols will be visible outside this |
| 3127 | module. |
| 3128 | |
| 3129 | |
| 3130 | File: ld.info, Node: PROVIDE, Next: PROVIDE_HIDDEN, Prev: HIDDEN, Up: Assignments |
| 3131 | |
| 3132 | 3.5.3 PROVIDE |
| 3133 | ------------- |
| 3134 | |
| 3135 | In some cases, it is desirable for a linker script to define a symbol |
| 3136 | only if it is referenced and is not defined by any object included in |
| 3137 | the link. For example, traditional linkers defined the symbol `etext'. |
| 3138 | However, ANSI C requires that the user be able to use `etext' as a |
| 3139 | function name without encountering an error. The `PROVIDE' keyword may |
| 3140 | be used to define a symbol, such as `etext', only if it is referenced |
| 3141 | but not defined. The syntax is `PROVIDE(SYMBOL = EXPRESSION)'. |
| 3142 | |
| 3143 | Here is an example of using `PROVIDE' to define `etext': |
| 3144 | SECTIONS |
| 3145 | { |
| 3146 | .text : |
| 3147 | { |
| 3148 | *(.text) |
| 3149 | _etext = .; |
| 3150 | PROVIDE(etext = .); |
| 3151 | } |
| 3152 | } |
| 3153 | |
| 3154 | In this example, if the program defines `_etext' (with a leading |
| 3155 | underscore), the linker will give a multiple definition error. If, on |
| 3156 | the other hand, the program defines `etext' (with no leading |
| 3157 | underscore), the linker will silently use the definition in the program. |
| 3158 | If the program references `etext' but does not define it, the linker |
| 3159 | will use the definition in the linker script. |
| 3160 | |
| 3161 | |
| 3162 | File: ld.info, Node: PROVIDE_HIDDEN, Next: Source Code Reference, Prev: PROVIDE, Up: Assignments |
| 3163 | |
| 3164 | 3.5.4 PROVIDE_HIDDEN |
| 3165 | -------------------- |
| 3166 | |
| 3167 | Similar to `PROVIDE'. For ELF targeted ports, the symbol will be |
| 3168 | hidden and won't be exported. |
| 3169 | |
| 3170 | |
| 3171 | File: ld.info, Node: Source Code Reference, Prev: PROVIDE_HIDDEN, Up: Assignments |
| 3172 | |
| 3173 | 3.5.5 Source Code Reference |
| 3174 | --------------------------- |
| 3175 | |
| 3176 | Accessing a linker script defined variable from source code is not |
| 3177 | intuitive. In particular a linker script symbol is not equivalent to a |
| 3178 | variable declaration in a high level language, it is instead a symbol |
| 3179 | that does not have a value. |
| 3180 | |
| 3181 | Before going further, it is important to note that compilers often |
| 3182 | transform names in the source code into different names when they are |
| 3183 | stored in the symbol table. For example, Fortran compilers commonly |
| 3184 | prepend or append an underscore, and C++ performs extensive `name |
| 3185 | mangling'. Therefore there might be a discrepancy between the name of |
| 3186 | a variable as it is used in source code and the name of the same |
| 3187 | variable as it is defined in a linker script. For example in C a |
| 3188 | linker script variable might be referred to as: |
| 3189 | |
| 3190 | extern int foo; |
| 3191 | |
| 3192 | But in the linker script it might be defined as: |
| 3193 | |
| 3194 | _foo = 1000; |
| 3195 | |
| 3196 | In the remaining examples however it is assumed that no name |
| 3197 | transformation has taken place. |
| 3198 | |
| 3199 | When a symbol is declared in a high level language such as C, two |
| 3200 | things happen. The first is that the compiler reserves enough space in |
| 3201 | the program's memory to hold the _value_ of the symbol. The second is |
| 3202 | that the compiler creates an entry in the program's symbol table which |
| 3203 | holds the symbol's _address_. ie the symbol table contains the address |
| 3204 | of the block of memory holding the symbol's value. So for example the |
| 3205 | following C declaration, at file scope: |
| 3206 | |
| 3207 | int foo = 1000; |
| 3208 | |
| 3209 | creates an entry called `foo' in the symbol table. This entry holds |
| 3210 | the address of an `int' sized block of memory where the number 1000 is |
| 3211 | initially stored. |
| 3212 | |
| 3213 | When a program references a symbol the compiler generates code that |
| 3214 | first accesses the symbol table to find the address of the symbol's |
| 3215 | memory block and then code to read the value from that memory block. |
| 3216 | So: |
| 3217 | |
| 3218 | foo = 1; |
| 3219 | |
| 3220 | looks up the symbol `foo' in the symbol table, gets the address |
| 3221 | associated with this symbol and then writes the value 1 into that |
| 3222 | address. Whereas: |
| 3223 | |
| 3224 | int * a = & foo; |
| 3225 | |
| 3226 | looks up the symbol `foo' in the symbol table, gets its address and |
| 3227 | then copies this address into the block of memory associated with the |
| 3228 | variable `a'. |
| 3229 | |
| 3230 | Linker scripts symbol declarations, by contrast, create an entry in |
| 3231 | the symbol table but do not assign any memory to them. Thus they are |
| 3232 | an address without a value. So for example the linker script |
| 3233 | definition: |
| 3234 | |
| 3235 | foo = 1000; |
| 3236 | |
| 3237 | creates an entry in the symbol table called `foo' which holds the |
| 3238 | address of memory location 1000, but nothing special is stored at |
| 3239 | address 1000. This means that you cannot access the _value_ of a |
| 3240 | linker script defined symbol - it has no value - all you can do is |
| 3241 | access the _address_ of a linker script defined symbol. |
| 3242 | |
| 3243 | Hence when you are using a linker script defined symbol in source |
| 3244 | code you should always take the address of the symbol, and never |
| 3245 | attempt to use its value. For example suppose you want to copy the |
| 3246 | contents of a section of memory called .ROM into a section called |
| 3247 | .FLASH and the linker script contains these declarations: |
| 3248 | |
| 3249 | start_of_ROM = .ROM; |
| 3250 | end_of_ROM = .ROM + sizeof (.ROM); |
| 3251 | start_of_FLASH = .FLASH; |
| 3252 | |
| 3253 | Then the C source code to perform the copy would be: |
| 3254 | |
| 3255 | extern char start_of_ROM, end_of_ROM, start_of_FLASH; |
| 3256 | |
| 3257 | memcpy (& start_of_FLASH, & start_of_ROM, & end_of_ROM - & start_of_ROM); |
| 3258 | |
| 3259 | Note the use of the `&' operators. These are correct. |
| 3260 | Alternatively the symbols can be treated as the names of vectors or |
| 3261 | arrays and then the code will again work as expected: |
| 3262 | |
| 3263 | extern char start_of_ROM[], end_of_ROM[], start_of_FLASH[]; |
| 3264 | |
| 3265 | memcpy (start_of_FLASH, start_of_ROM, end_of_ROM - start_of_ROM); |
| 3266 | |
| 3267 | Note how using this method does not require the use of `&' operators. |
| 3268 | |
| 3269 | |
| 3270 | File: ld.info, Node: SECTIONS, Next: MEMORY, Prev: Assignments, Up: Scripts |
| 3271 | |
| 3272 | 3.6 SECTIONS Command |
| 3273 | ==================== |
| 3274 | |
| 3275 | The `SECTIONS' command tells the linker how to map input sections into |
| 3276 | output sections, and how to place the output sections in memory. |
| 3277 | |
| 3278 | The format of the `SECTIONS' command is: |
| 3279 | SECTIONS |
| 3280 | { |
| 3281 | SECTIONS-COMMAND |
| 3282 | SECTIONS-COMMAND |
| 3283 | ... |
| 3284 | } |
| 3285 | |
| 3286 | Each SECTIONS-COMMAND may of be one of the following: |
| 3287 | |
| 3288 | * an `ENTRY' command (*note Entry command: Entry Point.) |
| 3289 | |
| 3290 | * a symbol assignment (*note Assignments::) |
| 3291 | |
| 3292 | * an output section description |
| 3293 | |
| 3294 | * an overlay description |
| 3295 | |
| 3296 | The `ENTRY' command and symbol assignments are permitted inside the |
| 3297 | `SECTIONS' command for convenience in using the location counter in |
| 3298 | those commands. This can also make the linker script easier to |
| 3299 | understand because you can use those commands at meaningful points in |
| 3300 | the layout of the output file. |
| 3301 | |
| 3302 | Output section descriptions and overlay descriptions are described |
| 3303 | below. |
| 3304 | |
| 3305 | If you do not use a `SECTIONS' command in your linker script, the |
| 3306 | linker will place each input section into an identically named output |
| 3307 | section in the order that the sections are first encountered in the |
| 3308 | input files. If all input sections are present in the first file, for |
| 3309 | example, the order of sections in the output file will match the order |
| 3310 | in the first input file. The first section will be at address zero. |
| 3311 | |
| 3312 | * Menu: |
| 3313 | |
| 3314 | * Output Section Description:: Output section description |
| 3315 | * Output Section Name:: Output section name |
| 3316 | * Output Section Address:: Output section address |
| 3317 | * Input Section:: Input section description |
| 3318 | * Output Section Data:: Output section data |
| 3319 | * Output Section Keywords:: Output section keywords |
| 3320 | * Output Section Discarding:: Output section discarding |
| 3321 | * Output Section Attributes:: Output section attributes |
| 3322 | * Overlay Description:: Overlay description |
| 3323 | |
| 3324 | |
| 3325 | File: ld.info, Node: Output Section Description, Next: Output Section Name, Up: SECTIONS |
| 3326 | |
| 3327 | 3.6.1 Output Section Description |
| 3328 | -------------------------------- |
| 3329 | |
| 3330 | The full description of an output section looks like this: |
| 3331 | SECTION [ADDRESS] [(TYPE)] : |
| 3332 | [AT(LMA)] |
| 3333 | [ALIGN(SECTION_ALIGN) | ALIGN_WITH_INPUT] |
| 3334 | [SUBALIGN(SUBSECTION_ALIGN)] |
| 3335 | [CONSTRAINT] |
| 3336 | { |
| 3337 | OUTPUT-SECTION-COMMAND |
| 3338 | OUTPUT-SECTION-COMMAND |
| 3339 | ... |
| 3340 | } [>REGION] [AT>LMA_REGION] [:PHDR :PHDR ...] [=FILLEXP] [,] |
| 3341 | |
| 3342 | Most output sections do not use most of the optional section |
| 3343 | attributes. |
| 3344 | |
| 3345 | The whitespace around SECTION is required, so that the section name |
| 3346 | is unambiguous. The colon and the curly braces are also required. The |
| 3347 | comma at the end may be required if a FILLEXP is used and the next |
| 3348 | SECTIONS-COMMAND looks like a continuation of the expression. The line |
| 3349 | breaks and other white space are optional. |
| 3350 | |
| 3351 | Each OUTPUT-SECTION-COMMAND may be one of the following: |
| 3352 | |
| 3353 | * a symbol assignment (*note Assignments::) |
| 3354 | |
| 3355 | * an input section description (*note Input Section::) |
| 3356 | |
| 3357 | * data values to include directly (*note Output Section Data::) |
| 3358 | |
| 3359 | * a special output section keyword (*note Output Section Keywords::) |
| 3360 | |
| 3361 | |
| 3362 | File: ld.info, Node: Output Section Name, Next: Output Section Address, Prev: Output Section Description, Up: SECTIONS |
| 3363 | |
| 3364 | 3.6.2 Output Section Name |
| 3365 | ------------------------- |
| 3366 | |
| 3367 | The name of the output section is SECTION. SECTION must meet the |
| 3368 | constraints of your output format. In formats which only support a |
| 3369 | limited number of sections, such as `a.out', the name must be one of |
| 3370 | the names supported by the format (`a.out', for example, allows only |
| 3371 | `.text', `.data' or `.bss'). If the output format supports any number |
| 3372 | of sections, but with numbers and not names (as is the case for Oasys), |
| 3373 | the name should be supplied as a quoted numeric string. A section name |
| 3374 | may consist of any sequence of characters, but a name which contains |
| 3375 | any unusual characters such as commas must be quoted. |
| 3376 | |
| 3377 | The output section name `/DISCARD/' is special; *Note Output Section |
| 3378 | Discarding::. |
| 3379 | |
| 3380 | |
| 3381 | File: ld.info, Node: Output Section Address, Next: Input Section, Prev: Output Section Name, Up: SECTIONS |
| 3382 | |
| 3383 | 3.6.3 Output Section Address |
| 3384 | ---------------------------- |
| 3385 | |
| 3386 | The ADDRESS is an expression for the VMA (the virtual memory address) |
| 3387 | of the output section. This address is optional, but if it is provided |
| 3388 | then the output address will be set exactly as specified. |
| 3389 | |
| 3390 | If the output address is not specified then one will be chosen for |
| 3391 | the section, based on the heuristic below. This address will be |
| 3392 | adjusted to fit the alignment requirement of the output section. The |
| 3393 | alignment requirement is the strictest alignment of any input section |
| 3394 | contained within the output section. |
| 3395 | |
| 3396 | The output section address heuristic is as follows: |
| 3397 | |
| 3398 | * If an output memory REGION is set for the section then it is added |
| 3399 | to this region and its address will be the next free address in |
| 3400 | that region. |
| 3401 | |
| 3402 | * If the MEMORY command has been used to create a list of memory |
| 3403 | regions then the first region which has attributes compatible with |
| 3404 | the section is selected to contain it. The section's output |
| 3405 | address will be the next free address in that region; *Note |
| 3406 | MEMORY::. |
| 3407 | |
| 3408 | * If no memory regions were specified, or none match the section then |
| 3409 | the output address will be based on the current value of the |
| 3410 | location counter. |
| 3411 | |
| 3412 | For example: |
| 3413 | |
| 3414 | .text . : { *(.text) } |
| 3415 | |
| 3416 | and |
| 3417 | |
| 3418 | .text : { *(.text) } |
| 3419 | |
| 3420 | are subtly different. The first will set the address of the `.text' |
| 3421 | output section to the current value of the location counter. The |
| 3422 | second will set it to the current value of the location counter aligned |
| 3423 | to the strictest alignment of any of the `.text' input sections. |
| 3424 | |
| 3425 | The ADDRESS may be an arbitrary expression; *Note Expressions::. |
| 3426 | For example, if you want to align the section on a 0x10 byte boundary, |
| 3427 | so that the lowest four bits of the section address are zero, you could |
| 3428 | do something like this: |
| 3429 | .text ALIGN(0x10) : { *(.text) } |
| 3430 | This works because `ALIGN' returns the current location counter |
| 3431 | aligned upward to the specified value. |
| 3432 | |
| 3433 | Specifying ADDRESS for a section will change the value of the |
| 3434 | location counter, provided that the section is non-empty. (Empty |
| 3435 | sections are ignored). |
| 3436 | |
| 3437 | |
| 3438 | File: ld.info, Node: Input Section, Next: Output Section Data, Prev: Output Section Address, Up: SECTIONS |
| 3439 | |
| 3440 | 3.6.4 Input Section Description |
| 3441 | ------------------------------- |
| 3442 | |
| 3443 | The most common output section command is an input section description. |
| 3444 | |
| 3445 | The input section description is the most basic linker script |
| 3446 | operation. You use output sections to tell the linker how to lay out |
| 3447 | your program in memory. You use input section descriptions to tell the |
| 3448 | linker how to map the input files into your memory layout. |
| 3449 | |
| 3450 | * Menu: |
| 3451 | |
| 3452 | * Input Section Basics:: Input section basics |
| 3453 | * Input Section Wildcards:: Input section wildcard patterns |
| 3454 | * Input Section Common:: Input section for common symbols |
| 3455 | * Input Section Keep:: Input section and garbage collection |
| 3456 | * Input Section Example:: Input section example |
| 3457 | |
| 3458 | |
| 3459 | File: ld.info, Node: Input Section Basics, Next: Input Section Wildcards, Up: Input Section |
| 3460 | |
| 3461 | 3.6.4.1 Input Section Basics |
| 3462 | ............................ |
| 3463 | |
| 3464 | An input section description consists of a file name optionally followed |
| 3465 | by a list of section names in parentheses. |
| 3466 | |
| 3467 | The file name and the section name may be wildcard patterns, which we |
| 3468 | describe further below (*note Input Section Wildcards::). |
| 3469 | |
| 3470 | The most common input section description is to include all input |
| 3471 | sections with a particular name in the output section. For example, to |
| 3472 | include all input `.text' sections, you would write: |
| 3473 | *(.text) |
| 3474 | Here the `*' is a wildcard which matches any file name. To exclude |
| 3475 | a list of files from matching the file name wildcard, EXCLUDE_FILE may |
| 3476 | be used to match all files except the ones specified in the |
| 3477 | EXCLUDE_FILE list. For example: |
| 3478 | *(EXCLUDE_FILE (*crtend.o *otherfile.o) .ctors) |
| 3479 | will cause all .ctors sections from all files except `crtend.o' and |
| 3480 | `otherfile.o' to be included. |
| 3481 | |
| 3482 | There are two ways to include more than one section: |
| 3483 | *(.text .rdata) |
| 3484 | *(.text) *(.rdata) |
| 3485 | The difference between these is the order in which the `.text' and |
| 3486 | `.rdata' input sections will appear in the output section. In the |
| 3487 | first example, they will be intermingled, appearing in the same order as |
| 3488 | they are found in the linker input. In the second example, all `.text' |
| 3489 | input sections will appear first, followed by all `.rdata' input |
| 3490 | sections. |
| 3491 | |
| 3492 | You can specify a file name to include sections from a particular |
| 3493 | file. You would do this if one or more of your files contain special |
| 3494 | data that needs to be at a particular location in memory. For example: |
| 3495 | data.o(.data) |
| 3496 | |
| 3497 | To refine the sections that are included based on the section flags |
| 3498 | of an input section, INPUT_SECTION_FLAGS may be used. |
| 3499 | |
| 3500 | Here is a simple example for using Section header flags for ELF |
| 3501 | sections: |
| 3502 | |
| 3503 | SECTIONS { |
| 3504 | .text : { INPUT_SECTION_FLAGS (SHF_MERGE & SHF_STRINGS) *(.text) } |
| 3505 | .text2 : { INPUT_SECTION_FLAGS (!SHF_WRITE) *(.text) } |
| 3506 | } |
| 3507 | |
| 3508 | In this example, the output section `.text' will be comprised of any |
| 3509 | input section matching the name *(.text) whose section header flags |
| 3510 | `SHF_MERGE' and `SHF_STRINGS' are set. The output section `.text2' |
| 3511 | will be comprised of any input section matching the name *(.text) whose |
| 3512 | section header flag `SHF_WRITE' is clear. |
| 3513 | |
| 3514 | You can also specify files within archives by writing a pattern |
| 3515 | matching the archive, a colon, then the pattern matching the file, with |
| 3516 | no whitespace around the colon. |
| 3517 | |
| 3518 | `archive:file' |
| 3519 | matches file within archive |
| 3520 | |
| 3521 | `archive:' |
| 3522 | matches the whole archive |
| 3523 | |
| 3524 | `:file' |
| 3525 | matches file but not one in an archive |
| 3526 | |
| 3527 | Either one or both of `archive' and `file' can contain shell |
| 3528 | wildcards. On DOS based file systems, the linker will assume that a |
| 3529 | single letter followed by a colon is a drive specifier, so `c:myfile.o' |
| 3530 | is a simple file specification, not `myfile.o' within an archive called |
| 3531 | `c'. `archive:file' filespecs may also be used within an |
| 3532 | `EXCLUDE_FILE' list, but may not appear in other linker script |
| 3533 | contexts. For instance, you cannot extract a file from an archive by |
| 3534 | using `archive:file' in an `INPUT' command. |
| 3535 | |
| 3536 | If you use a file name without a list of sections, then all sections |
| 3537 | in the input file will be included in the output section. This is not |
| 3538 | commonly done, but it may by useful on occasion. For example: |
| 3539 | data.o |
| 3540 | |
| 3541 | When you use a file name which is not an `archive:file' specifier |
| 3542 | and does not contain any wild card characters, the linker will first |
| 3543 | see if you also specified the file name on the linker command line or |
| 3544 | in an `INPUT' command. If you did not, the linker will attempt to open |
| 3545 | the file as an input file, as though it appeared on the command line. |
| 3546 | Note that this differs from an `INPUT' command, because the linker will |
| 3547 | not search for the file in the archive search path. |
| 3548 | |
| 3549 | |
| 3550 | File: ld.info, Node: Input Section Wildcards, Next: Input Section Common, Prev: Input Section Basics, Up: Input Section |
| 3551 | |
| 3552 | 3.6.4.2 Input Section Wildcard Patterns |
| 3553 | ....................................... |
| 3554 | |
| 3555 | In an input section description, either the file name or the section |
| 3556 | name or both may be wildcard patterns. |
| 3557 | |
| 3558 | The file name of `*' seen in many examples is a simple wildcard |
| 3559 | pattern for the file name. |
| 3560 | |
| 3561 | The wildcard patterns are like those used by the Unix shell. |
| 3562 | |
| 3563 | `*' |
| 3564 | matches any number of characters |
| 3565 | |
| 3566 | `?' |
| 3567 | matches any single character |
| 3568 | |
| 3569 | `[CHARS]' |
| 3570 | matches a single instance of any of the CHARS; the `-' character |
| 3571 | may be used to specify a range of characters, as in `[a-z]' to |
| 3572 | match any lower case letter |
| 3573 | |
| 3574 | `\' |
| 3575 | quotes the following character |
| 3576 | |
| 3577 | When a file name is matched with a wildcard, the wildcard characters |
| 3578 | will not match a `/' character (used to separate directory names on |
| 3579 | Unix). A pattern consisting of a single `*' character is an exception; |
| 3580 | it will always match any file name, whether it contains a `/' or not. |
| 3581 | In a section name, the wildcard characters will match a `/' character. |
| 3582 | |
| 3583 | File name wildcard patterns only match files which are explicitly |
| 3584 | specified on the command line or in an `INPUT' command. The linker |
| 3585 | does not search directories to expand wildcards. |
| 3586 | |
| 3587 | If a file name matches more than one wildcard pattern, or if a file |
| 3588 | name appears explicitly and is also matched by a wildcard pattern, the |
| 3589 | linker will use the first match in the linker script. For example, this |
| 3590 | sequence of input section descriptions is probably in error, because the |
| 3591 | `data.o' rule will not be used: |
| 3592 | .data : { *(.data) } |
| 3593 | .data1 : { data.o(.data) } |
| 3594 | |
| 3595 | Normally, the linker will place files and sections matched by |
| 3596 | wildcards in the order in which they are seen during the link. You can |
| 3597 | change this by using the `SORT_BY_NAME' keyword, which appears before a |
| 3598 | wildcard pattern in parentheses (e.g., `SORT_BY_NAME(.text*)'). When |
| 3599 | the `SORT_BY_NAME' keyword is used, the linker will sort the files or |
| 3600 | sections into ascending order by name before placing them in the output |
| 3601 | file. |
| 3602 | |
| 3603 | `SORT_BY_ALIGNMENT' is very similar to `SORT_BY_NAME'. The |
| 3604 | difference is `SORT_BY_ALIGNMENT' will sort sections into descending |
| 3605 | order by alignment before placing them in the output file. Larger |
| 3606 | alignments are placed before smaller alignments in order to reduce the |
| 3607 | amount of padding necessary. |
| 3608 | |
| 3609 | `SORT_BY_INIT_PRIORITY' is very similar to `SORT_BY_NAME'. The |
| 3610 | difference is `SORT_BY_INIT_PRIORITY' will sort sections into ascending |
| 3611 | order by numerical value of the GCC init_priority attribute encoded in |
| 3612 | the section name before placing them in the output file. |
| 3613 | |
| 3614 | `SORT' is an alias for `SORT_BY_NAME'. |
| 3615 | |
| 3616 | When there are nested section sorting commands in linker script, |
| 3617 | there can be at most 1 level of nesting for section sorting commands. |
| 3618 | |
| 3619 | 1. `SORT_BY_NAME' (`SORT_BY_ALIGNMENT' (wildcard section pattern)). |
| 3620 | It will sort the input sections by name first, then by alignment |
| 3621 | if two sections have the same name. |
| 3622 | |
| 3623 | 2. `SORT_BY_ALIGNMENT' (`SORT_BY_NAME' (wildcard section pattern)). |
| 3624 | It will sort the input sections by alignment first, then by name |
| 3625 | if two sections have the same alignment. |
| 3626 | |
| 3627 | 3. `SORT_BY_NAME' (`SORT_BY_NAME' (wildcard section pattern)) is |
| 3628 | treated the same as `SORT_BY_NAME' (wildcard section pattern). |
| 3629 | |
| 3630 | 4. `SORT_BY_ALIGNMENT' (`SORT_BY_ALIGNMENT' (wildcard section |
| 3631 | pattern)) is treated the same as `SORT_BY_ALIGNMENT' (wildcard |
| 3632 | section pattern). |
| 3633 | |
| 3634 | 5. All other nested section sorting commands are invalid. |
| 3635 | |
| 3636 | When both command line section sorting option and linker script |
| 3637 | section sorting command are used, section sorting command always takes |
| 3638 | precedence over the command line option. |
| 3639 | |
| 3640 | If the section sorting command in linker script isn't nested, the |
| 3641 | command line option will make the section sorting command to be treated |
| 3642 | as nested sorting command. |
| 3643 | |
| 3644 | 1. `SORT_BY_NAME' (wildcard section pattern ) with `--sort-sections |
| 3645 | alignment' is equivalent to `SORT_BY_NAME' (`SORT_BY_ALIGNMENT' |
| 3646 | (wildcard section pattern)). |
| 3647 | |
| 3648 | 2. `SORT_BY_ALIGNMENT' (wildcard section pattern) with |
| 3649 | `--sort-section name' is equivalent to `SORT_BY_ALIGNMENT' |
| 3650 | (`SORT_BY_NAME' (wildcard section pattern)). |
| 3651 | |
| 3652 | If the section sorting command in linker script is nested, the |
| 3653 | command line option will be ignored. |
| 3654 | |
| 3655 | `SORT_NONE' disables section sorting by ignoring the command line |
| 3656 | section sorting option. |
| 3657 | |
| 3658 | If you ever get confused about where input sections are going, use |
| 3659 | the `-M' linker option to generate a map file. The map file shows |
| 3660 | precisely how input sections are mapped to output sections. |
| 3661 | |
| 3662 | This example shows how wildcard patterns might be used to partition |
| 3663 | files. This linker script directs the linker to place all `.text' |
| 3664 | sections in `.text' and all `.bss' sections in `.bss'. The linker will |
| 3665 | place the `.data' section from all files beginning with an upper case |
| 3666 | character in `.DATA'; for all other files, the linker will place the |
| 3667 | `.data' section in `.data'. |
| 3668 | SECTIONS { |
| 3669 | .text : { *(.text) } |
| 3670 | .DATA : { [A-Z]*(.data) } |
| 3671 | .data : { *(.data) } |
| 3672 | .bss : { *(.bss) } |
| 3673 | } |
| 3674 | |
| 3675 | |
| 3676 | File: ld.info, Node: Input Section Common, Next: Input Section Keep, Prev: Input Section Wildcards, Up: Input Section |
| 3677 | |
| 3678 | 3.6.4.3 Input Section for Common Symbols |
| 3679 | ........................................ |
| 3680 | |
| 3681 | A special notation is needed for common symbols, because in many object |
| 3682 | file formats common symbols do not have a particular input section. The |
| 3683 | linker treats common symbols as though they are in an input section |
| 3684 | named `COMMON'. |
| 3685 | |
| 3686 | You may use file names with the `COMMON' section just as with any |
| 3687 | other input sections. You can use this to place common symbols from a |
| 3688 | particular input file in one section while common symbols from other |
| 3689 | input files are placed in another section. |
| 3690 | |
| 3691 | In most cases, common symbols in input files will be placed in the |
| 3692 | `.bss' section in the output file. For example: |
| 3693 | .bss { *(.bss) *(COMMON) } |
| 3694 | |
| 3695 | Some object file formats have more than one type of common symbol. |
| 3696 | For example, the MIPS ELF object file format distinguishes standard |
| 3697 | common symbols and small common symbols. In this case, the linker will |
| 3698 | use a different special section name for other types of common symbols. |
| 3699 | In the case of MIPS ELF, the linker uses `COMMON' for standard common |
| 3700 | symbols and `.scommon' for small common symbols. This permits you to |
| 3701 | map the different types of common symbols into memory at different |
| 3702 | locations. |
| 3703 | |
| 3704 | You will sometimes see `[COMMON]' in old linker scripts. This |
| 3705 | notation is now considered obsolete. It is equivalent to `*(COMMON)'. |
| 3706 | |
| 3707 | |
| 3708 | File: ld.info, Node: Input Section Keep, Next: Input Section Example, Prev: Input Section Common, Up: Input Section |
| 3709 | |
| 3710 | 3.6.4.4 Input Section and Garbage Collection |
| 3711 | ............................................ |
| 3712 | |
| 3713 | When link-time garbage collection is in use (`--gc-sections'), it is |
| 3714 | often useful to mark sections that should not be eliminated. This is |
| 3715 | accomplished by surrounding an input section's wildcard entry with |
| 3716 | `KEEP()', as in `KEEP(*(.init))' or `KEEP(SORT_BY_NAME(*)(.ctors))'. |
| 3717 | |
| 3718 | |
| 3719 | File: ld.info, Node: Input Section Example, Prev: Input Section Keep, Up: Input Section |
| 3720 | |
| 3721 | 3.6.4.5 Input Section Example |
| 3722 | ............................. |
| 3723 | |
| 3724 | The following example is a complete linker script. It tells the linker |
| 3725 | to read all of the sections from file `all.o' and place them at the |
| 3726 | start of output section `outputa' which starts at location `0x10000'. |
| 3727 | All of section `.input1' from file `foo.o' follows immediately, in the |
| 3728 | same output section. All of section `.input2' from `foo.o' goes into |
| 3729 | output section `outputb', followed by section `.input1' from `foo1.o'. |
| 3730 | All of the remaining `.input1' and `.input2' sections from any files |
| 3731 | are written to output section `outputc'. |
| 3732 | |
| 3733 | SECTIONS { |
| 3734 | outputa 0x10000 : |
| 3735 | { |
| 3736 | all.o |
| 3737 | foo.o (.input1) |
| 3738 | } |
| 3739 | outputb : |
| 3740 | { |
| 3741 | foo.o (.input2) |
| 3742 | foo1.o (.input1) |
| 3743 | } |
| 3744 | outputc : |
| 3745 | { |
| 3746 | *(.input1) |
| 3747 | *(.input2) |
| 3748 | } |
| 3749 | } |
| 3750 | |
| 3751 | |
| 3752 | File: ld.info, Node: Output Section Data, Next: Output Section Keywords, Prev: Input Section, Up: SECTIONS |
| 3753 | |
| 3754 | 3.6.5 Output Section Data |
| 3755 | ------------------------- |
| 3756 | |
| 3757 | You can include explicit bytes of data in an output section by using |
| 3758 | `BYTE', `SHORT', `LONG', `QUAD', or `SQUAD' as an output section |
| 3759 | command. Each keyword is followed by an expression in parentheses |
| 3760 | providing the value to store (*note Expressions::). The value of the |
| 3761 | expression is stored at the current value of the location counter. |
| 3762 | |
| 3763 | The `BYTE', `SHORT', `LONG', and `QUAD' commands store one, two, |
| 3764 | four, and eight bytes (respectively). After storing the bytes, the |
| 3765 | location counter is incremented by the number of bytes stored. |
| 3766 | |
| 3767 | For example, this will store the byte 1 followed by the four byte |
| 3768 | value of the symbol `addr': |
| 3769 | BYTE(1) |
| 3770 | LONG(addr) |
| 3771 | |
| 3772 | When using a 64 bit host or target, `QUAD' and `SQUAD' are the same; |
| 3773 | they both store an 8 byte, or 64 bit, value. When both host and target |
| 3774 | are 32 bits, an expression is computed as 32 bits. In this case `QUAD' |
| 3775 | stores a 32 bit value zero extended to 64 bits, and `SQUAD' stores a 32 |
| 3776 | bit value sign extended to 64 bits. |
| 3777 | |
| 3778 | If the object file format of the output file has an explicit |
| 3779 | endianness, which is the normal case, the value will be stored in that |
| 3780 | endianness. When the object file format does not have an explicit |
| 3781 | endianness, as is true of, for example, S-records, the value will be |
| 3782 | stored in the endianness of the first input object file. |
| 3783 | |
| 3784 | Note--these commands only work inside a section description and not |
| 3785 | between them, so the following will produce an error from the linker: |
| 3786 | SECTIONS { .text : { *(.text) } LONG(1) .data : { *(.data) } } |
| 3787 | whereas this will work: |
| 3788 | SECTIONS { .text : { *(.text) ; LONG(1) } .data : { *(.data) } } |
| 3789 | |
| 3790 | You may use the `FILL' command to set the fill pattern for the |
| 3791 | current section. It is followed by an expression in parentheses. Any |
| 3792 | otherwise unspecified regions of memory within the section (for example, |
| 3793 | gaps left due to the required alignment of input sections) are filled |
| 3794 | with the value of the expression, repeated as necessary. A `FILL' |
| 3795 | statement covers memory locations after the point at which it occurs in |
| 3796 | the section definition; by including more than one `FILL' statement, |
| 3797 | you can have different fill patterns in different parts of an output |
| 3798 | section. |
| 3799 | |
| 3800 | This example shows how to fill unspecified regions of memory with the |
| 3801 | value `0x90': |
| 3802 | FILL(0x90909090) |
| 3803 | |
| 3804 | The `FILL' command is similar to the `=FILLEXP' output section |
| 3805 | attribute, but it only affects the part of the section following the |
| 3806 | `FILL' command, rather than the entire section. If both are used, the |
| 3807 | `FILL' command takes precedence. *Note Output Section Fill::, for |
| 3808 | details on the fill expression. |
| 3809 | |
| 3810 | |
| 3811 | File: ld.info, Node: Output Section Keywords, Next: Output Section Discarding, Prev: Output Section Data, Up: SECTIONS |
| 3812 | |
| 3813 | 3.6.6 Output Section Keywords |
| 3814 | ----------------------------- |
| 3815 | |
| 3816 | There are a couple of keywords which can appear as output section |
| 3817 | commands. |
| 3818 | |
| 3819 | `CREATE_OBJECT_SYMBOLS' |
| 3820 | The command tells the linker to create a symbol for each input |
| 3821 | file. The name of each symbol will be the name of the |
| 3822 | corresponding input file. The section of each symbol will be the |
| 3823 | output section in which the `CREATE_OBJECT_SYMBOLS' command |
| 3824 | appears. |
| 3825 | |
| 3826 | This is conventional for the a.out object file format. It is not |
| 3827 | normally used for any other object file format. |
| 3828 | |
| 3829 | `CONSTRUCTORS' |
| 3830 | When linking using the a.out object file format, the linker uses an |
| 3831 | unusual set construct to support C++ global constructors and |
| 3832 | destructors. When linking object file formats which do not support |
| 3833 | arbitrary sections, such as ECOFF and XCOFF, the linker will |
| 3834 | automatically recognize C++ global constructors and destructors by |
| 3835 | name. For these object file formats, the `CONSTRUCTORS' command |
