| 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 | |
| 7 | @c @smallbook |
| 8 | |
| 9 | @ifinfo |
| 10 | @format |
| 11 | START-INFO-DIR-ENTRY |
| 12 | * Ld: (ld). The GNU linker. |
| 13 | END-INFO-DIR-ENTRY |
| 14 | @end format |
| 15 | @end ifinfo |
| 16 | |
| 17 | @ifinfo |
| 18 | This file documents the @sc{gnu} linker LD. |
| 19 | |
| 20 | Copyright (C) 1991, 92, 93, 94, 95, 1996 Free Software Foundation, Inc. |
| 21 | |
| 22 | Permission is granted to make and distribute verbatim copies of |
| 23 | this manual provided the copyright notice and this permission notice |
| 24 | are preserved on all copies. |
| 25 | |
| 26 | Permission is granted to copy and distribute modified versions of this |
| 27 | manual under the conditions for verbatim copying, provided also that |
| 28 | the entire resulting derived work is distributed under the terms of a |
| 29 | permission notice identical to this one. |
| 30 | |
| 31 | Permission is granted to copy and distribute translations of this manual |
| 32 | into another language, under the above conditions for modified versions. |
| 33 | |
| 34 | @ignore |
| 35 | Permission is granted to process this file through Tex and print the |
| 36 | results, provided the printed document carries copying permission |
| 37 | notice identical to this one except for the removal of this paragraph |
| 38 | (this paragraph not being relevant to the printed manual). |
| 39 | |
| 40 | @end ignore |
| 41 | @end ifinfo |
| 42 | @iftex |
| 43 | @finalout |
| 44 | @setchapternewpage odd |
| 45 | @settitle Using LD, the GNU linker |
| 46 | @titlepage |
| 47 | @title Using ld |
| 48 | @subtitle The GNU linker |
| 49 | @sp 1 |
| 50 | @subtitle @code{ld} version 2 |
| 51 | @subtitle January 1994 |
| 52 | @author Steve Chamberlain |
| 53 | @author Cygnus Support |
| 54 | @page |
| 55 | |
| 56 | @tex |
| 57 | {\parskip=0pt |
| 58 | \hfill Cygnus Support\par |
| 59 | \hfill steve\@cygnus.com, doc\@cygnus.com\par |
| 60 | \hfill {\it Using LD, the GNU linker}\par |
| 61 | \hfill Edited by Jeffrey Osier (jeffrey\@cygnus.com)\par |
| 62 | } |
| 63 | \global\parindent=0pt % Steve likes it this way. |
| 64 | @end tex |
| 65 | |
| 66 | @vskip 0pt plus 1filll |
| 67 | Copyright @copyright{} 1991, 92, 93, 94, 95, 1996 Free Software Foundation, Inc. |
| 68 | |
| 69 | Permission is granted to make and distribute verbatim copies of |
| 70 | this manual provided the copyright notice and this permission notice |
| 71 | are preserved on all copies. |
| 72 | |
| 73 | Permission is granted to copy and distribute modified versions of this |
| 74 | manual under the conditions for verbatim copying, provided also that |
| 75 | the entire resulting derived work is distributed under the terms of a |
| 76 | permission notice identical to this one. |
| 77 | |
| 78 | Permission is granted to copy and distribute translations of this manual |
| 79 | into another language, under the above conditions for modified versions. |
| 80 | @end titlepage |
| 81 | @end iftex |
| 82 | @c FIXME: Talk about importance of *order* of args, cmds to linker! |
| 83 | |
| 84 | @ifinfo |
| 85 | @node Top |
| 86 | @top Using ld |
| 87 | This file documents the @sc{gnu} linker ld. |
| 88 | |
| 89 | @menu |
| 90 | * Overview:: Overview |
| 91 | * Invocation:: Invocation |
| 92 | * Commands:: Command Language |
| 93 | @ifset GENERIC |
| 94 | * Machine Dependent:: Machine Dependent Features |
| 95 | @end ifset |
| 96 | @ifclear GENERIC |
| 97 | @ifset H8300 |
| 98 | * H8/300:: ld and the H8/300 |
| 99 | @end ifset |
| 100 | @ifset Hitachi |
| 101 | * Hitachi:: ld and other Hitachi micros |
| 102 | @end ifset |
| 103 | @ifset I960 |
| 104 | * i960:: ld and the Intel 960 family |
| 105 | @end ifset |
| 106 | @end ifclear |
| 107 | @ifclear SingleFormat |
| 108 | * BFD:: BFD |
| 109 | @end ifclear |
| 110 | @c Following blank line required for remaining bug in makeinfo conds/menus |
| 111 | |
| 112 | * Reporting Bugs:: Reporting Bugs |
| 113 | * MRI:: MRI Compatible Script Files |
| 114 | * Index:: Index |
| 115 | @end menu |
| 116 | @end ifinfo |
| 117 | |
| 118 | @node Overview |
| 119 | @chapter Overview |
| 120 | |
| 121 | @cindex @sc{gnu} linker |
| 122 | @cindex what is this? |
| 123 | @code{ld} combines a number of object and archive files, relocates |
| 124 | their data and ties up symbol references. Usually the last step in |
| 125 | compiling a program is to run @code{ld}. |
| 126 | |
| 127 | @code{ld} accepts Linker Command Language files written in |
| 128 | a superset of AT&T's Link Editor Command Language syntax, |
| 129 | to provide explicit and total control over the linking process. |
| 130 | |
| 131 | @ifclear SingleFormat |
| 132 | This version of @code{ld} uses the general purpose BFD libraries |
| 133 | to operate on object files. This allows @code{ld} to read, combine, and |
| 134 | write object files in many different formats---for example, COFF or |
| 135 | @code{a.out}. Different formats may be linked together to produce any |
| 136 | available kind of object file. @xref{BFD}, for more information. |
| 137 | @end ifclear |
| 138 | |
| 139 | Aside from its flexibility, the @sc{gnu} linker is more helpful than other |
| 140 | linkers in providing diagnostic information. Many linkers abandon |
| 141 | execution immediately upon encountering an error; whenever possible, |
| 142 | @code{ld} continues executing, allowing you to identify other errors |
| 143 | (or, in some cases, to get an output file in spite of the error). |
| 144 | |
| 145 | @node Invocation |
| 146 | @chapter Invocation |
| 147 | |
| 148 | The @sc{gnu} linker @code{ld} is meant to cover a broad range of situations, |
| 149 | and to be as compatible as possible with other linkers. As a result, |
| 150 | you have many choices to control its behavior. |
| 151 | |
| 152 | @ifset UsesEnvVars |
| 153 | @menu |
| 154 | * Options:: Command Line Options |
| 155 | * Environment:: Environment Variables |
| 156 | @end menu |
| 157 | |
| 158 | @node Options |
| 159 | @section Command Line Options |
| 160 | @end ifset |
| 161 | |
| 162 | @cindex command line |
| 163 | @cindex options |
| 164 | The linker supports a plethora of command-line options, but in actual |
| 165 | practice few of them are used in any particular context. |
| 166 | @cindex standard Unix system |
| 167 | For instance, a frequent use of @code{ld} is to link standard Unix |
| 168 | object files on a standard, supported Unix system. On such a system, to |
| 169 | link a file @code{hello.o}: |
| 170 | |
| 171 | @smallexample |
| 172 | ld -o @var{output} /lib/crt0.o hello.o -lc |
| 173 | @end smallexample |
| 174 | |
| 175 | This tells @code{ld} to produce a file called @var{output} as the |
| 176 | result of linking the file @code{/lib/crt0.o} with @code{hello.o} and |
| 177 | the library @code{libc.a}, which will come from the standard search |
| 178 | directories. (See the discussion of the @samp{-l} option below.) |
| 179 | |
| 180 | The command-line options to @code{ld} may be specified in any order, and |
| 181 | may be repeated at will. Repeating most options with a different |
| 182 | argument will either have no further effect, or override prior |
| 183 | occurrences (those further to the left on the command line) of that |
| 184 | option. Options which may be meaningfully specified more than once are |
| 185 | noted in the descriptions below. |
| 186 | |
| 187 | @cindex object files |
| 188 | Non-option arguments are objects files which are to be linked together. |
| 189 | They may follow, precede, or be mixed in with command-line options, |
| 190 | except that an object file argument may not be placed between an option |
| 191 | and its argument. |
| 192 | |
| 193 | Usually the linker is invoked with at least one object file, but you can |
| 194 | specify other forms of binary input files using @samp{-l}, @samp{-R}, |
| 195 | and the script command language. If @emph{no} binary input files at all |
| 196 | are specified, the linker does not produce any output, and issues the |
| 197 | message @samp{No input files}. |
| 198 | |
| 199 | If the linker can not recognize the format of an object file, it will |
| 200 | assume that it is a linker script. A script specified in this way |
| 201 | augments the main linker script used for the link (either the default |
| 202 | linker script or the one specified by using @samp{-T}). This feature |
| 203 | permits the linker to link against a file which appears to be an object |
| 204 | or an archive, but actually merely defines some symbol values, or uses |
| 205 | @code{INPUT} or @code{GROUP} to load other objects. Note that |
| 206 | specifying a script in this way should only be used to augment the main |
| 207 | linker script; if you want to use some command that logically can only |
| 208 | appear once, such as the @code{SECTIONS} or @code{MEMORY} command, you |
| 209 | must replace the default linker script using the @samp{-T} option. |
| 210 | @xref{Commands}. |
| 211 | |
| 212 | For options whose names are a single letter, |
| 213 | option arguments must either follow the option letter without intervening |
| 214 | whitespace, or be given as separate arguments immediately following the |
| 215 | option that requires them. |
| 216 | |
| 217 | For options whose names are multiple letters, either one dash or two can |
| 218 | precede the option name; for example, @samp{--oformat} and |
| 219 | @samp{--oformat} are equivalent. Arguments to multiple-letter options |
| 220 | must either be separated from the option name by an equals sign, or be |
| 221 | given as separate arguments immediately following the option that |
| 222 | requires them. For example, @samp{--oformat srec} and |
| 223 | @samp{--oformat=srec} are equivalent. Unique abbreviations of the names |
| 224 | of multiple-letter options are accepted. |
| 225 | |
| 226 | @table @code |
| 227 | @kindex -a@var{keyword} |
| 228 | @item -a@var{keyword} |
| 229 | This option is supported for HP/UX compatibility. The @var{keyword} |
| 230 | argument must be one of the strings @samp{archive}, @samp{shared}, or |
| 231 | @samp{default}. @samp{-aarchive} is functionally equivalent to |
| 232 | @samp{-Bstatic}, and the other two keywords are functionally equivalent |
| 233 | to @samp{-Bdynamic}. This option may be used any number of times. |
| 234 | |
| 235 | @ifset I960 |
| 236 | @cindex architectures |
| 237 | @kindex -A@var{arch} |
| 238 | @item -A@var{architecture} |
| 239 | @kindex --architecture=@var{arch} |
| 240 | @itemx --architecture=@var{architecture} |
| 241 | In the current release of @code{ld}, this option is useful only for the |
| 242 | Intel 960 family of architectures. In that @code{ld} configuration, the |
| 243 | @var{architecture} argument identifies the particular architecture in |
| 244 | the 960 family, enabling some safeguards and modifying the |
| 245 | archive-library search path. @xref{i960,,@code{ld} and the Intel 960 |
| 246 | family}, for details. |
| 247 | |
| 248 | Future releases of @code{ld} may support similar functionality for |
| 249 | other architecture families. |
| 250 | @end ifset |
| 251 | |
| 252 | @ifclear SingleFormat |
| 253 | @cindex binary input format |
| 254 | @kindex -b @var{format} |
| 255 | @kindex --format=@var{format} |
| 256 | @cindex input format |
| 257 | @cindex input format |
| 258 | @item -b @var{input-format} |
| 259 | @itemx --format=@var{input-format} |
| 260 | @code{ld} may be configured to support more than one kind of object |
| 261 | file. If your @code{ld} is configured this way, you can use the |
| 262 | @samp{-b} option to specify the binary format for input object files |
| 263 | that follow this option on the command line. Even when @code{ld} is |
| 264 | configured to support alternative object formats, you don't usually need |
| 265 | to specify this, as @code{ld} should be configured to expect as a |
| 266 | default input format the most usual format on each machine. |
| 267 | @var{input-format} is a text string, the name of a particular format |
| 268 | supported by the BFD libraries. (You can list the available binary |
| 269 | formats with @samp{objdump -i}.) |
| 270 | @xref{BFD}. |
| 271 | |
| 272 | You may want to use this option if you are linking files with an unusual |
| 273 | binary format. You can also use @samp{-b} to switch formats explicitly (when |
| 274 | linking object files of different formats), by including |
| 275 | @samp{-b @var{input-format}} before each group of object files in a |
| 276 | particular format. |
| 277 | |
| 278 | The default format is taken from the environment variable |
| 279 | @code{GNUTARGET}. |
| 280 | @ifset UsesEnvVars |
| 281 | @xref{Environment}. |
| 282 | @end ifset |
| 283 | You can also define the input |
| 284 | format from a script, using the command @code{TARGET}; see @ref{Option |
| 285 | Commands}. |
| 286 | @end ifclear |
| 287 | |
| 288 | @kindex -c @var{MRI-cmdfile} |
| 289 | @kindex --mri-script=@var{MRI-cmdfile} |
| 290 | @cindex compatibility, MRI |
| 291 | @item -c @var{MRI-commandfile} |
| 292 | @itemx --mri-script=@var{MRI-commandfile} |
| 293 | For compatibility with linkers produced by MRI, @code{ld} accepts script |
| 294 | files written in an alternate, restricted command language, described in |
| 295 | @ref{MRI,,MRI Compatible Script Files}. Introduce MRI script files with |
| 296 | the option @samp{-c}; use the @samp{-T} option to run linker |
| 297 | scripts written in the general-purpose @code{ld} scripting language. |
| 298 | If @var{MRI-cmdfile} does not exist, @code{ld} looks for it in the directories |
| 299 | specified by any @samp{-L} options. |
| 300 | |
| 301 | @cindex common allocation |
| 302 | @kindex -d |
| 303 | @kindex -dc |
| 304 | @kindex -dp |
| 305 | @item -d |
| 306 | @itemx -dc |
| 307 | @itemx -dp |
| 308 | These three options are equivalent; multiple forms are supported for |
| 309 | compatibility with other linkers. They |
| 310 | assign space to common symbols even if a relocatable output file is |
| 311 | specified (with @samp{-r}). The script command |
| 312 | @code{FORCE_COMMON_ALLOCATION} has the same effect. @xref{Option |
| 313 | 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. @xref{Entry Point}, for a |
| 322 | discussion of defaults and other ways of specifying the |
| 323 | entry point. |
| 324 | |
| 325 | @cindex dynamic symbol table |
| 326 | @kindex -E |
| 327 | @kindex --export-dynamic |
| 328 | @item -E |
| 329 | @itemx --export-dynamic |
| 330 | When creating a dynamically linked executable, add all symbols to the |
| 331 | dynamic symbol table. Normally, the dynamic symbol table contains only |
| 332 | symbols which are used by a dynamic object. This option is needed for |
| 333 | some uses of @code{dlopen}. |
| 334 | |
| 335 | @kindex -f |
| 336 | @kindex --auxiliary |
| 337 | @item -f |
| 338 | @itemx --auxiliary @var{name} |
| 339 | When creating an ELF shared object, set the internal DT_AUXILIARY field |
| 340 | to the specified name. This tells the dynamic linker that the symbol |
| 341 | table of the shared object should be used as an auxiliary filter on the |
| 342 | symbol table of the shared object @var{name}. |
| 343 | |
| 344 | If you later link a program against this filter object, then, when you |
| 345 | run the program, the dynamic linker will see the DT_AUXILIARY field. If |
| 346 | the dynamic linker resolves any symbols from the filter object, it will |
| 347 | first check whether there is a definition in the shared object |
| 348 | @var{name}. If there is one, it will be used instead of the definition |
| 349 | in the filter object. The shared object @var{name} need not exist. |
| 350 | Thus the shared object @var{name} may be used to provide an alternative |
| 351 | implementation of certain functions, perhaps for debugging or for |
| 352 | machine specific performance. |
| 353 | |
| 354 | @kindex -F |
| 355 | @kindex --filter |
| 356 | @item -F @var{name} |
| 357 | @itemx --filter @var{name} |
| 358 | When creating an ELF shared object, set the internal DT_FILTER field to |
| 359 | the specified name. This tells the dynamic linker that the symbol table |
| 360 | of the shared object which is being created should be used as a filter |
| 361 | on the symbol table of the shared object @var{name}. |
| 362 | |
| 363 | If you later link a program against this filter object, then, when you |
| 364 | run the program, the dynamic linker will see the DT_FILTER field. The |
| 365 | dynamic linker will resolve symbols according to the symbol table of the |
| 366 | filter object as usual, but it will actually link to the definitions |
| 367 | found in the shared object @var{name}. Thus the filter object can be |
| 368 | used to select a subset of the symbols provided by the object |
| 369 | @var{name}. |
| 370 | |
| 371 | Some older linkers used the @code{-F} option throughout a compilation |
| 372 | toolchain for specifying object-file format for both input and output |
| 373 | object files. The @sc{gnu} linker uses other mechanisms for this |
| 374 | purpose: the @code{-b}, @code{--format}, @code{--oformat} options, the |
| 375 | @code{TARGET} command in linker scripts, and the @code{GNUTARGET} |
| 376 | environment variable. The @sc{gnu} linker will ignore the @code{-F} |
| 377 | option when not creating an ELF shared object. |
| 378 | |
| 379 | @kindex --force-exe-suffix |
| 380 | @item --force-exe-suffix |
| 381 | Make sure that an output file has a .exe suffix. |
| 382 | |
| 383 | If a successfully built fully linked output file does not have a |
| 384 | @code{.exe} or @code{.dll} suffix, this option forces the linker to copy |
| 385 | the output file to one of the same name with a @code{.exe} suffix. This |
| 386 | option is useful when using unmodified Unix makefiles on a Microsoft |
| 387 | Windows host, since some versions of Windows won't run an image unless |
| 388 | it ends in a @code{.exe} suffix. |
| 389 | |
| 390 | @kindex -g |
| 391 | @item -g |
| 392 | Ignored. Provided for compatibility with other tools. |
| 393 | |
| 394 | @kindex -G |
| 395 | @kindex --gpsize |
| 396 | @cindex object size |
| 397 | @item -G@var{value} |
| 398 | @itemx --gpsize=@var{value} |
| 399 | Set the maximum size of objects to be optimized using the GP register to |
| 400 | @var{size}. This is only meaningful for object file formats such as |
| 401 | MIPS ECOFF which supports putting large and small objects into different |
| 402 | sections. This is ignored for other object file formats. |
| 403 | |
| 404 | @cindex runtime library name |
| 405 | @kindex -h@var{name} |
| 406 | @kindex -soname=@var{name} |
| 407 | @item -h@var{name} |
| 408 | @itemx -soname=@var{name} |
| 409 | When creating an ELF shared object, set the internal DT_SONAME field to |
| 410 | the specified name. When an executable is linked with a shared object |
| 411 | which has a DT_SONAME field, then when the executable is run the dynamic |
| 412 | linker will attempt to load the shared object specified by the DT_SONAME |
| 413 | field rather than the using the file name given to the linker. |
| 414 | |
| 415 | @kindex -i |
| 416 | @cindex incremental link |
| 417 | @item -i |
| 418 | Perform an incremental link (same as option @samp{-r}). |
| 419 | |
| 420 | @cindex archive files, from cmd line |
| 421 | @kindex -l@var{archive} |
| 422 | @kindex --library=@var{archive} |
| 423 | @item -l@var{archive} |
| 424 | @itemx --library=@var{archive} |
| 425 | Add archive file @var{archive} to the list of files to link. This |
| 426 | option may be used any number of times. @code{ld} will search its |
| 427 | path-list for occurrences of @code{lib@var{archive}.a} for every |
| 428 | @var{archive} specified. |
| 429 | |
| 430 | On systems which support shared libraries, @code{ld} may also search for |
| 431 | libraries with extensions other than @code{.a}. Specifically, on ELF |
| 432 | and SunOS systems, @code{ld} will search a directory for a library with |
