| 1 | \input texinfo |
| 2 | @setfilename ld.info |
| 3 | @syncodeindex ky cp |
| 4 | @include configdoc.texi |
| 5 | @c (configdoc.texi is generated by the Makefile) |
| 6 | @include ldver.texi |
| 7 | |
| 8 | @c @smallbook |
| 9 | |
| 10 | @ifinfo |
| 11 | @format |
| 12 | START-INFO-DIR-ENTRY |
| 13 | * Ld: (ld). The GNU linker. |
| 14 | END-INFO-DIR-ENTRY |
| 15 | @end format |
| 16 | @end ifinfo |
| 17 | |
| 18 | @ifinfo |
| 19 | This file documents the @sc{gnu} linker LD version @value{VERSION}. |
| 20 | |
| 21 | Copyright (C) 1991, 92, 93, 94, 95, 96, 97, 1998 Free Software Foundation, Inc. |
| 22 | |
| 23 | Permission is granted to make and distribute verbatim copies of |
| 24 | this manual provided the copyright notice and this permission notice |
| 25 | are preserved on all copies. |
| 26 | |
| 27 | Permission is granted to copy and distribute modified versions of this |
| 28 | manual under the conditions for verbatim copying, provided also that |
| 29 | the entire resulting derived work is distributed under the terms of a |
| 30 | permission notice identical to this one. |
| 31 | |
| 32 | Permission is granted to copy and distribute translations of this manual |
| 33 | into another language, under the above conditions for modified versions. |
| 34 | |
| 35 | @ignore |
| 36 | Permission is granted to process this file through Tex and print the |
| 37 | results, provided the printed document carries copying permission |
| 38 | notice identical to this one except for the removal of this paragraph |
| 39 | (this paragraph not being relevant to the printed manual). |
| 40 | |
| 41 | @end ignore |
| 42 | @end ifinfo |
| 43 | @iftex |
| 44 | @finalout |
| 45 | @setchapternewpage odd |
| 46 | @settitle Using LD, the GNU linker |
| 47 | @titlepage |
| 48 | @title Using ld |
| 49 | @subtitle The GNU linker |
| 50 | @sp 1 |
| 51 | @subtitle @code{ld} version 2 |
| 52 | @subtitle Version @value{VERSION} |
| 53 | @author Steve Chamberlain |
| 54 | @author Ian Lance Taylor |
| 55 | @author Cygnus Solutions |
| 56 | @page |
| 57 | |
| 58 | @tex |
| 59 | {\parskip=0pt |
| 60 | \hfill Cygnus Solutions\par |
| 61 | \hfill ian\@cygnus.com, doc\@cygnus.com\par |
| 62 | \hfill {\it Using LD, the GNU linker}\par |
| 63 | \hfill Edited by Jeffrey Osier (jeffrey\@cygnus.com)\par |
| 64 | } |
| 65 | \global\parindent=0pt % Steve likes it this way. |
| 66 | @end tex |
| 67 | |
| 68 | @vskip 0pt plus 1filll |
| 69 | Copyright @copyright{} 1991, 92, 93, 94, 95, 96, 97, 1998 Free Software Foundation, Inc. |
| 70 | |
| 71 | Permission is granted to make and distribute verbatim copies of |
| 72 | this manual provided the copyright notice and this permission notice |
| 73 | are preserved on all copies. |
| 74 | |
| 75 | Permission is granted to copy and distribute modified versions of this |
| 76 | manual under the conditions for verbatim copying, provided also that |
| 77 | the entire resulting derived work is distributed under the terms of a |
| 78 | permission notice identical to this one. |
| 79 | |
| 80 | Permission is granted to copy and distribute translations of this manual |
| 81 | into another language, under the above conditions for modified versions. |
| 82 | @end titlepage |
| 83 | @end iftex |
| 84 | @c FIXME: Talk about importance of *order* of args, cmds to linker! |
| 85 | |
| 86 | @ifinfo |
| 87 | @node Top |
| 88 | @top Using ld |
| 89 | This file documents the @sc{gnu} linker ld version @value{VERSION}. |
| 90 | |
| 91 | @menu |
| 92 | * Overview:: Overview |
| 93 | * Invocation:: Invocation |
| 94 | * Scripts:: Linker Scripts |
| 95 | @ifset GENERIC |
| 96 | * Machine Dependent:: Machine Dependent Features |
| 97 | @end ifset |
| 98 | @ifclear GENERIC |
| 99 | @ifset H8300 |
| 100 | * H8/300:: ld and the H8/300 |
| 101 | @end ifset |
| 102 | @ifset Hitachi |
| 103 | * Hitachi:: ld and other Hitachi micros |
| 104 | @end ifset |
| 105 | @ifset I960 |
| 106 | * i960:: ld and the Intel 960 family |
| 107 | @end ifset |
| 108 | @end ifclear |
| 109 | @ifclear SingleFormat |
| 110 | * BFD:: BFD |
| 111 | @end ifclear |
| 112 | @c Following blank line required for remaining bug in makeinfo conds/menus |
| 113 | |
| 114 | * Reporting Bugs:: Reporting Bugs |
| 115 | * MRI:: MRI Compatible Script Files |
| 116 | * Index:: Index |
| 117 | @end menu |
| 118 | @end ifinfo |
| 119 | |
| 120 | @node Overview |
| 121 | @chapter Overview |
| 122 | |
| 123 | @cindex @sc{gnu} linker |
| 124 | @cindex what is this? |
| 125 | @code{ld} combines a number of object and archive files, relocates |
| 126 | their data and ties up symbol references. Usually the last step in |
| 127 | compiling a program is to run @code{ld}. |
| 128 | |
| 129 | @code{ld} accepts Linker Command Language files written in |
| 130 | a superset of AT&T's Link Editor Command Language syntax, |
| 131 | to provide explicit and total control over the linking process. |
| 132 | |
| 133 | @ifclear SingleFormat |
| 134 | This version of @code{ld} uses the general purpose BFD libraries |
| 135 | to operate on object files. This allows @code{ld} to read, combine, and |
| 136 | write object files in many different formats---for example, COFF or |
| 137 | @code{a.out}. Different formats may be linked together to produce any |
| 138 | available kind of object file. @xref{BFD}, for more information. |
| 139 | @end ifclear |
| 140 | |
| 141 | Aside from its flexibility, the @sc{gnu} linker is more helpful than other |
| 142 | linkers in providing diagnostic information. Many linkers abandon |
| 143 | execution immediately upon encountering an error; whenever possible, |
| 144 | @code{ld} continues executing, allowing you to identify other errors |
| 145 | (or, in some cases, to get an output file in spite of the error). |
| 146 | |
| 147 | @node Invocation |
| 148 | @chapter Invocation |
| 149 | |
| 150 | The @sc{gnu} linker @code{ld} is meant to cover a broad range of situations, |
| 151 | and to be as compatible as possible with other linkers. As a result, |
| 152 | you have many choices to control its behavior. |
| 153 | |
| 154 | @ifset UsesEnvVars |
| 155 | @menu |
| 156 | * Options:: Command Line Options |
| 157 | * Environment:: Environment Variables |
| 158 | @end menu |
| 159 | |
| 160 | @node Options |
| 161 | @section Command Line Options |
| 162 | @end ifset |
| 163 | |
| 164 | @cindex command line |
| 165 | @cindex options |
| 166 | The linker supports a plethora of command-line options, but in actual |
| 167 | practice few of them are used in any particular context. |
| 168 | @cindex standard Unix system |
| 169 | For instance, a frequent use of @code{ld} is to link standard Unix |
| 170 | object files on a standard, supported Unix system. On such a system, to |
| 171 | link a file @code{hello.o}: |
| 172 | |
| 173 | @smallexample |
| 174 | ld -o @var{output} /lib/crt0.o hello.o -lc |
| 175 | @end smallexample |
| 176 | |
| 177 | This tells @code{ld} to produce a file called @var{output} as the |
| 178 | result of linking the file @code{/lib/crt0.o} with @code{hello.o} and |
| 179 | the library @code{libc.a}, which will come from the standard search |
| 180 | directories. (See the discussion of the @samp{-l} option below.) |
| 181 | |
| 182 | The command-line options to @code{ld} may be specified in any order, and |
| 183 | may be repeated at will. Repeating most options with a different |
| 184 | argument will either have no further effect, or override prior |
| 185 | occurrences (those further to the left on the command line) of that |
| 186 | option. Options which may be meaningfully specified more than once are |
| 187 | noted in the descriptions below. |
| 188 | |
| 189 | @cindex object files |
| 190 | Non-option arguments are objects files which are to be linked together. |
| 191 | They may follow, precede, or be mixed in with command-line options, |
| 192 | except that an object file argument may not be placed between an option |
| 193 | and its argument. |
| 194 | |
| 195 | Usually the linker is invoked with at least one object file, but you can |
| 196 | specify other forms of binary input files using @samp{-l}, @samp{-R}, |
| 197 | and the script command language. If @emph{no} binary input files at all |
| 198 | are specified, the linker does not produce any output, and issues the |
| 199 | message @samp{No input files}. |
| 200 | |
| 201 | If the linker can not recognize the format of an object file, it will |
| 202 | assume that it is a linker script. A script specified in this way |
| 203 | augments the main linker script used for the link (either the default |
| 204 | linker script or the one specified by using @samp{-T}). This feature |
| 205 | permits the linker to link against a file which appears to be an object |
| 206 | or an archive, but actually merely defines some symbol values, or uses |
| 207 | @code{INPUT} or @code{GROUP} to load other objects. Note that |
| 208 | specifying a script in this way should only be used to augment the main |
| 209 | linker script; if you want to use some command that logically can only |
| 210 | appear once, such as the @code{SECTIONS} or @code{MEMORY} command, you |
| 211 | must replace the default linker script using the @samp{-T} option. |
| 212 | @xref{Scripts}. |
| 213 | |
| 214 | For options whose names are a single letter, |
| 215 | option arguments must either follow the option letter without intervening |
| 216 | whitespace, or be given as separate arguments immediately following the |
| 217 | option that requires them. |
| 218 | |
| 219 | For options whose names are multiple letters, either one dash or two can |
| 220 | precede the option name; for example, @samp{--oformat} and |
| 221 | @samp{--oformat} are equivalent. Arguments to multiple-letter options |
| 222 | must either be separated from the option name by an equals sign, or be |
| 223 | given as separate arguments immediately following the option that |
| 224 | requires them. For example, @samp{--oformat srec} and |
| 225 | @samp{--oformat=srec} are equivalent. Unique abbreviations of the names |
| 226 | of multiple-letter options are accepted. |
| 227 | |
| 228 | @table @code |
| 229 | @kindex -a@var{keyword} |
| 230 | @item -a@var{keyword} |
| 231 | This option is supported for HP/UX compatibility. The @var{keyword} |
| 232 | argument must be one of the strings @samp{archive}, @samp{shared}, or |
| 233 | @samp{default}. @samp{-aarchive} is functionally equivalent to |
| 234 | @samp{-Bstatic}, and the other two keywords are functionally equivalent |
| 235 | to @samp{-Bdynamic}. This option may be used any number of times. |
| 236 | |
| 237 | @ifset I960 |
| 238 | @cindex architectures |
| 239 | @kindex -A@var{arch} |
| 240 | @item -A@var{architecture} |
| 241 | @kindex --architecture=@var{arch} |
| 242 | @itemx --architecture=@var{architecture} |
| 243 | In the current release of @code{ld}, this option is useful only for the |
| 244 | Intel 960 family of architectures. In that @code{ld} configuration, the |
| 245 | @var{architecture} argument identifies the particular architecture in |
| 246 | the 960 family, enabling some safeguards and modifying the |
| 247 | archive-library search path. @xref{i960,,@code{ld} and the Intel 960 |
| 248 | family}, for details. |
| 249 | |
| 250 | Future releases of @code{ld} may support similar functionality for |
| 251 | other architecture families. |
| 252 | @end ifset |
| 253 | |
| 254 | @ifclear SingleFormat |
| 255 | @cindex binary input format |
| 256 | @kindex -b @var{format} |
| 257 | @kindex --format=@var{format} |
| 258 | @cindex input format |
| 259 | @cindex input format |
| 260 | @item -b @var{input-format} |
| 261 | @itemx --format=@var{input-format} |
| 262 | @code{ld} may be configured to support more than one kind of object |
| 263 | file. If your @code{ld} is configured this way, you can use the |
| 264 | @samp{-b} option to specify the binary format for input object files |
| 265 | that follow this option on the command line. Even when @code{ld} is |
| 266 | configured to support alternative object formats, you don't usually need |
| 267 | to specify this, as @code{ld} should be configured to expect as a |
| 268 | default input format the most usual format on each machine. |
| 269 | @var{input-format} is a text string, the name of a particular format |
| 270 | supported by the BFD libraries. (You can list the available binary |
| 271 | formats with @samp{objdump -i}.) |
| 272 | @xref{BFD}. |
| 273 | |
| 274 | You may want to use this option if you are linking files with an unusual |
| 275 | binary format. You can also use @samp{-b} to switch formats explicitly (when |
| 276 | linking object files of different formats), by including |
| 277 | @samp{-b @var{input-format}} before each group of object files in a |
| 278 | particular format. |
| 279 | |
| 280 | The default format is taken from the environment variable |
| 281 | @code{GNUTARGET}. |
| 282 | @ifset UsesEnvVars |
| 283 | @xref{Environment}. |
| 284 | @end ifset |
| 285 | You can also define the input format from a script, using the command |
| 286 | @code{TARGET}; see @ref{Format Commands}. |
| 287 | @end ifclear |
| 288 | |
| 289 | @kindex -c @var{MRI-cmdfile} |
| 290 | @kindex --mri-script=@var{MRI-cmdfile} |
| 291 | @cindex compatibility, MRI |
| 292 | @item -c @var{MRI-commandfile} |
| 293 | @itemx --mri-script=@var{MRI-commandfile} |
| 294 | For compatibility with linkers produced by MRI, @code{ld} accepts script |
| 295 | files written in an alternate, restricted command language, described in |
| 296 | @ref{MRI,,MRI Compatible Script Files}. Introduce MRI script files with |
| 297 | the option @samp{-c}; use the @samp{-T} option to run linker |
| 298 | scripts written in the general-purpose @code{ld} scripting language. |
| 299 | If @var{MRI-cmdfile} does not exist, @code{ld} looks for it in the directories |
| 300 | specified by any @samp{-L} options. |
| 301 | |
| 302 | @cindex common allocation |
| 303 | @kindex -d |
| 304 | @kindex -dc |
| 305 | @kindex -dp |
| 306 | @item -d |
| 307 | @itemx -dc |
| 308 | @itemx -dp |
| 309 | These three options are equivalent; multiple forms are supported for |
| 310 | compatibility with other linkers. They assign space to common symbols |
| 311 | even if a relocatable output file is specified (with @samp{-r}). The |
| 312 | script command @code{FORCE_COMMON_ALLOCATION} has the same effect. |
| 313 | @xref{Miscellaneous Commands}. |
| 314 | |
| 315 | @cindex entry point, from command line |
| 316 | @kindex -e @var{entry} |
| 317 | @kindex --entry=@var{entry} |
| 318 | @item -e @var{entry} |
| 319 | @itemx --entry=@var{entry} |
| 320 | Use @var{entry} as the explicit symbol for beginning execution of your |
| 321 | program, rather than the default entry point. If there is no symbol |
| 322 | named @var{entry}, the linker will try to parse @var{entry} as a number, |
| 323 | and use that as the entry address (the number will be interpreted in |
| 324 | base 10; you may use a leading @samp{0x} for base 16, or a leading |
| 325 | @samp{0} for base 8). @xref{Entry Point}, for a discussion of defaults |
| 326 | and other ways of specifying the entry point. |
| 327 | |
| 328 | @cindex dynamic symbol table |
| 329 | @kindex -E |
| 330 | @kindex --export-dynamic |
| 331 | @item -E |
| 332 | @itemx --export-dynamic |
| 333 | When creating a dynamically linked executable, add all symbols to the |
| 334 | dynamic symbol table. The dynamic symbol table is the set of symbols |
| 335 | which are visible from dynamic objects at run time. |
| 336 | |
| 337 | If you do not use this option, the dynamic symbol table will normally |
| 338 | contain only those symbols which are referenced by some dynamic object |
| 339 | mentioned in the link. |
| 340 | |
| 341 | If you use @code{dlopen} to load a dynamic object which needs to refer |
| 342 | back to the symbols defined by the program, rather than some other |
| 343 | dynamic object, then you will probably need to use this option when |
| 344 | linking the program itself. |
| 345 | |
| 346 | @kindex -f |
| 347 | @kindex --auxiliary |
| 348 | @item -f |
| 349 | @itemx --auxiliary @var{name} |
| 350 | When creating an ELF shared object, set the internal DT_AUXILIARY field |
| 351 | to the specified name. This tells the dynamic linker that the symbol |
| 352 | table of the shared object should be used as an auxiliary filter on the |
| 353 | symbol table of the shared object @var{name}. |
| 354 | |
| 355 | If you later link a program against this filter object, then, when you |
| 356 | run the program, the dynamic linker will see the DT_AUXILIARY field. If |
| 357 | the dynamic linker resolves any symbols from the filter object, it will |
| 358 | first check whether there is a definition in the shared object |
| 359 | @var{name}. If there is one, it will be used instead of the definition |
| 360 | in the filter object. The shared object @var{name} need not exist. |
| 361 | Thus the shared object @var{name} may be used to provide an alternative |
| 362 | implementation of certain functions, perhaps for debugging or for |
| 363 | machine specific performance. |
| 364 | |
| 365 | This option may be specified more than once. The DT_AUXILIARY entries |
| 366 | will be created in the order in which they appear on the command line. |
| 367 | |
| 368 | @kindex -F |
| 369 | @kindex --filter |
| 370 | @item -F @var{name} |
| 371 | @itemx --filter @var{name} |
| 372 | When creating an ELF shared object, set the internal DT_FILTER field to |
| 373 | the specified name. This tells the dynamic linker that the symbol table |
| 374 | of the shared object which is being created should be used as a filter |
| 375 | on the symbol table of the shared object @var{name}. |
| 376 | |
| 377 | If you later link a program against this filter object, then, when you |
| 378 | run the program, the dynamic linker will see the DT_FILTER field. The |
| 379 | dynamic linker will resolve symbols according to the symbol table of the |
| 380 | filter object as usual, but it will actually link to the definitions |
| 381 | found in the shared object @var{name}. Thus the filter object can be |
| 382 | used to select a subset of the symbols provided by the object |
| 383 | @var{name}. |
| 384 | |
| 385 | Some older linkers used the @code{-F} option throughout a compilation |
| 386 | toolchain for specifying object-file format for both input and output |
| 387 | object files. The @sc{gnu} linker uses other mechanisms for this |
| 388 | purpose: the @code{-b}, @code{--format}, @code{--oformat} options, the |
| 389 | @code{TARGET} command in linker scripts, and the @code{GNUTARGET} |
| 390 | environment variable. The @sc{gnu} linker will ignore the @code{-F} |
| 391 | option when not creating an ELF shared object. |
| 392 | |
| 393 | @kindex --force-exe-suffix |
| 394 | @item --force-exe-suffix |
| 395 | Make sure that an output file has a .exe suffix. |
| 396 | |
| 397 | If a successfully built fully linked output file does not have a |
| 398 | @code{.exe} or @code{.dll} suffix, this option forces the linker to copy |
| 399 | the output file to one of the same name with a @code{.exe} suffix. This |
| 400 | option is useful when using unmodified Unix makefiles on a Microsoft |
| 401 | Windows host, since some versions of Windows won't run an image unless |
| 402 | it ends in a @code{.exe} suffix. |
| 403 | |
| 404 | @kindex -g |
| 405 | @item -g |
| 406 | Ignored. Provided for compatibility with other tools. |
| 407 | |
| 408 | @kindex -G |
| 409 | @kindex --gpsize |
| 410 | @cindex object size |
| 411 | @item -G@var{value} |
| 412 | @itemx --gpsize=@var{value} |
| 413 | Set the maximum size of objects to be optimized using the GP register to |
| 414 | @var{size}. This is only meaningful for object file formats such as |
| 415 | MIPS ECOFF which supports putting large and small objects into different |
| 416 | sections. This is ignored for other object file formats. |
| 417 | |
| 418 | @kindex --gc-sections |
| 419 | @cindex garbage collection |
| 420 | @item --gc-sections |
| 421 | Enable garbage collection of unused input sections. It is ignored on |
| 422 | targets that do not support this option. This option is not compatible |
| 423 | with @samp{-r}, nor should it be used with dynamic linking. |
| 424 | |
| 425 | @cindex runtime library name |
| 426 | @kindex -h@var{name} |
| 427 | @kindex -soname=@var{name} |
| 428 | @item -h@var{name} |
| 429 | @itemx -soname=@var{name} |
| 430 | When creating an ELF shared object, set the internal DT_SONAME field to |
| 431 | the specified name. When an executable is linked with a shared object |
| 432 | which has a DT_SONAME field, then when the executable is run the dynamic |
| 433 | linker will attempt to load the shared object specified by the DT_SONAME |
| 434 | field rather than the using the file name given to the linker. |
| 435 | |
| 436 | @kindex -i |
| 437 | @cindex incremental link |
| 438 | @item -i |
| 439 | Perform an incremental link (same as option @samp{-r}). |
| 440 | |
| 441 | @cindex archive files, from cmd line |
| 442 | @kindex -l@var{archive} |
| 443 | @kindex --library=@var{archive} |
| 444 | @item -l@var{archive} |
| 445 | @itemx --library=@var{archive} |
| 446 | Add archive file @var{archive} to the list of files to link. This |
| 447 | option may be used any number of times. @code{ld} will search its |
| 448 | path-list for occurrences of @code{lib@var{archive}.a} for every |
| 449 | @var{archive} specified. |
| 450 | |
| 451 | On systems which support shared libraries, @code{ld} may also search for |
| 452 | libraries with extensions other than @code{.a}. Specifically, on ELF |
| 453 | and SunOS systems, @code{ld} will search a directory for a library with |
| 454 | an extension of @code{.so} before searching for one with an extension of |
| 455 | @code{.a}. By convention, a @code{.so} extension indicates a shared |
| 456 | library. |
| 457 | |
| 458 | The linker will search an archive only once, at the location where it is |
| 459 | specified on the command line. If the archive defines a symbol which |
| 460 | was undefined in some object which appeared before the archive on the |
| 461 | command line, the linker will include the appropriate file(s) from the |
| 462 | archive. However, an undefined symbol in an object appearing later on |
| 463 | the command line will not cause the linker to search the archive again. |
| 464 | |
| 465 | See the @code{-(} option for a way to force the linker to search |
| 466 | archives multiple times. |
| 467 | |
| 468 | You may list the same archive multiple times on the command line. |
| 469 | |
| 470 | @ifset GENERIC |
| 471 | This type of archive searching is standard for Unix linkers. However, |
| 472 | if you are using @code{ld} on AIX, note that it is different from the |
| 473 | behaviour of the AIX linker. |
| 474 | @end ifset |
| 475 | |
| 476 | @cindex search directory, from cmd line |
| 477 | @kindex -L@var{dir} |
| 478 | @kindex --library-path=@var{dir} |
| 479 | @item -L@var{searchdir} |
| 480 | @itemx --library-path=@var{searchdir} |
| 481 | Add path @var{searchdir} to the list of paths that @code{ld} will search |
| 482 | for archive libraries and @code{ld} control scripts. You may use this |
| 483 | option any number of times. The directories are searched in the order |
| 484 | in which they are specified on the command line. Directories specified |
| 485 | on the command line are searched before the default directories. All |
| 486 | @code{-L} options apply to all @code{-l} options, regardless of the |
| 487 | order in which the options appear. |
| 488 | |
| 489 | @ifset UsesEnvVars |
| 490 | The default set of paths searched (without being specified with |
| 491 | @samp{-L}) depends on which emulation mode @code{ld} is using, and in |
| 492 | some cases also on how it was configured. @xref{Environment}. |
| 493 | @end ifset |
| 494 | |
| 495 | The paths can also be specified in a link script with the |
| 496 | @code{SEARCH_DIR} command. Directories specified this way are searched |
| 497 | at the point in which the linker script appears in the command line. |
| 498 | |
| 499 | @cindex emulation |
| 500 | @kindex -m @var{emulation} |
| 501 | @item -m@var{emulation} |
| 502 | Emulate the @var{emulation} linker. You can list the available |
| 503 | emulations with the @samp{--verbose} or @samp{-V} options. |
| 504 | |
| 505 | If the @samp{-m} option is not used, the emulation is taken from the |
| 506 | @code{LDEMULATION} environment variable, if that is defined. |