| 3836 | tells the linker to place constructor information in the output |
| 3837 | section where the `CONSTRUCTORS' command appears. The |
| 3838 | `CONSTRUCTORS' command is ignored for other object file formats. |
| 3839 | |
| 3840 | The symbol `__CTOR_LIST__' marks the start of the global |
| 3841 | constructors, and the symbol `__CTOR_END__' marks the end. |
| 3842 | Similarly, `__DTOR_LIST__' and `__DTOR_END__' mark the start and |
| 3843 | end of the global destructors. The first word in the list is the |
| 3844 | number of entries, followed by the address of each constructor or |
| 3845 | destructor, followed by a zero word. The compiler must arrange to |
| 3846 | actually run the code. For these object file formats GNU C++ |
| 3847 | normally calls constructors from a subroutine `__main'; a call to |
| 3848 | `__main' is automatically inserted into the startup code for |
| 3849 | `main'. GNU C++ normally runs destructors either by using |
| 3850 | `atexit', or directly from the function `exit'. |
| 3851 | |
| 3852 | For object file formats such as `COFF' or `ELF' which support |
| 3853 | arbitrary section names, GNU C++ will normally arrange to put the |
| 3854 | addresses of global constructors and destructors into the `.ctors' |
| 3855 | and `.dtors' sections. Placing the following sequence into your |
| 3856 | linker script will build the sort of table which the GNU C++ |
| 3857 | runtime code expects to see. |
| 3858 | |
| 3859 | __CTOR_LIST__ = .; |
| 3860 | LONG((__CTOR_END__ - __CTOR_LIST__) / 4 - 2) |
| 3861 | *(.ctors) |
| 3862 | LONG(0) |
| 3863 | __CTOR_END__ = .; |
| 3864 | __DTOR_LIST__ = .; |
| 3865 | LONG((__DTOR_END__ - __DTOR_LIST__) / 4 - 2) |
| 3866 | *(.dtors) |
| 3867 | LONG(0) |
| 3868 | __DTOR_END__ = .; |
| 3869 | |
| 3870 | If you are using the GNU C++ support for initialization priority, |
| 3871 | which provides some control over the order in which global |
| 3872 | constructors are run, you must sort the constructors at link time |
| 3873 | to ensure that they are executed in the correct order. When using |
| 3874 | the `CONSTRUCTORS' command, use `SORT_BY_NAME(CONSTRUCTORS)' |
| 3875 | instead. When using the `.ctors' and `.dtors' sections, use |
| 3876 | `*(SORT_BY_NAME(.ctors))' and `*(SORT_BY_NAME(.dtors))' instead of |
| 3877 | just `*(.ctors)' and `*(.dtors)'. |
| 3878 | |
| 3879 | Normally the compiler and linker will handle these issues |
| 3880 | automatically, and you will not need to concern yourself with |
| 3881 | them. However, you may need to consider this if you are using C++ |
| 3882 | and writing your own linker scripts. |
| 3883 | |
| 3884 | |
| 3885 | |
| 3886 | File: ld.info, Node: Output Section Discarding, Next: Output Section Attributes, Prev: Output Section Keywords, Up: SECTIONS |
| 3887 | |
| 3888 | 3.6.7 Output Section Discarding |
| 3889 | ------------------------------- |
| 3890 | |
| 3891 | The linker will not normally create output sections with no contents. |
| 3892 | This is for convenience when referring to input sections that may or |
| 3893 | may not be present in any of the input files. For example: |
| 3894 | .foo : { *(.foo) } |
| 3895 | will only create a `.foo' section in the output file if there is a |
| 3896 | `.foo' section in at least one input file, and if the input sections |
| 3897 | are not all empty. Other link script directives that allocate space in |
| 3898 | an output section will also create the output section. So too will |
| 3899 | assignments to dot even if the assignment does not create space, except |
| 3900 | for `. = 0', `. = . + 0', `. = sym', `. = . + sym' and `. = ALIGN (. != |
| 3901 | 0, expr, 1)' when `sym' is an absolute symbol of value 0 defined in the |
| 3902 | script. This allows you to force output of an empty section with `. = |
| 3903 | .'. |
| 3904 | |
| 3905 | The linker will ignore address assignments (*note Output Section |
| 3906 | Address::) on discarded output sections, except when the linker script |
| 3907 | defines symbols in the output section. In that case the linker will |
| 3908 | obey the address assignments, possibly advancing dot even though the |
| 3909 | section is discarded. |
| 3910 | |
| 3911 | The special output section name `/DISCARD/' may be used to discard |
| 3912 | input sections. Any input sections which are assigned to an output |
| 3913 | section named `/DISCARD/' are not included in the output file. |
| 3914 | |
| 3915 | |
| 3916 | File: ld.info, Node: Output Section Attributes, Next: Overlay Description, Prev: Output Section Discarding, Up: SECTIONS |
| 3917 | |
| 3918 | 3.6.8 Output Section Attributes |
| 3919 | ------------------------------- |
| 3920 | |
| 3921 | We showed above that the full description of an output section looked |
| 3922 | like this: |
| 3923 | |
| 3924 | SECTION [ADDRESS] [(TYPE)] : |
| 3925 | [AT(LMA)] |
| 3926 | [ALIGN(SECTION_ALIGN)] |
| 3927 | [SUBALIGN(SUBSECTION_ALIGN)] |
| 3928 | [CONSTRAINT] |
| 3929 | { |
| 3930 | OUTPUT-SECTION-COMMAND |
| 3931 | OUTPUT-SECTION-COMMAND |
| 3932 | ... |
| 3933 | } [>REGION] [AT>LMA_REGION] [:PHDR :PHDR ...] [=FILLEXP] |
| 3934 | |
| 3935 | We've already described SECTION, ADDRESS, and |
| 3936 | OUTPUT-SECTION-COMMAND. In this section we will describe the remaining |
| 3937 | section attributes. |
| 3938 | |
| 3939 | * Menu: |
| 3940 | |
| 3941 | * Output Section Type:: Output section type |
| 3942 | * Output Section LMA:: Output section LMA |
| 3943 | * Forced Output Alignment:: Forced Output Alignment |
| 3944 | * Forced Input Alignment:: Forced Input Alignment |
| 3945 | * Output Section Constraint:: Output section constraint |
| 3946 | * Output Section Region:: Output section region |
| 3947 | * Output Section Phdr:: Output section phdr |
| 3948 | * Output Section Fill:: Output section fill |
| 3949 | |
| 3950 | |
| 3951 | File: ld.info, Node: Output Section Type, Next: Output Section LMA, Up: Output Section Attributes |
| 3952 | |
| 3953 | 3.6.8.1 Output Section Type |
| 3954 | ........................... |
| 3955 | |
| 3956 | Each output section may have a type. The type is a keyword in |
| 3957 | parentheses. The following types are defined: |
| 3958 | |
| 3959 | `NOLOAD' |
| 3960 | The section should be marked as not loadable, so that it will not |
| 3961 | be loaded into memory when the program is run. |
| 3962 | |
| 3963 | `DSECT' |
| 3964 | `COPY' |
| 3965 | `INFO' |
| 3966 | `OVERLAY' |
| 3967 | These type names are supported for backward compatibility, and are |
| 3968 | rarely used. They all have the same effect: the section should be |
| 3969 | marked as not allocatable, so that no memory is allocated for the |
| 3970 | section when the program is run. |
| 3971 | |
| 3972 | The linker normally sets the attributes of an output section based on |
| 3973 | the input sections which map into it. You can override this by using |
| 3974 | the section type. For example, in the script sample below, the `ROM' |
| 3975 | section is addressed at memory location `0' and does not need to be |
| 3976 | loaded when the program is run. |
| 3977 | SECTIONS { |
| 3978 | ROM 0 (NOLOAD) : { ... } |
| 3979 | ... |
| 3980 | } |
| 3981 | |
| 3982 | |
| 3983 | File: ld.info, Node: Output Section LMA, Next: Forced Output Alignment, Prev: Output Section Type, Up: Output Section Attributes |
| 3984 | |
| 3985 | 3.6.8.2 Output Section LMA |
| 3986 | .......................... |
| 3987 | |
| 3988 | Every section has a virtual address (VMA) and a load address (LMA); see |
| 3989 | *Note Basic Script Concepts::. The virtual address is specified by the |
| 3990 | *note Output Section Address:: described earlier. The load address is |
| 3991 | specified by the `AT' or `AT>' keywords. Specifying a load address is |
| 3992 | optional. |
| 3993 | |
| 3994 | The `AT' keyword takes an expression as an argument. This specifies |
| 3995 | the exact load address of the section. The `AT>' keyword takes the |
| 3996 | name of a memory region as an argument. *Note MEMORY::. The load |
| 3997 | address of the section is set to the next free address in the region, |
| 3998 | aligned to the section's alignment requirements. |
| 3999 | |
| 4000 | If neither `AT' nor `AT>' is specified for an allocatable section, |
| 4001 | the linker will use the following heuristic to determine the load |
| 4002 | address: |
| 4003 | |
| 4004 | * If the section has a specific VMA address, then this is used as |
| 4005 | the LMA address as well. |
| 4006 | |
| 4007 | * If the section is not allocatable then its LMA is set to its VMA. |
| 4008 | |
| 4009 | * Otherwise if a memory region can be found that is compatible with |
| 4010 | the current section, and this region contains at least one |
| 4011 | section, then the LMA is set so the difference between the VMA and |
| 4012 | LMA is the same as the difference between the VMA and LMA of the |
| 4013 | last section in the located region. |
| 4014 | |
| 4015 | * If no memory regions have been declared then a default region that |
| 4016 | covers the entire address space is used in the previous step. |
| 4017 | |
| 4018 | * If no suitable region could be found, or there was no previous |
| 4019 | section then the LMA is set equal to the VMA. |
| 4020 | |
| 4021 | This feature is designed to make it easy to build a ROM image. For |
| 4022 | example, the following linker script creates three output sections: one |
| 4023 | called `.text', which starts at `0x1000', one called `.mdata', which is |
| 4024 | loaded at the end of the `.text' section even though its VMA is |
| 4025 | `0x2000', and one called `.bss' to hold uninitialized data at address |
| 4026 | `0x3000'. The symbol `_data' is defined with the value `0x2000', which |
| 4027 | shows that the location counter holds the VMA value, not the LMA value. |
| 4028 | |
| 4029 | SECTIONS |
| 4030 | { |
| 4031 | .text 0x1000 : { *(.text) _etext = . ; } |
| 4032 | .mdata 0x2000 : |
| 4033 | AT ( ADDR (.text) + SIZEOF (.text) ) |
| 4034 | { _data = . ; *(.data); _edata = . ; } |
| 4035 | .bss 0x3000 : |
| 4036 | { _bstart = . ; *(.bss) *(COMMON) ; _bend = . ;} |
| 4037 | } |
| 4038 | |
| 4039 | The run-time initialization code for use with a program generated |
| 4040 | with this linker script would include something like the following, to |
| 4041 | copy the initialized data from the ROM image to its runtime address. |
| 4042 | Notice how this code takes advantage of the symbols defined by the |
| 4043 | linker script. |
| 4044 | |
| 4045 | extern char _etext, _data, _edata, _bstart, _bend; |
| 4046 | char *src = &_etext; |
| 4047 | char *dst = &_data; |
| 4048 | |
| 4049 | /* ROM has data at end of text; copy it. */ |
| 4050 | while (dst < &_edata) |
| 4051 | *dst++ = *src++; |
| 4052 | |
| 4053 | /* Zero bss. */ |
| 4054 | for (dst = &_bstart; dst< &_bend; dst++) |
| 4055 | *dst = 0; |
| 4056 | |
| 4057 | |
| 4058 | File: ld.info, Node: Forced Output Alignment, Next: Forced Input Alignment, Prev: Output Section LMA, Up: Output Section Attributes |
| 4059 | |
| 4060 | 3.6.8.3 Forced Output Alignment |
| 4061 | ............................... |
| 4062 | |
| 4063 | You can increase an output section's alignment by using ALIGN. As an |
| 4064 | alternative you can enforce that the difference between the VMA and LMA |
| 4065 | remains intact throughout this output section with the ALIGN_WITH_INPUT |
| 4066 | attribute. |
| 4067 | |
| 4068 | |
| 4069 | File: ld.info, Node: Forced Input Alignment, Next: Output Section Constraint, Prev: Forced Output Alignment, Up: Output Section Attributes |
| 4070 | |
| 4071 | 3.6.8.4 Forced Input Alignment |
| 4072 | .............................. |
| 4073 | |
| 4074 | You can force input section alignment within an output section by using |
| 4075 | SUBALIGN. The value specified overrides any alignment given by input |
| 4076 | sections, whether larger or smaller. |
| 4077 | |
| 4078 | |
| 4079 | File: ld.info, Node: Output Section Constraint, Next: Output Section Region, Prev: Forced Input Alignment, Up: Output Section Attributes |
| 4080 | |
| 4081 | 3.6.8.5 Output Section Constraint |
| 4082 | ................................. |
| 4083 | |
| 4084 | You can specify that an output section should only be created if all of |
| 4085 | its input sections are read-only or all of its input sections are |
| 4086 | read-write by using the keyword `ONLY_IF_RO' and `ONLY_IF_RW' |
| 4087 | respectively. |
| 4088 | |
| 4089 | |
| 4090 | File: ld.info, Node: Output Section Region, Next: Output Section Phdr, Prev: Output Section Constraint, Up: Output Section Attributes |
| 4091 | |
| 4092 | 3.6.8.6 Output Section Region |
| 4093 | ............................. |
| 4094 | |
| 4095 | You can assign a section to a previously defined region of memory by |
| 4096 | using `>REGION'. *Note MEMORY::. |
| 4097 | |
| 4098 | Here is a simple example: |
| 4099 | MEMORY { rom : ORIGIN = 0x1000, LENGTH = 0x1000 } |
| 4100 | SECTIONS { ROM : { *(.text) } >rom } |
| 4101 | |
| 4102 | |
| 4103 | File: ld.info, Node: Output Section Phdr, Next: Output Section Fill, Prev: Output Section Region, Up: Output Section Attributes |
| 4104 | |
| 4105 | 3.6.8.7 Output Section Phdr |
| 4106 | ........................... |
| 4107 | |
| 4108 | You can assign a section to a previously defined program segment by |
| 4109 | using `:PHDR'. *Note PHDRS::. If a section is assigned to one or more |
| 4110 | segments, then all subsequent allocated sections will be assigned to |
| 4111 | those segments as well, unless they use an explicitly `:PHDR' modifier. |
| 4112 | You can use `:NONE' to tell the linker to not put the section in any |
| 4113 | segment at all. |
| 4114 | |
| 4115 | Here is a simple example: |
| 4116 | PHDRS { text PT_LOAD ; } |
| 4117 | SECTIONS { .text : { *(.text) } :text } |
| 4118 | |
| 4119 | |
| 4120 | File: ld.info, Node: Output Section Fill, Prev: Output Section Phdr, Up: Output Section Attributes |
| 4121 | |
| 4122 | 3.6.8.8 Output Section Fill |
| 4123 | ........................... |
| 4124 | |
| 4125 | You can set the fill pattern for an entire section by using `=FILLEXP'. |
| 4126 | FILLEXP is an expression (*note Expressions::). Any otherwise |
| 4127 | unspecified regions of memory within the output section (for example, |
| 4128 | gaps left due to the required alignment of input sections) will be |
| 4129 | filled with the value, repeated as necessary. If the fill expression |
| 4130 | is a simple hex number, ie. a string of hex digit starting with `0x' |
| 4131 | and without a trailing `k' or `M', then an arbitrarily long sequence of |
| 4132 | hex digits can be used to specify the fill pattern; Leading zeros |
| 4133 | become part of the pattern too. For all other cases, including extra |
| 4134 | parentheses or a unary `+', the fill pattern is the four least |
| 4135 | significant bytes of the value of the expression. In all cases, the |
| 4136 | number is big-endian. |
| 4137 | |
| 4138 | You can also change the fill value with a `FILL' command in the |
| 4139 | output section commands; (*note Output Section Data::). |
| 4140 | |
| 4141 | Here is a simple example: |
| 4142 | SECTIONS { .text : { *(.text) } =0x90909090 } |
| 4143 | |
| 4144 | |
| 4145 | File: ld.info, Node: Overlay Description, Prev: Output Section Attributes, Up: SECTIONS |
| 4146 | |
| 4147 | 3.6.9 Overlay Description |
| 4148 | ------------------------- |
| 4149 | |
| 4150 | An overlay description provides an easy way to describe sections which |
| 4151 | are to be loaded as part of a single memory image but are to be run at |
| 4152 | the same memory address. At run time, some sort of overlay manager will |
| 4153 | copy the overlaid sections in and out of the runtime memory address as |
| 4154 | required, perhaps by simply manipulating addressing bits. This approach |
| 4155 | can be useful, for example, when a certain region of memory is faster |
| 4156 | than another. |
| 4157 | |
| 4158 | Overlays are described using the `OVERLAY' command. The `OVERLAY' |
| 4159 | command is used within a `SECTIONS' command, like an output section |
| 4160 | description. The full syntax of the `OVERLAY' command is as follows: |
| 4161 | OVERLAY [START] : [NOCROSSREFS] [AT ( LDADDR )] |
| 4162 | { |
| 4163 | SECNAME1 |
| 4164 | { |
| 4165 | OUTPUT-SECTION-COMMAND |
| 4166 | OUTPUT-SECTION-COMMAND |
| 4167 | ... |
| 4168 | } [:PHDR...] [=FILL] |
| 4169 | SECNAME2 |
| 4170 | { |
| 4171 | OUTPUT-SECTION-COMMAND |
| 4172 | OUTPUT-SECTION-COMMAND |
| 4173 | ... |
| 4174 | } [:PHDR...] [=FILL] |
| 4175 | ... |
| 4176 | } [>REGION] [:PHDR...] [=FILL] [,] |
| 4177 | |
| 4178 | Everything is optional except `OVERLAY' (a keyword), and each |
| 4179 | section must have a name (SECNAME1 and SECNAME2 above). The section |
| 4180 | definitions within the `OVERLAY' construct are identical to those |
| 4181 | within the general `SECTIONS' construct (*note SECTIONS::), except that |
| 4182 | no addresses and no memory regions may be defined for sections within |
| 4183 | an `OVERLAY'. |
| 4184 | |
| 4185 | The comma at the end may be required if a FILL is used and the next |
| 4186 | SECTIONS-COMMAND looks like a continuation of the expression. |
| 4187 | |
| 4188 | The sections are all defined with the same starting address. The |
| 4189 | load addresses of the sections are arranged such that they are |
| 4190 | consecutive in memory starting at the load address used for the |
| 4191 | `OVERLAY' as a whole (as with normal section definitions, the load |
| 4192 | address is optional, and defaults to the start address; the start |
| 4193 | address is also optional, and defaults to the current value of the |
| 4194 | location counter). |
| 4195 | |
| 4196 | If the `NOCROSSREFS' keyword is used, and there are any references |
| 4197 | among the sections, the linker will report an error. Since the |
| 4198 | sections all run at the same address, it normally does not make sense |
| 4199 | for one section to refer directly to another. *Note NOCROSSREFS: |
| 4200 | Miscellaneous Commands. |
| 4201 | |
| 4202 | For each section within the `OVERLAY', the linker automatically |
| 4203 | provides two symbols. The symbol `__load_start_SECNAME' is defined as |
| 4204 | the starting load address of the section. The symbol |
| 4205 | `__load_stop_SECNAME' is defined as the final load address of the |
| 4206 | section. Any characters within SECNAME which are not legal within C |
| 4207 | identifiers are removed. C (or assembler) code may use these symbols |
| 4208 | to move the overlaid sections around as necessary. |
| 4209 | |
| 4210 | At the end of the overlay, the value of the location counter is set |
| 4211 | to the start address of the overlay plus the size of the largest |
| 4212 | section. |
| 4213 | |
| 4214 | Here is an example. Remember that this would appear inside a |
| 4215 | `SECTIONS' construct. |
| 4216 | OVERLAY 0x1000 : AT (0x4000) |
| 4217 | { |
| 4218 | .text0 { o1/*.o(.text) } |
| 4219 | .text1 { o2/*.o(.text) } |
| 4220 | } |
| 4221 | This will define both `.text0' and `.text1' to start at address |
| 4222 | 0x1000. `.text0' will be loaded at address 0x4000, and `.text1' will |
| 4223 | be loaded immediately after `.text0'. The following symbols will be |
| 4224 | defined if referenced: `__load_start_text0', `__load_stop_text0', |
| 4225 | `__load_start_text1', `__load_stop_text1'. |
| 4226 | |
| 4227 | C code to copy overlay `.text1' into the overlay area might look |
| 4228 | like the following. |
| 4229 | |
| 4230 | extern char __load_start_text1, __load_stop_text1; |
| 4231 | memcpy ((char *) 0x1000, &__load_start_text1, |
| 4232 | &__load_stop_text1 - &__load_start_text1); |
| 4233 | |
| 4234 | Note that the `OVERLAY' command is just syntactic sugar, since |
| 4235 | everything it does can be done using the more basic commands. The above |
| 4236 | example could have been written identically as follows. |
| 4237 | |
| 4238 | .text0 0x1000 : AT (0x4000) { o1/*.o(.text) } |
| 4239 | PROVIDE (__load_start_text0 = LOADADDR (.text0)); |
| 4240 | PROVIDE (__load_stop_text0 = LOADADDR (.text0) + SIZEOF (.text0)); |
| 4241 | .text1 0x1000 : AT (0x4000 + SIZEOF (.text0)) { o2/*.o(.text) } |
| 4242 | PROVIDE (__load_start_text1 = LOADADDR (.text1)); |
| 4243 | PROVIDE (__load_stop_text1 = LOADADDR (.text1) + SIZEOF (.text1)); |
| 4244 | . = 0x1000 + MAX (SIZEOF (.text0), SIZEOF (.text1)); |
| 4245 | |
| 4246 | |
| 4247 | File: ld.info, Node: MEMORY, Next: PHDRS, Prev: SECTIONS, Up: Scripts |
| 4248 | |
| 4249 | 3.7 MEMORY Command |
| 4250 | ================== |
| 4251 | |
| 4252 | The linker's default configuration permits allocation of all available |
| 4253 | memory. You can override this by using the `MEMORY' command. |
| 4254 | |
| 4255 | The `MEMORY' command describes the location and size of blocks of |
| 4256 | memory in the target. You can use it to describe which memory regions |
| 4257 | may be used by the linker, and which memory regions it must avoid. You |
| 4258 | can then assign sections to particular memory regions. The linker will |
| 4259 | set section addresses based on the memory regions, and will warn about |
| 4260 | regions that become too full. The linker will not shuffle sections |
| 4261 | around to fit into the available regions. |
| 4262 | |
| 4263 | A linker script may contain many uses of the `MEMORY' command, |
| 4264 | however, all memory blocks defined are treated as if they were |
| 4265 | specified inside a single `MEMORY' command. The syntax for `MEMORY' is: |
| 4266 | MEMORY |
| 4267 | { |
| 4268 | NAME [(ATTR)] : ORIGIN = ORIGIN, LENGTH = LEN |
| 4269 | ... |
| 4270 | } |
| 4271 | |
| 4272 | The NAME is a name used in the linker script to refer to the region. |
| 4273 | The region name has no meaning outside of the linker script. Region |
| 4274 | names are stored in a separate name space, and will not conflict with |
| 4275 | symbol names, file names, or section names. Each memory region must |
| 4276 | have a distinct name within the `MEMORY' command. However you can add |
| 4277 | later alias names to existing memory regions with the *Note |
| 4278 | REGION_ALIAS:: command. |
| 4279 | |
| 4280 | The ATTR string is an optional list of attributes that specify |
| 4281 | whether to use a particular memory region for an input section which is |
| 4282 | not explicitly mapped in the linker script. As described in *Note |
| 4283 | SECTIONS::, if you do not specify an output section for some input |
| 4284 | section, the linker will create an output section with the same name as |
| 4285 | the input section. If you define region attributes, the linker will use |
| 4286 | them to select the memory region for the output section that it creates. |
| 4287 | |
| 4288 | The ATTR string must consist only of the following characters: |
| 4289 | `R' |
| 4290 | Read-only section |
| 4291 | |
| 4292 | `W' |
| 4293 | Read/write section |
| 4294 | |
| 4295 | `X' |
| 4296 | Executable section |
| 4297 | |
| 4298 | `A' |
| 4299 | Allocatable section |
| 4300 | |
| 4301 | `I' |
| 4302 | Initialized section |
| 4303 | |
| 4304 | `L' |
| 4305 | Same as `I' |
| 4306 | |
| 4307 | `!' |
| 4308 | Invert the sense of any of the attributes that follow |
| 4309 | |
| 4310 | If a unmapped section matches any of the listed attributes other than |
| 4311 | `!', it will be placed in the memory region. The `!' attribute |
| 4312 | reverses this test, so that an unmapped section will be placed in the |
| 4313 | memory region only if it does not match any of the listed attributes. |
| 4314 | |
| 4315 | The ORIGIN is an numerical expression for the start address of the |
| 4316 | memory region. The expression must evaluate to a constant and it |
| 4317 | cannot involve any symbols. The keyword `ORIGIN' may be abbreviated to |
| 4318 | `org' or `o' (but not, for example, `ORG'). |
| 4319 | |
| 4320 | The LEN is an expression for the size in bytes of the memory region. |
| 4321 | As with the ORIGIN expression, the expression must be numerical only |
| 4322 | and must evaluate to a constant. The keyword `LENGTH' may be |
| 4323 | abbreviated to `len' or `l'. |
| 4324 | |
| 4325 | In the following example, we specify that there are two memory |
| 4326 | regions available for allocation: one starting at `0' for 256 kilobytes, |
| 4327 | and the other starting at `0x40000000' for four megabytes. The linker |
| 4328 | will place into the `rom' memory region every section which is not |
| 4329 | explicitly mapped into a memory region, and is either read-only or |
| 4330 | executable. The linker will place other sections which are not |
| 4331 | explicitly mapped into a memory region into the `ram' memory region. |
| 4332 | |
| 4333 | MEMORY |
| 4334 | { |
| 4335 | rom (rx) : ORIGIN = 0, LENGTH = 256K |
| 4336 | ram (!rx) : org = 0x40000000, l = 4M |
| 4337 | } |
| 4338 | |
| 4339 | Once you define a memory region, you can direct the linker to place |
| 4340 | specific output sections into that memory region by using the `>REGION' |
| 4341 | output section attribute. For example, if you have a memory region |
| 4342 | named `mem', you would use `>mem' in the output section definition. |
| 4343 | *Note Output Section Region::. If no address was specified for the |
| 4344 | output section, the linker will set the address to the next available |
| 4345 | address within the memory region. If the combined output sections |
| 4346 | directed to a memory region are too large for the region, the linker |
| 4347 | will issue an error message. |
| 4348 | |
| 4349 | It is possible to access the origin and length of a memory in an |
| 4350 | expression via the `ORIGIN(MEMORY)' and `LENGTH(MEMORY)' functions: |
| 4351 | |
| 4352 | _fstack = ORIGIN(ram) + LENGTH(ram) - 4; |
| 4353 | |
| 4354 | |
| 4355 | File: ld.info, Node: PHDRS, Next: VERSION, Prev: MEMORY, Up: Scripts |
| 4356 | |
| 4357 | 3.8 PHDRS Command |
| 4358 | ================= |
| 4359 | |
| 4360 | The ELF object file format uses "program headers", also knows as |
| 4361 | "segments". The program headers describe how the program should be |
| 4362 | loaded into memory. You can print them out by using the `objdump' |
| 4363 | program with the `-p' option. |
| 4364 | |
| 4365 | When you run an ELF program on a native ELF system, the system loader |
| 4366 | reads the program headers in order to figure out how to load the |
| 4367 | program. This will only work if the program headers are set correctly. |
| 4368 | This manual does not describe the details of how the system loader |
| 4369 | interprets program headers; for more information, see the ELF ABI. |
| 4370 | |
| 4371 | The linker will create reasonable program headers by default. |
| 4372 | However, in some cases, you may need to specify the program headers more |
| 4373 | precisely. You may use the `PHDRS' command for this purpose. When the |
| 4374 | linker sees the `PHDRS' command in the linker script, it will not |
| 4375 | create any program headers other than the ones specified. |
| 4376 | |
| 4377 | The linker only pays attention to the `PHDRS' command when |
| 4378 | generating an ELF output file. In other cases, the linker will simply |
| 4379 | ignore `PHDRS'. |
| 4380 | |
| 4381 | This is the syntax of the `PHDRS' command. The words `PHDRS', |
| 4382 | `FILEHDR', `AT', and `FLAGS' are keywords. |
| 4383 | |
| 4384 | PHDRS |
| 4385 | { |
| 4386 | NAME TYPE [ FILEHDR ] [ PHDRS ] [ AT ( ADDRESS ) ] |
| 4387 | [ FLAGS ( FLAGS ) ] ; |
| 4388 | } |
| 4389 | |
| 4390 | The NAME is used only for reference in the `SECTIONS' command of the |
| 4391 | linker script. It is not put into the output file. Program header |
| 4392 | names are stored in a separate name space, and will not conflict with |
| 4393 | symbol names, file names, or section names. Each program header must |
| 4394 | have a distinct name. The headers are processed in order and it is |
| 4395 | usual for them to map to sections in ascending load address order. |
| 4396 | |
| 4397 | Certain program header types describe segments of memory which the |
| 4398 | system loader will load from the file. In the linker script, you |
| 4399 | specify the contents of these segments by placing allocatable output |
| 4400 | sections in the segments. You use the `:PHDR' output section attribute |
| 4401 | to place a section in a particular segment. *Note Output Section |
| 4402 | Phdr::. |
| 4403 | |
| 4404 | It is normal to put certain sections in more than one segment. This |
| 4405 | merely implies that one segment of memory contains another. You may |
| 4406 | repeat `:PHDR', using it once for each segment which should contain the |
| 4407 | section. |
| 4408 | |
| 4409 | If you place a section in one or more segments using `:PHDR', then |
| 4410 | the linker will place all subsequent allocatable sections which do not |
| 4411 | specify `:PHDR' in the same segments. This is for convenience, since |
| 4412 | generally a whole set of contiguous sections will be placed in a single |
| 4413 | segment. You can use `:NONE' to override the default segment and tell |
| 4414 | the linker to not put the section in any segment at all. |
| 4415 | |
| 4416 | You may use the `FILEHDR' and `PHDRS' keywords after the program |
| 4417 | header type to further describe the contents of the segment. The |
| 4418 | `FILEHDR' keyword means that the segment should include the ELF file |
| 4419 | header. The `PHDRS' keyword means that the segment should include the |
| 4420 | ELF program headers themselves. If applied to a loadable segment |
| 4421 | (`PT_LOAD'), all prior loadable segments must have one of these |
| 4422 | keywords. |
| 4423 | |
| 4424 | The TYPE may be one of the following. The numbers indicate the |
| 4425 | value of the keyword. |
| 4426 | |
| 4427 | `PT_NULL' (0) |
| 4428 | Indicates an unused program header. |
| 4429 | |
| 4430 | `PT_LOAD' (1) |
| 4431 | Indicates that this program header describes a segment to be |
| 4432 | loaded from the file. |
| 4433 | |
| 4434 | `PT_DYNAMIC' (2) |
| 4435 | Indicates a segment where dynamic linking information can be found. |
| 4436 | |
| 4437 | `PT_INTERP' (3) |
| 4438 | Indicates a segment where the name of the program interpreter may |
| 4439 | be found. |
| 4440 | |
| 4441 | `PT_NOTE' (4) |
| 4442 | Indicates a segment holding note information. |
| 4443 | |
| 4444 | `PT_SHLIB' (5) |
| 4445 | A reserved program header type, defined but not specified by the |
| 4446 | ELF ABI. |
| 4447 | |
| 4448 | `PT_PHDR' (6) |
| 4449 | Indicates a segment where the program headers may be found. |
| 4450 | |
| 4451 | EXPRESSION |
| 4452 | An expression giving the numeric type of the program header. This |
| 4453 | may be used for types not defined above. |
| 4454 | |
| 4455 | You can specify that a segment should be loaded at a particular |
| 4456 | address in memory by using an `AT' expression. This is identical to the |
| 4457 | `AT' command used as an output section attribute (*note Output Section |
| 4458 | LMA::). The `AT' command for a program header overrides the output |
| 4459 | section attribute. |
| 4460 | |
| 4461 | The linker will normally set the segment flags based on the sections |
| 4462 | which comprise the segment. You may use the `FLAGS' keyword to |
| 4463 | explicitly specify the segment flags. The value of FLAGS must be an |
| 4464 | integer. It is used to set the `p_flags' field of the program header. |
| 4465 | |
| 4466 | Here is an example of `PHDRS'. This shows a typical set of program |
| 4467 | headers used on a native ELF system. |
| 4468 | |
| 4469 | PHDRS |
| 4470 | { |
| 4471 | headers PT_PHDR PHDRS ; |
| 4472 | interp PT_INTERP ; |
| 4473 | text PT_LOAD FILEHDR PHDRS ; |
| 4474 | data PT_LOAD ; |
| 4475 | dynamic PT_DYNAMIC ; |
| 4476 | } |
| 4477 | |
| 4478 | SECTIONS |
| 4479 | { |
| 4480 | . = SIZEOF_HEADERS; |
| 4481 | .interp : { *(.interp) } :text :interp |
| 4482 | .text : { *(.text) } :text |
| 4483 | .rodata : { *(.rodata) } /* defaults to :text */ |
| 4484 | ... |
| 4485 | . = . + 0x1000; /* move to a new page in memory */ |
| 4486 | .data : { *(.data) } :data |
| 4487 | .dynamic : { *(.dynamic) } :data :dynamic |
| 4488 | ... |
| 4489 | } |
| 4490 | |
| 4491 | |
| 4492 | File: ld.info, Node: VERSION, Next: Expressions, Prev: PHDRS, Up: Scripts |
| 4493 | |
| 4494 | 3.9 VERSION Command |
| 4495 | =================== |
| 4496 | |
| 4497 | The linker supports symbol versions when using ELF. Symbol versions are |
| 4498 | only useful when using shared libraries. The dynamic linker can use |
| 4499 | symbol versions to select a specific version of a function when it runs |
| 4500 | a program that may have been linked against an earlier version of the |
| 4501 | shared library. |
| 4502 | |
| 4503 | You can include a version script directly in the main linker script, |
| 4504 | or you can supply the version script as an implicit linker script. You |
| 4505 | can also use the `--version-script' linker option. |
| 4506 | |
| 4507 | The syntax of the `VERSION' command is simply |
| 4508 | VERSION { version-script-commands } |
| 4509 | |
| 4510 | The format of the version script commands is identical to that used |
| 4511 | by Sun's linker in Solaris 2.5. The version script defines a tree of |
| 4512 | version nodes. You specify the node names and interdependencies in the |
| 4513 | version script. You can specify which symbols are bound to which |
| 4514 | version nodes, and you can reduce a specified set of symbols to local |
| 4515 | scope so that they are not globally visible outside of the shared |
| 4516 | library. |
| 4517 | |
| 4518 | The easiest way to demonstrate the version script language is with a |
| 4519 | few examples. |
| 4520 | |
| 4521 | VERS_1.1 { |
| 4522 | global: |
| 4523 | foo1; |
| 4524 | local: |
| 4525 | old*; |
| 4526 | original*; |
| 4527 | new*; |
| 4528 | }; |
| 4529 | |
| 4530 | VERS_1.2 { |
| 4531 | foo2; |
| 4532 | } VERS_1.1; |
| 4533 | |
| 4534 | VERS_2.0 { |
| 4535 | bar1; bar2; |
| 4536 | extern "C++" { |
| 4537 | ns::*; |
| 4538 | "f(int, double)"; |
| 4539 | }; |
| 4540 | } VERS_1.2; |
| 4541 | |
| 4542 | This example version script defines three version nodes. The first |
| 4543 | version node defined is `VERS_1.1'; it has no other dependencies. The |
| 4544 | script binds the symbol `foo1' to `VERS_1.1'. It reduces a number of |
| 4545 | symbols to local scope so that they are not visible outside of the |
| 4546 | shared library; this is done using wildcard patterns, so that any |
| 4547 | symbol whose name begins with `old', `original', or `new' is matched. |
| 4548 | The wildcard patterns available are the same as those used in the shell |
| 4549 | when matching filenames (also known as "globbing"). However, if you |