| 433 | an extension of @code{.so} before searching for one with an extension of |
| 434 | @code{.a}. By convention, a @code{.so} extension indicates a shared |
| 435 | library. |
| 436 | |
| 437 | The linker will search an archive only once, at the location where it is |
| 438 | specified on the command line. If the archive defines a symbol which |
| 439 | was undefined in some object which appeared before the archive on the |
| 440 | command line, the linker will include the appropriate file(s) from the |
| 441 | archive. However, an undefined symbol in an object appearing later on |
| 442 | the command line will not cause the linker to search the archive again. |
| 443 | |
| 444 | See the @code{-(} option for a way to force the linker to search |
| 445 | archives multiple times. |
| 446 | |
| 447 | You may list the same archive multiple times on the command line. |
| 448 | |
| 449 | @ifset GENERIC |
| 450 | This type of archive searching is standard for Unix linkers. However, |
| 451 | if you are using @code{ld} on AIX, note that it is different from the |
| 452 | behaviour of the AIX linker. |
| 453 | @end ifset |
| 454 | |
| 455 | @cindex search directory, from cmd line |
| 456 | @kindex -L@var{dir} |
| 457 | @kindex --library-path=@var{dir} |
| 458 | @item -L@var{searchdir} |
| 459 | @itemx --library-path=@var{searchdir} |
| 460 | Add path @var{searchdir} to the list of paths that @code{ld} will search |
| 461 | for archive libraries and @code{ld} control scripts. You may use this |
| 462 | option any number of times. The directories are searched in the order |
| 463 | in which they are specified on the command line. Directories specified |
| 464 | on the command line are searched before the default directories. All |
| 465 | @code{-L} options apply to all @code{-l} options, regardless of the |
| 466 | order in which the options appear. |
| 467 | |
| 468 | @ifset UsesEnvVars |
| 469 | The default set of paths searched (without being specified with |
| 470 | @samp{-L}) depends on which emulation mode @code{ld} is using, and in |
| 471 | some cases also on how it was configured. @xref{Environment}. |
| 472 | @end ifset |
| 473 | |
| 474 | The paths can also be specified in a link script with the |
| 475 | @code{SEARCH_DIR} command. Directories specified this way are searched |
| 476 | at the point in which the linker script appears in the command line. |
| 477 | |
| 478 | @cindex emulation |
| 479 | @kindex -m @var{emulation} |
| 480 | @item -m@var{emulation} |
| 481 | Emulate the @var{emulation} linker. You can list the available |
| 482 | emulations with the @samp{--verbose} or @samp{-V} options. The default |
| 483 | depends on how your @code{ld} was configured. |
| 484 | |
| 485 | @cindex link map |
| 486 | @kindex -M |
| 487 | @kindex --print-map |
| 488 | @item -M |
| 489 | @itemx --print-map |
| 490 | Print (to the standard output) a link map---diagnostic information about |
| 491 | where symbols are mapped by @code{ld}, and information on global common |
| 492 | storage allocation. |
| 493 | |
| 494 | @kindex -n |
| 495 | @cindex read-only text |
| 496 | @cindex NMAGIC |
| 497 | @kindex --nmagic |
| 498 | @item -n |
| 499 | @itemx --nmagic |
| 500 | Set the text segment to be read only, and mark the output as |
| 501 | @code{NMAGIC} if possible. |
| 502 | |
| 503 | @kindex -N |
| 504 | @kindex --omagic |
| 505 | @cindex read/write from cmd line |
| 506 | @cindex OMAGIC |
| 507 | @item -N |
| 508 | @itemx --omagic |
| 509 | Set the text and data sections to be readable and writable. Also, do |
| 510 | not page-align the data segment. If the output format supports Unix |
| 511 | style magic numbers, mark the output as @code{OMAGIC}. |
| 512 | |
| 513 | @kindex -o @var{output} |
| 514 | @kindex --output=@var{output} |
| 515 | @cindex naming the output file |
| 516 | @item -o @var{output} |
| 517 | @itemx --output=@var{output} |
| 518 | Use @var{output} as the name for the program produced by @code{ld}; if this |
| 519 | option is not specified, the name @file{a.out} is used by default. The |
| 520 | script command @code{OUTPUT} can also specify the output file name. |
| 521 | |
| 522 | @cindex partial link |
| 523 | @cindex relocatable output |
| 524 | @kindex -r |
| 525 | @kindex --relocateable |
| 526 | @item -r |
| 527 | @itemx --relocateable |
| 528 | Generate relocatable output---i.e., generate an output file that can in |
| 529 | turn serve as input to @code{ld}. This is often called @dfn{partial |
| 530 | linking}. As a side effect, in environments that support standard Unix |
| 531 | magic numbers, this option also sets the output file's magic number to |
| 532 | @code{OMAGIC}. |
| 533 | @c ; see @code{-N}. |
| 534 | If this option is not specified, an absolute file is produced. When |
| 535 | linking C++ programs, this option @emph{will not} resolve references to |
| 536 | constructors; to do that, use @samp{-Ur}. |
| 537 | |
| 538 | This option does the same thing as @samp{-i}. |
| 539 | |
| 540 | @kindex -R @var{file} |
| 541 | @kindex --just-symbols=@var{file} |
| 542 | @cindex symbol-only input |
| 543 | @item -R @var{filename} |
| 544 | @itemx --just-symbols=@var{filename} |
| 545 | Read symbol names and their addresses from @var{filename}, but do not |
| 546 | relocate it or include it in the output. This allows your output file |
| 547 | to refer symbolically to absolute locations of memory defined in other |
| 548 | programs. You may use this option more than once. |
| 549 | |
| 550 | For compatibility with other ELF linkers, if the @code{-R} option is |
| 551 | followed by a directory name, rather than a file name, it is treated as |
| 552 | the @code{-rpath} option. |
| 553 | |
| 554 | @kindex -s |
| 555 | @kindex --strip-all |
| 556 | @cindex strip all symbols |
| 557 | @item -s |
| 558 | @itemx --strip-all |
| 559 | Omit all symbol information from the output file. |
| 560 | |
| 561 | @kindex -S |
| 562 | @kindex --strip-debug |
| 563 | @cindex strip debugger symbols |
| 564 | @item -S |
| 565 | @itemx --strip-debug |
| 566 | Omit debugger symbol information (but not all symbols) from the output file. |
| 567 | |
| 568 | @kindex -t |
| 569 | @kindex --trace |
| 570 | @cindex input files, displaying |
| 571 | @item -t |
| 572 | @itemx --trace |
| 573 | Print the names of the input files as @code{ld} processes them. |
| 574 | |
| 575 | @kindex -T @var{script} |
| 576 | @kindex --script=@var{script} |
| 577 | @cindex script files |
| 578 | @item -T @var{commandfile} |
| 579 | @itemx --script=@var{commandfile} |
| 580 | Read link commands from the file @var{commandfile}. These commands |
| 581 | replace @code{ld}'s default link script (rather than adding to it), so |
| 582 | @var{commandfile} must specify everything necessary to describe the |
| 583 | target format. You must use this option if you want to use a command |
| 584 | which can only appear once in a linker script, such as the |
| 585 | @code{SECTIONS} or @code{MEMORY} command. @xref{Commands}. If |
| 586 | @var{commandfile} does not exist, @code{ld} looks for it in the |
| 587 | directories specified by any preceding @samp{-L} options. Multiple |
| 588 | @samp{-T} options accumulate. |
| 589 | |
| 590 | @kindex -u @var{symbol} |
| 591 | @kindex --undefined=@var{symbol} |
| 592 | @cindex undefined symbol |
| 593 | @item -u @var{symbol} |
| 594 | @itemx --undefined=@var{symbol} |
| 595 | Force @var{symbol} to be entered in the output file as an undefined symbol. |
| 596 | Doing this may, for example, trigger linking of additional modules from |
| 597 | standard libraries. @samp{-u} may be repeated with different option |
| 598 | arguments to enter additional undefined symbols. |
| 599 | @c Nice idea, but no such command: This option is equivalent |
| 600 | @c to the @code{EXTERN} linker command. |
| 601 | |
| 602 | @kindex -v |
| 603 | @kindex -V |
| 604 | @kindex --version |
| 605 | @cindex version |
| 606 | @item -v |
| 607 | @itemx --version |
| 608 | @itemx -V |
| 609 | Display the version number for @code{ld}. The @code{-V} option also |
| 610 | lists the supported emulations. |
| 611 | |
| 612 | @kindex -x |
| 613 | @kindex --discard-all |
| 614 | @cindex deleting local symbols |
| 615 | @item -x |
| 616 | @itemx --discard-all |
| 617 | Delete all local symbols. |
| 618 | |
| 619 | @kindex -X |
| 620 | @kindex --discard-locals |
| 621 | @cindex local symbols, deleting |
| 622 | @cindex L, deleting symbols beginning |
| 623 | @item -X |
| 624 | @itemx --discard-locals |
| 625 | Delete all temporary local symbols. For most targets, this is all local |
| 626 | symbols whose names begin with @samp{L}. |
| 627 | |
| 628 | @kindex -y @var{symbol} |
| 629 | @kindex --trace-symbol=@var{symbol} |
| 630 | @cindex symbol tracing |
| 631 | @item -y @var{symbol} |
| 632 | @itemx --trace-symbol=@var{symbol} |
| 633 | Print the name of each linked file in which @var{symbol} appears. This |
| 634 | option may be given any number of times. On many systems it is necessary |
| 635 | to prepend an underscore. |
| 636 | |
| 637 | This option is useful when you have an undefined symbol in your link but |
| 638 | don't know where the reference is coming from. |
| 639 | |
| 640 | @kindex -Y @var{path} |
| 641 | @item -Y @var{path} |
| 642 | Add @var{path} to the default library search path. This option exists |
| 643 | for Solaris compatibility. |
| 644 | |
| 645 | @kindex -z @var{keyword} |
| 646 | @item -z @var{keyword} |
| 647 | This option is ignored for Solaris compatibility. |
| 648 | |
| 649 | @kindex -( |
| 650 | @cindex groups of archives |
| 651 | @item -( @var{archives} -) |
| 652 | @itemx --start-group @var{archives} --end-group |
| 653 | The @var{archives} should be a list of archive files. They may be |
| 654 | either explicit file names, or @samp{-l} options. |
| 655 | |
| 656 | The specified archives are searched repeatedly until no new undefined |
| 657 | references are created. Normally, an archive is searched only once in |
| 658 | the order that it is specified on the command line. If a symbol in that |
| 659 | archive is needed to resolve an undefined symbol referred to by an |
| 660 | object in an archive that appears later on the command line, the linker |
| 661 | would not be able to resolve that reference. By grouping the archives, |
| 662 | they all be searched repeatedly until all possible references are |
| 663 | resolved. |
| 664 | |
| 665 | Using this option has a significant performance cost. It is best to use |
| 666 | it only when there are unavoidable circular references between two or |
| 667 | more archives. |
| 668 | |
| 669 | @kindex -assert @var{keyword} |
| 670 | @item -assert @var{keyword} |
| 671 | This option is ignored for SunOS compatibility. |
| 672 | |
| 673 | @kindex -Bdynamic |
| 674 | @kindex -dy |
| 675 | @kindex -call_shared |
| 676 | @item -Bdynamic |
| 677 | @itemx -dy |
| 678 | @itemx -call_shared |
| 679 | Link against dynamic libraries. This is only meaningful on platforms |
| 680 | for which shared libraries are supported. This option is normally the |
| 681 | default on such platforms. The different variants of this option are |
| 682 | for compatibility with various systems. You may use this option |
| 683 | multiple times on the command line: it affects library searching for |
| 684 | @code{-l} options which follow it. |
| 685 | |
| 686 | @kindex -Bstatic |
| 687 | @kindex -dn |
| 688 | @kindex -non_shared |
| 689 | @kindex -static |
| 690 | @item -Bstatic |
| 691 | @itemx -dn |
| 692 | @itemx -non_shared |
| 693 | @itemx -static |
| 694 | Do not link against shared libraries. This is only meaningful on |
| 695 | platforms for which shared libraries are supported. The different |
| 696 | variants of this option are for compatibility with various systems. You |
| 697 | may use this option multiple times on the command line: it affects |
| 698 | library searching for @code{-l} options which follow it. |
| 699 | |
| 700 | @kindex -Bsymbolic |
| 701 | @item -Bsymbolic |
| 702 | When creating a shared library, bind references to global symbols to the |
| 703 | definition within the shared library, if any. Normally, it is possible |
| 704 | for a program linked against a shared library to override the definition |
| 705 | within the shared library. This option is only meaningful on ELF |
| 706 | platforms which support shared libraries. |
| 707 | |
| 708 | @cindex cross reference table |
| 709 | @kindex --cref |
| 710 | @item --cref |
| 711 | Output a cross reference table. If a linker map file is being |
| 712 | generated, the cross reference table is printed to the map file. |
| 713 | Otherwise, it is printed on the standard output. |
| 714 | |
| 715 | The format of the table is intentionally simple, so that it may be |
| 716 | easily processed by a script if necessary. The symbols are printed out, |
| 717 | sorted by name. For each symbol, a list of file names is given. If the |
| 718 | symbol is defined, the first file listed is the location of the |
| 719 | definition. The remaining files contain references to the symbol. |
| 720 | |
| 721 | @cindex symbols, from command line |
| 722 | @kindex --defsym @var{symbol}=@var{exp} |
| 723 | @item --defsym @var{symbol}=@var{expression} |
| 724 | Create a global symbol in the output file, containing the absolute |
| 725 | address given by @var{expression}. You may use this option as many |
| 726 | times as necessary to define multiple symbols in the command line. A |
| 727 | limited form of arithmetic is supported for the @var{expression} in this |
| 728 | context: you may give a hexadecimal constant or the name of an existing |
| 729 | symbol, or use @code{+} and @code{-} to add or subtract hexadecimal |
| 730 | constants or symbols. If you need more elaborate expressions, consider |
| 731 | using the linker command language from a script (@pxref{Assignment, , |
| 732 | Assignment: Symbol Definitions}). @emph{Note:} there should be no |
| 733 | white space between @var{symbol}, the equals sign (``@key{=}''), and |
| 734 | @var{expression}. |
| 735 | |
| 736 | @cindex dynamic linker, from command line |
| 737 | @kindex --dynamic-linker @var{file} |
| 738 | @item --dynamic-linker @var{file} |
| 739 | Set the name of the dynamic linker. This is only meaningful when |
| 740 | generating dynamically linked ELF executables. The default dynamic |
| 741 | linker is normally correct; don't use this unless you know what you are |
| 742 | doing. |
| 743 | |
| 744 | @cindex big-endian objects |
| 745 | @cindex endianness |
| 746 | @kindex -EB |
| 747 | @item -EB |
| 748 | Link big-endian objects. This affects the default output format. |
| 749 | |
| 750 | @cindex little-endian objects |
| 751 | @kindex -EL |
| 752 | @item -EL |
| 753 | Link little-endian objects. This affects the default output format. |
| 754 | |
| 755 | @cindex MIPS embedded PIC code |
| 756 | @kindex --embedded-relocs |
| 757 | @item --embedded-relocs |
| 758 | This option is only meaningful when linking MIPS embedded PIC code, |
| 759 | generated by the -membedded-pic option to the @sc{gnu} compiler and |
| 760 | assembler. It causes the linker to create a table which may be used at |
| 761 | runtime to relocate any data which was statically initialized to pointer |
| 762 | values. See the code in testsuite/ld-empic for details. |
| 763 | |
| 764 | @cindex help |
| 765 | @cindex usage |
| 766 | @kindex --help |
| 767 | @item --help |
| 768 | Print a summary of the command-line options on the standard output and exit. |
| 769 | |
| 770 | @cindex link map |
| 771 | @kindex -Map |
| 772 | @item -Map @var{mapfile} |
| 773 | Print to the file @var{mapfile} a link map---diagnostic information |
| 774 | about where symbols are mapped by @code{ld}, and information on global |
| 775 | common storage allocation. |
| 776 | |
| 777 | @cindex memory usage |
| 778 | @kindex --no-keep-memory |
| 779 | @item --no-keep-memory |
| 780 | @code{ld} normally optimizes for speed over memory usage by caching the |
| 781 | symbol tables of input files in memory. This option tells @code{ld} to |
| 782 | instead optimize for memory usage, by rereading the symbol tables as |
| 783 | necessary. This may be required if @code{ld} runs out of memory space |
| 784 | while linking a large executable. |
| 785 | |
| 786 | @kindex --no-whole-archive |
| 787 | @item --no-whole-archive |
| 788 | Turn off the effect of the @code{--whole-archive} option for subsequent |
| 789 | archive files. |
| 790 | |
| 791 | @cindex output file after errors |
| 792 | @kindex --noinhibit-exec |
| 793 | @item --noinhibit-exec |
| 794 | Retain the executable output file whenever it is still usable. |
| 795 | Normally, the linker will not produce an output file if it encounters |
| 796 | errors during the link process; it exits without writing an output file |
| 797 | when it issues any error whatsoever. |
| 798 | |
| 799 | @ifclear SingleFormat |
| 800 | @kindex --oformat |
| 801 | @item --oformat @var{output-format} |
| 802 | @code{ld} may be configured to support more than one kind of object |
| 803 | file. If your @code{ld} is configured this way, you can use the |
| 804 | @samp{--oformat} option to specify the binary format for the output |
| 805 | object file. Even when @code{ld} is configured to support alternative |
| 806 | object formats, you don't usually need to specify this, as @code{ld} |
| 807 | should be configured to produce as a default output format the most |
| 808 | usual format on each machine. @var{output-format} is a text string, the |
| 809 | name of a particular format supported by the BFD libraries. (You can |
| 810 | list the available binary formats with @samp{objdump -i}.) The script |
| 811 | command @code{OUTPUT_FORMAT} can also specify the output format, but |
| 812 | this option overrides it. @xref{BFD}. |
| 813 | @end ifclear |
| 814 | |
| 815 | @kindex -qmagic |
| 816 | @item -qmagic |
| 817 | This option is ignored for Linux compatibility. |
| 818 | |
| 819 | @kindex -Qy |
| 820 | @item -Qy |
| 821 | This option is ignored for SVR4 compatibility. |
| 822 | |
| 823 | @kindex --relax |
| 824 | @cindex synthesizing linker |
| 825 | @cindex relaxing addressing modes |
| 826 | @item --relax |
| 827 | An option with machine dependent effects. |
| 828 | @ifset GENERIC |
| 829 | This option is only supported on a few targets. |
| 830 | @end ifset |
| 831 | @ifset H8300 |
| 832 | @xref{H8/300,,@code{ld} and the H8/300}. |
| 833 | @end ifset |
| 834 | @ifset I960 |
| 835 | @xref{i960,, @code{ld} and the Intel 960 family}. |
| 836 | @end ifset |
| 837 | |
| 838 | On some platforms, the @samp{--relax} option performs global |
| 839 | optimizations that become possible when the linker resolves addressing |
| 840 | in the program, such as relaxing address modes and synthesizing new |
| 841 | instructions in the output object file. |
| 842 | |
| 843 | @ifset GENERIC |
| 844 | On platforms where this is not supported, @samp{--relax} is accepted, |
| 845 | but ignored. |
| 846 | @end ifset |
| 847 | |
| 848 | @cindex retaining specified symbols |
| 849 | @cindex stripping all but some symbols |
| 850 | @cindex symbols, retaining selectively |
| 851 | @item --retain-symbols-file @var{filename} |
| 852 | Retain @emph{only} the symbols listed in the file @var{filename}, |
| 853 | discarding all others. @var{filename} is simply a flat file, with one |
| 854 | symbol name per line. This option is especially useful in environments |
| 855 | @ifset GENERIC |
| 856 | (such as VxWorks) |
| 857 | @end ifset |
| 858 | where a large global symbol table is accumulated gradually, to conserve |
| 859 | run-time memory. |
| 860 | |
| 861 | @samp{--retain-symbols-file} does @emph{not} discard undefined symbols, |
| 862 | or symbols needed for relocations. |
| 863 | |
| 864 | You may only specify @samp{--retain-symbols-file} once in the command |
| 865 | line. It overrides @samp{-s} and @samp{-S}. |
| 866 | |
| 867 | @ifset GENERIC |