| 507 | |
| 508 | Otherwise, the default emulation depends upon how the linker was |
| 509 | configured. |
| 510 | |
| 511 | @cindex link map |
| 512 | @kindex -M |
| 513 | @kindex --print-map |
| 514 | @item -M |
| 515 | @itemx --print-map |
| 516 | Print a link map to the standard output. A link map provides |
| 517 | information about the link, including the following: |
| 518 | |
| 519 | @itemize @bullet |
| 520 | @item |
| 521 | Where object files and symbols are mapped into memory. |
| 522 | @item |
| 523 | How common symbols are allocated. |
| 524 | @item |
| 525 | All archive members included in the link, with a mention of the symbol |
| 526 | which caused the archive member to be brought in. |
| 527 | @end itemize |
| 528 | |
| 529 | @kindex -n |
| 530 | @cindex read-only text |
| 531 | @cindex NMAGIC |
| 532 | @kindex --nmagic |
| 533 | @item -n |
| 534 | @itemx --nmagic |
| 535 | Set the text segment to be read only, and mark the output as |
| 536 | @code{NMAGIC} if possible. |
| 537 | |
| 538 | @kindex -N |
| 539 | @kindex --omagic |
| 540 | @cindex read/write from cmd line |
| 541 | @cindex OMAGIC |
| 542 | @item -N |
| 543 | @itemx --omagic |
| 544 | Set the text and data sections to be readable and writable. Also, do |
| 545 | not page-align the data segment. If the output format supports Unix |
| 546 | style magic numbers, mark the output as @code{OMAGIC}. |
| 547 | |
| 548 | @kindex -o @var{output} |
| 549 | @kindex --output=@var{output} |
| 550 | @cindex naming the output file |
| 551 | @item -o @var{output} |
| 552 | @itemx --output=@var{output} |
| 553 | Use @var{output} as the name for the program produced by @code{ld}; if this |
| 554 | option is not specified, the name @file{a.out} is used by default. The |
| 555 | script command @code{OUTPUT} can also specify the output file name. |
| 556 | |
| 557 | @kindex -O @var{level} |
| 558 | @cindex generating optimized output |
| 559 | @item -O @var{level} |
| 560 | If @var{level} is a numeric values greater than zero @code{ld} optimizes |
| 561 | the output. This might take significantly longer and therefore probably |
| 562 | should only be enabled for the final binary. |
| 563 | |
| 564 | @cindex partial link |
| 565 | @cindex relocatable output |
| 566 | @kindex -r |
| 567 | @kindex --relocateable |
| 568 | @item -r |
| 569 | @itemx --relocateable |
| 570 | Generate relocatable output---i.e., generate an output file that can in |
| 571 | turn serve as input to @code{ld}. This is often called @dfn{partial |
| 572 | linking}. As a side effect, in environments that support standard Unix |
| 573 | magic numbers, this option also sets the output file's magic number to |
| 574 | @code{OMAGIC}. |
| 575 | @c ; see @code{-N}. |
| 576 | If this option is not specified, an absolute file is produced. When |
| 577 | linking C++ programs, this option @emph{will not} resolve references to |
| 578 | constructors; to do that, use @samp{-Ur}. |
| 579 | |
| 580 | This option does the same thing as @samp{-i}. |
| 581 | |
| 582 | @kindex -R @var{file} |
| 583 | @kindex --just-symbols=@var{file} |
| 584 | @cindex symbol-only input |
| 585 | @item -R @var{filename} |
| 586 | @itemx --just-symbols=@var{filename} |
| 587 | Read symbol names and their addresses from @var{filename}, but do not |
| 588 | relocate it or include it in the output. This allows your output file |
| 589 | to refer symbolically to absolute locations of memory defined in other |
| 590 | programs. You may use this option more than once. |
| 591 | |
| 592 | For compatibility with other ELF linkers, if the @code{-R} option is |
| 593 | followed by a directory name, rather than a file name, it is treated as |
| 594 | the @code{-rpath} option. |
| 595 | |
| 596 | @kindex -s |
| 597 | @kindex --strip-all |
| 598 | @cindex strip all symbols |
| 599 | @item -s |
| 600 | @itemx --strip-all |
| 601 | Omit all symbol information from the output file. |
| 602 | |
| 603 | @kindex -S |
| 604 | @kindex --strip-debug |
| 605 | @cindex strip debugger symbols |
| 606 | @item -S |
| 607 | @itemx --strip-debug |
| 608 | Omit debugger symbol information (but not all symbols) from the output file. |
| 609 | |
| 610 | @kindex -t |
| 611 | @kindex --trace |
| 612 | @cindex input files, displaying |
| 613 | @item -t |
| 614 | @itemx --trace |
| 615 | Print the names of the input files as @code{ld} processes them. |
| 616 | |
| 617 | @kindex -T @var{script} |
| 618 | @kindex --script=@var{script} |
| 619 | @cindex script files |
| 620 | @item -T @var{scriptfile} |
| 621 | @itemx --script=@var{scriptfile} |
| 622 | Use @var{scriptfile} as the linker script. This script replaces |
| 623 | @code{ld}'s default linker script (rather than adding to it), so |
| 624 | @var{commandfile} must specify everything necessary to describe the |
| 625 | output file. You must use this option if you want to use a command |
| 626 | which can only appear once in a linker script, such as the |
| 627 | @code{SECTIONS} or @code{MEMORY} command. @xref{Scripts}. If |
| 628 | @var{scriptfile} does not exist in the current directory, @code{ld} |
| 629 | looks for it in the directories specified by any preceding @samp{-L} |
| 630 | options. Multiple @samp{-T} options accumulate. |
| 631 | |
| 632 | @kindex -u @var{symbol} |
| 633 | @kindex --undefined=@var{symbol} |
| 634 | @cindex undefined symbol |
| 635 | @item -u @var{symbol} |
| 636 | @itemx --undefined=@var{symbol} |
| 637 | Force @var{symbol} to be entered in the output file as an undefined |
| 638 | symbol. Doing this may, for example, trigger linking of additional |
| 639 | modules from standard libraries. @samp{-u} may be repeated with |
| 640 | different option arguments to enter additional undefined symbols. This |
| 641 | option is equivalent to the @code{EXTERN} linker script command. |
| 642 | |
| 643 | @kindex -v |
| 644 | @kindex -V |
| 645 | @kindex --version |
| 646 | @cindex version |
| 647 | @item -v |
| 648 | @itemx --version |
| 649 | @itemx -V |
| 650 | Display the version number for @code{ld}. The @code{-V} option also |
| 651 | lists the supported emulations. |
| 652 | |
| 653 | @kindex -x |
| 654 | @kindex --discard-all |
| 655 | @cindex deleting local symbols |
| 656 | @item -x |
| 657 | @itemx --discard-all |
| 658 | Delete all local symbols. |
| 659 | |
| 660 | @kindex -X |
| 661 | @kindex --discard-locals |
| 662 | @cindex local symbols, deleting |
| 663 | @cindex L, deleting symbols beginning |
| 664 | @item -X |
| 665 | @itemx --discard-locals |
| 666 | Delete all temporary local symbols. For most targets, this is all local |
| 667 | symbols whose names begin with @samp{L}. |
| 668 | |
| 669 | @kindex -y @var{symbol} |
| 670 | @kindex --trace-symbol=@var{symbol} |
| 671 | @cindex symbol tracing |
| 672 | @item -y @var{symbol} |
| 673 | @itemx --trace-symbol=@var{symbol} |
| 674 | Print the name of each linked file in which @var{symbol} appears. This |
| 675 | option may be given any number of times. On many systems it is necessary |
| 676 | to prepend an underscore. |
| 677 | |
| 678 | This option is useful when you have an undefined symbol in your link but |
| 679 | don't know where the reference is coming from. |
| 680 | |
| 681 | @kindex -Y @var{path} |
| 682 | @item -Y @var{path} |
| 683 | Add @var{path} to the default library search path. This option exists |
| 684 | for Solaris compatibility. |
| 685 | |
| 686 | @kindex -z @var{keyword} |
| 687 | @item -z @var{keyword} |
| 688 | This option is ignored for Solaris compatibility. |
| 689 | |
| 690 | @kindex -( |
| 691 | @cindex groups of archives |
| 692 | @item -( @var{archives} -) |
| 693 | @itemx --start-group @var{archives} --end-group |
| 694 | The @var{archives} should be a list of archive files. They may be |
| 695 | either explicit file names, or @samp{-l} options. |
| 696 | |
| 697 | The specified archives are searched repeatedly until no new undefined |
| 698 | references are created. Normally, an archive is searched only once in |
| 699 | the order that it is specified on the command line. If a symbol in that |
| 700 | archive is needed to resolve an undefined symbol referred to by an |
| 701 | object in an archive that appears later on the command line, the linker |
| 702 | would not be able to resolve that reference. By grouping the archives, |
| 703 | they all be searched repeatedly until all possible references are |
| 704 | resolved. |
| 705 | |
| 706 | Using this option has a significant performance cost. It is best to use |
| 707 | it only when there are unavoidable circular references between two or |
| 708 | more archives. |
| 709 | |
| 710 | @kindex -assert @var{keyword} |
| 711 | @item -assert @var{keyword} |
| 712 | This option is ignored for SunOS compatibility. |
| 713 | |
| 714 | @kindex -Bdynamic |
| 715 | @kindex -dy |
| 716 | @kindex -call_shared |
| 717 | @item -Bdynamic |
| 718 | @itemx -dy |
| 719 | @itemx -call_shared |
| 720 | Link against dynamic libraries. This is only meaningful on platforms |
| 721 | for which shared libraries are supported. This option is normally the |
| 722 | default on such platforms. The different variants of this option are |
| 723 | for compatibility with various systems. You may use this option |
| 724 | multiple times on the command line: it affects library searching for |
| 725 | @code{-l} options which follow it. |
| 726 | |
| 727 | @kindex -Bstatic |
| 728 | @kindex -dn |
| 729 | @kindex -non_shared |
| 730 | @kindex -static |
| 731 | @item -Bstatic |
| 732 | @itemx -dn |
| 733 | @itemx -non_shared |
| 734 | @itemx -static |
| 735 | Do not link against shared libraries. This is only meaningful on |
| 736 | platforms for which shared libraries are supported. The different |
| 737 | variants of this option are for compatibility with various systems. You |
| 738 | may use this option multiple times on the command line: it affects |
| 739 | library searching for @code{-l} options which follow it. |
| 740 | |
| 741 | @kindex -Bsymbolic |
| 742 | @item -Bsymbolic |
| 743 | When creating a shared library, bind references to global symbols to the |
| 744 | definition within the shared library, if any. Normally, it is possible |
| 745 | for a program linked against a shared library to override the definition |
| 746 | within the shared library. This option is only meaningful on ELF |
| 747 | platforms which support shared libraries. |
| 748 | |
| 749 | @cindex cross reference table |
| 750 | @kindex --cref |
| 751 | @item --cref |
| 752 | Output a cross reference table. If a linker map file is being |
| 753 | generated, the cross reference table is printed to the map file. |
| 754 | Otherwise, it is printed on the standard output. |
| 755 | |
| 756 | The format of the table is intentionally simple, so that it may be |
| 757 | easily processed by a script if necessary. The symbols are printed out, |
| 758 | sorted by name. For each symbol, a list of file names is given. If the |
| 759 | symbol is defined, the first file listed is the location of the |
| 760 | definition. The remaining files contain references to the symbol. |
| 761 | |
| 762 | @cindex symbols, from command line |
| 763 | @kindex --defsym @var{symbol}=@var{exp} |
| 764 | @item --defsym @var{symbol}=@var{expression} |
| 765 | Create a global symbol in the output file, containing the absolute |
| 766 | address given by @var{expression}. You may use this option as many |
| 767 | times as necessary to define multiple symbols in the command line. A |
| 768 | limited form of arithmetic is supported for the @var{expression} in this |
| 769 | context: you may give a hexadecimal constant or the name of an existing |
| 770 | symbol, or use @code{+} and @code{-} to add or subtract hexadecimal |
| 771 | constants or symbols. If you need more elaborate expressions, consider |
| 772 | using the linker command language from a script (@pxref{Assignments,, |
| 773 | Assignment: Symbol Definitions}). @emph{Note:} there should be no white |
| 774 | space between @var{symbol}, the equals sign (``@key{=}''), and |
| 775 | @var{expression}. |
| 776 | |
| 777 | @cindex dynamic linker, from command line |
| 778 | @kindex --dynamic-linker @var{file} |
| 779 | @item --dynamic-linker @var{file} |
| 780 | Set the name of the dynamic linker. This is only meaningful when |
| 781 | generating dynamically linked ELF executables. The default dynamic |
| 782 | linker is normally correct; don't use this unless you know what you are |
| 783 | doing. |
| 784 | |
| 785 | @cindex big-endian objects |
| 786 | @cindex endianness |
| 787 | @kindex -EB |
| 788 | @item -EB |
| 789 | Link big-endian objects. This affects the default output format. |
| 790 | |
| 791 | @cindex little-endian objects |
| 792 | @kindex -EL |
| 793 | @item -EL |
| 794 | Link little-endian objects. This affects the default output format. |
| 795 | |
| 796 | @cindex MIPS embedded PIC code |
| 797 | @kindex --embedded-relocs |
| 798 | @item --embedded-relocs |
| 799 | This option is only meaningful when linking MIPS embedded PIC code, |
| 800 | generated by the -membedded-pic option to the @sc{gnu} compiler and |
| 801 | assembler. It causes the linker to create a table which may be used at |
| 802 | runtime to relocate any data which was statically initialized to pointer |
| 803 | values. See the code in testsuite/ld-empic for details. |
| 804 | |
| 805 | @cindex help |
| 806 | @cindex usage |
| 807 | @kindex --help |
| 808 | @item --help |
| 809 | Print a summary of the command-line options on the standard output and exit. |
| 810 | |
| 811 | @kindex -Map |
| 812 | @item -Map @var{mapfile} |
| 813 | Print a link map to the file @var{mapfile}. See the description of the |
| 814 | @samp{-M} option, above. |
| 815 | |
| 816 | @cindex memory usage |
| 817 | @kindex --no-keep-memory |
| 818 | @item --no-keep-memory |
| 819 | @code{ld} normally optimizes for speed over memory usage by caching the |
| 820 | symbol tables of input files in memory. This option tells @code{ld} to |
| 821 | instead optimize for memory usage, by rereading the symbol tables as |
| 822 | necessary. This may be required if @code{ld} runs out of memory space |
| 823 | while linking a large executable. |
| 824 | |
| 825 | @kindex --no-warn-mismatch |
| 826 | @item --no-warn-mismatch |
| 827 | Normally @code{ld} will give an error if you try to link together input |
| 828 | files that are mismatched for some reason, perhaps because they have |
| 829 | been compiled for different processors or for different endiannesses. |
| 830 | This option tells @code{ld} that it should silently permit such possible |
| 831 | errors. This option should only be used with care, in cases when you |
| 832 | have taken some special action that ensures that the linker errors are |
| 833 | inappropriate. |
| 834 | |
| 835 | @kindex --no-whole-archive |
| 836 | @item --no-whole-archive |
| 837 | Turn off the effect of the @code{--whole-archive} option for subsequent |
| 838 | archive files. |
| 839 | |
| 840 | @cindex output file after errors |
| 841 | @kindex --noinhibit-exec |
| 842 | @item --noinhibit-exec |
| 843 | Retain the executable output file whenever it is still usable. |
| 844 | Normally, the linker will not produce an output file if it encounters |
| 845 | errors during the link process; it exits without writing an output file |
| 846 | when it issues any error whatsoever. |
| 847 | |
| 848 | @ifclear SingleFormat |
| 849 | @kindex --oformat |
| 850 | @item --oformat @var{output-format} |
| 851 | @code{ld} may be configured to support more than one kind of object |
| 852 | file. If your @code{ld} is configured this way, you can use the |
| 853 | @samp{--oformat} option to specify the binary format for the output |
| 854 | object file. Even when @code{ld} is configured to support alternative |
| 855 | object formats, you don't usually need to specify this, as @code{ld} |
| 856 | should be configured to produce as a default output format the most |
| 857 | usual format on each machine. @var{output-format} is a text string, the |
| 858 | name of a particular format supported by the BFD libraries. (You can |
| 859 | list the available binary formats with @samp{objdump -i}.) The script |
| 860 | command @code{OUTPUT_FORMAT} can also specify the output format, but |
| 861 | this option overrides it. @xref{BFD}. |
| 862 | @end ifclear |
| 863 | |
| 864 | @kindex -qmagic |
| 865 | @item -qmagic |
| 866 | This option is ignored for Linux compatibility. |
| 867 | |
| 868 | @kindex -Qy |
| 869 | @item -Qy |
| 870 | This option is ignored for SVR4 compatibility. |
| 871 | |
| 872 | @kindex --relax |
| 873 | @cindex synthesizing linker |
| 874 | @cindex relaxing addressing modes |
| 875 | @item --relax |
| 876 | An option with machine dependent effects. |
| 877 | @ifset GENERIC |
| 878 | This option is only supported on a few targets. |
| 879 | @end ifset |
| 880 | @ifset H8300 |
| 881 | @xref{H8/300,,@code{ld} and the H8/300}. |
| 882 | @end ifset |
| 883 | @ifset I960 |
| 884 | @xref{i960,, @code{ld} and the Intel 960 family}. |
| 885 | @end ifset |
| 886 | |
| 887 | |
| 888 | On some platforms, the @samp{--relax} option performs global |
| 889 | optimizations that become possible when the linker resolves addressing |
| 890 | in the program, such as relaxing address modes and synthesizing new |
| 891 | instructions in the output object file. |
| 892 | |
| 893 | On some platforms these link time global optimizations may make symbolic |
| 894 | debugging of the resulting executable impossible. |
| 895 | @ifset GENERIC |
| 896 | This is known to be |
| 897 | the case for the Matsushita MN10200 and MN10300 family of processors. |
| 898 | @end ifset |
| 899 | |
| 900 | @ifset GENERIC |
| 901 | On platforms where this is not supported, @samp{--relax} is accepted, |
| 902 | but ignored. |
| 903 | @end ifset |
| 904 | |
| 905 | @cindex retaining specified symbols |
| 906 | @cindex stripping all but some symbols |
| 907 | @cindex symbols, retaining selectively |
| 908 | @item --retain-symbols-file @var{filename} |
| 909 | Retain @emph{only} the symbols listed in the file @var{filename}, |
| 910 | discarding all others. @var{filename} is simply a flat file, with one |
| 911 | symbol name per line. This option is especially useful in environments |
| 912 | @ifset GENERIC |
| 913 | (such as VxWorks) |
| 914 | @end ifset |
| 915 | where a large global symbol table is accumulated gradually, to conserve |
| 916 | run-time memory. |
| 917 | |
| 918 | @samp{--retain-symbols-file} does @emph{not} discard undefined symbols, |
| 919 | or symbols needed for relocations. |
| 920 | |
| 921 | You may only specify @samp{--retain-symbols-file} once in the command |
| 922 | line. It overrides @samp{-s} and @samp{-S}. |
| 923 | |
| 924 | @ifset GENERIC |
| 925 | @item -rpath @var{dir} |
| 926 | @cindex runtime library search path |
| 927 | @kindex -rpath |
| 928 | Add a directory to the runtime library search path. This is used when |
| 929 | linking an ELF executable with shared objects. All @code{-rpath} |
| 930 | arguments are concatenated and passed to the runtime linker, which uses |
| 931 | them to locate shared objects at runtime. The @code{-rpath} option is |
| 932 | also used when locating shared objects which are needed by shared |
| 933 | objects explicitly included in the link; see the description of the |
| 934 | @code{-rpath-link} option. If @code{-rpath} is not used when linking an |
| 935 | ELF executable, the contents of the environment variable |
| 936 | @code{LD_RUN_PATH} will be used if it is defined. |
| 937 | |
| 938 | The @code{-rpath} option may also be used on SunOS. By default, on |
| 939 | SunOS, the linker will form a runtime search patch out of all the |
| 940 | @code{-L} options it is given. If a @code{-rpath} option is used, the |
| 941 | runtime search path will be formed exclusively using the @code{-rpath} |
| 942 | options, ignoring the @code{-L} options. This can be useful when using |
| 943 | gcc, which adds many @code{-L} options which may be on NFS mounted |
| 944 | filesystems. |
| 945 | |
| 946 | For compatibility with other ELF linkers, if the @code{-R} option is |
| 947 | followed by a directory name, rather than a file name, it is treated as |
| 948 | the @code{-rpath} option. |
| 949 | @end ifset |
| 950 | |
| 951 | @ifset GENERIC |
| 952 | @cindex link-time runtime library search path |
| 953 | @kindex -rpath-link |
| 954 | @item -rpath-link @var{DIR} |
| 955 | When using ELF or SunOS, one shared library may require another. This |
| 956 | happens when an @code{ld -shared} link includes a shared library as one |
| 957 | of the input files. |
| 958 | |
| 959 | When the linker encounters such a dependency when doing a non-shared, |
| 960 | non-relocatable link, it will automatically try to locate the required |
| 961 | shared library and include it in the link, if it is not included |
| 962 | explicitly. In such a case, the @code{-rpath-link} option |
| 963 | specifies the first set of directories to search. The |
| 964 | @code{-rpath-link} option may specify a sequence of directory names |
| 965 | either by specifying a list of names separated by colons, or by |
| 966 | appearing multiple times. |
| 967 | |
| 968 | The linker uses the following search paths to locate required shared |
| 969 | libraries. |
| 970 | @enumerate |
| 971 | @item |
| 972 | Any directories specified by @code{-rpath-link} options. |
| 973 | @item |
| 974 | Any directories specified by @code{-rpath} options. The difference |
| 975 | between @code{-rpath} and @code{-rpath-link} is that directories |
| 976 | specified by @code{-rpath} options are included in the executable and |
| 977 | used at runtime, whereas the @code{-rpath-link} option is only effective |
| 978 | at link time. |
| 979 | @item |
| 980 | On an ELF system, if the @code{-rpath} and @code{rpath-link} options |
| 981 | were not used, search the contents of the environment variable |
| 982 | @code{LD_RUN_PATH}. |
| 983 | @item |
| 984 | On SunOS, if the @code{-rpath} option was not used, search any |
| 985 | directories specified using @code{-L} options. |
| 986 | @item |
| 987 | For a native linker, the contents of the environment variable |
| 988 | @code{LD_LIBRARY_PATH}. |
| 989 | @item |
| 990 | The default directories, normally @file{/lib} and @file{/usr/lib}. |
| 991 | @item |
| 992 | For a native linker on an ELF system, if the file @file{/etc/ld.so.conf} |
| 993 | exists, the list of directories found in that file. |
| 994 | @end enumerate |
| 995 | |
| 996 | If the required shared library is not found, the linker will issue a |
| 997 | warning and continue with the link. |
| 998 | @end ifset |
| 999 | |
| 1000 | @kindex -shared |
| 1001 | @kindex -Bshareable |
| 1002 | @item -shared |
| 1003 | @itemx -Bshareable |
| 1004 | @cindex shared libraries |
| 1005 | Create a shared library. This is currently only supported on ELF, XCOFF |
| 1006 | and SunOS platforms. On SunOS, the linker will automatically create a |
| 1007 | shared library if the @code{-e} option is not used and there are |
| 1008 | undefined symbols in the link. |
| 1009 | |
| 1010 | @item --sort-common |
| 1011 | @kindex --sort-common |
| 1012 | This option tells @code{ld} to sort the common symbols by size when it |
| 1013 | places them in the appropriate output sections. First come all the one |
| 1014 | byte symbols, then all the two bytes, then all the four bytes, and then |
| 1015 | everything else. This is to prevent gaps between symbols due to |
| 1016 | alignment constraints. |
| 1017 | |
| 1018 | @kindex --split-by-file |
| 1019 | @item --split-by-file |
| 1020 | Similar to @code{--split-by-reloc} but creates a new output section for |
| 1021 | each input file. |
| 1022 | |
| 1023 | @kindex --split-by-reloc |
| 1024 | @item --split-by-reloc @var{count} |
| 1025 | Trys to creates extra sections in the output file so that no single |
| 1026 | output section in the file contains more than @var{count} relocations. |
| 1027 | This is useful when generating huge relocatable for downloading into |
| 1028 | certain real time kernels with the COFF object file format; since COFF |
| 1029 | cannot represent more than 65535 relocations in a single section. Note |
| 1030 | that this will fail to work with object file formats which do not |
| 1031 | support arbitrary sections. The linker will not split up individual |
| 1032 | input sections for redistribution, so if a single input section contains |