| 4550 | specify the symbol name inside double quotes, then the name is treated |
| 4551 | as literal, rather than as a glob pattern. |
| 4552 | |
| 4553 | Next, the version script defines node `VERS_1.2'. This node depends |
| 4554 | upon `VERS_1.1'. The script binds the symbol `foo2' to the version |
| 4555 | node `VERS_1.2'. |
| 4556 | |
| 4557 | Finally, the version script defines node `VERS_2.0'. This node |
| 4558 | depends upon `VERS_1.2'. The scripts binds the symbols `bar1' and |
| 4559 | `bar2' are bound to the version node `VERS_2.0'. |
| 4560 | |
| 4561 | When the linker finds a symbol defined in a library which is not |
| 4562 | specifically bound to a version node, it will effectively bind it to an |
| 4563 | unspecified base version of the library. You can bind all otherwise |
| 4564 | unspecified symbols to a given version node by using `global: *;' |
| 4565 | somewhere in the version script. Note that it's slightly crazy to use |
| 4566 | wildcards in a global spec except on the last version node. Global |
| 4567 | wildcards elsewhere run the risk of accidentally adding symbols to the |
| 4568 | set exported for an old version. That's wrong since older versions |
| 4569 | ought to have a fixed set of symbols. |
| 4570 | |
| 4571 | The names of the version nodes have no specific meaning other than |
| 4572 | what they might suggest to the person reading them. The `2.0' version |
| 4573 | could just as well have appeared in between `1.1' and `1.2'. However, |
| 4574 | this would be a confusing way to write a version script. |
| 4575 | |
| 4576 | Node name can be omitted, provided it is the only version node in |
| 4577 | the version script. Such version script doesn't assign any versions to |
| 4578 | symbols, only selects which symbols will be globally visible out and |
| 4579 | which won't. |
| 4580 | |
| 4581 | { global: foo; bar; local: *; }; |
| 4582 | |
| 4583 | When you link an application against a shared library that has |
| 4584 | versioned symbols, the application itself knows which version of each |
| 4585 | symbol it requires, and it also knows which version nodes it needs from |
| 4586 | each shared library it is linked against. Thus at runtime, the dynamic |
| 4587 | loader can make a quick check to make sure that the libraries you have |
| 4588 | linked against do in fact supply all of the version nodes that the |
| 4589 | application will need to resolve all of the dynamic symbols. In this |
| 4590 | way it is possible for the dynamic linker to know with certainty that |
| 4591 | all external symbols that it needs will be resolvable without having to |
| 4592 | search for each symbol reference. |
| 4593 | |
| 4594 | The symbol versioning is in effect a much more sophisticated way of |
| 4595 | doing minor version checking that SunOS does. The fundamental problem |
| 4596 | that is being addressed here is that typically references to external |
| 4597 | functions are bound on an as-needed basis, and are not all bound when |
| 4598 | the application starts up. If a shared library is out of date, a |
| 4599 | required interface may be missing; when the application tries to use |
| 4600 | that interface, it may suddenly and unexpectedly fail. With symbol |
| 4601 | versioning, the user will get a warning when they start their program if |
| 4602 | the libraries being used with the application are too old. |
| 4603 | |
| 4604 | There are several GNU extensions to Sun's versioning approach. The |
| 4605 | first of these is the ability to bind a symbol to a version node in the |
| 4606 | source file where the symbol is defined instead of in the versioning |
| 4607 | script. This was done mainly to reduce the burden on the library |
| 4608 | maintainer. You can do this by putting something like: |
| 4609 | __asm__(".symver original_foo,foo@VERS_1.1"); |
| 4610 | in the C source file. This renames the function `original_foo' to |
| 4611 | be an alias for `foo' bound to the version node `VERS_1.1'. The |
| 4612 | `local:' directive can be used to prevent the symbol `original_foo' |
| 4613 | from being exported. A `.symver' directive takes precedence over a |
| 4614 | version script. |
| 4615 | |
| 4616 | The second GNU extension is to allow multiple versions of the same |
| 4617 | function to appear in a given shared library. In this way you can make |
| 4618 | an incompatible change to an interface without increasing the major |
| 4619 | version number of the shared library, while still allowing applications |
| 4620 | linked against the old interface to continue to function. |
| 4621 | |
| 4622 | To do this, you must use multiple `.symver' directives in the source |
| 4623 | file. Here is an example: |
| 4624 | |
| 4625 | __asm__(".symver original_foo,foo@"); |
| 4626 | __asm__(".symver old_foo,foo@VERS_1.1"); |
| 4627 | __asm__(".symver old_foo1,foo@VERS_1.2"); |
| 4628 | __asm__(".symver new_foo,foo@@VERS_2.0"); |
| 4629 | |
| 4630 | In this example, `foo@' represents the symbol `foo' bound to the |
| 4631 | unspecified base version of the symbol. The source file that contains |
| 4632 | this example would define 4 C functions: `original_foo', `old_foo', |
| 4633 | `old_foo1', and `new_foo'. |
| 4634 | |
| 4635 | When you have multiple definitions of a given symbol, there needs to |
| 4636 | be some way to specify a default version to which external references to |
| 4637 | this symbol will be bound. You can do this with the `foo@@VERS_2.0' |
| 4638 | type of `.symver' directive. You can only declare one version of a |
| 4639 | symbol as the default in this manner; otherwise you would effectively |
| 4640 | have multiple definitions of the same symbol. |
| 4641 | |
| 4642 | If you wish to bind a reference to a specific version of the symbol |
| 4643 | within the shared library, you can use the aliases of convenience |
| 4644 | (i.e., `old_foo'), or you can use the `.symver' directive to |
| 4645 | specifically bind to an external version of the function in question. |
| 4646 | |
| 4647 | You can also specify the language in the version script: |
| 4648 | |
| 4649 | VERSION extern "lang" { version-script-commands } |
| 4650 | |
| 4651 | The supported `lang's are `C', `C++', and `Java'. The linker will |
| 4652 | iterate over the list of symbols at the link time and demangle them |
| 4653 | according to `lang' before matching them to the patterns specified in |
| 4654 | `version-script-commands'. The default `lang' is `C'. |
| 4655 | |
| 4656 | Demangled names may contains spaces and other special characters. As |
| 4657 | described above, you can use a glob pattern to match demangled names, |
| 4658 | or you can use a double-quoted string to match the string exactly. In |
| 4659 | the latter case, be aware that minor differences (such as differing |
| 4660 | whitespace) between the version script and the demangler output will |
| 4661 | cause a mismatch. As the exact string generated by the demangler might |
| 4662 | change in the future, even if the mangled name does not, you should |
| 4663 | check that all of your version directives are behaving as you expect |
| 4664 | when you upgrade. |
| 4665 | |
| 4666 | |
| 4667 | File: ld.info, Node: Expressions, Next: Implicit Linker Scripts, Prev: VERSION, Up: Scripts |
| 4668 | |
| 4669 | 3.10 Expressions in Linker Scripts |
| 4670 | ================================== |
| 4671 | |
| 4672 | The syntax for expressions in the linker script language is identical to |
| 4673 | that of C expressions. All expressions are evaluated as integers. All |
| 4674 | expressions are evaluated in the same size, which is 32 bits if both the |
| 4675 | host and target are 32 bits, and is otherwise 64 bits. |
| 4676 | |
| 4677 | You can use and set symbol values in expressions. |
| 4678 | |
| 4679 | The linker defines several special purpose builtin functions for use |
| 4680 | in expressions. |
| 4681 | |
| 4682 | * Menu: |
| 4683 | |
| 4684 | * Constants:: Constants |
| 4685 | * Symbolic Constants:: Symbolic constants |
| 4686 | * Symbols:: Symbol Names |
| 4687 | * Orphan Sections:: Orphan Sections |
| 4688 | * Location Counter:: The Location Counter |
| 4689 | * Operators:: Operators |
| 4690 | * Evaluation:: Evaluation |
| 4691 | * Expression Section:: The Section of an Expression |
| 4692 | * Builtin Functions:: Builtin Functions |
| 4693 | |
| 4694 | |
| 4695 | File: ld.info, Node: Constants, Next: Symbolic Constants, Up: Expressions |
| 4696 | |
| 4697 | 3.10.1 Constants |
| 4698 | ---------------- |
| 4699 | |
| 4700 | All constants are integers. |
| 4701 | |
| 4702 | As in C, the linker considers an integer beginning with `0' to be |
| 4703 | octal, and an integer beginning with `0x' or `0X' to be hexadecimal. |
| 4704 | Alternatively the linker accepts suffixes of `h' or `H' for |
| 4705 | hexadecimal, `o' or `O' for octal, `b' or `B' for binary and `d' or `D' |
| 4706 | for decimal. Any integer value without a prefix or a suffix is |
| 4707 | considered to be decimal. |
| 4708 | |
| 4709 | In addition, you can use the suffixes `K' and `M' to scale a |
| 4710 | constant by `1024' or `1024*1024' respectively. For example, the |
| 4711 | following all refer to the same quantity: |
| 4712 | |
| 4713 | _fourk_1 = 4K; |
| 4714 | _fourk_2 = 4096; |
| 4715 | _fourk_3 = 0x1000; |
| 4716 | _fourk_4 = 10000o; |
| 4717 | |
| 4718 | Note - the `K' and `M' suffixes cannot be used in conjunction with |
| 4719 | the base suffixes mentioned above. |
| 4720 | |
| 4721 | |
| 4722 | File: ld.info, Node: Symbolic Constants, Next: Symbols, Prev: Constants, Up: Expressions |
| 4723 | |
| 4724 | 3.10.2 Symbolic Constants |
| 4725 | ------------------------- |
| 4726 | |
| 4727 | It is possible to refer to target specific constants via the use of the |
| 4728 | `CONSTANT(NAME)' operator, where NAME is one of: |
| 4729 | |
| 4730 | `MAXPAGESIZE' |
| 4731 | The target's maximum page size. |
| 4732 | |
| 4733 | `COMMONPAGESIZE' |
| 4734 | The target's default page size. |
| 4735 | |
| 4736 | So for example: |
| 4737 | |
| 4738 | .text ALIGN (CONSTANT (MAXPAGESIZE)) : { *(.text) } |
| 4739 | |
| 4740 | will create a text section aligned to the largest page boundary |
| 4741 | supported by the target. |
| 4742 | |
| 4743 | |
| 4744 | File: ld.info, Node: Symbols, Next: Orphan Sections, Prev: Symbolic Constants, Up: Expressions |
| 4745 | |
| 4746 | 3.10.3 Symbol Names |
| 4747 | ------------------- |
| 4748 | |
| 4749 | Unless quoted, symbol names start with a letter, underscore, or period |
| 4750 | and may include letters, digits, underscores, periods, and hyphens. |
| 4751 | Unquoted symbol names must not conflict with any keywords. You can |
| 4752 | specify a symbol which contains odd characters or has the same name as a |
| 4753 | keyword by surrounding the symbol name in double quotes: |
| 4754 | "SECTION" = 9; |
| 4755 | "with a space" = "also with a space" + 10; |
| 4756 | |
| 4757 | Since symbols can contain many non-alphabetic characters, it is |
| 4758 | safest to delimit symbols with spaces. For example, `A-B' is one |
| 4759 | symbol, whereas `A - B' is an expression involving subtraction. |
| 4760 | |
| 4761 | |
| 4762 | File: ld.info, Node: Orphan Sections, Next: Location Counter, Prev: Symbols, Up: Expressions |
| 4763 | |
| 4764 | 3.10.4 Orphan Sections |
| 4765 | ---------------------- |
| 4766 | |
| 4767 | Orphan sections are sections present in the input files which are not |
| 4768 | explicitly placed into the output file by the linker script. The |
| 4769 | linker will still copy these sections into the output file, but it has |
| 4770 | to guess as to where they should be placed. The linker uses a simple |
| 4771 | heuristic to do this. It attempts to place orphan sections after |
| 4772 | non-orphan sections of the same attribute, such as code vs data, |
| 4773 | loadable vs non-loadable, etc. If there is not enough room to do this |
| 4774 | then it places at the end of the file. |
| 4775 | |
| 4776 | For ELF targets, the attribute of the section includes section type |
| 4777 | as well as section flag. |
| 4778 | |
| 4779 | The command line options `--orphan-handling' and `--unique' (*note |
| 4780 | Command Line Options: Options.) can be used to control which output |
| 4781 | sections an orphan is placed in. |
| 4782 | |
| 4783 | If an orphaned section's name is representable as a C identifier then |
| 4784 | the linker will automatically *note PROVIDE:: two symbols: |
| 4785 | __start_SECNAME and __stop_SECNAME, where SECNAME is the name of the |
| 4786 | section. These indicate the start address and end address of the |
| 4787 | orphaned section respectively. Note: most section names are not |
| 4788 | representable as C identifiers because they contain a `.' character. |
| 4789 | |
| 4790 | |
| 4791 | File: ld.info, Node: Location Counter, Next: Operators, Prev: Orphan Sections, Up: Expressions |
| 4792 | |
| 4793 | 3.10.5 The Location Counter |
| 4794 | --------------------------- |
| 4795 | |
| 4796 | The special linker variable "dot" `.' always contains the current |
| 4797 | output location counter. Since the `.' always refers to a location in |
| 4798 | an output section, it may only appear in an expression within a |
| 4799 | `SECTIONS' command. The `.' symbol may appear anywhere that an |
| 4800 | ordinary symbol is allowed in an expression. |
| 4801 | |
| 4802 | Assigning a value to `.' will cause the location counter to be |
| 4803 | moved. This may be used to create holes in the output section. The |
| 4804 | location counter may not be moved backwards inside an output section, |
| 4805 | and may not be moved backwards outside of an output section if so doing |
| 4806 | creates areas with overlapping LMAs. |
| 4807 | |
| 4808 | SECTIONS |
| 4809 | { |
| 4810 | output : |
| 4811 | { |
| 4812 | file1(.text) |
| 4813 | . = . + 1000; |
| 4814 | file2(.text) |
| 4815 | . += 1000; |
| 4816 | file3(.text) |
| 4817 | } = 0x12345678; |
| 4818 | } |
| 4819 | In the previous example, the `.text' section from `file1' is located |
| 4820 | at the beginning of the output section `output'. It is followed by a |
| 4821 | 1000 byte gap. Then the `.text' section from `file2' appears, also |
| 4822 | with a 1000 byte gap following before the `.text' section from `file3'. |
| 4823 | The notation `= 0x12345678' specifies what data to write in the gaps |
| 4824 | (*note Output Section Fill::). |
| 4825 | |
| 4826 | Note: `.' actually refers to the byte offset from the start of the |
| 4827 | current containing object. Normally this is the `SECTIONS' statement, |
| 4828 | whose start address is 0, hence `.' can be used as an absolute address. |
| 4829 | If `.' is used inside a section description however, it refers to the |
| 4830 | byte offset from the start of that section, not an absolute address. |
| 4831 | Thus in a script like this: |
| 4832 | |
| 4833 | SECTIONS |
| 4834 | { |
| 4835 | . = 0x100 |
| 4836 | .text: { |
| 4837 | *(.text) |
| 4838 | . = 0x200 |
| 4839 | } |
| 4840 | . = 0x500 |
| 4841 | .data: { |
| 4842 | *(.data) |
| 4843 | . += 0x600 |
| 4844 | } |
| 4845 | } |
| 4846 | |
| 4847 | The `.text' section will be assigned a starting address of 0x100 and |
| 4848 | a size of exactly 0x200 bytes, even if there is not enough data in the |
| 4849 | `.text' input sections to fill this area. (If there is too much data, |
| 4850 | an error will be produced because this would be an attempt to move `.' |
| 4851 | backwards). The `.data' section will start at 0x500 and it will have |
| 4852 | an extra 0x600 bytes worth of space after the end of the values from |
| 4853 | the `.data' input sections and before the end of the `.data' output |
| 4854 | section itself. |
| 4855 | |
| 4856 | Setting symbols to the value of the location counter outside of an |
| 4857 | output section statement can result in unexpected values if the linker |
| 4858 | needs to place orphan sections. For example, given the following: |
| 4859 | |
| 4860 | SECTIONS |
| 4861 | { |
| 4862 | start_of_text = . ; |
| 4863 | .text: { *(.text) } |
| 4864 | end_of_text = . ; |
| 4865 | |
| 4866 | start_of_data = . ; |
| 4867 | .data: { *(.data) } |
| 4868 | end_of_data = . ; |
| 4869 | } |
| 4870 | |
| 4871 | If the linker needs to place some input section, e.g. `.rodata', not |
| 4872 | mentioned in the script, it might choose to place that section between |
| 4873 | `.text' and `.data'. You might think the linker should place `.rodata' |
| 4874 | on the blank line in the above script, but blank lines are of no |
| 4875 | particular significance to the linker. As well, the linker doesn't |
| 4876 | associate the above symbol names with their sections. Instead, it |
| 4877 | assumes that all assignments or other statements belong to the previous |
| 4878 | output section, except for the special case of an assignment to `.'. |
| 4879 | I.e., the linker will place the orphan `.rodata' section as if the |
| 4880 | script was written as follows: |
| 4881 | |
| 4882 | SECTIONS |
| 4883 | { |
| 4884 | start_of_text = . ; |
| 4885 | .text: { *(.text) } |
| 4886 | end_of_text = . ; |
| 4887 | |
| 4888 | start_of_data = . ; |
| 4889 | .rodata: { *(.rodata) } |
| 4890 | .data: { *(.data) } |
| 4891 | end_of_data = . ; |
| 4892 | } |
| 4893 | |
| 4894 | This may or may not be the script author's intention for the value of |
| 4895 | `start_of_data'. One way to influence the orphan section placement is |
| 4896 | to assign the location counter to itself, as the linker assumes that an |
| 4897 | assignment to `.' is setting the start address of a following output |
| 4898 | section and thus should be grouped with that section. So you could |
| 4899 | write: |
| 4900 | |
| 4901 | SECTIONS |
| 4902 | { |
| 4903 | start_of_text = . ; |
| 4904 | .text: { *(.text) } |
| 4905 | end_of_text = . ; |
| 4906 | |
| 4907 | . = . ; |
| 4908 | start_of_data = . ; |
| 4909 | .data: { *(.data) } |
| 4910 | end_of_data = . ; |
| 4911 | } |
| 4912 | |
| 4913 | Now, the orphan `.rodata' section will be placed between |
| 4914 | `end_of_text' and `start_of_data'. |
| 4915 | |
| 4916 | |
| 4917 | File: ld.info, Node: Operators, Next: Evaluation, Prev: Location Counter, Up: Expressions |
| 4918 | |
| 4919 | 3.10.6 Operators |
| 4920 | ---------------- |
| 4921 | |
| 4922 | The linker recognizes the standard C set of arithmetic operators, with |
| 4923 | the standard bindings and precedence levels: |
| 4924 | precedence associativity Operators Notes |
| 4925 | (highest) |
| 4926 | 1 left ! - ~ (1) |
| 4927 | 2 left * / % |
| 4928 | 3 left + - |
| 4929 | 4 left >> << |
| 4930 | 5 left == != > < <= >= |
| 4931 | 6 left & |
| 4932 | 7 left | |
| 4933 | 8 left && |
| 4934 | 9 left || |
| 4935 | 10 right ? : |
| 4936 | 11 right &= += -= *= /= (2) |
| 4937 | (lowest) |
| 4938 | Notes: (1) Prefix operators (2) *Note Assignments::. |
| 4939 | |
| 4940 | |
| 4941 | File: ld.info, Node: Evaluation, Next: Expression Section, Prev: Operators, Up: Expressions |
| 4942 | |
| 4943 | 3.10.7 Evaluation |
| 4944 | ----------------- |
| 4945 | |
| 4946 | The linker evaluates expressions lazily. It only computes the value of |
| 4947 | an expression when absolutely necessary. |
| 4948 | |
| 4949 | The linker needs some information, such as the value of the start |
| 4950 | address of the first section, and the origins and lengths of memory |
| 4951 | regions, in order to do any linking at all. These values are computed |
| 4952 | as soon as possible when the linker reads in the linker script. |
| 4953 | |
| 4954 | However, other values (such as symbol values) are not known or needed |
| 4955 | until after storage allocation. Such values are evaluated later, when |
| 4956 | other information (such as the sizes of output sections) is available |
| 4957 | for use in the symbol assignment expression. |
| 4958 | |
| 4959 | The sizes of sections cannot be known until after allocation, so |
| 4960 | assignments dependent upon these are not performed until after |
| 4961 | allocation. |
| 4962 | |
| 4963 | Some expressions, such as those depending upon the location counter |
| 4964 | `.', must be evaluated during section allocation. |
| 4965 | |
| 4966 | If the result of an expression is required, but the value is not |
| 4967 | available, then an error results. For example, a script like the |
| 4968 | following |
| 4969 | SECTIONS |
| 4970 | { |
| 4971 | .text 9+this_isnt_constant : |
| 4972 | { *(.text) } |
| 4973 | } |
| 4974 | will cause the error message `non constant expression for initial |
| 4975 | address'. |
| 4976 | |
| 4977 | |
| 4978 | File: ld.info, Node: Expression Section, Next: Builtin Functions, Prev: Evaluation, Up: Expressions |
| 4979 | |
| 4980 | 3.10.8 The Section of an Expression |
| 4981 | ----------------------------------- |
| 4982 | |
| 4983 | Addresses and symbols may be section relative, or absolute. A section |
| 4984 | relative symbol is relocatable. If you request relocatable output |
| 4985 | using the `-r' option, a further link operation may change the value of |
| 4986 | a section relative symbol. On the other hand, an absolute symbol will |
| 4987 | retain the same value throughout any further link operations. |
| 4988 | |
| 4989 | Some terms in linker expressions are addresses. This is true of |
| 4990 | section relative symbols and for builtin functions that return an |
| 4991 | address, such as `ADDR', `LOADADDR', `ORIGIN' and `SEGMENT_START'. |
| 4992 | Other terms are simply numbers, or are builtin functions that return a |
| 4993 | non-address value, such as `LENGTH'. One complication is that unless |
| 4994 | you set `LD_FEATURE ("SANE_EXPR")' (*note Miscellaneous Commands::), |
| 4995 | numbers and absolute symbols are treated differently depending on their |
| 4996 | location, for compatibility with older versions of `ld'. Expressions |
| 4997 | appearing outside an output section definition treat all numbers as |
| 4998 | absolute addresses. Expressions appearing inside an output section |
| 4999 | definition treat absolute symbols as numbers. If `LD_FEATURE |
| 5000 | ("SANE_EXPR")' is given, then absolute symbols and numbers are simply |
| 5001 | treated as numbers everywhere. |
| 5002 | |
| 5003 | In the following simple example, |
| 5004 | |
| 5005 | SECTIONS |
| 5006 | { |
| 5007 | . = 0x100; |
| 5008 | __executable_start = 0x100; |
| 5009 | .data : |
| 5010 | { |
| 5011 | . = 0x10; |
| 5012 | __data_start = 0x10; |
| 5013 | *(.data) |
| 5014 | } |
| 5015 | ... |
| 5016 | } |
| 5017 | |
| 5018 | both `.' and `__executable_start' are set to the absolute address |
| 5019 | 0x100 in the first two assignments, then both `.' and `__data_start' |
| 5020 | are set to 0x10 relative to the `.data' section in the second two |
| 5021 | assignments. |
| 5022 | |
| 5023 | For expressions involving numbers, relative addresses and absolute |
| 5024 | addresses, ld follows these rules to evaluate terms: |
| 5025 | |
| 5026 | * Unary operations on an absolute address or number, and binary |
| 5027 | operations on two absolute addresses or two numbers, or between one |
| 5028 | absolute address and a number, apply the operator to the value(s). |
| 5029 | |
| 5030 | * Unary operations on a relative address, and binary operations on |
| 5031 | two relative addresses in the same section or between one relative |
| 5032 | address and a number, apply the operator to the offset part of the |
| 5033 | address(es). |
| 5034 | |
| 5035 | * Other binary operations, that is, between two relative addresses |
| 5036 | not in the same section, or between a relative address and an |
| 5037 | absolute address, first convert any non-absolute term to an |
| 5038 | absolute address before applying the operator. |
| 5039 | |
| 5040 | The result section of each sub-expression is as follows: |
| 5041 | |
| 5042 | * An operation involving only numbers results in a number. |
| 5043 | |
| 5044 | * The result of comparisons, `&&' and `||' is also a number. |
| 5045 | |
| 5046 | * The result of other binary arithmetic and logical operations on two |
| 5047 | relative addresses in the same section or two absolute addresses |
| 5048 | (after above conversions) is also a number. |
| 5049 | |
| 5050 | * The result of other operations on relative addresses or one |
| 5051 | relative address and a number, is a relative address in the same |
| 5052 | section as the relative operand(s). |
| 5053 | |
| 5054 | * The result of other operations on absolute addresses (after above |
| 5055 | conversions) is an absolute address. |
| 5056 | |
| 5057 | You can use the builtin function `ABSOLUTE' to force an expression |
| 5058 | to be absolute when it would otherwise be relative. For example, to |
| 5059 | create an absolute symbol set to the address of the end of the output |
| 5060 | section `.data': |
| 5061 | SECTIONS |
| 5062 | { |
| 5063 | .data : { *(.data) _edata = ABSOLUTE(.); } |
| 5064 | } |
| 5065 | If `ABSOLUTE' were not used, `_edata' would be relative to the |
| 5066 | `.data' section. |
| 5067 | |
| 5068 | Using `LOADADDR' also forces an expression absolute, since this |
| 5069 | particular builtin function returns an absolute address. |
| 5070 | |
| 5071 | |
| 5072 | File: ld.info, Node: Builtin Functions, Prev: Expression Section, Up: Expressions |
| 5073 | |
| 5074 | 3.10.9 Builtin Functions |
| 5075 | ------------------------ |
| 5076 | |
| 5077 | The linker script language includes a number of builtin functions for |
| 5078 | use in linker script expressions. |
| 5079 | |
| 5080 | `ABSOLUTE(EXP)' |
| 5081 | Return the absolute (non-relocatable, as opposed to non-negative) |
| 5082 | value of the expression EXP. Primarily useful to assign an |
| 5083 | absolute value to a symbol within a section definition, where |
| 5084 | symbol values are normally section relative. *Note Expression |
| 5085 | Section::. |
| 5086 | |
| 5087 | `ADDR(SECTION)' |
| 5088 | Return the address (VMA) of the named SECTION. Your script must |
| 5089 | previously have defined the location of that section. In the |
| 5090 | following example, `start_of_output_1', `symbol_1' and `symbol_2' |
| 5091 | are assigned equivalent values, except that `symbol_1' will be |
| 5092 | relative to the `.output1' section while the other two will be |
| 5093 | absolute: |
| 5094 | SECTIONS { ... |
| 5095 | .output1 : |
| 5096 | { |
| 5097 | start_of_output_1 = ABSOLUTE(.); |
| 5098 | ... |
| 5099 | } |
| 5100 | .output : |
| 5101 | { |
| 5102 | symbol_1 = ADDR(.output1); |
| 5103 | symbol_2 = start_of_output_1; |
| 5104 | } |
| 5105 | ... } |
| 5106 | |
| 5107 | `ALIGN(ALIGN)' |
| 5108 | `ALIGN(EXP,ALIGN)' |
| 5109 | Return the location counter (`.') or arbitrary expression aligned |
| 5110 | to the next ALIGN boundary. The single operand `ALIGN' doesn't |
| 5111 | change the value of the location counter--it just does arithmetic |
| 5112 | on it. The two operand `ALIGN' allows an arbitrary expression to |
| 5113 | be aligned upwards (`ALIGN(ALIGN)' is equivalent to |
| 5114 | `ALIGN(ABSOLUTE(.), ALIGN)'). |
| 5115 | |
| 5116 | Here is an example which aligns the output `.data' section to the |
| 5117 | next `0x2000' byte boundary after the preceding section and sets a |
| 5118 | variable within the section to the next `0x8000' boundary after the |
| 5119 | input sections: |
| 5120 | SECTIONS { ... |
| 5121 | .data ALIGN(0x2000): { |
| 5122 | *(.data) |
| 5123 | variable = ALIGN(0x8000); |
| 5124 | } |
| 5125 | ... } |
| 5126 | The first use of `ALIGN' in this example specifies the |
| 5127 | location of a section because it is used as the optional ADDRESS |
| 5128 | attribute of a section definition (*note Output Section |
| 5129 | Address::). The second use of `ALIGN' is used to defines the |
| 5130 | value of a symbol. |
| 5131 | |
| 5132 | The builtin function `NEXT' is closely related to `ALIGN'. |
| 5133 | |
| 5134 | `ALIGNOF(SECTION)' |
| 5135 | Return the alignment in bytes of the named SECTION, if that |
| 5136 | section has been allocated. If the section has not been allocated |
| 5137 | when this is evaluated, the linker will report an error. In the |
| 5138 | following example, the alignment of the `.output' section is |
| 5139 | stored as the first value in that section. |
| 5140 | SECTIONS{ ... |
| 5141 | .output { |
| 5142 | LONG (ALIGNOF (.output)) |
| 5143 | ... |
| 5144 | } |
| 5145 | ... } |
| 5146 | |
| 5147 | `BLOCK(EXP)' |
| 5148 | This is a synonym for `ALIGN', for compatibility with older linker |
| 5149 | scripts. It is most often seen when setting the address of an |
| 5150 | output section. |
| 5151 | |
| 5152 | `DATA_SEGMENT_ALIGN(MAXPAGESIZE, COMMONPAGESIZE)' |
| 5153 | This is equivalent to either |
| 5154 | (ALIGN(MAXPAGESIZE) + (. & (MAXPAGESIZE - 1))) |
| 5155 | or |
| 5156 | (ALIGN(MAXPAGESIZE) |
| 5157 | + ((. + COMMONPAGESIZE - 1) & (MAXPAGESIZE - COMMONPAGESIZE))) |
| 5158 | depending on whether the latter uses fewer COMMONPAGESIZE sized |
| 5159 | pages for the data segment (area between the result of this |
| 5160 | expression and `DATA_SEGMENT_END') than the former or not. If the |
| 5161 | latter form is used, it means COMMONPAGESIZE bytes of runtime |
| 5162 | memory will be saved at the expense of up to COMMONPAGESIZE wasted |
| 5163 | bytes in the on-disk file. |
| 5164 | |
| 5165 | This expression can only be used directly in `SECTIONS' commands, |
| 5166 | not in any output section descriptions and only once in the linker |
| 5167 | script. COMMONPAGESIZE should be less or equal to MAXPAGESIZE and |
| 5168 | should be the system page size the object wants to be optimized |
| 5169 | for (while still working on system page sizes up to MAXPAGESIZE). |
| 5170 | |
| 5171 | Example: |
| 5172 | . = DATA_SEGMENT_ALIGN(0x10000, 0x2000); |
| 5173 | |
| 5174 | `DATA_SEGMENT_END(EXP)' |
| 5175 | This defines the end of data segment for `DATA_SEGMENT_ALIGN' |
| 5176 | evaluation purposes. |
| 5177 | |
| 5178 | . = DATA_SEGMENT_END(.); |
| 5179 | |
| 5180 | `DATA_SEGMENT_RELRO_END(OFFSET, EXP)' |
| 5181 | This defines the end of the `PT_GNU_RELRO' segment when `-z relro' |
| 5182 | option is used. When `-z relro' option is not present, |
| 5183 | `DATA_SEGMENT_RELRO_END' does nothing, otherwise |
| 5184 | `DATA_SEGMENT_ALIGN' is padded so that EXP + OFFSET is aligned to |
| 5185 | the most commonly used page boundary for particular target. If |
| 5186 | present in the linker script, it must always come in between |
| 5187 | `DATA_SEGMENT_ALIGN' and `DATA_SEGMENT_END'. Evaluates to the |
| 5188 | second argument plus any padding needed at the end of the |
| 5189 | `PT_GNU_RELRO' segment due to section alignment. |
| 5190 | |
| 5191 | . = DATA_SEGMENT_RELRO_END(24, .); |
| 5192 | |
| 5193 | `DEFINED(SYMBOL)' |
| 5194 | Return 1 if SYMBOL is in the linker global symbol table and is |
| 5195 | defined before the statement using DEFINED in the script, otherwise |
| 5196 | return 0. You can use this function to provide default values for |
| 5197 | symbols. For example, the following script fragment shows how to |
| 5198 | set a global symbol `begin' to the first location in the `.text' |
| 5199 | section--but if a symbol called `begin' already existed, its value |
| 5200 | is preserved: |
| 5201 | |
| 5202 | SECTIONS { ... |
| 5203 | .text : { |
| 5204 | begin = DEFINED(begin) ? begin : . ; |
| 5205 | ... |
| 5206 | } |
| 5207 | ... |
| 5208 | } |
| 5209 | |
| 5210 | `LENGTH(MEMORY)' |
| 5211 | Return the length of the memory region named MEMORY. |
| 5212 | |
| 5213 | `LOADADDR(SECTION)' |
| 5214 | Return the absolute LMA of the named SECTION. (*note Output |
| 5215 | Section LMA::). |
| 5216 | |
| 5217 | `LOG2CEIL(EXP)' |
| 5218 | Return the binary logarithm of EXP rounded towards infinity. |
| 5219 | `LOG2CEIL(0)' returns 0. |
| 5220 | |
| 5221 | `MAX(EXP1, EXP2)' |
| 5222 | Returns the maximum of EXP1 and EXP2. |
| 5223 | |
| 5224 | `MIN(EXP1, EXP2)' |
| 5225 | Returns the minimum of EXP1 and EXP2. |
| 5226 | |
| 5227 | `NEXT(EXP)' |
| 5228 | Return the next unallocated address that is a multiple of EXP. |
| 5229 | This function is closely related to `ALIGN(EXP)'; unless you use |
| 5230 | the `MEMORY' command to define discontinuous memory for the output |
| 5231 | file, the two functions are equivalent. |
| 5232 | |
| 5233 | `ORIGIN(MEMORY)' |
| 5234 | Return the origin of the memory region named MEMORY. |
| 5235 | |
| 5236 | `SEGMENT_START(SEGMENT, DEFAULT)' |
| 5237 | Return the base address of the named SEGMENT. If an explicit |
| 5238 | value has already been given for this segment (with a command-line |
| 5239 | `-T' option) then that value will be returned otherwise the value |
| 5240 | will be DEFAULT. At present, the `-T' command-line option can |
| 5241 | only be used to set the base address for the "text", "data", and |
| 5242 | "bss" sections, but you can use `SEGMENT_START' with any segment |
| 5243 | name. |
| 5244 | |
| 5245 | `SIZEOF(SECTION)' |
| 5246 | Return the size in bytes of the named SECTION, if that section has |
| 5247 | been allocated. If the section has not been allocated when this is |
| 5248 | evaluated, the linker will report an error. In the following |
| 5249 | example, `symbol_1' and `symbol_2' are assigned identical values: |
| 5250 | SECTIONS{ ... |
| 5251 | .output { |
| 5252 | .start = . ; |
| 5253 | ... |
| 5254 | .end = . ; |
| 5255 | } |
| 5256 | symbol_1 = .end - .start ; |
| 5257 | symbol_2 = SIZEOF(.output); |
| 5258 | ... } |
| 5259 | |
| 5260 | `SIZEOF_HEADERS' |
| 5261 | `sizeof_headers' |
| 5262 | Return the size in bytes of the output file's headers. This is |
| 5263 | information which appears at the start of the output file. You |
| 5264 | can use this number when setting the start address of the first |
| 5265 | section, if you choose, to facilitate paging. |
| 5266 | |
| 5267 | When producing an ELF output file, if the linker script uses the |
| 5268 | `SIZEOF_HEADERS' builtin function, the linker must compute the |
| 5269 | number of program headers before it has determined all the section |
| 5270 | addresses and sizes. If the linker later discovers that it needs |
| 5271 | additional program headers, it will report an error `not enough |
| 5272 | room for program headers'. To avoid this error, you must avoid |
| 5273 | using the `SIZEOF_HEADERS' function, or you must rework your linker |
| 5274 | script to avoid forcing the linker to use additional program |
| 5275 | headers, or you must define the program headers yourself using the |
| 5276 | `PHDRS' command (*note PHDRS::). |
| 5277 | |
| 5278 | |
| 5279 | File: ld.info, Node: Implicit Linker Scripts, Prev: Expressions, Up: Scripts |
| 5280 | |
| 5281 | 3.11 Implicit Linker Scripts |
| 5282 | ============================ |
| 5283 | |