| 868 | @item -rpath @var{dir} |
| 869 | @cindex runtime library search path |
| 870 | @kindex -rpath |
| 871 | Add a directory to the runtime library search path. This is used when |
| 872 | linking an ELF executable with shared objects. All @code{-rpath} |
| 873 | arguments are concatenated and passed to the runtime linker, which uses |
| 874 | them to locate shared objects at runtime. The @code{-rpath} option is |
| 875 | also used when locating shared objects which are needed by shared |
| 876 | objects explicitly included in the link; see the description of the |
| 877 | @code{-rpath-link} option. If @code{-rpath} is not used when linking an |
| 878 | ELF executable, the contents of the environment variable |
| 879 | @code{LD_RUN_PATH} will be used if it is defined. |
| 880 | |
| 881 | The @code{-rpath} option may also be used on SunOS. By default, on |
| 882 | SunOS, the linker will form a runtime search patch out of all the |
| 883 | @code{-L} options it is given. If a @code{-rpath} option is used, the |
| 884 | runtime search path will be formed exclusively using the @code{-rpath} |
| 885 | options, ignoring the @code{-L} options. This can be useful when using |
| 886 | gcc, which adds many @code{-L} options which may be on NFS mounted |
| 887 | filesystems. |
| 888 | |
| 889 | For compatibility with other ELF linkers, if the @code{-R} option is |
| 890 | followed by a directory name, rather than a file name, it is treated as |
| 891 | the @code{-rpath} option. |
| 892 | @end ifset |
| 893 | |
| 894 | @ifset GENERIC |
| 895 | @cindex link-time runtime library search path |
| 896 | @kindex -rpath-link |
| 897 | @item -rpath-link @var{DIR} |
| 898 | When using ELF or SunOS, one shared library may require another. This |
| 899 | happens when an @code{ld -shared} link includes a shared library as one |
| 900 | of the input files. |
| 901 | |
| 902 | When the linker encounters such a dependency when doing a non-shared, |
| 903 | non-relocateable link, it will automatically try to locate the required |
| 904 | shared library and include it in the link, if it is not included |
| 905 | explicitly. In such a case, the @code{-rpath-link} option |
| 906 | specifies the first set of directories to search. The |
| 907 | @code{-rpath-link} option may specify a sequence of directory names |
| 908 | either by specifying a list of names separated by colons, or by |
| 909 | appearing multiple times. |
| 910 | |
| 911 | The linker uses the following search paths to locate required shared |
| 912 | libraries. |
| 913 | @enumerate |
| 914 | @item |
| 915 | Any directories specified by @code{-rpath-link} options. |
| 916 | @item |
| 917 | Any directories specified by @code{-rpath} options. The difference |
| 918 | between @code{-rpath} and @code{-rpath-link} is that directories |
| 919 | specified by @code{-rpath} options are included in the executable and |
| 920 | used at runtime, whereas the @code{-rpath-link} option is only effective |
| 921 | at link time. |
| 922 | @item |
| 923 | On an ELF system, if the @code{-rpath} and @code{rpath-link} options |
| 924 | were not used, search the contents of the environment variable |
| 925 | @code{LD_RUN_PATH}. |
| 926 | @item |
| 927 | On SunOS, if the @code{-rpath} option was not used, search any |
| 928 | directories specified using @code{-L} options. |
| 929 | @item |
| 930 | For a native linker, the contents of the environment variable |
| 931 | @code{LD_LIBRARY_PATH}. |
| 932 | @item |
| 933 | The default directories, normally @file{/lib} and @file{/usr/lib}. |
| 934 | @end enumerate |
| 935 | |
| 936 | If the required shared library is not found, the linker will issue a |
| 937 | warning and continue with the link. |
| 938 | @end ifset |
| 939 | |
| 940 | @kindex -shared |
| 941 | @kindex -Bshareable |
| 942 | @item -shared |
| 943 | @itemx -Bshareable |
| 944 | @cindex shared libraries |
| 945 | Create a shared library. This is currently only supported on ELF, XCOFF |
| 946 | and SunOS platforms. On SunOS, the linker will automatically create a |
| 947 | shared library if the @code{-e} option is not used and there are |
| 948 | undefined symbols in the link. |
| 949 | |
| 950 | @item --sort-common |
| 951 | @kindex --sort-common |
| 952 | This option tells @code{ld} to sort the common symbols by size when it |
| 953 | places them in the appropriate output sections. First come all the one |
| 954 | byte symbols, then all the two bytes, then all the four bytes, and then |
| 955 | everything else. This is to prevent gaps between symbols due to |
| 956 | alignment constraints. |
| 957 | |
| 958 | @kindex --split-by-file |
| 959 | @item --split-by-file |
| 960 | Similar to @code{--split-by-reloc} but creates a new output section for |
| 961 | each input file. |
| 962 | |
| 963 | @kindex --split-by-reloc |
| 964 | @item --split-by-reloc @var{count} |
| 965 | Trys to creates extra sections in the output file so that no single |
| 966 | output section in the file contains more than @var{count} relocations. |
| 967 | This is useful when generating huge relocatable for downloading into |
| 968 | certain real time kernels with the COFF object file format; since COFF |
| 969 | cannot represent more than 65535 relocations in a single section. Note |
| 970 | that this will fail to work with object file formats which do not |
| 971 | support arbitrary sections. The linker will not split up individual |
| 972 | input sections for redistribution, so if a single input section contains |
| 973 | more than @var{count} relocations one output section will contain that |
| 974 | many relocations. |
| 975 | |
| 976 | @kindex --stats |
| 977 | @item --stats |
| 978 | Compute and display statistics about the operation of the linker, such |
| 979 | as execution time and memory usage. |
| 980 | |
| 981 | @kindex --traditional-format |
| 982 | @cindex traditional format |
| 983 | @item --traditional-format |
| 984 | For some targets, the output of @code{ld} is different in some ways from |
| 985 | the output of some existing linker. This switch requests @code{ld} to |
| 986 | use the traditional format instead. |
| 987 | |
| 988 | @cindex dbx |
| 989 | For example, on SunOS, @code{ld} combines duplicate entries in the |
| 990 | symbol string table. This can reduce the size of an output file with |
| 991 | full debugging information by over 30 percent. Unfortunately, the SunOS |
| 992 | @code{dbx} program can not read the resulting program (@code{gdb} has no |
| 993 | trouble). The @samp{--traditional-format} switch tells @code{ld} to not |
| 994 | combine duplicate entries. |
| 995 | |
| 996 | @kindex -Tbss @var{org} |
| 997 | @kindex -Tdata @var{org} |
| 998 | @kindex -Ttext @var{org} |
| 999 | @cindex segment origins, cmd line |
| 1000 | @item -Tbss @var{org} |
| 1001 | @itemx -Tdata @var{org} |
| 1002 | @itemx -Ttext @var{org} |
| 1003 | Use @var{org} as the starting address for---respectively---the |
| 1004 | @code{bss}, @code{data}, or the @code{text} segment of the output file. |
| 1005 | @var{org} must be a single hexadecimal integer; |
| 1006 | for compatibility with other linkers, you may omit the leading |
| 1007 | @samp{0x} usually associated with hexadecimal values. |
| 1008 | |
| 1009 | @kindex -Ur |
| 1010 | @cindex constructors |
| 1011 | @item -Ur |
| 1012 | For anything other than C++ programs, this option is equivalent to |
| 1013 | @samp{-r}: it generates relocatable output---i.e., an output file that can in |
| 1014 | turn serve as input to @code{ld}. When linking C++ programs, @samp{-Ur} |
| 1015 | @emph{does} resolve references to constructors, unlike @samp{-r}. |
| 1016 | It does not work to use @samp{-Ur} on files that were themselves linked |
| 1017 | with @samp{-Ur}; once the constructor table has been built, it cannot |
| 1018 | be added to. Use @samp{-Ur} only for the last partial link, and |
| 1019 | @samp{-r} for the others. |
| 1020 | |
| 1021 | @kindex --verbose |
| 1022 | @cindex verbose |
| 1023 | @item --verbose |
| 1024 | Display the version number for @code{ld} and list the linker emulations |
| 1025 | supported. Display which input files can and cannot be opened. Display |
| 1026 | the linker script if using a default builtin script. |
| 1027 | |
| 1028 | @kindex --warn-comon |
| 1029 | @cindex warnings, on combining symbols |
| 1030 | @cindex combining symbols, warnings on |
| 1031 | @item --warn-common |
| 1032 | Warn when a common symbol is combined with another common symbol or with |
| 1033 | a symbol definition. Unix linkers allow this somewhat sloppy practice, |
| 1034 | but linkers on some other operating systems do not. This option allows |
| 1035 | you to find potential problems from combining global symbols. |
| 1036 | Unfortunately, some C libraries use this practice, so you may get some |
| 1037 | warnings about symbols in the libraries as well as in your programs. |
| 1038 | |
| 1039 | There are three kinds of global symbols, illustrated here by C examples: |
| 1040 | |
| 1041 | @table @samp |
| 1042 | @item int i = 1; |
| 1043 | A definition, which goes in the initialized data section of the output |
| 1044 | file. |
| 1045 | |
| 1046 | @item extern int i; |
| 1047 | An undefined reference, which does not allocate space. |
| 1048 | There must be either a definition or a common symbol for the |
| 1049 | variable somewhere. |
| 1050 | |
| 1051 | @item int i; |
| 1052 | A common symbol. If there are only (one or more) common symbols for a |
| 1053 | variable, it goes in the uninitialized data area of the output file. |
| 1054 | The linker merges multiple common symbols for the same variable into a |
| 1055 | single symbol. If they are of different sizes, it picks the largest |
| 1056 | size. The linker turns a common symbol into a declaration, if there is |
| 1057 | a definition of the same variable. |
| 1058 | @end table |
| 1059 | |
| 1060 | The @samp{--warn-common} option can produce five kinds of warnings. |
| 1061 | Each warning consists of a pair of lines: the first describes the symbol |
| 1062 | just encountered, and the second describes the previous symbol |
| 1063 | encountered with the same name. One or both of the two symbols will be |
| 1064 | a common symbol. |
| 1065 | |
| 1066 | @enumerate |
| 1067 | @item |
| 1068 | Turning a common symbol into a reference, because there is already a |
| 1069 | definition for the symbol. |
| 1070 | @smallexample |
| 1071 | @var{file}(@var{section}): warning: common of `@var{symbol}' |
| 1072 | overridden by definition |
| 1073 | @var{file}(@var{section}): warning: defined here |
| 1074 | @end smallexample |
| 1075 | |
| 1076 | @item |
| 1077 | Turning a common symbol into a reference, because a later definition for |
| 1078 | the symbol is encountered. This is the same as the previous case, |
| 1079 | except that the symbols are encountered in a different order. |
| 1080 | @smallexample |
| 1081 | @var{file}(@var{section}): warning: definition of `@var{symbol}' |
| 1082 | overriding common |
| 1083 | @var{file}(@var{section}): warning: common is here |
| 1084 | @end smallexample |
| 1085 | |
| 1086 | @item |
| 1087 | Merging a common symbol with a previous same-sized common symbol. |
| 1088 | @smallexample |
| 1089 | @var{file}(@var{section}): warning: multiple common |
| 1090 | of `@var{symbol}' |
| 1091 | @var{file}(@var{section}): warning: previous common is here |
| 1092 | @end smallexample |
| 1093 | |
| 1094 | @item |
| 1095 | Merging a common symbol with a previous larger common symbol. |
| 1096 | @smallexample |
| 1097 | @var{file}(@var{section}): warning: common of `@var{symbol}' |
| 1098 | overridden by larger common |
| 1099 | @var{file}(@var{section}): warning: larger common is here |
| 1100 | @end smallexample |
| 1101 | |
| 1102 | @item |
| 1103 | Merging a common symbol with a previous smaller common symbol. This is |
| 1104 | the same as the previous case, except that the symbols are |
| 1105 | encountered in a different order. |
| 1106 | @smallexample |
| 1107 | @var{file}(@var{section}): warning: common of `@var{symbol}' |
| 1108 | overriding smaller common |
| 1109 | @var{file}(@var{section}): warning: smaller common is here |
| 1110 | @end smallexample |
| 1111 | @end enumerate |
| 1112 | |
| 1113 | @kindex --warn-constructors |
| 1114 | @item --warn-constructors |
| 1115 | Warn if any global constructors are used. This is only useful for a few |
| 1116 | object file formats. For formats like COFF or ELF, the linker can not |
| 1117 | detect the use of global constructors. |
| 1118 | |
| 1119 | @kindex --warn-multiple-gp |
| 1120 | @item --warn-multiple-gp |
| 1121 | Warn if multiple global pointer values are required in the output file. |
| 1122 | This is only meaningful for certain processors, such as the Alpha. |
| 1123 | Specifically, some processors put large-valued constants in a special |
| 1124 | section. A special register (the global pointer) points into the middle |
| 1125 | of this section, so that constants can be loaded efficiently via a |
| 1126 | base-register relative addressing mode. Since the offset in |
| 1127 | base-register relative mode is fixed and relatively small (e.g., 16 |
| 1128 | bits), this limits the maximum size of the constant pool. Thus, in |
| 1129 | large programs, it is often necessary to use multiple global pointer |
| 1130 | values in order to be able to address all possible constants. This |
| 1131 | option causes a warning to be issued whenever this case occurs. |
| 1132 | |
| 1133 | @kindex --warn-once |
| 1134 | @cindex warnings, on undefined symbols |
| 1135 | @cindex undefined symbols, warnings on |
| 1136 | @item --warn-once |
| 1137 | Only warn once for each undefined symbol, rather than once per module |
| 1138 | which refers to it. |
| 1139 | |
| 1140 | @kindex --warn-section-align |
| 1141 | @cindex warnings, on section alignment |
| 1142 | @cindex section alignment, warnings on |
| 1143 | @item --warn-section-align |
| 1144 | Warn if the address of an output section is changed because of |
| 1145 | alignment. Typically, the alignment will be set by an input section. |
| 1146 | The address will only be changed if it not explicitly specified; that |
| 1147 | is, if the @code{SECTIONS} command does not specify a start address for |
| 1148 | the section (@pxref{SECTIONS}). |
| 1149 | |
| 1150 | @kindex --whole-archive |
| 1151 | @cindex including an entire archive |
| 1152 | @item --whole-archive |
| 1153 | For each archive mentioned on the command line after the |
| 1154 | @code{--whole-archive} option, include every object file in the archive |
| 1155 | in the link, rather than searching the archive for the required object |
| 1156 | files. This is normally used to turn an archive file into a shared |
| 1157 | library, forcing every object to be included in the resulting shared |
| 1158 | library. This option may be used more than once. |
| 1159 | |
| 1160 | @kindex --wrap |
| 1161 | @item --wrap @var{symbol} |
| 1162 | Use a wrapper function for @var{symbol}. Any undefined reference to |
| 1163 | @var{symbol} will be resolved to @code{__wrap_@var{symbol}}. Any |
| 1164 | undefined reference to @code{__real_@var{symbol}} will be resolved to |
| 1165 | @var{symbol}. |
| 1166 | |
| 1167 | This can be used to provide a wrapper for a system function. The |
| 1168 | wrapper function should be called @code{__wrap_@var{symbol}}. If it |
| 1169 | wishes to call the system function, it should call |
| 1170 | @code{__real_@var{symbol}}. |
| 1171 | |
| 1172 | Here is a trivial example: |
| 1173 | |
| 1174 | @smallexample |
| 1175 | void * |
| 1176 | __wrap_malloc (int c) |
| 1177 | @{ |
| 1178 | printf ("malloc called with %ld\n", c); |
| 1179 | return __real_malloc (c); |
| 1180 | @} |
| 1181 | @end smallexample |
| 1182 | |
| 1183 | If you link other code with this file using @code{--wrap malloc}, then |
| 1184 | all calls to @code{malloc} will call the function @code{__wrap_malloc} |
| 1185 | instead. The call to @code{__real_malloc} in @code{__wrap_malloc} will |
| 1186 | call the real @code{malloc} function. |
| 1187 | |
| 1188 | You may wish to provide a @code{__real_malloc} function as well, so that |
| 1189 | links without the @code{--wrap} option will succeed. If you do this, |
| 1190 | you should not put the definition of @code{__real_malloc} in the same |
| 1191 | file as @code{__wrap_malloc}; if you do, the assembler may resolve the |
| 1192 | call before the linker has a chance to wrap it to @code{malloc}. |
| 1193 | |
| 1194 | @end table |
| 1195 | |
| 1196 | @ifset UsesEnvVars |
| 1197 | @node Environment |
| 1198 | @section Environment Variables |
| 1199 | |
| 1200 | You can change the behavior of @code{ld} with the environment |
| 1201 | variable @code{GNUTARGET}. |
| 1202 | |
| 1203 | @kindex GNUTARGET |
| 1204 | @cindex default input format |
| 1205 | @code{GNUTARGET} determines the input-file object format if you don't |
| 1206 | use @samp{-b} (or its synonym @samp{--format}). Its value should be one |
| 1207 | of the BFD names for an input format (@pxref{BFD}). If there is no |
| 1208 | @code{GNUTARGET} in the environment, @code{ld} uses the natural format |
| 1209 | of the target. If @code{GNUTARGET} is set to @code{default} then BFD |
| 1210 | attempts to discover the input format by examining binary input files; |
| 1211 | this method often succeeds, but there are potential ambiguities, since |
| 1212 | there is no method of ensuring that the magic number used to specify |
| 1213 | object-file formats is unique. However, the configuration procedure for |
| 1214 | BFD on each system places the conventional format for that system first |
| 1215 | in the search-list, so ambiguities are resolved in favor of convention. |
| 1216 | @end ifset |
| 1217 | |
| 1218 | @node Commands |
| 1219 | @chapter Command Language |
| 1220 | |
| 1221 | @cindex command files |
| 1222 | The command language provides explicit control over the link process, |
| 1223 | allowing complete specification of the mapping between the linker's |
| 1224 | input files and its output. It controls: |
| 1225 | @itemize @bullet |
| 1226 | @item |
| 1227 | input files |
| 1228 | @item |
| 1229 | file formats |
| 1230 | @item |
| 1231 | output file layout |
| 1232 | @item |
| 1233 | addresses of sections |
| 1234 | @item |
| 1235 | placement of common blocks |
| 1236 | @end itemize |
| 1237 | |
| 1238 | You may supply a command file (also known as a linker script) to the |
| 1239 | linker either explicitly through the @samp{-T} option, or implicitly as |
| 1240 | an ordinary file. Normally you should use the @samp{-T} option. An |
| 1241 | implicit linker script should only be used when you want to augment, |
| 1242 | rather than replace, the default linker script; typically an implicit |
| 1243 | linker script would consist only of @code{INPUT} or @code{GROUP} |
| 1244 | commands. |
| 1245 | |
| 1246 | If the linker opens a file which it cannot recognize as a supported |
| 1247 | object or archive format, nor as a linker script, it reports an error. |
| 1248 | |
| 1249 | @menu |
| 1250 | * Scripts:: Linker Scripts |
| 1251 | * Expressions:: Expressions |
| 1252 | * MEMORY:: MEMORY Command |
| 1253 | * SECTIONS:: SECTIONS Command |
| 1254 | * PHDRS:: PHDRS Command |
| 1255 | * Entry Point:: The Entry Point |
| 1256 | * Option Commands:: Option Commands |
| 1257 | @end menu |
| 1258 | |
| 1259 | @node Scripts |
| 1260 | @section Linker Scripts |
| 1261 | The @code{ld} command language is a collection of statements; some are |
| 1262 | simple keywords setting a particular option, some are used to select and |
| 1263 | group input files or name output files; and two statement |
| 1264 | types have a fundamental and pervasive impact on the linking process. |
| 1265 | |
| 1266 | @cindex fundamental script commands |
| 1267 | @cindex commands, fundamental |
| 1268 | @cindex output file layout |
| 1269 | @cindex layout of output file |