| 1033 | more than @var{count} relocations one output section will contain that |
| 1034 | many relocations. |
| 1035 | |
| 1036 | @kindex --stats |
| 1037 | @item --stats |
| 1038 | Compute and display statistics about the operation of the linker, such |
| 1039 | as execution time and memory usage. |
| 1040 | |
| 1041 | @kindex --traditional-format |
| 1042 | @cindex traditional format |
| 1043 | @item --traditional-format |
| 1044 | For some targets, the output of @code{ld} is different in some ways from |
| 1045 | the output of some existing linker. This switch requests @code{ld} to |
| 1046 | use the traditional format instead. |
| 1047 | |
| 1048 | @cindex dbx |
| 1049 | For example, on SunOS, @code{ld} combines duplicate entries in the |
| 1050 | symbol string table. This can reduce the size of an output file with |
| 1051 | full debugging information by over 30 percent. Unfortunately, the SunOS |
| 1052 | @code{dbx} program can not read the resulting program (@code{gdb} has no |
| 1053 | trouble). The @samp{--traditional-format} switch tells @code{ld} to not |
| 1054 | combine duplicate entries. |
| 1055 | |
| 1056 | @kindex -Tbss @var{org} |
| 1057 | @kindex -Tdata @var{org} |
| 1058 | @kindex -Ttext @var{org} |
| 1059 | @cindex segment origins, cmd line |
| 1060 | @item -Tbss @var{org} |
| 1061 | @itemx -Tdata @var{org} |
| 1062 | @itemx -Ttext @var{org} |
| 1063 | Use @var{org} as the starting address for---respectively---the |
| 1064 | @code{bss}, @code{data}, or the @code{text} segment of the output file. |
| 1065 | @var{org} must be a single hexadecimal integer; |
| 1066 | for compatibility with other linkers, you may omit the leading |
| 1067 | @samp{0x} usually associated with hexadecimal values. |
| 1068 | |
| 1069 | @kindex -Ur |
| 1070 | @cindex constructors |
| 1071 | @item -Ur |
| 1072 | For anything other than C++ programs, this option is equivalent to |
| 1073 | @samp{-r}: it generates relocatable output---i.e., an output file that can in |
| 1074 | turn serve as input to @code{ld}. When linking C++ programs, @samp{-Ur} |
| 1075 | @emph{does} resolve references to constructors, unlike @samp{-r}. |
| 1076 | It does not work to use @samp{-Ur} on files that were themselves linked |
| 1077 | with @samp{-Ur}; once the constructor table has been built, it cannot |
| 1078 | be added to. Use @samp{-Ur} only for the last partial link, and |
| 1079 | @samp{-r} for the others. |
| 1080 | |
| 1081 | @kindex --verbose |
| 1082 | @cindex verbose |
| 1083 | @item --verbose |
| 1084 | Display the version number for @code{ld} and list the linker emulations |
| 1085 | supported. Display which input files can and cannot be opened. Display |
| 1086 | the linker script if using a default builtin script. |
| 1087 | |
| 1088 | @kindex --version-script=@var{version-scriptfile} |
| 1089 | @cindex version script, symbol versions |
| 1090 | @itemx --version-script=@var{version-scriptfile} |
| 1091 | Specify the name of a version script to the linker. This is typically |
| 1092 | used when creating shared libraries to specify additional information |
| 1093 | about the version heirarchy for the library being created. This option |
| 1094 | is only meaningful on ELF platforms which support shared libraries. |
| 1095 | @xref{VERSION}. |
| 1096 | |
| 1097 | @kindex --warn-comon |
| 1098 | @cindex warnings, on combining symbols |
| 1099 | @cindex combining symbols, warnings on |
| 1100 | @item --warn-common |
| 1101 | Warn when a common symbol is combined with another common symbol or with |
| 1102 | a symbol definition. Unix linkers allow this somewhat sloppy practice, |
| 1103 | but linkers on some other operating systems do not. This option allows |
| 1104 | you to find potential problems from combining global symbols. |
| 1105 | Unfortunately, some C libraries use this practice, so you may get some |
| 1106 | warnings about symbols in the libraries as well as in your programs. |
| 1107 | |
| 1108 | There are three kinds of global symbols, illustrated here by C examples: |
| 1109 | |
| 1110 | @table @samp |
| 1111 | @item int i = 1; |
| 1112 | A definition, which goes in the initialized data section of the output |
| 1113 | file. |
| 1114 | |
| 1115 | @item extern int i; |
| 1116 | An undefined reference, which does not allocate space. |
| 1117 | There must be either a definition or a common symbol for the |
| 1118 | variable somewhere. |
| 1119 | |
| 1120 | @item int i; |
| 1121 | A common symbol. If there are only (one or more) common symbols for a |
| 1122 | variable, it goes in the uninitialized data area of the output file. |
| 1123 | The linker merges multiple common symbols for the same variable into a |
| 1124 | single symbol. If they are of different sizes, it picks the largest |
| 1125 | size. The linker turns a common symbol into a declaration, if there is |
| 1126 | a definition of the same variable. |
| 1127 | @end table |
| 1128 | |
| 1129 | The @samp{--warn-common} option can produce five kinds of warnings. |
| 1130 | Each warning consists of a pair of lines: the first describes the symbol |
| 1131 | just encountered, and the second describes the previous symbol |
| 1132 | encountered with the same name. One or both of the two symbols will be |
| 1133 | a common symbol. |
| 1134 | |
| 1135 | @enumerate |
| 1136 | @item |
| 1137 | Turning a common symbol into a reference, because there is already a |
| 1138 | definition for the symbol. |
| 1139 | @smallexample |
| 1140 | @var{file}(@var{section}): warning: common of `@var{symbol}' |
| 1141 | overridden by definition |
| 1142 | @var{file}(@var{section}): warning: defined here |
| 1143 | @end smallexample |
| 1144 | |
| 1145 | @item |
| 1146 | Turning a common symbol into a reference, because a later definition for |
| 1147 | the symbol is encountered. This is the same as the previous case, |
| 1148 | except that the symbols are encountered in a different order. |
| 1149 | @smallexample |
| 1150 | @var{file}(@var{section}): warning: definition of `@var{symbol}' |
| 1151 | overriding common |
| 1152 | @var{file}(@var{section}): warning: common is here |
| 1153 | @end smallexample |
| 1154 | |
| 1155 | @item |
| 1156 | Merging a common symbol with a previous same-sized common symbol. |
| 1157 | @smallexample |
| 1158 | @var{file}(@var{section}): warning: multiple common |
| 1159 | of `@var{symbol}' |
| 1160 | @var{file}(@var{section}): warning: previous common is here |
| 1161 | @end smallexample |
| 1162 | |
| 1163 | @item |
| 1164 | Merging a common symbol with a previous larger common symbol. |
| 1165 | @smallexample |
| 1166 | @var{file}(@var{section}): warning: common of `@var{symbol}' |
| 1167 | overridden by larger common |
| 1168 | @var{file}(@var{section}): warning: larger common is here |
| 1169 | @end smallexample |
| 1170 | |
| 1171 | @item |
| 1172 | Merging a common symbol with a previous smaller common symbol. This is |
| 1173 | the same as the previous case, except that the symbols are |
| 1174 | encountered in a different order. |
| 1175 | @smallexample |
| 1176 | @var{file}(@var{section}): warning: common of `@var{symbol}' |
| 1177 | overriding smaller common |
| 1178 | @var{file}(@var{section}): warning: smaller common is here |
| 1179 | @end smallexample |
| 1180 | @end enumerate |
| 1181 | |
| 1182 | @kindex --warn-constructors |
| 1183 | @item --warn-constructors |
| 1184 | Warn if any global constructors are used. This is only useful for a few |
| 1185 | object file formats. For formats like COFF or ELF, the linker can not |
| 1186 | detect the use of global constructors. |
| 1187 | |
| 1188 | @kindex --warn-multiple-gp |
| 1189 | @item --warn-multiple-gp |
| 1190 | Warn if multiple global pointer values are required in the output file. |
| 1191 | This is only meaningful for certain processors, such as the Alpha. |
| 1192 | Specifically, some processors put large-valued constants in a special |
| 1193 | section. A special register (the global pointer) points into the middle |
| 1194 | of this section, so that constants can be loaded efficiently via a |
| 1195 | base-register relative addressing mode. Since the offset in |
| 1196 | base-register relative mode is fixed and relatively small (e.g., 16 |
| 1197 | bits), this limits the maximum size of the constant pool. Thus, in |
| 1198 | large programs, it is often necessary to use multiple global pointer |
| 1199 | values in order to be able to address all possible constants. This |
| 1200 | option causes a warning to be issued whenever this case occurs. |
| 1201 | |
| 1202 | @kindex --warn-once |
| 1203 | @cindex warnings, on undefined symbols |
| 1204 | @cindex undefined symbols, warnings on |
| 1205 | @item --warn-once |
| 1206 | Only warn once for each undefined symbol, rather than once per module |
| 1207 | which refers to it. |
| 1208 | |
| 1209 | @kindex --warn-section-align |
| 1210 | @cindex warnings, on section alignment |
| 1211 | @cindex section alignment, warnings on |
| 1212 | @item --warn-section-align |
| 1213 | Warn if the address of an output section is changed because of |
| 1214 | alignment. Typically, the alignment will be set by an input section. |
| 1215 | The address will only be changed if it not explicitly specified; that |
| 1216 | is, if the @code{SECTIONS} command does not specify a start address for |
| 1217 | the section (@pxref{SECTIONS}). |
| 1218 | |
| 1219 | @kindex --whole-archive |
| 1220 | @cindex including an entire archive |
| 1221 | @item --whole-archive |
| 1222 | For each archive mentioned on the command line after the |
| 1223 | @code{--whole-archive} option, include every object file in the archive |
| 1224 | in the link, rather than searching the archive for the required object |
| 1225 | files. This is normally used to turn an archive file into a shared |
| 1226 | library, forcing every object to be included in the resulting shared |
| 1227 | library. This option may be used more than once. |
| 1228 | |
| 1229 | @kindex --wrap |
| 1230 | @item --wrap @var{symbol} |
| 1231 | Use a wrapper function for @var{symbol}. Any undefined reference to |
| 1232 | @var{symbol} will be resolved to @code{__wrap_@var{symbol}}. Any |
| 1233 | undefined reference to @code{__real_@var{symbol}} will be resolved to |
| 1234 | @var{symbol}. |
| 1235 | |
| 1236 | This can be used to provide a wrapper for a system function. The |
| 1237 | wrapper function should be called @code{__wrap_@var{symbol}}. If it |
| 1238 | wishes to call the system function, it should call |
| 1239 | @code{__real_@var{symbol}}. |
| 1240 | |
| 1241 | Here is a trivial example: |
| 1242 | |
| 1243 | @smallexample |
| 1244 | void * |
| 1245 | __wrap_malloc (int c) |
| 1246 | @{ |
| 1247 | printf ("malloc called with %ld\n", c); |
| 1248 | return __real_malloc (c); |
| 1249 | @} |
| 1250 | @end smallexample |
| 1251 | |
| 1252 | If you link other code with this file using @code{--wrap malloc}, then |
| 1253 | all calls to @code{malloc} will call the function @code{__wrap_malloc} |
| 1254 | instead. The call to @code{__real_malloc} in @code{__wrap_malloc} will |
| 1255 | call the real @code{malloc} function. |
| 1256 | |
| 1257 | You may wish to provide a @code{__real_malloc} function as well, so that |
| 1258 | links without the @code{--wrap} option will succeed. If you do this, |
| 1259 | you should not put the definition of @code{__real_malloc} in the same |
| 1260 | file as @code{__wrap_malloc}; if you do, the assembler may resolve the |
| 1261 | call before the linker has a chance to wrap it to @code{malloc}. |
| 1262 | |
| 1263 | @end table |
| 1264 | |
| 1265 | @ifset UsesEnvVars |
| 1266 | @node Environment |
| 1267 | @section Environment Variables |
| 1268 | |
| 1269 | You can change the behavior of @code{ld} with the environment variables |
| 1270 | @code{GNUTARGET} and @code{LDEMULATION}. |
| 1271 | |
| 1272 | @kindex GNUTARGET |
| 1273 | @cindex default input format |
| 1274 | @code{GNUTARGET} determines the input-file object format if you don't |
| 1275 | use @samp{-b} (or its synonym @samp{--format}). Its value should be one |
| 1276 | of the BFD names for an input format (@pxref{BFD}). If there is no |
| 1277 | @code{GNUTARGET} in the environment, @code{ld} uses the natural format |
| 1278 | of the target. If @code{GNUTARGET} is set to @code{default} then BFD |
| 1279 | attempts to discover the input format by examining binary input files; |
| 1280 | this method often succeeds, but there are potential ambiguities, since |
| 1281 | there is no method of ensuring that the magic number used to specify |
| 1282 | object-file formats is unique. However, the configuration procedure for |
| 1283 | BFD on each system places the conventional format for that system first |
| 1284 | in the search-list, so ambiguities are resolved in favor of convention. |
| 1285 | |
| 1286 | @kindex LDEMULATION |
| 1287 | @cindex default emulation |
| 1288 | @cindex emulation, default |
| 1289 | @code{LDEMULATION} determines the default emulation if you don't use the |
| 1290 | @samp{-m} option. The emulation can affect various aspects of linker |
| 1291 | behaviour, particularly the default linker script. You can list the |
| 1292 | available emulations with the @samp{--verbose} or @samp{-V} options. If |
| 1293 | the @samp{-m} option is not used, and the @code{LDEMULATION} environment |
| 1294 | variable is not defined, the default emulation depends upon how the |
| 1295 | linker was configured. |
| 1296 | @end ifset |
| 1297 | |
| 1298 | @node Scripts |
| 1299 | @chapter Linker Scripts |
| 1300 | |
| 1301 | @cindex scripts |
| 1302 | @cindex linker scripts |
| 1303 | @cindex command files |
| 1304 | Every link is controlled by a @dfn{linker script}. This script is |
| 1305 | written in the linker command language. |
| 1306 | |
| 1307 | The main purpose of the linker script is to describe how the sections in |
| 1308 | the input files should be mapped into the output file, and to control |
| 1309 | the memory layout of the output file. Most linker scripts do nothing |
| 1310 | more than this. However, when necessary, the linker script can also |
| 1311 | direct the linker to perform many other operations, using the commands |
| 1312 | described below. |
| 1313 | |
| 1314 | The linker always uses a linker script. If you do not supply one |
| 1315 | yourself, the linker will use a default script that is compiled into the |
| 1316 | linker executable. You can use the @samp{--verbose} command line option |
| 1317 | to display the default linker script. Certain command line options, |
| 1318 | such as @samp{-r} or @samp{-N}, will affect the default linker script. |
| 1319 | |
| 1320 | You may supply your own linker script by using the @samp{-T} command |
| 1321 | line option. When you do this, your linker script will replace the |
| 1322 | default linker script. |
| 1323 | |
| 1324 | You may also use linker scripts implicitly by naming them as input files |
| 1325 | to the linker, as though they were files to be linked. @xref{Implicit |
| 1326 | Linker Scripts}. |
| 1327 | |
| 1328 | @menu |
| 1329 | * Basic Script Concepts:: Basic Linker Script Concepts |
| 1330 | * Script Format:: Linker Script Format |
| 1331 | * Simple Example:: Simple Linker Script Example |
| 1332 | * Simple Commands:: Simple Linker Script Commands |
| 1333 | * Assignments:: Assigning Values to Symbols |
| 1334 | * SECTIONS:: SECTIONS Command |
| 1335 | * MEMORY:: MEMORY Command |
| 1336 | * PHDRS:: PHDRS Command |
| 1337 | * VERSION:: VERSION Command |
| 1338 | * Expressions:: Expressions in Linker Scripts |
| 1339 | * Implicit Linker Scripts:: Implicit Linker Scripts |
| 1340 | @end menu |
| 1341 | |
| 1342 | @node Basic Script Concepts |
| 1343 | @section Basic Linker Script Concepts |
| 1344 | @cindex linker script concepts |
| 1345 | We need to define some basic concepts and vocabulary in order to |
| 1346 | describe the linker script language. |
| 1347 | |
| 1348 | The linker combines input files into a single output file. The output |
| 1349 | file and each input file are in a special data format known as an |
| 1350 | @dfn{object file format}. Each file is called an @dfn{object file}. |
| 1351 | The output file is often called an @dfn{executable}, but for our |
| 1352 | purposes we will also call it an object file. Each object file has, |
| 1353 | among other things, a list of @dfn{sections}. We sometimes refer to a |
| 1354 | section in an input file as an @dfn{input section}; similarly, a section |
| 1355 | in the output file is an @dfn{output section}. |
| 1356 | |
| 1357 | Each section in an object file has a name and a size. Most sections |
| 1358 | also have an associated block of data, known as the @dfn{section |
| 1359 | contents}. A section may be marked as @dfn{loadable}, which mean that |
| 1360 | the contents should be loaded into memory when the output file is run. |
| 1361 | A section with no contents may be @dfn{allocatable}, which means that an |
| 1362 | area in memory should be set aside, but nothing in particular should be |
| 1363 | loaded there (in some cases this memory must be zeroed out). A section |
| 1364 | which is neither loadable nor allocatable typically contains some sort |
| 1365 | of debugging information. |
| 1366 | |
| 1367 | Every loadable or allocatable output section has two addresses. The |
| 1368 | first is the @dfn{VMA}, or virtual memory address. This is the address |
| 1369 | the section will have when the output file is run. The second is the |
| 1370 | @dfn{LMA}, or load memory address. This is the address at which the |
| 1371 | section will be loaded. In most cases the two addresses will be the |
| 1372 | same. An example of when they might be different is when a data section |
| 1373 | is loaded into ROM, and then copied into RAM when the program starts up |
| 1374 | (this technique is often used to initialize global variables in a ROM |
| 1375 | based system). In this case the ROM address would be the LMA, and the |
| 1376 | RAM address would be the VMA. |
| 1377 | |
| 1378 | You can see the sections in an object file by using the @code{objdump} |
| 1379 | program with the @samp{-h} option. |
| 1380 | |
| 1381 | Every object file also has a list of @dfn{symbols}, known as the |
| 1382 | @dfn{symbol table}. A symbol may be defined or undefined. Each symbol |
| 1383 | has a name, and each defined symbol has an address, among other |
| 1384 | information. If you compile a C or C++ program into an object file, you |
| 1385 | will get a defined symbol for every defined function and global or |
| 1386 | static variable. Every undefined function or global variable which is |
| 1387 | referenced in the input file will become an undefined symbol. |
| 1388 | |
| 1389 | You can see the symbols in an object file by using the @code{nm} |
| 1390 | program, or by using the @code{objdump} program with the @samp{-t} |
| 1391 | option. |
| 1392 | |
| 1393 | @node Script Format |
| 1394 | @section Linker Script Format |
| 1395 | @cindex linker script format |
| 1396 | Linker scripts are text files. |
| 1397 | |
| 1398 | You write a linker script as a series of commands. Each command is |
| 1399 | either a keyword, possibly followed by arguments, or an assignment to a |
| 1400 | symbol. You may separate commands using semicolons. Whitespace is |
| 1401 | generally ignored. |
| 1402 | |
| 1403 | Strings such as file or format names can normally be entered directly. |
| 1404 | If the file name contains a character such as a comma which would |
| 1405 | otherwise serve to separate file names, you may put the file name in |
| 1406 | double quotes. There is no way to use a double quote character in a |
| 1407 | file name. |
| 1408 | |
| 1409 | You may include comments in linker scripts just as in C, delimited by |
| 1410 | @samp{/*} and @samp{*/}. As in C, comments are syntactically equivalent |
| 1411 | to whitespace. |
| 1412 | |
| 1413 | @node Simple Example |
| 1414 | @section Simple Linker Script Example |
| 1415 | @cindex linker script example |
| 1416 | @cindex example of linker script |
| 1417 | Many linker scripts are fairly simple. |
| 1418 | |
| 1419 | The simplest possible linker script has just one command: |
| 1420 | @samp{SECTIONS}. You use the @samp{SECTIONS} command to describe the |
| 1421 | memory layout of the output file. |
| 1422 | |
| 1423 | The @samp{SECTIONS} command is a powerful command. Here we will |
| 1424 | describe a simple use of it. Let's assume your program consists only of |
| 1425 | code, initialized data, and uninitialized data. These will be in the |
| 1426 | @samp{.text}, @samp{.data}, and @samp{.bss} sections, respectively. |
| 1427 | Let's assume further that these are the only sections which appear in |
| 1428 | your input files. |
| 1429 | |
| 1430 | For this example, let's say that the code should be loaded at address |
| 1431 | 0x10000, and that the data should start at address 0x8000000. Here is a |
| 1432 | linker script which will do that: |
| 1433 | @smallexample |
| 1434 | SECTIONS |
| 1435 | @{ |
| 1436 | . = 0x10000; |
| 1437 | .text : @{ *(.text) @} |
| 1438 | . = 0x8000000; |
| 1439 | .data : @{ *(.data) @} |
| 1440 | .bss : @{ *(.bss) @} |
| 1441 | @} |
| 1442 | @end smallexample |
| 1443 | |
| 1444 | You write the @samp{SECTIONS} command as the keyword @samp{SECTIONS}, |
| 1445 | followed by a series of symbol assignments and output section |
| 1446 | descriptions enclosed in curly braces. |
| 1447 | |
| 1448 | The first line in the above example sets the special symbol @samp{.}, |
| 1449 | which is the location counter. If you do not specify the address of an |
| 1450 | output section in some other way (other ways are described later), the |
| 1451 | address is set from the current value of the location counter. The |
| 1452 | location counter is then incremented by the size of the output section. |
| 1453 | |
| 1454 | The first line inside the @samp{SECTIONS} command of the above example |
| 1455 | sets the value of the special symbol @samp{.}, which is the location |
| 1456 | counter. If you do not specify the address of an output section in some |
| 1457 | other way (other ways are described later), the address is set from the |
| 1458 | current value of the location counter. The location counter is then |
| 1459 | incremented by the size of the output section. At the start of the |
| 1460 | @samp{SECTIONS} command, the location counter has the value @samp{0}. |
| 1461 | |
| 1462 | The second line defines an output section, @samp{.text}. The colon is |
| 1463 | required syntax which may be ignored for now. Within the curly braces |
| 1464 | after the output section name, you list the names of the input sections |
| 1465 | which should be placed into this output section. The @samp{*} is a |
| 1466 | wildcard which matches any file name. The expression @samp{*(.text)} |
| 1467 | means all @samp{.text} input sections in all input files. |
| 1468 | |
| 1469 | Since the location counter is @samp{0x10000} when the output section |
| 1470 | @samp{.text} is defined, the linker will set the address of the |
| 1471 | @samp{.text} section in the output file to be @samp{0x10000}. |
| 1472 | |
| 1473 | The remaining lines define the @samp{.data} and @samp{.bss} sections in |
| 1474 | the output file. The linker will place the @samp{.data} output section |
| 1475 | at address @samp{0x8000000}. After the linker places the @samp{.data} |
| 1476 | output section, the value of the location counter will be |
| 1477 | @samp{0x8000000} plus the size of the @samp{.data} output section. The |
| 1478 | effect is that the linker will place the @samp{.bss} output section |
| 1479 | immediately after the @samp{.data} output section in memory |
| 1480 | |
| 1481 | The linker will ensure that each output section has the required |
| 1482 | alignment, by increasing the location counter if necessary. In this |
| 1483 | example, the specified addresses for the @samp{.text} and @samp{.data} |
| 1484 | sections will probably satisfy any alignment constraints, but the linker |
| 1485 | may have to create a small gap between the @samp{.data} and @samp{.bss} |
| 1486 | sections. |
| 1487 | |
| 1488 | That's it! That's a simple and complete linker script. |
| 1489 | |
| 1490 | @node Simple Commands |
| 1491 | @section Simple Linker Script Commands |
| 1492 | @cindex linker script simple commands |
| 1493 | In this section we describe the simple linker script commands. |
| 1494 | |
| 1495 | @menu |
| 1496 | * Entry Point:: Setting the entry point |
| 1497 | * File Commands:: Commands dealing with files |