| 5284 | If you specify a linker input file which the linker can not recognize as |
| 5285 | an object file or an archive file, it will try to read the file as a |
| 5286 | linker script. If the file can not be parsed as a linker script, the |
| 5287 | linker will report an error. |
| 5288 | |
| 5289 | An implicit linker script will not replace the default linker script. |
| 5290 | |
| 5291 | Typically an implicit linker script would contain only symbol |
| 5292 | assignments, or the `INPUT', `GROUP', or `VERSION' commands. |
| 5293 | |
| 5294 | Any input files read because of an implicit linker script will be |
| 5295 | read at the position in the command line where the implicit linker |
| 5296 | script was read. This can affect archive searching. |
| 5297 | |
| 5298 | |
| 5299 | File: ld.info, Node: Machine Dependent, Next: BFD, Prev: Scripts, Up: Top |
| 5300 | |
| 5301 | 4 Machine Dependent Features |
| 5302 | **************************** |
| 5303 | |
| 5304 | `ld' has additional features on some platforms; the following sections |
| 5305 | describe them. Machines where `ld' has no additional functionality are |
| 5306 | not listed. |
| 5307 | |
| 5308 | * Menu: |
| 5309 | |
| 5310 | |
| 5311 | * H8/300:: `ld' and the H8/300 |
| 5312 | |
| 5313 | * i960:: `ld' and the Intel 960 family |
| 5314 | |
| 5315 | * M68HC11/68HC12:: `ld' and the Motorola 68HC11 and 68HC12 families |
| 5316 | |
| 5317 | * ARM:: `ld' and the ARM family |
| 5318 | |
| 5319 | * HPPA ELF32:: `ld' and HPPA 32-bit ELF |
| 5320 | |
| 5321 | * M68K:: `ld' and the Motorola 68K family |
| 5322 | |
| 5323 | * MIPS:: `ld' and the MIPS family |
| 5324 | |
| 5325 | * MMIX:: `ld' and MMIX |
| 5326 | |
| 5327 | * MSP430:: `ld' and MSP430 |
| 5328 | |
| 5329 | * NDS32:: `ld' and NDS32 |
| 5330 | |
| 5331 | * Nios II:: `ld' and the Altera Nios II |
| 5332 | |
| 5333 | * PowerPC ELF32:: `ld' and PowerPC 32-bit ELF Support |
| 5334 | |
| 5335 | * PowerPC64 ELF64:: `ld' and PowerPC64 64-bit ELF Support |
| 5336 | |
| 5337 | * SPU ELF:: `ld' and SPU ELF Support |
| 5338 | |
| 5339 | * TI COFF:: `ld' and TI COFF |
| 5340 | |
| 5341 | * WIN32:: `ld' and WIN32 (cygwin/mingw) |
| 5342 | |
| 5343 | * Xtensa:: `ld' and Xtensa Processors |
| 5344 | |
| 5345 | |
| 5346 | File: ld.info, Node: H8/300, Next: i960, Up: Machine Dependent |
| 5347 | |
| 5348 | 4.1 `ld' and the H8/300 |
| 5349 | ======================= |
| 5350 | |
| 5351 | For the H8/300, `ld' can perform these global optimizations when you |
| 5352 | specify the `--relax' command-line option. |
| 5353 | |
| 5354 | _relaxing address modes_ |
| 5355 | `ld' finds all `jsr' and `jmp' instructions whose targets are |
| 5356 | within eight bits, and turns them into eight-bit program-counter |
| 5357 | relative `bsr' and `bra' instructions, respectively. |
| 5358 | |
| 5359 | _synthesizing instructions_ |
| 5360 | `ld' finds all `mov.b' instructions which use the sixteen-bit |
| 5361 | absolute address form, but refer to the top page of memory, and |
| 5362 | changes them to use the eight-bit address form. (That is: the |
| 5363 | linker turns `mov.b `@'AA:16' into `mov.b `@'AA:8' whenever the |
| 5364 | address AA is in the top page of memory). |
| 5365 | |
| 5366 | `ld' finds all `mov' instructions which use the register indirect |
| 5367 | with 32-bit displacement addressing mode, but use a small |
| 5368 | displacement inside 16-bit displacement range, and changes them to |
| 5369 | use the 16-bit displacement form. (That is: the linker turns |
| 5370 | `mov.b `@'D:32,ERx' into `mov.b `@'D:16,ERx' whenever the |
| 5371 | displacement D is in the 16 bit signed integer range. Only |
| 5372 | implemented in ELF-format ld). |
| 5373 | |
| 5374 | _bit manipulation instructions_ |
| 5375 | `ld' finds all bit manipulation instructions like `band, bclr, |
| 5376 | biand, bild, bior, bist, bixor, bld, bnot, bor, bset, bst, btst, |
| 5377 | bxor' which use 32 bit and 16 bit absolute address form, but refer |
| 5378 | to the top page of memory, and changes them to use the 8 bit |
| 5379 | address form. (That is: the linker turns `bset #xx:3,`@'AA:32' |
| 5380 | into `bset #xx:3,`@'AA:8' whenever the address AA is in the top |
| 5381 | page of memory). |
| 5382 | |
| 5383 | _system control instructions_ |
| 5384 | `ld' finds all `ldc.w, stc.w' instructions which use the 32 bit |
| 5385 | absolute address form, but refer to the top page of memory, and |
| 5386 | changes them to use 16 bit address form. (That is: the linker |
| 5387 | turns `ldc.w `@'AA:32,ccr' into `ldc.w `@'AA:16,ccr' whenever the |
| 5388 | address AA is in the top page of memory). |
| 5389 | |
| 5390 | |
| 5391 | File: ld.info, Node: i960, Next: M68HC11/68HC12, Prev: H8/300, Up: Machine Dependent |
| 5392 | |
| 5393 | 4.2 `ld' and the Intel 960 Family |
| 5394 | ================================= |
| 5395 | |
| 5396 | You can use the `-AARCHITECTURE' command line option to specify one of |
| 5397 | the two-letter names identifying members of the 960 family; the option |
| 5398 | specifies the desired output target, and warns of any incompatible |
| 5399 | instructions in the input files. It also modifies the linker's search |
| 5400 | strategy for archive libraries, to support the use of libraries |
| 5401 | specific to each particular architecture, by including in the search |
| 5402 | loop names suffixed with the string identifying the architecture. |
| 5403 | |
| 5404 | For example, if your `ld' command line included `-ACA' as well as |
| 5405 | `-ltry', the linker would look (in its built-in search paths, and in |
| 5406 | any paths you specify with `-L') for a library with the names |
| 5407 | |
| 5408 | try |
| 5409 | libtry.a |
| 5410 | tryca |
| 5411 | libtryca.a |
| 5412 | |
| 5413 | The first two possibilities would be considered in any event; the last |
| 5414 | two are due to the use of `-ACA'. |
| 5415 | |
| 5416 | You can meaningfully use `-A' more than once on a command line, since |
| 5417 | the 960 architecture family allows combination of target architectures; |
| 5418 | each use will add another pair of name variants to search for when `-l' |
| 5419 | specifies a library. |
| 5420 | |
| 5421 | `ld' supports the `--relax' option for the i960 family. If you |
| 5422 | specify `--relax', `ld' finds all `balx' and `calx' instructions whose |
| 5423 | targets are within 24 bits, and turns them into 24-bit program-counter |
| 5424 | relative `bal' and `cal' instructions, respectively. `ld' also turns |
| 5425 | `cal' instructions into `bal' instructions when it determines that the |
| 5426 | target subroutine is a leaf routine (that is, the target subroutine does |
| 5427 | not itself call any subroutines). |
| 5428 | |
| 5429 | |
| 5430 | File: ld.info, Node: M68HC11/68HC12, Next: ARM, Prev: i960, Up: Machine Dependent |
| 5431 | |
| 5432 | 4.3 `ld' and the Motorola 68HC11 and 68HC12 families |
| 5433 | ==================================================== |
| 5434 | |
| 5435 | 4.3.1 Linker Relaxation |
| 5436 | ----------------------- |
| 5437 | |
| 5438 | For the Motorola 68HC11, `ld' can perform these global optimizations |
| 5439 | when you specify the `--relax' command-line option. |
| 5440 | |
| 5441 | _relaxing address modes_ |
| 5442 | `ld' finds all `jsr' and `jmp' instructions whose targets are |
| 5443 | within eight bits, and turns them into eight-bit program-counter |
| 5444 | relative `bsr' and `bra' instructions, respectively. |
| 5445 | |
| 5446 | `ld' also looks at all 16-bit extended addressing modes and |
| 5447 | transforms them in a direct addressing mode when the address is in |
| 5448 | page 0 (between 0 and 0x0ff). |
| 5449 | |
| 5450 | _relaxing gcc instruction group_ |
| 5451 | When `gcc' is called with `-mrelax', it can emit group of |
| 5452 | instructions that the linker can optimize to use a 68HC11 direct |
| 5453 | addressing mode. These instructions consists of `bclr' or `bset' |
| 5454 | instructions. |
| 5455 | |
| 5456 | |
| 5457 | 4.3.2 Trampoline Generation |
| 5458 | --------------------------- |
| 5459 | |
| 5460 | For 68HC11 and 68HC12, `ld' can generate trampoline code to call a far |
| 5461 | function using a normal `jsr' instruction. The linker will also change |
| 5462 | the relocation to some far function to use the trampoline address |
| 5463 | instead of the function address. This is typically the case when a |
| 5464 | pointer to a function is taken. The pointer will in fact point to the |
| 5465 | function trampoline. |
| 5466 | |
| 5467 | |
| 5468 | File: ld.info, Node: ARM, Next: HPPA ELF32, Prev: M68HC11/68HC12, Up: Machine Dependent |
| 5469 | |
| 5470 | 4.4 `ld' and the ARM family |
| 5471 | =========================== |
| 5472 | |
| 5473 | For the ARM, `ld' will generate code stubs to allow functions calls |
| 5474 | between ARM and Thumb code. These stubs only work with code that has |
| 5475 | been compiled and assembled with the `-mthumb-interwork' command line |
| 5476 | option. If it is necessary to link with old ARM object files or |
| 5477 | libraries, which have not been compiled with the -mthumb-interwork |
| 5478 | option then the `--support-old-code' command line switch should be |
| 5479 | given to the linker. This will make it generate larger stub functions |
| 5480 | which will work with non-interworking aware ARM code. Note, however, |
| 5481 | the linker does not support generating stubs for function calls to |
| 5482 | non-interworking aware Thumb code. |
| 5483 | |
| 5484 | The `--thumb-entry' switch is a duplicate of the generic `--entry' |
| 5485 | switch, in that it sets the program's starting address. But it also |
| 5486 | sets the bottom bit of the address, so that it can be branched to using |
| 5487 | a BX instruction, and the program will start executing in Thumb mode |
| 5488 | straight away. |
| 5489 | |
| 5490 | The `--use-nul-prefixed-import-tables' switch is specifying, that |
| 5491 | the import tables idata4 and idata5 have to be generated with a zero |
| 5492 | element prefix for import libraries. This is the old style to generate |
| 5493 | import tables. By default this option is turned off. |
| 5494 | |
| 5495 | The `--be8' switch instructs `ld' to generate BE8 format |
| 5496 | executables. This option is only valid when linking big-endian objects |
| 5497 | - ie ones which have been assembled with the `-EB' option. The |
| 5498 | resulting image will contain big-endian data and little-endian code. |
| 5499 | |
| 5500 | The `R_ARM_TARGET1' relocation is typically used for entries in the |
| 5501 | `.init_array' section. It is interpreted as either `R_ARM_REL32' or |
| 5502 | `R_ARM_ABS32', depending on the target. The `--target1-rel' and |
| 5503 | `--target1-abs' switches override the default. |
| 5504 | |
| 5505 | The `--target2=type' switch overrides the default definition of the |
| 5506 | `R_ARM_TARGET2' relocation. Valid values for `type', their meanings, |
| 5507 | and target defaults are as follows: |
| 5508 | `rel' |
| 5509 | `R_ARM_REL32' (arm*-*-elf, arm*-*-eabi) |
| 5510 | |
| 5511 | `abs' |
| 5512 | `R_ARM_ABS32' (arm*-*-symbianelf) |
| 5513 | |
| 5514 | `got-rel' |
| 5515 | `R_ARM_GOT_PREL' (arm*-*-linux, arm*-*-*bsd) |
| 5516 | |
| 5517 | The `R_ARM_V4BX' relocation (defined by the ARM AAELF specification) |
| 5518 | enables objects compiled for the ARMv4 architecture to be |
| 5519 | interworking-safe when linked with other objects compiled for ARMv4t, |
| 5520 | but also allows pure ARMv4 binaries to be built from the same ARMv4 |
| 5521 | objects. |
| 5522 | |
| 5523 | In the latter case, the switch `--fix-v4bx' must be passed to the |
| 5524 | linker, which causes v4t `BX rM' instructions to be rewritten as `MOV |
| 5525 | PC,rM', since v4 processors do not have a `BX' instruction. |
| 5526 | |
| 5527 | In the former case, the switch should not be used, and `R_ARM_V4BX' |
| 5528 | relocations are ignored. |
| 5529 | |
| 5530 | Replace `BX rM' instructions identified by `R_ARM_V4BX' relocations |
| 5531 | with a branch to the following veneer: |
| 5532 | |
| 5533 | TST rM, #1 |
| 5534 | MOVEQ PC, rM |
| 5535 | BX Rn |
| 5536 | |
| 5537 | This allows generation of libraries/applications that work on ARMv4 |
| 5538 | cores and are still interworking safe. Note that the above veneer |
| 5539 | clobbers the condition flags, so may cause incorrect program behavior |
| 5540 | in rare cases. |
| 5541 | |
| 5542 | The `--use-blx' switch enables the linker to use ARM/Thumb BLX |
| 5543 | instructions (available on ARMv5t and above) in various situations. |
| 5544 | Currently it is used to perform calls via the PLT from Thumb code using |
| 5545 | BLX rather than using BX and a mode-switching stub before each PLT |
| 5546 | entry. This should lead to such calls executing slightly faster. |
| 5547 | |
| 5548 | This option is enabled implicitly for SymbianOS, so there is no need |
| 5549 | to specify it if you are using that target. |
| 5550 | |
| 5551 | The `--vfp11-denorm-fix' switch enables a link-time workaround for a |
| 5552 | bug in certain VFP11 coprocessor hardware, which sometimes allows |
| 5553 | instructions with denorm operands (which must be handled by support |
| 5554 | code) to have those operands overwritten by subsequent instructions |
| 5555 | before the support code can read the intended values. |
| 5556 | |
| 5557 | The bug may be avoided in scalar mode if you allow at least one |
| 5558 | intervening instruction between a VFP11 instruction which uses a |
| 5559 | register and another instruction which writes to the same register, or |
| 5560 | at least two intervening instructions if vector mode is in use. The bug |
| 5561 | only affects full-compliance floating-point mode: you do not need this |
| 5562 | workaround if you are using "runfast" mode. Please contact ARM for |
| 5563 | further details. |
| 5564 | |
| 5565 | If you know you are using buggy VFP11 hardware, you can enable this |
| 5566 | workaround by specifying the linker option `--vfp-denorm-fix=scalar' if |
| 5567 | you are using the VFP11 scalar mode only, or `--vfp-denorm-fix=vector' |
| 5568 | if you are using vector mode (the latter also works for scalar code). |
| 5569 | The default is `--vfp-denorm-fix=none'. |
| 5570 | |
| 5571 | If the workaround is enabled, instructions are scanned for |
| 5572 | potentially-troublesome sequences, and a veneer is created for each |
| 5573 | such sequence which may trigger the erratum. The veneer consists of the |
| 5574 | first instruction of the sequence and a branch back to the subsequent |
| 5575 | instruction. The original instruction is then replaced with a branch to |
| 5576 | the veneer. The extra cycles required to call and return from the veneer |
| 5577 | are sufficient to avoid the erratum in both the scalar and vector cases. |
| 5578 | |
| 5579 | The `--fix-arm1176' switch enables a link-time workaround for an |
| 5580 | erratum in certain ARM1176 processors. The workaround is enabled by |
| 5581 | default if you are targeting ARM v6 (excluding ARM v6T2) or earlier. |
| 5582 | It can be disabled unconditionally by specifying `--no-fix-arm1176'. |
| 5583 | |
| 5584 | Further information is available in the "ARM1176JZ-S and ARM1176JZF-S |
| 5585 | Programmer Advice Notice" available on the ARM documentation website at: |
| 5586 | http://infocenter.arm.com/. |
| 5587 | |
| 5588 | The `--fix-stm32l4xx-629360' switch enables a link-time workaround |
| 5589 | for a bug in the bus matrix / memory controller for some of the STM32 |
| 5590 | Cortex-M4 based products (STM32L4xx). When accessing off-chip memory |
| 5591 | via the affected bus for bus reads of 9 words or more, the bus can |
| 5592 | generate corrupt data and/or abort. These are only core-initiated |
| 5593 | accesses (not DMA), and might affect any access: integer loads such as |
| 5594 | LDM, POP and floating-point loads such as VLDM, VPOP. Stores are not |
| 5595 | affected. |
| 5596 | |
| 5597 | The bug can be avoided by splitting memory accesses into the |
| 5598 | necessary chunks to keep bus reads below 8 words. |
| 5599 | |
| 5600 | The workaround is not enabled by default, this is equivalent to use |
| 5601 | `--fix-stm32l4xx-629360=none'. If you know you are using buggy |
| 5602 | STM32L4xx hardware, you can enable the workaround by specifying the |
| 5603 | linker option `--fix-stm32l4xx-629360', or the equivalent |
| 5604 | `--fix-stm32l4xx-629360=default'. |
| 5605 | |
| 5606 | If the workaround is enabled, instructions are scanned for |
| 5607 | potentially-troublesome sequences, and a veneer is created for each |
| 5608 | such sequence which may trigger the erratum. The veneer consists in a |
| 5609 | replacement sequence emulating the behaviour of the original one and a |
| 5610 | branch back to the subsequent instruction. The original instruction is |
| 5611 | then replaced with a branch to the veneer. |
| 5612 | |
| 5613 | The workaround does not always preserve the memory access order for |
| 5614 | the LDMDB instruction, when the instruction loads the PC. |
| 5615 | |
| 5616 | The workaround is not able to handle problematic instructions when |
| 5617 | they are in the middle of an IT block, since a branch is not allowed |
| 5618 | there. In that case, the linker reports a warning and no replacement |
| 5619 | occurs. |
| 5620 | |
| 5621 | The workaround is not able to replace problematic instructions with a |
| 5622 | PC-relative branch instruction if the `.text' section is too large. In |
| 5623 | that case, when the branch that replaces the original code cannot be |
| 5624 | encoded, the linker reports a warning and no replacement occurs. |
| 5625 | |
| 5626 | The `--no-enum-size-warning' switch prevents the linker from warning |
| 5627 | when linking object files that specify incompatible EABI enumeration |
| 5628 | size attributes. For example, with this switch enabled, linking of an |
| 5629 | object file using 32-bit enumeration values with another using |
| 5630 | enumeration values fitted into the smallest possible space will not be |
| 5631 | diagnosed. |
| 5632 | |
| 5633 | The `--no-wchar-size-warning' switch prevents the linker from |
| 5634 | warning when linking object files that specify incompatible EABI |
| 5635 | `wchar_t' size attributes. For example, with this switch enabled, |
| 5636 | linking of an object file using 32-bit `wchar_t' values with another |
| 5637 | using 16-bit `wchar_t' values will not be diagnosed. |
| 5638 | |
| 5639 | The `--pic-veneer' switch makes the linker use PIC sequences for |
| 5640 | ARM/Thumb interworking veneers, even if the rest of the binary is not |
| 5641 | PIC. This avoids problems on uClinux targets where `--emit-relocs' is |
| 5642 | used to generate relocatable binaries. |
| 5643 | |
| 5644 | The linker will automatically generate and insert small sequences of |
| 5645 | code into a linked ARM ELF executable whenever an attempt is made to |
| 5646 | perform a function call to a symbol that is too far away. The |
| 5647 | placement of these sequences of instructions - called stubs - is |
| 5648 | controlled by the command line option `--stub-group-size=N'. The |
| 5649 | placement is important because a poor choice can create a need for |
| 5650 | duplicate stubs, increasing the code size. The linker will try to |
| 5651 | group stubs together in order to reduce interruptions to the flow of |
| 5652 | code, but it needs guidance as to how big these groups should be and |
| 5653 | where they should be placed. |
| 5654 | |
| 5655 | The value of `N', the parameter to the `--stub-group-size=' option |
| 5656 | controls where the stub groups are placed. If it is negative then all |
| 5657 | stubs are placed after the first branch that needs them. If it is |
| 5658 | positive then the stubs can be placed either before or after the |
| 5659 | branches that need them. If the value of `N' is 1 (either +1 or -1) |
| 5660 | then the linker will choose exactly where to place groups of stubs, |
| 5661 | using its built in heuristics. A value of `N' greater than 1 (or |
| 5662 | smaller than -1) tells the linker that a single group of stubs can |
| 5663 | service at most `N' bytes from the input sections. |
| 5664 | |
| 5665 | The default, if `--stub-group-size=' is not specified, is `N = +1'. |
| 5666 | |
| 5667 | Farcalls stubs insertion is fully supported for the ARM-EABI target |
| 5668 | only, because it relies on object files properties not present |
| 5669 | otherwise. |
| 5670 | |
| 5671 | The `--fix-cortex-a8' switch enables a link-time workaround for an |
| 5672 | erratum in certain Cortex-A8 processors. The workaround is enabled by |
| 5673 | default if you are targeting the ARM v7-A architecture profile. It can |
| 5674 | be enabled otherwise by specifying `--fix-cortex-a8', or disabled |
| 5675 | unconditionally by specifying `--no-fix-cortex-a8'. |
| 5676 | |
| 5677 | The erratum only affects Thumb-2 code. Please contact ARM for |
| 5678 | further details. |
| 5679 | |
| 5680 | The `--fix-cortex-a53-835769' switch enables a link-time workaround |
| 5681 | for erratum 835769 present on certain early revisions of Cortex-A53 |
| 5682 | processors. The workaround is disabled by default. It can be enabled |
| 5683 | by specifying `--fix-cortex-a53-835769', or disabled unconditionally by |
| 5684 | specifying `--no-fix-cortex-a53-835769'. |
| 5685 | |
| 5686 | Please contact ARM for further details. |
| 5687 | |
| 5688 | The `--no-merge-exidx-entries' switch disables the merging of |
| 5689 | adjacent exidx entries in debuginfo. |
| 5690 | |
| 5691 | The `--long-plt' option enables the use of 16 byte PLT entries which |
| 5692 | support up to 4Gb of code. The default is to use 12 byte PLT entries |
| 5693 | which only support 512Mb of code. |
| 5694 | |
| 5695 | The `--no-apply-dynamic-relocs' option makes AArch64 linker do not |
| 5696 | apply link-time values for dynamic relocations. |
| 5697 | |
| 5698 | |
| 5699 | File: ld.info, Node: HPPA ELF32, Next: M68K, Prev: ARM, Up: Machine Dependent |
| 5700 | |
| 5701 | 4.5 `ld' and HPPA 32-bit ELF Support |
| 5702 | ==================================== |
| 5703 | |
| 5704 | When generating a shared library, `ld' will by default generate import |
| 5705 | stubs suitable for use with a single sub-space application. The |
| 5706 | `--multi-subspace' switch causes `ld' to generate export stubs, and |
| 5707 | different (larger) import stubs suitable for use with multiple |
| 5708 | sub-spaces. |
| 5709 | |
| 5710 | Long branch stubs and import/export stubs are placed by `ld' in stub |
| 5711 | sections located between groups of input sections. `--stub-group-size' |
| 5712 | specifies the maximum size of a group of input sections handled by one |
| 5713 | stub section. Since branch offsets are signed, a stub section may |
| 5714 | serve two groups of input sections, one group before the stub section, |
| 5715 | and one group after it. However, when using conditional branches that |
| 5716 | require stubs, it may be better (for branch prediction) that stub |
| 5717 | sections only serve one group of input sections. A negative value for |
| 5718 | `N' chooses this scheme, ensuring that branches to stubs always use a |
| 5719 | negative offset. Two special values of `N' are recognized, `1' and |
| 5720 | `-1'. These both instruct `ld' to automatically size input section |
| 5721 | groups for the branch types detected, with the same behaviour regarding |
| 5722 | stub placement as other positive or negative values of `N' respectively. |
| 5723 | |
| 5724 | Note that `--stub-group-size' does not split input sections. A |
| 5725 | single input section larger than the group size specified will of course |
| 5726 | create a larger group (of one section). If input sections are too |
| 5727 | large, it may not be possible for a branch to reach its stub. |
| 5728 | |
| 5729 | |
| 5730 | File: ld.info, Node: M68K, Next: MIPS, Prev: HPPA ELF32, Up: Machine Dependent |
| 5731 | |
| 5732 | 4.6 `ld' and the Motorola 68K family |
| 5733 | ==================================== |
| 5734 | |
| 5735 | The `--got=TYPE' option lets you choose the GOT generation scheme. The |
| 5736 | choices are `single', `negative', `multigot' and `target'. When |
| 5737 | `target' is selected the linker chooses the default GOT generation |
| 5738 | scheme for the current target. `single' tells the linker to generate a |
| 5739 | single GOT with entries only at non-negative offsets. `negative' |
| 5740 | instructs the linker to generate a single GOT with entries at both |
| 5741 | negative and positive offsets. Not all environments support such GOTs. |
| 5742 | `multigot' allows the linker to generate several GOTs in the output |
| 5743 | file. All GOT references from a single input object file access the |
| 5744 | same GOT, but references from different input object files might access |
| 5745 | different GOTs. Not all environments support such GOTs. |
| 5746 | |
| 5747 | |
| 5748 | File: ld.info, Node: MIPS, Next: MMIX, Prev: M68K, Up: Machine Dependent |
| 5749 | |
| 5750 | 4.7 `ld' and the MIPS family |
| 5751 | ============================ |
| 5752 | |
| 5753 | The `--insn32' and `--no-insn32' options control the choice of |
| 5754 | microMIPS instructions used in code generated by the linker, such as |
| 5755 | that in the PLT or lazy binding stubs, or in relaxation. If `--insn32' |
| 5756 | is used, then the linker only uses 32-bit instruction encodings. By |
| 5757 | default or if `--no-insn32' is used, all instruction encodings are used, |
| 5758 | including 16-bit ones where possible. |
| 5759 | |
| 5760 | |
| 5761 | File: ld.info, Node: MMIX, Next: MSP430, Prev: MIPS, Up: Machine Dependent |
| 5762 | |
| 5763 | 4.8 `ld' and MMIX |
| 5764 | ================= |
| 5765 | |
| 5766 | For MMIX, there is a choice of generating `ELF' object files or `mmo' |
| 5767 | object files when linking. The simulator `mmix' understands the `mmo' |
| 5768 | format. The binutils `objcopy' utility can translate between the two |
| 5769 | formats. |
| 5770 | |
| 5771 | There is one special section, the `.MMIX.reg_contents' section. |
| 5772 | Contents in this section is assumed to correspond to that of global |
| 5773 | registers, and symbols referring to it are translated to special |
| 5774 | symbols, equal to registers. In a final link, the start address of the |
| 5775 | `.MMIX.reg_contents' section corresponds to the first allocated global |
| 5776 | register multiplied by 8. Register `$255' is not included in this |
| 5777 | section; it is always set to the program entry, which is at the symbol |
| 5778 | `Main' for `mmo' files. |
| 5779 | |
| 5780 | Global symbols with the prefix `__.MMIX.start.', for example |
| 5781 | `__.MMIX.start..text' and `__.MMIX.start..data' are special. The |
| 5782 | default linker script uses these to set the default start address of a |
| 5783 | section. |
| 5784 | |
| 5785 | Initial and trailing multiples of zero-valued 32-bit words in a |
| 5786 | section, are left out from an mmo file. |
| 5787 | |
| 5788 | |
| 5789 | File: ld.info, Node: MSP430, Next: NDS32, Prev: MMIX, Up: Machine Dependent |
| 5790 | |
| 5791 | 4.9 `ld' and MSP430 |
| 5792 | =================== |
| 5793 | |
| 5794 | For the MSP430 it is possible to select the MPU architecture. The flag |
| 5795 | `-m [mpu type]' will select an appropriate linker script for selected |
| 5796 | MPU type. (To get a list of known MPUs just pass `-m help' option to |
| 5797 | the linker). |
| 5798 | |
| 5799 | The linker will recognize some extra sections which are MSP430 |
| 5800 | specific: |
| 5801 | |
| 5802 | ``.vectors'' |
| 5803 | Defines a portion of ROM where interrupt vectors located. |
| 5804 | |
| 5805 | ``.bootloader'' |
| 5806 | Defines the bootloader portion of the ROM (if applicable). Any |
| 5807 | code in this section will be uploaded to the MPU. |
| 5808 | |
| 5809 | ``.infomem'' |
| 5810 | Defines an information memory section (if applicable). Any code in |
| 5811 | this section will be uploaded to the MPU. |
| 5812 | |
| 5813 | ``.infomemnobits'' |
| 5814 | This is the same as the `.infomem' section except that any code in |
| 5815 | this section will not be uploaded to the MPU. |
| 5816 | |
| 5817 | ``.noinit'' |
| 5818 | Denotes a portion of RAM located above `.bss' section. |
| 5819 | |
| 5820 | The last two sections are used by gcc. |
| 5821 | |
| 5822 | |
| 5823 | File: ld.info, Node: NDS32, Next: Nios II, Prev: MSP430, Up: Machine Dependent |
| 5824 | |
| 5825 | 4.10 `ld' and NDS32 |
| 5826 | =================== |
| 5827 | |
| 5828 | For NDS32, there are some options to select relaxation behavior. The |
| 5829 | linker relaxes objects according to these options. |
| 5830 | |
| 5831 | ``--m[no-]fp-as-gp'' |
| 5832 | Disable/enable fp-as-gp relaxation. |
| 5833 | |
| 5834 | ``--mexport-symbols=FILE'' |
| 5835 | Exporting symbols and their address into FILE as linker script. |
| 5836 | |
| 5837 | ``--m[no-]ex9'' |
| 5838 | Disable/enable link-time EX9 relaxation. |
| 5839 | |
| 5840 | ``--mexport-ex9=FILE'' |
| 5841 | Export the EX9 table after linking. |
| 5842 | |
| 5843 | ``--mimport-ex9=FILE'' |
| 5844 | Import the Ex9 table for EX9 relaxation. |
| 5845 | |
| 5846 | ``--mupdate-ex9'' |
| 5847 | Update the existing EX9 table. |
| 5848 | |
| 5849 | ``--mex9-limit=NUM'' |
| 5850 | Maximum number of entries in the ex9 table. |
| 5851 | |
| 5852 | ``--mex9-loop-aware'' |
| 5853 | Avoid generating the EX9 instruction inside the loop. |
| 5854 | |
| 5855 | ``--m[no-]ifc'' |
| 5856 | Disable/enable the link-time IFC optimization. |
| 5857 | |
| 5858 | ``--mifc-loop-aware'' |
| 5859 | Avoid generating the IFC instruction inside the loop. |
| 5860 | |
| 5861 | |
| 5862 | File: ld.info, Node: Nios II, Next: PowerPC ELF32, Prev: NDS32, Up: Machine Dependent |
| 5863 | |
| 5864 | 4.11 `ld' and the Altera Nios II |
| 5865 | ================================ |
| 5866 | |
| 5867 | Call and immediate jump instructions on Nios II processors are limited |
| 5868 | to transferring control to addresses in the same 256MB memory segment, |
| 5869 | which may result in `ld' giving `relocation truncated to fit' errors |
| 5870 | with very large programs. The command-line option `--relax' enables |
| 5871 | the generation of trampolines that can access the entire 32-bit address |
| 5872 | space for calls outside the normal `call' and `jmpi' address range. |
| 5873 | These trampolines are inserted at section boundaries, so may not |
| 5874 | themselves be reachable if an input section and its associated call |
| 5875 | trampolines are larger than 256MB. |
| 5876 | |
| 5877 | The `--relax' option is enabled by default unless `-r' is also |
| 5878 | specified. You can disable trampoline generation by using the |
| 5879 | `--no-relax' linker option. You can also disable this optimization |
| 5880 | locally by using the `set .noat' directive in assembly-language source |
| 5881 | files, as the linker-inserted trampolines use the `at' register as a |
| 5882 | temporary. |
| 5883 | |
| 5884 | Note that the linker `--relax' option is independent of assembler |
| 5885 | relaxation options, and that using the GNU assembler's `-relax-all' |
| 5886 | option interferes with the linker's more selective call instruction |
| 5887 | relaxation. |
| 5888 | |
| 5889 | |
| 5890 | File: ld.info, Node: PowerPC ELF32, Next: PowerPC64 ELF64, Prev: Nios II, Up: Machine Dependent |
| 5891 | |
| 5892 | 4.12 `ld' and PowerPC 32-bit ELF Support |
| 5893 | ======================================== |
| 5894 | |
| 5895 | Branches on PowerPC processors are limited to a signed 26-bit |
| 5896 | displacement, which may result in `ld' giving `relocation truncated to |
| 5897 | fit' errors with very large programs. `--relax' enables the generation |
| 5898 | of trampolines that can access the entire 32-bit address space. These |
| 5899 | trampolines are inserted at section boundaries, so may not themselves |
| 5900 | be reachable if an input section exceeds 33M in size. You may combine |
| 5901 | `-r' and `--relax' to add trampolines in a partial link. In that case |
| 5902 | both branches to undefined symbols and inter-section branches are also |
| 5903 | considered potentially out of range, and trampolines inserted. |
| 5904 | |
| 5905 | `--bss-plt' |
| 5906 | Current PowerPC GCC accepts a `-msecure-plt' option that generates |
| 5907 | code capable of using a newer PLT and GOT layout that has the |
| 5908 | security advantage of no executable section ever needing to be |
| 5909 | writable and no writable section ever being executable. PowerPC |
| 5910 | `ld' will generate this layout, including stubs to access the PLT, |
| 5911 | if all input files (including startup and static libraries) were |
| 5912 | compiled with `-msecure-plt'. `--bss-plt' forces the old BSS PLT |
| 5913 | (and GOT layout) which can give slightly better performance. |
| 5914 | |
| 5915 | `--secure-plt' |
| 5916 | `ld' will use the new PLT and GOT layout if it is linking new |
| 5917 | `-fpic' or `-fPIC' code, but does not do so automatically when |
| 5918 | linking non-PIC code. This option requests the new PLT and GOT |
| 5919 | layout. A warning will be given if some object file requires the |
| 5920 | old style BSS PLT. |
| 5921 | |
| 5922 | `--sdata-got' |
| 5923 | The new secure PLT and GOT are placed differently relative to other |
| 5924 | sections compared to older BSS PLT and GOT placement. The |
| 5925 | location of `.plt' must change because the new secure PLT is an |
| 5926 | initialized section while the old PLT is uninitialized. The |
| 5927 | reason for the `.got' change is more subtle: The new placement |
| 5928 | allows `.got' to be read-only in applications linked with `-z |
| 5929 | relro -z now'. However, this placement means that `.sdata' cannot |
| 5930 | always be used in shared libraries, because the PowerPC ABI |
| 5931 | accesses `.sdata' in shared libraries from the GOT pointer. |
| 5932 | `--sdata-got' forces the old GOT placement. PowerPC GCC doesn't |
| 5933 | use `.sdata' in shared libraries, so this option is really only |
| 5934 | useful for other compilers that may do so. |
| 5935 | |
| 5936 | `--emit-stub-syms' |
| 5937 | This option causes `ld' to label linker stubs with a local symbol |
| 5938 | that encodes the stub type and destination. |
| 5939 | |
| 5940 | `--no-tls-optimize' |
| 5941 | PowerPC `ld' normally performs some optimization of code sequences |
| 5942 | used to access Thread-Local Storage. Use this option to disable |
| 5943 | the optimization. |
| 5944 | |
| 5945 | |
| 5946 | File: ld.info, Node: PowerPC64 ELF64, Next: SPU ELF, Prev: PowerPC ELF32, Up: Machine Dependent |
| 5947 | |
| 5948 | 4.13 `ld' and PowerPC64 64-bit ELF Support |