| 1270 | The most fundamental command of the @code{ld} command language is the |
| 1271 | @code{SECTIONS} command (@pxref{SECTIONS}). Every meaningful command |
| 1272 | script must have a @code{SECTIONS} command: it specifies a |
| 1273 | ``picture'' of the output file's layout, in varying degrees of detail. |
| 1274 | No other command is required in all cases. |
| 1275 | |
| 1276 | The @code{MEMORY} command complements @code{SECTIONS} by describing the |
| 1277 | available memory in the target architecture. This command is optional; |
| 1278 | if you don't use a @code{MEMORY} command, @code{ld} assumes sufficient |
| 1279 | memory is available in a contiguous block for all output. |
| 1280 | @xref{MEMORY}. |
| 1281 | |
| 1282 | @cindex comments |
| 1283 | You may include comments in linker scripts just as in C: delimited |
| 1284 | by @samp{/*} and @samp{*/}. As in C, comments are syntactically |
| 1285 | equivalent to whitespace. |
| 1286 | |
| 1287 | @node Expressions |
| 1288 | @section Expressions |
| 1289 | @cindex expression syntax |
| 1290 | @cindex arithmetic |
| 1291 | Many useful commands involve arithmetic expressions. The syntax for |
| 1292 | expressions in the command language is identical to that of C |
| 1293 | expressions, with the following features: |
| 1294 | @itemize @bullet |
| 1295 | @item |
| 1296 | All expressions evaluated as integers and |
| 1297 | are of ``long'' or ``unsigned long'' type. |
| 1298 | @item |
| 1299 | All constants are integers. |
| 1300 | @item |
| 1301 | All of the C arithmetic operators are provided. |
| 1302 | @item |
| 1303 | You may reference, define, and create global variables. |
| 1304 | @item |
| 1305 | You may call special purpose built-in functions. |
| 1306 | @end itemize |
| 1307 | |
| 1308 | @menu |
| 1309 | * Integers:: Integers |
| 1310 | * Symbols:: Symbol Names |
| 1311 | * Location Counter:: The Location Counter |
| 1312 | * Operators:: Operators |
| 1313 | * Evaluation:: Evaluation |
| 1314 | * Assignment:: Assignment: Defining Symbols |
| 1315 | * Arithmetic Functions:: Built-In Functions |
| 1316 | * Semicolons:: Semicolon Usage |
| 1317 | @end menu |
| 1318 | |
| 1319 | @node Integers |
| 1320 | @subsection Integers |
| 1321 | @cindex integer notation |
| 1322 | @cindex octal integers |
| 1323 | An octal integer is @samp{0} followed by zero or more of the octal |
| 1324 | digits (@samp{01234567}). |
| 1325 | @smallexample |
| 1326 | _as_octal = 0157255; |
| 1327 | @end smallexample |
| 1328 | |
| 1329 | @cindex decimal integers |
| 1330 | A decimal integer starts with a non-zero digit followed by zero or |
| 1331 | more digits (@samp{0123456789}). |
| 1332 | @smallexample |
| 1333 | _as_decimal = 57005; |
| 1334 | @end smallexample |
| 1335 | |
| 1336 | @cindex hexadecimal integers |
| 1337 | @kindex 0x |
| 1338 | A hexadecimal integer is @samp{0x} or @samp{0X} followed by one or |
| 1339 | more hexadecimal digits chosen from @samp{0123456789abcdefABCDEF}. |
| 1340 | @smallexample |
| 1341 | _as_hex = 0xdead; |
| 1342 | @end smallexample |
| 1343 | |
| 1344 | @cindex negative integers |
| 1345 | To write a negative integer, use |
| 1346 | the prefix operator @samp{-} (@pxref{Operators}). |
| 1347 | @smallexample |
| 1348 | _as_neg = -57005; |
| 1349 | @end smallexample |
| 1350 | |
| 1351 | @cindex scaled integers |
| 1352 | @cindex K and M integer suffixes |
| 1353 | @cindex M and K integer suffixes |
| 1354 | @cindex suffixes for integers |
| 1355 | @cindex integer suffixes |
| 1356 | Additionally the suffixes @code{K} and @code{M} may be used to scale a |
| 1357 | constant by |
| 1358 | @c TEXI2ROFF-KILL |
| 1359 | @ifinfo |
| 1360 | @c END TEXI2ROFF-KILL |
| 1361 | @code{1024} or @code{1024*1024} |
| 1362 | @c TEXI2ROFF-KILL |
| 1363 | @end ifinfo |
| 1364 | @tex |
| 1365 | ${\rm 1024}$ or ${\rm 1024}^2$ |
| 1366 | @end tex |
| 1367 | @c END TEXI2ROFF-KILL |
| 1368 | respectively. For example, the following all refer to the same quantity: |
| 1369 | |
| 1370 | @smallexample |
| 1371 | _fourk_1 = 4K; |
| 1372 | _fourk_2 = 4096; |
| 1373 | _fourk_3 = 0x1000; |
| 1374 | @end smallexample |
| 1375 | |
| 1376 | @node Symbols |
| 1377 | @subsection Symbol Names |
| 1378 | @cindex symbol names |
| 1379 | @cindex names |
| 1380 | @cindex quoted symbol names |
| 1381 | @kindex " |
| 1382 | Unless quoted, symbol names start with a letter, underscore, or point |
| 1383 | and may include any letters, underscores, digits, points, |
| 1384 | and hyphens. Unquoted symbol names must not conflict with any |
| 1385 | keywords. You can specify a symbol which contains odd characters or has |
| 1386 | the same name as a keyword, by surrounding the symbol name in double quotes: |
| 1387 | @smallexample |
| 1388 | "SECTION" = 9; |
| 1389 | "with a space" = "also with a space" + 10; |
| 1390 | @end smallexample |
| 1391 | |
| 1392 | Since symbols can contain many non-alphabetic characters, it is safest |
| 1393 | to delimit symbols with spaces. For example, @samp{A-B} is one symbol, |
| 1394 | whereas @samp{A - B} is an expression involving subtraction. |
| 1395 | |
| 1396 | @node Location Counter |
| 1397 | @subsection The Location Counter |
| 1398 | @kindex . |
| 1399 | @cindex dot |
| 1400 | @cindex location counter |
| 1401 | @cindex current output location |
| 1402 | The special linker variable @dfn{dot} @samp{.} always contains the |
| 1403 | current output location counter. Since the @code{.} always refers to |
| 1404 | a location in an output section, it must always appear in an |
| 1405 | expression within a @code{SECTIONS} command. The @code{.} symbol |
| 1406 | may appear anywhere that an ordinary symbol is allowed in an |
| 1407 | expression, but its assignments have a side effect. Assigning a value |
| 1408 | to the @code{.} symbol will cause the location counter to be moved. |
| 1409 | @cindex holes |
| 1410 | This may be used to create holes in the output section. The location |
| 1411 | counter may never be moved backwards. |
| 1412 | @smallexample |
| 1413 | SECTIONS |
| 1414 | @{ |
| 1415 | output : |
| 1416 | @{ |
| 1417 | file1(.text) |
| 1418 | . = . + 1000; |
| 1419 | file2(.text) |
| 1420 | . += 1000; |
| 1421 | file3(.text) |
| 1422 | @} = 0x1234; |
| 1423 | @} |
| 1424 | @end smallexample |
| 1425 | @noindent |
| 1426 | In the previous example, @code{file1} is located at the beginning of the |
| 1427 | output section, then there is a 1000 byte gap. Then @code{file2} |
| 1428 | appears, also with a 1000 byte gap following before @code{file3} is |
| 1429 | loaded. The notation @samp{= 0x1234} specifies what data to write in |
| 1430 | the gaps (@pxref{Section Options}). |
| 1431 | |
| 1432 | @iftex |
| 1433 | @vfill |
| 1434 | @end iftex |
| 1435 | |
| 1436 | @need 2000 |
| 1437 | @node Operators |
| 1438 | @subsection Operators |
| 1439 | @cindex Operators for arithmetic |
| 1440 | @cindex arithmetic operators |
| 1441 | @cindex precedence in expressions |
| 1442 | The linker recognizes the standard C set of arithmetic operators, with |
| 1443 | the standard bindings and precedence levels: |
| 1444 | @c TEXI2ROFF-KILL |
| 1445 | @ifinfo |
| 1446 | @c END TEXI2ROFF-KILL |
| 1447 | @smallexample |
| 1448 | precedence associativity Operators Notes |
| 1449 | (highest) |
| 1450 | 1 left ! - ~ (1) |
| 1451 | 2 left * / % |
| 1452 | 3 left + - |
| 1453 | 4 left >> << |
| 1454 | 5 left == != > < <= >= |
| 1455 | 6 left & |
| 1456 | 7 left | |
| 1457 | 8 left && |
| 1458 | 9 left || |
| 1459 | 10 right ? : |
| 1460 | 11 right &= += -= *= /= (2) |
| 1461 | (lowest) |
| 1462 | @end smallexample |
| 1463 | Notes: |
| 1464 | (1) Prefix operators |
| 1465 | (2) @xref{Assignment}. |
| 1466 | @c TEXI2ROFF-KILL |
| 1467 | @end ifinfo |
| 1468 | @tex |
| 1469 | \vskip \baselineskip |
| 1470 | %"lispnarrowing" is the extra indent used generally for smallexample |
| 1471 | \hskip\lispnarrowing\vbox{\offinterlineskip |
| 1472 | \hrule |
| 1473 | \halign |
| 1474 | {\vrule#&\strut\hfil\ #\ \hfil&\vrule#&\strut\hfil\ #\ \hfil&\vrule#&\strut\hfil\ {\tt #}\ \hfil&\vrule#\cr |
| 1475 | height2pt&\omit&&\omit&&\omit&\cr |
| 1476 | &Precedence&& Associativity &&{\rm Operators}&\cr |
| 1477 | height2pt&\omit&&\omit&&\omit&\cr |
| 1478 | \noalign{\hrule} |
| 1479 | height2pt&\omit&&\omit&&\omit&\cr |
| 1480 | &highest&&&&&\cr |
| 1481 | % '176 is tilde, '~' in tt font |
| 1482 | &1&&left&&\qquad- \char'176\ !\qquad\dag&\cr |
| 1483 | &2&&left&&* / \%&\cr |
| 1484 | &3&&left&&+ -&\cr |
| 1485 | &4&&left&&>> <<&\cr |
| 1486 | &5&&left&&== != > < <= >=&\cr |
| 1487 | &6&&left&&\&&\cr |
| 1488 | &7&&left&&|&\cr |
| 1489 | &8&&left&&{\&\&}&\cr |
| 1490 | &9&&left&&||&\cr |
| 1491 | &10&&right&&? :&\cr |
| 1492 | &11&&right&&\qquad\&= += -= *= /=\qquad\ddag&\cr |
| 1493 | &lowest&&&&&\cr |
| 1494 | height2pt&\omit&&\omit&&\omit&\cr} |
| 1495 | \hrule} |
| 1496 | @end tex |
| 1497 | @iftex |
| 1498 | { |
| 1499 | @obeylines@parskip=0pt@parindent=0pt |
| 1500 | @dag@quad Prefix operators. |
| 1501 | @ddag@quad @xref{Assignment}. |
| 1502 | } |
| 1503 | @end iftex |
| 1504 | @c END TEXI2ROFF-KILL |
| 1505 | |
| 1506 | @node Evaluation |
| 1507 | @subsection Evaluation |
| 1508 | |
| 1509 | @cindex lazy evaluation |
| 1510 | @cindex expression evaluation order |
| 1511 | The linker uses ``lazy evaluation'' for expressions; it only calculates |
| 1512 | an expression when absolutely necessary. The linker needs the value of |
| 1513 | the start address, and the lengths of memory regions, in order to do any |
| 1514 | linking at all; these values are computed as soon as possible when the |
| 1515 | linker reads in the command file. However, other values (such as symbol |
| 1516 | values) are not known or needed until after storage allocation. Such |
| 1517 | values are evaluated later, when other information (such as the sizes of |
| 1518 | output sections) is available for use in the symbol assignment |
| 1519 | expression. |
| 1520 | |
| 1521 | @node Assignment |
| 1522 | @subsection Assignment: Defining Symbols |
| 1523 | @cindex assignment in scripts |
| 1524 | @cindex symbol definition, scripts |
| 1525 | @cindex variables, defining |
| 1526 | You may create global symbols, and assign values (addresses) to global |
| 1527 | symbols, using any of the C assignment operators: |
| 1528 | |
| 1529 | @table @code |
| 1530 | @item @var{symbol} = @var{expression} ; |
| 1531 | @itemx @var{symbol} &= @var{expression} ; |
| 1532 | @itemx @var{symbol} += @var{expression} ; |
| 1533 | @itemx @var{symbol} -= @var{expression} ; |
| 1534 | @itemx @var{symbol} *= @var{expression} ; |
| 1535 | @itemx @var{symbol} /= @var{expression} ; |
| 1536 | @end table |
| 1537 | |
| 1538 | Two things distinguish assignment from other operators in @code{ld} |
| 1539 | expressions. |
| 1540 | @itemize @bullet |
| 1541 | @item |
| 1542 | Assignment may only be used at the root of an expression; |
| 1543 | @samp{a=b+3;} is allowed, but @samp{a+b=3;} is an error. |
| 1544 | |
| 1545 | @kindex ; |
| 1546 | @cindex semicolon |
| 1547 | @item |
| 1548 | You must place a trailing semicolon (``@key{;}'') at the end of an |
| 1549 | assignment statement. |
| 1550 | @end itemize |
| 1551 | |
| 1552 | Assignment statements may appear: |
| 1553 | @itemize @bullet |
| 1554 | @item |
| 1555 | as commands in their own right in an @code{ld} script; or |
| 1556 | @item |
| 1557 | as independent statements within a @code{SECTIONS} command; or |
| 1558 | @item |
| 1559 | as part of the contents of a section definition in a |
| 1560 | @code{SECTIONS} command. |
| 1561 | @end itemize |
| 1562 | |
| 1563 | The first two cases are equivalent in effect---both define a symbol with |
| 1564 | an absolute address. The last case defines a symbol whose address is |
| 1565 | relative to a particular section (@pxref{SECTIONS}). |
| 1566 | |
| 1567 | @cindex absolute and relocatable symbols |
| 1568 | @cindex relocatable and absolute symbols |
| 1569 | @cindex symbols, relocatable and absolute |
| 1570 | When a linker expression is evaluated and assigned to a variable, it is |
| 1571 | given either an absolute or a relocatable type. An absolute expression |
| 1572 | type is one in which the symbol contains the value that it will have in |
| 1573 | the output file; a relocatable expression type is one in which the |
| 1574 | value is expressed as a fixed offset from the base of a section. |
| 1575 | |
| 1576 | The type of the expression is controlled by its position in the script |
| 1577 | file. A symbol assigned within a section definition is created relative |
| 1578 | to the base of the section; a symbol assigned in any other place is |
| 1579 | created as an absolute symbol. Since a symbol created within a |
| 1580 | section definition is relative to the base of the section, it |
| 1581 | will remain relocatable if relocatable output is requested. A symbol |
| 1582 | may be created with an absolute value even when assigned to within a |
| 1583 | section definition by using the absolute assignment function |
| 1584 | @code{ABSOLUTE}. For example, to create an absolute symbol whose address |
| 1585 | is the last byte of an output section named @code{.data}: |
| 1586 | @smallexample |
| 1587 | SECTIONS@{ @dots{} |
| 1588 | .data : |
| 1589 | @{ |
| 1590 | *(.data) |
| 1591 | _edata = ABSOLUTE(.) ; |
| 1592 | @} |
| 1593 | @dots{} @} |
| 1594 | @end smallexample |
| 1595 | |
| 1596 | The linker tries to put off the evaluation of an assignment until all |
| 1597 | the terms in the source expression are known (@pxref{Evaluation}). For |
| 1598 | instance, the sizes of sections cannot be known until after allocation, |
| 1599 | so assignments dependent upon these are not performed until after |
| 1600 | allocation. Some expressions, such as those depending upon the location |
| 1601 | counter @dfn{dot}, @samp{.} must be evaluated during allocation. If the |
| 1602 | result of an expression is required, but the value is not available, |
| 1603 | then an error results. For example, a script like the following |
| 1604 | @smallexample |
| 1605 | SECTIONS @{ @dots{} |
| 1606 | text 9+this_isnt_constant : |
| 1607 | @{ @dots{} |
| 1608 | @} |
| 1609 | @dots{} @} |
| 1610 | @end smallexample |
| 1611 | @kindex Non constant expression |
| 1612 | @noindent |
| 1613 | will cause the error message ``@code{Non constant expression for initial |
| 1614 | address}''. |
| 1615 | |
| 1616 | @cindex provide |
| 1617 | In some cases, it is desirable for a linker script to define a symbol |
| 1618 | only if it is referenced, and only if it is not defined by any object |
| 1619 | included in the link. For example, traditional linkers defined the |
| 1620 | symbol @samp{etext}. However, ANSI C requires that the user be able to |
| 1621 | use @samp{etext} as a function name without encountering an error. |
| 1622 | The @code{PROVIDE} keyword may be used to define a symbol, such as |
| 1623 | @samp{etext}, only if it is referenced but not defined. The syntax is |
| 1624 | @code{PROVIDE(@var{symbol} = @var{expression})}. |
| 1625 | |
| 1626 | @node Arithmetic Functions |
| 1627 | @subsection Arithmetic Functions |
| 1628 | @cindex functions in expression language |
| 1629 | The command language includes a number of built-in |
| 1630 | functions for use in link script expressions. |
| 1631 | @table @code |
| 1632 | @kindex ABSOLUTE(@var{exp}) |
| 1633 | @cindex expression, absolute |
| 1634 | @item ABSOLUTE(@var{exp}) |
| 1635 | Return the absolute (non-relocatable, as opposed to non-negative) value |
| 1636 | of the expression @var{exp}. Primarily useful to assign an absolute |
| 1637 | value to a symbol within a section definition, where symbol values are |
| 1638 | normally section-relative. |
| 1639 | |
| 1640 | @kindex ADDR(@var{section}) |
| 1641 | @cindex section address |
| 1642 | @item ADDR(@var{section}) |
| 1643 | Return the absolute address of the named @var{section}. Your script must |
| 1644 | previously have defined the location of that section. In the following |
| 1645 | example, @code{symbol_1} and @code{symbol_2} are assigned identical |
| 1646 | values: |
| 1647 | @smallexample |
| 1648 | @group |
| 1649 | SECTIONS@{ @dots{} |
| 1650 | .output1 : |
| 1651 | @{ |
| 1652 | start_of_output_1 = ABSOLUTE(.); |
| 1653 | @dots{} |
| 1654 | @} |
| 1655 | .output : |
| 1656 | @{ |
| 1657 | symbol_1 = ADDR(.output1); |
| 1658 | symbol_2 = start_of_output_1; |
| 1659 | @} |
| 1660 | @dots{} @} |
| 1661 | @end group |
| 1662 | @end smallexample |
| 1663 | |
| 1664 | @kindex LOADADDR(@var{section}) |
| 1665 | @cindex section load address |
| 1666 | @item LOADADDR(@var{section}) |
| 1667 | Return the absolute load address of the named @var{section}. This is |
| 1668 | normally the same as @code{ADDR}, but it may be different if the |
| 1669 | @code{AT} keyword is used in the section definition (@pxref{Section |
| 1670 | Options}). |
| 1671 | |
| 1672 | @kindex ALIGN(@var{exp}) |
| 1673 | @cindex rounding up location counter |
| 1674 | @item ALIGN(@var{exp}) |
| 1675 | Return the result of the current location counter (@code{.}) aligned to |
| 1676 | the next @var{exp} boundary. @var{exp} must be an expression whose |
| 1677 | value is a power of two. This is equivalent to |
| 1678 | @smallexample |
| 1679 | (. + @var{exp} - 1) & ~(@var{exp} - 1) |
| 1680 | @end smallexample |
| 1681 | |
| 1682 | @code{ALIGN} doesn't change the value of the location counter---it just |
| 1683 | does arithmetic on it. As an example, to align the output @code{.data} |
| 1684 | section to the next @code{0x2000} byte boundary after the preceding |
| 1685 | section and to set a variable within the section to the next |
| 1686 | @code{0x8000} boundary after the input sections: |
| 1687 | @smallexample |
| 1688 | @group |
| 1689 | SECTIONS@{ @dots{} |
| 1690 | .data ALIGN(0x2000): @{ |
| 1691 | *(.data) |
| 1692 | variable = ALIGN(0x8000); |
| 1693 | @} |
| 1694 | @dots{} @} |
| 1695 | @end group |
| 1696 | @end smallexample |
| 1697 | @noindent |
| 1698 | The first use of @code{ALIGN} in this example specifies the location of |
| 1699 | a section because it is used as the optional @var{start} attribute of a |
| 1700 | section definition (@pxref{Section Options}). The second use simply |
| 1701 | defines the value of a variable. |
| 1702 | |
| 1703 | The built-in @code{NEXT} is closely related to @code{ALIGN}. |
| 1704 | |
| 1705 | @kindex DEFINED(@var{symbol}) |
| 1706 | @cindex symbol defaults |
| 1707 | @item DEFINED(@var{symbol}) |
| 1708 | Return 1 if @var{symbol} is in the linker global symbol table and is |
| 1709 | defined, otherwise return 0. You can use this function to provide default |
| 1710 | values for symbols. For example, the following command-file fragment shows how |
| 1711 | to set a global symbol @code{begin} to the first location in the |
| 1712 | @code{.text} section---but if a symbol called @code{begin} already |
| 1713 | existed, its value is preserved: |
| 1714 | |
| 1715 | @smallexample |
| 1716 | @group |
| 1717 | SECTIONS@{ @dots{} |
| 1718 | .text : @{ |
| 1719 | begin = DEFINED(begin) ? begin : . ; |
| 1720 | @dots{} |
| 1721 | @} |
| 1722 | @dots{} @} |
| 1723 | @end group |
| 1724 | @end smallexample |
| 1725 | |
| 1726 | @kindex NEXT(@var{exp}) |
| 1727 | @cindex unallocated address, next |
| 1728 | @item NEXT(@var{exp}) |
| 1729 | Return the next unallocated address that is a multiple of @var{exp}. |
| 1730 | This function is closely related to @code{ALIGN(@var{exp})}; unless you |
| 1731 | use the @code{MEMORY} command to define discontinuous memory for the |
| 1732 | output file, the two functions are equivalent. |
| 1733 | |
| 1734 | @kindex SIZEOF(@var{section}) |
| 1735 | @cindex section size |
| 1736 | @item SIZEOF(@var{section}) |
| 1737 | Return the size in bytes of the named @var{section}, if that section has |
| 1738 | been allocated. In the following example, @code{symbol_1} and |
| 1739 | @code{symbol_2} are assigned identical values: |