| 1498 | @ifclear SingleFormat |
| 1499 | * Format Commands:: Commands dealing with object file formats |
| 1500 | @end ifclear |
| 1501 | |
| 1502 | * Miscellaneous Commands:: Other linker script commands |
| 1503 | @end menu |
| 1504 | |
| 1505 | @node Entry Point |
| 1506 | @subsection Setting the entry point |
| 1507 | @kindex ENTRY(@var{symbol}) |
| 1508 | @cindex start of execution |
| 1509 | @cindex first instruction |
| 1510 | @cindex entry point |
| 1511 | The first instruction to execute in a program is called the @dfn{entry |
| 1512 | point}. You can use the @code{ENTRY} linker script command to set the |
| 1513 | entry point. The argument is a symbol name: |
| 1514 | @smallexample |
| 1515 | ENTRY(@var{symbol}) |
| 1516 | @end smallexample |
| 1517 | |
| 1518 | There are several ways to set the entry point. The linker will set the |
| 1519 | entry point by trying each of the following methods in order, and |
| 1520 | stopping when one of them succeeds: |
| 1521 | @itemize @bullet |
| 1522 | @item |
| 1523 | the @samp{-e} @var{entry} command-line option; |
| 1524 | @item |
| 1525 | the @code{ENTRY(@var{symbol})} command in a linker script; |
| 1526 | @item |
| 1527 | the value of the symbol @code{start}, if defined; |
| 1528 | @item |
| 1529 | the address of the first byte of the @samp{.text} section, if present; |
| 1530 | @item |
| 1531 | The address @code{0}. |
| 1532 | @end itemize |
| 1533 | |
| 1534 | @node File Commands |
| 1535 | @subsection Commands dealing with files |
| 1536 | @cindex linker script file commands |
| 1537 | Several linker script commands deal with files. |
| 1538 | |
| 1539 | @table @code |
| 1540 | @item INCLUDE @var{filename} |
| 1541 | @kindex INCLUDE @var{filename} |
| 1542 | @cindex including a linker script |
| 1543 | Include the linker script @var{filename} at this point. The file will |
| 1544 | be searched for in the current directory, and in any directory specified |
| 1545 | with the @code{-L} option. You can nest calls to @code{INCLUDE} up to |
| 1546 | 10 levels deep. |
| 1547 | |
| 1548 | @item INPUT(@var{file}, @var{file}, @dots{}) |
| 1549 | @itemx INPUT(@var{file} @var{file} @dots{}) |
| 1550 | @kindex INPUT(@var{files}) |
| 1551 | @cindex input files in linker scripts |
| 1552 | @cindex input object files in linker scripts |
| 1553 | @cindex linker script input object files |
| 1554 | The @code{INPUT} command directs the linker to include the named files |
| 1555 | in the link, as though they were named on the command line. |
| 1556 | |
| 1557 | For example, if you always want to include @file{subr.o} any time you do |
| 1558 | a link, but you can't be bothered to put it on every link command line, |
| 1559 | then you can put @samp{INPUT (subr.o)} in your linker script. |
| 1560 | |
| 1561 | In fact, if you like, you can list all of your input files in the linker |
| 1562 | script, and then invoke the linker with nothing but a @samp{-T} option. |
| 1563 | |
| 1564 | The linker will first try to open the file in the current directory. If |
| 1565 | it is not found, the linker will search through the archive library |
| 1566 | search path. See the description of @samp{-L} in @ref{Options,,Command |
| 1567 | Line Options}. |
| 1568 | |
| 1569 | If you use @samp{INPUT (-l@var{file})}, @code{ld} will transform the |
| 1570 | name to @code{lib@var{file}.a}, as with the command line argument |
| 1571 | @samp{-l}. |
| 1572 | |
| 1573 | When you use the @code{INPUT} command in an implicit linker script, the |
| 1574 | files will be included in the link at the point at which the linker |
| 1575 | script file is included. This can affect archive searching. |
| 1576 | |
| 1577 | @item GROUP(@var{file}, @var{file}, @dots{}) |
| 1578 | @itemx GROUP(@var{file} @var{file} @dots{}) |
| 1579 | @kindex GROUP(@var{files}) |
| 1580 | @cindex grouping input files |
| 1581 | The @code{GROUP} command is like @code{INPUT}, except that the named |
| 1582 | files should all be archives, and they are searched repeatedly until no |
| 1583 | new undefined references are created. See the description of @samp{-(} |
| 1584 | in @ref{Options,,Command Line Options}. |
| 1585 | |
| 1586 | @item OUTPUT(@var{filename}) |
| 1587 | @kindex OUTPUT(@var{filename}) |
| 1588 | @cindex output file name in linker scripot |
| 1589 | The @code{OUTPUT} command names the output file. Using |
| 1590 | @code{OUTPUT(@var{filename})} in the linker script is exactly like using |
| 1591 | @samp{-o @var{filename}} on the command line (@pxref{Options,,Command |
| 1592 | Line Options}). If both are used, the command line option takes |
| 1593 | precedence. |
| 1594 | |
| 1595 | You can use the @code{OUTPUT} command to define a default name for the |
| 1596 | output file other than the usual default of @file{a.out}. |
| 1597 | |
| 1598 | @item SEARCH_DIR(@var{path}) |
| 1599 | @kindex SEARCH_DIR(@var{path}) |
| 1600 | @cindex library search path in linker script |
| 1601 | @cindex archive search path in linker script |
| 1602 | @cindex search path in linker script |
| 1603 | The @code{SEARCH_DIR} command adds @var{path} to the list of paths where |
| 1604 | @code{ld} looks for archive libraries. Using |
| 1605 | @code{SEARCH_DIR(@var{path})} is exactly like using @samp{-L @var{path}} |
| 1606 | on the command line (@pxref{Options,,Command Line Options}). If both |
| 1607 | are used, then the linker will search both paths. Paths specified using |
| 1608 | the command line option are searched first. |
| 1609 | |
| 1610 | @item STARTUP(@var{filename}) |
| 1611 | @kindex STARTUP(@var{filename}) |
| 1612 | @cindex first input file |
| 1613 | The @code{STARTUP} command is just like the @code{INPUT} command, except |
| 1614 | that @var{filename} will become the first input file to be linked, as |
| 1615 | though it were specified first on the command line. This may be useful |
| 1616 | when using a system in which the entry point is always the start of the |
| 1617 | first file. |
| 1618 | @end table |
| 1619 | |
| 1620 | @ifclear SingleFormat |
| 1621 | @node Format Commands |
| 1622 | @subsection Commands dealing with object file formats |
| 1623 | A couple of linker script commands deal with object file formats. |
| 1624 | |
| 1625 | @table @code |
| 1626 | @item OUTPUT_FORMAT(@var{bfdname}) |
| 1627 | @itemx OUTPUT_FORMAT(@var{default}, @var{big}, @var{little}) |
| 1628 | @kindex OUTPUT_FORMAT(@var{bfdname}) |
| 1629 | @cindex output file format in linker script |
| 1630 | The @code{OUTPUT_FORMAT} command names the BFD format to use for the |
| 1631 | output file (@pxref{BFD}). Using @code{OUTPUT_FORMAT(@var{bfdname})} is |
| 1632 | exactly like using @samp{-oformat @var{bfdname}} on the command line |
| 1633 | (@pxref{Options,,Command Line Options}). If both are used, the command |
| 1634 | line option takes precedence. |
| 1635 | |
| 1636 | You can use @code{OUTPUT_FORMAT} with three arguments to use different |
| 1637 | formats based on the @samp{-EB} and @samp{-EL} command line options. |
| 1638 | This permits the linker script to set the output format based on the |
| 1639 | desired endianness. |
| 1640 | |
| 1641 | If neither @samp{-EB} nor @samp{-EL} are used, then the output format |
| 1642 | will be the first argument, @var{default}. If @samp{-EB} is used, the |
| 1643 | output format will be the second argument, @var{big}. If @samp{-EL} is |
| 1644 | used, the output format will be the third argument, @var{little}. |
| 1645 | |
| 1646 | For example, the default linker script for the MIPS ELF target uses this |
| 1647 | command: |
| 1648 | @smallexample |
| 1649 | OUTPUT_FORMAT(elf32-bigmips, elf32-bigmips, elf32-littlemips) |
| 1650 | @end smallexample |
| 1651 | This says that the default format for the output file is |
| 1652 | @samp{elf32-bigmips}, but if the user uses the @samp{-EL} command line |
| 1653 | option, the output file will be created in the @samp{elf32-littlemips} |
| 1654 | format. |
| 1655 | |
| 1656 | @item TARGET(@var{bfdname}) |
| 1657 | @kindex TARGET(@var{bfdname}) |
| 1658 | @cindex input file format in linker script |
| 1659 | The @code{TARGET} command names the BFD format to use when reading input |
| 1660 | files. It affects subsequent @code{INPUT} and @code{GROUP} commands. |
| 1661 | This command is like using @samp{-b @var{bfdname}} on the command line |
| 1662 | (@pxref{Options,,Command Line Options}). If the @code{TARGET} command |
| 1663 | is used but @code{OUTPUT_FORMAT} is not, then the last @code{TARGET} |
| 1664 | command is also used to set the format for the output file. @xref{BFD}. |
| 1665 | @end table |
| 1666 | @end ifclear |
| 1667 | |
| 1668 | @node Miscellaneous Commands |
| 1669 | @subsection Other linker script commands |
| 1670 | There are a few other linker scripts commands. |
| 1671 | |
| 1672 | @table @code |
| 1673 | @item ASSERT(@var{exp}, @var{message}) |
| 1674 | @kindex ASSERT |
| 1675 | @cindex assertion in linker script |
| 1676 | Ensure that @var{exp} is non-zero. If it is zero, then exit the linker |
| 1677 | with an error code, and print @var{message}. |
| 1678 | |
| 1679 | @item EXTERN(@var{symbol} @var{symbol} @dots{}) |
| 1680 | @kindex EXTERN |
| 1681 | @cindex undefined symbol in linker script |
| 1682 | Force @var{symbol} to be entered in the output file as an undefined |
| 1683 | symbol. Doing this may, for example, trigger linking of additional |
| 1684 | modules from standard libraries. You may list several @var{symbol}s for |
| 1685 | each @code{EXTERN}, and you may use @code{EXTERN} multiple times. This |
| 1686 | command has the same effect as the @samp{-u} command-line option. |
| 1687 | |
| 1688 | @item FORCE_COMMON_ALLOCATION |
| 1689 | @kindex FORCE_COMMON_ALLOCATION |
| 1690 | @cindex common allocation in linker script |
| 1691 | This command has the same effect as the @samp{-d} command-line option: |
| 1692 | to make @code{ld} assign space to common symbols even if a relocatable |
| 1693 | output file is specified (@samp{-r}). |
| 1694 | |
| 1695 | @item NOCROSSREFS(@var{section} @var{section} @dots{}) |
| 1696 | @kindex NOCROSSREFS(@var{sections}) |
| 1697 | @cindex cross references |
| 1698 | This command may be used to tell @code{ld} to issue an error about any |
| 1699 | references among certain output sections. |
| 1700 | |
| 1701 | In certain types of programs, particularly on embedded systems when |
| 1702 | using overlays, when one section is loaded into memory, another section |
| 1703 | will not be. Any direct references between the two sections would be |
| 1704 | errors. For example, it would be an error if code in one section called |
| 1705 | a function defined in the other section. |
| 1706 | |
| 1707 | The @code{NOCROSSREFS} command takes a list of output section names. If |
| 1708 | @code{ld} detects any cross references between the sections, it reports |
| 1709 | an error and returns a non-zero exit status. Note that the |
| 1710 | @code{NOCROSSREFS} command uses output section names, not input section |
| 1711 | names. |
| 1712 | |
| 1713 | @ifclear SingleFormat |
| 1714 | @item OUTPUT_ARCH(@var{bfdarch}) |
| 1715 | @kindex OUTPUT_ARCH(@var{bfdarch}) |
| 1716 | @cindex machine architecture |
| 1717 | @cindex architecture |
| 1718 | Specify a particular output machine architecture. The argument is one |
| 1719 | of the names used by the BFD library (@pxref{BFD}). You can see the |
| 1720 | architecture of an object file by using the @code{objdump} program with |
| 1721 | the @samp{-f} option. |
| 1722 | @end ifclear |
| 1723 | @end table |
| 1724 | |
| 1725 | @node Assignments |
| 1726 | @section Assigning Values to Symbols |
| 1727 | @cindex assignment in scripts |
| 1728 | @cindex symbol definition, scripts |
| 1729 | @cindex variables, defining |
| 1730 | You may assign a value to a symbol in a linker script. This will define |
| 1731 | the symbol as a global symbol. |
| 1732 | |
| 1733 | @menu |
| 1734 | * Simple Assignments:: Simple Assignments |
| 1735 | * PROVIDE:: PROVIDE |
| 1736 | @end menu |
| 1737 | |
| 1738 | @node Simple Assignments |
| 1739 | @subsection Simple Assignments |
| 1740 | |
| 1741 | You may assign to a symbol using any of the C assignment operators: |
| 1742 | |
| 1743 | @table @code |
| 1744 | @item @var{symbol} = @var{expression} ; |
| 1745 | @itemx @var{symbol} += @var{expression} ; |
| 1746 | @itemx @var{symbol} -= @var{expression} ; |
| 1747 | @itemx @var{symbol} *= @var{expression} ; |
| 1748 | @itemx @var{symbol} /= @var{expression} ; |
| 1749 | @itemx @var{symbol} <<= @var{expression} ; |
| 1750 | @itemx @var{symbol} >>= @var{expression} ; |
| 1751 | @itemx @var{symbol} &= @var{expression} ; |
| 1752 | @itemx @var{symbol} |= @var{expression} ; |
| 1753 | @end table |
| 1754 | |
| 1755 | The first case will define @var{symbol} to the value of |
| 1756 | @var{expression}. In the other cases, @var{symbol} must already be |
| 1757 | defined, and the value will be adjusted accordingly. |
| 1758 | |
| 1759 | The special symbol name @samp{.} indicates the location counter. You |
| 1760 | may only use this within a @code{SECTIONS} command. |
| 1761 | |
| 1762 | The semicolon after @var{expression} is required. |
| 1763 | |
| 1764 | Expressions are defined below; see @ref{Expressions}. |
| 1765 | |
| 1766 | You may write symbol assignments as commands in their own right, or as |
| 1767 | statements within a @code{SECTIONS} command, or as part of an output |
| 1768 | section description in a @code{SECTIONS} command. |
| 1769 | |
| 1770 | The section of the symbol will be set from the section of the |
| 1771 | expression; for more information, see @ref{Expression Section}. |
| 1772 | |
| 1773 | Here is an example showing the three different places that symbol |
| 1774 | assignments may be used: |
| 1775 | |
| 1776 | @smallexample |
| 1777 | floating_point = 0; |
| 1778 | SECTIONS |
| 1779 | @{ |
| 1780 | .text : |
| 1781 | @{ |
| 1782 | *(.text) |
| 1783 | _etext = .; |
| 1784 | @} |
| 1785 | _bdata = (. + 3) & ~ 4; |
| 1786 | .data : @{ *(.data) @} |
| 1787 | @} |
| 1788 | @end smallexample |
| 1789 | @noindent |
| 1790 | In this example, the symbol @samp{floating_point} will be defined as |
| 1791 | zero. The symbol @samp{_etext} will be defined as the address following |
| 1792 | the last @samp{.text} input section. The symbol @samp{_bdata} will be |
| 1793 | defined as the address following the @samp{.text} output section aligned |
| 1794 | upward to a 4 byte boundary. |
| 1795 | |
| 1796 | @node PROVIDE |
| 1797 | @subsection PROVIDE |
| 1798 | @cindex PROVIDE |
| 1799 | In some cases, it is desirable for a linker script to define a symbol |
| 1800 | only if it is referenced and is not defined by any object included in |
| 1801 | the link. For example, traditional linkers defined the symbol |
| 1802 | @samp{etext}. However, ANSI C requires that the user be able to use |
| 1803 | @samp{etext} as a function name without encountering an error. The |
| 1804 | @code{PROVIDE} keyword may be used to define a symbol, such as |
| 1805 | @samp{etext}, only if it is referenced but not defined. The syntax is |
| 1806 | @code{PROVIDE(@var{symbol} = @var{expression})}. |
| 1807 | |
| 1808 | Here is an example of using @code{PROVIDE} to define @samp{etext}: |
| 1809 | @smallexample |
| 1810 | SECTIONS |
| 1811 | @{ |
| 1812 | .text : |
| 1813 | @{ |
| 1814 | *(.text) |
| 1815 | _etext = .; |
| 1816 | PROVIDE(etext = .); |
| 1817 | @} |
| 1818 | @} |
| 1819 | @end smallexample |
| 1820 | |
| 1821 | In this example, if the program defines @samp{_etext} (with a leading |
| 1822 | underscore), the linker will give a multiple definition error. If, on |
| 1823 | the other hand, the program defines @samp{etext} (with no leading |
| 1824 | underscore), the linker will silently use the definition in the program. |
| 1825 | If the program references @samp{etext} but does not define it, the |
| 1826 | linker will use the definition in the linker script. |
| 1827 | |
| 1828 | @node SECTIONS |
| 1829 | @section SECTIONS command |
| 1830 | @kindex SECTIONS |
| 1831 | The @code{SECTIONS} command tells the linker how to map input sections |
| 1832 | into output sections, and how to place the output sections in memory. |
| 1833 | |
| 1834 | The format of the @code{SECTIONS} command is: |
| 1835 | @smallexample |
| 1836 | SECTIONS |
| 1837 | @{ |
| 1838 | @var{sections-command} |
| 1839 | @var{sections-command} |
| 1840 | @dots{} |
| 1841 | @} |
| 1842 | @end smallexample |
| 1843 | |
| 1844 | Each @var{sections-command} may of be one of the following: |
| 1845 | |
| 1846 | @itemize @bullet |
| 1847 | @item |
| 1848 | an @code{ENTRY} command (@pxref{Entry Point,,Entry command}) |
| 1849 | @item |
| 1850 | a symbol assignment (@pxref{Assignments}) |
| 1851 | @item |
| 1852 | an output section description |
| 1853 | @item |
| 1854 | an overlay description |
| 1855 | @end itemize |
| 1856 | |
| 1857 | The @code{ENTRY} command and symbol assignments are permitted inside the |
| 1858 | @code{SECTIONS} command for convenience in using the location counter in |
| 1859 | those commands. This can also make the linker script easier to |
| 1860 | understand because you can use those commands at meaningful points in |
| 1861 | the layout of the output file. |
| 1862 | |
| 1863 | Output section descriptions and overlay descriptions are described |
| 1864 | below. |
| 1865 | |
| 1866 | If you do not use a @code{SECTIONS} command in your linker script, the |
| 1867 | linker will place each input section into an identically named output |
| 1868 | section in the order that the sections are first encountered in the |
| 1869 | input files. If all input sections are present in the first file, for |
| 1870 | example, the order of sections in the output file will match the order |
| 1871 | in the first input file. The first section will be at address zero. |
| 1872 | |
| 1873 | @menu |
| 1874 | * Output Section Description:: Output section description |
| 1875 | * Output Section Name:: Output section name |
| 1876 | * Output Section Address:: Output section address |
| 1877 | * Input Section:: Input section description |
| 1878 | * Output Section Data:: Output section data |
| 1879 | * Output Section Keywords:: Output section keywords |
| 1880 | * Output Section Discarding:: Output section discarding |
| 1881 | * Output Section Attributes:: Output section attributes |
| 1882 | * Overlay Description:: Overlay description |
| 1883 | @end menu |
| 1884 | |
| 1885 | @node Output Section Description |
| 1886 | @subsection Output section description |
| 1887 | The full description of an output section looks like this: |
| 1888 | @smallexample |
| 1889 | @group |
| 1890 | @var{section} [@var{address}] [(@var{type})] : [AT(@var{lma})] |
| 1891 | @{ |
| 1892 | @var{output-section-command} |
| 1893 | @var{output-section-command} |
| 1894 | @dots{} |
| 1895 | @} [>@var{region}] [:@var{phdr} :@var{phdr} @dots{}] [=@var{fillexp}] |
| 1896 | @end group |
| 1897 | @end smallexample |
| 1898 | |
| 1899 | Most output sections do not use most of the optional section attributes. |
| 1900 | |
| 1901 | The whitespace around @var{section} is required, so that the section |
| 1902 | name is unambiguous. The colon and the curly braces are also required. |
| 1903 | The line breaks and other white space are optional. |
| 1904 | |
| 1905 | Each @var{output-section-command} may be one of the following: |
| 1906 | |
| 1907 | @itemize @bullet |
| 1908 | @item |
| 1909 | a symbol assignment (@pxref{Assignments}) |
| 1910 | @item |
| 1911 | an input section description (@pxref{Input Section}) |
| 1912 | @item |
| 1913 | data values to include directly (@pxref{Output Section Data}) |
| 1914 | @item |
| 1915 | a special output section keyword (@pxref{Output Section Keywords}) |
| 1916 | @end itemize |
| 1917 | |
| 1918 | @node Output Section Name |
| 1919 | @subsection Output section name |
| 1920 | @cindex name, section |
| 1921 | @cindex section name |
| 1922 | The name of the output section is @var{section}. @var{section} must |
| 1923 | meet the constraints of your output format. In formats which only |
| 1924 | support a limited number of sections, such as @code{a.out}, the name |
| 1925 | must be one of the names supported by the format (@code{a.out}, for |
| 1926 | example, allows only @samp{.text}, @samp{.data} or @samp{.bss}). If the |
| 1927 | output format supports any number of sections, but with numbers and not |
| 1928 | names (as is the case for Oasys), the name should be supplied as a |
| 1929 | quoted numeric string. A section name may consist of any sequence of |
| 1930 | characters, but a name which contains any unusual characters such as |
| 1931 | commas must be quoted. |
| 1932 | |
| 1933 | The output section name @samp{/DISCARD/} is special; @ref{Output Section |
| 1934 | Discarding}. |
| 1935 | |
| 1936 | @node Output Section Address |
| 1937 | @subsection Output section address |
| 1938 | @cindex address, section |
| 1939 | @cindex section address |
| 1940 | The @var{address} is an expression for the VMA (the virtual memory |
| 1941 | address) of the output section. If you do not provide @var{address}, |
| 1942 | the linker will set it based on @var{region} if present, or otherwise |
| 1943 | based on the current value of the location counter. |
| 1944 | |
| 1945 | If you provide @var{address}, the address of the output section will be |
| 1946 | set to precisely that. If you provide neither @var{address} nor |
| 1947 | @var{region}, then the address of the output section will be set to the |
| 1948 | current value of the location counter aligned to the alignment |
| 1949 | requirements of the output section. The alignment requirement of the |
| 1950 | output section is the strictest alignment of any input section contained |
| 1951 | within the output section. |
| 1952 | |
| 1953 | For example, |
| 1954 | @smallexample |
| 1955 | .text . : @{ *(.text) @} |
| 1956 | @end smallexample |
| 1957 | @noindent |
| 1958 | and |
| 1959 | @smallexample |
| 1960 | .text : @{ *(.text) @} |
| 1961 | @end smallexample |
| 1962 | @noindent |
| 1963 | are subtly different. The first will set the address of the |
| 1964 | @samp{.text} output section to the current value of the location |
| 1965 | counter. The second will set it to the current value of the location |
| 1966 | counter aligned to the strictest alignment of a @samp{.text} input |
| 1967 | section. |
| 1968 | |
| 1969 | The @var{address} may be an arbitrary expression; @ref{Expressions}. |
| 1970 | For example, if you want to align the section on a 0x10 byte boundary, |
| 1971 | so that the lowest four bits of the section address are zero, you could |
| 1972 | do something like this: |
| 1973 | @smallexample |
| 1974 | .text ALIGN(0x10) : @{ *(.text) @} |
| 1975 | @end smallexample |
| 1976 | @noindent |
| 1977 | This works because @code{ALIGN} returns the current location counter |
| 1978 | aligned upward to the specified value. |
| 1979 | |
| 1980 | Specifying @var{address} for a section will change the value of the |
| 1981 | location counter. |
| 1982 | |
| 1983 | @node Input Section |
| 1984 | @subsection Input section description |
| 1985 | @cindex input sections |
| 1986 | @cindex mapping input sections to output sections |
| 1987 | The most common output section command is an input section description. |
| 1988 | |
| 1989 | The input section description is the most basic linker script operation. |
| 1990 | You use output sections to tell the linker how to lay out your program |
| 1991 | in memory. You use input section descriptions to tell the linker how to |
| 1992 | map the input files into your memory layout. |
| 1993 | |
| 1994 | @menu |
| 1995 | * Input Section Basics:: Input section basics |
| 1996 | * Input Section Wildcards:: Input section wildcard patterns |
| 1997 | * Input Section Common:: Input section for common symbols |
| 1998 | * Input Section Keep:: Input section and garbage collection |
| 1999 | * Input Section Example:: Input section example |
| 2000 | @end menu |
| 2001 | |
| 2002 | @node Input Section Basics |
| 2003 | @subsubsection Input section basics |
| 2004 | @cindex input section basics |
| 2005 | An input section description consists of a file name optionally followed |
| 2006 | by a list of section names in parentheses. |
| 2007 | |
| 2008 | The file name and the section name may be wildcard patterns, which we |
| 2009 | describe further below (@pxref{Input Section Wildcards}). |
| 2010 | |