| 5949 | ========================================== |
| 5950 | |
| 5951 | `--stub-group-size' |
| 5952 | Long branch stubs, PLT call stubs and TOC adjusting stubs are |
| 5953 | placed by `ld' in stub sections located between groups of input |
| 5954 | sections. `--stub-group-size' specifies the maximum size of a |
| 5955 | group of input sections handled by one stub section. Since branch |
| 5956 | offsets are signed, a stub section may serve two groups of input |
| 5957 | sections, one group before the stub section, and one group after |
| 5958 | it. However, when using conditional branches that require stubs, |
| 5959 | it may be better (for branch prediction) that stub sections only |
| 5960 | serve one group of input sections. A negative value for `N' |
| 5961 | chooses this scheme, ensuring that branches to stubs always use a |
| 5962 | negative offset. Two special values of `N' are recognized, `1' |
| 5963 | and `-1'. These both instruct `ld' to automatically size input |
| 5964 | section groups for the branch types detected, with the same |
| 5965 | behaviour regarding stub placement as other positive or negative |
| 5966 | values of `N' respectively. |
| 5967 | |
| 5968 | Note that `--stub-group-size' does not split input sections. A |
| 5969 | single input section larger than the group size specified will of |
| 5970 | course create a larger group (of one section). If input sections |
| 5971 | are too large, it may not be possible for a branch to reach its |
| 5972 | stub. |
| 5973 | |
| 5974 | `--emit-stub-syms' |
| 5975 | This option causes `ld' to label linker stubs with a local symbol |
| 5976 | that encodes the stub type and destination. |
| 5977 | |
| 5978 | `--dotsyms' |
| 5979 | `--no-dotsyms' |
| 5980 | These two options control how `ld' interprets version patterns in |
| 5981 | a version script. Older PowerPC64 compilers emitted both a |
| 5982 | function descriptor symbol with the same name as the function, and |
| 5983 | a code entry symbol with the name prefixed by a dot (`.'). To |
| 5984 | properly version a function `foo', the version script thus needs |
| 5985 | to control both `foo' and `.foo'. The option `--dotsyms', on by |
| 5986 | default, automatically adds the required dot-prefixed patterns. |
| 5987 | Use `--no-dotsyms' to disable this feature. |
| 5988 | |
| 5989 | `--save-restore-funcs' |
| 5990 | `--no-save-restore-funcs' |
| 5991 | These two options control whether PowerPC64 `ld' automatically |
| 5992 | provides out-of-line register save and restore functions used by |
| 5993 | `-Os' code. The default is to provide any such referenced |
| 5994 | function for a normal final link, and to not do so for a |
| 5995 | relocatable link. |
| 5996 | |
| 5997 | `--no-tls-optimize' |
| 5998 | PowerPC64 `ld' normally performs some optimization of code |
| 5999 | sequences used to access Thread-Local Storage. Use this option to |
| 6000 | disable the optimization. |
| 6001 | |
| 6002 | `--tls-get-addr-optimize' |
| 6003 | `--no-tls-get-addr-optimize' |
| 6004 | These options control whether PowerPC64 `ld' uses a special stub |
| 6005 | to call __tls_get_addr. PowerPC64 glibc 2.22 and later support an |
| 6006 | optimization that allows the second and subsequent calls to |
| 6007 | `__tls_get_addr' for a given symbol to be resolved by the special |
| 6008 | stub without calling in to glibc. By default the linker enables |
| 6009 | this option when glibc advertises the availability of |
| 6010 | __tls_get_addr_opt. Forcing this option on when using an older |
| 6011 | glibc won't do much besides slow down your applications, but may |
| 6012 | be useful if linking an application against an older glibc with |
| 6013 | the expectation that it will normally be used on systems having a |
| 6014 | newer glibc. |
| 6015 | |
| 6016 | `--no-opd-optimize' |
| 6017 | PowerPC64 `ld' normally removes `.opd' section entries |
| 6018 | corresponding to deleted link-once functions, or functions removed |
| 6019 | by the action of `--gc-sections' or linker script `/DISCARD/'. |
| 6020 | Use this option to disable `.opd' optimization. |
| 6021 | |
| 6022 | `--non-overlapping-opd' |
| 6023 | Some PowerPC64 compilers have an option to generate compressed |
| 6024 | `.opd' entries spaced 16 bytes apart, overlapping the third word, |
| 6025 | the static chain pointer (unused in C) with the first word of the |
| 6026 | next entry. This option expands such entries to the full 24 bytes. |
| 6027 | |
| 6028 | `--no-toc-optimize' |
| 6029 | PowerPC64 `ld' normally removes unused `.toc' section entries. |
| 6030 | Such entries are detected by examining relocations that reference |
| 6031 | the TOC in code sections. A reloc in a deleted code section marks |
| 6032 | a TOC word as unneeded, while a reloc in a kept code section marks |
| 6033 | a TOC word as needed. Since the TOC may reference itself, TOC |
| 6034 | relocs are also examined. TOC words marked as both needed and |
| 6035 | unneeded will of course be kept. TOC words without any referencing |
| 6036 | reloc are assumed to be part of a multi-word entry, and are kept or |
| 6037 | discarded as per the nearest marked preceding word. This works |
| 6038 | reliably for compiler generated code, but may be incorrect if |
| 6039 | assembly code is used to insert TOC entries. Use this option to |
| 6040 | disable the optimization. |
| 6041 | |
| 6042 | `--no-multi-toc' |
| 6043 | If given any toc option besides `-mcmodel=medium' or |
| 6044 | `-mcmodel=large', PowerPC64 GCC generates code for a TOC model |
| 6045 | where TOC entries are accessed with a 16-bit offset from r2. This |
| 6046 | limits the total TOC size to 64K. PowerPC64 `ld' extends this |
| 6047 | limit by grouping code sections such that each group uses less |
| 6048 | than 64K for its TOC entries, then inserts r2 adjusting stubs |
| 6049 | between inter-group calls. `ld' does not split apart input |
| 6050 | sections, so cannot help if a single input file has a `.toc' |
| 6051 | section that exceeds 64K, most likely from linking multiple files |
| 6052 | with `ld -r'. Use this option to turn off this feature. |
| 6053 | |
| 6054 | `--no-toc-sort' |
| 6055 | By default, `ld' sorts TOC sections so that those whose file |
| 6056 | happens to have a section called `.init' or `.fini' are placed |
| 6057 | first, followed by TOC sections referenced by code generated with |
| 6058 | PowerPC64 gcc's `-mcmodel=small', and lastly TOC sections |
| 6059 | referenced only by code generated with PowerPC64 gcc's |
| 6060 | `-mcmodel=medium' or `-mcmodel=large' options. Doing this results |
| 6061 | in better TOC grouping for multi-TOC. Use this option to turn off |
| 6062 | this feature. |
| 6063 | |
| 6064 | `--plt-align' |
| 6065 | `--no-plt-align' |
| 6066 | Use these options to control whether individual PLT call stubs are |
| 6067 | padded so that they don't cross a 32-byte boundary, or to the |
| 6068 | specified power of two boundary when using `--plt-align='. Note |
| 6069 | that this isn't alignment in the usual sense. By default PLT call |
| 6070 | stubs are packed tightly. |
| 6071 | |
| 6072 | `--plt-static-chain' |
| 6073 | `--no-plt-static-chain' |
| 6074 | Use these options to control whether PLT call stubs load the static |
| 6075 | chain pointer (r11). `ld' defaults to not loading the static |
| 6076 | chain since there is never any need to do so on a PLT call. |
| 6077 | |
| 6078 | `--plt-thread-safe' |
| 6079 | `--no-thread-safe' |
| 6080 | With power7's weakly ordered memory model, it is possible when |
| 6081 | using lazy binding for ld.so to update a plt entry in one thread |
| 6082 | and have another thread see the individual plt entry words update |
| 6083 | in the wrong order, despite ld.so carefully writing in the correct |
| 6084 | order and using memory write barriers. To avoid this we need some |
| 6085 | sort of read barrier in the call stub, or use LD_BIND_NOW=1. By |
| 6086 | default, `ld' looks for calls to commonly used functions that |
| 6087 | create threads, and if seen, adds the necessary barriers. Use |
| 6088 | these options to change the default behaviour. |
| 6089 | |
| 6090 | |
| 6091 | File: ld.info, Node: SPU ELF, Next: TI COFF, Prev: PowerPC64 ELF64, Up: Machine Dependent |
| 6092 | |
| 6093 | 4.14 `ld' and SPU ELF Support |
| 6094 | ============================= |
| 6095 | |
| 6096 | `--plugin' |
| 6097 | This option marks an executable as a PIC plugin module. |
| 6098 | |
| 6099 | `--no-overlays' |
| 6100 | Normally, `ld' recognizes calls to functions within overlay |
| 6101 | regions, and redirects such calls to an overlay manager via a stub. |
| 6102 | `ld' also provides a built-in overlay manager. This option turns |
| 6103 | off all this special overlay handling. |
| 6104 | |
| 6105 | `--emit-stub-syms' |
| 6106 | This option causes `ld' to label overlay stubs with a local symbol |
| 6107 | that encodes the stub type and destination. |
| 6108 | |
| 6109 | `--extra-overlay-stubs' |
| 6110 | This option causes `ld' to add overlay call stubs on all function |
| 6111 | calls out of overlay regions. Normally stubs are not added on |
| 6112 | calls to non-overlay regions. |
| 6113 | |
| 6114 | `--local-store=lo:hi' |
| 6115 | `ld' usually checks that a final executable for SPU fits in the |
| 6116 | address range 0 to 256k. This option may be used to change the |
| 6117 | range. Disable the check entirely with `--local-store=0:0'. |
| 6118 | |
| 6119 | `--stack-analysis' |
| 6120 | SPU local store space is limited. Over-allocation of stack space |
| 6121 | unnecessarily limits space available for code and data, while |
| 6122 | under-allocation results in runtime failures. If given this |
| 6123 | option, `ld' will provide an estimate of maximum stack usage. |
| 6124 | `ld' does this by examining symbols in code sections to determine |
| 6125 | the extents of functions, and looking at function prologues for |
| 6126 | stack adjusting instructions. A call-graph is created by looking |
| 6127 | for relocations on branch instructions. The graph is then searched |
| 6128 | for the maximum stack usage path. Note that this analysis does not |
| 6129 | find calls made via function pointers, and does not handle |
| 6130 | recursion and other cycles in the call graph. Stack usage may be |
| 6131 | under-estimated if your code makes such calls. Also, stack usage |
| 6132 | for dynamic allocation, e.g. alloca, will not be detected. If a |
| 6133 | link map is requested, detailed information about each function's |
| 6134 | stack usage and calls will be given. |
| 6135 | |
| 6136 | `--emit-stack-syms' |
| 6137 | This option, if given along with `--stack-analysis' will result in |
| 6138 | `ld' emitting stack sizing symbols for each function. These take |
| 6139 | the form `__stack_<function_name>' for global functions, and |
| 6140 | `__stack_<number>_<function_name>' for static functions. |
| 6141 | `<number>' is the section id in hex. The value of such symbols is |
| 6142 | the stack requirement for the corresponding function. The symbol |
| 6143 | size will be zero, type `STT_NOTYPE', binding `STB_LOCAL', and |
| 6144 | section `SHN_ABS'. |
| 6145 | |
| 6146 | |
| 6147 | File: ld.info, Node: TI COFF, Next: WIN32, Prev: SPU ELF, Up: Machine Dependent |
| 6148 | |
| 6149 | 4.15 `ld''s Support for Various TI COFF Versions |
| 6150 | ================================================ |
| 6151 | |
| 6152 | The `--format' switch allows selection of one of the various TI COFF |
| 6153 | versions. The latest of this writing is 2; versions 0 and 1 are also |
| 6154 | supported. The TI COFF versions also vary in header byte-order format; |
| 6155 | `ld' will read any version or byte order, but the output header format |
| 6156 | depends on the default specified by the specific target. |
| 6157 | |
| 6158 | |
| 6159 | File: ld.info, Node: WIN32, Next: Xtensa, Prev: TI COFF, Up: Machine Dependent |
| 6160 | |
| 6161 | 4.16 `ld' and WIN32 (cygwin/mingw) |
| 6162 | ================================== |
| 6163 | |
| 6164 | This section describes some of the win32 specific `ld' issues. See |
| 6165 | *Note Command Line Options: Options. for detailed description of the |
| 6166 | command line options mentioned here. |
| 6167 | |
| 6168 | _import libraries_ |
| 6169 | The standard Windows linker creates and uses so-called import |
| 6170 | libraries, which contains information for linking to dll's. They |
| 6171 | are regular static archives and are handled as any other static |
| 6172 | archive. The cygwin and mingw ports of `ld' have specific support |
| 6173 | for creating such libraries provided with the `--out-implib' |
| 6174 | command line option. |
| 6175 | |
| 6176 | _exporting DLL symbols_ |
| 6177 | The cygwin/mingw `ld' has several ways to export symbols for dll's. |
| 6178 | |
| 6179 | _using auto-export functionality_ |
| 6180 | By default `ld' exports symbols with the auto-export |
| 6181 | functionality, which is controlled by the following command |
| 6182 | line options: |
| 6183 | |
| 6184 | * -export-all-symbols [This is the default] |
| 6185 | |
| 6186 | * -exclude-symbols |
| 6187 | |
| 6188 | * -exclude-libs |
| 6189 | |
| 6190 | * -exclude-modules-for-implib |
| 6191 | |
| 6192 | * -version-script |
| 6193 | |
| 6194 | When auto-export is in operation, `ld' will export all the |
| 6195 | non-local (global and common) symbols it finds in a DLL, with |
| 6196 | the exception of a few symbols known to belong to the |
| 6197 | system's runtime and libraries. As it will often not be |
| 6198 | desirable to export all of a DLL's symbols, which may include |
| 6199 | private functions that are not part of any public interface, |
| 6200 | the command-line options listed above may be used to filter |
| 6201 | symbols out from the list for exporting. The `--output-def' |
| 6202 | option can be used in order to see the final list of exported |
| 6203 | symbols with all exclusions taken into effect. |
| 6204 | |
| 6205 | If `--export-all-symbols' is not given explicitly on the |
| 6206 | command line, then the default auto-export behavior will be |
| 6207 | _disabled_ if either of the following are true: |
| 6208 | |
| 6209 | * A DEF file is used. |
| 6210 | |
| 6211 | * Any symbol in any object file was marked with the |
| 6212 | __declspec(dllexport) attribute. |
| 6213 | |
| 6214 | _using a DEF file_ |
| 6215 | Another way of exporting symbols is using a DEF file. A DEF |
| 6216 | file is an ASCII file containing definitions of symbols which |
| 6217 | should be exported when a dll is created. Usually it is |
| 6218 | named `<dll name>.def' and is added as any other object file |
| 6219 | to the linker's command line. The file's name must end in |
| 6220 | `.def' or `.DEF'. |
| 6221 | |
| 6222 | gcc -o <output> <objectfiles> <dll name>.def |
| 6223 | |
| 6224 | Using a DEF file turns off the normal auto-export behavior, |
| 6225 | unless the `--export-all-symbols' option is also used. |
| 6226 | |
| 6227 | Here is an example of a DEF file for a shared library called |
| 6228 | `xyz.dll': |
| 6229 | |
| 6230 | LIBRARY "xyz.dll" BASE=0x20000000 |
| 6231 | |
| 6232 | EXPORTS |
| 6233 | foo |
| 6234 | bar |
| 6235 | _bar = bar |
| 6236 | another_foo = abc.dll.afoo |
| 6237 | var1 DATA |
| 6238 | doo = foo == foo2 |
| 6239 | eoo DATA == var1 |
| 6240 | |
| 6241 | This example defines a DLL with a non-default base address |
| 6242 | and seven symbols in the export table. The third exported |
| 6243 | symbol `_bar' is an alias for the second. The fourth symbol, |
| 6244 | `another_foo' is resolved by "forwarding" to another module |
| 6245 | and treating it as an alias for `afoo' exported from the DLL |
| 6246 | `abc.dll'. The final symbol `var1' is declared to be a data |
| 6247 | object. The `doo' symbol in export library is an alias of |
| 6248 | `foo', which gets the string name in export table `foo2'. The |
| 6249 | `eoo' symbol is an data export symbol, which gets in export |
| 6250 | table the name `var1'. |
| 6251 | |
| 6252 | The optional `LIBRARY <name>' command indicates the _internal_ |
| 6253 | name of the output DLL. If `<name>' does not include a suffix, |
| 6254 | the default library suffix, `.DLL' is appended. |
| 6255 | |
| 6256 | When the .DEF file is used to build an application, rather |
| 6257 | than a library, the `NAME <name>' command should be used |
| 6258 | instead of `LIBRARY'. If `<name>' does not include a suffix, |
| 6259 | the default executable suffix, `.EXE' is appended. |
| 6260 | |
| 6261 | With either `LIBRARY <name>' or `NAME <name>' the optional |
| 6262 | specification `BASE = <number>' may be used to specify a |
| 6263 | non-default base address for the image. |
| 6264 | |
| 6265 | If neither `LIBRARY <name>' nor `NAME <name>' is specified, |
| 6266 | or they specify an empty string, the internal name is the |
| 6267 | same as the filename specified on the command line. |
| 6268 | |
| 6269 | The complete specification of an export symbol is: |
| 6270 | |
| 6271 | EXPORTS |
| 6272 | ( ( ( <name1> [ = <name2> ] ) |
| 6273 | | ( <name1> = <module-name> . <external-name>)) |
| 6274 | [ @ <integer> ] [NONAME] [DATA] [CONSTANT] [PRIVATE] [== <name3>] ) * |
| 6275 | |
| 6276 | Declares `<name1>' as an exported symbol from the DLL, or |
| 6277 | declares `<name1>' as an exported alias for `<name2>'; or |
| 6278 | declares `<name1>' as a "forward" alias for the symbol |
| 6279 | `<external-name>' in the DLL `<module-name>'. Optionally, |
| 6280 | the symbol may be exported by the specified ordinal |
| 6281 | `<integer>' alias. The optional `<name3>' is the to be used |
| 6282 | string in import/export table for the symbol. |
| 6283 | |
| 6284 | The optional keywords that follow the declaration indicate: |
| 6285 | |
| 6286 | `NONAME': Do not put the symbol name in the DLL's export |
| 6287 | table. It will still be exported by its ordinal alias |
| 6288 | (either the value specified by the .def specification or, |
| 6289 | otherwise, the value assigned by the linker). The symbol |
| 6290 | name, however, does remain visible in the import library (if |
| 6291 | any), unless `PRIVATE' is also specified. |
| 6292 | |
| 6293 | `DATA': The symbol is a variable or object, rather than a |
| 6294 | function. The import lib will export only an indirect |
| 6295 | reference to `foo' as the symbol `_imp__foo' (ie, `foo' must |
| 6296 | be resolved as `*_imp__foo'). |
| 6297 | |
| 6298 | `CONSTANT': Like `DATA', but put the undecorated `foo' as |
| 6299 | well as `_imp__foo' into the import library. Both refer to the |
| 6300 | read-only import address table's pointer to the variable, not |
| 6301 | to the variable itself. This can be dangerous. If the user |
| 6302 | code fails to add the `dllimport' attribute and also fails to |
| 6303 | explicitly add the extra indirection that the use of the |
| 6304 | attribute enforces, the application will behave unexpectedly. |
| 6305 | |
| 6306 | `PRIVATE': Put the symbol in the DLL's export table, but do |
| 6307 | not put it into the static import library used to resolve |
| 6308 | imports at link time. The symbol can still be imported using |
| 6309 | the `LoadLibrary/GetProcAddress' API at runtime or by by |
| 6310 | using the GNU ld extension of linking directly to the DLL |
| 6311 | without an import library. |
| 6312 | |
| 6313 | See ld/deffilep.y in the binutils sources for the full |
| 6314 | specification of other DEF file statements |
| 6315 | |
| 6316 | While linking a shared dll, `ld' is able to create a DEF file |
| 6317 | with the `--output-def <file>' command line option. |
| 6318 | |
| 6319 | _Using decorations_ |
| 6320 | Another way of marking symbols for export is to modify the |
| 6321 | source code itself, so that when building the DLL each symbol |
| 6322 | to be exported is declared as: |
| 6323 | |
| 6324 | __declspec(dllexport) int a_variable |
| 6325 | __declspec(dllexport) void a_function(int with_args) |
| 6326 | |
| 6327 | All such symbols will be exported from the DLL. If, however, |
| 6328 | any of the object files in the DLL contain symbols decorated |
| 6329 | in this way, then the normal auto-export behavior is |
| 6330 | disabled, unless the `--export-all-symbols' option is also |
| 6331 | used. |
| 6332 | |
| 6333 | Note that object files that wish to access these symbols must |
| 6334 | _not_ decorate them with dllexport. Instead, they should use |
| 6335 | dllimport, instead: |
| 6336 | |
| 6337 | __declspec(dllimport) int a_variable |
| 6338 | __declspec(dllimport) void a_function(int with_args) |
| 6339 | |
| 6340 | This complicates the structure of library header files, |
| 6341 | because when included by the library itself the header must |
| 6342 | declare the variables and functions as dllexport, but when |
| 6343 | included by client code the header must declare them as |
| 6344 | dllimport. There are a number of idioms that are typically |
| 6345 | used to do this; often client code can omit the __declspec() |
| 6346 | declaration completely. See `--enable-auto-import' and |
| 6347 | `automatic data imports' for more information. |
| 6348 | |
| 6349 | _automatic data imports_ |
| 6350 | The standard Windows dll format supports data imports from dlls |
| 6351 | only by adding special decorations (dllimport/dllexport), which |
| 6352 | let the compiler produce specific assembler instructions to deal |
| 6353 | with this issue. This increases the effort necessary to port |
| 6354 | existing Un*x code to these platforms, especially for large c++ |
| 6355 | libraries and applications. The auto-import feature, which was |
| 6356 | initially provided by Paul Sokolovsky, allows one to omit the |
| 6357 | decorations to achieve a behavior that conforms to that on |
| 6358 | POSIX/Un*x platforms. This feature is enabled with the |
| 6359 | `--enable-auto-import' command-line option, although it is enabled |
| 6360 | by default on cygwin/mingw. The `--enable-auto-import' option |
| 6361 | itself now serves mainly to suppress any warnings that are |
| 6362 | ordinarily emitted when linked objects trigger the feature's use. |
| 6363 | |
| 6364 | auto-import of variables does not always work flawlessly without |
| 6365 | additional assistance. Sometimes, you will see this message |
| 6366 | |
| 6367 | "variable '<var>' can't be auto-imported. Please read the |
| 6368 | documentation for ld's `--enable-auto-import' for details." |
| 6369 | |
| 6370 | The `--enable-auto-import' documentation explains why this error |
| 6371 | occurs, and several methods that can be used to overcome this |
| 6372 | difficulty. One of these methods is the _runtime pseudo-relocs_ |
| 6373 | feature, described below. |
| 6374 | |
| 6375 | For complex variables imported from DLLs (such as structs or |
| 6376 | classes), object files typically contain a base address for the |
| 6377 | variable and an offset (_addend_) within the variable-to specify a |
| 6378 | particular field or public member, for instance. Unfortunately, |
| 6379 | the runtime loader used in win32 environments is incapable of |
| 6380 | fixing these references at runtime without the additional |
| 6381 | information supplied by dllimport/dllexport decorations. The |
| 6382 | standard auto-import feature described above is unable to resolve |
| 6383 | these references. |
| 6384 | |
| 6385 | The `--enable-runtime-pseudo-relocs' switch allows these |
| 6386 | references to be resolved without error, while leaving the task of |
| 6387 | adjusting the references themselves (with their non-zero addends) |
| 6388 | to specialized code provided by the runtime environment. Recent |
| 6389 | versions of the cygwin and mingw environments and compilers |
| 6390 | provide this runtime support; older versions do not. However, the |
| 6391 | support is only necessary on the developer's platform; the |
| 6392 | compiled result will run without error on an older system. |
| 6393 | |
| 6394 | `--enable-runtime-pseudo-relocs' is not the default; it must be |
| 6395 | explicitly enabled as needed. |
| 6396 | |
| 6397 | _direct linking to a dll_ |
| 6398 | The cygwin/mingw ports of `ld' support the direct linking, |
| 6399 | including data symbols, to a dll without the usage of any import |
| 6400 | libraries. This is much faster and uses much less memory than |
| 6401 | does the traditional import library method, especially when |
| 6402 | linking large libraries or applications. When `ld' creates an |
| 6403 | import lib, each function or variable exported from the dll is |
| 6404 | stored in its own bfd, even though a single bfd could contain many |
| 6405 | exports. The overhead involved in storing, loading, and |
| 6406 | processing so many bfd's is quite large, and explains the |
| 6407 | tremendous time, memory, and storage needed to link against |
| 6408 | particularly large or complex libraries when using import libs. |
| 6409 | |
| 6410 | Linking directly to a dll uses no extra command-line switches |
| 6411 | other than `-L' and `-l', because `ld' already searches for a |
| 6412 | number of names to match each library. All that is needed from |
| 6413 | the developer's perspective is an understanding of this search, in |
| 6414 | order to force ld to select the dll instead of an import library. |
| 6415 | |
| 6416 | For instance, when ld is called with the argument `-lxxx' it will |
| 6417 | attempt to find, in the first directory of its search path, |
| 6418 | |
| 6419 | libxxx.dll.a |
| 6420 | xxx.dll.a |
| 6421 | libxxx.a |
| 6422 | xxx.lib |
| 6423 | cygxxx.dll (*) |
| 6424 | libxxx.dll |
| 6425 | xxx.dll |
| 6426 | |
| 6427 | before moving on to the next directory in the search path. |
| 6428 | |
| 6429 | (*) Actually, this is not `cygxxx.dll' but in fact is |
| 6430 | `<prefix>xxx.dll', where `<prefix>' is set by the `ld' option |
| 6431 | `--dll-search-prefix=<prefix>'. In the case of cygwin, the |
| 6432 | standard gcc spec file includes `--dll-search-prefix=cyg', so in |
| 6433 | effect we actually search for `cygxxx.dll'. |
| 6434 | |
| 6435 | Other win32-based unix environments, such as mingw or pw32, may |
| 6436 | use other `<prefix>'es, although at present only cygwin makes use |
| 6437 | of this feature. It was originally intended to help avoid name |
| 6438 | conflicts among dll's built for the various win32/un*x |
| 6439 | environments, so that (for example) two versions of a zlib dll |
| 6440 | could coexist on the same machine. |
| 6441 | |
| 6442 | The generic cygwin/mingw path layout uses a `bin' directory for |
| 6443 | applications and dll's and a `lib' directory for the import |
| 6444 | libraries (using cygwin nomenclature): |
| 6445 | |
| 6446 | bin/ |
| 6447 | cygxxx.dll |
| 6448 | lib/ |
| 6449 | libxxx.dll.a (in case of dll's) |
| 6450 | libxxx.a (in case of static archive) |
| 6451 | |
| 6452 | Linking directly to a dll without using the import library can be |
| 6453 | done two ways: |
| 6454 | |
| 6455 | 1. Use the dll directly by adding the `bin' path to the link line |
| 6456 | gcc -Wl,-verbose -o a.exe -L../bin/ -lxxx |
| 6457 | |
| 6458 | However, as the dll's often have version numbers appended to their |
| 6459 | names (`cygncurses-5.dll') this will often fail, unless one |
| 6460 | specifies `-L../bin -lncurses-5' to include the version. Import |
| 6461 | libs are generally not versioned, and do not have this difficulty. |
| 6462 | |
| 6463 | 2. Create a symbolic link from the dll to a file in the `lib' |
| 6464 | directory according to the above mentioned search pattern. This |
| 6465 | should be used to avoid unwanted changes in the tools needed for |
| 6466 | making the app/dll. |
| 6467 | |
| 6468 | ln -s bin/cygxxx.dll lib/[cyg|lib|]xxx.dll[.a] |
| 6469 | |
| 6470 | Then you can link without any make environment changes. |
| 6471 | |
| 6472 | gcc -Wl,-verbose -o a.exe -L../lib/ -lxxx |
| 6473 | |
| 6474 | This technique also avoids the version number problems, because |
| 6475 | the following is perfectly legal |
| 6476 | |
| 6477 | bin/ |
| 6478 | cygxxx-5.dll |
| 6479 | lib/ |
| 6480 | libxxx.dll.a -> ../bin/cygxxx-5.dll |
| 6481 | |
| 6482 | Linking directly to a dll without using an import lib will work |
| 6483 | even when auto-import features are exercised, and even when |
| 6484 | `--enable-runtime-pseudo-relocs' is used. |
| 6485 | |
| 6486 | Given the improvements in speed and memory usage, one might |
| 6487 | justifiably wonder why import libraries are used at all. There |
| 6488 | are three reasons: |
| 6489 | |
| 6490 | 1. Until recently, the link-directly-to-dll functionality did _not_ |
| 6491 | work with auto-imported data. |
| 6492 | |
| 6493 | 2. Sometimes it is necessary to include pure static objects within |
| 6494 | the import library (which otherwise contains only bfd's for |
| 6495 | indirection symbols that point to the exports of a dll). Again, |
| 6496 | the import lib for the cygwin kernel makes use of this ability, |
| 6497 | and it is not possible to do this without an import lib. |
| 6498 | |
| 6499 | 3. Symbol aliases can only be resolved using an import lib. This |
| 6500 | is critical when linking against OS-supplied dll's (eg, the win32 |
| 6501 | API) in which symbols are usually exported as undecorated aliases |
| 6502 | of their stdcall-decorated assembly names. |
| 6503 | |
| 6504 | So, import libs are not going away. But the ability to replace |
| 6505 | true import libs with a simple symbolic link to (or a copy of) a |
| 6506 | dll, in many cases, is a useful addition to the suite of tools |
| 6507 | binutils makes available to the win32 developer. Given the |
| 6508 | massive improvements in memory requirements during linking, storage |
| 6509 | requirements, and linking speed, we expect that many developers |
| 6510 | will soon begin to use this feature whenever possible. |
| 6511 | |
| 6512 | _symbol aliasing_ |
| 6513 | |
| 6514 | _adding additional names_ |
| 6515 | Sometimes, it is useful to export symbols with additional |
| 6516 | names. A symbol `foo' will be exported as `foo', but it can |
| 6517 | also be exported as `_foo' by using special directives in the |
| 6518 | DEF file when creating the dll. This will affect also the |
| 6519 | optional created import library. Consider the following DEF |
| 6520 | file: |
| 6521 | |
| 6522 | LIBRARY "xyz.dll" BASE=0x61000000 |
| 6523 | |
| 6524 | EXPORTS |
| 6525 | foo |
| 6526 | _foo = foo |
| 6527 | |
| 6528 | The line `_foo = foo' maps the symbol `foo' to `_foo'. |
| 6529 | |
| 6530 | Another method for creating a symbol alias is to create it in |
| 6531 | the source code using the "weak" attribute: |
| 6532 | |
| 6533 | void foo () { /* Do something. */; } |
| 6534 | void _foo () __attribute__ ((weak, alias ("foo"))); |
| 6535 | |
| 6536 | See the gcc manual for more information about attributes and |
| 6537 | weak symbols. |
| 6538 | |
| 6539 | _renaming symbols_ |
| 6540 | Sometimes it is useful to rename exports. For instance, the |
| 6541 | cygwin kernel does this regularly. A symbol `_foo' can be |
| 6542 | exported as `foo' but not as `_foo' by using special |
| 6543 | directives in the DEF file. (This will also affect the import |
| 6544 | library, if it is created). In the following example: |
| 6545 | |
| 6546 | LIBRARY "xyz.dll" BASE=0x61000000 |
| 6547 | |
| 6548 | EXPORTS |
| 6549 | _foo = foo |
| 6550 | |
| 6551 | The line `_foo = foo' maps the exported symbol `foo' to |
| 6552 | `_foo'. |
| 6553 | |
| 6554 | Note: using a DEF file disables the default auto-export behavior, |
| 6555 | unless the `--export-all-symbols' command line option is used. |
| 6556 | If, however, you are trying to rename symbols, then you should list |
| 6557 | _all_ desired exports in the DEF file, including the symbols that |
| 6558 | are not being renamed, and do _not_ use the `--export-all-symbols' |
| 6559 | option. If you list only the renamed symbols in the DEF file, and |
| 6560 | use `--export-all-symbols' to handle the other symbols, then the |
| 6561 | both the new names _and_ the original names for the renamed |
| 6562 | symbols will be exported. In effect, you'd be aliasing those |
| 6563 | symbols, not renaming them, which is probably not what you wanted. |
| 6564 | |
| 6565 | _weak externals_ |
| 6566 | The Windows object format, PE, specifies a form of weak symbols |
| 6567 | called weak externals. When a weak symbol is linked and the |
| 6568 | symbol is not defined, the weak symbol becomes an alias for some |
| 6569 | other symbol. There are three variants of weak externals: |
| 6570 | * Definition is searched for in objects and libraries, |
| 6571 | historically called lazy externals. |
| 6572 | |
| 6573 | * Definition is searched for only in other objects, not in |
| 6574 | libraries. This form is not presently implemented. |
| 6575 | |
| 6576 | * No search; the symbol is an alias. This form is not presently |
| 6577 | implemented. |
| 6578 | As a GNU extension, weak symbols that do not specify an alternate |
| 6579 | symbol are supported. If the symbol is undefined when linking, |
| 6580 | the symbol uses a default value. |
| 6581 | |
| 6582 | _aligned common symbols_ |
| 6583 | As a GNU extension to the PE file format, it is possible to |
| 6584 | specify the desired alignment for a common symbol. This |
| 6585 | information is conveyed from the assembler or compiler to the |
| 6586 | linker by means of GNU-specific commands carried in the object |
| 6587 | file's `.drectve' section, which are recognized by `ld' and |
| 6588 | respected when laying out the common symbols. Native tools will |
| 6589 | be able to process object files employing this GNU extension, but |
| 6590 | will fail to respect the alignment instructions, and may issue |
| 6591 | noisy warnings about unknown linker directives. |
| 6592 | |
| 6593 | |
| 6594 | |
| 6595 | File: ld.info, Node: Xtensa, Prev: WIN32, Up: Machine Dependent |
| 6596 | |
| 6597 | 4.17 `ld' and Xtensa Processors |
| 6598 | =============================== |
| 6599 | |
| 6600 | The default `ld' behavior for Xtensa processors is to interpret |
| 6601 | `SECTIONS' commands so that lists of explicitly named sections in a |
| 6602 | specification with a wildcard file will be interleaved when necessary to |
| 6603 | keep literal pools within the range of PC-relative load offsets. For |
| 6604 | example, with the command: |
| 6605 | |
| 6606 | SECTIONS |
| 6607 | { |
| 6608 | .text : { |
| 6609 | *(.literal .text) |
| 6610 | } |
| 6611 | } |
| 6612 | |
| 6613 | `ld' may interleave some of the `.literal' and `.text' sections from |
| 6614 | different object files to ensure that the literal pools are within the |
| 6615 | range of PC-relative load offsets. A valid interleaving might place |