| 1740 | @c What does it return if the section hasn't been allocated? 0? |
| 1741 | @smallexample |
| 1742 | @group |
| 1743 | SECTIONS@{ @dots{} |
| 1744 | .output @{ |
| 1745 | .start = . ; |
| 1746 | @dots{} |
| 1747 | .end = . ; |
| 1748 | @} |
| 1749 | symbol_1 = .end - .start ; |
| 1750 | symbol_2 = SIZEOF(.output); |
| 1751 | @dots{} @} |
| 1752 | @end group |
| 1753 | @end smallexample |
| 1754 | |
| 1755 | @kindex SIZEOF_HEADERS |
| 1756 | @cindex header size |
| 1757 | @kindex sizeof_headers |
| 1758 | @item SIZEOF_HEADERS |
| 1759 | @itemx sizeof_headers |
| 1760 | Return the size in bytes of the output file's headers. You can use this number |
| 1761 | as the start address of the first section, if you choose, to facilitate |
| 1762 | paging. |
| 1763 | |
| 1764 | @kindex MAX |
| 1765 | @item MAX(@var{exp1}, @var{exp2}) |
| 1766 | Returns the maximum of @var{exp1} and @var{exp2}. |
| 1767 | |
| 1768 | @kindex MIN |
| 1769 | @item MIN(@var{exp1}, @var{exp2}) |
| 1770 | Returns the minimum of @var{exp1} and @var{exp2}. |
| 1771 | |
| 1772 | @end table |
| 1773 | |
| 1774 | @node Semicolons |
| 1775 | @subsection Semicolons |
| 1776 | |
| 1777 | Semicolons (``@key{;}'') are required in the following places. In all |
| 1778 | other places they can appear for aesthetic reasons but are otherwise ignored. |
| 1779 | |
| 1780 | @table @code |
| 1781 | @item Assignment |
| 1782 | Semicolons must appear at the end of assignment expressions. |
| 1783 | @xref{Assignment} |
| 1784 | |
| 1785 | @item PHDRS |
| 1786 | Semicolons must appear at the end of a @code{PHDRS} statement. |
| 1787 | @xref{PHDRS} |
| 1788 | @end table |
| 1789 | |
| 1790 | @node MEMORY |
| 1791 | @section Memory Layout |
| 1792 | @kindex MEMORY |
| 1793 | @cindex regions of memory |
| 1794 | @cindex discontinuous memory |
| 1795 | @cindex allocating memory |
| 1796 | The linker's default configuration permits allocation of all available memory. |
| 1797 | You can override this configuration by using the @code{MEMORY} command. The |
| 1798 | @code{MEMORY} command describes the location and size of blocks of |
| 1799 | memory in the target. By using it carefully, you can describe which |
| 1800 | memory regions may be used by the linker, and which memory regions it |
| 1801 | must avoid. The linker does not shuffle sections to fit into the |
| 1802 | available regions, but does move the requested sections into the correct |
| 1803 | regions and issue errors when the regions become too full. |
| 1804 | |
| 1805 | A command file may contain at most one use of the @code{MEMORY} |
| 1806 | command; however, you can define as many blocks of memory within it as |
| 1807 | you wish. The syntax is: |
| 1808 | |
| 1809 | @smallexample |
| 1810 | @group |
| 1811 | MEMORY |
| 1812 | @{ |
| 1813 | @var{name} (@var{attr}) : ORIGIN = @var{origin}, LENGTH = @var{len} |
| 1814 | @dots{} |
| 1815 | @} |
| 1816 | @end group |
| 1817 | @end smallexample |
| 1818 | @table @code |
| 1819 | @cindex naming memory regions |
| 1820 | @item @var{name} |
| 1821 | is a name used internally by the linker to refer to the region. Any |
| 1822 | symbol name may be used. The region names are stored in a separate |
| 1823 | name space, and will not conflict with symbols, file names or section |
| 1824 | names. Use distinct names to specify multiple regions. |
| 1825 | |
| 1826 | @cindex memory region attributes |
| 1827 | @item (@var{attr}) |
| 1828 | is an optional list of attributes, permitted for compatibility with the |
| 1829 | AT&T linker but not used by @code{ld} beyond checking that the |
| 1830 | attribute list is valid. Valid attribute lists must be made up of the |
| 1831 | characters ``@code{LIRWX}''. If you omit the attribute list, you may |
| 1832 | omit the parentheses around it as well. |
| 1833 | |
| 1834 | @kindex ORIGIN = |
| 1835 | @kindex o = |
| 1836 | @kindex org = |
| 1837 | @item @var{origin} |
| 1838 | is the start address of the region in physical memory. It is |
| 1839 | an expression that must evaluate to a constant before |
| 1840 | memory allocation is performed. The keyword @code{ORIGIN} may be |
| 1841 | abbreviated to @code{org} or @code{o} (but not, for example, @samp{ORG}). |
| 1842 | |
| 1843 | @kindex LENGTH = |
| 1844 | @kindex len = |
| 1845 | @kindex l = |
| 1846 | @item @var{len} |
| 1847 | is the size in bytes of the region (an expression). |
| 1848 | The keyword @code{LENGTH} may be abbreviated to @code{len} or @code{l}. |
| 1849 | @end table |
| 1850 | |
| 1851 | For example, to specify that memory has two regions available for |
| 1852 | allocation---one starting at 0 for 256 kilobytes, and the other |
| 1853 | starting at @code{0x40000000} for four megabytes: |
| 1854 | |
| 1855 | @smallexample |
| 1856 | @group |
| 1857 | MEMORY |
| 1858 | @{ |
| 1859 | rom : ORIGIN = 0, LENGTH = 256K |
| 1860 | ram : org = 0x40000000, l = 4M |
| 1861 | @} |
| 1862 | @end group |
| 1863 | @end smallexample |
| 1864 | |
| 1865 | Once you have defined a region of memory named @var{mem}, you can direct |
| 1866 | specific output sections there by using a command ending in |
| 1867 | @samp{>@var{mem}} within the @code{SECTIONS} command (@pxref{Section |
| 1868 | Options}). If the combined output sections directed to a region are too |
| 1869 | big for the region, the linker will issue an error message. |
| 1870 | |
| 1871 | @node SECTIONS |
| 1872 | @section Specifying Output Sections |
| 1873 | |
| 1874 | @kindex SECTIONS |
| 1875 | The @code{SECTIONS} command controls exactly where input sections are |
| 1876 | placed into output sections, their order in the output file, and to |
| 1877 | which output sections they are allocated. |
| 1878 | |
| 1879 | You may use at most one @code{SECTIONS} command in a script file, |
| 1880 | but you can have as many statements within it as you wish. Statements |
| 1881 | within the @code{SECTIONS} command can do one of three things: |
| 1882 | |
| 1883 | @itemize @bullet |
| 1884 | @item |
| 1885 | define the entry point; |
| 1886 | |
| 1887 | @item |
| 1888 | assign a value to a symbol; |
| 1889 | |
| 1890 | @item |
| 1891 | describe the placement of a named output section, and which input |
| 1892 | sections go into it. |
| 1893 | @end itemize |
| 1894 | |
| 1895 | You can also use the first two operations---defining the entry point and |
| 1896 | defining symbols---outside the @code{SECTIONS} command: @pxref{Entry |
| 1897 | Point}, and @ref{Assignment}. They are permitted here as well for |
| 1898 | your convenience in reading the script, so that symbols and the entry |
| 1899 | point can be defined at meaningful points in your output-file layout. |
| 1900 | |
| 1901 | If you do not use a @code{SECTIONS} command, the linker places each input |
| 1902 | section into an identically named output section in the order that the |
| 1903 | sections are first encountered in the input files. If all input sections |
| 1904 | are present in the first file, for example, the order of sections in the |
| 1905 | output file will match the order in the first input file. |
| 1906 | |
| 1907 | @menu |
| 1908 | * Section Definition:: Section Definitions |
| 1909 | * Section Placement:: Section Placement |
| 1910 | * Section Data Expressions:: Section Data Expressions |
| 1911 | * Section Options:: Optional Section Attributes |
| 1912 | * Overlays:: Overlays |
| 1913 | @end menu |
| 1914 | |
| 1915 | @node Section Definition |
| 1916 | @subsection Section Definitions |
| 1917 | @cindex section definition |
| 1918 | The most frequently used statement in the @code{SECTIONS} command is |
| 1919 | the @dfn{section definition}, which specifies the |
| 1920 | properties of an output section: its location, alignment, contents, |
| 1921 | fill pattern, and target memory region. Most of |
| 1922 | these specifications are optional; the simplest form of a section |
| 1923 | definition is |
| 1924 | @smallexample |
| 1925 | SECTIONS @{ @dots{} |
| 1926 | @var{secname} : @{ |
| 1927 | @var{contents} |
| 1928 | @} |
| 1929 | @dots{} @} |
| 1930 | @end smallexample |
| 1931 | @cindex naming output sections |
| 1932 | @noindent |
| 1933 | @var{secname} is the name of the output section, and @var{contents} a |
| 1934 | specification of what goes there---for example, a list of input files or |
| 1935 | sections of input files (@pxref{Section Placement}). As you might |
| 1936 | assume, the whitespace shown is optional. You do need the colon |
| 1937 | @samp{:} and the braces @samp{@{@}}, however. |
| 1938 | |
| 1939 | @var{secname} must meet the constraints of your output format. In |
| 1940 | formats which only support a limited number of sections, such as |
| 1941 | @code{a.out}, the name must be one of the names supported by the format |
| 1942 | (@code{a.out}, for example, allows only @code{.text}, @code{.data} or |
| 1943 | @code{.bss}). If the output format supports any number of sections, but |
| 1944 | with numbers and not names (as is the case for Oasys), the name should be |
| 1945 | supplied as a quoted numeric string. A section name may consist of any |
| 1946 | sequence of characters, but any name which does not conform to the standard |
| 1947 | @code{ld} symbol name syntax must be quoted. |
| 1948 | @xref{Symbols, , Symbol Names}. |
| 1949 | |
| 1950 | The special @var{secname} @samp{/DISCARD/} may be used to discard input |
| 1951 | sections. Any sections which are assigned to an output section named |
| 1952 | @samp{/DISCARD/} are not included in the final link output. |
| 1953 | |
| 1954 | The linker will not create output sections which do not have any |
| 1955 | contents. This is for convenience when referring to input sections that |
| 1956 | may or may not exist. For example, |
| 1957 | @smallexample |
| 1958 | .foo @{ *(.foo) @} |
| 1959 | @end smallexample |
| 1960 | will only create a @samp{.foo} section in the output file if there is a |
| 1961 | @samp{.foo} section in at least one input file. |
| 1962 | |
| 1963 | @node Section Placement |
| 1964 | @subsection Section Placement |
| 1965 | |
| 1966 | @cindex contents of a section |
| 1967 | In a section definition, you can specify the contents of an output |
| 1968 | section by listing particular input files, by listing particular |
| 1969 | input-file sections, or by a combination of the two. You can also place |
| 1970 | arbitrary data in the section, and define symbols relative to the |
| 1971 | beginning of the section. |
| 1972 | |
| 1973 | The @var{contents} of a section definition may include any of the |
| 1974 | following kinds of statement. You can include as many of these as you |
| 1975 | like in a single section definition, separated from one another by |
| 1976 | whitespace. |
| 1977 | |
| 1978 | @table @code |
| 1979 | @kindex @var{filename} |
| 1980 | @cindex input files, section defn |
| 1981 | @cindex files, including in output sections |
| 1982 | @item @var{filename} |
| 1983 | You may simply name a particular input file to be placed in the current |
| 1984 | output section; @emph{all} sections from that file are placed in the |
| 1985 | current section definition. If the file name has already been mentioned |
| 1986 | in another section definition, with an explicit section name list, then |
| 1987 | only those sections which have not yet been allocated are used. |
| 1988 | |
| 1989 | To specify a list of particular files by name: |
| 1990 | @smallexample |
| 1991 | .data : @{ afile.o bfile.o cfile.o @} |
| 1992 | @end smallexample |
| 1993 | @noindent |
| 1994 | The example also illustrates that multiple statements can be included in |
| 1995 | the contents of a section definition, since each file name is a separate |
| 1996 | statement. |
| 1997 | |
| 1998 | @kindex @var{filename}(@var{section}) |
| 1999 | @cindex files and sections, section defn |
| 2000 | @item @var{filename}( @var{section} ) |
| 2001 | @itemx @var{filename}( @var{section} , @var{section}, @dots{} ) |
| 2002 | @itemx @var{filename}( @var{section} @var{section} @dots{} ) |
| 2003 | You can name one or more sections from your input files, for |
| 2004 | insertion in the current output section. If you wish to specify a list |
| 2005 | of input-file sections inside the parentheses, you may separate the |
| 2006 | section names by either commas or whitespace. |
| 2007 | |
| 2008 | @cindex input sections to output section |
| 2009 | @kindex *(@var{section}) |
| 2010 | @item * (@var{section}) |
| 2011 | @itemx * (@var{section}, @var{section}, @dots{}) |
| 2012 | @itemx * (@var{section} @var{section} @dots{}) |
| 2013 | Instead of explicitly naming particular input files in a link control |
| 2014 | script, you can refer to @emph{all} files from the @code{ld} command |
| 2015 | line: use @samp{*} instead of a particular file name before the |
| 2016 | parenthesized input-file section list. |
| 2017 | |
| 2018 | If you have already explicitly included some files by name, @samp{*} |
| 2019 | refers to all @emph{remaining} files---those whose places in the output |
| 2020 | file have not yet been defined. |
| 2021 | |
| 2022 | For example, to copy sections @code{1} through @code{4} from an Oasys file |
| 2023 | into the @code{.text} section of an @code{a.out} file, and sections @code{13} |
| 2024 | and @code{14} into the @code{.data} section: |
| 2025 | @smallexample |
| 2026 | @group |
| 2027 | SECTIONS @{ |
| 2028 | .text :@{ |
| 2029 | *("1" "2" "3" "4") |
| 2030 | @} |
| 2031 | |
| 2032 | .data :@{ |
| 2033 | *("13" "14") |
| 2034 | @} |
| 2035 | @} |
| 2036 | @end group |
| 2037 | @end smallexample |
| 2038 | |
| 2039 | @cindex @code{[@var{section}@dots{}]}, not supported |
| 2040 | @samp{[ @var{section} @dots{} ]} used to be accepted as an alternate way |
| 2041 | to specify named sections from all unallocated input files. Because |
| 2042 | some operating systems (VMS) allow brackets in file names, that notation |
| 2043 | is no longer supported. |
| 2044 | |
| 2045 | @cindex uninitialized data |
| 2046 | @cindex commons in output |
| 2047 | @kindex *( COMMON ) |
| 2048 | @item @var{filename}@code{( COMMON )} |
| 2049 | @itemx *( COMMON ) |
| 2050 | Specify where in your output file to place uninitialized data |
| 2051 | with this notation. @code{*(COMMON)} by itself refers to all |
| 2052 | uninitialized data from all input files (so far as it is not yet |
| 2053 | allocated); @var{filename}@code{(COMMON)} refers to uninitialized data |
| 2054 | from a particular file. Both are special cases of the general |
| 2055 | mechanisms for specifying where to place input-file sections: |
| 2056 | @code{ld} permits you to refer to uninitialized data as if it |
| 2057 | were in an input-file section named @code{COMMON}, regardless of the |
| 2058 | input file's format. |
| 2059 | @end table |
| 2060 | |
| 2061 | In any place where you may use a specific file or section name, you may |
| 2062 | also use a wildcard pattern. The linker handles wildcards much as the |
| 2063 | Unix shell does. A @samp{*} character matches any number of characters. |
| 2064 | A @samp{?} character matches any single character. The sequence |
| 2065 | @samp{[@var{chars}]} will match a single instance of any of the |
| 2066 | @var{chars}; the @samp{-} character may be used to specify a range of |
| 2067 | characters, as in @samp{[a-z]} to match any lower case letter. A |
| 2068 | @samp{\} character may be used to quote the following character. |
| 2069 | |
| 2070 | When a file name is matched with a wildcard, the wildcard characters |
| 2071 | will not match a @samp{/} character (used to separate directory names on |
| 2072 | Unix). A pattern consisting of a single @samp{*} character is an |
| 2073 | exception; it will always match any file name. In a section name, the |
| 2074 | wildcard characters will match a @samp{/} character. |
| 2075 | |
| 2076 | Wildcards only match files which are explicitly specified on the command |
| 2077 | line. The linker does not search directories to expand wildcards. |
| 2078 | However, if you specify a simple file name---a name with no wildcard |
| 2079 | characters---in a linker script, and the file name is not also specified |
| 2080 | on the command line, the linker will attempt to open the file as though |
| 2081 | it appeared on the command line. |
| 2082 | |
| 2083 | In the following example, the command script arranges the output file |
| 2084 | into three consecutive sections, named @code{.text}, @code{.data}, and |
| 2085 | @code{.bss}, taking the input for each from the correspondingly named |
| 2086 | sections of all the input files: |
| 2087 | |
| 2088 | @smallexample |
| 2089 | @group |
| 2090 | SECTIONS @{ |
| 2091 | .text : @{ *(.text) @} |
| 2092 | .data : @{ *(.data) @} |
| 2093 | .bss : @{ *(.bss) *(COMMON) @} |
| 2094 | @} |
| 2095 | @end group |
| 2096 | @end smallexample |
| 2097 | |
| 2098 | The following example reads all of the sections from file @code{all.o} |
| 2099 | and places them at the start of output section @code{outputa} which |
| 2100 | starts at location @code{0x10000}. All of section @code{.input1} from |
| 2101 | file @code{foo.o} follows immediately, in the same output section. All |
| 2102 | of section @code{.input2} from @code{foo.o} goes into output section |
| 2103 | @code{outputb}, followed by section @code{.input1} from @code{foo1.o}. |
| 2104 | All of the remaining @code{.input1} and @code{.input2} sections from any |
| 2105 | files are written to output section @code{outputc}. |
| 2106 | |
| 2107 | @smallexample |
| 2108 | @group |
| 2109 | SECTIONS @{ |
| 2110 | outputa 0x10000 : |
| 2111 | @{ |
| 2112 | all.o |
| 2113 | foo.o (.input1) |
| 2114 | @} |
| 2115 | outputb : |
| 2116 | @{ |
| 2117 | foo.o (.input2) |
| 2118 | foo1.o (.input1) |
| 2119 | @} |
| 2120 | outputc : |
| 2121 | @{ |
| 2122 | *(.input1) |
| 2123 | *(.input2) |
| 2124 | @} |
| 2125 | @} |
| 2126 | @end group |
| 2127 | @end smallexample |
| 2128 | |
| 2129 | This example shows how wildcard patterns might be used to partition |
| 2130 | files. All @code{.text} sections are placed in @code{.text}, and all |
| 2131 | @code{.bss} sections are placed in @code{.bss}. For all files beginning |
| 2132 | with an upper case character, the @code{.data} section is placed into |
| 2133 | @code{.DATA}; for all other files, the @code{.data} section is placed |
| 2134 | into @code{.data}. |
| 2135 | |
| 2136 | @smallexample |
| 2137 | @group |
| 2138 | SECTIONS @{ |
| 2139 | .text : @{ *(.text) @} |
| 2140 | .DATA : @{ [A-Z]*(.data) @} |
| 2141 | .data : @{ *(.data) @} |
| 2142 | .bss : @{ *(.bss) @} |
| 2143 | @} |
| 2144 | @end group |
| 2145 | @end smallexample |
| 2146 | |
| 2147 | @node Section Data Expressions |
| 2148 | @subsection Section Data Expressions |
| 2149 | |
| 2150 | @cindex expressions in a section |
| 2151 | The foregoing statements arrange, in your output file, data originating |
| 2152 | from your input files. You can also place data directly in an output |
| 2153 | section from the link command script. Most of these additional |
| 2154 | statements involve expressions (@pxref{Expressions}). Although these |
| 2155 | statements are shown separately here for ease of presentation, no such |
| 2156 | segregation is needed within a section definition in the @code{SECTIONS} |
| 2157 | command; you can intermix them freely with any of the statements we've |
| 2158 | just described. |
| 2159 | |
| 2160 | @table @code |
| 2161 | @cindex input filename symbols |
| 2162 | @cindex filename symbols |
| 2163 | @kindex CREATE_OBJECT_SYMBOLS |
| 2164 | @item CREATE_OBJECT_SYMBOLS |
| 2165 | Create a symbol for each input file |
| 2166 | in the current section, set to the address of the first byte of |
| 2167 | data written from that input file. For instance, with @code{a.out} |
| 2168 | files it is conventional to have a symbol for each input file. You can |
| 2169 | accomplish this by defining the output @code{.text} section as follows: |
| 2170 | @smallexample |
| 2171 | @group |
| 2172 | SECTIONS @{ |