| 2011 | The most common input section description is to include all input |
| 2012 | sections with a particular name in the output section. For example, to |
| 2013 | include all input @samp{.text} sections, you would write: |
| 2014 | @smallexample |
| 2015 | *(.text) |
| 2016 | @end smallexample |
| 2017 | @noindent |
| 2018 | Here the @samp{*} is a wildcard which matches any file name. |
| 2019 | |
| 2020 | There are two ways to include more than one section: |
| 2021 | @smallexample |
| 2022 | *(.text .rdata) |
| 2023 | *(.text) *(.rdata) |
| 2024 | @end smallexample |
| 2025 | @noindent |
| 2026 | The difference between these is the order in which the @samp{.text} and |
| 2027 | @samp{.rdata} input sections will appear in the output section. In the |
| 2028 | first example, they will be intermingled. In the second example, all |
| 2029 | @samp{.text} input sections will appear first, followed by all |
| 2030 | @samp{.rdata} input sections. |
| 2031 | |
| 2032 | You can specify a file name to include sections from a particular file. |
| 2033 | You would do this if one or more of your files contain special data that |
| 2034 | needs to be at a particular location in memory. For example: |
| 2035 | @smallexample |
| 2036 | data.o(.data) |
| 2037 | @end smallexample |
| 2038 | |
| 2039 | If you use a file name without a list of sections, then all sections in |
| 2040 | the input file will be included in the output section. This is not |
| 2041 | commonly done, but it may by useful on occasion. For example: |
| 2042 | @smallexample |
| 2043 | data.o |
| 2044 | @end smallexample |
| 2045 | |
| 2046 | When you use a file name which does not contain any wild card |
| 2047 | characters, the linker will first see if you also specified the file |
| 2048 | name on the linker command line or in an @code{INPUT} command. If you |
| 2049 | did not, the linker will attempt to open the file as an input file, as |
| 2050 | though it appeared on the command line. Note that this differs from an |
| 2051 | @code{INPUT} command, because the linker will not search for the file in |
| 2052 | the archive search path. |
| 2053 | |
| 2054 | @node Input Section Wildcards |
| 2055 | @subsubsection Input section wildcard patterns |
| 2056 | @cindex input section wildcards |
| 2057 | @cindex wildcard file name patterns |
| 2058 | @cindex file name wildcard patterns |
| 2059 | @cindex section name wildcard patterns |
| 2060 | In an input section description, either the file name or the section |
| 2061 | name or both may be wildcard patterns. |
| 2062 | |
| 2063 | The file name of @samp{*} seen in many examples is a simple wildcard |
| 2064 | pattern for the file name. |
| 2065 | |
| 2066 | The wildcard patterns are like those used by the Unix shell. |
| 2067 | |
| 2068 | @table @samp |
| 2069 | @item * |
| 2070 | matches any number of characters |
| 2071 | @item ? |
| 2072 | matches any single character |
| 2073 | @item [@var{chars}] |
| 2074 | matches a single instance of any of the @var{chars}; the @samp{-} |
| 2075 | character may be used to specify a range of characters, as in |
| 2076 | @samp{[a-z]} to match any lower case letter |
| 2077 | @item \ |
| 2078 | quotes the following character |
| 2079 | @end table |
| 2080 | |
| 2081 | When a file name is matched with a wildcard, the wildcard characters |
| 2082 | will not match a @samp{/} character (used to separate directory names on |
| 2083 | Unix). A pattern consisting of a single @samp{*} character is an |
| 2084 | exception; it will always match any file name, whether it contains a |
| 2085 | @samp{/} or not. In a section name, the wildcard characters will match |
| 2086 | a @samp{/} character. |
| 2087 | |
| 2088 | File name wildcard patterns only match files which are explicitly |
| 2089 | specified on the command line or in an @code{INPUT} command. The linker |
| 2090 | does not search directories to expand wildcards. |
| 2091 | |
| 2092 | If a file name matches more than one wildcard pattern, or if a file name |
| 2093 | appears explicitly and is also matched by a wildcard pattern, the linker |
| 2094 | will use the first match in the linker script. For example, this |
| 2095 | sequence of input section descriptions is probably in error, because the |
| 2096 | @file{data.o} rule will not be used: |
| 2097 | @smallexample |
| 2098 | .data : @{ *(.data) @} |
| 2099 | .data1 : @{ data.o(.data) @} |
| 2100 | @end smallexample |
| 2101 | |
| 2102 | @cindex SORT |
| 2103 | Normally, the linker will place files and sections matched by wildcards |
| 2104 | in the order in which they are seen during the link. You can change |
| 2105 | this by using the @code{SORT} keyword, which appears before a wildcard |
| 2106 | pattern in parentheses (e.g., @code{SORT(.text*)}). When the |
| 2107 | @code{SORT} keyword is used, the linker will sort the files or sections |
| 2108 | into ascending order by name before placing them in the output file. |
| 2109 | |
| 2110 | If you ever get confused about where input sections are going, use the |
| 2111 | @samp{-M} linker option to generate a map file. The map file shows |
| 2112 | precisely how input sections are mapped to output sections. |
| 2113 | |
| 2114 | This example shows how wildcard patterns might be used to partition |
| 2115 | files. This linker script directs the linker to place all @samp{.text} |
| 2116 | sections in @samp{.text} and all @samp{.bss} sections in @samp{.bss}. |
| 2117 | The linker will place the @samp{.data} section from all files beginning |
| 2118 | with an upper case character in @samp{.DATA}; for all other files, the |
| 2119 | linker will place the @samp{.data} section in @samp{.data}. |
| 2120 | @smallexample |
| 2121 | @group |
| 2122 | SECTIONS @{ |
| 2123 | .text : @{ *(.text) @} |
| 2124 | .DATA : @{ [A-Z]*(.data) @} |
| 2125 | .data : @{ *(.data) @} |
| 2126 | .bss : @{ *(.bss) @} |
| 2127 | @} |
| 2128 | @end group |
| 2129 | @end smallexample |
| 2130 | |
| 2131 | @node Input Section Common |
| 2132 | @subsubsection Input section for common symbols |
| 2133 | @cindex common symbol placement |
| 2134 | @cindex uninitialized data placement |
| 2135 | A special notation is needed for common symbols, because in many object |
| 2136 | file formats common symbols do not have a particular input section. The |
| 2137 | linker treats common symbols as though they are in an input section |
| 2138 | named @samp{COMMON}. |
| 2139 | |
| 2140 | You may use file names with the @samp{COMMON} section just as with any |
| 2141 | other input sections. You can use this to place common symbols from a |
| 2142 | particular input file in one section while common symbols from other |
| 2143 | input files are placed in another section. |
| 2144 | |
| 2145 | In most cases, common symbols in input files will be placed in the |
| 2146 | @samp{.bss} section in the output file. For example: |
| 2147 | @smallexample |
| 2148 | .bss @{ *(.bss) *(COMMON) @} |
| 2149 | @end smallexample |
| 2150 | |
| 2151 | @cindex scommon section |
| 2152 | @cindex small common symbols |
| 2153 | Some object file formats have more than one type of common symbol. For |
| 2154 | example, the MIPS ELF object file format distinguishes standard common |
| 2155 | symbols and small common symbols. In this case, the linker will use a |
| 2156 | different special section name for other types of common symbols. In |
| 2157 | the case of MIPS ELF, the linker uses @samp{COMMON} for standard common |
| 2158 | symbols and @samp{.scommon} for small common symbols. This permits you |
| 2159 | to map the different types of common symbols into memory at different |
| 2160 | locations. |
| 2161 | |
| 2162 | @cindex [COMMON] |
| 2163 | You will sometimes see @samp{[COMMON]} in old linker scripts. This |
| 2164 | notation is now considered obsolete. It is equivalent to |
| 2165 | @samp{*(COMMON)}. |
| 2166 | |
| 2167 | @node Input Section Keep |
| 2168 | @subsubsection Input section and garbage collection |
| 2169 | @cindex KEEP |
| 2170 | @cindex garbage collection |
| 2171 | When link-time garbage collection is in use (@samp{--gc-sections}), |
| 2172 | it is often useful to mark sections that should not be eliminated. |
| 2173 | This is accomplished by surrounding an input section's wildcard entry |
| 2174 | with @code{KEEP()}, as in @code{KEEP(*(.init))} or |
| 2175 | @code{KEEP(SORT(*)(.ctors))}. |
| 2176 | |
| 2177 | @node Input Section Example |
| 2178 | @subsubsection Input section example |
| 2179 | The following example is a complete linker script. It tells the linker |
| 2180 | to read all of the sections from file @file{all.o} and place them at the |
| 2181 | start of output section @samp{outputa} which starts at location |
| 2182 | @samp{0x10000}. All of section @samp{.input1} from file @file{foo.o} |
| 2183 | follows immediately, in the same output section. All of section |
| 2184 | @samp{.input2} from @file{foo.o} goes into output section |
| 2185 | @samp{outputb}, followed by section @samp{.input1} from @file{foo1.o}. |
| 2186 | All of the remaining @samp{.input1} and @samp{.input2} sections from any |
| 2187 | files are written to output section @samp{outputc}. |
| 2188 | |
| 2189 | @smallexample |
| 2190 | @group |
| 2191 | SECTIONS @{ |
| 2192 | outputa 0x10000 : |
| 2193 | @{ |
| 2194 | all.o |
| 2195 | foo.o (.input1) |
| 2196 | @} |
| 2197 | outputb : |
| 2198 | @{ |
| 2199 | foo.o (.input2) |
| 2200 | foo1.o (.input1) |
| 2201 | @} |
| 2202 | outputc : |
| 2203 | @{ |
| 2204 | *(.input1) |
| 2205 | *(.input2) |
| 2206 | @} |
| 2207 | @} |
| 2208 | @end group |
| 2209 | @end smallexample |
| 2210 | |
| 2211 | @node Output Section Data |
| 2212 | @subsection Output section data |
| 2213 | @cindex data |
| 2214 | @cindex section data |
| 2215 | @cindex output section data |
| 2216 | @kindex BYTE(@var{expression}) |
| 2217 | @kindex SHORT(@var{expression}) |
| 2218 | @kindex LONG(@var{expression}) |
| 2219 | @kindex QUAD(@var{expression}) |
| 2220 | @kindex SQUAD(@var{expression}) |
| 2221 | You can include explicit bytes of data in an output section by using |
| 2222 | @code{BYTE}, @code{SHORT}, @code{LONG}, @code{QUAD}, or @code{SQUAD} as |
| 2223 | an output section command. Each keyword is followed by an expression in |
| 2224 | parentheses providing the value to store (@pxref{Expressions}). The |
| 2225 | value of the expression is stored at the current value of the location |
| 2226 | counter. |
| 2227 | |
| 2228 | The @code{BYTE}, @code{SHORT}, @code{LONG}, and @code{QUAD} commands |
| 2229 | store one, two, four, and eight bytes (respectively). After storing the |
| 2230 | bytes, the location counter is incremented by the number of bytes |
| 2231 | stored. |
| 2232 | |
| 2233 | For example, this will store the byte 1 followed by the four byte value |
| 2234 | of the symbol @samp{addr}: |
| 2235 | @smallexample |
| 2236 | BYTE(1) |
| 2237 | LONG(addr) |
| 2238 | @end smallexample |
| 2239 | |
| 2240 | When using a 64 bit host or target, @code{QUAD} and @code{SQUAD} are the |
| 2241 | same; they both store an 8 byte, or 64 bit, value. When both host and |
| 2242 | target are 32 bits, an expression is computed as 32 bits. In this case |
| 2243 | @code{QUAD} stores a 32 bit value zero extended to 64 bits, and |
| 2244 | @code{SQUAD} stores a 32 bit value sign extended to 64 bits. |
| 2245 | |
| 2246 | If the object file format of the output file has an explicit endianness, |
| 2247 | which is the normal case, the value will be stored in that endianness. |
| 2248 | When the object file format does not have an explicit endianness, as is |
| 2249 | true of, for example, S-records, the value will be stored in the |
| 2250 | endianness of the first input object file. |
| 2251 | |
| 2252 | @kindex FILL(@var{expression}) |
| 2253 | @cindex holes, filling |
| 2254 | @cindex unspecified memory |
| 2255 | You may use the @code{FILL} command to set the fill pattern for the |
| 2256 | current section. It is followed by an expression in parentheses. Any |
| 2257 | otherwise unspecified regions of memory within the section (for example, |
| 2258 | gaps left due to the required alignment of input sections) are filled |
| 2259 | with the two least significant bytes of the expression, repeated as |
| 2260 | necessary. A @code{FILL} statement covers memory locations after the |
| 2261 | point at which it occurs in the section definition; by including more |
| 2262 | than one @code{FILL} statement, you can have different fill patterns in |
| 2263 | different parts of an output section. |
| 2264 | |
| 2265 | This example shows how to fill unspecified regions of memory with the |
| 2266 | value @samp{0x9090}: |
| 2267 | @smallexample |
| 2268 | FILL(0x9090) |
| 2269 | @end smallexample |
| 2270 | |
| 2271 | The @code{FILL} command is similar to the @samp{=@var{fillexp}} output |
| 2272 | section attribute (@pxref{Output Section Fill}), but it only affects the |
| 2273 | part of the section following the @code{FILL} command, rather than the |
| 2274 | entire section. If both are used, the @code{FILL} command takes |
| 2275 | precedence. |
| 2276 | |
| 2277 | @node Output Section Keywords |
| 2278 | @subsection Output section keywords |
| 2279 | There are a couple of keywords which can appear as output section |
| 2280 | commands. |
| 2281 | |
| 2282 | @table @code |
| 2283 | @kindex CREATE_OBJECT_SYMBOLS |
| 2284 | @cindex input filename symbols |
| 2285 | @cindex filename symbols |
| 2286 | @item CREATE_OBJECT_SYMBOLS |
| 2287 | The command tells the linker to create a symbol for each input file. |
| 2288 | The name of each symbol will be the name of the corresponding input |
| 2289 | file. The section of each symbol will be the output section in which |
| 2290 | the @code{CREATE_OBJECT_SYMBOLS} command appears. |
| 2291 | |
| 2292 | This is conventional for the a.out object file format. It is not |
| 2293 | normally used for any other object file format. |
| 2294 | |
| 2295 | @kindex CONSTRUCTORS |
| 2296 | @cindex C++ constructors, arranging in link |
| 2297 | @cindex constructors, arranging in link |
| 2298 | @item CONSTRUCTORS |
| 2299 | When linking using the a.out object file format, the linker uses an |
| 2300 | unusual set construct to support C++ global constructors and |
| 2301 | destructors. When linking object file formats which do not support |
| 2302 | arbitrary sections, such as ECOFF and XCOFF, the linker will |
| 2303 | automatically recognize C++ global constructors and destructors by name. |
| 2304 | For these object file formats, the @code{CONSTRUCTORS} command tells the |
| 2305 | linker to place constructor information in the output section where the |
| 2306 | @code{CONSTRUCTORS} command appears. The @code{CONSTRUCTORS} command is |
| 2307 | ignored for other object file formats. |
| 2308 | |
| 2309 | The symbol @w{@code{__CTOR_LIST__}} marks the start of the global |
| 2310 | constructors, and the symbol @w{@code{__DTOR_LIST}} marks the end. The |
| 2311 | first word in the list is the number of entries, followed by the address |
| 2312 | of each constructor or destructor, followed by a zero word. The |
| 2313 | compiler must arrange to actually run the code. For these object file |
| 2314 | formats @sc{gnu} C++ normally calls constructors from a subroutine |
| 2315 | @code{__main}; a call to @code{__main} is automatically inserted into |
| 2316 | the startup code for @code{main}. @sc{gnu} C++ normally runs |
| 2317 | destructors either by using @code{atexit}, or directly from the function |
| 2318 | @code{exit}. |
| 2319 | |
| 2320 | For object file formats such as @code{COFF} or @code{ELF} which support |
| 2321 | arbitrary section names, @sc{gnu} C++ will normally arrange to put the |
| 2322 | addresses of global constructors and destructors into the @code{.ctors} |
| 2323 | and @code{.dtors} sections. Placing the following sequence into your |
| 2324 | linker script will build the sort of table which the @sc{gnu} C++ |
| 2325 | runtime code expects to see. |
| 2326 | |
| 2327 | @smallexample |
| 2328 | __CTOR_LIST__ = .; |
| 2329 | LONG((__CTOR_END__ - __CTOR_LIST__) / 4 - 2) |
| 2330 | *(.ctors) |
| 2331 | LONG(0) |
| 2332 | __CTOR_END__ = .; |
| 2333 | __DTOR_LIST__ = .; |
| 2334 | LONG((__DTOR_END__ - __DTOR_LIST__) / 4 - 2) |
| 2335 | *(.dtors) |
| 2336 | LONG(0) |
| 2337 | __DTOR_END__ = .; |
| 2338 | @end smallexample |
| 2339 | |
| 2340 | If you are using the @sc{gnu} C++ support for initialization priority, |
| 2341 | which provides some control over the order in which global constructors |
| 2342 | are run, you must sort the constructors at link time to ensure that they |
| 2343 | are executed in the correct order. When using the @code{CONSTRUCTORS} |
| 2344 | command, use @samp{SORT(CONSTRUCTORS)} instead. When using the |
| 2345 | @code{.ctors} and @code{.dtors} sections, use @samp{*(SORT(.ctors))} and |
| 2346 | @samp{*(SORT(.dtors))} instead of just @samp{*(.ctors)} and |
| 2347 | @samp{*(.dtors)}. |
| 2348 | |
| 2349 | Normally the compiler and linker will handle these issues automatically, |
| 2350 | and you will not need to concern yourself with them. However, you may |
| 2351 | need to consider this if you are using C++ and writing your own linker |
| 2352 | scripts. |
| 2353 | |
| 2354 | @end table |
| 2355 | |
| 2356 | @node Output Section Discarding |
| 2357 | @subsection Output section discarding |
| 2358 | @cindex discarding sections |
| 2359 | @cindex sections, discarding |
| 2360 | @cindex removing sections |
| 2361 | The linker will not create output section which do not have any |
| 2362 | contents. This is for convenience when referring to input sections that |
| 2363 | may or may not be present in any of the input files. For example: |
| 2364 | @smallexample |
| 2365 | .foo @{ *(.foo) @} |
| 2366 | @end smallexample |
| 2367 | @noindent |
| 2368 | will only create a @samp{.foo} section in the output file if there is a |
| 2369 | @samp{.foo} section in at least one input file. |
| 2370 | |
| 2371 | If you use anything other than an input section description as an output |
| 2372 | section command, such as a symbol assignment, then the output section |
| 2373 | will always be created, even if there are no matching input sections. |
| 2374 | |
| 2375 | @cindex /DISCARD/ |
| 2376 | The special output section name @samp{/DISCARD/} may be used to discard |
| 2377 | input sections. Any input sections which are assigned to an output |
| 2378 | section named @samp{/DISCARD/} are not included in the output file. |
| 2379 | |
| 2380 | @node Output Section Attributes |
| 2381 | @subsection Output section attributes |
| 2382 | @cindex output section attributes |
| 2383 | We showed above that the full description of an output section looked |
| 2384 | like this: |
| 2385 | @smallexample |
| 2386 | @group |
| 2387 | @var{section} [@var{address}] [(@var{type})] : [AT(@var{lma})] |
| 2388 | @{ |
| 2389 | @var{output-section-command} |
| 2390 | @var{output-section-command} |
| 2391 | @dots{} |
| 2392 | @} [>@var{region}] [:@var{phdr} :@var{phdr} @dots{}] [=@var{fillexp}] |
| 2393 | @end group |
| 2394 | @end smallexample |
| 2395 | We've already described @var{section}, @var{address}, and |
| 2396 | @var{output-section-command}. In this section we will describe the |
| 2397 | remaining section attributes. |
| 2398 | |
| 2399 | @menu |
| 2400 | * Output Section Type:: Output section type |
| 2401 | * Output Section LMA:: Output section LMA |
| 2402 | * Output Section Region:: Output section region |
| 2403 | * Output Section Phdr:: Output section phdr |
| 2404 | * Output Section Fill:: Output section fill |
| 2405 | @end menu |
| 2406 | |
| 2407 | @node Output Section Type |
| 2408 | @subsubsection Output section type |
| 2409 | Each output section may have a type. The type is a keyword in |
| 2410 | parentheses. The following types are defined: |
| 2411 | |
| 2412 | @table @code |
| 2413 | @item NOLOAD |
| 2414 | The section should be marked as not loadable, so that it will not be |
| 2415 | loaded into memory when the program is run. |
| 2416 | @item DSECT |
| 2417 | @itemx COPY |
| 2418 | @itemx INFO |
| 2419 | @itemx OVERLAY |
| 2420 | These type names are supported for backward compatibility, and are |
| 2421 | rarely used. They all have the same effect: the section should be |
| 2422 | marked as not allocatable, so that no memory is allocated for the |
| 2423 | section when the program is run. |
| 2424 | @end table |
| 2425 | |
| 2426 | @kindex NOLOAD |
| 2427 | @cindex prevent unnecessary loading |
| 2428 | @cindex loading, preventing |
| 2429 | The linker normally sets the attributes of an output section based on |
| 2430 | the input sections which map into it. You can override this by using |
| 2431 | the section type. For example, in the script sample below, the |
| 2432 | @samp{ROM} section is addressed at memory location @samp{0} and does not |
| 2433 | need to be loaded when the program is run. The contents of the |
| 2434 | @samp{ROM} section will appear in the linker output file as usual. |
| 2435 | @smallexample |
| 2436 | @group |
| 2437 | SECTIONS @{ |
| 2438 | ROM 0 (NOLOAD) : @{ @dots{} @} |
| 2439 | @dots{} |
| 2440 | @} |
| 2441 | @end group |
| 2442 | @end smallexample |
| 2443 | |
| 2444 | @node Output Section LMA |
| 2445 | @subsubsection Output section LMA |
| 2446 | @kindex AT(@var{lma}) |
| 2447 | @cindex load address |
| 2448 | @cindex section load address |
| 2449 | Every section has a virtual address (VMA) and a load address (LMA); see |
| 2450 | @ref{Basic Script Concepts}. The address expression which may appear in |
| 2451 | an output section description sets the VMA (@pxref{Output Section |
| 2452 | Address}). |
| 2453 | |
| 2454 | The linker will normally set the LMA equal to the VMA. You can change |
| 2455 | that by using the @code{AT} keyword. The expression @var{lma} that |
| 2456 | follows the @code{AT} keyword specifies the load address of the section. |
| 2457 | |
| 2458 | @cindex ROM initialized data |
| 2459 | @cindex initialized data in ROM |
| 2460 | This feature is designed to make it easy to build a ROM image. For |
| 2461 | example, the following linker script creates three output sections: one |
| 2462 | called @samp{.text}, which starts at @code{0x1000}, one called |
| 2463 | @samp{.mdata}, which is loaded at the end of the @samp{.text} section |
| 2464 | even though its VMA is @code{0x2000}, and one called @samp{.bss} to hold |
| 2465 | uninitialized data at address @code{0x3000}. The symbol @code{_data} is |
| 2466 | defined with the value @code{0x2000}, which shows that the location |
| 2467 | counter holds the VMA value, not the LMA value. |
| 2468 | |
| 2469 | @smallexample |
| 2470 | @group |
| 2471 | SECTIONS |
| 2472 | @{ |
| 2473 | .text 0x1000 : @{ *(.text) _etext = . ; @} |
| 2474 | .mdata 0x2000 : |
| 2475 | AT ( ADDR (.text) + SIZEOF (.text) ) |
| 2476 | @{ _data = . ; *(.data); _edata = . ; @} |
| 2477 | .bss 0x3000 : |
| 2478 | @{ _bstart = . ; *(.bss) *(COMMON) ; _bend = . ;@} |
| 2479 | @} |
| 2480 | @end group |
| 2481 | @end smallexample |
| 2482 | |
| 2483 | The run-time initialization code for use with a program generated with |
| 2484 | this linker script would include something like the following, to copy |
| 2485 | the initialized data from the ROM image to its runtime address. Notice |
| 2486 | how this code takes advantage of the symbols defined by the linker |
| 2487 | script. |
| 2488 | |
| 2489 | @smallexample |
| 2490 | @group |
| 2491 | extern char _etext, _data, _edata, _bstart, _bend; |
| 2492 | char *src = &_etext; |
| 2493 | char *dst = &_data; |
| 2494 | |
| 2495 | /* ROM has data at end of text; copy it. */ |
| 2496 | while (dst < &_edata) @{ |
| 2497 | *dst++ = *src++; |
| 2498 | @} |
| 2499 | |
| 2500 | /* Zero bss */ |
| 2501 | for (dst = &_bstart; dst< &_bend; dst++) |
| 2502 | *dst = 0; |
| 2503 | @end group |
| 2504 | @end smallexample |
| 2505 | |
| 2506 | @node Output Section Region |
| 2507 | @subsubsection Output section region |
| 2508 | @kindex >@var{region} |
| 2509 | @cindex section, assigning to memory region |
| 2510 | @cindex memory regions and sections |
| 2511 | You can assign a section to a previously defined region of memory by |
| 2512 | using @samp{>@var{region}}. @xref{MEMORY}. |
| 2513 | |
| 2514 | Here is a simple example: |
| 2515 | @smallexample |
| 2516 | @group |
| 2517 | MEMORY @{ rom : ORIGIN = 0x1000, LENGTH = 0x1000 @} |
| 2518 | SECTIONS @{ ROM : @{ *(.text) @} >rom @} |
| 2519 | @end group |
| 2520 | @end smallexample |