| 6616 | the `.literal' sections from an initial group of files followed by the |
| 6617 | `.text' sections of that group of files. Then, the `.literal' sections |
| 6618 | from the rest of the files and the `.text' sections from the rest of |
| 6619 | the files would follow. |
| 6620 | |
| 6621 | Relaxation is enabled by default for the Xtensa version of `ld' and |
| 6622 | provides two important link-time optimizations. The first optimization |
| 6623 | is to combine identical literal values to reduce code size. A redundant |
| 6624 | literal will be removed and all the `L32R' instructions that use it |
| 6625 | will be changed to reference an identical literal, as long as the |
| 6626 | location of the replacement literal is within the offset range of all |
| 6627 | the `L32R' instructions. The second optimization is to remove |
| 6628 | unnecessary overhead from assembler-generated "longcall" sequences of |
| 6629 | `L32R'/`CALLXN' when the target functions are within range of direct |
| 6630 | `CALLN' instructions. |
| 6631 | |
| 6632 | For each of these cases where an indirect call sequence can be |
| 6633 | optimized to a direct call, the linker will change the `CALLXN' |
| 6634 | instruction to a `CALLN' instruction, remove the `L32R' instruction, |
| 6635 | and remove the literal referenced by the `L32R' instruction if it is |
| 6636 | not used for anything else. Removing the `L32R' instruction always |
| 6637 | reduces code size but can potentially hurt performance by changing the |
| 6638 | alignment of subsequent branch targets. By default, the linker will |
| 6639 | always preserve alignments, either by switching some instructions |
| 6640 | between 24-bit encodings and the equivalent density instructions or by |
| 6641 | inserting a no-op in place of the `L32R' instruction that was removed. |
| 6642 | If code size is more important than performance, the `--size-opt' |
| 6643 | option can be used to prevent the linker from widening density |
| 6644 | instructions or inserting no-ops, except in a few cases where no-ops |
| 6645 | are required for correctness. |
| 6646 | |
| 6647 | The following Xtensa-specific command-line options can be used to |
| 6648 | control the linker: |
| 6649 | |
| 6650 | `--size-opt' |
| 6651 | When optimizing indirect calls to direct calls, optimize for code |
| 6652 | size more than performance. With this option, the linker will not |
| 6653 | insert no-ops or widen density instructions to preserve branch |
| 6654 | target alignment. There may still be some cases where no-ops are |
| 6655 | required to preserve the correctness of the code. |
| 6656 | |
| 6657 | |
| 6658 | File: ld.info, Node: BFD, Next: Reporting Bugs, Prev: Machine Dependent, Up: Top |
| 6659 | |
| 6660 | 5 BFD |
| 6661 | ***** |
| 6662 | |
| 6663 | The linker accesses object and archive files using the BFD libraries. |
| 6664 | These libraries allow the linker to use the same routines to operate on |
| 6665 | object files whatever the object file format. A different object file |
| 6666 | format can be supported simply by creating a new BFD back end and adding |
| 6667 | it to the library. To conserve runtime memory, however, the linker and |
| 6668 | associated tools are usually configured to support only a subset of the |
| 6669 | object file formats available. You can use `objdump -i' (*note |
| 6670 | objdump: (binutils.info)objdump.) to list all the formats available for |
| 6671 | your configuration. |
| 6672 | |
| 6673 | As with most implementations, BFD is a compromise between several |
| 6674 | conflicting requirements. The major factor influencing BFD design was |
| 6675 | efficiency: any time used converting between formats is time which |
| 6676 | would not have been spent had BFD not been involved. This is partly |
| 6677 | offset by abstraction payback; since BFD simplifies applications and |
| 6678 | back ends, more time and care may be spent optimizing algorithms for a |
| 6679 | greater speed. |
| 6680 | |
| 6681 | One minor artifact of the BFD solution which you should bear in mind |
| 6682 | is the potential for information loss. There are two places where |
| 6683 | useful information can be lost using the BFD mechanism: during |
| 6684 | conversion and during output. *Note BFD information loss::. |
| 6685 | |
| 6686 | * Menu: |
| 6687 | |
| 6688 | * BFD outline:: How it works: an outline of BFD |
| 6689 | |
| 6690 | |
| 6691 | File: ld.info, Node: BFD outline, Up: BFD |
| 6692 | |
| 6693 | 5.1 How It Works: An Outline of BFD |
| 6694 | =================================== |
| 6695 | |
| 6696 | When an object file is opened, BFD subroutines automatically determine |
| 6697 | the format of the input object file. They then build a descriptor in |
| 6698 | memory with pointers to routines that will be used to access elements of |
| 6699 | the object file's data structures. |
| 6700 | |
| 6701 | As different information from the object files is required, BFD |
| 6702 | reads from different sections of the file and processes them. For |
| 6703 | example, a very common operation for the linker is processing symbol |
| 6704 | tables. Each BFD back end provides a routine for converting between |
| 6705 | the object file's representation of symbols and an internal canonical |
| 6706 | format. When the linker asks for the symbol table of an object file, it |
| 6707 | calls through a memory pointer to the routine from the relevant BFD |
| 6708 | back end which reads and converts the table into a canonical form. The |
| 6709 | linker then operates upon the canonical form. When the link is finished |
| 6710 | and the linker writes the output file's symbol table, another BFD back |
| 6711 | end routine is called to take the newly created symbol table and |
| 6712 | convert it into the chosen output format. |
| 6713 | |
| 6714 | * Menu: |
| 6715 | |
| 6716 | * BFD information loss:: Information Loss |
| 6717 | * Canonical format:: The BFD canonical object-file format |
| 6718 | |
| 6719 | |
| 6720 | File: ld.info, Node: BFD information loss, Next: Canonical format, Up: BFD outline |
| 6721 | |
| 6722 | 5.1.1 Information Loss |
| 6723 | ---------------------- |
| 6724 | |
| 6725 | _Information can be lost during output._ The output formats supported |
| 6726 | by BFD do not provide identical facilities, and information which can |
| 6727 | be described in one form has nowhere to go in another format. One |
| 6728 | example of this is alignment information in `b.out'. There is nowhere |
| 6729 | in an `a.out' format file to store alignment information on the |
| 6730 | contained data, so when a file is linked from `b.out' and an `a.out' |
| 6731 | image is produced, alignment information will not propagate to the |
| 6732 | output file. (The linker will still use the alignment information |
| 6733 | internally, so the link is performed correctly). |
| 6734 | |
| 6735 | Another example is COFF section names. COFF files may contain an |
| 6736 | unlimited number of sections, each one with a textual section name. If |
| 6737 | the target of the link is a format which does not have many sections |
| 6738 | (e.g., `a.out') or has sections without names (e.g., the Oasys format), |
| 6739 | the link cannot be done simply. You can circumvent this problem by |
| 6740 | describing the desired input-to-output section mapping with the linker |
| 6741 | command language. |
| 6742 | |
| 6743 | _Information can be lost during canonicalization._ The BFD internal |
| 6744 | canonical form of the external formats is not exhaustive; there are |
| 6745 | structures in input formats for which there is no direct representation |
| 6746 | internally. This means that the BFD back ends cannot maintain all |
| 6747 | possible data richness through the transformation between external to |
| 6748 | internal and back to external formats. |
| 6749 | |
| 6750 | This limitation is only a problem when an application reads one |
| 6751 | format and writes another. Each BFD back end is responsible for |
| 6752 | maintaining as much data as possible, and the internal BFD canonical |
| 6753 | form has structures which are opaque to the BFD core, and exported only |
| 6754 | to the back ends. When a file is read in one format, the canonical form |
| 6755 | is generated for BFD and the application. At the same time, the back |
| 6756 | end saves away any information which may otherwise be lost. If the data |
| 6757 | is then written back in the same format, the back end routine will be |
| 6758 | able to use the canonical form provided by the BFD core as well as the |
| 6759 | information it prepared earlier. Since there is a great deal of |
| 6760 | commonality between back ends, there is no information lost when |
| 6761 | linking or copying big endian COFF to little endian COFF, or `a.out' to |
| 6762 | `b.out'. When a mixture of formats is linked, the information is only |
| 6763 | lost from the files whose format differs from the destination. |
| 6764 | |
| 6765 | |
| 6766 | File: ld.info, Node: Canonical format, Prev: BFD information loss, Up: BFD outline |
| 6767 | |
| 6768 | 5.1.2 The BFD canonical object-file format |
| 6769 | ------------------------------------------ |
| 6770 | |
| 6771 | The greatest potential for loss of information occurs when there is the |
| 6772 | least overlap between the information provided by the source format, |
| 6773 | that stored by the canonical format, and that needed by the destination |
| 6774 | format. A brief description of the canonical form may help you |
| 6775 | understand which kinds of data you can count on preserving across |
| 6776 | conversions. |
| 6777 | |
| 6778 | _files_ |
| 6779 | Information stored on a per-file basis includes target machine |
| 6780 | architecture, particular implementation format type, a demand |
| 6781 | pageable bit, and a write protected bit. Information like Unix |
| 6782 | magic numbers is not stored here--only the magic numbers' meaning, |
| 6783 | so a `ZMAGIC' file would have both the demand pageable bit and the |
| 6784 | write protected text bit set. The byte order of the target is |
| 6785 | stored on a per-file basis, so that big- and little-endian object |
| 6786 | files may be used with one another. |
| 6787 | |
| 6788 | _sections_ |
| 6789 | Each section in the input file contains the name of the section, |
| 6790 | the section's original address in the object file, size and |
| 6791 | alignment information, various flags, and pointers into other BFD |
| 6792 | data structures. |
| 6793 | |
| 6794 | _symbols_ |
| 6795 | Each symbol contains a pointer to the information for the object |
| 6796 | file which originally defined it, its name, its value, and various |
| 6797 | flag bits. When a BFD back end reads in a symbol table, it |
| 6798 | relocates all symbols to make them relative to the base of the |
| 6799 | section where they were defined. Doing this ensures that each |
| 6800 | symbol points to its containing section. Each symbol also has a |
| 6801 | varying amount of hidden private data for the BFD back end. Since |
| 6802 | the symbol points to the original file, the private data format |
| 6803 | for that symbol is accessible. `ld' can operate on a collection |
| 6804 | of symbols of wildly different formats without problems. |
| 6805 | |
| 6806 | Normal global and simple local symbols are maintained on output, |
| 6807 | so an output file (no matter its format) will retain symbols |
| 6808 | pointing to functions and to global, static, and common variables. |
| 6809 | Some symbol information is not worth retaining; in `a.out', type |
| 6810 | information is stored in the symbol table as long symbol names. |
| 6811 | This information would be useless to most COFF debuggers; the |
| 6812 | linker has command line switches to allow users to throw it away. |
| 6813 | |
| 6814 | There is one word of type information within the symbol, so if the |
| 6815 | format supports symbol type information within symbols (for |
| 6816 | example, COFF, IEEE, Oasys) and the type is simple enough to fit |
| 6817 | within one word (nearly everything but aggregates), the |
| 6818 | information will be preserved. |
| 6819 | |
| 6820 | _relocation level_ |
| 6821 | Each canonical BFD relocation record contains a pointer to the |
| 6822 | symbol to relocate to, the offset of the data to relocate, the |
| 6823 | section the data is in, and a pointer to a relocation type |
| 6824 | descriptor. Relocation is performed by passing messages through |
| 6825 | the relocation type descriptor and the symbol pointer. Therefore, |
| 6826 | relocations can be performed on output data using a relocation |
| 6827 | method that is only available in one of the input formats. For |
| 6828 | instance, Oasys provides a byte relocation format. A relocation |
| 6829 | record requesting this relocation type would point indirectly to a |
| 6830 | routine to perform this, so the relocation may be performed on a |
| 6831 | byte being written to a 68k COFF file, even though 68k COFF has no |
| 6832 | such relocation type. |
| 6833 | |
| 6834 | _line numbers_ |
| 6835 | Object formats can contain, for debugging purposes, some form of |
| 6836 | mapping between symbols, source line numbers, and addresses in the |
| 6837 | output file. These addresses have to be relocated along with the |
| 6838 | symbol information. Each symbol with an associated list of line |
| 6839 | number records points to the first record of the list. The head |
| 6840 | of a line number list consists of a pointer to the symbol, which |
| 6841 | allows finding out the address of the function whose line number |
| 6842 | is being described. The rest of the list is made up of pairs: |
| 6843 | offsets into the section and line numbers. Any format which can |
| 6844 | simply derive this information can pass it successfully between |
| 6845 | formats (COFF, IEEE and Oasys). |
| 6846 | |
| 6847 | |
| 6848 | File: ld.info, Node: Reporting Bugs, Next: MRI, Prev: BFD, Up: Top |
| 6849 | |
| 6850 | 6 Reporting Bugs |
| 6851 | **************** |
| 6852 | |
| 6853 | Your bug reports play an essential role in making `ld' reliable. |
| 6854 | |
| 6855 | Reporting a bug may help you by bringing a solution to your problem, |
| 6856 | or it may not. But in any case the principal function of a bug report |
| 6857 | is to help the entire community by making the next version of `ld' work |
| 6858 | better. Bug reports are your contribution to the maintenance of `ld'. |
| 6859 | |
| 6860 | In order for a bug report to serve its purpose, you must include the |
| 6861 | information that enables us to fix the bug. |
| 6862 | |
| 6863 | * Menu: |
| 6864 | |
| 6865 | * Bug Criteria:: Have you found a bug? |
| 6866 | * Bug Reporting:: How to report bugs |
| 6867 | |
| 6868 | |
| 6869 | File: ld.info, Node: Bug Criteria, Next: Bug Reporting, Up: Reporting Bugs |
| 6870 | |
| 6871 | 6.1 Have You Found a Bug? |
| 6872 | ========================= |
| 6873 | |
| 6874 | If you are not sure whether you have found a bug, here are some |
| 6875 | guidelines: |
| 6876 | |
| 6877 | * If the linker gets a fatal signal, for any input whatever, that is |
| 6878 | a `ld' bug. Reliable linkers never crash. |
| 6879 | |
| 6880 | * If `ld' produces an error message for valid input, that is a bug. |
| 6881 | |
| 6882 | * If `ld' does not produce an error message for invalid input, that |
| 6883 | may be a bug. In the general case, the linker can not verify that |
| 6884 | object files are correct. |
| 6885 | |
| 6886 | * If you are an experienced user of linkers, your suggestions for |
| 6887 | improvement of `ld' are welcome in any case. |
| 6888 | |
| 6889 | |
| 6890 | File: ld.info, Node: Bug Reporting, Prev: Bug Criteria, Up: Reporting Bugs |
| 6891 | |
| 6892 | 6.2 How to Report Bugs |
| 6893 | ====================== |
| 6894 | |
| 6895 | A number of companies and individuals offer support for GNU products. |
| 6896 | If you obtained `ld' from a support organization, we recommend you |
| 6897 | contact that organization first. |
| 6898 | |
| 6899 | You can find contact information for many support companies and |
| 6900 | individuals in the file `etc/SERVICE' in the GNU Emacs distribution. |
| 6901 | |
| 6902 | Otherwise, send bug reports for `ld' to |
| 6903 | `http://www.sourceware.org/bugzilla/'. |
| 6904 | |
| 6905 | The fundamental principle of reporting bugs usefully is this: |
| 6906 | *report all the facts*. If you are not sure whether to state a fact or |
| 6907 | leave it out, state it! |
| 6908 | |
| 6909 | Often people omit facts because they think they know what causes the |
| 6910 | problem and assume that some details do not matter. Thus, you might |
| 6911 | assume that the name of a symbol you use in an example does not matter. |
| 6912 | Well, probably it does not, but one cannot be sure. Perhaps the bug |
| 6913 | is a stray memory reference which happens to fetch from the location |
| 6914 | where that name is stored in memory; perhaps, if the name were |
| 6915 | different, the contents of that location would fool the linker into |
| 6916 | doing the right thing despite the bug. Play it safe and give a |
| 6917 | specific, complete example. That is the easiest thing for you to do, |
| 6918 | and the most helpful. |
| 6919 | |
| 6920 | Keep in mind that the purpose of a bug report is to enable us to fix |
| 6921 | the bug if it is new to us. Therefore, always write your bug reports |
| 6922 | on the assumption that the bug has not been reported previously. |
| 6923 | |
| 6924 | Sometimes people give a few sketchy facts and ask, "Does this ring a |
| 6925 | bell?" This cannot help us fix a bug, so it is basically useless. We |
| 6926 | respond by asking for enough details to enable us to investigate. You |
| 6927 | might as well expedite matters by sending them to begin with. |
| 6928 | |
| 6929 | To enable us to fix the bug, you should include all these things: |
| 6930 | |
| 6931 | * The version of `ld'. `ld' announces it if you start it with the |
| 6932 | `--version' argument. |
| 6933 | |
| 6934 | Without this, we will not know whether there is any point in |
| 6935 | looking for the bug in the current version of `ld'. |
| 6936 | |
| 6937 | * Any patches you may have applied to the `ld' source, including any |
| 6938 | patches made to the `BFD' library. |
| 6939 | |
| 6940 | * The type of machine you are using, and the operating system name |
| 6941 | and version number. |
| 6942 | |
| 6943 | * What compiler (and its version) was used to compile `ld'--e.g. |
| 6944 | "`gcc-2.7'". |
| 6945 | |
| 6946 | * The command arguments you gave the linker to link your example and |
| 6947 | observe the bug. To guarantee you will not omit something |
| 6948 | important, list them all. A copy of the Makefile (or the output |
| 6949 | from make) is sufficient. |
| 6950 | |
| 6951 | If we were to try to guess the arguments, we would probably guess |
| 6952 | wrong and then we might not encounter the bug. |
| 6953 | |
| 6954 | * A complete input file, or set of input files, that will reproduce |
| 6955 | the bug. It is generally most helpful to send the actual object |
| 6956 | files provided that they are reasonably small. Say no more than |
| 6957 | 10K. For bigger files you can either make them available by FTP |
| 6958 | or HTTP or else state that you are willing to send the object |
| 6959 | file(s) to whomever requests them. (Note - your email will be |
| 6960 | going to a mailing list, so we do not want to clog it up with |
| 6961 | large attachments). But small attachments are best. |
| 6962 | |
| 6963 | If the source files were assembled using `gas' or compiled using |
| 6964 | `gcc', then it may be OK to send the source files rather than the |
| 6965 | object files. In this case, be sure to say exactly what version of |
| 6966 | `gas' or `gcc' was used to produce the object files. Also say how |
| 6967 | `gas' or `gcc' were configured. |
| 6968 | |
| 6969 | * A description of what behavior you observe that you believe is |
| 6970 | incorrect. For example, "It gets a fatal signal." |
| 6971 | |
| 6972 | Of course, if the bug is that `ld' gets a fatal signal, then we |
| 6973 | will certainly notice it. But if the bug is incorrect output, we |
| 6974 | might not notice unless it is glaringly wrong. You might as well |
| 6975 | not give us a chance to make a mistake. |
| 6976 | |
| 6977 | Even if the problem you experience is a fatal signal, you should |
| 6978 | still say so explicitly. Suppose something strange is going on, |
| 6979 | such as, your copy of `ld' is out of sync, or you have encountered |
| 6980 | a bug in the C library on your system. (This has happened!) Your |
| 6981 | copy might crash and ours would not. If you told us to expect a |
| 6982 | crash, then when ours fails to crash, we would know that the bug |
| 6983 | was not happening for us. If you had not told us to expect a |
| 6984 | crash, then we would not be able to draw any conclusion from our |
| 6985 | observations. |
| 6986 | |
| 6987 | * If you wish to suggest changes to the `ld' source, send us context |
| 6988 | diffs, as generated by `diff' with the `-u', `-c', or `-p' option. |
| 6989 | Always send diffs from the old file to the new file. If you even |
| 6990 | discuss something in the `ld' source, refer to it by context, not |
| 6991 | by line number. |
| 6992 | |
| 6993 | The line numbers in our development sources will not match those |
| 6994 | in your sources. Your line numbers would convey no useful |
| 6995 | information to us. |
| 6996 | |
| 6997 | Here are some things that are not necessary: |
| 6998 | |
| 6999 | * A description of the envelope of the bug. |
| 7000 | |
| 7001 | Often people who encounter a bug spend a lot of time investigating |
| 7002 | which changes to the input file will make the bug go away and which |
| 7003 | changes will not affect it. |
| 7004 | |
| 7005 | This is often time consuming and not very useful, because the way |
| 7006 | we will find the bug is by running a single example under the |
| 7007 | debugger with breakpoints, not by pure deduction from a series of |
| 7008 | examples. We recommend that you save your time for something else. |
| 7009 | |
| 7010 | Of course, if you can find a simpler example to report _instead_ |
| 7011 | of the original one, that is a convenience for us. Errors in the |
| 7012 | output will be easier to spot, running under the debugger will take |
| 7013 | less time, and so on. |
| 7014 | |
| 7015 | However, simplification is not vital; if you do not want to do |
| 7016 | this, report the bug anyway and send us the entire test case you |
| 7017 | used. |
| 7018 | |
| 7019 | * A patch for the bug. |
| 7020 | |
| 7021 | A patch for the bug does help us if it is a good one. But do not |
| 7022 | omit the necessary information, such as the test case, on the |
| 7023 | assumption that a patch is all we need. We might see problems |
| 7024 | with your patch and decide to fix the problem another way, or we |
| 7025 | might not understand it at all. |
| 7026 | |
| 7027 | Sometimes with a program as complicated as `ld' it is very hard to |
| 7028 | construct an example that will make the program follow a certain |
| 7029 | path through the code. If you do not send us the example, we will |
| 7030 | not be able to construct one, so we will not be able to verify |
| 7031 | that the bug is fixed. |
| 7032 | |
| 7033 | And if we cannot understand what bug you are trying to fix, or why |
| 7034 | your patch should be an improvement, we will not install it. A |
| 7035 | test case will help us to understand. |
| 7036 | |
| 7037 | * A guess about what the bug is or what it depends on. |
| 7038 | |
| 7039 | Such guesses are usually wrong. Even we cannot guess right about |
| 7040 | such things without first using the debugger to find the facts. |
| 7041 | |
| 7042 | |
| 7043 | File: ld.info, Node: MRI, Next: GNU Free Documentation License, Prev: Reporting Bugs, Up: Top |
| 7044 | |
| 7045 | Appendix A MRI Compatible Script Files |
| 7046 | ************************************** |
| 7047 | |
| 7048 | To aid users making the transition to GNU `ld' from the MRI linker, |
| 7049 | `ld' can use MRI compatible linker scripts as an alternative to the |
| 7050 | more general-purpose linker scripting language described in *Note |
| 7051 | Scripts::. MRI compatible linker scripts have a much simpler command |
| 7052 | set than the scripting language otherwise used with `ld'. GNU `ld' |
| 7053 | supports the most commonly used MRI linker commands; these commands are |
| 7054 | described here. |
| 7055 | |
| 7056 | In general, MRI scripts aren't of much use with the `a.out' object |
| 7057 | file format, since it only has three sections and MRI scripts lack some |
| 7058 | features to make use of them. |
| 7059 | |
| 7060 | You can specify a file containing an MRI-compatible script using the |
| 7061 | `-c' command-line option. |
| 7062 | |
| 7063 | Each command in an MRI-compatible script occupies its own line; each |
| 7064 | command line starts with the keyword that identifies the command (though |
| 7065 | blank lines are also allowed for punctuation). If a line of an |
| 7066 | MRI-compatible script begins with an unrecognized keyword, `ld' issues |
| 7067 | a warning message, but continues processing the script. |
| 7068 | |
| 7069 | Lines beginning with `*' are comments. |
| 7070 | |
| 7071 | You can write these commands using all upper-case letters, or all |
| 7072 | lower case; for example, `chip' is the same as `CHIP'. The following |
| 7073 | list shows only the upper-case form of each command. |
| 7074 | |
| 7075 | `ABSOLUTE SECNAME' |
| 7076 | `ABSOLUTE SECNAME, SECNAME, ... SECNAME' |
| 7077 | Normally, `ld' includes in the output file all sections from all |
| 7078 | the input files. However, in an MRI-compatible script, you can |
| 7079 | use the `ABSOLUTE' command to restrict the sections that will be |
| 7080 | present in your output program. If the `ABSOLUTE' command is used |
| 7081 | at all in a script, then only the sections named explicitly in |
| 7082 | `ABSOLUTE' commands will appear in the linker output. You can |
| 7083 | still use other input sections (whatever you select on the command |
| 7084 | line, or using `LOAD') to resolve addresses in the output file. |
| 7085 | |
| 7086 | `ALIAS OUT-SECNAME, IN-SECNAME' |
| 7087 | Use this command to place the data from input section IN-SECNAME |
| 7088 | in a section called OUT-SECNAME in the linker output file. |
| 7089 | |
| 7090 | IN-SECNAME may be an integer. |
| 7091 | |
| 7092 | `ALIGN SECNAME = EXPRESSION' |
| 7093 | Align the section called SECNAME to EXPRESSION. The EXPRESSION |
| 7094 | should be a power of two. |
| 7095 | |
| 7096 | `BASE EXPRESSION' |
| 7097 | Use the value of EXPRESSION as the lowest address (other than |
| 7098 | absolute addresses) in the output file. |
| 7099 | |
| 7100 | `CHIP EXPRESSION' |
| 7101 | `CHIP EXPRESSION, EXPRESSION' |
| 7102 | This command does nothing; it is accepted only for compatibility. |
| 7103 | |
| 7104 | `END' |
| 7105 | This command does nothing whatever; it's only accepted for |
| 7106 | compatibility. |
| 7107 | |
| 7108 | `FORMAT OUTPUT-FORMAT' |
| 7109 | Similar to the `OUTPUT_FORMAT' command in the more general linker |
| 7110 | language, but restricted to one of these output formats: |
| 7111 | |
| 7112 | 1. S-records, if OUTPUT-FORMAT is `S' |
| 7113 | |
| 7114 | 2. IEEE, if OUTPUT-FORMAT is `IEEE' |
| 7115 | |
| 7116 | 3. COFF (the `coff-m68k' variant in BFD), if OUTPUT-FORMAT is |
| 7117 | `COFF' |
| 7118 | |
| 7119 | `LIST ANYTHING...' |
| 7120 | Print (to the standard output file) a link map, as produced by the |
| 7121 | `ld' command-line option `-M'. |
| 7122 | |
| 7123 | The keyword `LIST' may be followed by anything on the same line, |
| 7124 | with no change in its effect. |
| 7125 | |
| 7126 | `LOAD FILENAME' |
| 7127 | `LOAD FILENAME, FILENAME, ... FILENAME' |
| 7128 | Include one or more object file FILENAME in the link; this has the |
| 7129 | same effect as specifying FILENAME directly on the `ld' command |
| 7130 | line. |
| 7131 | |
| 7132 | `NAME OUTPUT-NAME' |
| 7133 | OUTPUT-NAME is the name for the program produced by `ld'; the |
| 7134 | MRI-compatible command `NAME' is equivalent to the command-line |
| 7135 | option `-o' or the general script language command `OUTPUT'. |
| 7136 | |
| 7137 | `ORDER SECNAME, SECNAME, ... SECNAME' |
| 7138 | `ORDER SECNAME SECNAME SECNAME' |
| 7139 | Normally, `ld' orders the sections in its output file in the order |
| 7140 | in which they first appear in the input files. In an |
| 7141 | MRI-compatible script, you can override this ordering with the |
| 7142 | `ORDER' command. The sections you list with `ORDER' will appear |
| 7143 | first in your output file, in the order specified. |
| 7144 | |
| 7145 | `PUBLIC NAME=EXPRESSION' |
| 7146 | `PUBLIC NAME,EXPRESSION' |
| 7147 | `PUBLIC NAME EXPRESSION' |
| 7148 | Supply a value (EXPRESSION) for external symbol NAME used in the |
| 7149 | linker input files. |
| 7150 | |
| 7151 | `SECT SECNAME, EXPRESSION' |
| 7152 | `SECT SECNAME=EXPRESSION' |
| 7153 | `SECT SECNAME EXPRESSION' |
| 7154 | You can use any of these three forms of the `SECT' command to |
| 7155 | specify the start address (EXPRESSION) for section SECNAME. If |
| 7156 | you have more than one `SECT' statement for the same SECNAME, only |
| 7157 | the _first_ sets the start address. |
| 7158 | |
| 7159 | |
| 7160 | File: ld.info, Node: GNU Free Documentation License, Next: LD Index, Prev: MRI, Up: Top |
| 7161 | |
| 7162 | Appendix B GNU Free Documentation License |
| 7163 | ***************************************** |
| 7164 | |
| 7165 | Version 1.3, 3 November 2008 |
| 7166 | |
| 7167 | Copyright (C) 2000, 2001, 2002, 2007, 2008 Free Software Foundation, Inc. |
| 7168 | `http://fsf.org/' |
| 7169 | |
| 7170 | Everyone is permitted to copy and distribute verbatim copies |
| 7171 | of this license document, but changing it is not allowed. |
| 7172 | |
| 7173 | 0. PREAMBLE |
| 7174 | |
| 7175 | The purpose of this License is to make a manual, textbook, or other |
| 7176 | functional and useful document "free" in the sense of freedom: to |
| 7177 | assure everyone the effective freedom to copy and redistribute it, |
| 7178 | with or without modifying it, either commercially or |
| 7179 | noncommercially. Secondarily, this License preserves for the |
| 7180 | author and publisher a way to get credit for their work, while not |
| 7181 | being considered responsible for modifications made by others. |
| 7182 | |
| 7183 | This License is a kind of "copyleft", which means that derivative |
| 7184 | works of the document must themselves be free in the same sense. |
| 7185 | It complements the GNU General Public License, which is a copyleft |
| 7186 | license designed for free software. |
| 7187 | |
| 7188 | We have designed this License in order to use it for manuals for |
| 7189 | free software, because free software needs free documentation: a |
| 7190 | free program should come with manuals providing the same freedoms |
| 7191 | that the software does. But this License is not limited to |
| 7192 | software manuals; it can be used for any textual work, regardless |
| 7193 | of subject matter or whether it is published as a printed book. |
| 7194 | We recommend this License principally for works whose purpose is |
| 7195 | instruction or reference. |
| 7196 | |
| 7197 | 1. APPLICABILITY AND DEFINITIONS |
| 7198 | |
| 7199 | This License applies to any manual or other work, in any medium, |
| 7200 | that contains a notice placed by the copyright holder saying it |
| 7201 | can be distributed under the terms of this License. Such a notice |
| 7202 | grants a world-wide, royalty-free license, unlimited in duration, |
| 7203 | to use that work under the conditions stated herein. The |
| 7204 | "Document", below, refers to any such manual or work. Any member |
| 7205 | of the public is a licensee, and is addressed as "you". You |
| 7206 | accept the license if you copy, modify or distribute the work in a |
| 7207 | way requiring permission under copyright law. |
| 7208 | |
| 7209 | A "Modified Version" of the Document means any work containing the |
| 7210 | Document or a portion of it, either copied verbatim, or with |
| 7211 | modifications and/or translated into another language. |
| 7212 | |
| 7213 | A "Secondary Section" is a named appendix or a front-matter section |
| 7214 | of the Document that deals exclusively with the relationship of the |
| 7215 | publishers or authors of the Document to the Document's overall |
| 7216 | subject (or to related matters) and contains nothing that could |
| 7217 | fall directly within that overall subject. (Thus, if the Document |
| 7218 | is in part a textbook of mathematics, a Secondary Section may not |
| 7219 | explain any mathematics.) The relationship could be a matter of |
| 7220 | historical connection with the subject or with related matters, or |
| 7221 | of legal, commercial, philosophical, ethical or political position |
| 7222 | regarding them. |
| 7223 | |
| 7224 | The "Invariant Sections" are certain Secondary Sections whose |
| 7225 | titles are designated, as being those of Invariant Sections, in |
| 7226 | the notice that says that the Document is released under this |
| 7227 | License. If a section does not fit the above definition of |
| 7228 | Secondary then it is not allowed to be designated as Invariant. |
| 7229 | The Document may contain zero Invariant Sections. If the Document |
| 7230 | does not identify any Invariant Sections then there are none. |
| 7231 | |
| 7232 | The "Cover Texts" are certain short passages of text that are |
| 7233 | listed, as Front-Cover Texts or Back-Cover Texts, in the notice |
| 7234 | that says that the Document is released under this License. A |
| 7235 | Front-Cover Text may be at most 5 words, and a Back-Cover Text may |
| 7236 | be at most 25 words. |
| 7237 | |
| 7238 | A "Transparent" copy of the Document means a machine-readable copy, |
| 7239 | represented in a format whose specification is available to the |
| 7240 | general public, that is suitable for revising the document |
| 7241 | straightforwardly with generic text editors or (for images |
| 7242 | composed of pixels) generic paint programs or (for drawings) some |
| 7243 | widely available drawing editor, and that is suitable for input to |
| 7244 | text formatters or for automatic translation to a variety of |
| 7245 | formats suitable for input to text formatters. A copy made in an |
| 7246 | otherwise Transparent file format whose markup, or absence of |