| 2173 | .text 0x2020 : |
| 2174 | @{ |
| 2175 | CREATE_OBJECT_SYMBOLS |
| 2176 | *(.text) |
| 2177 | _etext = ALIGN(0x2000); |
| 2178 | @} |
| 2179 | @dots{} |
| 2180 | @} |
| 2181 | @end group |
| 2182 | @end smallexample |
| 2183 | |
| 2184 | If @code{sample.ld} is a file containing this script, and @code{a.o}, |
| 2185 | @code{b.o}, @code{c.o}, and @code{d.o} are four input files with |
| 2186 | contents like the following--- |
| 2187 | @smallexample |
| 2188 | @group |
| 2189 | /* a.c */ |
| 2190 | |
| 2191 | afunction() @{ @} |
| 2192 | int adata=1; |
| 2193 | int abss; |
| 2194 | @end group |
| 2195 | @end smallexample |
| 2196 | |
| 2197 | @noindent |
| 2198 | @samp{ld -M -T sample.ld a.o b.o c.o d.o} would create a map like this, |
| 2199 | containing symbols matching the object file names: |
| 2200 | @smallexample |
| 2201 | 00000000 A __DYNAMIC |
| 2202 | 00004020 B _abss |
| 2203 | 00004000 D _adata |
| 2204 | 00002020 T _afunction |
| 2205 | 00004024 B _bbss |
| 2206 | 00004008 D _bdata |
| 2207 | 00002038 T _bfunction |
| 2208 | 00004028 B _cbss |
| 2209 | 00004010 D _cdata |
| 2210 | 00002050 T _cfunction |
| 2211 | 0000402c B _dbss |
| 2212 | 00004018 D _ddata |
| 2213 | 00002068 T _dfunction |
| 2214 | 00004020 D _edata |
| 2215 | 00004030 B _end |
| 2216 | 00004000 T _etext |
| 2217 | 00002020 t a.o |
| 2218 | 00002038 t b.o |
| 2219 | 00002050 t c.o |
| 2220 | 00002068 t d.o |
| 2221 | @end smallexample |
| 2222 | |
| 2223 | @kindex @var{symbol} = @var{expression} ; |
| 2224 | @kindex @var{symbol} @var{f}= @var{expression} ; |
| 2225 | @item @var{symbol} = @var{expression} ; |
| 2226 | @itemx @var{symbol} @var{f}= @var{expression} ; |
| 2227 | @var{symbol} is any symbol name (@pxref{Symbols}). ``@var{f}='' |
| 2228 | refers to any of the operators @code{&= += -= *= /=} which combine |
| 2229 | arithmetic and assignment. |
| 2230 | |
| 2231 | @cindex assignment, in section defn |
| 2232 | When you assign a value to a symbol within a particular section |
| 2233 | definition, the value is relative to the beginning of the section |
| 2234 | (@pxref{Assignment}). If you write |
| 2235 | |
| 2236 | @smallexample |
| 2237 | @group |
| 2238 | SECTIONS @{ |
| 2239 | abs = 14 ; |
| 2240 | @dots{} |
| 2241 | .data : @{ @dots{} rel = 14 ; @dots{} @} |
| 2242 | abs2 = 14 + ADDR(.data); |
| 2243 | @dots{} |
| 2244 | @} |
| 2245 | @end group |
| 2246 | @end smallexample |
| 2247 | |
| 2248 | @c FIXME: Try above example! |
| 2249 | @noindent |
| 2250 | @code{abs} and @code{rel} do not have the same value; @code{rel} has the |
| 2251 | same value as @code{abs2}. |
| 2252 | |
| 2253 | @kindex BYTE(@var{expression}) |
| 2254 | @kindex SHORT(@var{expression}) |
| 2255 | @kindex LONG(@var{expression}) |
| 2256 | @kindex QUAD(@var{expression}) |
| 2257 | @cindex direct output |
| 2258 | @item BYTE(@var{expression}) |
| 2259 | @itemx SHORT(@var{expression}) |
| 2260 | @itemx LONG(@var{expression}) |
| 2261 | @itemx QUAD(@var{expression}) |
| 2262 | By including one of these four statements in a section definition, you |
| 2263 | can explicitly place one, two, four, or eight bytes (respectively) at |
| 2264 | the current address of that section. @code{QUAD} is only supported when |
| 2265 | using a 64 bit host or target. |
| 2266 | |
| 2267 | @ifclear SingleFormat |
| 2268 | Multiple-byte quantities are represented in whatever byte order is |
| 2269 | appropriate for the output file format (@pxref{BFD}). |
| 2270 | @end ifclear |
| 2271 | |
| 2272 | @kindex FILL(@var{expression}) |
| 2273 | @cindex holes, filling |
| 2274 | @cindex unspecified memory |
| 2275 | @item FILL(@var{expression}) |
| 2276 | Specify the ``fill pattern'' for the current section. Any otherwise |
| 2277 | unspecified regions of memory within the section (for example, regions |
| 2278 | you skip over by assigning a new value to the location counter @samp{.}) |
| 2279 | are filled with the two least significant bytes from the |
| 2280 | @var{expression} argument. A @code{FILL} statement covers memory |
| 2281 | locations @emph{after} the point it occurs in the section definition; by |
| 2282 | including more than one @code{FILL} statement, you can have different |
| 2283 | fill patterns in different parts of an output section. |
| 2284 | @end table |
| 2285 | |
| 2286 | @node Section Options |
| 2287 | @subsection Optional Section Attributes |
| 2288 | @cindex section defn, full syntax |
| 2289 | Here is the full syntax of a section definition, including all the |
| 2290 | optional portions: |
| 2291 | |
| 2292 | @smallexample |
| 2293 | @group |
| 2294 | SECTIONS @{ |
| 2295 | @dots{} |
| 2296 | @var{secname} @var{start} BLOCK(@var{align}) (NOLOAD) : AT ( @var{ldadr} ) |
| 2297 | @{ @var{contents} @} >@var{region} :@var{phdr} =@var{fill} |
| 2298 | @dots{} |
| 2299 | @} |
| 2300 | @end group |
| 2301 | @end smallexample |
| 2302 | |
| 2303 | @var{secname} and @var{contents} are required. @xref{Section |
| 2304 | Definition}, and @ref{Section Placement}, for details on |
| 2305 | @var{contents}. The remaining elements---@var{start}, |
| 2306 | @code{BLOCK(@var{align)}}, @code{(NOLOAD)}, @code{AT ( @var{ldadr} )}, |
| 2307 | @code{>@var{region}}, @code{:@var{phdr}}, and @code{=@var{fill}}---are |
| 2308 | all optional. |
| 2309 | |
| 2310 | @table @code |
| 2311 | @cindex start address, section |
| 2312 | @cindex section start |
| 2313 | @cindex section address |
| 2314 | @item @var{start} |
| 2315 | You can force the output section to be loaded at a specified address by |
| 2316 | specifying @var{start} immediately following the section name. |
| 2317 | @var{start} can be represented as any expression. The following |
| 2318 | example generates section @var{output} at location |
| 2319 | @code{0x40000000}: |
| 2320 | |
| 2321 | @smallexample |
| 2322 | @group |
| 2323 | SECTIONS @{ |
| 2324 | @dots{} |
| 2325 | output 0x40000000: @{ |
| 2326 | @dots{} |
| 2327 | @} |
| 2328 | @dots{} |
| 2329 | @} |
| 2330 | @end group |
| 2331 | @end smallexample |
| 2332 | |
| 2333 | @kindex BLOCK(@var{align}) |
| 2334 | @cindex section alignment |
| 2335 | @cindex aligning sections |
| 2336 | @item BLOCK(@var{align}) |
| 2337 | You can include @code{BLOCK()} specification to advance |
| 2338 | the location counter @code{.} prior to the beginning of the section, so |
| 2339 | that the section will begin at the specified alignment. @var{align} is |
| 2340 | an expression. |
| 2341 | |
| 2342 | @kindex NOLOAD |
| 2343 | @cindex prevent unnecessary loading |
| 2344 | @cindex loading, preventing |
| 2345 | @item (NOLOAD) |
| 2346 | Use @samp{(NOLOAD)} to prevent a section from being loaded into memory |
| 2347 | each time it is accessed. For example, in the script sample below, the |
| 2348 | @code{ROM} segment is addressed at memory location @samp{0} and does not |
| 2349 | need to be loaded into each object file: |
| 2350 | |
| 2351 | @smallexample |
| 2352 | @group |
| 2353 | SECTIONS @{ |
| 2354 | ROM 0 (NOLOAD) : @{ @dots{} @} |
| 2355 | @dots{} |
| 2356 | @} |
| 2357 | @end group |
| 2358 | @end smallexample |
| 2359 | |
| 2360 | @kindex AT ( @var{ldadr} ) |
| 2361 | @cindex specify load address |
| 2362 | @cindex load address, specifying |
| 2363 | @item AT ( @var{ldadr} ) |
| 2364 | The expression @var{ldadr} that follows the @code{AT} keyword specifies |
| 2365 | the load address of the section. The default (if you do not use the |
| 2366 | @code{AT} keyword) is to make the load address the same as the |
| 2367 | relocation address. This feature is designed to make it easy to build a |
| 2368 | ROM image. For example, this @code{SECTIONS} definition creates two |
| 2369 | output sections: one called @samp{.text}, which starts at @code{0x1000}, |
| 2370 | and one called @samp{.mdata}, which is loaded at the end of the |
| 2371 | @samp{.text} section even though its relocation address is |
| 2372 | @code{0x2000}. The symbol @code{_data} is defined with the value |
| 2373 | @code{0x2000}: |
| 2374 | |
| 2375 | @smallexample |
| 2376 | @group |
| 2377 | SECTIONS |
| 2378 | @{ |
| 2379 | .text 0x1000 : @{ *(.text) _etext = . ; @} |
| 2380 | .mdata 0x2000 : |
| 2381 | AT ( ADDR(.text) + SIZEOF ( .text ) ) |
| 2382 | @{ _data = . ; *(.data); _edata = . ; @} |
| 2383 | .bss 0x3000 : |
| 2384 | @{ _bstart = . ; *(.bss) *(COMMON) ; _bend = . ;@} |
| 2385 | @} |
| 2386 | @end group |
| 2387 | @end smallexample |
| 2388 | |
| 2389 | The run-time initialization code (for C programs, usually @code{crt0}) |
| 2390 | for use with a ROM generated this way has to include something like |
| 2391 | the following, to copy the initialized data from the ROM image to its runtime |
| 2392 | address: |
| 2393 | |
| 2394 | @smallexample |
| 2395 | @group |
| 2396 | char *src = _etext; |
| 2397 | char *dst = _data; |
| 2398 | |
| 2399 | /* ROM has data at end of text; copy it. */ |
| 2400 | while (dst < _edata) @{ |
| 2401 | *dst++ = *src++; |
| 2402 | @} |
| 2403 | |
| 2404 | /* Zero bss */ |
| 2405 | for (dst = _bstart; dst< _bend; dst++) |
| 2406 | *dst = 0; |
| 2407 | @end group |
| 2408 | @end smallexample |
| 2409 | |
| 2410 | @kindex >@var{region} |
| 2411 | @cindex section, assigning to memory region |
| 2412 | @cindex memory regions and sections |
| 2413 | @item >@var{region} |
| 2414 | Assign this section to a previously defined region of memory. |
| 2415 | @xref{MEMORY}. |
| 2416 | |
| 2417 | @kindex :@var{phdr} |
| 2418 | @cindex section, assigning to program header |
| 2419 | @cindex program headers and sections |
| 2420 | @item :@var{phdr} |
| 2421 | Assign this section to a segment described by a program header. |
| 2422 | @xref{PHDRS}. If a section is assigned to one or more segments, then |
| 2423 | all subsequent allocated sections will be assigned to those segments as |
| 2424 | well, unless they use an explicitly @code{:@var{phdr}} modifier. To |
| 2425 | prevent a section from being assigned to a segment when it would |
| 2426 | normally default to one, use @code{:NONE}. |
| 2427 | |
| 2428 | @kindex =@var{fill} |
| 2429 | @cindex section fill pattern |
| 2430 | @cindex fill pattern, entire section |
| 2431 | @item =@var{fill} |
| 2432 | Including @code{=@var{fill}} in a section definition specifies the |
| 2433 | initial fill value for that section. You may use any expression to |
| 2434 | specify @var{fill}. Any unallocated holes in the current output section |
| 2435 | when written to the output file will be filled with the two least |
| 2436 | significant bytes of the value, repeated as necessary. You can also |
| 2437 | change the fill value with a @code{FILL} statement in the @var{contents} |
| 2438 | of a section definition. |
| 2439 | |
| 2440 | @end table |
| 2441 | |
| 2442 | @node Overlays |
| 2443 | @subsection Overlays |
| 2444 | @kindex OVERLAY |
| 2445 | @cindex overlays |
| 2446 | |
| 2447 | The @code{OVERLAY} command provides an easy way to describe sections |
| 2448 | which are to be loaded as part of a single memory image but are to be |
| 2449 | run at the same memory address. At run time, some sort of overlay |
| 2450 | manager will copy the overlaid sections in and out of the runtime memory |
| 2451 | address as required, perhaps by simply manipulating addressing bits. |
| 2452 | This approach can be useful, for example, when a certain region of |
| 2453 | memory is faster than another. |
| 2454 | |
| 2455 | The @code{OVERLAY} command is used within a @code{SECTIONS} command. It |
| 2456 | appears as follows: |
| 2457 | @smallexample |
| 2458 | @group |
| 2459 | OVERLAY @var{start} : [ NOCROSSREFS ] AT ( @var{ldaddr} ) |
| 2460 | @{ |
| 2461 | @var{secname1} @{ @var{contents} @} :@var{phdr} =@var{fill} |
| 2462 | @var{secname2} @{ @var{contents} @} :@var{phdr} =@var{fill} |
| 2463 | @dots{} |
| 2464 | @} >@var{region} :@var{phdr} =@var{fill} |
| 2465 | @end group |
| 2466 | @end smallexample |
| 2467 | |
| 2468 | Everything is optional except @code{OVERLAY} (a keyword), and each |
| 2469 | section must have a name (@var{secname1} and @var{secname2} above). The |
| 2470 | section definitions within the @code{OVERLAY} construct are identical to |
| 2471 | those within the general @code{SECTIONS} contruct (@pxref{SECTIONS}), |
| 2472 | except that no addresses and no memory regions may be defined for |
| 2473 | sections within an @code{OVERLAY}. |
| 2474 | |
| 2475 | The sections are all defined with the same starting address. The load |
| 2476 | addresses of the sections are arranged such that they are consecutive in |
| 2477 | memory starting at the load address used for the @code{OVERLAY} as a |
| 2478 | whole (as with normal section definitions, the load address is optional, |
| 2479 | and defaults to the start address; the start address is also optional, |
| 2480 | and defaults to @code{.}). |
| 2481 | |
| 2482 | If the @code{NOCROSSREFS} keyword is used, and there any references |
| 2483 | among the sections, the linker will report an error. Since the sections |
| 2484 | all run at the same address, it normally does not make sense for one |
| 2485 | section to refer directly to another. @xref{Option Commands, |
| 2486 | NOCROSSREFS}. |
| 2487 | |
| 2488 | For each section within the @code{OVERLAY}, the linker automatically |
| 2489 | defines two symbols. The symbol @code{__load_start_@var{secname}} is |
| 2490 | defined as the starting load address of the section. The symbol |
| 2491 | @code{__load_stop_@var{secname}} is defined as the final load address of |
| 2492 | the section. Any characters within @var{secname} which are not legal |
| 2493 | within C identifiers are removed. C (or assembler) code may use these |
| 2494 | symbols to move the overlaid sections around as necessary. |
| 2495 | |
| 2496 | At the end of the overlay, the value of @code{.} is set to the start |
| 2497 | address of the overlay plus the size of the largest section. |
| 2498 | |
| 2499 | Here is an example. Remember that this would appear inside a |
| 2500 | @code{SECTIONS} construct. |
| 2501 | |
| 2502 | @smallexample |
| 2503 | @group |
| 2504 | OVERLAY 0x1000 : AT (0x4000) |
| 2505 | @{ |
| 2506 | .text0 @{ o1/*.o(.text) @} |
| 2507 | .text1 @{ o2/*.o(.text) @} |
| 2508 | @} |
| 2509 | @end group |
| 2510 | @end smallexample |
| 2511 | |
| 2512 | This will define both @code{.text0} and @code{.text1} to start at |
| 2513 | address 0x1000. @code{.text0} will be loaded at address 0x4000, and |
| 2514 | @code{.text1} will be loaded immediately after @code{.text0}. The |
| 2515 | following symbols will be defined: @code{__load_start_text0}, |
| 2516 | @code{__load_stop_text0}, @code{__load_start_text1}, |
| 2517 | @code{__load_stop_text1}. |
| 2518 | |
| 2519 | C code to copy overlay @code{.text1} into the overlay area might look |
| 2520 | like the following. |
| 2521 | |
| 2522 | @smallexample |
| 2523 | @group |
| 2524 | extern char __load_start_text1, __load_stop_text1; |
| 2525 | memcpy ((char *) 0x1000, &__load_start_text1, |
| 2526 | &__load_stop_text1 - &__load_start_text1); |
| 2527 | @end group |
| 2528 | @end smallexample |
| 2529 | |
| 2530 | Note that the @code{OVERLAY} command is just syntactic sugar, since |
| 2531 | everything it does can be done using the more basic commands. The above |
| 2532 | example could have been written identically as follows. |
| 2533 | |
| 2534 | @smallexample |
| 2535 | @group |
| 2536 | .text0 0x1000 : AT (0x4000) @{ o1/*.o(.text) @} |
| 2537 | __load_start_text0 = LOADADDR (.text0); |
| 2538 | __load_stop_text0 = LOADADDR (.text0) + SIZEOF (.text0); |
| 2539 | .text1 0x1000 : AT (0x4000 + SIZEOF (.text0)) @{ o2/*.o(.text) @} |
| 2540 | __load_start_text1 = LOADADDR (.text1); |
| 2541 | __load_stop_text1 = LOADADDR (.text1) + SIZEOF (.text1); |
| 2542 | . = 0x1000 + MAX (SIZEOF (.text0), SIZEOF (.text1)); |
| 2543 | @end group |
| 2544 | @end smallexample |
| 2545 | |
| 2546 | @node PHDRS |
| 2547 | @section ELF Program Headers |
| 2548 | @kindex PHDRS |
| 2549 | @cindex program headers |
| 2550 | @cindex ELF program headers |
| 2551 | |
| 2552 | The ELF object file format uses @dfn{program headers}, which are read by |
| 2553 | the system loader and describe how the program should be loaded into |
| 2554 | memory. These program headers must be set correctly in order to run the |
| 2555 | program on a native ELF system. The linker will create reasonable |
| 2556 | program headers by default. However, in some cases, it is desirable to |
| 2557 | specify the program headers more precisely; the @code{PHDRS} command may |
| 2558 | be used for this purpose. When the @code{PHDRS} command is used, the |
| 2559 | linker will not generate any program headers itself. |
| 2560 | |
| 2561 | The @code{PHDRS} command is only meaningful when generating an ELF |
| 2562 | output file. It is ignored in other cases. This manual does not |
| 2563 | describe the details of how the system loader interprets program |
| 2564 | headers; for more information, see the ELF ABI. The program headers of |
| 2565 | an ELF file may be displayed using the @samp{-p} option of the |
| 2566 | @code{objdump} command. |
| 2567 | |
| 2568 | This is the syntax of the @code{PHDRS} command. The words @code{PHDRS}, |
| 2569 | @code{FILEHDR}, @code{AT}, and @code{FLAGS} are keywords. |
| 2570 | |
| 2571 | @smallexample |
| 2572 | @group |
| 2573 | PHDRS |
| 2574 | @{ |
| 2575 | @var{name} @var{type} [ FILEHDR ] [ PHDRS ] [ AT ( @var{address} ) ] |
| 2576 | [ FLAGS ( @var{flags} ) ] ; |
| 2577 | @} |
| 2578 | @end group |
| 2579 | @end smallexample |
| 2580 | |
| 2581 | The @var{name} is used only for reference in the @code{SECTIONS} command |
| 2582 | of the linker script. It does not get put into the output file. |
| 2583 | |
| 2584 | Certain program header types describe segments of memory which are |
| 2585 | loaded from the file by the system loader. In the linker script, the |
| 2586 | contents of these segments are specified by directing allocated output |
| 2587 | sections to be placed in the segment. To do this, the command |
| 2588 | describing the output section in the @code{SECTIONS} command should use |
| 2589 | @samp{:@var{name}}, where @var{name} is the name of the program header |
| 2590 | as it appears in the @code{PHDRS} command. @xref{Section Options}. |
| 2591 | |
| 2592 | It is normal for certain sections to appear in more than one segment. |
| 2593 | This merely implies that one segment of memory contains another. This |
| 2594 | is specified by repeating @samp{:@var{name}}, using it once for each |
| 2595 | program header in which the section is to appear. |
| 2596 | |
| 2597 | If a section is placed in one or more segments using @samp{:@var{name}}, |
| 2598 | then all subsequent allocated sections which do not specify |
| 2599 | @samp{:@var{name}} are placed in the same segments. This is for |
| 2600 | convenience, since generally a whole set of contiguous sections will be |
| 2601 | placed in a single segment. To prevent a section from being assigned to |
| 2602 | a segment when it would normally default to one, use @code{:NONE}. |
| 2603 | |
| 2604 | The @code{FILEHDR} and @code{PHDRS} keywords which may appear after the |
| 2605 | program header type also indicate contents of the segment of memory. |
| 2606 | The @code{FILEHDR} keyword means that the segment should include the ELF |
| 2607 | file header. The @code{PHDRS} keyword means that the segment should |
| 2608 | include the ELF program headers themselves. |
| 2609 | |
| 2610 | The @var{type} may be one of the following. The numbers indicate the |
| 2611 | value of the keyword. |
| 2612 | |
| 2613 | @table @asis |
| 2614 | @item @code{PT_NULL} (0) |
| 2615 | Indicates an unused program header. |
| 2616 | |
| 2617 | @item @code{PT_LOAD} (1) |
| 2618 | Indicates that this program header describes a segment to be loaded from |
| 2619 | the file. |
| 2620 | |
| 2621 | @item @code{PT_DYNAMIC} (2) |
| 2622 | Indicates a segment where dynamic linking information can be found. |
| 2623 | |