| 2521 | |
| 2522 | @node Output Section Phdr |
| 2523 | @subsubsection Output section phdr |
| 2524 | @kindex :@var{phdr} |
| 2525 | @cindex section, assigning to program header |
| 2526 | @cindex program headers and sections |
| 2527 | You can assign a section to a previously defined program segment by |
| 2528 | using @samp{:@var{phdr}}. @xref{PHDRS}. If a section is assigned to |
| 2529 | one or more segments, then all subsequent allocated sections will be |
| 2530 | assigned to those segments as well, unless they use an explicitly |
| 2531 | @code{:@var{phdr}} modifier. You can use @code{:NONE} to tell the |
| 2532 | linker to not put the section in any segment at all. |
| 2533 | |
| 2534 | Here is a simple example: |
| 2535 | @smallexample |
| 2536 | @group |
| 2537 | PHDRS @{ text PT_LOAD ; @} |
| 2538 | SECTIONS @{ .text : @{ *(.text) @} :text @} |
| 2539 | @end group |
| 2540 | @end smallexample |
| 2541 | |
| 2542 | @node Output Section Fill |
| 2543 | @subsubsection Output section fill |
| 2544 | @kindex =@var{fillexp} |
| 2545 | @cindex section fill pattern |
| 2546 | @cindex fill pattern, entire section |
| 2547 | You can set the fill pattern for an entire section by using |
| 2548 | @samp{=@var{fillexp}}. @var{fillexp} is an expression |
| 2549 | (@pxref{Expressions}). Any otherwise unspecified regions of memory |
| 2550 | within the output section (for example, gaps left due to the required |
| 2551 | alignment of input sections) will be filled with the two least |
| 2552 | significant bytes of the value, repeated as necessary. |
| 2553 | |
| 2554 | You can also change the fill value with a @code{FILL} command in the |
| 2555 | output section commands; see @ref{Output Section Data}. |
| 2556 | |
| 2557 | Here is a simple example: |
| 2558 | @smallexample |
| 2559 | @group |
| 2560 | SECTIONS @{ .text : @{ *(.text) @} =0x9090 @} |
| 2561 | @end group |
| 2562 | @end smallexample |
| 2563 | |
| 2564 | @node Overlay Description |
| 2565 | @subsection Overlay description |
| 2566 | @kindex OVERLAY |
| 2567 | @cindex overlays |
| 2568 | An overlay description provides an easy way to describe sections which |
| 2569 | are to be loaded as part of a single memory image but are to be run at |
| 2570 | the same memory address. At run time, some sort of overlay manager will |
| 2571 | copy the overlaid sections in and out of the runtime memory address as |
| 2572 | required, perhaps by simply manipulating addressing bits. This approach |
| 2573 | can be useful, for example, when a certain region of memory is faster |
| 2574 | than another. |
| 2575 | |
| 2576 | Overlays are described using the @code{OVERLAY} command. The |
| 2577 | @code{OVERLAY} command is used within a @code{SECTIONS} command, like an |
| 2578 | output section description. The full syntax of the @code{OVERLAY} |
| 2579 | command is as follows: |
| 2580 | @smallexample |
| 2581 | @group |
| 2582 | OVERLAY [@var{start}] : [NOCROSSREFS] [AT ( @var{ldaddr} )] |
| 2583 | @{ |
| 2584 | @var{secname1} |
| 2585 | @{ |
| 2586 | @var{output-section-command} |
| 2587 | @var{output-section-command} |
| 2588 | @dots{} |
| 2589 | @} [:@var{phdr}@dots{}] [=@var{fill}] |
| 2590 | @var{secname2} |
| 2591 | @{ |
| 2592 | @var{output-section-command} |
| 2593 | @var{output-section-command} |
| 2594 | @dots{} |
| 2595 | @} [:@var{phdr}@dots{}] [=@var{fill}] |
| 2596 | @dots{} |
| 2597 | @} [>@var{region}] [:@var{phdr}@dots{}] [=@var{fill}] |
| 2598 | @end group |
| 2599 | @end smallexample |
| 2600 | |
| 2601 | Everything is optional except @code{OVERLAY} (a keyword), and each |
| 2602 | section must have a name (@var{secname1} and @var{secname2} above). The |
| 2603 | section definitions within the @code{OVERLAY} construct are identical to |
| 2604 | those within the general @code{SECTIONS} contruct (@pxref{SECTIONS}), |
| 2605 | except that no addresses and no memory regions may be defined for |
| 2606 | sections within an @code{OVERLAY}. |
| 2607 | |
| 2608 | The sections are all defined with the same starting address. The load |
| 2609 | addresses of the sections are arranged such that they are consecutive in |
| 2610 | memory starting at the load address used for the @code{OVERLAY} as a |
| 2611 | whole (as with normal section definitions, the load address is optional, |
| 2612 | and defaults to the start address; the start address is also optional, |
| 2613 | and defaults to the current value of the location counter). |
| 2614 | |
| 2615 | If the @code{NOCROSSREFS} keyword is used, and there any references |
| 2616 | among the sections, the linker will report an error. Since the sections |
| 2617 | all run at the same address, it normally does not make sense for one |
| 2618 | section to refer directly to another. @xref{Miscellaneous Commands, |
| 2619 | NOCROSSREFS}. |
| 2620 | |
| 2621 | For each section within the @code{OVERLAY}, the linker automatically |
| 2622 | defines two symbols. The symbol @code{__load_start_@var{secname}} is |
| 2623 | defined as the starting load address of the section. The symbol |
| 2624 | @code{__load_stop_@var{secname}} is defined as the final load address of |
| 2625 | the section. Any characters within @var{secname} which are not legal |
| 2626 | within C identifiers are removed. C (or assembler) code may use these |
| 2627 | symbols to move the overlaid sections around as necessary. |
| 2628 | |
| 2629 | At the end of the overlay, the value of the location counter is set to |
| 2630 | the start address of the overlay plus the size of the largest section. |
| 2631 | |
| 2632 | Here is an example. Remember that this would appear inside a |
| 2633 | @code{SECTIONS} construct. |
| 2634 | @smallexample |
| 2635 | @group |
| 2636 | OVERLAY 0x1000 : AT (0x4000) |
| 2637 | @{ |
| 2638 | .text0 @{ o1/*.o(.text) @} |
| 2639 | .text1 @{ o2/*.o(.text) @} |
| 2640 | @} |
| 2641 | @end group |
| 2642 | @end smallexample |
| 2643 | @noindent |
| 2644 | This will define both @samp{.text0} and @samp{.text1} to start at |
| 2645 | address 0x1000. @samp{.text0} will be loaded at address 0x4000, and |
| 2646 | @samp{.text1} will be loaded immediately after @samp{.text0}. The |
| 2647 | following symbols will be defined: @code{__load_start_text0}, |
| 2648 | @code{__load_stop_text0}, @code{__load_start_text1}, |
| 2649 | @code{__load_stop_text1}. |
| 2650 | |
| 2651 | C code to copy overlay @code{.text1} into the overlay area might look |
| 2652 | like the following. |
| 2653 | |
| 2654 | @smallexample |
| 2655 | @group |
| 2656 | extern char __load_start_text1, __load_stop_text1; |
| 2657 | memcpy ((char *) 0x1000, &__load_start_text1, |
| 2658 | &__load_stop_text1 - &__load_start_text1); |
| 2659 | @end group |
| 2660 | @end smallexample |
| 2661 | |
| 2662 | Note that the @code{OVERLAY} command is just syntactic sugar, since |
| 2663 | everything it does can be done using the more basic commands. The above |
| 2664 | example could have been written identically as follows. |
| 2665 | |
| 2666 | @smallexample |
| 2667 | @group |
| 2668 | .text0 0x1000 : AT (0x4000) @{ o1/*.o(.text) @} |
| 2669 | __load_start_text0 = LOADADDR (.text0); |
| 2670 | __load_stop_text0 = LOADADDR (.text0) + SIZEOF (.text0); |
| 2671 | .text1 0x1000 : AT (0x4000 + SIZEOF (.text0)) @{ o2/*.o(.text) @} |
| 2672 | __load_start_text1 = LOADADDR (.text1); |
| 2673 | __load_stop_text1 = LOADADDR (.text1) + SIZEOF (.text1); |
| 2674 | . = 0x1000 + MAX (SIZEOF (.text0), SIZEOF (.text1)); |
| 2675 | @end group |
| 2676 | @end smallexample |
| 2677 | |
| 2678 | @node MEMORY |
| 2679 | @section MEMORY command |
| 2680 | @kindex MEMORY |
| 2681 | @cindex memory regions |
| 2682 | @cindex regions of memory |
| 2683 | @cindex allocating memory |
| 2684 | @cindex discontinuous memory |
| 2685 | The linker's default configuration permits allocation of all available |
| 2686 | memory. You can override this by using the @code{MEMORY} command. |
| 2687 | |
| 2688 | The @code{MEMORY} command describes the location and size of blocks of |
| 2689 | memory in the target. You can use it to describe which memory regions |
| 2690 | may be used by the linker, and which memory regions it must avoid. You |
| 2691 | can then assign sections to particular memory regions. The linker will |
| 2692 | set section addresses based on the memory regions, and will warn about |
| 2693 | regions that become too full. The linker will not shuffle sections |
| 2694 | around to fit into the available regions. |
| 2695 | |
| 2696 | A linker script may contain at most one use of the @code{MEMORY} |
| 2697 | command. However, you can define as many blocks of memory within it as |
| 2698 | you wish. The syntax is: |
| 2699 | @smallexample |
| 2700 | @group |
| 2701 | MEMORY |
| 2702 | @{ |
| 2703 | @var{name} [(@var{attr})] : ORIGIN = @var{origin}, LENGTH = @var{len} |
| 2704 | @dots{} |
| 2705 | @} |
| 2706 | @end group |
| 2707 | @end smallexample |
| 2708 | |
| 2709 | The @var{name} is a name used in the linker script to refer to the |
| 2710 | region. The region name has no meaning outside of the linker script. |
| 2711 | Region names are stored in a separate name space, and will not conflict |
| 2712 | with symbol names, file names, or section names. Each memory region |
| 2713 | must have a distinct name. |
| 2714 | |
| 2715 | @cindex memory region attributes |
| 2716 | The @var{attr} string is an optional list of attributes that specify |
| 2717 | whether to use a particular memory region for an input section which is |
| 2718 | not explicitly mapped in the linker script. As described in |
| 2719 | @ref{SECTIONS}, if you do not specify an output section for some input |
| 2720 | section, the linker will create an output section with the same name as |
| 2721 | the input section. If you define region attributes, the linker will use |
| 2722 | them to select the memory region for the output section that it creates. |
| 2723 | |
| 2724 | The @var{attr} string must consist only of the following characters: |
| 2725 | @table @samp |
| 2726 | @item R |
| 2727 | Read-only section |
| 2728 | @item W |
| 2729 | Read/write section |
| 2730 | @item X |
| 2731 | Executable section |
| 2732 | @item A |
| 2733 | Allocatable section |
| 2734 | @item I |
| 2735 | Initialized section |
| 2736 | @item L |
| 2737 | Same as @samp{I} |
| 2738 | @item ! |
| 2739 | Invert the sense of any of the preceding attributes |
| 2740 | @end table |
| 2741 | |
| 2742 | If a unmapped section matches any of the listed attributes other than |
| 2743 | @samp{!}, it will be placed in the memory region. The @samp{!} |
| 2744 | attribute reverses this test, so that an unmapped section will be placed |
| 2745 | in the memory region only if it does not match any of the listed |
| 2746 | attributes. |
| 2747 | |
| 2748 | @kindex ORIGIN = |
| 2749 | @kindex o = |
| 2750 | @kindex org = |
| 2751 | The @var{origin} is an expression for the start address of the memory |
| 2752 | region. The expression must evaluate to a constant before memory |
| 2753 | allocation is performed, which means that you may not use any section |
| 2754 | relative symbols. The keyword @code{ORIGIN} may be abbreviated to |
| 2755 | @code{org} or @code{o} (but not, for example, @code{ORG}). |
| 2756 | |
| 2757 | @kindex LENGTH = |
| 2758 | @kindex len = |
| 2759 | @kindex l = |
| 2760 | The @var{len} is an expression for the size in bytes of the memory |
| 2761 | region. As with the @var{origin} expression, the expression must |
| 2762 | evaluate to a constant before memory allocation is performed. The |
| 2763 | keyword @code{LENGTH} may be abbreviated to @code{len} or @code{l}. |
| 2764 | |
| 2765 | In the following example, we specify that there are two memory regions |
| 2766 | available for allocation: one starting at @samp{0} for 256 kilobytes, |
| 2767 | and the other starting at @samp{0x40000000} for four megabytes. The |
| 2768 | linker will place into the @samp{rom} memory region every section which |
| 2769 | is not explicitly mapped into a memory region, and is either read-only |
| 2770 | or executable. The linker will place other sections which are not |
| 2771 | explicitly mapped into a memory region into the @samp{ram} memory |
| 2772 | region. |
| 2773 | |
| 2774 | @smallexample |
| 2775 | @group |
| 2776 | MEMORY |
| 2777 | @{ |
| 2778 | rom (rx) : ORIGIN = 0, LENGTH = 256K |
| 2779 | ram (!rx) : org = 0x40000000, l = 4M |
| 2780 | @} |
| 2781 | @end group |
| 2782 | @end smallexample |
| 2783 | |
| 2784 | Once you define a memory region, you can direct the linker to place |
| 2785 | specific output sections into that memory region by using the |
| 2786 | @samp{>@var{region}} output section attribute. For example, if you have |
| 2787 | a memory region named @samp{mem}, you would use @samp{>mem} in the |
| 2788 | output section definition. @xref{Output Section Region}. If no address |
| 2789 | was specified for the output section, the linker will set the address to |
| 2790 | the next available address within the memory region. If the combined |
| 2791 | output sections directed to a memory region are too large for the |
| 2792 | region, the linker will issue an error message. |
| 2793 | |
| 2794 | @node PHDRS |
| 2795 | @section PHDRS Command |
| 2796 | @kindex PHDRS |
| 2797 | @cindex program headers |
| 2798 | @cindex ELF program headers |
| 2799 | @cindex program segments |
| 2800 | @cindex segments, ELF |
| 2801 | The ELF object file format uses @dfn{program headers}, also knows as |
| 2802 | @dfn{segments}. The program headers describe how the program should be |
| 2803 | loaded into memory. You can print them out by using the @code{objdump} |
| 2804 | program with the @samp{-p} option. |
| 2805 | |
| 2806 | When you run an ELF program on a native ELF system, the system loader |
| 2807 | reads the program headers in order to figure out how to load the |
| 2808 | program. This will only work if the program headers are set correctly. |
| 2809 | This manual does not describe the details of how the system loader |
| 2810 | interprets program headers; for more information, see the ELF ABI. |
| 2811 | |
| 2812 | The linker will create reasonable program headers by default. However, |
| 2813 | in some cases, you may need to specify the program headers more |
| 2814 | precisely. You may use the @code{PHDRS} command for this purpose. When |
| 2815 | the linker sees the @code{PHDRS} command in the linker script, it will |
| 2816 | not create any program headers other than the ones specified. |
| 2817 | |
| 2818 | The linker only pays attention to the @code{PHDRS} command when |
| 2819 | generating an ELF output file. In other cases, the linker will simply |
| 2820 | ignore @code{PHDRS}. |
| 2821 | |
| 2822 | This is the syntax of the @code{PHDRS} command. The words @code{PHDRS}, |
| 2823 | @code{FILEHDR}, @code{AT}, and @code{FLAGS} are keywords. |
| 2824 | |
| 2825 | @smallexample |
| 2826 | @group |
| 2827 | PHDRS |
| 2828 | @{ |
| 2829 | @var{name} @var{type} [ FILEHDR ] [ PHDRS ] [ AT ( @var{address} ) ] |
| 2830 | [ FLAGS ( @var{flags} ) ] ; |
| 2831 | @} |
| 2832 | @end group |
| 2833 | @end smallexample |
| 2834 | |
| 2835 | The @var{name} is used only for reference in the @code{SECTIONS} command |
| 2836 | of the linker script. It is not put into the output file. Program |
| 2837 | header names are stored in a separate name space, and will not conflict |
| 2838 | with symbol names, file names, or section names. Each program header |
| 2839 | must have a distinct name. |
| 2840 | |
| 2841 | Certain program header types describe segments of memory which the |
| 2842 | system loader will load from the file. In the linker script, you |
| 2843 | specify the contents of these segments by placing allocatable output |
| 2844 | sections in the segments. You use the @samp{:@var{phdr}} output section |
| 2845 | attribute to place a section in a particular segment. @xref{Output |
| 2846 | Section Phdr}. |
| 2847 | |
| 2848 | It is normal to put certain sections in more than one segment. This |
| 2849 | merely implies that one segment of memory contains another. You may |
| 2850 | repeat @samp{:@var{phdr}}, using it once for each segment which should |
| 2851 | contain the section. |
| 2852 | |
| 2853 | If you place a section in one or more segments using @samp{:@var{phdr}}, |
| 2854 | then the linker will place all subsequent allocatable sections which do |
| 2855 | not specify @samp{:@var{phdr}} in the same segments. This is for |
| 2856 | convenience, since generally a whole set of contiguous sections will be |
| 2857 | placed in a single segment. You can use @code{:NONE} to override the |
| 2858 | default segment and tell the linker to not put the section in any |
| 2859 | segment at all. |
| 2860 | |
| 2861 | @kindex FILEHDR |
| 2862 | @kindex PHDRS |
| 2863 | You may use the @code{FILEHDR} and @code{PHDRS} keywords appear after |
| 2864 | the program header type to further describe the contents of the segment. |
| 2865 | The @code{FILEHDR} keyword means that the segment should include the ELF |
| 2866 | file header. The @code{PHDRS} keyword means that the segment should |
| 2867 | include the ELF program headers themselves. |
| 2868 | |
| 2869 | The @var{type} may be one of the following. The numbers indicate the |
| 2870 | value of the keyword. |
| 2871 | |
| 2872 | @table @asis |
| 2873 | @item @code{PT_NULL} (0) |
| 2874 | Indicates an unused program header. |
| 2875 | |
| 2876 | @item @code{PT_LOAD} (1) |
| 2877 | Indicates that this program header describes a segment to be loaded from |
| 2878 | the file. |
| 2879 | |
| 2880 | @item @code{PT_DYNAMIC} (2) |
| 2881 | Indicates a segment where dynamic linking information can be found. |
| 2882 | |
| 2883 | @item @code{PT_INTERP} (3) |
| 2884 | Indicates a segment where the name of the program interpreter may be |
| 2885 | found. |
| 2886 | |
| 2887 | @item @code{PT_NOTE} (4) |
| 2888 | Indicates a segment holding note information. |
| 2889 | |
| 2890 | @item @code{PT_SHLIB} (5) |
| 2891 | A reserved program header type, defined but not specified by the ELF |
| 2892 | ABI. |
| 2893 | |
| 2894 | @item @code{PT_PHDR} (6) |
| 2895 | Indicates a segment where the program headers may be found. |
| 2896 | |
| 2897 | @item @var{expression} |
| 2898 | An expression giving the numeric type of the program header. This may |
| 2899 | be used for types not defined above. |
| 2900 | @end table |
| 2901 | |
| 2902 | You can specify that a segment should be loaded at a particular address |
| 2903 | in memory by using an @code{AT} expression. This is identical to the |
| 2904 | @code{AT} command used as an output section attribute (@pxref{Output |
| 2905 | Section LMA}). The @code{AT} command for a program header overrides the |
| 2906 | output section attribute. |
| 2907 | |
| 2908 | The linker will normally set the segment flags based on the sections |
| 2909 | which comprise the segment. You may use the @code{FLAGS} keyword to |
| 2910 | explicitly specify the segment flags. The value of @var{flags} must be |
| 2911 | an integer. It is used to set the @code{p_flags} field of the program |
| 2912 | header. |
| 2913 | |
| 2914 | Here is an example of @code{PHDRS}. This shows a typical set of program |
| 2915 | headers used on a native ELF system. |
| 2916 | |
| 2917 | @example |
| 2918 | @group |
| 2919 | PHDRS |
| 2920 | @{ |
| 2921 | headers PT_PHDR PHDRS ; |
| 2922 | interp PT_INTERP ; |
| 2923 | text PT_LOAD FILEHDR PHDRS ; |
| 2924 | data PT_LOAD ; |
| 2925 | dynamic PT_DYNAMIC ; |
| 2926 | @} |
| 2927 | |
| 2928 | SECTIONS |
| 2929 | @{ |
| 2930 | . = SIZEOF_HEADERS; |
| 2931 | .interp : @{ *(.interp) @} :text :interp |
| 2932 | .text : @{ *(.text) @} :text |
| 2933 | .rodata : @{ *(.rodata) @} /* defaults to :text */ |
| 2934 | @dots{} |
| 2935 | . = . + 0x1000; /* move to a new page in memory */ |
| 2936 | .data : @{ *(.data) @} :data |
| 2937 | .dynamic : @{ *(.dynamic) @} :data :dynamic |
| 2938 | @dots{} |
| 2939 | @} |
| 2940 | @end group |
| 2941 | @end example |
| 2942 | |
| 2943 | @node VERSION |
| 2944 | @section VERSION Command |
| 2945 | @kindex VERSION @{script text@} |
| 2946 | @cindex symbol versions |
| 2947 | @cindex version script |
| 2948 | @cindex versions of symbols |
| 2949 | The linker supports symbol versions when using ELF. Symbol versions are |
| 2950 | only useful when using shared libraries. The dynamic linker can use |
| 2951 | symbol versions to select a specific version of a function when it runs |
| 2952 | a program that may have been linked against an earlier version of the |
| 2953 | shared library. |
| 2954 | |
| 2955 | You can include a version script directly in the main linker script, or |
| 2956 | you can supply the version script as an implicit linker script. You can |
| 2957 | also use the @samp{--version-script} linker option. |
| 2958 | |
| 2959 | The syntax of the @code{VERSION} command is simply |
| 2960 | @smallexample |
| 2961 | VERSION @{ version-script-commands @} |
| 2962 | @end smallexample |
| 2963 | |
| 2964 | The format of the version script commands is identical to that used by |
| 2965 | Sun's linker in Solaris 2.5. The version script defines a tree of |
| 2966 | version nodes. You specify the node names and interdependencies in the |
| 2967 | version script. You can specify which symbols are bound to which |
| 2968 | version nodes, and you can reduce a specified set of symbols to local |
| 2969 | scope so that they are not globally visible outside of the shared |
| 2970 | library. |
| 2971 | |
| 2972 | The easiest way to demonstrate the version script language is with a few |
| 2973 | examples. |
| 2974 | |
| 2975 | @smallexample |
| 2976 | VERS_1.1 @{ |
| 2977 | global: |
| 2978 | foo1; |
| 2979 | local: |
| 2980 | old*; |
| 2981 | original*; |
| 2982 | new*; |
| 2983 | @}; |
| 2984 | |
| 2985 | VERS_1.2 @{ |
| 2986 | foo2; |
| 2987 | @} VERS_1.1; |
| 2988 | |
| 2989 | VERS_2.0 @{ |
| 2990 | bar1; bar2; |
| 2991 | @} VERS_1.2; |
| 2992 | @end smallexample |
| 2993 | |
| 2994 | This example version script defines three version nodes. The first |
| 2995 | version node defined is @samp{VERS_1.1}; it has no other dependencies. |
| 2996 | The script binds the symbol @samp{foo1} to @samp{VERS_1.1}. It reduces |
| 2997 | a number of symbols to local scope so that they are not visible outside |
| 2998 | of the shared library. |
| 2999 | |
| 3000 | Next, the version script defines node @samp{VERS_1.2}. This node |
| 3001 | depends upon @samp{VERS_1.1}. The script binds the symbol @samp{foo2} |
| 3002 | to the version node @samp{VERS_1.2}. |
| 3003 | |
| 3004 | Finally, the version script defines node @samp{VERS_2.0}. This node |
| 3005 | depends upon @samp{VERS_1.2}. The scripts binds the symbols @samp{bar1} |
| 3006 | and @samp{bar2} are bound to the version node @samp{VERS_2.0}. |
| 3007 | |
| 3008 | When the linker finds a symbol defined in a library which is not |
| 3009 | specifically bound to a version node, it will effectively bind it to an |
| 3010 | unspecified base version of the library. You can bind all otherwise |
| 3011 | unspecified symbols to a given version node by using @samp{global: *} |
| 3012 | somewhere in the version script. |
| 3013 | |
| 3014 | The names of the version nodes have no specific meaning other than what |
| 3015 | they might suggest to the person reading them. The @samp{2.0} version |
| 3016 | could just as well have appeared in between @samp{1.1} and @samp{1.2}. |
| 3017 | However, this would be a confusing way to write a version script. |
| 3018 | |
| 3019 | When you link an application against a shared library that has versioned |
| 3020 | symbols, the application itself knows which version of each symbol it |
| 3021 | requires, and it also knows which version nodes it needs from each |
| 3022 | shared library it is linked against. Thus at runtime, the dynamic |
| 3023 | loader can make a quick check to make sure that the libraries you have |
| 3024 | linked against do in fact supply all of the version nodes that the |
| 3025 | application will need to resolve all of the dynamic symbols. In this |
| 3026 | way it is possible for the dynamic linker to know with certainty that |
| 3027 | all external symbols that it needs will be resolvable without having to |
| 3028 | search for each symbol reference. |
| 3029 | |
| 3030 | The symbol versioning is in effect a much more sophisticated way of |
| 3031 | doing minor version checking that SunOS does. The fundamental problem |