| 7247 | markup, has been arranged to thwart or discourage subsequent |
| 7248 | modification by readers is not Transparent. An image format is |
| 7249 | not Transparent if used for any substantial amount of text. A |
| 7250 | copy that is not "Transparent" is called "Opaque". |
| 7251 | |
| 7252 | Examples of suitable formats for Transparent copies include plain |
| 7253 | ASCII without markup, Texinfo input format, LaTeX input format, |
| 7254 | SGML or XML using a publicly available DTD, and |
| 7255 | standard-conforming simple HTML, PostScript or PDF designed for |
| 7256 | human modification. Examples of transparent image formats include |
| 7257 | PNG, XCF and JPG. Opaque formats include proprietary formats that |
| 7258 | can be read and edited only by proprietary word processors, SGML or |
| 7259 | XML for which the DTD and/or processing tools are not generally |
| 7260 | available, and the machine-generated HTML, PostScript or PDF |
| 7261 | produced by some word processors for output purposes only. |
| 7262 | |
| 7263 | The "Title Page" means, for a printed book, the title page itself, |
| 7264 | plus such following pages as are needed to hold, legibly, the |
| 7265 | material this License requires to appear in the title page. For |
| 7266 | works in formats which do not have any title page as such, "Title |
| 7267 | Page" means the text near the most prominent appearance of the |
| 7268 | work's title, preceding the beginning of the body of the text. |
| 7269 | |
| 7270 | The "publisher" means any person or entity that distributes copies |
| 7271 | of the Document to the public. |
| 7272 | |
| 7273 | A section "Entitled XYZ" means a named subunit of the Document |
| 7274 | whose title either is precisely XYZ or contains XYZ in parentheses |
| 7275 | following text that translates XYZ in another language. (Here XYZ |
| 7276 | stands for a specific section name mentioned below, such as |
| 7277 | "Acknowledgements", "Dedications", "Endorsements", or "History".) |
| 7278 | To "Preserve the Title" of such a section when you modify the |
| 7279 | Document means that it remains a section "Entitled XYZ" according |
| 7280 | to this definition. |
| 7281 | |
| 7282 | The Document may include Warranty Disclaimers next to the notice |
| 7283 | which states that this License applies to the Document. These |
| 7284 | Warranty Disclaimers are considered to be included by reference in |
| 7285 | this License, but only as regards disclaiming warranties: any other |
| 7286 | implication that these Warranty Disclaimers may have is void and |
| 7287 | has no effect on the meaning of this License. |
| 7288 | |
| 7289 | 2. VERBATIM COPYING |
| 7290 | |
| 7291 | You may copy and distribute the Document in any medium, either |
| 7292 | commercially or noncommercially, provided that this License, the |
| 7293 | copyright notices, and the license notice saying this License |
| 7294 | applies to the Document are reproduced in all copies, and that you |
| 7295 | add no other conditions whatsoever to those of this License. You |
| 7296 | may not use technical measures to obstruct or control the reading |
| 7297 | or further copying of the copies you make or distribute. However, |
| 7298 | you may accept compensation in exchange for copies. If you |
| 7299 | distribute a large enough number of copies you must also follow |
| 7300 | the conditions in section 3. |
| 7301 | |
| 7302 | You may also lend copies, under the same conditions stated above, |
| 7303 | and you may publicly display copies. |
| 7304 | |
| 7305 | 3. COPYING IN QUANTITY |
| 7306 | |
| 7307 | If you publish printed copies (or copies in media that commonly |
| 7308 | have printed covers) of the Document, numbering more than 100, and |
| 7309 | the Document's license notice requires Cover Texts, you must |
| 7310 | enclose the copies in covers that carry, clearly and legibly, all |
| 7311 | these Cover Texts: Front-Cover Texts on the front cover, and |
| 7312 | Back-Cover Texts on the back cover. Both covers must also clearly |
| 7313 | and legibly identify you as the publisher of these copies. The |
| 7314 | front cover must present the full title with all words of the |
| 7315 | title equally prominent and visible. You may add other material |
| 7316 | on the covers in addition. Copying with changes limited to the |
| 7317 | covers, as long as they preserve the title of the Document and |
| 7318 | satisfy these conditions, can be treated as verbatim copying in |
| 7319 | other respects. |
| 7320 | |
| 7321 | If the required texts for either cover are too voluminous to fit |
| 7322 | legibly, you should put the first ones listed (as many as fit |
| 7323 | reasonably) on the actual cover, and continue the rest onto |
| 7324 | adjacent pages. |
| 7325 | |
| 7326 | If you publish or distribute Opaque copies of the Document |
| 7327 | numbering more than 100, you must either include a |
| 7328 | machine-readable Transparent copy along with each Opaque copy, or |
| 7329 | state in or with each Opaque copy a computer-network location from |
| 7330 | which the general network-using public has access to download |
| 7331 | using public-standard network protocols a complete Transparent |
| 7332 | copy of the Document, free of added material. If you use the |
| 7333 | latter option, you must take reasonably prudent steps, when you |
| 7334 | begin distribution of Opaque copies in quantity, to ensure that |
| 7335 | this Transparent copy will remain thus accessible at the stated |
| 7336 | location until at least one year after the last time you |
| 7337 | distribute an Opaque copy (directly or through your agents or |
| 7338 | retailers) of that edition to the public. |
| 7339 | |
| 7340 | It is requested, but not required, that you contact the authors of |
| 7341 | the Document well before redistributing any large number of |
| 7342 | copies, to give them a chance to provide you with an updated |
| 7343 | version of the Document. |
| 7344 | |
| 7345 | 4. MODIFICATIONS |
| 7346 | |
| 7347 | You may copy and distribute a Modified Version of the Document |
| 7348 | under the conditions of sections 2 and 3 above, provided that you |
| 7349 | release the Modified Version under precisely this License, with |
| 7350 | the Modified Version filling the role of the Document, thus |
| 7351 | licensing distribution and modification of the Modified Version to |
| 7352 | whoever possesses a copy of it. In addition, you must do these |
| 7353 | things in the Modified Version: |
| 7354 | |
| 7355 | A. Use in the Title Page (and on the covers, if any) a title |
| 7356 | distinct from that of the Document, and from those of |
| 7357 | previous versions (which should, if there were any, be listed |
| 7358 | in the History section of the Document). You may use the |
| 7359 | same title as a previous version if the original publisher of |
| 7360 | that version gives permission. |
| 7361 | |
| 7362 | B. List on the Title Page, as authors, one or more persons or |
| 7363 | entities responsible for authorship of the modifications in |
| 7364 | the Modified Version, together with at least five of the |
| 7365 | principal authors of the Document (all of its principal |
| 7366 | authors, if it has fewer than five), unless they release you |
| 7367 | from this requirement. |
| 7368 | |
| 7369 | C. State on the Title page the name of the publisher of the |
| 7370 | Modified Version, as the publisher. |
| 7371 | |
| 7372 | D. Preserve all the copyright notices of the Document. |
| 7373 | |
| 7374 | E. Add an appropriate copyright notice for your modifications |
| 7375 | adjacent to the other copyright notices. |
| 7376 | |
| 7377 | F. Include, immediately after the copyright notices, a license |
| 7378 | notice giving the public permission to use the Modified |
| 7379 | Version under the terms of this License, in the form shown in |
| 7380 | the Addendum below. |
| 7381 | |
| 7382 | G. Preserve in that license notice the full lists of Invariant |
| 7383 | Sections and required Cover Texts given in the Document's |
| 7384 | license notice. |
| 7385 | |
| 7386 | H. Include an unaltered copy of this License. |
| 7387 | |
| 7388 | I. Preserve the section Entitled "History", Preserve its Title, |
| 7389 | and add to it an item stating at least the title, year, new |
| 7390 | authors, and publisher of the Modified Version as given on |
| 7391 | the Title Page. If there is no section Entitled "History" in |
| 7392 | the Document, create one stating the title, year, authors, |
| 7393 | and publisher of the Document as given on its Title Page, |
| 7394 | then add an item describing the Modified Version as stated in |
| 7395 | the previous sentence. |
| 7396 | |
| 7397 | J. Preserve the network location, if any, given in the Document |
| 7398 | for public access to a Transparent copy of the Document, and |
| 7399 | likewise the network locations given in the Document for |
| 7400 | previous versions it was based on. These may be placed in |
| 7401 | the "History" section. You may omit a network location for a |
| 7402 | work that was published at least four years before the |
| 7403 | Document itself, or if the original publisher of the version |
| 7404 | it refers to gives permission. |
| 7405 | |
| 7406 | K. For any section Entitled "Acknowledgements" or "Dedications", |
| 7407 | Preserve the Title of the section, and preserve in the |
| 7408 | section all the substance and tone of each of the contributor |
| 7409 | acknowledgements and/or dedications given therein. |
| 7410 | |
| 7411 | L. Preserve all the Invariant Sections of the Document, |
| 7412 | unaltered in their text and in their titles. Section numbers |
| 7413 | or the equivalent are not considered part of the section |
| 7414 | titles. |
| 7415 | |
| 7416 | M. Delete any section Entitled "Endorsements". Such a section |
| 7417 | may not be included in the Modified Version. |
| 7418 | |
| 7419 | N. Do not retitle any existing section to be Entitled |
| 7420 | "Endorsements" or to conflict in title with any Invariant |
| 7421 | Section. |
| 7422 | |
| 7423 | O. Preserve any Warranty Disclaimers. |
| 7424 | |
| 7425 | If the Modified Version includes new front-matter sections or |
| 7426 | appendices that qualify as Secondary Sections and contain no |
| 7427 | material copied from the Document, you may at your option |
| 7428 | designate some or all of these sections as invariant. To do this, |
| 7429 | add their titles to the list of Invariant Sections in the Modified |
| 7430 | Version's license notice. These titles must be distinct from any |
| 7431 | other section titles. |
| 7432 | |
| 7433 | You may add a section Entitled "Endorsements", provided it contains |
| 7434 | nothing but endorsements of your Modified Version by various |
| 7435 | parties--for example, statements of peer review or that the text |
| 7436 | has been approved by an organization as the authoritative |
| 7437 | definition of a standard. |
| 7438 | |
| 7439 | You may add a passage of up to five words as a Front-Cover Text, |
| 7440 | and a passage of up to 25 words as a Back-Cover Text, to the end |
| 7441 | of the list of Cover Texts in the Modified Version. Only one |
| 7442 | passage of Front-Cover Text and one of Back-Cover Text may be |
| 7443 | added by (or through arrangements made by) any one entity. If the |
| 7444 | Document already includes a cover text for the same cover, |
| 7445 | previously added by you or by arrangement made by the same entity |
| 7446 | you are acting on behalf of, you may not add another; but you may |
| 7447 | replace the old one, on explicit permission from the previous |
| 7448 | publisher that added the old one. |
| 7449 | |
| 7450 | The author(s) and publisher(s) of the Document do not by this |
| 7451 | License give permission to use their names for publicity for or to |
| 7452 | assert or imply endorsement of any Modified Version. |
| 7453 | |
| 7454 | 5. COMBINING DOCUMENTS |
| 7455 | |
| 7456 | You may combine the Document with other documents released under |
| 7457 | this License, under the terms defined in section 4 above for |
| 7458 | modified versions, provided that you include in the combination |
| 7459 | all of the Invariant Sections of all of the original documents, |
| 7460 | unmodified, and list them all as Invariant Sections of your |
| 7461 | combined work in its license notice, and that you preserve all |
| 7462 | their Warranty Disclaimers. |
| 7463 | |
| 7464 | The combined work need only contain one copy of this License, and |
| 7465 | multiple identical Invariant Sections may be replaced with a single |
| 7466 | copy. If there are multiple Invariant Sections with the same name |
| 7467 | but different contents, make the title of each such section unique |
| 7468 | by adding at the end of it, in parentheses, the name of the |
| 7469 | original author or publisher of that section if known, or else a |
| 7470 | unique number. Make the same adjustment to the section titles in |
| 7471 | the list of Invariant Sections in the license notice of the |
| 7472 | combined work. |
| 7473 | |
| 7474 | In the combination, you must combine any sections Entitled |
| 7475 | "History" in the various original documents, forming one section |
| 7476 | Entitled "History"; likewise combine any sections Entitled |
| 7477 | "Acknowledgements", and any sections Entitled "Dedications". You |
| 7478 | must delete all sections Entitled "Endorsements." |
| 7479 | |
| 7480 | 6. COLLECTIONS OF DOCUMENTS |
| 7481 | |
| 7482 | You may make a collection consisting of the Document and other |
| 7483 | documents released under this License, and replace the individual |
| 7484 | copies of this License in the various documents with a single copy |
| 7485 | that is included in the collection, provided that you follow the |
| 7486 | rules of this License for verbatim copying of each of the |
| 7487 | documents in all other respects. |
| 7488 | |
| 7489 | You may extract a single document from such a collection, and |
| 7490 | distribute it individually under this License, provided you insert |
| 7491 | a copy of this License into the extracted document, and follow |
| 7492 | this License in all other respects regarding verbatim copying of |
| 7493 | that document. |
| 7494 | |
| 7495 | 7. AGGREGATION WITH INDEPENDENT WORKS |
| 7496 | |
| 7497 | A compilation of the Document or its derivatives with other |
| 7498 | separate and independent documents or works, in or on a volume of |
| 7499 | a storage or distribution medium, is called an "aggregate" if the |
| 7500 | copyright resulting from the compilation is not used to limit the |
| 7501 | legal rights of the compilation's users beyond what the individual |
| 7502 | works permit. When the Document is included in an aggregate, this |
| 7503 | License does not apply to the other works in the aggregate which |
| 7504 | are not themselves derivative works of the Document. |
| 7505 | |
| 7506 | If the Cover Text requirement of section 3 is applicable to these |
| 7507 | copies of the Document, then if the Document is less than one half |
| 7508 | of the entire aggregate, the Document's Cover Texts may be placed |
| 7509 | on covers that bracket the Document within the aggregate, or the |
| 7510 | electronic equivalent of covers if the Document is in electronic |
| 7511 | form. Otherwise they must appear on printed covers that bracket |
| 7512 | the whole aggregate. |
| 7513 | |
| 7514 | 8. TRANSLATION |
| 7515 | |
| 7516 | Translation is considered a kind of modification, so you may |
| 7517 | distribute translations of the Document under the terms of section |
| 7518 | 4. Replacing Invariant Sections with translations requires special |
| 7519 | permission from their copyright holders, but you may include |
| 7520 | translations of some or all Invariant Sections in addition to the |
| 7521 | original versions of these Invariant Sections. You may include a |
| 7522 | translation of this License, and all the license notices in the |
| 7523 | Document, and any Warranty Disclaimers, provided that you also |
| 7524 | include the original English version of this License and the |
| 7525 | original versions of those notices and disclaimers. In case of a |
| 7526 | disagreement between the translation and the original version of |
| 7527 | this License or a notice or disclaimer, the original version will |
| 7528 | prevail. |
| 7529 | |
| 7530 | If a section in the Document is Entitled "Acknowledgements", |
| 7531 | "Dedications", or "History", the requirement (section 4) to |
| 7532 | Preserve its Title (section 1) will typically require changing the |
| 7533 | actual title. |
| 7534 | |
| 7535 | 9. TERMINATION |
| 7536 | |
| 7537 | You may not copy, modify, sublicense, or distribute the Document |
| 7538 | except as expressly provided under this License. Any attempt |
| 7539 | otherwise to copy, modify, sublicense, or distribute it is void, |
| 7540 | and will automatically terminate your rights under this License. |
| 7541 | |
| 7542 | However, if you cease all violation of this License, then your |
| 7543 | license from a particular copyright holder is reinstated (a) |
| 7544 | provisionally, unless and until the copyright holder explicitly |
| 7545 | and finally terminates your license, and (b) permanently, if the |
| 7546 | copyright holder fails to notify you of the violation by some |
| 7547 | reasonable means prior to 60 days after the cessation. |
| 7548 | |
| 7549 | Moreover, your license from a particular copyright holder is |
| 7550 | reinstated permanently if the copyright holder notifies you of the |
| 7551 | violation by some reasonable means, this is the first time you have |
| 7552 | received notice of violation of this License (for any work) from |
| 7553 | that copyright holder, and you cure the violation prior to 30 days |
| 7554 | after your receipt of the notice. |
| 7555 | |
| 7556 | Termination of your rights under this section does not terminate |
| 7557 | the licenses of parties who have received copies or rights from |
| 7558 | you under this License. If your rights have been terminated and |
| 7559 | not permanently reinstated, receipt of a copy of some or all of |
| 7560 | the same material does not give you any rights to use it. |
| 7561 | |
| 7562 | 10. FUTURE REVISIONS OF THIS LICENSE |
| 7563 | |
| 7564 | The Free Software Foundation may publish new, revised versions of |
| 7565 | the GNU Free Documentation License from time to time. Such new |
| 7566 | versions will be similar in spirit to the present version, but may |
| 7567 | differ in detail to address new problems or concerns. See |
| 7568 | `http://www.gnu.org/copyleft/'. |
| 7569 | |
| 7570 | Each version of the License is given a distinguishing version |
| 7571 | number. If the Document specifies that a particular numbered |
| 7572 | version of this License "or any later version" applies to it, you |
| 7573 | have the option of following the terms and conditions either of |
| 7574 | that specified version or of any later version that has been |
| 7575 | published (not as a draft) by the Free Software Foundation. If |
| 7576 | the Document does not specify a version number of this License, |
| 7577 | you may choose any version ever published (not as a draft) by the |
| 7578 | Free Software Foundation. If the Document specifies that a proxy |
| 7579 | can decide which future versions of this License can be used, that |
| 7580 | proxy's public statement of acceptance of a version permanently |
| 7581 | authorizes you to choose that version for the Document. |
| 7582 | |
| 7583 | 11. RELICENSING |
| 7584 | |
| 7585 | "Massive Multiauthor Collaboration Site" (or "MMC Site") means any |
| 7586 | World Wide Web server that publishes copyrightable works and also |
| 7587 | provides prominent facilities for anybody to edit those works. A |
| 7588 | public wiki that anybody can edit is an example of such a server. |
| 7589 | A "Massive Multiauthor Collaboration" (or "MMC") contained in the |
| 7590 | site means any set of copyrightable works thus published on the MMC |
| 7591 | site. |
| 7592 | |
| 7593 | "CC-BY-SA" means the Creative Commons Attribution-Share Alike 3.0 |
| 7594 | license published by Creative Commons Corporation, a not-for-profit |
| 7595 | corporation with a principal place of business in San Francisco, |
| 7596 | California, as well as future copyleft versions of that license |
| 7597 | published by that same organization. |
| 7598 | |
| 7599 | "Incorporate" means to publish or republish a Document, in whole or |
| 7600 | in part, as part of another Document. |
| 7601 | |
| 7602 | An MMC is "eligible for relicensing" if it is licensed under this |
| 7603 | License, and if all works that were first published under this |
| 7604 | License somewhere other than this MMC, and subsequently |
| 7605 | incorporated in whole or in part into the MMC, (1) had no cover |
| 7606 | texts or invariant sections, and (2) were thus incorporated prior |
| 7607 | to November 1, 2008. |
| 7608 | |
| 7609 | The operator of an MMC Site may republish an MMC contained in the |
| 7610 | site under CC-BY-SA on the same site at any time before August 1, |
| 7611 | 2009, provided the MMC is eligible for relicensing. |
| 7612 | |
| 7613 | |
| 7614 | ADDENDUM: How to use this License for your documents |
| 7615 | ==================================================== |
| 7616 | |
| 7617 | To use this License in a document you have written, include a copy of |
| 7618 | the License in the document and put the following copyright and license |
| 7619 | notices just after the title page: |
| 7620 | |
| 7621 | Copyright (C) YEAR YOUR NAME. |
| 7622 | Permission is granted to copy, distribute and/or modify this document |
| 7623 | under the terms of the GNU Free Documentation License, Version 1.3 |
| 7624 | or any later version published by the Free Software Foundation; |
| 7625 | with no Invariant Sections, no Front-Cover Texts, and no Back-Cover |
| 7626 | Texts. A copy of the license is included in the section entitled ``GNU |
| 7627 | Free Documentation License''. |
| 7628 | |
| 7629 | If you have Invariant Sections, Front-Cover Texts and Back-Cover |
| 7630 | Texts, replace the "with...Texts." line with this: |
| 7631 | |
| 7632 | with the Invariant Sections being LIST THEIR TITLES, with |
| 7633 | the Front-Cover Texts being LIST, and with the Back-Cover Texts |
| 7634 | being LIST. |
| 7635 | |
| 7636 | If you have Invariant Sections without Cover Texts, or some other |
| 7637 | combination of the three, merge those two alternatives to suit the |
| 7638 | situation. |
| 7639 | |
| 7640 | If your document contains nontrivial examples of program code, we |
| 7641 | recommend releasing these examples in parallel under your choice of |
| 7642 | free software license, such as the GNU General Public License, to |
| 7643 | permit their use in free software. |
| 7644 | |
| 7645 | |
| 7646 | File: ld.info, Node: LD Index, Prev: GNU Free Documentation License, Up: Top |
| 7647 | |
| 7648 | LD Index |
| 7649 | ******** |
| 7650 | |
| 7651 | [index] |
| 7652 | * Menu: |
| 7653 | |
| 7654 | * ": Symbols. (line 6) |
| 7655 | * -(: Options. (line 823) |
| 7656 | * --accept-unknown-input-arch: Options. (line 841) |
| 7657 | * --add-needed: Options. (line 869) |
| 7658 | * --add-stdcall-alias: Options. (line 1785) |
| 7659 | * --allow-multiple-definition: Options. (line 1147) |
| 7660 | * --allow-shlib-undefined: Options. (line 1153) |
| 7661 | * --architecture=ARCH: Options. (line 123) |
| 7662 | * --as-needed: Options. (line 851) |
| 7663 | * --audit AUDITLIB: Options. (line 112) |
| 7664 | * --auxiliary=NAME: Options. (line 255) |
| 7665 | * --bank-window: Options. (line 2239) |
| 7666 | * --base-file: Options. (line 1790) |
| 7667 | * --be8: ARM. (line 28) |
| 7668 | * --bss-plt: PowerPC ELF32. (line 16) |
| 7669 | * --build-id: Options. (line 1731) |
| 7670 | * --build-id=STYLE: Options. (line 1731) |
| 7671 | * --check-sections: Options. (line 955) |
| 7672 | * --compress-debug-sections=none: Options. (line 1701) |
| 7673 | * --compress-debug-sections=zlib: Options. (line 1701) |
| 7674 | * --compress-debug-sections=zlib-gabi: Options. (line 1701) |
| 7675 | * --compress-debug-sections=zlib-gnu: Options. (line 1701) |
| 7676 | * --copy-dt-needed-entries: Options. (line 967) |
| 7677 | * --cref: Options. (line 987) |
| 7678 | * --default-imported-symver: Options. (line 1190) |
| 7679 | * --default-script=SCRIPT: Options. (line 562) |
| 7680 | * --default-symver: Options. (line 1186) |
| 7681 | * --defsym=SYMBOL=EXP: Options. (line 1016) |
| 7682 | * --demangle[=STYLE]: Options. (line 1028) |
| 7683 | * --depaudit AUDITLIB: Options. (line 177) |
| 7684 | * --disable-auto-image-base: Options. (line 1977) |
| 7685 | * --disable-auto-import: Options. (line 2112) |
| 7686 | * --disable-large-address-aware: Options. (line 1916) |
| 7687 | * --disable-long-section-names: Options. (line 1800) |
| 7688 | * --disable-new-dtags: Options. (line 1677) |
| 7689 | * --disable-runtime-pseudo-reloc: Options. (line 2125) |
| 7690 | * --disable-stdcall-fixup: Options. (line 1822) |
| 7691 | * --discard-all: Options. (line 647) |
| 7692 | * --discard-locals: Options. (line 651) |
| 7693 | * --dll: Options. (line 1795) |
| 7694 | * --dll-search-prefix: Options. (line 1983) |
| 7695 | * --dotsyms: PowerPC64 ELF64. (line 33) |
| 7696 | * --dsbt-index: Options. (line 2216) |
| 7697 | * --dsbt-size: Options. (line 2211) |
| 7698 | * --dynamic-linker=FILE: Options. (line 1041) |
| 7699 | * --dynamic-list-cpp-new: Options. (line 947) |
| 7700 | * --dynamic-list-cpp-typeinfo: Options. (line 951) |
| 7701 | * --dynamic-list-data: Options. (line 944) |
| 7702 | * --dynamic-list=DYNAMIC-LIST-FILE: Options. (line 931) |
| 7703 | * --dynamicbase: Options. (line 2165) |
| 7704 | * --eh-frame-hdr: Options. (line 1668) |
| 7705 | * --emit-relocs: Options. (line 497) |
| 7706 | * --emit-stack-syms: SPU ELF. (line 46) |
| 7707 | * --emit-stub-syms <1>: PowerPC64 ELF64. (line 29) |
| 7708 | * --emit-stub-syms <2>: PowerPC ELF32. (line 47) |
| 7709 | * --emit-stub-syms: SPU ELF. (line 15) |
| 7710 | * --enable-auto-image-base: Options. (line 1968) |
| 7711 | * --enable-auto-import: Options. (line 1992) |
| 7712 | * --enable-extra-pe-debug: Options. (line 2130) |
| 7713 | * --enable-long-section-names: Options. (line 1800) |
| 7714 | * --enable-new-dtags: Options. (line 1677) |
| 7715 | * --enable-runtime-pseudo-reloc: Options. (line 2117) |
| 7716 | * --enable-stdcall-fixup: Options. (line 1822) |
| 7717 | * --entry=ENTRY: Options. (line 187) |
| 7718 | * --error-poison-system-directories: Options. (line 1765) |
| 7719 | * --error-unresolved-symbols: Options. (line 1621) |
| 7720 | * --exclude-all-symbols: Options. (line 1876) |
| 7721 | * --exclude-libs: Options. (line 197) |
| 7722 | * --exclude-modules-for-implib: Options. (line 208) |
| 7723 | * --exclude-symbols: Options. (line 1870) |
| 7724 | * --export-all-symbols: Options. (line 1846) |
| 7725 | * --export-dynamic: Options. (line 221) |
| 7726 | * --extra-overlay-stubs: SPU ELF. (line 19) |
| 7727 | * --fatal-warnings: Options. (line 1054) |
| 7728 | * --file-alignment: Options. (line 1880) |
| 7729 | * --filter=NAME: Options. (line 276) |
| 7730 | * --fix-arm1176: ARM. (line 112) |
| 7731 | * --fix-cortex-a53-835769: ARM. (line 213) |
| 7732 | * --fix-cortex-a8: ARM. (line 204) |
| 7733 | * --fix-stm32l4xx-629360: ARM. (line 121) |
| 7734 | * --fix-v4bx: ARM. (line 50) |
| 7735 | * --fix-v4bx-interworking: ARM. (line 63) |
| 7736 | * --force-dynamic: Options. (line 506) |
| 7737 | * --force-exe-suffix: Options. (line 1059) |
| 7738 | * --forceinteg: Options. (line 2170) |
| 7739 | * --format=FORMAT: Options. (line 134) |
| 7740 | * --format=VERSION: TI COFF. (line 6) |
| 7741 | * --gc-sections: Options. (line 1069) |
| 7742 | * --got: Options. (line 2252) |
| 7743 | * --got=TYPE: M68K. (line 6) |
| 7744 | * --gpsize=VALUE: Options. (line 309) |
| 7745 | * --hash-size=NUMBER: Options. (line 1687) |
| 7746 | * --hash-style=STYLE: Options. (line 1695) |
| 7747 | * --heap: Options. (line 1886) |
| 7748 | * --help: Options. (line 1120) |
| 7749 | * --high-entropy-va: Options. (line 2161) |
| 7750 | * --image-base: Options. (line 1893) |
| 7751 | * --insert-timestamp: Options. (line 2193) |
| 7752 | * --insn32 <1>: Options. (line 2264) |
| 7753 | * --insn32: MIPS. (line 6) |
| 7754 | * --just-symbols=FILE: Options. (line 529) |
| 7755 | * --kill-at: Options. (line 1902) |
| 7756 | * --large-address-aware: Options. (line 1907) |
| 7757 | * --ld-generated-unwind-info: Options. (line 1672) |
| 7758 | * --leading-underscore: Options. (line 1840) |
| 7759 | * --library-path=DIR: Options. (line 367) |
| 7760 | * --library=NAMESPEC: Options. (line 334) |
| 7761 | * --local-store=lo:hi: SPU ELF. (line 24) |
| 7762 | * --long-plt: ARM. (line 224) |
| 7763 | * --major-image-version: Options. (line 1923) |
| 7764 | * --major-os-version: Options. (line 1928) |
| 7765 | * --major-subsystem-version: Options. (line 1932) |
| 7766 | * --merge-exidx-entries: ARM. (line 221) |
| 7767 | * --minor-image-version: Options. (line 1937) |
| 7768 | * --minor-os-version: Options. (line 1942) |
| 7769 | * --minor-subsystem-version: Options. (line 1946) |
| 7770 | * --mri-script=MRI-CMDFILE: Options. (line 158) |
| 7771 | * --multi-subspace: HPPA ELF32. (line 6) |
| 7772 | * --nmagic: Options. (line 439) |
| 7773 | * --no-accept-unknown-input-arch: Options. (line 841) |
| 7774 | * --no-add-needed: Options. (line 869) |
| 7775 | * --no-allow-shlib-undefined: Options. (line 1153) |
| 7776 | * --no-apply-dynamic-relocs: ARM. (line 228) |
| 7777 | * --no-as-needed: Options. (line 851) |
| 7778 | * --no-bind: Options. (line 2184) |
| 7779 | * --no-check-sections: Options. (line 955) |
| 7780 | * --no-copy-dt-needed-entries: Options. (line 967) |
| 7781 | * --no-define-common: Options. (line 1000) |
| 7782 | * --no-demangle: Options. (line 1028) |
| 7783 | * --no-dotsyms: PowerPC64 ELF64. (line 33) |
| 7784 | * --no-dynamic-linker: Options. (line 1048) |
| 7785 | * --no-enum-size-warning: ARM. (line 159) |
| 7786 | * --no-export-dynamic: Options. (line 221) |
| 7787 | * --no-fatal-warnings: Options. (line 1054) |
| 7788 | * --no-fix-arm1176: ARM. (line 112) |
| 7789 | * --no-fix-cortex-a53-835769: ARM. (line 213) |
| 7790 | * --no-fix-cortex-a8: ARM. (line 204) |
| 7791 | * --no-gc-sections: Options. (line 1069) |
| 7792 | * --no-insn32 <1>: Options. (line 2265) |
| 7793 | * --no-insn32: MIPS. (line 6) |
| 7794 | * --no-isolation: Options. (line 2177) |
| 7795 | * --no-keep-memory: Options. (line 1132) |
| 7796 | * --no-leading-underscore: Options. (line 1840) |
| 7797 | * --no-merge-exidx-entries <1>: Options. (line 2223) |
| 7798 | * --no-merge-exidx-entries: ARM. (line 221) |
| 7799 | * --no-multi-toc: PowerPC64 ELF64. (line 97) |
| 7800 | * --no-omagic: Options. (line 454) |
| 7801 | * --no-opd-optimize: PowerPC64 ELF64. (line 71) |
| 7802 | * --no-overlays: SPU ELF. (line 9) |
| 7803 | * --no-plt-align: PowerPC64 ELF64. (line 119) |
| 7804 | * --no-plt-static-chain: PowerPC64 ELF64. (line 127) |
| 7805 | * --no-plt-thread-safe: PowerPC64 ELF64. (line 133) |
| 7806 | * --no-print-gc-sections: Options. (line 1093) |
| 7807 | * --no-save-restore-funcs: PowerPC64 ELF64. (line 44) |
| 7808 | * --no-seh: Options. (line 2180) |
| 7809 | * --no-tls-get-addr-optimize: PowerPC64 ELF64. (line 57) |
| 7810 | * --no-tls-optimize <1>: PowerPC ELF32. (line 51) |
| 7811 | * --no-tls-optimize: PowerPC64 ELF64. (line 52) |
| 7812 | * --no-toc-optimize: PowerPC64 ELF64. (line 83) |
| 7813 | * --no-toc-sort: PowerPC64 ELF64. (line 109) |
| 7814 | * --no-trampoline: Options. (line 2233) |
| 7815 | * --no-undefined: Options. (line 1139) |
| 7816 | * --no-undefined-version: Options. (line 1181) |
| 7817 | * --no-warn-mismatch: Options. (line 1194) |
| 7818 | * --no-warn-poison-system-directories: Options. (line 1759) |
| 7819 | * --no-warn-search-mismatch: Options. (line 1203) |
| 7820 | * --no-wchar-size-warning: ARM. (line 166) |
| 7821 | * --no-whole-archive: Options. (line 1207) |
| 7822 | * --noinhibit-exec: Options. (line 1211) |
| 7823 | * --non-overlapping-opd: PowerPC64 ELF64. (line 77) |
| 7824 | * --nxcompat: Options. (line 2173) |
| 7825 | * --oformat=OUTPUT-FORMAT: Options. (line 1223) |
| 7826 | * --omagic: Options. (line 445) |
| 7827 | * --orphan-handling=MODE: Options. (line 606) |
| 7828 | * --out-implib: Options. (line 1959) |
| 7829 | * --output-def: Options. (line 1951) |
| 7830 | * --output=OUTPUT: Options. (line 460) |
| 7831 | * --pic-executable: Options. (line 1236) |
| 7832 | * --pic-veneer: ARM. (line 172) |
| 7833 | * --plt-align: PowerPC64 ELF64. (line 119) |
| 7834 | * --plt-static-chain: PowerPC64 ELF64. (line 127) |
| 7835 | * --plt-thread-safe: PowerPC64 ELF64. (line 133) |
| 7836 | * --plugin: SPU ELF. (line 6) |
| 7837 | * --pop-state: Options. (line 494) |
| 7838 | * --print-gc-sections: Options. (line 1093) |
| 7839 | * --print-map: Options. (line 402) |
| 7840 | * --print-memory-usage: Options. (line 1108) |
| 7841 | * --print-output-format: Options. (line 1102) |
| 7842 | * --push-state: Options. (line 476) |
| 7843 | * --reduce-memory-overheads: Options. (line 1717) |
| 7844 | * --relax: Options. (line 1252) |
| 7845 | * --relax on i960: i960. (line 31) |
| 7846 | * --relax on Nios II: Nios II. (line 6) |
| 7847 | * --relax on PowerPC: PowerPC ELF32. (line 6) |
| 7848 | * --relax on Xtensa: Xtensa. (line 27) |
| 7849 | * --relocatable: Options. (line 510) |
| 7850 | * --require-defined=SYMBOL: Options. (line 588) |
| 7851 | * --retain-symbols-file=FILENAME: Options. (line 1278) |
| 7852 | * --save-restore-funcs: PowerPC64 ELF64. (line 44) |
| 7853 | * --script=SCRIPT: Options. (line 553) |
| 7854 | * --sdata-got: PowerPC ELF32. (line 33) |
| 7855 | * --section-alignment: Options. (line 2135) |
| 7856 | * --section-start=SECTIONNAME=ORG: Options. (line 1434) |
| 7857 | * --secure-plt: PowerPC ELF32. (line 26) |
| 7858 | * --sort-common: Options. (line 1376) |
| 7859 | * --sort-section=alignment: Options. (line 1391) |
| 7860 | * --sort-section=name: Options. (line 1387) |
| 7861 | * --split-by-file: Options. (line 1395) |
| 7862 | * --split-by-reloc: Options. (line 1400) |
| 7863 | * --stack: Options. (line 2141) |
| 7864 | * --stack-analysis: SPU ELF. (line 29) |
| 7865 | * --stats: Options. (line 1413) |
| 7866 | * --strip-all: Options. (line 540) |
| 7867 | * --strip-debug: Options. (line 544) |
| 7868 | * --stub-group-size: PowerPC64 ELF64. (line 6) |
| 7869 | * --stub-group-size=N <1>: ARM. (line 177) |
| 7870 | * --stub-group-size=N: HPPA ELF32. (line 12) |