| 2624 | @item @code{PT_INTERP} (3) |
| 2625 | Indicates a segment where the name of the program interpreter may be |
| 2626 | found. |
| 2627 | |
| 2628 | @item @code{PT_NOTE} (4) |
| 2629 | Indicates a segment holding note information. |
| 2630 | |
| 2631 | @item @code{PT_SHLIB} (5) |
| 2632 | A reserved program header type, defined but not specified by the ELF |
| 2633 | ABI. |
| 2634 | |
| 2635 | @item @code{PT_PHDR} (6) |
| 2636 | Indicates a segment where the program headers may be found. |
| 2637 | |
| 2638 | @item @var{expression} |
| 2639 | An expression giving the numeric type of the program header. This may |
| 2640 | be used for types not defined above. |
| 2641 | @end table |
| 2642 | |
| 2643 | It is possible to specify that a segment should be loaded at a |
| 2644 | particular address in memory. This is done using an @code{AT} |
| 2645 | expression. This is identical to the @code{AT} command used in the |
| 2646 | @code{SECTIONS} command (@pxref{Section Options}). Using the @code{AT} |
| 2647 | command for a program header overrides any information in the |
| 2648 | @code{SECTIONS} command. |
| 2649 | |
| 2650 | Normally the segment flags are set based on the sections. The |
| 2651 | @code{FLAGS} keyword may be used to explicitly specify the segment |
| 2652 | flags. The value of @var{flags} must be an integer. It is used to |
| 2653 | set the @code{p_flags} field of the program header. |
| 2654 | |
| 2655 | Here is an example of the use of @code{PHDRS}. This shows a typical set |
| 2656 | of program headers used on a native ELF system. |
| 2657 | |
| 2658 | @example |
| 2659 | @group |
| 2660 | PHDRS |
| 2661 | @{ |
| 2662 | headers PT_PHDR PHDRS ; |
| 2663 | interp PT_INTERP ; |
| 2664 | text PT_LOAD FILEHDR PHDRS ; |
| 2665 | data PT_LOAD ; |
| 2666 | dynamic PT_DYNAMIC ; |
| 2667 | @} |
| 2668 | |
| 2669 | SECTIONS |
| 2670 | @{ |
| 2671 | . = SIZEOF_HEADERS; |
| 2672 | .interp : @{ *(.interp) @} :text :interp |
| 2673 | .text : @{ *(.text) @} :text |
| 2674 | .rodata : @{ *(.rodata) @} /* defaults to :text */ |
| 2675 | @dots{} |
| 2676 | . = . + 0x1000; /* move to a new page in memory */ |
| 2677 | .data : @{ *(.data) @} :data |
| 2678 | .dynamic : @{ *(.dynamic) @} :data :dynamic |
| 2679 | @dots{} |
| 2680 | @} |
| 2681 | @end group |
| 2682 | @end example |
| 2683 | |
| 2684 | @node Entry Point |
| 2685 | @section The Entry Point |
| 2686 | @kindex ENTRY(@var{symbol}) |
| 2687 | @cindex start of execution |
| 2688 | @cindex first instruction |
| 2689 | The linker command language includes a command specifically for |
| 2690 | defining the first executable instruction in an output file (its |
| 2691 | @dfn{entry point}). Its argument is a symbol name: |
| 2692 | @smallexample |
| 2693 | ENTRY(@var{symbol}) |
| 2694 | @end smallexample |
| 2695 | |
| 2696 | Like symbol assignments, the @code{ENTRY} command may be placed either |
| 2697 | as an independent command in the command file, or among the section |
| 2698 | definitions within the @code{SECTIONS} command---whatever makes the most |
| 2699 | sense for your layout. |
| 2700 | |
| 2701 | @cindex entry point, defaults |
| 2702 | @code{ENTRY} is only one of several ways of choosing the entry point. |
| 2703 | You may indicate it in any of the following ways (shown in descending |
| 2704 | order of priority: methods higher in the list override methods lower down). |
| 2705 | @itemize @bullet |
| 2706 | @item |
| 2707 | the @samp{-e} @var{entry} command-line option; |
| 2708 | @item |
| 2709 | the @code{ENTRY(@var{symbol})} command in a linker control script; |
| 2710 | @item |
| 2711 | the value of the symbol @code{start}, if present; |
| 2712 | @item |
| 2713 | the address of the first byte of the @code{.text} section, if present; |
| 2714 | @item |
| 2715 | The address @code{0}. |
| 2716 | @end itemize |
| 2717 | |
| 2718 | For example, you can use these rules to generate an entry point with an |
| 2719 | assignment statement: if no symbol @code{start} is defined within your |
| 2720 | input files, you can simply define it, assigning it an appropriate |
| 2721 | value--- |
| 2722 | |
| 2723 | @smallexample |
| 2724 | start = 0x2020; |
| 2725 | @end smallexample |
| 2726 | |
| 2727 | @noindent |
| 2728 | The example shows an absolute address, but you can use any expression. |
| 2729 | For example, if your input object files use some other symbol-name |
| 2730 | convention for the entry point, you can just assign the value of |
| 2731 | whatever symbol contains the start address to @code{start}: |
| 2732 | |
| 2733 | @smallexample |
| 2734 | start = other_symbol ; |
| 2735 | @end smallexample |
| 2736 | |
| 2737 | @node Option Commands |
| 2738 | @section Option Commands |
| 2739 | The command language includes a number of other commands that you can |
| 2740 | use for specialized purposes. They are similar in purpose to |
| 2741 | command-line options. |
| 2742 | |
| 2743 | @table @code |
| 2744 | @kindex CONSTRUCTORS |
| 2745 | @cindex C++ constructors, arranging in link |
| 2746 | @cindex constructors, arranging in link |
| 2747 | @item CONSTRUCTORS |
| 2748 | When linking using the @code{a.out} object file format, the linker uses |
| 2749 | an unusual set construct to support C++ global constructors and |
| 2750 | destructors. When linking object file formats which do not support |
| 2751 | arbitrary sections, such as @code{ECOFF} and @code{XCOFF}, the linker |
| 2752 | will automatically recognize C++ global constructors and destructors by |
| 2753 | name. For these object file formats, the @code{CONSTRUCTORS} command |
| 2754 | tells the linker where this information should be placed. The |
| 2755 | @code{CONSTRUCTORS} command is ignored for other object file formats. |
| 2756 | |
| 2757 | The symbol @w{@code{__CTOR_LIST__}} marks the start of the global |
| 2758 | constructors, and the symbol @w{@code{__DTOR_LIST}} marks the end. The |
| 2759 | first word in the list is the number of entries, followed by the address |
| 2760 | of each constructor or destructor, followed by a zero word. The |
| 2761 | compiler must arrange to actually run the code. For these object file |
| 2762 | formats @sc{gnu} C++ calls constructors from a subroutine @code{__main}; |
| 2763 | a call to @code{__main} is automatically inserted into the startup code |
| 2764 | for @code{main}. @sc{gnu} C++ runs destructors either by using |
| 2765 | @code{atexit}, or directly from the function @code{exit}. |
| 2766 | |
| 2767 | For object file formats such as @code{COFF} or @code{ELF} which support |
| 2768 | multiple sections, @sc{gnu} C++ will normally arrange to put the |
| 2769 | addresses of global constructors and destructors into the @code{.ctors} |
| 2770 | and @code{.dtors} sections. Placing the following sequence into your |
| 2771 | linker script will build the sort of table which the @sc{gnu} C++ |
| 2772 | runtime code expects to see. |
| 2773 | |
| 2774 | @smallexample |
| 2775 | __CTOR_LIST__ = .; |
| 2776 | LONG((__CTOR_END__ - __CTOR_LIST__) / 4 - 2) |
| 2777 | *(.ctors) |
| 2778 | LONG(0) |
| 2779 | __CTOR_END__ = .; |
| 2780 | __DTOR_LIST__ = .; |
| 2781 | LONG((__DTOR_END__ - __DTOR_LIST__) / 4 - 2) |
| 2782 | *(.dtors) |
| 2783 | LONG(0) |
| 2784 | __DTOR_END__ = .; |
| 2785 | @end smallexample |
| 2786 | |
| 2787 | Normally the compiler and linker will handle these issues automatically, |
| 2788 | and you will not need to concern yourself with them. However, you may |
| 2789 | need to consider this if you are using C++ and writing your own linker |
| 2790 | scripts. |
| 2791 | |
| 2792 | @need 1000 |
| 2793 | @kindex FLOAT |
| 2794 | @kindex NOFLOAT |
| 2795 | @item FLOAT |
| 2796 | @itemx NOFLOAT |
| 2797 | These keywords were used in some older linkers to request a particular |
| 2798 | math subroutine library. @code{ld} doesn't use the keywords, assuming |
| 2799 | instead that any necessary subroutines are in libraries specified using |
| 2800 | the general mechanisms for linking to archives; but to permit the use of |
| 2801 | scripts that were written for the older linkers, the keywords |
| 2802 | @code{FLOAT} and @code{NOFLOAT} are accepted and ignored. |
| 2803 | |
| 2804 | @kindex FORCE_COMMON_ALLOCATION |
| 2805 | @cindex common allocation |
| 2806 | @item FORCE_COMMON_ALLOCATION |
| 2807 | This command has the same effect as the @samp{-d} command-line option: |
| 2808 | to make @code{ld} assign space to common symbols even if a relocatable |
| 2809 | output file is specified (@samp{-r}). |
| 2810 | |
| 2811 | @kindex INPUT ( @var{files} ) |
| 2812 | @cindex binary input files |
| 2813 | @item INPUT ( @var{file}, @var{file}, @dots{} ) |
| 2814 | @itemx INPUT ( @var{file} @var{file} @dots{} ) |
| 2815 | Use this command to include binary input files in the link, without |
| 2816 | including them in a particular section definition. |
| 2817 | Specify the full name for each @var{file}, including @samp{.a} if |
| 2818 | required. |
| 2819 | |
| 2820 | @code{ld} searches for each @var{file} through the archive-library |
| 2821 | search path, just as for files you specify on the command line. |
| 2822 | See the description of @samp{-L} in @ref{Options,,Command Line |
| 2823 | Options}. |
| 2824 | |
| 2825 | If you use @samp{-l@var{file}}, @code{ld} will transform the name to |
| 2826 | @code{lib@var{file}.a} as with the command line argument @samp{-l}. |
| 2827 | |
| 2828 | @kindex GROUP ( @var{files} ) |
| 2829 | @cindex grouping input files |
| 2830 | @item GROUP ( @var{file}, @var{file}, @dots{} ) |
| 2831 | @itemx GROUP ( @var{file} @var{file} @dots{} ) |
| 2832 | This command is like @code{INPUT}, except that the named files should |
| 2833 | all be archives, and they are searched repeatedly until no new undefined |
| 2834 | references are created. See the description of @samp{-(} in |
| 2835 | @ref{Options,,Command Line Options}. |
| 2836 | |
| 2837 | @ignore |
| 2838 | @kindex MAP ( @var{name} ) |
| 2839 | @item MAP ( @var{name} ) |
| 2840 | @c MAP(...) appears to look for an F in the arg, ignoring all other |
| 2841 | @c chars; if it finds one, it sets "map_option_f" to true. But nothing |
| 2842 | @c checks map_option_f. Apparently a stub for the future... |
| 2843 | @end ignore |
| 2844 | |
| 2845 | @kindex OUTPUT ( @var{filename} ) |
| 2846 | @cindex naming the output file |
| 2847 | @item OUTPUT ( @var{filename} ) |
| 2848 | Use this command to name the link output file @var{filename}. The |
| 2849 | effect of @code{OUTPUT(@var{filename})} is identical to the effect of |
| 2850 | @w{@samp{-o @var{filename}}}, which overrides it. You can use this |
| 2851 | command to supply a default output-file name other than @code{a.out}. |
| 2852 | |
| 2853 | @ifclear SingleFormat |
| 2854 | @kindex OUTPUT_ARCH ( @var{bfdname} ) |
| 2855 | @cindex machine architecture, output |
| 2856 | @item OUTPUT_ARCH ( @var{bfdname} ) |
| 2857 | Specify a particular output machine architecture, with one of the names |
| 2858 | used by the BFD back-end routines (@pxref{BFD}). This command is often |
| 2859 | unnecessary; the architecture is most often set implicitly by either the |
| 2860 | system BFD configuration or as a side effect of the @code{OUTPUT_FORMAT} |
| 2861 | command. |
| 2862 | |
| 2863 | @kindex OUTPUT_FORMAT ( @var{bfdname} ) |
| 2864 | @cindex format, output file |
| 2865 | @item OUTPUT_FORMAT ( @var{bfdname} ) |
| 2866 | When @code{ld} is configured to support multiple object code formats, |
| 2867 | you can use this command to specify a particular output format. |
| 2868 | @var{bfdname} is one of the names used by the BFD back-end routines |
| 2869 | (@pxref{BFD}). The effect is identical to the effect of the |
| 2870 | @samp{--oformat} command-line option. This selection affects only the |
| 2871 | output file; the related command @code{TARGET} affects primarily input |
| 2872 | files. |
| 2873 | @end ifclear |
| 2874 | |
| 2875 | @kindex SEARCH_DIR ( @var{path} ) |
| 2876 | @cindex path for libraries |
| 2877 | @cindex search path, libraries |
| 2878 | @item SEARCH_DIR ( @var{path} ) |
| 2879 | Add @var{path} to the list of paths where @code{ld} looks for |
| 2880 | archive libraries. @code{SEARCH_DIR(@var{path})} has the same |
| 2881 | effect as @samp{-L@var{path}} on the command line. |
| 2882 | |
| 2883 | @kindex STARTUP ( @var{filename} ) |
| 2884 | @cindex first input file |
| 2885 | @item STARTUP ( @var{filename} ) |
| 2886 | Ensure that @var{filename} is the first input file used in the link |
| 2887 | process. |
| 2888 | |
| 2889 | @ifclear SingleFormat |
| 2890 | @cindex input file format |
| 2891 | @kindex TARGET ( @var{format} ) |
| 2892 | @item TARGET ( @var{format} ) |
| 2893 | When @code{ld} is configured to support multiple object code formats, |
| 2894 | you can use this command to change the input-file object code format |
| 2895 | (like the command-line option @samp{-b} or its synonym @samp{--format}). |
| 2896 | The argument @var{format} is one of the strings used by BFD to name |
| 2897 | binary formats. If @code{TARGET} is specified but @code{OUTPUT_FORMAT} |
| 2898 | is not, the last @code{TARGET} argument is also used as the default |
| 2899 | format for the @code{ld} output file. @xref{BFD}. |
| 2900 | |
| 2901 | @kindex GNUTARGET |
| 2902 | If you don't use the @code{TARGET} command, @code{ld} uses the value of |
| 2903 | the environment variable @code{GNUTARGET}, if available, to select the |
| 2904 | output file format. If that variable is also absent, @code{ld} uses |
| 2905 | the default format configured for your machine in the BFD libraries. |
| 2906 | @end ifclear |
| 2907 | |
| 2908 | @cindex cross references |
| 2909 | @kindex NOCROSSREFS ( @var{sections} ) |
| 2910 | @item NOCROSSREFS ( @var{section} @var{section} @dots{} ) |
| 2911 | This command may be used to tell @code{ld} to issue an error about any |
| 2912 | references among certain sections. |
| 2913 | |
| 2914 | In certain types of programs, particularly on embedded systems, when one |
| 2915 | section is loaded into memory, another section will not be. Any direct |
| 2916 | references between the two sections would be errors. For example, it |
| 2917 | would be an error if code in one section called a function defined in |
| 2918 | the other section. |
| 2919 | |
| 2920 | The @code{NOCROSSREFS} command takes a list of section names. If |
| 2921 | @code{ld} detects any cross references between the sections, it reports |
| 2922 | an error and returns a non-zero exit status. The @code{NOCROSSREFS} |
| 2923 | command uses output section names, defined in the @code{SECTIONS} |
| 2924 | command. It does not use the names of input sections. |
| 2925 | @end table |
| 2926 | |
| 2927 | @ifset GENERIC |
| 2928 | @node Machine Dependent |
| 2929 | @chapter Machine Dependent Features |
| 2930 | |
| 2931 | @cindex machine dependencies |
| 2932 | @code{ld} has additional features on some platforms; the following |
| 2933 | sections describe them. Machines where @code{ld} has no additional |
| 2934 | functionality are not listed. |
| 2935 | |
| 2936 | @menu |
| 2937 | * H8/300:: @code{ld} and the H8/300 |
| 2938 | * i960:: @code{ld} and the Intel 960 family |
| 2939 | @end menu |
| 2940 | @end ifset |
| 2941 | |
| 2942 | @c FIXME! This could use @raisesections/@lowersections, but there seems to be a conflict |
| 2943 | @c between those and node-defaulting. |
| 2944 | @ifset H8300 |
| 2945 | @ifclear GENERIC |
| 2946 | @raisesections |
| 2947 | @end ifclear |
| 2948 | @node H8/300 |
| 2949 | @section @code{ld} and the H8/300 |
| 2950 | |
| 2951 | @cindex H8/300 support |
| 2952 | For the H8/300, @code{ld} can perform these global optimizations when |
| 2953 | you specify the @samp{--relax} command-line option. |
| 2954 | |
| 2955 | @table @emph |
| 2956 | @cindex relaxing on H8/300 |
| 2957 | @item relaxing address modes |
| 2958 | @code{ld} finds all @code{jsr} and @code{jmp} instructions whose |
| 2959 | targets are within eight bits, and turns them into eight-bit |
| 2960 | program-counter relative @code{bsr} and @code{bra} instructions, |
| 2961 | respectively. |
| 2962 | |
| 2963 | @cindex synthesizing on H8/300 |
| 2964 | @item synthesizing instructions |
| 2965 | @c FIXME: specifically mov.b, or any mov instructions really? |
| 2966 | @code{ld} finds all @code{mov.b} instructions which use the |
| 2967 | sixteen-bit absolute address form, but refer to the top |
| 2968 | page of memory, and changes them to use the eight-bit address form. |
| 2969 | (That is: the linker turns @samp{mov.b @code{@@}@var{aa}:16} into |
| 2970 | @samp{mov.b @code{@@}@var{aa}:8} whenever the address @var{aa} is in the |
| 2971 | top page of memory). |
| 2972 | @end table |
| 2973 | @ifclear GENERIC |
| 2974 | @lowersections |
| 2975 | @end ifclear |
| 2976 | @end ifset |
| 2977 | |
| 2978 | @ifclear GENERIC |
| 2979 | @ifset Hitachi |
| 2980 | @c This stuff is pointless to say unless you're especially concerned |
| 2981 | @c with Hitachi chips; don't enable it for generic case, please. |
| 2982 | @node Hitachi |
| 2983 | @chapter @code{ld} and other Hitachi chips |
| 2984 | |
| 2985 | @code{ld} also supports the H8/300H, the H8/500, and the Hitachi SH. No |
| 2986 | special features, commands, or command-line options are required for |
| 2987 | these chips. |
| 2988 | @end ifset |
| 2989 | @end ifclear |
| 2990 | |
| 2991 | @ifset I960 |
| 2992 | @ifclear GENERIC |
| 2993 | @raisesections |
| 2994 | @end ifclear |
| 2995 | @node i960 |
| 2996 | @section @code{ld} and the Intel 960 family |
| 2997 | |
| 2998 | @cindex i960 support |
| 2999 | |
| 3000 | You can use the @samp{-A@var{architecture}} command line option to |
| 3001 | specify one of the two-letter names identifying members of the 960 |
| 3002 | family; the option specifies the desired output target, and warns of any |
| 3003 | incompatible instructions in the input files. It also modifies the |
| 3004 | linker's search strategy for archive libraries, to support the use of |
| 3005 | libraries specific to each particular architecture, by including in the |
| 3006 | search loop names suffixed with the string identifying the architecture. |
| 3007 | |
| 3008 | For example, if your @code{ld} command line included @w{@samp{-ACA}} as |
| 3009 | well as @w{@samp{-ltry}}, the linker would look (in its built-in search |
| 3010 | paths, and in any paths you specify with @samp{-L}) for a library with |
| 3011 | the names |
| 3012 | |
| 3013 | @smallexample |
| 3014 | @group |
| 3015 | try |
| 3016 | libtry.a |
| 3017 | tryca |
| 3018 | libtryca.a |
| 3019 | @end group |
| 3020 | @end smallexample |
| 3021 | |
| 3022 | @noindent |
| 3023 | The first two possibilities would be considered in any event; the last |
| 3024 | two are due to the use of @w{@samp{-ACA}}. |
| 3025 | |
| 3026 | You can meaningfully use @samp{-A} more than once on a command line, since |
| 3027 | the 960 architecture family allows combination of target architectures; each |
| 3028 | use will add another pair of name variants to search for when @w{@samp{-l}} |
| 3029 | specifies a library. |
| 3030 | |
| 3031 | @cindex @code{--relax} on i960 |
| 3032 | @cindex relaxing on i960 |
| 3033 | @code{ld} supports the @samp{--relax} option for the i960 family. If |
| 3034 | you specify @samp{--relax}, @code{ld} finds all @code{balx} and |
| 3035 | @code{calx} instructions whose targets are within 24 bits, and turns |
| 3036 | them into 24-bit program-counter relative @code{bal} and @code{cal} |
| 3037 | instructions, respectively. @code{ld} also turns @code{cal} |
| 3038 | instructions into @code{bal} instructions when it determines that the |
| 3039 | target subroutine is a leaf routine (that is, the target subroutine does |
| 3040 | not itself call any subroutines). |
| 3041 | |
| 3042 | @ifclear GENERIC |
| 3043 | @lowersections |
| 3044 | @end ifclear |
| 3045 | @end ifset |
| 3046 | |
| 3047 | @ifclear SingleFormat |
| 3048 | @node BFD |