| 3032 | that is being addressed here is that typically references to external |
| 3033 | functions are bound on an as-needed basis, and are not all bound when |
| 3034 | the application starts up. If a shared library is out of date, a |
| 3035 | required interface may be missing; when the application tries to use |
| 3036 | that interface, it may suddenly and unexpectedly fail. With symbol |
| 3037 | versioning, the user will get a warning when they start their program if |
| 3038 | the libraries being used with the application are too old. |
| 3039 | |
| 3040 | There are several GNU extensions to Sun's versioning approach. The |
| 3041 | first of these is the ability to bind a symbol to a version node in the |
| 3042 | source file where the symbol is defined instead of in the versioning |
| 3043 | script. This was done mainly to reduce the burden on the library |
| 3044 | maintainer. You can do this by putting something like: |
| 3045 | @smallexample |
| 3046 | __asm__(".symver original_foo,foo@@VERS_1.1"); |
| 3047 | @end smallexample |
| 3048 | @noindent |
| 3049 | in the C source file. This renames the function @samp{original_foo} to |
| 3050 | be an alias for @samp{foo} bound to the version node @samp{VERS_1.1}. |
| 3051 | The @samp{local:} directive can be used to prevent the symbol |
| 3052 | @samp{original_foo} from being exported. |
| 3053 | |
| 3054 | The second GNU extension is to allow multiple versions of the same |
| 3055 | function to appear in a given shared library. In this way you can make |
| 3056 | an incompatible change to an interface without increasing the major |
| 3057 | version number of the shared library, while still allowing applications |
| 3058 | linked against the old interface to continue to function. |
| 3059 | |
| 3060 | To do this, you must use multiple @samp{.symver} directives in the |
| 3061 | source file. Here is an example: |
| 3062 | |
| 3063 | @smallexample |
| 3064 | __asm__(".symver original_foo,foo@@"); |
| 3065 | __asm__(".symver old_foo,foo@@VERS_1.1"); |
| 3066 | __asm__(".symver old_foo1,foo@@VERS_1.2"); |
| 3067 | __asm__(".symver new_foo,foo@@@@VERS_2.0"); |
| 3068 | @end smallexample |
| 3069 | |
| 3070 | In this example, @samp{foo@@} represents the symbol @samp{foo} bound to the |
| 3071 | unspecified base version of the symbol. The source file that contains this |
| 3072 | example would define 4 C functions: @samp{original_foo}, @samp{old_foo}, |
| 3073 | @samp{old_foo1}, and @samp{new_foo}. |
| 3074 | |
| 3075 | When you have multiple definitions of a given symbol, there needs to be |
| 3076 | some way to specify a default version to which external references to |
| 3077 | this symbol will be bound. You can do this with the |
| 3078 | @samp{foo@@@@VERS_2.0} type of @samp{.symver} directive. You can only |
| 3079 | declare one version of a symbol as the default in this manner; otherwise |
| 3080 | you would effectively have multiple definitions of the same symbol. |
| 3081 | |
| 3082 | If you wish to bind a reference to a specific version of the symbol |
| 3083 | within the shared library, you can use the aliases of convenience |
| 3084 | (i.e. @samp{old_foo}), or you can use the @samp{.symver} directive to |
| 3085 | specifically bind to an external version of the function in question. |
| 3086 | |
| 3087 | @node Expressions |
| 3088 | @section Expressions in Linker Scripts |
| 3089 | @cindex expressions |
| 3090 | @cindex arithmetic |
| 3091 | The syntax for expressions in the linker script language is identical to |
| 3092 | that of C expressions. All expressions are evaluated as integers. All |
| 3093 | expressions are evaluated in the same size, which is 32 bits if both the |
| 3094 | host and target are 32 bits, and is otherwise 64 bits. |
| 3095 | |
| 3096 | You can use and set symbol values in expressions. |
| 3097 | |
| 3098 | The linker defines several special purpose builtin functions for use in |
| 3099 | expressions. |
| 3100 | |
| 3101 | @menu |
| 3102 | * Constants:: Constants |
| 3103 | * Symbols:: Symbol Names |
| 3104 | * Location Counter:: The Location Counter |
| 3105 | * Operators:: Operators |
| 3106 | * Evaluation:: Evaluation |
| 3107 | * Expression Section:: The Section of an Expression |
| 3108 | * Builtin Functions:: Builtin Functions |
| 3109 | @end menu |
| 3110 | |
| 3111 | @node Constants |
| 3112 | @subsection Constants |
| 3113 | @cindex integer notation |
| 3114 | @cindex constants in linker scripts |
| 3115 | All constants are integers. |
| 3116 | |
| 3117 | As in C, the linker considers an integer beginning with @samp{0} to be |
| 3118 | octal, and an integer beginning with @samp{0x} or @samp{0X} to be |
| 3119 | hexadecimal. The linker considers other integers to be decimal. |
| 3120 | |
| 3121 | @cindex scaled integers |
| 3122 | @cindex K and M integer suffixes |
| 3123 | @cindex M and K integer suffixes |
| 3124 | @cindex suffixes for integers |
| 3125 | @cindex integer suffixes |
| 3126 | In addition, you can use the suffixes @code{K} and @code{M} to scale a |
| 3127 | constant by |
| 3128 | @c TEXI2ROFF-KILL |
| 3129 | @ifinfo |
| 3130 | @c END TEXI2ROFF-KILL |
| 3131 | @code{1024} or @code{1024*1024} |
| 3132 | @c TEXI2ROFF-KILL |
| 3133 | @end ifinfo |
| 3134 | @tex |
| 3135 | ${\rm 1024}$ or ${\rm 1024}^2$ |
| 3136 | @end tex |
| 3137 | @c END TEXI2ROFF-KILL |
| 3138 | respectively. For example, the following all refer to the same quantity: |
| 3139 | @smallexample |
| 3140 | _fourk_1 = 4K; |
| 3141 | _fourk_2 = 4096; |
| 3142 | _fourk_3 = 0x1000; |
| 3143 | @end smallexample |
| 3144 | |
| 3145 | @node Symbols |
| 3146 | @subsection Symbol Names |
| 3147 | @cindex symbol names |
| 3148 | @cindex names |
| 3149 | @cindex quoted symbol names |
| 3150 | @kindex " |
| 3151 | Unless quoted, symbol names start with a letter, underscore, or period |
| 3152 | and may include letters, digits, underscores, periods, and hyphens. |
| 3153 | Unquoted symbol names must not conflict with any keywords. You can |
| 3154 | specify a symbol which contains odd characters or has the same name as a |
| 3155 | keyword by surrounding the symbol name in double quotes: |
| 3156 | @smallexample |
| 3157 | "SECTION" = 9; |
| 3158 | "with a space" = "also with a space" + 10; |
| 3159 | @end smallexample |
| 3160 | |
| 3161 | Since symbols can contain many non-alphabetic characters, it is safest |
| 3162 | to delimit symbols with spaces. For example, @samp{A-B} is one symbol, |
| 3163 | whereas @samp{A - B} is an expression involving subtraction. |
| 3164 | |
| 3165 | @node Location Counter |
| 3166 | @subsection The Location Counter |
| 3167 | @kindex . |
| 3168 | @cindex dot |
| 3169 | @cindex location counter |
| 3170 | @cindex current output location |
| 3171 | The special linker variable @dfn{dot} @samp{.} always contains the |
| 3172 | current output location counter. Since the @code{.} always refers to a |
| 3173 | location in an output section, it may only appear in an expression |
| 3174 | within a @code{SECTIONS} command. The @code{.} symbol may appear |
| 3175 | anywhere that an ordinary symbol is allowed in an expression. |
| 3176 | |
| 3177 | @cindex holes |
| 3178 | Assigning a value to @code{.} will cause the location counter to be |
| 3179 | moved. This may be used to create holes in the output section. The |
| 3180 | location counter may never be moved backwards. |
| 3181 | |
| 3182 | @smallexample |
| 3183 | SECTIONS |
| 3184 | @{ |
| 3185 | output : |
| 3186 | @{ |
| 3187 | file1(.text) |
| 3188 | . = . + 1000; |
| 3189 | file2(.text) |
| 3190 | . += 1000; |
| 3191 | file3(.text) |
| 3192 | @} = 0x1234; |
| 3193 | @} |
| 3194 | @end smallexample |
| 3195 | @noindent |
| 3196 | In the previous example, the @samp{.text} section from @file{file1} is |
| 3197 | located at the beginning of the output section @samp{output}. It is |
| 3198 | followed by a 1000 byte gap. Then the @samp{.text} section from |
| 3199 | @file{file2} appears, also with a 1000 byte gap following before the |
| 3200 | @samp{.text} section from @file{file3}. The notation @samp{= 0x1234} |
| 3201 | specifies what data to write in the gaps (@pxref{Output Section Fill}). |
| 3202 | |
| 3203 | @need 2000 |
| 3204 | @node Operators |
| 3205 | @subsection Operators |
| 3206 | @cindex operators for arithmetic |
| 3207 | @cindex arithmetic operators |
| 3208 | @cindex precedence in expressions |
| 3209 | The linker recognizes the standard C set of arithmetic operators, with |
| 3210 | the standard bindings and precedence levels: |
| 3211 | @c TEXI2ROFF-KILL |
| 3212 | @ifinfo |
| 3213 | @c END TEXI2ROFF-KILL |
| 3214 | @smallexample |
| 3215 | precedence associativity Operators Notes |
| 3216 | (highest) |
| 3217 | 1 left ! - ~ (1) |
| 3218 | 2 left * / % |
| 3219 | 3 left + - |
| 3220 | 4 left >> << |
| 3221 | 5 left == != > < <= >= |
| 3222 | 6 left & |
| 3223 | 7 left | |
| 3224 | 8 left && |
| 3225 | 9 left || |
| 3226 | 10 right ? : |
| 3227 | 11 right &= += -= *= /= (2) |
| 3228 | (lowest) |
| 3229 | @end smallexample |
| 3230 | Notes: |
| 3231 | (1) Prefix operators |
| 3232 | (2) @xref{Assignments}. |
| 3233 | @c TEXI2ROFF-KILL |
| 3234 | @end ifinfo |
| 3235 | @tex |
| 3236 | \vskip \baselineskip |
| 3237 | %"lispnarrowing" is the extra indent used generally for smallexample |
| 3238 | \hskip\lispnarrowing\vbox{\offinterlineskip |
| 3239 | \hrule |
| 3240 | \halign |
| 3241 | {\vrule#&\strut\hfil\ #\ \hfil&\vrule#&\strut\hfil\ #\ \hfil&\vrule#&\strut\hfil\ {\tt #}\ \hfil&\vrule#\cr |
| 3242 | height2pt&\omit&&\omit&&\omit&\cr |
| 3243 | &Precedence&& Associativity &&{\rm Operators}&\cr |
| 3244 | height2pt&\omit&&\omit&&\omit&\cr |
| 3245 | \noalign{\hrule} |
| 3246 | height2pt&\omit&&\omit&&\omit&\cr |
| 3247 | &highest&&&&&\cr |
| 3248 | % '176 is tilde, '~' in tt font |
| 3249 | &1&&left&&\qquad- \char'176\ !\qquad\dag&\cr |
| 3250 | &2&&left&&* / \%&\cr |
| 3251 | &3&&left&&+ -&\cr |
| 3252 | &4&&left&&>> <<&\cr |
| 3253 | &5&&left&&== != > < <= >=&\cr |
| 3254 | &6&&left&&\&&\cr |
| 3255 | &7&&left&&|&\cr |
| 3256 | &8&&left&&{\&\&}&\cr |
| 3257 | &9&&left&&||&\cr |
| 3258 | &10&&right&&? :&\cr |
| 3259 | &11&&right&&\qquad\&= += -= *= /=\qquad\ddag&\cr |
| 3260 | &lowest&&&&&\cr |
| 3261 | height2pt&\omit&&\omit&&\omit&\cr} |
| 3262 | \hrule} |
| 3263 | @end tex |
| 3264 | @iftex |
| 3265 | { |
| 3266 | @obeylines@parskip=0pt@parindent=0pt |
| 3267 | @dag@quad Prefix operators. |
| 3268 | @ddag@quad @xref{Assignments}. |
| 3269 | } |
| 3270 | @end iftex |
| 3271 | @c END TEXI2ROFF-KILL |
| 3272 | |
| 3273 | @node Evaluation |
| 3274 | @subsection Evaluation |
| 3275 | @cindex lazy evaluation |
| 3276 | @cindex expression evaluation order |
| 3277 | The linker evaluates expressions lazily. It only computes the value of |
| 3278 | an expression when absolutely necessary. |
| 3279 | |
| 3280 | The linker needs some information, such as the value of the start |
| 3281 | address of the first section, and the origins and lengths of memory |
| 3282 | regions, in order to do any linking at all. These values are computed |
| 3283 | as soon as possible when the linker reads in the linker script. |
| 3284 | |
| 3285 | However, other values (such as symbol values) are not known or needed |
| 3286 | until after storage allocation. Such values are evaluated later, when |
| 3287 | other information (such as the sizes of output sections) is available |
| 3288 | for use in the symbol assignment expression. |
| 3289 | |
| 3290 | The sizes of sections cannot be known until after allocation, so |
| 3291 | assignments dependent upon these are not performed until after |
| 3292 | allocation. |
| 3293 | |
| 3294 | Some expressions, such as those depending upon the location counter |
| 3295 | @samp{.}, must be evaluated during section allocation. |
| 3296 | |
| 3297 | If the result of an expression is required, but the value is not |
| 3298 | available, then an error results. For example, a script like the |
| 3299 | following |
| 3300 | @smallexample |
| 3301 | @group |
| 3302 | SECTIONS |
| 3303 | @{ |
| 3304 | .text 9+this_isnt_constant : |
| 3305 | @{ *(.text) @} |
| 3306 | @} |
| 3307 | @end group |
| 3308 | @end smallexample |
| 3309 | @noindent |
| 3310 | will cause the error message @samp{non constant expression for initial |
| 3311 | address}. |
| 3312 | |
| 3313 | @node Expression Section |
| 3314 | @subsection The Section of an Expression |
| 3315 | @cindex expression sections |
| 3316 | @cindex absolute expressions |
| 3317 | @cindex relative expressions |
| 3318 | @cindex absolute and relocatable symbols |
| 3319 | @cindex relocatable and absolute symbols |
| 3320 | @cindex symbols, relocatable and absolute |
| 3321 | When the linker evaluates an expression, the result is either absolute |
| 3322 | or relative to some section. A relative expression is expressed as a |
| 3323 | fixed offset from the base of a section. |
| 3324 | |
| 3325 | The position of the expression within the linker script determines |
| 3326 | whether it is absolute or relative. An expression which appears within |
| 3327 | an output section definition is relative to the base of the output |
| 3328 | section. An expression which appears elsewhere will be absolute. |
| 3329 | |
| 3330 | A symbol set to a relative expression will be relocatable if you request |
| 3331 | relocatable output using the @samp{-r} option. That means that a |
| 3332 | further link operation may change the value of the symbol. The symbol's |
| 3333 | section will be the section of the relative expression. |
| 3334 | |
| 3335 | A symbol set to an absolute expression will retain the same value |
| 3336 | through any further link operation. The symbol will be absolute, and |
| 3337 | will not have any particular associated section. |
| 3338 | |
| 3339 | You can use the builtin function @code{ABSOLUTE} to force an expression |
| 3340 | to be absolute when it would otherwise be relative. For example, to |
| 3341 | create an absolute symbol set to the address of the end of the output |
| 3342 | section @samp{.data}: |
| 3343 | @smallexample |
| 3344 | SECTIONS |
| 3345 | @{ |
| 3346 | .data : @{ *(.data) _edata = ABSOLUTE(.); @} |
| 3347 | @} |
| 3348 | @end smallexample |
| 3349 | @noindent |
| 3350 | If @samp{ABSOLUTE} were not used, @samp{_edata} would be relative to the |
| 3351 | @samp{.data} section. |
| 3352 | |
| 3353 | @node Builtin Functions |
| 3354 | @subsection Builtin Functions |
| 3355 | @cindex functions in expressions |
| 3356 | The linker script language includes a number of builtin functions for |
| 3357 | use in linker script expressions. |
| 3358 | |
| 3359 | @table @code |
| 3360 | @item ABSOLUTE(@var{exp}) |
| 3361 | @kindex ABSOLUTE(@var{exp}) |
| 3362 | @cindex expression, absolute |
| 3363 | Return the absolute (non-relocatable, as opposed to non-negative) value |
| 3364 | of the expression @var{exp}. Primarily useful to assign an absolute |
| 3365 | value to a symbol within a section definition, where symbol values are |
| 3366 | normally section relative. @xref{Expression Section}. |
| 3367 | |
| 3368 | @item ADDR(@var{section}) |
| 3369 | @kindex ADDR(@var{section}) |
| 3370 | @cindex section address in expression |
| 3371 | Return the absolute address (the VMA) of the named @var{section}. Your |
| 3372 | script must previously have defined the location of that section. In |
| 3373 | the following example, @code{symbol_1} and @code{symbol_2} are assigned |
| 3374 | identical values: |
| 3375 | @smallexample |
| 3376 | @group |
| 3377 | SECTIONS @{ @dots{} |
| 3378 | .output1 : |
| 3379 | @{ |
| 3380 | start_of_output_1 = ABSOLUTE(.); |
| 3381 | @dots{} |
| 3382 | @} |
| 3383 | .output : |
| 3384 | @{ |
| 3385 | symbol_1 = ADDR(.output1); |
| 3386 | symbol_2 = start_of_output_1; |
| 3387 | @} |
| 3388 | @dots{} @} |
| 3389 | @end group |
| 3390 | @end smallexample |
| 3391 | |
| 3392 | @item ALIGN(@var{exp}) |
| 3393 | @kindex ALIGN(@var{exp}) |
| 3394 | @cindex round up location counter |
| 3395 | @cindex align location counter |
| 3396 | Return the location counter (@code{.}) aligned to the next @var{exp} |
| 3397 | boundary. @var{exp} must be an expression whose value is a power of |
| 3398 | two. This is equivalent to |
| 3399 | @smallexample |
| 3400 | (. + @var{exp} - 1) & ~(@var{exp} - 1) |
| 3401 | @end smallexample |
| 3402 | |
| 3403 | @code{ALIGN} doesn't change the value of the location counter---it just |
| 3404 | does arithmetic on it. Here is an example which aligns the output |
| 3405 | @code{.data} section to the next @code{0x2000} byte boundary after the |
| 3406 | preceding section and sets a variable within the section to the next |
| 3407 | @code{0x8000} boundary after the input sections: |
| 3408 | @smallexample |
| 3409 | @group |
| 3410 | SECTIONS @{ @dots{} |
| 3411 | .data ALIGN(0x2000): @{ |
| 3412 | *(.data) |
| 3413 | variable = ALIGN(0x8000); |
| 3414 | @} |
| 3415 | @dots{} @} |
| 3416 | @end group |
| 3417 | @end smallexample |
| 3418 | @noindent |
| 3419 | The first use of @code{ALIGN} in this example specifies the location of |
| 3420 | a section because it is used as the optional @var{address} attribute of |
| 3421 | a section definition (@pxref{Output Section Address}). The second use |
| 3422 | of @code{ALIGN} is used to defines the value of a symbol. |
| 3423 | |
| 3424 | The builtin function @code{NEXT} is closely related to @code{ALIGN}. |
| 3425 | |
| 3426 | @item BLOCK(@var{exp}) |
| 3427 | @kindex BLOCK(@var{exp}) |
| 3428 | This is a synonym for @code{ALIGN}, for compatibility with older linker |
| 3429 | scripts. It is most often seen when setting the address of an output |
| 3430 | section. |
| 3431 | |
| 3432 | @item DEFINED(@var{symbol}) |
| 3433 | @kindex DEFINED(@var{symbol}) |
| 3434 | @cindex symbol defaults |
| 3435 | Return 1 if @var{symbol} is in the linker global symbol table and is |
| 3436 | defined, otherwise return 0. You can use this function to provide |
| 3437 | default values for symbols. For example, the following script fragment |
| 3438 | shows how to set a global symbol @samp{begin} to the first location in |
| 3439 | the @samp{.text} section---but if a symbol called @samp{begin} already |
| 3440 | existed, its value is preserved: |
| 3441 | |
| 3442 | @smallexample |
| 3443 | @group |
| 3444 | SECTIONS @{ @dots{} |
| 3445 | .text : @{ |
| 3446 | begin = DEFINED(begin) ? begin : . ; |
| 3447 | @dots{} |
| 3448 | @} |
| 3449 | @dots{} |
| 3450 | @} |
| 3451 | @end group |
| 3452 | @end smallexample |
| 3453 | |
| 3454 | @item LOADADDR(@var{section}) |
| 3455 | @kindex LOADADDR(@var{section}) |
| 3456 | @cindex section load address in expression |
| 3457 | Return the absolute LMA of the named @var{section}. This is normally |
| 3458 | the same as @code{ADDR}, but it may be different if the @code{AT} |
| 3459 | attribute is used in the output section definition (@pxref{Output |
| 3460 | Section LMA}). |
| 3461 | |
| 3462 | @kindex MAX |
| 3463 | @item MAX(@var{exp1}, @var{exp2}) |
| 3464 | Returns the maximum of @var{exp1} and @var{exp2}. |
| 3465 | |
| 3466 | @kindex MIN |
| 3467 | @item MIN(@var{exp1}, @var{exp2}) |
| 3468 | Returns the minimum of @var{exp1} and @var{exp2}. |
| 3469 | |
| 3470 | @item NEXT(@var{exp}) |
| 3471 | @kindex NEXT(@var{exp}) |
| 3472 | @cindex unallocated address, next |
| 3473 | Return the next unallocated address that is a multiple of @var{exp}. |
| 3474 | This function is closely related to @code{ALIGN(@var{exp})}; unless you |
| 3475 | use the @code{MEMORY} command to define discontinuous memory for the |
| 3476 | output file, the two functions are equivalent. |
| 3477 | |
| 3478 | @item SIZEOF(@var{section}) |
| 3479 | @kindex SIZEOF(@var{section}) |
| 3480 | @cindex section size |
| 3481 | Return the size in bytes of the named @var{section}, if that section has |
| 3482 | been allocated. If the section has not been allocated when this is |
| 3483 | evaluated, the linker will report an error. In the following example, |
| 3484 | @code{symbol_1} and @code{symbol_2} are assigned identical values: |
| 3485 | @smallexample |
| 3486 | @group |
| 3487 | SECTIONS@{ @dots{} |
| 3488 | .output @{ |
| 3489 | .start = . ; |
| 3490 | @dots{} |
| 3491 | .end = . ; |
| 3492 | @} |
| 3493 | symbol_1 = .end - .start ; |
| 3494 | symbol_2 = SIZEOF(.output); |
| 3495 | @dots{} @} |
| 3496 | @end group |
| 3497 | @end smallexample |
| 3498 | |
| 3499 | @item SIZEOF_HEADERS |
| 3500 | @itemx sizeof_headers |
| 3501 | @kindex SIZEOF_HEADERS |
| 3502 | @cindex header size |
| 3503 | Return the size in bytes of the output file's headers. This is |
| 3504 | information which appears at the start of the output file. You can use |
| 3505 | this number when setting the start address of the first section, if you |
| 3506 | choose, to facilitate paging. |
| 3507 | |
| 3508 | @cindex not enough room for program headers |
| 3509 | @cindex program headers, not enough room |
| 3510 | When producing an ELF output file, if the linker script uses the |
| 3511 | @code{SIZEOF_HEADERS} builtin function, the linker must compute the |
| 3512 | number of program headers before it has determined all the section |
| 3513 | addresses and sizes. If the linker later discovers that it needs |
| 3514 | additional program headers, it will report an error @samp{not enough |
| 3515 | room for program headers}. To avoid this error, you must avoid using |
| 3516 | the @code{SIZEOF_HEADERS} function, or you must rework your linker |
| 3517 | script to avoid forcing the linker to use additional program headers, or |
| 3518 | you must define the program headers yourself using the @code{PHDRS} |
| 3519 | command (@pxref{PHDRS}). |
| 3520 | @end table |
| 3521 | |
| 3522 | @node Implicit Linker Scripts |
| 3523 | @section Implicit Linker Scripts |
| 3524 | @cindex implicit linker scripts |
| 3525 | If you specify a linker input file which the linker can not recognize as |
| 3526 | an object file or an archive file, it will try to read the file as a |
| 3527 | linker script. If the file can not be parsed as a linker script, the |
| 3528 | linker will report an error. |
| 3529 | |
| 3530 | An implicit linker script will not replace the default linker script. |
| 3531 | |
| 3532 | Typically an implicit linker script would contain only symbol |
| 3533 | assignments, or the @code{INPUT}, @code{GROUP}, or @code{VERSION} |
| 3534 | commands. |
| 3535 | |
| 3536 | Any input files read because of an implicit linker script will be read |
| 3537 | at the position in the command line where the implicit linker script was |
| 3538 | read. This can affect archive searching. |
| 3539 | |
| 3540 | @ifset GENERIC |
| 3541 | @node Machine Dependent |
| 3542 | @chapter Machine Dependent Features |
| 3543 | |
| 3544 | @cindex machine dependencies |
| 3545 | @code{ld} has additional features on some platforms; the following |
| 3546 | sections describe them. Machines where @code{ld} has no additional |
| 3547 | functionality are not listed. |
| 3548 | |
| 3549 | @menu |
| 3550 | * H8/300:: @code{ld} and the H8/300 |
| 3551 | * i960:: @code{ld} and the Intel 960 family |
| 3552 | * ARM:: @code{ld} and the ARM family |
| 3553 | @end menu |
| 3554 | @end ifset |
| 3555 | |
| 3556 | @c FIXME! This could use @raisesections/@lowersections, but there seems to be a conflict |
| 3557 | @c between those and node-defaulting. |
| 3558 | @ifset H8300 |
| 3559 | @ifclear GENERIC |
| 3560 | @raisesections |
| 3561 | @end ifclear |
| 3562 | |
| 3563 | @node H8/300 |
| 3564 | @section @code{ld} and the H8/300 |
| 3565 | |
| 3566 | @cindex H8/300 support |
| 3567 | For the H8/300, @code{ld} can perform these global optimizations when |
| 3568 | you specify the @samp{--relax} command-line option. |
| 3569 | |
| 3570 | @table @emph |
| 3571 | @cindex relaxing on H8/300 |
| 3572 | @item relaxing address modes |
| 3573 | @code{ld} finds all @code{jsr} and @code{jmp} instructions whose |
| 3574 | targets are within eight bits, and turns them into eight-bit |
| 3575 | program-counter relative @code{bsr} and @code{bra} instructions, |
| 3576 | respectively. |
| 3577 | |
| 3578 | @cindex synthesizing on H8/300 |
| 3579 | @item synthesizing instructions |
| 3580 | @c FIXME: specifically mov.b, or any mov instructions really? |
| 3581 | @code{ld} finds all @code{mov.b} instructions which use the |
| 3582 | sixteen-bit absolute address form, but refer to the top |
| 3583 | page of memory, and changes them to use the eight-bit address form. |