| 7871 | * --subsystem: Options. (line 2148) |
| 7872 | * --support-old-code: ARM. (line 6) |
| 7873 | * --sysroot=DIRECTORY: Options. (line 1417) |
| 7874 | * --target-help: Options. (line 1124) |
| 7875 | * --target1-abs: ARM. (line 33) |
| 7876 | * --target1-rel: ARM. (line 33) |
| 7877 | * --target2=TYPE: ARM. (line 38) |
| 7878 | * --thumb-entry=ENTRY: ARM. (line 17) |
| 7879 | * --tls-get-addr-optimize: PowerPC64 ELF64. (line 57) |
| 7880 | * --trace: Options. (line 549) |
| 7881 | * --trace-symbol=SYMBOL: Options. (line 657) |
| 7882 | * --traditional-format: Options. (line 1422) |
| 7883 | * --tsaware: Options. (line 2190) |
| 7884 | * --undefined=SYMBOL: Options. (line 575) |
| 7885 | * --unique[=SECTION]: Options. (line 632) |
| 7886 | * --unresolved-symbols: Options. (line 1464) |
| 7887 | * --use-blx: ARM. (line 75) |
| 7888 | * --use-nul-prefixed-import-tables: ARM. (line 23) |
| 7889 | * --verbose[=NUMBER]: Options. (line 1493) |
| 7890 | * --version: Options. (line 641) |
| 7891 | * --version-script=VERSION-SCRIPTFILE: Options. (line 1501) |
| 7892 | * --vfp11-denorm-fix: ARM. (line 84) |
| 7893 | * --warn-alternate-em: Options. (line 1613) |
| 7894 | * --warn-common: Options. (line 1512) |
| 7895 | * --warn-constructors: Options. (line 1580) |
| 7896 | * --warn-multiple-gp: Options. (line 1585) |
| 7897 | * --warn-once: Options. (line 1599) |
| 7898 | * --warn-poison-system-directories: Options. (line 1753) |
| 7899 | * --warn-section-align: Options. (line 1603) |
| 7900 | * --warn-shared-textrel: Options. (line 1610) |
| 7901 | * --warn-unresolved-symbols: Options. (line 1616) |
| 7902 | * --wdmdriver: Options. (line 2187) |
| 7903 | * --whole-archive: Options. (line 1625) |
| 7904 | * --wrap=SYMBOL: Options. (line 1639) |
| 7905 | * -A ARCH: Options. (line 122) |
| 7906 | * -a KEYWORD: Options. (line 105) |
| 7907 | * -assert KEYWORD: Options. (line 876) |
| 7908 | * -b FORMAT: Options. (line 134) |
| 7909 | * -Bdynamic: Options. (line 879) |
| 7910 | * -Bgroup: Options. (line 889) |
| 7911 | * -Bshareable: Options. (line 1369) |
| 7912 | * -Bstatic: Options. (line 896) |
| 7913 | * -Bsymbolic: Options. (line 911) |
| 7914 | * -Bsymbolic-functions: Options. (line 922) |
| 7915 | * -c MRI-CMDFILE: Options. (line 158) |
| 7916 | * -call_shared: Options. (line 879) |
| 7917 | * -d: Options. (line 168) |
| 7918 | * -dc: Options. (line 168) |
| 7919 | * -dn: Options. (line 896) |
| 7920 | * -dp: Options. (line 168) |
| 7921 | * -dT SCRIPT: Options. (line 562) |
| 7922 | * -dy: Options. (line 879) |
| 7923 | * -E: Options. (line 221) |
| 7924 | * -e ENTRY: Options. (line 187) |
| 7925 | * -EB: Options. (line 248) |
| 7926 | * -EL: Options. (line 251) |
| 7927 | * -f NAME: Options. (line 255) |
| 7928 | * -F NAME: Options. (line 276) |
| 7929 | * -fini=NAME: Options. (line 300) |
| 7930 | * -g: Options. (line 306) |
| 7931 | * -G VALUE: Options. (line 309) |
| 7932 | * -h NAME: Options. (line 316) |
| 7933 | * -i: Options. (line 325) |
| 7934 | * -IFILE: Options. (line 1041) |
| 7935 | * -init=NAME: Options. (line 328) |
| 7936 | * -L DIR: Options. (line 367) |
| 7937 | * -l NAMESPEC: Options. (line 334) |
| 7938 | * -M: Options. (line 402) |
| 7939 | * -m EMULATION: Options. (line 392) |
| 7940 | * -Map=MAPFILE: Options. (line 1128) |
| 7941 | * -n: Options. (line 439) |
| 7942 | * -N: Options. (line 445) |
| 7943 | * -no-relax: Options. (line 1252) |
| 7944 | * -non_shared: Options. (line 896) |
| 7945 | * -nostdlib: Options. (line 1217) |
| 7946 | * -O LEVEL: Options. (line 466) |
| 7947 | * -o OUTPUT: Options. (line 460) |
| 7948 | * -P AUDITLIB: Options. (line 177) |
| 7949 | * -pie: Options. (line 1236) |
| 7950 | * -q: Options. (line 497) |
| 7951 | * -qmagic: Options. (line 1246) |
| 7952 | * -Qy: Options. (line 1249) |
| 7953 | * -r: Options. (line 510) |
| 7954 | * -R FILE: Options. (line 529) |
| 7955 | * -rpath-link=DIR: Options. (line 1314) |
| 7956 | * -rpath=DIR: Options. (line 1292) |
| 7957 | * -s: Options. (line 540) |
| 7958 | * -S: Options. (line 544) |
| 7959 | * -shared: Options. (line 1369) |
| 7960 | * -soname=NAME: Options. (line 316) |
| 7961 | * -static: Options. (line 896) |
| 7962 | * -t: Options. (line 549) |
| 7963 | * -T SCRIPT: Options. (line 553) |
| 7964 | * -Tbss=ORG: Options. (line 1443) |
| 7965 | * -Tdata=ORG: Options. (line 1443) |
| 7966 | * -Tldata-segment=ORG: Options. (line 1459) |
| 7967 | * -Trodata-segment=ORG: Options. (line 1453) |
| 7968 | * -Ttext-segment=ORG: Options. (line 1449) |
| 7969 | * -Ttext=ORG: Options. (line 1443) |
| 7970 | * -u SYMBOL: Options. (line 575) |
| 7971 | * -Ur: Options. (line 596) |
| 7972 | * -V: Options. (line 641) |
| 7973 | * -v: Options. (line 641) |
| 7974 | * -X: Options. (line 651) |
| 7975 | * -x: Options. (line 647) |
| 7976 | * -Y PATH: Options. (line 666) |
| 7977 | * -y SYMBOL: Options. (line 657) |
| 7978 | * -z defs: Options. (line 1139) |
| 7979 | * -z KEYWORD: Options. (line 670) |
| 7980 | * -z muldefs: Options. (line 1147) |
| 7981 | * .: Location Counter. (line 6) |
| 7982 | * /DISCARD/: Output Section Discarding. |
| 7983 | (line 26) |
| 7984 | * 32-bit PLT entries: ARM. (line 224) |
| 7985 | * :PHDR: Output Section Phdr. |
| 7986 | (line 6) |
| 7987 | * =FILLEXP: Output Section Fill. |
| 7988 | (line 6) |
| 7989 | * >REGION: Output Section Region. |
| 7990 | (line 6) |
| 7991 | * [COMMON]: Input Section Common. |
| 7992 | (line 29) |
| 7993 | * AArch64 rela addend: ARM. (line 228) |
| 7994 | * ABSOLUTE (MRI): MRI. (line 33) |
| 7995 | * absolute and relocatable symbols: Expression Section. (line 6) |
| 7996 | * absolute expressions: Expression Section. (line 6) |
| 7997 | * ABSOLUTE(EXP): Builtin Functions. (line 10) |
| 7998 | * ADDR(SECTION): Builtin Functions. (line 17) |
| 7999 | * address, section: Output Section Address. |
| 8000 | (line 6) |
| 8001 | * ALIAS (MRI): MRI. (line 44) |
| 8002 | * ALIGN (MRI): MRI. (line 50) |
| 8003 | * align expression: Builtin Functions. (line 38) |
| 8004 | * align location counter: Builtin Functions. (line 38) |
| 8005 | * ALIGN(ALIGN): Builtin Functions. (line 38) |
| 8006 | * ALIGN(EXP,ALIGN): Builtin Functions. (line 38) |
| 8007 | * ALIGN(SECTION_ALIGN): Forced Output Alignment. |
| 8008 | (line 6) |
| 8009 | * aligned common symbols: WIN32. (line 424) |
| 8010 | * ALIGNOF(SECTION): Builtin Functions. (line 64) |
| 8011 | * allocating memory: MEMORY. (line 6) |
| 8012 | * architecture: Miscellaneous Commands. |
| 8013 | (line 115) |
| 8014 | * architectures: Options. (line 122) |
| 8015 | * archive files, from cmd line: Options. (line 334) |
| 8016 | * archive search path in linker script: File Commands. (line 76) |
| 8017 | * arithmetic: Expressions. (line 6) |
| 8018 | * arithmetic operators: Operators. (line 6) |
| 8019 | * ARM interworking support: ARM. (line 6) |
| 8020 | * ARM1176 erratum workaround: ARM. (line 112) |
| 8021 | * AS_NEEDED(FILES): File Commands. (line 56) |
| 8022 | * ASSERT: Miscellaneous Commands. |
| 8023 | (line 9) |
| 8024 | * assertion in linker script: Miscellaneous Commands. |
| 8025 | (line 9) |
| 8026 | * assignment in scripts: Assignments. (line 6) |
| 8027 | * AT(LMA): Output Section LMA. (line 6) |
| 8028 | * AT>LMA_REGION: Output Section LMA. (line 6) |
| 8029 | * automatic data imports: WIN32. (line 191) |
| 8030 | * back end: BFD. (line 6) |
| 8031 | * BASE (MRI): MRI. (line 54) |
| 8032 | * BE8: ARM. (line 28) |
| 8033 | * BFD canonical format: Canonical format. (line 11) |
| 8034 | * BFD requirements: BFD. (line 16) |
| 8035 | * big-endian objects: Options. (line 248) |
| 8036 | * binary input format: Options. (line 134) |
| 8037 | * BLOCK(EXP): Builtin Functions. (line 77) |
| 8038 | * bug criteria: Bug Criteria. (line 6) |
| 8039 | * bug reports: Bug Reporting. (line 6) |
| 8040 | * bugs in ld: Reporting Bugs. (line 6) |
| 8041 | * BYTE(EXPRESSION): Output Section Data. |
| 8042 | (line 6) |
| 8043 | * C++ constructors, arranging in link: Output Section Keywords. |
| 8044 | (line 19) |
| 8045 | * CHIP (MRI): MRI. (line 58) |
| 8046 | * COLLECT_NO_DEMANGLE: Environment. (line 29) |
| 8047 | * combining symbols, warnings on: Options. (line 1512) |
| 8048 | * command files: Scripts. (line 6) |
| 8049 | * command line: Options. (line 6) |
| 8050 | * common allocation: Options. (line 168) |
| 8051 | * common allocation in linker script: Miscellaneous Commands. |
| 8052 | (line 51) |
| 8053 | * common symbol placement: Input Section Common. |
| 8054 | (line 6) |
| 8055 | * COMMONPAGESIZE: Symbolic Constants. (line 13) |
| 8056 | * compatibility, MRI: Options. (line 158) |
| 8057 | * CONSTANT: Symbolic Constants. (line 6) |
| 8058 | * constants in linker scripts: Constants. (line 6) |
| 8059 | * constraints on output sections: Output Section Constraint. |
| 8060 | (line 6) |
| 8061 | * constructors: Options. (line 596) |
| 8062 | * CONSTRUCTORS: Output Section Keywords. |
| 8063 | (line 19) |
| 8064 | * constructors, arranging in link: Output Section Keywords. |
| 8065 | (line 19) |
| 8066 | * Cortex-A53 erratum 835769 workaround: ARM. (line 213) |
| 8067 | * Cortex-A8 erratum workaround: ARM. (line 204) |
| 8068 | * crash of linker: Bug Criteria. (line 9) |
| 8069 | * CREATE_OBJECT_SYMBOLS: Output Section Keywords. |
| 8070 | (line 9) |
| 8071 | * creating a DEF file: WIN32. (line 158) |
| 8072 | * cross reference table: Options. (line 987) |
| 8073 | * cross references: Miscellaneous Commands. |
| 8074 | (line 98) |
| 8075 | * current output location: Location Counter. (line 6) |
| 8076 | * data: Output Section Data. |
| 8077 | (line 6) |
| 8078 | * DATA_SEGMENT_ALIGN(MAXPAGESIZE, COMMONPAGESIZE): Builtin Functions. |
| 8079 | (line 82) |
| 8080 | * DATA_SEGMENT_END(EXP): Builtin Functions. (line 104) |
| 8081 | * DATA_SEGMENT_RELRO_END(OFFSET, EXP): Builtin Functions. (line 110) |
| 8082 | * dbx: Options. (line 1427) |
| 8083 | * DEF files, creating: Options. (line 1951) |
| 8084 | * default emulation: Environment. (line 21) |
| 8085 | * default input format: Environment. (line 9) |
| 8086 | * defined symbol: Options. (line 588) |
| 8087 | * DEFINED(SYMBOL): Builtin Functions. (line 123) |
| 8088 | * deleting local symbols: Options. (line 647) |
| 8089 | * demangling, default: Environment. (line 29) |
| 8090 | * demangling, from command line: Options. (line 1028) |
| 8091 | * direct linking to a dll: WIN32. (line 239) |
| 8092 | * discarding sections: Output Section Discarding. |
| 8093 | (line 6) |
| 8094 | * discontinuous memory: MEMORY. (line 6) |
| 8095 | * DLLs, creating: Options. (line 1846) |
| 8096 | * DLLs, linking to: Options. (line 1983) |
| 8097 | * dot: Location Counter. (line 6) |
| 8098 | * dot inside sections: Location Counter. (line 36) |
| 8099 | * dot outside sections: Location Counter. (line 66) |
| 8100 | * dynamic linker, from command line: Options. (line 1041) |
| 8101 | * dynamic symbol table: Options. (line 221) |
| 8102 | * ELF program headers: PHDRS. (line 6) |
| 8103 | * emulation: Options. (line 392) |
| 8104 | * emulation, default: Environment. (line 21) |
| 8105 | * END (MRI): MRI. (line 62) |
| 8106 | * endianness: Options. (line 248) |
| 8107 | * entry point: Entry Point. (line 6) |
| 8108 | * entry point, from command line: Options. (line 187) |
| 8109 | * entry point, thumb: ARM. (line 17) |
| 8110 | * ENTRY(SYMBOL): Entry Point. (line 6) |
| 8111 | * error on valid input: Bug Criteria. (line 12) |
| 8112 | * example of linker script: Simple Example. (line 6) |
| 8113 | * exporting DLL symbols: WIN32. (line 19) |
| 8114 | * expression evaluation order: Evaluation. (line 6) |
| 8115 | * expression sections: Expression Section. (line 6) |
| 8116 | * expression, absolute: Builtin Functions. (line 10) |
| 8117 | * expressions: Expressions. (line 6) |
| 8118 | * EXTERN: Miscellaneous Commands. |
| 8119 | (line 39) |
| 8120 | * fatal signal: Bug Criteria. (line 9) |
| 8121 | * file name wildcard patterns: Input Section Wildcards. |
| 8122 | (line 6) |
| 8123 | * FILEHDR: PHDRS. (line 62) |
| 8124 | * filename symbols: Output Section Keywords. |
| 8125 | (line 9) |
| 8126 | * fill pattern, entire section: Output Section Fill. |
| 8127 | (line 6) |
| 8128 | * FILL(EXPRESSION): Output Section Data. |
| 8129 | (line 39) |
| 8130 | * finalization function: Options. (line 300) |
| 8131 | * first input file: File Commands. (line 84) |
| 8132 | * first instruction: Entry Point. (line 6) |
| 8133 | * FIX_V4BX: ARM. (line 50) |
| 8134 | * FIX_V4BX_INTERWORKING: ARM. (line 63) |
| 8135 | * FORCE_COMMON_ALLOCATION: Miscellaneous Commands. |
| 8136 | (line 46) |
| 8137 | * forcing input section alignment: Forced Input Alignment. |
| 8138 | (line 6) |
| 8139 | * forcing output section alignment: Forced Output Alignment. |
| 8140 | (line 6) |
| 8141 | * forcing the creation of dynamic sections: Options. (line 506) |
| 8142 | * FORMAT (MRI): MRI. (line 66) |
| 8143 | * functions in expressions: Builtin Functions. (line 6) |
| 8144 | * garbage collection <1>: Input Section Keep. (line 6) |
| 8145 | * garbage collection: Options. (line 1069) |
| 8146 | * generating optimized output: Options. (line 466) |
| 8147 | * GNU linker: Overview. (line 6) |
| 8148 | * GNUTARGET: Environment. (line 9) |
| 8149 | * GROUP(FILES): File Commands. (line 49) |
| 8150 | * grouping input files: File Commands. (line 49) |
| 8151 | * groups of archives: Options. (line 823) |
| 8152 | * H8/300 support: H8/300. (line 6) |
| 8153 | * header size: Builtin Functions. (line 191) |
| 8154 | * heap size: Options. (line 1886) |
| 8155 | * help: Options. (line 1120) |
| 8156 | * HIDDEN: HIDDEN. (line 6) |
| 8157 | * holes: Location Counter. (line 12) |
| 8158 | * holes, filling: Output Section Data. |
| 8159 | (line 39) |
| 8160 | * HPPA multiple sub-space stubs: HPPA ELF32. (line 6) |
| 8161 | * HPPA stub grouping: HPPA ELF32. (line 12) |
| 8162 | * i960 support: i960. (line 6) |
| 8163 | * image base: Options. (line 1893) |
| 8164 | * implicit linker scripts: Implicit Linker Scripts. |
| 8165 | (line 6) |
| 8166 | * import libraries: WIN32. (line 10) |
| 8167 | * INCLUDE FILENAME: File Commands. (line 9) |
| 8168 | * including a linker script: File Commands. (line 9) |
| 8169 | * including an entire archive: Options. (line 1625) |
| 8170 | * incremental link: Options. (line 325) |
| 8171 | * INHIBIT_COMMON_ALLOCATION: Miscellaneous Commands. |
| 8172 | (line 51) |
| 8173 | * initialization function: Options. (line 328) |
| 8174 | * initialized data in ROM: Output Section LMA. (line 39) |
| 8175 | * input file format in linker script: Format Commands. (line 35) |
| 8176 | * input filename symbols: Output Section Keywords. |
| 8177 | (line 9) |
| 8178 | * input files in linker scripts: File Commands. (line 19) |
| 8179 | * input files, displaying: Options. (line 549) |
| 8180 | * input format: Options. (line 134) |
| 8181 | * input object files in linker scripts: File Commands. (line 19) |
| 8182 | * input section alignment: Forced Input Alignment. |
| 8183 | (line 6) |
| 8184 | * input section basics: Input Section Basics. |
| 8185 | (line 6) |
| 8186 | * input section wildcards: Input Section Wildcards. |
| 8187 | (line 6) |
| 8188 | * input sections: Input Section. (line 6) |
| 8189 | * INPUT(FILES): File Commands. (line 19) |
| 8190 | * INSERT: Miscellaneous Commands. |
| 8191 | (line 56) |
| 8192 | * insert user script into default script: Miscellaneous Commands. |
| 8193 | (line 56) |
| 8194 | * integer notation: Constants. (line 6) |
| 8195 | * integer suffixes: Constants. (line 15) |
| 8196 | * internal object-file format: Canonical format. (line 11) |
| 8197 | * invalid input: Bug Criteria. (line 14) |
| 8198 | * K and M integer suffixes: Constants. (line 15) |
| 8199 | * KEEP: Input Section Keep. (line 6) |
| 8200 | * l =: MEMORY. (line 74) |
| 8201 | * lazy evaluation: Evaluation. (line 6) |
| 8202 | * ld bugs, reporting: Bug Reporting. (line 6) |
| 8203 | * LD_FEATURE(STRING): Miscellaneous Commands. |
| 8204 | (line 121) |
| 8205 | * ldata segment origin, cmd line: Options. (line 1460) |
| 8206 | * LDEMULATION: Environment. (line 21) |
| 8207 | * len =: MEMORY. (line 74) |
| 8208 | * LENGTH =: MEMORY. (line 74) |
| 8209 | * LENGTH(MEMORY): Builtin Functions. (line 140) |
| 8210 | * library search path in linker script: File Commands. (line 76) |
| 8211 | * link map: Options. (line 402) |
| 8212 | * link-time runtime library search path: Options. (line 1314) |
| 8213 | * linker crash: Bug Criteria. (line 9) |
| 8214 | * linker script concepts: Basic Script Concepts. |
| 8215 | (line 6) |
| 8216 | * linker script example: Simple Example. (line 6) |
| 8217 | * linker script file commands: File Commands. (line 6) |
| 8218 | * linker script format: Script Format. (line 6) |
| 8219 | * linker script input object files: File Commands. (line 19) |
| 8220 | * linker script simple commands: Simple Commands. (line 6) |
| 8221 | * linker scripts: Scripts. (line 6) |
| 8222 | * LIST (MRI): MRI. (line 77) |
| 8223 | * little-endian objects: Options. (line 251) |
| 8224 | * LOAD (MRI): MRI. (line 84) |
| 8225 | * load address: Output Section LMA. (line 6) |
| 8226 | * LOADADDR(SECTION): Builtin Functions. (line 143) |
| 8227 | * loading, preventing: Output Section Type. |
| 8228 | (line 22) |
| 8229 | * local symbols, deleting: Options. (line 651) |
| 8230 | * location counter: Location Counter. (line 6) |
| 8231 | * LOG2CEIL(EXP): Builtin Functions. (line 147) |
| 8232 | * LONG(EXPRESSION): Output Section Data. |
| 8233 | (line 6) |
| 8234 | * M and K integer suffixes: Constants. (line 15) |
| 8235 | * M68HC11 and 68HC12 support: M68HC11/68HC12. (line 6) |
| 8236 | * machine architecture: Miscellaneous Commands. |
| 8237 | (line 115) |
| 8238 | * machine dependencies: Machine Dependent. (line 6) |
| 8239 | * mapping input sections to output sections: Input Section. (line 6) |
| 8240 | * MAX: Builtin Functions. (line 150) |
| 8241 | * MAXPAGESIZE: Symbolic Constants. (line 10) |
| 8242 | * MEMORY: MEMORY. (line 6) |
| 8243 | * memory region attributes: MEMORY. (line 34) |
| 8244 | * memory regions: MEMORY. (line 6) |
| 8245 | * memory regions and sections: Output Section Region. |
| 8246 | (line 6) |
| 8247 | * memory usage: Options. (line 1108) |
| 8248 | * Merging exidx entries: ARM. (line 221) |
| 8249 | * MIN: Builtin Functions. (line 153) |
| 8250 | * MIPS microMIPS instruction choice selection: MIPS. (line 6) |
| 8251 | * Motorola 68K GOT generation: M68K. (line 6) |
| 8252 | * MRI compatibility: MRI. (line 6) |
| 8253 | * MSP430 extra sections: MSP430. (line 11) |
| 8254 | * NAME (MRI): MRI. (line 90) |
| 8255 | * name, section: Output Section Name. |
| 8256 | (line 6) |
| 8257 | * names: Symbols. (line 6) |
| 8258 | * naming the output file: Options. (line 460) |
| 8259 | * NEXT(EXP): Builtin Functions. (line 157) |
| 8260 | * Nios II call relaxation: Nios II. (line 6) |
| 8261 | * NMAGIC: Options. (line 439) |
| 8262 | * NO_ENUM_SIZE_WARNING: ARM. (line 159) |
| 8263 | * NO_WCHAR_SIZE_WARNING: ARM. (line 166) |
| 8264 | * NOCROSSREFS(SECTIONS): Miscellaneous Commands. |
| 8265 | (line 82) |
| 8266 | * NOCROSSREFS_TO(TOSECTION FROMSECTIONS): Miscellaneous Commands. |
| 8267 | (line 98) |
| 8268 | * NOLOAD: Output Section Type. |
| 8269 | (line 22) |
| 8270 | * not enough room for program headers: Builtin Functions. (line 196) |
| 8271 | * o =: MEMORY. (line 69) |
| 8272 | * objdump -i: BFD. (line 6) |
| 8273 | * object file management: BFD. (line 6) |
| 8274 | * object files: Options. (line 29) |
| 8275 | * object formats available: BFD. (line 6) |
| 8276 | * object size: Options. (line 309) |
| 8277 | * OMAGIC: Options. (line 454) |
| 8278 | * ONLY_IF_RO: Output Section Constraint. |
| 8279 | (line 6) |
| 8280 | * ONLY_IF_RW: Output Section Constraint. |
| 8281 | (line 6) |
| 8282 | * opening object files: BFD outline. (line 6) |
| 8283 | * operators for arithmetic: Operators. (line 6) |
| 8284 | * options: Options. (line 6) |
| 8285 | * ORDER (MRI): MRI. (line 95) |
| 8286 | * org =: MEMORY. (line 69) |
| 8287 | * ORIGIN =: MEMORY. (line 69) |
| 8288 | * ORIGIN(MEMORY): Builtin Functions. (line 163) |
| 8289 | * orphan: Orphan Sections. (line 6) |
| 8290 | * orphan sections: Options. (line 606) |
| 8291 | * output file after errors: Options. (line 1211) |
| 8292 | * output file format in linker script: Format Commands. (line 10) |
| 8293 | * output file name in linker script: File Commands. (line 66) |
| 8294 | * output format: Options. (line 1102) |
| 8295 | * output section alignment: Forced Output Alignment. |
| 8296 | (line 6) |
| 8297 | * output section attributes: Output Section Attributes. |
| 8298 | (line 6) |
| 8299 | * output section data: Output Section Data. |
| 8300 | (line 6) |
| 8301 | * OUTPUT(FILENAME): File Commands. (line 66) |
| 8302 | * OUTPUT_ARCH(BFDARCH): Miscellaneous Commands. |
| 8303 | (line 115) |
| 8304 | * OUTPUT_FORMAT(BFDNAME): Format Commands. (line 10) |
| 8305 | * OVERLAY: Overlay Description. |
| 8306 | (line 6) |
| 8307 | * overlays: Overlay Description. |
| 8308 | (line 6) |
| 8309 | * partial link: Options. (line 510) |
| 8310 | * PE import table prefixing: ARM. (line 23) |
| 8311 | * PHDRS: PHDRS. (line 6) |
| 8312 | * PIC_VENEER: ARM. (line 172) |
| 8313 | * pop state governing input file handling: Options. (line 494) |
| 8314 | * position independent executables: Options. (line 1238) |
| 8315 | * PowerPC ELF32 options: PowerPC ELF32. (line 16) |
| 8316 | * PowerPC GOT: PowerPC ELF32. (line 33) |
| 8317 | * PowerPC long branches: PowerPC ELF32. (line 6) |
| 8318 | * PowerPC PLT: PowerPC ELF32. (line 16) |
| 8319 | * PowerPC stub symbols: PowerPC ELF32. (line 47) |
| 8320 | * PowerPC TLS optimization: PowerPC ELF32. (line 51) |
| 8321 | * PowerPC64 __tls_get_addr optimization: PowerPC64 ELF64. (line 57) |
| 8322 | * PowerPC64 dot symbols: PowerPC64 ELF64. (line 33) |
| 8323 | * PowerPC64 ELF64 options: PowerPC64 ELF64. (line 6) |
| 8324 | * PowerPC64 multi-TOC: PowerPC64 ELF64. (line 97) |
| 8325 | * PowerPC64 OPD optimization: PowerPC64 ELF64. (line 71) |
| 8326 | * PowerPC64 OPD spacing: PowerPC64 ELF64. (line 77) |
| 8327 | * PowerPC64 PLT call stub static chain: PowerPC64 ELF64. (line 127) |
| 8328 | * PowerPC64 PLT call stub thread safety: PowerPC64 ELF64. (line 133) |
| 8329 | * PowerPC64 PLT stub alignment: PowerPC64 ELF64. (line 119) |
| 8330 | * PowerPC64 register save/restore functions: PowerPC64 ELF64. |
| 8331 | (line 44) |
| 8332 | * PowerPC64 stub grouping: PowerPC64 ELF64. (line 6) |
| 8333 | * PowerPC64 stub symbols: PowerPC64 ELF64. (line 29) |
| 8334 | * PowerPC64 TLS optimization: PowerPC64 ELF64. (line 52) |
| 8335 | * PowerPC64 TOC optimization: PowerPC64 ELF64. (line 83) |
| 8336 | * PowerPC64 TOC sorting: PowerPC64 ELF64. (line 109) |
| 8337 | * precedence in expressions: Operators. (line 6) |
| 8338 | * prevent unnecessary loading: Output Section Type. |
| 8339 | (line 22) |
| 8340 | * program headers: PHDRS. (line 6) |
| 8341 | * program headers and sections: Output Section Phdr. |
| 8342 | (line 6) |
| 8343 | * program headers, not enough room: Builtin Functions. (line 196) |
| 8344 | * program segments: PHDRS. (line 6) |
| 8345 | * PROVIDE: PROVIDE. (line 6) |
| 8346 | * PROVIDE_HIDDEN: PROVIDE_HIDDEN. (line 6) |
| 8347 | * PUBLIC (MRI): MRI. (line 103) |
| 8348 | * push state governing input file handling: Options. (line 476) |
| 8349 | * QUAD(EXPRESSION): Output Section Data. |
| 8350 | (line 6) |
| 8351 | * quoted symbol names: Symbols. (line 6) |
| 8352 | * read-only text: Options. (line 439) |
| 8353 | * read/write from cmd line: Options. (line 445) |
| 8354 | * region alias: REGION_ALIAS. (line 6) |
| 8355 | * region names: REGION_ALIAS. (line 6) |
| 8356 | * REGION_ALIAS(ALIAS, REGION): REGION_ALIAS. (line 6) |
| 8357 | * regions of memory: MEMORY. (line 6) |
| 8358 | * relative expressions: Expression Section. (line 6) |
| 8359 | * relaxing addressing modes: Options. (line 1252) |
| 8360 | * relaxing on H8/300: H8/300. (line 9) |
| 8361 | * relaxing on i960: i960. (line 31) |
| 8362 | * relaxing on M68HC11: M68HC11/68HC12. (line 12) |
| 8363 | * relaxing on NDS32: NDS32. (line 6) |
| 8364 | * relaxing on Xtensa: Xtensa. (line 27) |
| 8365 | * relocatable and absolute symbols: Expression Section. (line 6) |
| 8366 | * relocatable output: Options. (line 510) |
| 8367 | * removing sections: Output Section Discarding. |
| 8368 | (line 6) |
| 8369 | * reporting bugs in ld: Reporting Bugs. (line 6) |
| 8370 | * requirements for BFD: BFD. (line 16) |
| 8371 | * retain relocations in final executable: Options. (line 497) |
| 8372 | * retaining specified symbols: Options. (line 1278) |
| 8373 | * rodata segment origin, cmd line: Options. (line 1454) |
| 8374 | * ROM initialized data: Output Section LMA. (line 39) |
| 8375 | * round up expression: Builtin Functions. (line 38) |
| 8376 | * round up location counter: Builtin Functions. (line 38) |
| 8377 | * runtime library name: Options. (line 316) |
| 8378 | * runtime library search path: Options. (line 1292) |
| 8379 | * runtime pseudo-relocation: WIN32. (line 217) |
| 8380 | * scaled integers: Constants. (line 15) |
| 8381 | * scommon section: Input Section Common. |
| 8382 | (line 20) |
| 8383 | * script files: Options. (line 562) |
| 8384 | * scripts: Scripts. (line 6) |
| 8385 | * search directory, from cmd line: Options. (line 367) |
| 8386 | * search path in linker script: File Commands. (line 76) |
| 8387 | * SEARCH_DIR(PATH): File Commands. (line 76) |
| 8388 | * SECT (MRI): MRI. (line 109) |
| 8389 | * section address: Output Section Address. |
| 8390 | (line 6) |
| 8391 | * section address in expression: Builtin Functions. (line 17) |
| 8392 | * section alignment: Builtin Functions. (line 64) |
| 8393 | * section alignment, warnings on: Options. (line 1603) |
| 8394 | * section data: Output Section Data. |
| 8395 | (line 6) |
| 8396 | * section fill pattern: Output Section Fill. |
| 8397 | (line 6) |
| 8398 | * section load address: Output Section LMA. (line 6) |
| 8399 | * section load address in expression: Builtin Functions. (line 143) |
| 8400 | * section name: Output Section Name. |
| 8401 | (line 6) |
| 8402 | * section name wildcard patterns: Input Section Wildcards. |
| 8403 | (line 6) |
| 8404 | * section size: Builtin Functions. (line 175) |
| 8405 | * section, assigning to memory region: Output Section Region. |
| 8406 | (line 6) |
| 8407 | * section, assigning to program header: Output Section Phdr. |
| 8408 | (line 6) |
| 8409 | * SECTIONS: SECTIONS. (line 6) |
| 8410 | * sections, discarding: Output Section Discarding. |
| 8411 | (line 6) |
| 8412 | * sections, orphan: Options. (line 606) |
| 8413 | * segment origins, cmd line: Options. (line 1443) |
| 8414 | * SEGMENT_START(SEGMENT, DEFAULT): Builtin Functions. (line 166) |
| 8415 | * segments, ELF: PHDRS. (line 6) |
| 8416 | * shared libraries: Options. (line 1371) |
| 8417 | * SHORT(EXPRESSION): Output Section Data. |
| 8418 | (line 6) |
| 8419 | * SIZEOF(SECTION): Builtin Functions. (line 175) |
| 8420 | * SIZEOF_HEADERS: Builtin Functions. (line 191) |
| 8421 | * small common symbols: Input Section Common. |
| 8422 | (line 20) |
| 8423 | * SORT: Input Section Wildcards. |
| 8424 | (line 65) |
| 8425 | * SORT_BY_ALIGNMENT: Input Section Wildcards. |
| 8426 | (line 54) |
| 8427 | * SORT_BY_INIT_PRIORITY: Input Section Wildcards. |
| 8428 | (line 60) |
| 8429 | * SORT_BY_NAME: Input Section Wildcards. |
| 8430 | (line 46) |
| 8431 | * SORT_NONE: Input Section Wildcards. |
| 8432 | (line 106) |
| 8433 | * SPU: SPU ELF. (line 29) |
| 8434 | * SPU ELF options: SPU ELF. (line 6) |
| 8435 | * SPU extra overlay stubs: SPU ELF. (line 19) |
| 8436 | * SPU local store size: SPU ELF. (line 24) |
| 8437 | * SPU overlay stub symbols: SPU ELF. (line 15) |
| 8438 | * SPU overlays: SPU ELF. (line 9) |
| 8439 | * SPU plugins: SPU ELF. (line 6) |
| 8440 | * SQUAD(EXPRESSION): Output Section Data. |
| 8441 | (line 6) |
| 8442 | * stack size: Options. (line 2141) |
| 8443 | * standard Unix system: Options. (line 7) |
| 8444 | * start of execution: Entry Point. (line 6) |
| 8445 | * STARTUP(FILENAME): File Commands. (line 84) |
| 8446 | * STM32L4xx erratum workaround: ARM. (line 121) |
| 8447 | * strip all symbols: Options. (line 540) |
| 8448 | * strip debugger symbols: Options. (line 544) |
| 8449 | * stripping all but some symbols: Options. (line 1278) |
| 8450 | * STUB_GROUP_SIZE: ARM. (line 177) |
| 8451 | * SUBALIGN(SUBSECTION_ALIGN): Forced Input Alignment. |
| 8452 | (line 6) |
| 8453 | * suffixes for integers: Constants. (line 15) |
| 8454 | * symbol defaults: Builtin Functions. (line 123) |
| 8455 | * symbol definition, scripts: Assignments. (line 6) |
| 8456 | * symbol names: Symbols. (line 6) |
| 8457 | * symbol tracing: Options. (line 657) |
| 8458 | * symbol versions: VERSION. (line 6) |
| 8459 | * symbol-only input: Options. (line 529) |
| 8460 | * symbolic constants: Symbolic Constants. (line 6) |
| 8461 | * symbols, from command line: Options. (line 1016) |
| 8462 | * symbols, relocatable and absolute: Expression Section. (line 6) |
| 8463 | * symbols, require defined: Options. (line 588) |
| 8464 | * symbols, retaining selectively: Options. (line 1278) |
| 8465 | * synthesizing linker: Options. (line 1252) |
| 8466 | * synthesizing on H8/300: H8/300. (line 14) |
| 8467 | * TARGET(BFDNAME): Format Commands. (line 35) |
| 8468 | * TARGET1: ARM. (line 33) |
| 8469 | * TARGET2: ARM. (line 38) |
| 8470 | * text segment origin, cmd line: Options. (line 1450) |
| 8471 | * thumb entry point: ARM. (line 17) |
| 8472 | * TI COFF versions: TI COFF. (line 6) |
| 8473 | * traditional format: Options. (line 1422) |
| 8474 | * trampoline generation on M68HC11: M68HC11/68HC12. (line 31) |
| 8475 | * trampoline generation on M68HC12: M68HC11/68HC12. (line 31) |
| 8476 | * unallocated address, next: Builtin Functions. (line 157) |
| 8477 | * undefined symbol: Options. (line 575) |
| 8478 | * undefined symbol in linker script: Miscellaneous Commands. |
| 8479 | (line 39) |
| 8480 | * undefined symbols, warnings on: Options. (line 1599) |
| 8481 | * uninitialized data placement: Input Section Common. |
| 8482 | (line 6) |
| 8483 | * unspecified memory: Output Section Data. |
| 8484 | (line 39) |
| 8485 | * usage: Options. (line 1120) |
| 8486 | * USE_BLX: ARM. (line 75) |
| 8487 | * using a DEF file: WIN32. (line 57) |
| 8488 | * using auto-export functionality: WIN32. (line 22) |
| 8489 | * Using decorations: WIN32. (line 162) |
| 8490 | * variables, defining: Assignments. (line 6) |
| 8491 | * verbose[=NUMBER]: Options. (line 1493) |
| 8492 | * version: Options. (line 641) |
| 8493 | * version script: VERSION. (line 6) |
| 8494 | * version script, symbol versions: Options. (line 1501) |
| 8495 | * VERSION {script text}: VERSION. (line 6) |
| 8496 | * versions of symbols: VERSION. (line 6) |
| 8497 | * VFP11_DENORM_FIX: ARM. (line 84) |
| 8498 | * warnings, on combining symbols: Options. (line 1512) |
| 8499 | * warnings, on section alignment: Options. (line 1603) |
| 8500 | * warnings, on undefined symbols: Options. (line 1599) |
| 8501 | * weak externals: WIN32. (line 407) |
| 8502 | * what is this?: Overview. (line 6) |
| 8503 | * wildcard file name patterns: Input Section Wildcards. |
| 8504 | (line 6) |
| 8505 | * Xtensa options: Xtensa. (line 56) |
| 8506 | * Xtensa processors: Xtensa. (line 6) |
| 8507 | |
| 8508 | |
| 8509 | |
| 8510 | Tag Table: |
| 8511 | Node: Top795 |
| 8512 | Node: Overview1580 |
| 8513 | Node: Invocation2694 |
| 8514 | Node: Options3102 |
| 8515 | Node: Environment105699 |
| 8516 | Node: Scripts107459 |
| 8517 | Node: Basic Script Concepts109193 |
| 8518 | Node: Script Format111901 |
| 8519 | Node: Simple Example112764 |
| 8520 | Node: Simple Commands115860 |
| 8521 | Node: Entry Point116366 |
| 8522 | Node: File Commands117299 |
| 8523 | Node: Format Commands121419 |
| 8524 | Node: REGION_ALIAS123375 |
| 8525 | Node: Miscellaneous Commands128207 |
| 8526 | Node: Assignments133746 |
| 8527 | Node: Simple Assignments134257 |
| 8528 | Node: HIDDEN135992 |
| 8529 | Node: PROVIDE136622 |
| 8530 | Node: PROVIDE_HIDDEN137815 |
| 8531 | Node: Source Code Reference138059 |
| 8532 | Node: SECTIONS141975 |
| 8533 | Node: Output Section Description143866 |
| 8534 | Node: Output Section Name145110 |
| 8535 | Node: Output Section Address145986 |
| 8536 | Node: Input Section148221 |
| 8537 | Node: Input Section Basics149022 |
| 8538 | Node: Input Section Wildcards152928 |
| 8539 | Node: Input Section Common158135 |
| 8540 | Node: Input Section Keep159617 |
| 8541 | Node: Input Section Example160107 |
| 8542 | Node: Output Section Data161075 |
| 8543 | Node: Output Section Keywords163852 |
| 8544 | Node: Output Section Discarding167421 |
| 8545 | Node: Output Section Attributes168914 |
| 8546 | Node: Output Section Type170015 |
| 8547 | Node: Output Section LMA171086 |
| 8548 | Node: Forced Output Alignment174157 |
| 8549 | Node: Forced Input Alignment174587 |
| 8550 | Node: Output Section Constraint174976 |
| 8551 | Node: Output Section Region175404 |
| 8552 | Node: Output Section Phdr175837 |
| 8553 | Node: Output Section Fill176501 |
| 8554 | Node: Overlay Description177643 |
| 8555 | Node: MEMORY182089 |
| 8556 | Node: PHDRS186466 |
| 8557 | Node: VERSION191720 |
| 8558 | Node: Expressions199813 |
| 8559 | Node: Constants200742 |
| 8560 | Node: Symbolic Constants201617 |
| 8561 | Node: Symbols202168 |
| 8562 | Node: Orphan Sections202915 |
| 8563 | Node: Location Counter204252 |
| 8564 | Node: Operators208688 |
| 8565 | Node: Evaluation209610 |
| 8566 | Node: Expression Section210974 |
| 8567 | Node: Builtin Functions214838 |
| 8568 | Node: Implicit Linker Scripts223078 |
| 8569 | Node: Machine Dependent223853 |
| 8570 | Node: H8/300224977 |
| 8571 | Node: i960227039 |
| 8572 | Node: M68HC11/68HC12228735 |
| 8573 | Node: ARM230177 |
| 8574 | Node: HPPA ELF32241316 |
| 8575 | Node: M68K242939 |
| 8576 | Node: MIPS243848 |
| 8577 | Node: MMIX244372 |
| 8578 | Node: MSP430245537 |
| 8579 | Node: NDS32246577 |
| 8580 | Node: Nios II247543 |
| 8581 | Node: PowerPC ELF32248859 |
| 8582 | Node: PowerPC64 ELF64251690 |
| 8583 | Node: SPU ELF259022 |
| 8584 | Node: TI COFF261654 |
| 8585 | Node: WIN32262180 |
| 8586 | Node: Xtensa282306 |
| 8587 | Node: BFD285271 |
| 8588 | Node: BFD outline286726 |
| 8589 | Node: BFD information loss288012 |
| 8590 | Node: Canonical format290529 |
| 8591 | Node: Reporting Bugs294886 |
| 8592 | Node: Bug Criteria295580 |
| 8593 | Node: Bug Reporting296279 |
| 8594 | Node: MRI303318 |
| 8595 | Node: GNU Free Documentation License307961 |
| 8596 | Node: LD Index333117 |
| 8597 | |
| 8598 | End Tag Table |