| 3049 | @chapter BFD |
| 3050 | |
| 3051 | @cindex back end |
| 3052 | @cindex object file management |
| 3053 | @cindex object formats available |
| 3054 | @kindex objdump -i |
| 3055 | The linker accesses object and archive files using the BFD libraries. |
| 3056 | These libraries allow the linker to use the same routines to operate on |
| 3057 | object files whatever the object file format. A different object file |
| 3058 | format can be supported simply by creating a new BFD back end and adding |
| 3059 | it to the library. To conserve runtime memory, however, the linker and |
| 3060 | associated tools are usually configured to support only a subset of the |
| 3061 | object file formats available. You can use @code{objdump -i} |
| 3062 | (@pxref{objdump,,objdump,binutils.info,The GNU Binary Utilities}) to |
| 3063 | list all the formats available for your configuration. |
| 3064 | |
| 3065 | @cindex BFD requirements |
| 3066 | @cindex requirements for BFD |
| 3067 | As with most implementations, BFD is a compromise between |
| 3068 | several conflicting requirements. The major factor influencing |
| 3069 | BFD design was efficiency: any time used converting between |
| 3070 | formats is time which would not have been spent had BFD not |
| 3071 | been involved. This is partly offset by abstraction payback; since |
| 3072 | BFD simplifies applications and back ends, more time and care |
| 3073 | may be spent optimizing algorithms for a greater speed. |
| 3074 | |
| 3075 | One minor artifact of the BFD solution which you should bear in |
| 3076 | mind is the potential for information loss. There are two places where |
| 3077 | useful information can be lost using the BFD mechanism: during |
| 3078 | conversion and during output. @xref{BFD information loss}. |
| 3079 | |
| 3080 | @menu |
| 3081 | * BFD outline:: How it works: an outline of BFD |
| 3082 | @end menu |
| 3083 | |
| 3084 | @node BFD outline |
| 3085 | @section How it works: an outline of BFD |
| 3086 | @cindex opening object files |
| 3087 | @include bfdsumm.texi |
| 3088 | @end ifclear |
| 3089 | |
| 3090 | @node Reporting Bugs |
| 3091 | @chapter Reporting Bugs |
| 3092 | @cindex bugs in @code{ld} |
| 3093 | @cindex reporting bugs in @code{ld} |
| 3094 | |
| 3095 | Your bug reports play an essential role in making @code{ld} reliable. |
| 3096 | |
| 3097 | Reporting a bug may help you by bringing a solution to your problem, or |
| 3098 | it may not. But in any case the principal function of a bug report is |
| 3099 | to help the entire community by making the next version of @code{ld} |
| 3100 | work better. Bug reports are your contribution to the maintenance of |
| 3101 | @code{ld}. |
| 3102 | |
| 3103 | In order for a bug report to serve its purpose, you must include the |
| 3104 | information that enables us to fix the bug. |
| 3105 | |
| 3106 | @menu |
| 3107 | * Bug Criteria:: Have you found a bug? |
| 3108 | * Bug Reporting:: How to report bugs |
| 3109 | @end menu |
| 3110 | |
| 3111 | @node Bug Criteria |
| 3112 | @section Have you found a bug? |
| 3113 | @cindex bug criteria |
| 3114 | |
| 3115 | If you are not sure whether you have found a bug, here are some guidelines: |
| 3116 | |
| 3117 | @itemize @bullet |
| 3118 | @cindex fatal signal |
| 3119 | @cindex linker crash |
| 3120 | @cindex crash of linker |
| 3121 | @item |
| 3122 | If the linker gets a fatal signal, for any input whatever, that is a |
| 3123 | @code{ld} bug. Reliable linkers never crash. |
| 3124 | |
| 3125 | @cindex error on valid input |
| 3126 | @item |
| 3127 | If @code{ld} produces an error message for valid input, that is a bug. |
| 3128 | |
| 3129 | @cindex invalid input |
| 3130 | @item |
| 3131 | If @code{ld} does not produce an error message for invalid input, that |
| 3132 | may be a bug. In the general case, the linker can not verify that |
| 3133 | object files are correct. |
| 3134 | |
| 3135 | @item |
| 3136 | If you are an experienced user of linkers, your suggestions for |
| 3137 | improvement of @code{ld} are welcome in any case. |
| 3138 | @end itemize |
| 3139 | |
| 3140 | @node Bug Reporting |
| 3141 | @section How to report bugs |
| 3142 | @cindex bug reports |
| 3143 | @cindex @code{ld} bugs, reporting |
| 3144 | |
| 3145 | A number of companies and individuals offer support for @sc{gnu} |
| 3146 | products. If you obtained @code{ld} from a support organization, we |
| 3147 | recommend you contact that organization first. |
| 3148 | |
| 3149 | You can find contact information for many support companies and |
| 3150 | individuals in the file @file{etc/SERVICE} in the @sc{gnu} Emacs |
| 3151 | distribution. |
| 3152 | |
| 3153 | In any event, we also recommend that you send bug reports for @code{ld} |
| 3154 | to @samp{bug-gnu-utils@@prep.ai.mit.edu}. |
| 3155 | |
| 3156 | The fundamental principle of reporting bugs usefully is this: |
| 3157 | @strong{report all the facts}. If you are not sure whether to state a |
| 3158 | fact or leave it out, state it! |
| 3159 | |
| 3160 | Often people omit facts because they think they know what causes the |
| 3161 | problem and assume that some details do not matter. Thus, you might |
| 3162 | assume that the name of a symbol you use in an example does not matter. |
| 3163 | Well, probably it does not, but one cannot be sure. Perhaps the bug is |
| 3164 | a stray memory reference which happens to fetch from the location where |
| 3165 | that name is stored in memory; perhaps, if the name were different, the |
| 3166 | contents of that location would fool the linker into doing the right |
| 3167 | thing despite the bug. Play it safe and give a specific, complete |
| 3168 | example. That is the easiest thing for you to do, and the most helpful. |
| 3169 | |
| 3170 | Keep in mind that the purpose of a bug report is to enable us to fix the bug if |
| 3171 | it is new to us. Therefore, always write your bug reports on the assumption |
| 3172 | that the bug has not been reported previously. |
| 3173 | |
| 3174 | Sometimes people give a few sketchy facts and ask, ``Does this ring a |
| 3175 | bell?'' Those bug reports are useless, and we urge everyone to |
| 3176 | @emph{refuse to respond to them} except to chide the sender to report |
| 3177 | bugs properly. |
| 3178 | |
| 3179 | To enable us to fix the bug, you should include all these things: |
| 3180 | |
| 3181 | @itemize @bullet |
| 3182 | @item |
| 3183 | The version of @code{ld}. @code{ld} announces it if you start it with |
| 3184 | the @samp{--version} argument. |
| 3185 | |
| 3186 | Without this, we will not know whether there is any point in looking for |
| 3187 | the bug in the current version of @code{ld}. |
| 3188 | |
| 3189 | @item |
| 3190 | Any patches you may have applied to the @code{ld} source, including any |
| 3191 | patches made to the @code{BFD} library. |
| 3192 | |
| 3193 | @item |
| 3194 | The type of machine you are using, and the operating system name and |
| 3195 | version number. |
| 3196 | |
| 3197 | @item |
| 3198 | What compiler (and its version) was used to compile @code{ld}---e.g. |
| 3199 | ``@code{gcc-2.7}''. |
| 3200 | |
| 3201 | @item |
| 3202 | The command arguments you gave the linker to link your example and |
| 3203 | observe the bug. To guarantee you will not omit something important, |
| 3204 | list them all. A copy of the Makefile (or the output from make) is |
| 3205 | sufficient. |
| 3206 | |
| 3207 | If we were to try to guess the arguments, we would probably guess wrong |
| 3208 | and then we might not encounter the bug. |
| 3209 | |
| 3210 | @item |
| 3211 | A complete input file, or set of input files, that will reproduce the |
| 3212 | bug. It is generally most helpful to send the actual object files, |
| 3213 | uuencoded if necessary to get them through the mail system. Making them |
| 3214 | available for anonymous FTP is not as good, but may be the only |
| 3215 | reasonable choice for large object files. |
| 3216 | |
| 3217 | If the source files were assembled using @code{gas} or compiled using |
| 3218 | @code{gcc}, then it may be OK to send the source files rather than the |
| 3219 | object files. In this case, be sure to say exactly what version of |
| 3220 | @code{gas} or @code{gcc} was used to produce the object files. Also say |
| 3221 | how @code{gas} or @code{gcc} were configured. |
| 3222 | |
| 3223 | @item |
| 3224 | A description of what behavior you observe that you believe is |
| 3225 | incorrect. For example, ``It gets a fatal signal.'' |
| 3226 | |
| 3227 | Of course, if the bug is that @code{ld} gets a fatal signal, then we |
| 3228 | will certainly notice it. But if the bug is incorrect output, we might |
| 3229 | not notice unless it is glaringly wrong. You might as well not give us |
| 3230 | a chance to make a mistake. |
| 3231 | |
| 3232 | Even if the problem you experience is a fatal signal, you should still |
| 3233 | say so explicitly. Suppose something strange is going on, such as, your |
| 3234 | copy of @code{ld} is out of synch, or you have encountered a bug in the |
| 3235 | C library on your system. (This has happened!) Your copy might crash |
| 3236 | and ours would not. If you told us to expect a crash, then when ours |
| 3237 | fails to crash, we would know that the bug was not happening for us. If |
| 3238 | you had not told us to expect a crash, then we would not be able to draw |
| 3239 | any conclusion from our observations. |
| 3240 | |
| 3241 | @item |
| 3242 | If you wish to suggest changes to the @code{ld} source, send us context |
| 3243 | diffs, as generated by @code{diff} with the @samp{-u}, @samp{-c}, or |
| 3244 | @samp{-p} option. Always send diffs from the old file to the new file. |
| 3245 | If you even discuss something in the @code{ld} source, refer to it by |
| 3246 | context, not by line number. |
| 3247 | |
| 3248 | The line numbers in our development sources will not match those in your |
| 3249 | sources. Your line numbers would convey no useful information to us. |
| 3250 | @end itemize |
| 3251 | |
| 3252 | Here are some things that are not necessary: |
| 3253 | |
| 3254 | @itemize @bullet |
| 3255 | @item |
| 3256 | A description of the envelope of the bug. |
| 3257 | |
| 3258 | Often people who encounter a bug spend a lot of time investigating |
| 3259 | which changes to the input file will make the bug go away and which |
| 3260 | changes will not affect it. |
| 3261 | |
| 3262 | This is often time consuming and not very useful, because the way we |
| 3263 | will find the bug is by running a single example under the debugger |
| 3264 | with breakpoints, not by pure deduction from a series of examples. |
| 3265 | We recommend that you save your time for something else. |
| 3266 | |
| 3267 | Of course, if you can find a simpler example to report @emph{instead} |
| 3268 | of the original one, that is a convenience for us. Errors in the |
| 3269 | output will be easier to spot, running under the debugger will take |
| 3270 | less time, and so on. |
| 3271 | |
| 3272 | However, simplification is not vital; if you do not want to do this, |
| 3273 | report the bug anyway and send us the entire test case you used. |
| 3274 | |
| 3275 | @item |
| 3276 | A patch for the bug. |
| 3277 | |
| 3278 | A patch for the bug does help us if it is a good one. But do not omit |
| 3279 | the necessary information, such as the test case, on the assumption that |
| 3280 | a patch is all we need. We might see problems with your patch and decide |
| 3281 | to fix the problem another way, or we might not understand it at all. |
| 3282 | |
| 3283 | Sometimes with a program as complicated as @code{ld} it is very hard to |
| 3284 | construct an example that will make the program follow a certain path |
| 3285 | through the code. If you do not send us the example, we will not be |
| 3286 | able to construct one, so we will not be able to verify that the bug is |
| 3287 | fixed. |
| 3288 | |
| 3289 | And if we cannot understand what bug you are trying to fix, or why your |
| 3290 | patch should be an improvement, we will not install it. A test case will |
| 3291 | help us to understand. |
| 3292 | |
| 3293 | @item |
| 3294 | A guess about what the bug is or what it depends on. |
| 3295 | |
| 3296 | Such guesses are usually wrong. Even we cannot guess right about such |
| 3297 | things without first using the debugger to find the facts. |
| 3298 | @end itemize |
| 3299 | |
| 3300 | @node MRI |
| 3301 | @appendix MRI Compatible Script Files |
| 3302 | @cindex MRI compatibility |
| 3303 | To aid users making the transition to @sc{gnu} @code{ld} from the MRI |
| 3304 | linker, @code{ld} can use MRI compatible linker scripts as an |
| 3305 | alternative to the more general-purpose linker scripting language |
| 3306 | described in @ref{Commands,,Command Language}. MRI compatible linker |
| 3307 | scripts have a much simpler command set than the scripting language |
| 3308 | otherwise used with @code{ld}. @sc{gnu} @code{ld} supports the most |
| 3309 | commonly used MRI linker commands; these commands are described here. |
| 3310 | |
| 3311 | In general, MRI scripts aren't of much use with the @code{a.out} object |
| 3312 | file format, since it only has three sections and MRI scripts lack some |
| 3313 | features to make use of them. |
| 3314 | |
| 3315 | You can specify a file containing an MRI-compatible script using the |
| 3316 | @samp{-c} command-line option. |
| 3317 | |
| 3318 | Each command in an MRI-compatible script occupies its own line; each |
| 3319 | command line starts with the keyword that identifies the command (though |
| 3320 | blank lines are also allowed for punctuation). If a line of an |
| 3321 | MRI-compatible script begins with an unrecognized keyword, @code{ld} |
| 3322 | issues a warning message, but continues processing the script. |
| 3323 | |
| 3324 | Lines beginning with @samp{*} are comments. |
| 3325 | |
| 3326 | You can write these commands using all upper-case letters, or all |
| 3327 | lower case; for example, @samp{chip} is the same as @samp{CHIP}. |
| 3328 | The following list shows only the upper-case form of each command. |
| 3329 | |
| 3330 | @table @code |
| 3331 | @cindex @code{ABSOLUTE} (MRI) |
| 3332 | @item ABSOLUTE @var{secname} |
| 3333 | @itemx ABSOLUTE @var{secname}, @var{secname}, @dots{} @var{secname} |
| 3334 | Normally, @code{ld} includes in the output file all sections from all |
| 3335 | the input files. However, in an MRI-compatible script, you can use the |
| 3336 | @code{ABSOLUTE} command to restrict the sections that will be present in |
| 3337 | your output program. If the @code{ABSOLUTE} command is used at all in a |
| 3338 | script, then only the sections named explicitly in @code{ABSOLUTE} |
| 3339 | commands will appear in the linker output. You can still use other |
| 3340 | input sections (whatever you select on the command line, or using |
| 3341 | @code{LOAD}) to resolve addresses in the output file. |
| 3342 | |
| 3343 | @cindex @code{ALIAS} (MRI) |
| 3344 | @item ALIAS @var{out-secname}, @var{in-secname} |
| 3345 | Use this command to place the data from input section @var{in-secname} |
| 3346 | in a section called @var{out-secname} in the linker output file. |
| 3347 | |
| 3348 | @var{in-secname} may be an integer. |
| 3349 | |
| 3350 | @cindex @code{ALIGN} (MRI) |
| 3351 | @item ALIGN @var{secname} = @var{expression} |
| 3352 | Align the section called @var{secname} to @var{expression}. The |
| 3353 | @var{expression} should be a power of two. |
| 3354 | |
| 3355 | @cindex @code{BASE} (MRI) |
| 3356 | @item BASE @var{expression} |
| 3357 | Use the value of @var{expression} as the lowest address (other than |
| 3358 | absolute addresses) in the output file. |
| 3359 | |
| 3360 | @cindex @code{CHIP} (MRI) |
| 3361 | @item CHIP @var{expression} |
| 3362 | @itemx CHIP @var{expression}, @var{expression} |
| 3363 | This command does nothing; it is accepted only for compatibility. |
| 3364 | |
| 3365 | @cindex @code{END} (MRI) |
| 3366 | @item END |
| 3367 | This command does nothing whatever; it's only accepted for compatibility. |
| 3368 | |
| 3369 | @cindex @code{FORMAT} (MRI) |
| 3370 | @item FORMAT @var{output-format} |
| 3371 | Similar to the @code{OUTPUT_FORMAT} command in the more general linker |
| 3372 | language, but restricted to one of these output formats: |
| 3373 | |
| 3374 | @enumerate |
| 3375 | @item |
| 3376 | S-records, if @var{output-format} is @samp{S} |
| 3377 | |
| 3378 | @item |
| 3379 | IEEE, if @var{output-format} is @samp{IEEE} |
| 3380 | |
| 3381 | @item |
| 3382 | COFF (the @samp{coff-m68k} variant in BFD), if @var{output-format} is |
| 3383 | @samp{COFF} |
| 3384 | @end enumerate |
| 3385 | |
| 3386 | @cindex @code{LIST} (MRI) |
| 3387 | @item LIST @var{anything}@dots{} |
| 3388 | Print (to the standard output file) a link map, as produced by the |
| 3389 | @code{ld} command-line option @samp{-M}. |
| 3390 | |
| 3391 | The keyword @code{LIST} may be followed by anything on the |
| 3392 | same line, with no change in its effect. |
| 3393 | |
| 3394 | @cindex @code{LOAD} (MRI) |
| 3395 | @item LOAD @var{filename} |
| 3396 | @itemx LOAD @var{filename}, @var{filename}, @dots{} @var{filename} |
| 3397 | Include one or more object file @var{filename} in the link; this has the |
| 3398 | same effect as specifying @var{filename} directly on the @code{ld} |
| 3399 | command line. |
| 3400 | |
| 3401 | @cindex @code{NAME} (MRI) |
| 3402 | @item NAME @var{output-name} |
| 3403 | @var{output-name} is the name for the program produced by @code{ld}; the |
| 3404 | MRI-compatible command @code{NAME} is equivalent to the command-line |
| 3405 | option @samp{-o} or the general script language command @code{OUTPUT}. |
| 3406 | |
| 3407 | @cindex @code{ORDER} (MRI) |
| 3408 | @item ORDER @var{secname}, @var{secname}, @dots{} @var{secname} |
| 3409 | @itemx ORDER @var{secname} @var{secname} @var{secname} |
| 3410 | Normally, @code{ld} orders the sections in its output file in the |
| 3411 | order in which they first appear in the input files. In an MRI-compatible |
| 3412 | script, you can override this ordering with the @code{ORDER} command. The |
| 3413 | sections you list with @code{ORDER} will appear first in your output |
| 3414 | file, in the order specified. |
| 3415 | |
| 3416 | @cindex @code{PUBLIC} (MRI) |
| 3417 | @item PUBLIC @var{name}=@var{expression} |
| 3418 | @itemx PUBLIC @var{name},@var{expression} |
| 3419 | @itemx PUBLIC @var{name} @var{expression} |
| 3420 | Supply a value (@var{expression}) for external symbol |
| 3421 | @var{name} used in the linker input files. |
| 3422 | |
| 3423 | @cindex @code{SECT} (MRI) |
| 3424 | @item SECT @var{secname}, @var{expression} |
| 3425 | @itemx SECT @var{secname}=@var{expression} |
| 3426 | @itemx SECT @var{secname} @var{expression} |
| 3427 | You can use any of these three forms of the @code{SECT} command to |
| 3428 | specify the start address (@var{expression}) for section @var{secname}. |
| 3429 | If you have more than one @code{SECT} statement for the same |
| 3430 | @var{secname}, only the @emph{first} sets the start address. |
| 3431 | @end table |
| 3432 | |
| 3433 | @node Index |
| 3434 | @unnumbered Index |
| 3435 | |
| 3436 | @printindex cp |
| 3437 | |
| 3438 | @tex |
| 3439 | % I think something like @colophon should be in texinfo. In the |
| 3440 | % meantime: |
| 3441 | \long\def\colophon{\hbox to0pt{}\vfill |
| 3442 | \centerline{The body of this manual is set in} |
| 3443 | \centerline{\fontname\tenrm,} |
| 3444 | \centerline{with headings in {\bf\fontname\tenbf}} |
| 3445 | \centerline{and examples in {\tt\fontname\tentt}.} |
| 3446 | \centerline{{\it\fontname\tenit\/} and} |
| 3447 | \centerline{{\sl\fontname\tensl\/}} |
| 3448 | \centerline{are used for emphasis.}\vfill} |
| 3449 | \page\colophon |
| 3450 | % Blame: doc@cygnus.com, 28mar91. |
| 3451 | @end tex |
| 3452 | |
| 3453 | |
| 3454 | @contents |
| 3455 | @bye |
| 3456 | |
| 3457 | |