| 3584 | (That is: the linker turns @samp{mov.b @code{@@}@var{aa}:16} into |
| 3585 | @samp{mov.b @code{@@}@var{aa}:8} whenever the address @var{aa} is in the |
| 3586 | top page of memory). |
| 3587 | @end table |
| 3588 | |
| 3589 | @ifclear GENERIC |
| 3590 | @lowersections |
| 3591 | @end ifclear |
| 3592 | @end ifset |
| 3593 | |
| 3594 | @ifclear GENERIC |
| 3595 | @ifset Hitachi |
| 3596 | @c This stuff is pointless to say unless you're especially concerned |
| 3597 | @c with Hitachi chips; don't enable it for generic case, please. |
| 3598 | @node Hitachi |
| 3599 | @chapter @code{ld} and other Hitachi chips |
| 3600 | |
| 3601 | @code{ld} also supports the H8/300H, the H8/500, and the Hitachi SH. No |
| 3602 | special features, commands, or command-line options are required for |
| 3603 | these chips. |
| 3604 | @end ifset |
| 3605 | @end ifclear |
| 3606 | |
| 3607 | @ifset I960 |
| 3608 | @ifclear GENERIC |
| 3609 | @raisesections |
| 3610 | @end ifclear |
| 3611 | |
| 3612 | @node i960 |
| 3613 | @section @code{ld} and the Intel 960 family |
| 3614 | |
| 3615 | @cindex i960 support |
| 3616 | |
| 3617 | You can use the @samp{-A@var{architecture}} command line option to |
| 3618 | specify one of the two-letter names identifying members of the 960 |
| 3619 | family; the option specifies the desired output target, and warns of any |
| 3620 | incompatible instructions in the input files. It also modifies the |
| 3621 | linker's search strategy for archive libraries, to support the use of |
| 3622 | libraries specific to each particular architecture, by including in the |
| 3623 | search loop names suffixed with the string identifying the architecture. |
| 3624 | |
| 3625 | For example, if your @code{ld} command line included @w{@samp{-ACA}} as |
| 3626 | well as @w{@samp{-ltry}}, the linker would look (in its built-in search |
| 3627 | paths, and in any paths you specify with @samp{-L}) for a library with |
| 3628 | the names |
| 3629 | |
| 3630 | @smallexample |
| 3631 | @group |
| 3632 | try |
| 3633 | libtry.a |
| 3634 | tryca |
| 3635 | libtryca.a |
| 3636 | @end group |
| 3637 | @end smallexample |
| 3638 | |
| 3639 | @noindent |
| 3640 | The first two possibilities would be considered in any event; the last |
| 3641 | two are due to the use of @w{@samp{-ACA}}. |
| 3642 | |
| 3643 | You can meaningfully use @samp{-A} more than once on a command line, since |
| 3644 | the 960 architecture family allows combination of target architectures; each |
| 3645 | use will add another pair of name variants to search for when @w{@samp{-l}} |
| 3646 | specifies a library. |
| 3647 | |
| 3648 | @cindex @code{--relax} on i960 |
| 3649 | @cindex relaxing on i960 |
| 3650 | @code{ld} supports the @samp{--relax} option for the i960 family. If |
| 3651 | you specify @samp{--relax}, @code{ld} finds all @code{balx} and |
| 3652 | @code{calx} instructions whose targets are within 24 bits, and turns |
| 3653 | them into 24-bit program-counter relative @code{bal} and @code{cal} |
| 3654 | instructions, respectively. @code{ld} also turns @code{cal} |
| 3655 | instructions into @code{bal} instructions when it determines that the |
| 3656 | target subroutine is a leaf routine (that is, the target subroutine does |
| 3657 | not itself call any subroutines). |
| 3658 | |
| 3659 | @ifclear GENERIC |
| 3660 | @lowersections |
| 3661 | @end ifclear |
| 3662 | @end ifset |
| 3663 | |
| 3664 | @ifclear GENERIC |
| 3665 | @raisesections |
| 3666 | @end ifclear |
| 3667 | |
| 3668 | @node ARM |
| 3669 | @section @code{ld}'s support for interworking between ARM and Thumb code |
| 3670 | |
| 3671 | @cindex ARM interworking support |
| 3672 | @cindex --support-old-code |
| 3673 | For the ARM, @code{ld} will generate code stubs to allow functions calls |
| 3674 | betweem ARM and Thumb code. These stubs only work with code that has |
| 3675 | been compiled and assembled with the @samp{-mthumb-interwork} command |
| 3676 | line option. If it is necessary to link with old ARM object files or |
| 3677 | libraries, which have not been compiled with the -mthumb-interwork |
| 3678 | option then the @samp{--support-old-code} command line switch should be |
| 3679 | given to the linker. This will make it generate larger stub functions |
| 3680 | which will work with non-interworking aware ARM code. Note, however, |
| 3681 | the linker does not support generating stubs for function calls to |
| 3682 | non-interworking aware Thumb code. |
| 3683 | |
| 3684 | @ifclear GENERIC |
| 3685 | @lowersections |
| 3686 | @end ifclear |
| 3687 | |
| 3688 | @ifclear SingleFormat |
| 3689 | @node BFD |
| 3690 | @chapter BFD |
| 3691 | |
| 3692 | @cindex back end |
| 3693 | @cindex object file management |
| 3694 | @cindex object formats available |
| 3695 | @kindex objdump -i |
| 3696 | The linker accesses object and archive files using the BFD libraries. |
| 3697 | These libraries allow the linker to use the same routines to operate on |
| 3698 | object files whatever the object file format. A different object file |
| 3699 | format can be supported simply by creating a new BFD back end and adding |
| 3700 | it to the library. To conserve runtime memory, however, the linker and |
| 3701 | associated tools are usually configured to support only a subset of the |
| 3702 | object file formats available. You can use @code{objdump -i} |
| 3703 | (@pxref{objdump,,objdump,binutils.info,The GNU Binary Utilities}) to |
| 3704 | list all the formats available for your configuration. |
| 3705 | |
| 3706 | @cindex BFD requirements |
| 3707 | @cindex requirements for BFD |
| 3708 | As with most implementations, BFD is a compromise between |
| 3709 | several conflicting requirements. The major factor influencing |
| 3710 | BFD design was efficiency: any time used converting between |
| 3711 | formats is time which would not have been spent had BFD not |
| 3712 | been involved. This is partly offset by abstraction payback; since |
| 3713 | BFD simplifies applications and back ends, more time and care |
| 3714 | may be spent optimizing algorithms for a greater speed. |
| 3715 | |
| 3716 | One minor artifact of the BFD solution which you should bear in |
| 3717 | mind is the potential for information loss. There are two places where |
| 3718 | useful information can be lost using the BFD mechanism: during |
| 3719 | conversion and during output. @xref{BFD information loss}. |
| 3720 | |
| 3721 | @menu |
| 3722 | * BFD outline:: How it works: an outline of BFD |
| 3723 | @end menu |
| 3724 | |
| 3725 | @node BFD outline |
| 3726 | @section How it works: an outline of BFD |
| 3727 | @cindex opening object files |
| 3728 | @include bfdsumm.texi |
| 3729 | @end ifclear |
| 3730 | |
| 3731 | @node Reporting Bugs |
| 3732 | @chapter Reporting Bugs |
| 3733 | @cindex bugs in @code{ld} |
| 3734 | @cindex reporting bugs in @code{ld} |
| 3735 | |
| 3736 | Your bug reports play an essential role in making @code{ld} reliable. |
| 3737 | |
| 3738 | Reporting a bug may help you by bringing a solution to your problem, or |
| 3739 | it may not. But in any case the principal function of a bug report is |
| 3740 | to help the entire community by making the next version of @code{ld} |
| 3741 | work better. Bug reports are your contribution to the maintenance of |
| 3742 | @code{ld}. |
| 3743 | |
| 3744 | In order for a bug report to serve its purpose, you must include the |
| 3745 | information that enables us to fix the bug. |
| 3746 | |
| 3747 | @menu |
| 3748 | * Bug Criteria:: Have you found a bug? |
| 3749 | * Bug Reporting:: How to report bugs |
| 3750 | @end menu |
| 3751 | |
| 3752 | @node Bug Criteria |
| 3753 | @section Have you found a bug? |
| 3754 | @cindex bug criteria |
| 3755 | |
| 3756 | If you are not sure whether you have found a bug, here are some guidelines: |
| 3757 | |
| 3758 | @itemize @bullet |
| 3759 | @cindex fatal signal |
| 3760 | @cindex linker crash |
| 3761 | @cindex crash of linker |
| 3762 | @item |
| 3763 | If the linker gets a fatal signal, for any input whatever, that is a |
| 3764 | @code{ld} bug. Reliable linkers never crash. |
| 3765 | |
| 3766 | @cindex error on valid input |
| 3767 | @item |
| 3768 | If @code{ld} produces an error message for valid input, that is a bug. |
| 3769 | |
| 3770 | @cindex invalid input |
| 3771 | @item |
| 3772 | If @code{ld} does not produce an error message for invalid input, that |
| 3773 | may be a bug. In the general case, the linker can not verify that |
| 3774 | object files are correct. |
| 3775 | |
| 3776 | @item |
| 3777 | If you are an experienced user of linkers, your suggestions for |
| 3778 | improvement of @code{ld} are welcome in any case. |
| 3779 | @end itemize |
| 3780 | |
| 3781 | @node Bug Reporting |
| 3782 | @section How to report bugs |
| 3783 | @cindex bug reports |
| 3784 | @cindex @code{ld} bugs, reporting |
| 3785 | |
| 3786 | A number of companies and individuals offer support for @sc{gnu} |
| 3787 | products. If you obtained @code{ld} from a support organization, we |
| 3788 | recommend you contact that organization first. |
| 3789 | |
| 3790 | You can find contact information for many support companies and |
| 3791 | individuals in the file @file{etc/SERVICE} in the @sc{gnu} Emacs |
| 3792 | distribution. |
| 3793 | |
| 3794 | Otherwise, send bug reports for @code{ld} to |
| 3795 | @samp{bug-gnu-utils@@gnu.org}. |
| 3796 | |
| 3797 | The fundamental principle of reporting bugs usefully is this: |
| 3798 | @strong{report all the facts}. If you are not sure whether to state a |
| 3799 | fact or leave it out, state it! |
| 3800 | |
| 3801 | Often people omit facts because they think they know what causes the |
| 3802 | problem and assume that some details do not matter. Thus, you might |
| 3803 | assume that the name of a symbol you use in an example does not matter. |
| 3804 | Well, probably it does not, but one cannot be sure. Perhaps the bug is |
| 3805 | a stray memory reference which happens to fetch from the location where |
| 3806 | that name is stored in memory; perhaps, if the name were different, the |
| 3807 | contents of that location would fool the linker into doing the right |
| 3808 | thing despite the bug. Play it safe and give a specific, complete |
| 3809 | example. That is the easiest thing for you to do, and the most helpful. |
| 3810 | |
| 3811 | Keep in mind that the purpose of a bug report is to enable us to fix the bug if |
| 3812 | it is new to us. Therefore, always write your bug reports on the assumption |
| 3813 | that the bug has not been reported previously. |
| 3814 | |
| 3815 | Sometimes people give a few sketchy facts and ask, ``Does this ring a |
| 3816 | bell?'' Those bug reports are useless, and we urge everyone to |
| 3817 | @emph{refuse to respond to them} except to chide the sender to report |
| 3818 | bugs properly. |
| 3819 | |
| 3820 | To enable us to fix the bug, you should include all these things: |
| 3821 | |
| 3822 | @itemize @bullet |
| 3823 | @item |
| 3824 | The version of @code{ld}. @code{ld} announces it if you start it with |
| 3825 | the @samp{--version} argument. |
| 3826 | |
| 3827 | Without this, we will not know whether there is any point in looking for |
| 3828 | the bug in the current version of @code{ld}. |
| 3829 | |
| 3830 | @item |
| 3831 | Any patches you may have applied to the @code{ld} source, including any |
| 3832 | patches made to the @code{BFD} library. |
| 3833 | |
| 3834 | @item |
| 3835 | The type of machine you are using, and the operating system name and |
| 3836 | version number. |
| 3837 | |
| 3838 | @item |
| 3839 | What compiler (and its version) was used to compile @code{ld}---e.g. |
| 3840 | ``@code{gcc-2.7}''. |
| 3841 | |
| 3842 | @item |
| 3843 | The command arguments you gave the linker to link your example and |
| 3844 | observe the bug. To guarantee you will not omit something important, |
| 3845 | list them all. A copy of the Makefile (or the output from make) is |
| 3846 | sufficient. |
| 3847 | |
| 3848 | If we were to try to guess the arguments, we would probably guess wrong |
| 3849 | and then we might not encounter the bug. |
| 3850 | |
| 3851 | @item |
| 3852 | A complete input file, or set of input files, that will reproduce the |
| 3853 | bug. It is generally most helpful to send the actual object files, |
| 3854 | uuencoded if necessary to get them through the mail system. Making them |
| 3855 | available for anonymous FTP is not as good, but may be the only |
| 3856 | reasonable choice for large object files. |
| 3857 | |
| 3858 | If the source files were assembled using @code{gas} or compiled using |
| 3859 | @code{gcc}, then it may be OK to send the source files rather than the |
| 3860 | object files. In this case, be sure to say exactly what version of |
| 3861 | @code{gas} or @code{gcc} was used to produce the object files. Also say |
| 3862 | how @code{gas} or @code{gcc} were configured. |
| 3863 | |
| 3864 | @item |
| 3865 | A description of what behavior you observe that you believe is |
| 3866 | incorrect. For example, ``It gets a fatal signal.'' |
| 3867 | |
| 3868 | Of course, if the bug is that @code{ld} gets a fatal signal, then we |
| 3869 | will certainly notice it. But if the bug is incorrect output, we might |
| 3870 | not notice unless it is glaringly wrong. You might as well not give us |
| 3871 | a chance to make a mistake. |
| 3872 | |
| 3873 | Even if the problem you experience is a fatal signal, you should still |
| 3874 | say so explicitly. Suppose something strange is going on, such as, your |
| 3875 | copy of @code{ld} is out of synch, or you have encountered a bug in the |
| 3876 | C library on your system. (This has happened!) Your copy might crash |
| 3877 | and ours would not. If you told us to expect a crash, then when ours |
| 3878 | fails to crash, we would know that the bug was not happening for us. If |
| 3879 | you had not told us to expect a crash, then we would not be able to draw |
| 3880 | any conclusion from our observations. |
| 3881 | |
| 3882 | @item |
| 3883 | If you wish to suggest changes to the @code{ld} source, send us context |
| 3884 | diffs, as generated by @code{diff} with the @samp{-u}, @samp{-c}, or |
| 3885 | @samp{-p} option. Always send diffs from the old file to the new file. |
| 3886 | If you even discuss something in the @code{ld} source, refer to it by |
| 3887 | context, not by line number. |
| 3888 | |
| 3889 | The line numbers in our development sources will not match those in your |
| 3890 | sources. Your line numbers would convey no useful information to us. |
| 3891 | @end itemize |
| 3892 | |
| 3893 | Here are some things that are not necessary: |
| 3894 | |
| 3895 | @itemize @bullet |
| 3896 | @item |
| 3897 | A description of the envelope of the bug. |
| 3898 | |
| 3899 | Often people who encounter a bug spend a lot of time investigating |
| 3900 | which changes to the input file will make the bug go away and which |
| 3901 | changes will not affect it. |
| 3902 | |
| 3903 | This is often time consuming and not very useful, because the way we |
| 3904 | will find the bug is by running a single example under the debugger |
| 3905 | with breakpoints, not by pure deduction from a series of examples. |
| 3906 | We recommend that you save your time for something else. |
| 3907 | |
| 3908 | Of course, if you can find a simpler example to report @emph{instead} |
| 3909 | of the original one, that is a convenience for us. Errors in the |
| 3910 | output will be easier to spot, running under the debugger will take |
| 3911 | less time, and so on. |
| 3912 | |
| 3913 | However, simplification is not vital; if you do not want to do this, |
| 3914 | report the bug anyway and send us the entire test case you used. |
| 3915 | |
| 3916 | @item |
| 3917 | A patch for the bug. |
| 3918 | |
| 3919 | A patch for the bug does help us if it is a good one. But do not omit |
| 3920 | the necessary information, such as the test case, on the assumption that |
| 3921 | a patch is all we need. We might see problems with your patch and decide |
| 3922 | to fix the problem another way, or we might not understand it at all. |
| 3923 | |
| 3924 | Sometimes with a program as complicated as @code{ld} it is very hard to |
| 3925 | construct an example that will make the program follow a certain path |
| 3926 | through the code. If you do not send us the example, we will not be |
| 3927 | able to construct one, so we will not be able to verify that the bug is |
| 3928 | fixed. |
| 3929 | |
| 3930 | And if we cannot understand what bug you are trying to fix, or why your |
| 3931 | patch should be an improvement, we will not install it. A test case will |
| 3932 | help us to understand. |
| 3933 | |
| 3934 | @item |
| 3935 | A guess about what the bug is or what it depends on. |
| 3936 | |
| 3937 | Such guesses are usually wrong. Even we cannot guess right about such |
| 3938 | things without first using the debugger to find the facts. |
| 3939 | @end itemize |
| 3940 | |
| 3941 | @node MRI |
| 3942 | @appendix MRI Compatible Script Files |
| 3943 | @cindex MRI compatibility |
| 3944 | To aid users making the transition to @sc{gnu} @code{ld} from the MRI |
| 3945 | linker, @code{ld} can use MRI compatible linker scripts as an |
| 3946 | alternative to the more general-purpose linker scripting language |
| 3947 | described in @ref{Scripts}. MRI compatible linker scripts have a much |
| 3948 | simpler command set than the scripting language otherwise used with |
| 3949 | @code{ld}. @sc{gnu} @code{ld} supports the most commonly used MRI |
| 3950 | linker commands; these commands are described here. |
| 3951 | |
| 3952 | In general, MRI scripts aren't of much use with the @code{a.out} object |
| 3953 | file format, since it only has three sections and MRI scripts lack some |
| 3954 | features to make use of them. |
| 3955 | |
| 3956 | You can specify a file containing an MRI-compatible script using the |
| 3957 | @samp{-c} command-line option. |
| 3958 | |
| 3959 | Each command in an MRI-compatible script occupies its own line; each |
| 3960 | command line starts with the keyword that identifies the command (though |
| 3961 | blank lines are also allowed for punctuation). If a line of an |
| 3962 | MRI-compatible script begins with an unrecognized keyword, @code{ld} |
| 3963 | issues a warning message, but continues processing the script. |
| 3964 | |
| 3965 | Lines beginning with @samp{*} are comments. |
| 3966 | |
| 3967 | You can write these commands using all upper-case letters, or all |
| 3968 | lower case; for example, @samp{chip} is the same as @samp{CHIP}. |
| 3969 | The following list shows only the upper-case form of each command. |
| 3970 | |
| 3971 | @table @code |
| 3972 | @cindex @code{ABSOLUTE} (MRI) |
| 3973 | @item ABSOLUTE @var{secname} |
| 3974 | @itemx ABSOLUTE @var{secname}, @var{secname}, @dots{} @var{secname} |
| 3975 | Normally, @code{ld} includes in the output file all sections from all |
| 3976 | the input files. However, in an MRI-compatible script, you can use the |
| 3977 | @code{ABSOLUTE} command to restrict the sections that will be present in |
| 3978 | your output program. If the @code{ABSOLUTE} command is used at all in a |
| 3979 | script, then only the sections named explicitly in @code{ABSOLUTE} |
| 3980 | commands will appear in the linker output. You can still use other |
| 3981 | input sections (whatever you select on the command line, or using |
| 3982 | @code{LOAD}) to resolve addresses in the output file. |
| 3983 | |
| 3984 | @cindex @code{ALIAS} (MRI) |
| 3985 | @item ALIAS @var{out-secname}, @var{in-secname} |
| 3986 | Use this command to place the data from input section @var{in-secname} |
| 3987 | in a section called @var{out-secname} in the linker output file. |
| 3988 | |
| 3989 | @var{in-secname} may be an integer. |
| 3990 | |
| 3991 | @cindex @code{ALIGN} (MRI) |
| 3992 | @item ALIGN @var{secname} = @var{expression} |
| 3993 | Align the section called @var{secname} to @var{expression}. The |
| 3994 | @var{expression} should be a power of two. |
| 3995 | |
| 3996 | @cindex @code{BASE} (MRI) |
| 3997 | @item BASE @var{expression} |
| 3998 | Use the value of @var{expression} as the lowest address (other than |
| 3999 | absolute addresses) in the output file. |
| 4000 | |
| 4001 | @cindex @code{CHIP} (MRI) |
| 4002 | @item CHIP @var{expression} |
| 4003 | @itemx CHIP @var{expression}, @var{expression} |
| 4004 | This command does nothing; it is accepted only for compatibility. |
| 4005 | |
| 4006 | @cindex @code{END} (MRI) |
| 4007 | @item END |
| 4008 | This command does nothing whatever; it's only accepted for compatibility. |
| 4009 | |
| 4010 | @cindex @code{FORMAT} (MRI) |
| 4011 | @item FORMAT @var{output-format} |
| 4012 | Similar to the @code{OUTPUT_FORMAT} command in the more general linker |
| 4013 | language, but restricted to one of these output formats: |
| 4014 | |
| 4015 | @enumerate |
| 4016 | @item |
| 4017 | S-records, if @var{output-format} is @samp{S} |
| 4018 | |
| 4019 | @item |
| 4020 | IEEE, if @var{output-format} is @samp{IEEE} |
| 4021 | |
| 4022 | @item |
| 4023 | COFF (the @samp{coff-m68k} variant in BFD), if @var{output-format} is |
| 4024 | @samp{COFF} |
| 4025 | @end enumerate |
| 4026 | |
| 4027 | @cindex @code{LIST} (MRI) |
| 4028 | @item LIST @var{anything}@dots{} |
| 4029 | Print (to the standard output file) a link map, as produced by the |
| 4030 | @code{ld} command-line option @samp{-M}. |
| 4031 | |
| 4032 | The keyword @code{LIST} may be followed by anything on the |
| 4033 | same line, with no change in its effect. |
| 4034 | |
| 4035 | @cindex @code{LOAD} (MRI) |
| 4036 | @item LOAD @var{filename} |
| 4037 | @itemx LOAD @var{filename}, @var{filename}, @dots{} @var{filename} |
| 4038 | Include one or more object file @var{filename} in the link; this has the |
| 4039 | same effect as specifying @var{filename} directly on the @code{ld} |
| 4040 | command line. |
| 4041 | |
| 4042 | @cindex @code{NAME} (MRI) |
| 4043 | @item NAME @var{output-name} |
| 4044 | @var{output-name} is the name for the program produced by @code{ld}; the |
| 4045 | MRI-compatible command @code{NAME} is equivalent to the command-line |
| 4046 | option @samp{-o} or the general script language command @code{OUTPUT}. |
| 4047 | |
| 4048 | @cindex @code{ORDER} (MRI) |
| 4049 | @item ORDER @var{secname}, @var{secname}, @dots{} @var{secname} |
| 4050 | @itemx ORDER @var{secname} @var{secname} @var{secname} |
| 4051 | Normally, @code{ld} orders the sections in its output file in the |
| 4052 | order in which they first appear in the input files. In an MRI-compatible |
| 4053 | script, you can override this ordering with the @code{ORDER} command. The |
| 4054 | sections you list with @code{ORDER} will appear first in your output |
| 4055 | file, in the order specified. |
| 4056 | |
| 4057 | @cindex @code{PUBLIC} (MRI) |
| 4058 | @item PUBLIC @var{name}=@var{expression} |
| 4059 | @itemx PUBLIC @var{name},@var{expression} |
| 4060 | @itemx PUBLIC @var{name} @var{expression} |
| 4061 | Supply a value (@var{expression}) for external symbol |
| 4062 | @var{name} used in the linker input files. |
| 4063 | |
| 4064 | @cindex @code{SECT} (MRI) |
| 4065 | @item SECT @var{secname}, @var{expression} |
| 4066 | @itemx SECT @var{secname}=@var{expression} |
| 4067 | @itemx SECT @var{secname} @var{expression} |
| 4068 | You can use any of these three forms of the @code{SECT} command to |
| 4069 | specify the start address (@var{expression}) for section @var{secname}. |
| 4070 | If you have more than one @code{SECT} statement for the same |
| 4071 | @var{secname}, only the @emph{first} sets the start address. |
| 4072 | @end table |
| 4073 | |
| 4074 | @node Index |
| 4075 | @unnumbered Index |
| 4076 | |
| 4077 | @printindex cp |
| 4078 | |
| 4079 | @tex |
| 4080 | % I think something like @colophon should be in texinfo. In the |
| 4081 | % meantime: |
| 4082 | \long\def\colophon{\hbox to0pt{}\vfill |
| 4083 | \centerline{The body of this manual is set in} |
| 4084 | \centerline{\fontname\tenrm,} |
| 4085 | \centerline{with headings in {\bf\fontname\tenbf}} |
| 4086 | \centerline{and examples in {\tt\fontname\tentt}.} |
| 4087 | \centerline{{\it\fontname\tenit\/} and} |
| 4088 | \centerline{{\sl\fontname\tensl\/}} |
| 4089 | \centerline{are used for emphasis.}\vfill} |
| 4090 | \page\colophon |
| 4091 | % Blame: doc@cygnus.com, 28mar91. |
| 4092 | @end tex |
| 4093 | |
| 4094 | |
| 4095 | @contents |
| 4096 | @bye |
| 4097 | |
| 4098 | |